RE VIEW PA PER Analysis of Carotenoids by High Performance Liquid Chromatography and Diode-array Detection

Peter M. Bramley Department of Biochemistry. Royal Hollowey and Bcdford New Collcgc (Univcrsity of London), Egham Hill, Egham. Surrey TW20 OEX. U K

The use of diode-array detectors with high performance liquid chromatography enables the separation and analysis of carotenoids from biological samples to be achieved in a significantly shorter period than when conventional ultraviolet/visible detectors are used. This relates to the ability of diode-array detectors to provide full spectral analysis of the compounds during chromatography. The features of diode-array detectors, including their advantages and drawbacks, are discussed, particularly with respect to the identification and quantitation of carotenoids.

KcJywords: Carotenoids; high performance liquid chromatography: diode-array detection.

INTRODUCTION

The carotenoids represent one of the most important and widespread groups of pigments in Nature. The structures of well over S O 0 are known, and they are responsible for many of the yellow and red colours of flowers, fruits, birds. insects and other animals (Weedon, 1971; Goodwin, 1980, 1983, 1986; Schiedt, 1990). Their universal presence in photosynthetic tissue is on ly noticeable at the onset of leaf senescence when the chlorophylls disappear. The dramatic changes in the colour of ripening fruits also reflect the degradation of chlorophylls and a concomitant and massive increase in carotenoids.

Carotenoids play vital roles in photosynthetic tissues, acting as accessory light harvesting pigments (Jeffrey ei ul . , 1974; Larkum and Barrett, 1983; Moore et ul . , 1990), protectors against photoxidation (Knox and Dodge, 1985) and precursors of abscisic acid (Milbor- row, 1983). Their importance as dietary precursors of vitamin A (Goodwin, 198t)) as well as their increasing use as protectors against photosensitive skin disorders (Krinsky, 1989, 1990) and as anticancer agents (Mathews-Roth, 1989) has led to an increase in their commercial use, especially when derived from biologi- cal sources rather than produced synthetically (Taylor,

This upsurge of commercial interest, the on-going research into the enzymology and regulation of carote- noid biosynthesis and of their function, requires accur- ate techniques for the separation, identification and quantitation of these pigments from a variety of sources. Traditional chromatographic techniques of column and thin layer chromatography (TLC) are being replaced by high performance liquid chroma- tography (HPLC), and most recently the use of HPLC coupled to a diode-array detector. The purpose of this review is to explain the underlying principles of diode

0958-0344/92/030097-OX $09.00 0 1992 by John Wiley & Sons. Ltd.

1990).

array detection and to illustrate its advantages over conventional ultravioletfvisible (UViVIS) detectors, especially in the analysis of carotenoids.

STRUCTURE, NOMENCLATURE AND PROPERTIES OF CAROTENOIDS

Carotenoids are isoprenoids which generally contain eight isoprene units joined together so that the linking of the units is reversed at the centre of the molecule. The main feature of the symmetrical tetraterpenoid is a polyene chain which may extend from 3 to 1.5 conju- gated double bonds. Cyclization of the carbon skele- ton, at one o r both ends of the chromophore, occurs for example in (3-carotene (6,P-carotene; Fig. l ) , while xanthophylls are formed from the hydrocarbon caro- tenes by the introduction of oxygen functions, e.g. lutein (3,3’-dihydroxy-f!,~-carotene) and violaxanthin (5,6,5’ ,6‘ -diepoxy-S,6,5’ ,6‘ - tetrahydro- B$-carotene - 3,3’-diol; Fig. 1).

Common carotenoids have well established trivial names, and also IUPAC-recommended systematic nomenclature (Commission on Biochemical Nomenclature, 1971, 1975). Trivial names are used throughout this review; the corresponding semi- systematic names are quoted at the first mention of each carotenoid. The structures of the carotenoids can be found in Straub (1987), whilst regular reviews on advances in carotenoid chemistry are published by the Chemical Society of Great Britain (e.g. Britton, 1989). The structures of predominant plant carotenoids are shown in Fig. 1. Progress in the structural elucidation of plant carotenoids has been reviewed recently (Szabolcs, 1990).

The length of the carotenoid chromophore deter- mines the absorption spectrum of the molecule. Thus

Received 20 October 1991 Accepted freoised) 2 December f Y Y l

08 PETER M. BRAMLEY

I

OH

11 HO

OH

111 HO

OH

HO

OH

HO

Figure 1. Structures of carotenoids found in green leaves: ( I ) Ikarotene, (11) antheraxanthin, (Ill) lutein, (IV) neoxanthin, (V) violaxanthin and (VI) zeaxanthin.

phytoene (7,8,11,12,7’,8’, 11’,12’-octahydro-q~,~-caro- tene) has three conjugated double bonds, and absorbs in the UV at 285 nm, whilst b-carotene (11 conjugated double bonds) absorbs at 451 nm. Both the absorption wavelengths and spectral persistence are characteristics of carotenoids (e.g. Fig. 2 ) , and are frequently used to identify and quantify these compounds (Davies, 1976). The spectral properties are fully exploited in diode array detection. It should be noted, however, that carotenoids are unstable to heat, light and oxygen, and so care must be taken in the extraction and handling of samples throughout analysis. Detailed experimental procedures can be found elsewhere (Britton and Goodwin, 1971; Davies, 1976; Britton, 1985; Goodwin and Britton, 1988).

PRINCIPLES OF DIODE ARRAY DETECTORS

‘The diode- or photodiode-array detector was intro- duced i n the early 1980s by Hewlett-Packard. It is now available from several manufacurers (Table 1).

Although it shares many elements with a conventional UV/VIS detector, the essential difference is that it can record the entire spectral range (190-800 nm) during analysis, thus monitoring the chromatogram at selected wavelengths whilst recording the spectra of eluants simultaneously (Fig. 3 ) . Although conventional detec- tors are able to monitor at several wavelengths. they cannot record spectra simultaneously since they separ- ate the incident light with a diffraction grating before it reaches the sample.

In a diode-array detector, polychromatic light from the lamp source (usually deuterium) is transmitted through an aperture (typically 5 nm), lens and shutter directly onto the sample in a flow cell. This arrange- ment of optical components, with the flow cell posi- tioned before the grating, is often called reverse optics. The shutter closes when the measurement of dark current is required. Transmitted light is then focussed via a slit onto a diffraction grating which separates the light into a spectrum, which travels t o a n array of up t o 512 photodiodes (Fig. 4). The spacing of the diodes and the slit width determine the overall wavelength resolu- tion of the detector. The slit width also affects the intensity o f light reaching the array and the noise of the response. Decreasing the slit width increases optical resolution, but also increases noise.

The photodiode array is a series of light-sensitive elements on a silicon chip, and is normally fully charged. As each diode is struck by light of a given wavelength, it discharges. The amount o f current required to recharge the photodiode is a function of the quantity of light striking it. Both the dark current, i.e. the electrical current discharged from the array when no light is striking the photodiodes, and the reference current, the result of light transmitted by the solvent. are also measured. The reference energy spectrum is

350 400 450 500 550 Wavelength (nm)

Figure 2. Absorption spectra of (+-carotene (- - -) and lycopene (---I in hexane.

ANALYSIS OF C'AKOTENOIDS BY HPLC 99

Y

t

Table 1 . Features of selected diode array detectors Wavelength range Accuracy Number Minimum band

Manufacturerlmodel (nm) (nm) Lampa of diodes width inm)

Applied Biosystems 1000s 190-370; 370-555 1 .O D 76 5 Beckman 168 190-600 1 .o D 51 2 1 Hewlett-Packard 1040 190-600 1 .o D 21 1 4 Hitachi L-3000 200-520 5.0 D 70 10 LCD Analytical 5000 190-360 1 .o D 70 5 Merck-Hitachi L3000 200-360; 360-520 5.0 D 36 5 Phillips PU4121 190-590 1.5 D, T - Pharmacia-LKB 2140 190-370 0.9 D 200 4 Waters 990 190-800 1 .o D 512 1.4 a D=deuterium: T=tungsten.

I n m Absorbance

Chromatogram \ Wavelength

\ 2

Y

t Absorbance I

Figure 3. Chromatographic data obtainable with a diode-array detector. (Reproduced with permission of Millipore Corporation.)

used as a n indicator o f lamp intensity and instrument performance. The quantity of light at varying wavc- lengths is thus accurately and instantly measured after the diodes are recharged.

Holographic

grating piit t

\ S ' A L I ]

,Second order filter

Spectrum\

Figure 4. The optical assembly of a Waters 990 diode-array detector. A = aperture; L = lens; S = shutter. The bold, arrowed line indicates the light psth. (Reproduced with permission of Millipore Corporation.)

DATA COLLECTION AND ANALYSIS ~ ~~ ~ ~

Because of the enormous quantity of data generated during a typical chromatographic separation, diode- array detectors are coupled to computers with compre- hensive software for data analysis. Typically, the com- puter is programmed to scan the diodes (the sampling period) repeatedly during periods as short as 10ms. The entire array is recharged and read at the start of each sampling period as an analogue signal. Once the array is scanned, the diodes are allowed to discharge until the beginning of the next sampling period, when the scan is repeated. The process of recharging takes the same time, regardless of how much each diode was discharged, so that the total sampling time is the sum of the scan of the array and the waiting period whilst the diodes are recharging.

Scanning all wavelengths for the duration of one chromatogram and saving each data point requires a great deal of space on the hard disk of the computer. Therefore, all the data for each scan are not stored in memory. Several parameters can he optimized in order to minimize the size of files:

(a) Wurieleiigrh runge. Only specific wavelengths appropriate to the absorption spectra of the compounds need to be stored: for carotenoids a range of 250- 650 nm is sufficient.

(b) Meusirrrrnerif time urzd iiirerrd. Data need only he iiccumulated during the elution times of compound of interest. The interval defines how many data points are stored by setting the frequency at which data are saved: H smaller interval collects larger amounts of data. During periods of no interest in the separation, e.g. between peaks. the computer can be programmed not t o collect data.

(c) Acc~irrnir/ririoii. The data points from several scans can be averaged t o a single point. The accumu- lation parameter determines this number: increasing the value decreases the noise of the dispay.

(d) Hesolrrrion. For the identification of carotenoids by spectral analysis, the resolution of the diode array should be at ;I maximum. However. if absolute resolu- tion is not required, a larger value can be used to reduce disk space required for storage.

Post-run analyses. using new algorithms, provide extremely sophisticated methods of visualizing data, usually o n a colour monitor which can be connected to a colour printer. Isograms (contour plots. Fig. 5 ) and three-dimensional topographical plots (Fig. 6) are readily obtaincd. as well as composites of elution pro-

I00 PETER M. BRAMLEY

files and spectra of each component (Fig. 7). The data printout can be graphical or numerical. A spectral library feature enables comparison of experimental spectra with standards to be made, whilst integration, and hence quantitation, of elution peaks can be achieved once peak areas have been calibrated with standard amounts of solute.

ADVANTAGES AND DISADVANTAGES

Probably the greatest advantage of using a diode-array detector is that spectral and chromatographic profiles are recorded simultaneously (Fig. 3), thus eliminating the need to generate multiple chromatograms at differ- ent wavelengths and/or collecting fractions to record spectra later. This facility significantly reduces the time taken for analysis.

The data handling features enable peak priority to be determined quickly-a very useful feature in methods development and quality control. The high wavelength resolution of the detector enables identification of peaks to be achieved accurately, whilst contaminants can aIso be recognized. Spectra can be recorded on the up-slope, apex and down-slope of a chromatogram peak, normalized and overlaid, pinpointing differences in spectral shape and hence purity. Ratios of peak areas can also be calculated to determine purity. These features are often very useful in highlighting carotenoid

isomers which may co-elute, but have slightly different absorption spectra. Spectral libraries, stored on hard disk, make comparisons with standards easy, while baseline substructions enable comparisons to be made once a gradient blank has been stored.

A limitation of diode-array detectors is their sensitiv- ity in comparison to conventional UV/VIS detectors. This is most noticeable in the visible regions of the spectrum, as the deuterium lamp has relatively low spectral energy between 400 and 600nm. For this reason, some machines have been developed with both deuterium and tungsten lamps. However, the lamps cannot be used simultaneously, and require switching when needed. This is difficult during the chromatog- raphy of compounds whose spectra are unknown. In the UV region, instruments are sensitive as the best single or variable wavelength detectors and can be set at 0.001 AUFS. Generation of spectra at such levels of sensitivity is more problematical, however, due to signal-to-noise ratios, and may require a scale which is an order of magnitude less sensitive.

Other disadvantages relate to the complexity of the instruments themselves and of the software. The instru- ments do take longer to customize than conventional detectors, due to the large number of parameters to be set. Machines invariably require changes in parameter settings when required by another user. The time taken for alterations can be minimized by a mouse-driven system. The hard disk capacity is soon saturated, and so files must be deleted, or transferred to floppy disks or mainframe computers for storage purposes.

i.

3

1 5 2 'I 51

Figure 5. Contour plot of an HPLC profile, showing a three-dimensional plot of absorbance (AU) vs. wavelength (nm) vs. elution time (min).

ANALYSIS OF ('AROTENOIDS BY HPLC

we.., 31111 SlOl

1 AT( 1 4 7 1 A Q

Figure 6. Alternative representation of the three-dimensional plot of Fig. 5 showing absorbance (AU) vs. wavelength (nm) vs. elution time (min).

................................... Waters 990 Spectrum index plot ton17.fiT3 Y-scalr S l o p e Sanipl ing t i i u r Sense Re so I u t i on Tine raiigr I n t e l - y a l Baseline

4 scc OFF ~. . ........................................ - Chromato 2hO - - - ClOO nm(Max1

.. -.UO1 --- . 1 AU - - 0 .05 -

..................................

......................................... (peak) correct Water5

. 7 n

(I 7 30 2 1 m I n

300 400 600

Figure 7. Spectrum index plot showing the elution profile of the HPLC separation and the spectra of the eluting components of the sample represented in Figs. 5 and 6.

l o ? PETER M. BRAMLEY

Table 2, Examples of the separation of plant and algal carotenoids by HPLC

Algae 511 Spherisorb CN Gradient elution-heptane : ethyl acetate and

Barley 3p Silica Gradient elution-2-propanol and hexane Spinach lop Lichrosorb RP8 Methanol : acetonitrile Daffodil chromoplasts Spherisorb ASY-alumina Hexane, hexane : diathyl ether Tobacco 3p Spherisorb silica 2-propanol and hexane Fruit and vegetables Acetonitrile : methanol : tetrahydrofuran

Acetonitrile : methanol :ethyl acetate Paprika Zorbax C,a Acetone : methanol and acetone :water Flower petals 5p Supelcosill LC18-S Dichloromethane : 2-propanol and acetonitrile

Source support Moblle phase

dichloromethane

4p Novapak Cia Chromsil C18

Reference

Gillan and Johns (1983)

Britton (1991) Braumann and Grimrne (1981) Beyer et a/. (1985) Stuart etal. (1983) Bushway (1985) Daood etal. (1989) Fischer and Kocis (1987) Zonta etal. (1987)

APPLICATIONS OF HPLC DIODE-ARRAY DETECTION TO CAROTENOID ANALYSES

The separation of carotenoids by HPLC has now become the technique of choice in many laboratories, largely replacing TLC systems. Several reviews on the topic have been published in recent years (Taylor and Ikawa, 1980: Tayor, 1983; Ruddat and Will, 1985; Britton, 1985, 1989; Ruedi, 1985; Taylor et a[ . , 1990).

Both normal phase (adsorption) colums and reversed phase (partition) columns are used for the separation of carotenoids (Tables 2 and 3). Silica columns are very useful for HPLC separations, particularly of xantho- phylls but, because they are weakly adsorbed, caro- tenes are not well resolved on such columns. A particu- larly useful protocol for the separation of chloroplast carotenoids (and chlorophylls) has been reported by Britton (1991). Using a silica stationary phase and a gradient elution with hexane : propan-2-01 (4 : 1 v/v) and hexane, the pigments were resolved in some 3Smin, including the geometric isomers of the main leaf xan- thophylls. One major disadvantage of silica columns, however, is the tendency for chlorophylls to degrade and carotenoids to isomerize on the column. The degradation products can mask the presence of minor components of the mixture. Aluminium oxide columns have been used to separate the geometric isomers of 5- carotene and lycopene (Beyer et al., 1989).

An alternative normal phase material is a bonded nitrile phase such as Spherisorb SS-CN. The resolving

power of such columns has been well illustrated by Ruedi (1985) who has shown that structural isomers, geometric isomers, and 5,6- and S,K-epoxides can be separated on such columns. The best results have been obtained from a semi-purified mixture of xanthophylls rather than from crude extracts.

The most popular HPLC stationary phase used for the separation of carotenoids is a reversed phase parti- tion column with a C,, octadecylsilane (ODS) alkyl chain bonded to the silica. Although many columns have such bonded chains, the degree of end-capping and carbon loading varies between manufacturers, and so it is important to choose a column which gives efficient, reproducible separations. This often means the purchase of a column from the same supplier as that referred to in a standard literature protocol. Reversed phase columns not only give excellent resolution of carotenoids, but also there is virtually no danger of degradation or isomerization of the pigments on the column.

The order of elution from a reversed phase column is not, as might be predicted, the exact opposite of that from a normal phase system. For example, lutein is eluted before zeaxanthin in both methods. The reason is that, apart from the overall polarity of the carotenoid molecule, other physicochemical factors influence sep- arations. The strongest association with the C,, alkyl chain is that of an unsubstituted carotenoid end-group, so that a diol with both hydroxy groups in one end- group will be more strongly retained than one with a hydroxyl in each end-group. Carotenoids with a greater

Table 3. Example of the applications of HPLC diode-array detection in carotenoid analyses

Source of carotenoid

Carrot Radish Carrot and radish Carrot and barley Palm oil Green vegetables Orange juice Marigold flowers Soybean Fruits Food products Artemia Algae Brevibacterium, Rhodobacter and Rhodomicrobium spp.

Carotenoids analysed

a-$-Carotenes, retinoids Carotenoids, xanthophylls Carotenes, xanthophylls Chlorophyll, carotenoids a-,p-Carotenes, mono- and di-epoxides Chlorophyll, carotenoids tx-,(+-Carotenes, f3-cryptoxanthina Lutein and lutein esters Chlorophyll, carotenoids Carotenoids, carotenal esters Carotenoids, retinoids cis-Canthaxanthinsb Chlorophylls, phaeophytin, carotenoids

Carotenes, xanthophylls

Diode-array detector

HP 1090 HP 1040A HP 1040A HP 1040A HP 1090M HP 1040A HP 1040A Waters 990 HP 1090 HP 1040A HP 1040 HP 1040 HP 1040A

HP 1040A

Reference

Heinonen (1990) Britton eta/ . (1987) Barry and Pallett (1990) Young eta l . (1989) Ng and Tan (1988) Khachik etal. (1986) Fischer and Rouseff (1986) De las Rivas (1989) Mayonado etal. (1989). Khachik eta/ . (1989) Taylor et a/. (1990) Nelis etal. (1984) Mantoura and LleweHyn (1984)

Nelis and De Leenheer (1989) a (I-Cryptoxanthin : 3-hydroxy-(3,(3-carotene.

Canthaxanthin : 4,4'-diketo-fI-carotene.

ANALYSIS O F ( 'AROTENOILX BY HPLC 103

degree of saturation are usually more strongly retained. as are acyclic carotenoids in comparison to cyclic molecules.

A range of aqueous and non-aqueous solvent systems have been used for reversed phase HPLC o f cnroten- oids. Non-aqueous isocratic solvent systems include acetonitrile: dichloromethane : methanol (Nelis and DeLeenheer, 1983) and ethyl acetate : acetonitrile, con- taining 0.1 '% tz-decanol as modifier (Lauren et nl., 1986). The it-decanol is added to prevent the activation of silanol groups in the reversed phase packing, which leads to deterioration of the column. Most workers, however, use a solvent mixture containing water. A typical example is a linear gradient of ethyl acetate (O- 100'%)) in acetonitrile: water (9: 1 v /v) , containing 0 . 5 % triethylamine at a flow rate of 1 mL/min. Over a 25 niin period this gives a good separation of xanthophylls. carotenes, carotene epoxides and xanthophyll acyl esters in one chromatogram (Britton, 1991). Alternative mobile phases are listed in Table 2.

The resolution of optical isomers of carotenoids has been achieved by HPLC. The preparation o f diastereo- isomeric esters by reaction with optically activc ( - )-camphony1 chloride (Vecchi and Muller, 1979) or derivitization with (S)-( + )-u-( 1-naphthy1)ethyl isocyn- nate (Ruttimann et al. , 1983) allows subsequent sep- aration of the optical isomers. A chiral resolution column, Sumipax OA-2000, has been used for similar separations. after conversion of the isomers to the corresponding benzoates (Maoka and Matsuno. 1989).

Separations of carotenoids o n analytical columns can be scaled up t o preparative HPLC using self-packed axially compressed columns o f silica onto which up to 5 0 mg of carotenoids can be loaded and resolved (Isaken and Francis. 1989).

The use of on-line diode-array detectors has prob- ably been the most significant advance in the HPLC ot carotenoids over the past five years. The spectral properties of carotenoids are particulurly suited to diode-array detection. since their spectrn art' distinctive and this often enables the identification of individual c;i ro t e n o ids. Th c ah i I i t v to o b t ;I i 11 fu I I spec t r al ;I na I y sih and ;I print-out of eiich eluted carotenoid wi th in minutes o f completing the HPLC run has significantly decreased the time f o r iinalysis. I n addition. simult;i- neous q win t i t at ion of caro t e noid amounts . usi ng integration o f peak ;ire;is and ccmipnrison with ;I

calibration curve. is also of considerable value. The

increasing popularity of this procedure is shown by a number of recent reports o n its use (Table 3). In conjunction with an autosampler coupled to the HPLC, it is a technique which is now used successfully for the routine analysis of plant and food carotenoids. For example, the changes that occur in plant carotenoids following exposure to herbicides can be rapidly and accurately detected (Britton et al . , 1987; Young et ul., 1989; Barry and Pallett, 1990), whilst the analysis of carotenoids/retinoids in foods has been standardized using HPLC diode-array detection (Taylor er al.. 1990). Since carotenoids absorb light in both the UV and visible regions of the spectrum, a diode-array detector which has a spectral range including both these re&' lions must he used (Table 1).

While HPLC diode-array detectors are increasingly being shown to be invaluable in the analysis of caro- tenoids, it is important not to use the technique beyond its capabilities. Identification using only spectral max- ima can lead to inaccurate assignments, since the spec- tral accuracy of k 1 nm cannot, in all cases, distinguish between carotenoids with very similar absorption spec- tra. In addition, i t is well known that the absorption maxima of carotenoids do vary in different solvents (Davies. 1976), so that spectra taken on-line can be misleading. When isocratic solvent systems are used, standard spectra can be established and used as refer- ences, but this is more difficult in the case of gradient HPLC solvent systems. The problem of identification of unknown carotenoids has been addressed through the use of HPLC-mass spectrometry (Taylor et al . ,

With respect to the sensitivity of detection, care should be taken to minimize the computerized damping or smoothing of baselines during analysis, since this can lead to overestimation of the amounts detected. A recent. thorough investigation o f quantitation of caro- tenoids by HPLC diode-array detection has concluded that the working range for consistent reproducibility and accuracy is 50-200 ng per injection (Taylor et ul., 1990). This is especially true for biological samples which contain high background interference.

Despite these constraints, which can be overcome by careful quality control and standardization. HPLC diode-array detectors have enormously simplified the routine analysis of carotenoids from biological samples. I t is a technique which is likely to be used increasingly i n the future.

1990).

REFERENCES

Barry, P. and Pallet, K. E. (1990). Herbicidal inhibition of caro- tenogenesis detected by HPLC. Z. Nafurforsch. 45C. 492- 497.

Beyer, P., Weiss, G. and Kleinig, H. 11985). Solubilization and reconstitution of the membrane-bound carotenogenic enzymes from daffodil chromoplasts. Eur. J. Biochem. 153,

Beyer, P., Mayer, M. and Kleinig, H. (1989). Molecular oxygen and the state of geometric isomerism of intermediates are essential in the carotene desaturation and cyclization reac- tions in daffodil chromoplasts. Eur. J. Biochem. 184, 141- 150.

Braumann, T. and Grimme, L. H. (1981). Reversed-phase HPLC of chlorophylls and carotenoids. Biochim. Biophys. Acta

Britton, G. (1985). General carotenoid methods. Methods

341 -346.

637, 8-17.

En~ymol. 111, 113-147.

Britton, G. (1989). Carotenoids and polyterpenoids. Naf. Prod. Rep. 6, 359-392.

Britton, G. (1991). Carotenoids. In Methods in Plant Biochemistry (Charlwood, 6. V. and Banthorpe, D. V., eds.), Vol. 7, pp. 473-518. Academic Press, London.

Britton, G. and Goodwin, T. W. (1971). Biosynthesis of caro- tenoids. Methods Enzymol. 18, 654-701,

Britton, G., Barry, P. and Young, A. J. (1987). The mode of action of diflufenican: its evaluation by HPLC. Proc. Brit. Crop Protection Conf.-Weeds 9B-5, 1015-1022.

Bushway, R. J. (1985). Separation of carotenoids in fruits and vegetables by HPLC. J. Liq. Chromatogr. 8, 1527-1547.

Commission on Biochemical Nomenclature (1971). Tentative rules for the nomenclature of carotenoids. Biochemistry 10,

Commission on Biochemical Nomenclature (1975). Nomenclature of carotenoids. Biochemistry 14, 1803- 1804.

4827-4837.

I04 PETER M.

Daood, H. G., Czinkotai, B., Hoschke, A. and Biacs, P. (1989). High-performance liquid chromatography of chlorophylls and carotenoids from vegetables. J. Chromatogr. 472, 296- 302.

Davies, B. H. (1976). Carotenoids. In Chemistry and Biochemistry of Plant Pigments (Goodwin, T. W., ed.), 2nd edn., pp. 38-162. Academic Press, London.

De las Rivas, J. (1989). Reversed-phase HPLC separation of lutein fatty acid esters from marigold flower petal powders. J. Chromatogr. 464, 442-447.

Fisher, C. and Kocis, J. A. (1987). Separation of paprika pigments by HPLC. J. Agric. Food Chem. 35, 55-57.

Fischer, J. F. and Rouseff, R. L. (1986). Solid-phase extraction and HPLC determination of (3-cryptoxanthin and a-and+ carotene in orange juice. J. Agric. Food Chem. 34,985-989.

Gillan, F. T. and Johns, R. B. (1983). Normal-phase HPLC analysis of microbial carotenoids and neutral lipids. J. Chromatogr. Sci. 21, 34-38.

Goodwin, T. W. (1980). The Biochemistry of Carotenoids, Vol. 1. Chapman and Hall, London.

Goodwin, T. W. (1983). The Biochemistry of Carotenoids, Vol. 2. Chapman and Hall, London.

Goodwin, T. W. (1986). Metabolism, nutrition and function of carotenoids. Ann. Rev. Nutr. 6, 273-298.

Goodwin, T. W. and Britton, G. (1988). Distribution and analysis of carotenoids. In Plant Pigments (Goodwin, T. W., ed.), pp. 61-131. Academic Press, London.

Heinonen, M. I. (1990). Carotenoids and provitamin A activity of carrot (Daucus carota L.) cultivars. J. Agric. Food Chem. 38,

Isaken, M. and Francis, G. W. (1989). Preparative HPLC of carotenoids. Chromatographia 27, 325-327.

Jeffrey, S. W., Douce, R. and Benson, A. A. (1974). Carotenoid transformations in the chloroplast envelope. Proc. Natl. Acad. Sci. 71, 807-810.

Khachik, F., Beecher, G. R. and Whittaker, N. F. (1986). Separation, identification and quantification of the major carotenoid and chlorophyll constituents in extracts of several green vegetables by liquid chromatogrphy. J. Agric. Food Chem. 34, 603-61 6.

Khachik, F., Beecher, G. R. and Lasby, W. R. (1989). Separation, identification and quantification of the major carotenoids in extracts of apricots, peaches, cantaloupe and pink grapefruit by liquid chromatography. J. Agric. Food Chem. 37, 1465- 1473.

Knox, J. P. and Dodge, A. D. (1985). Singlet oxygen and plants. Phytochemistry 24, 889-896.

Krinsky, N. I. (1989). (3-Carotene: functions. In New Protective Roles for Selected Nutrients, pp. 1-15. A. R. Liss Inc., New York.

Krinsky, N. I. (1990). Carotenoids in medicine. In Carotenoids; Chemistry and Biology (Krinsky, N. I . , Mathews-Roth, M. M. and Taylor, R. F., eds.), pp. 279-291. Plenum Press, New York.

Larkum, A. W. D. and Barrett, J. (1983). Light-harvesting pro- cesses in algae. Adv. Bot. Res. 10, 1-219.

Lauren, D. R., Agnew, M. P. and McNaughton, 0. E. (1986). The use of decanol for improving chromatographic stability in isocratic non-aqueous reversed phase analysis of caro- tenoids by HPLC. J. Liq. Chromatogr. 9, 1997-2012.

Mantoura, R. F. C. and Llewellyn, C. A. (1984). Trace enrichment of marine algal pigments for use with HPLC-diode-array spectroscopy. J. High Res. Chromatogr., Chromatogr. Commun. 7, 632-635.

Maoka, T. and Matsuno, T. (1989). Diastereomeric resolution of carotenoids. 111. \3/3-Caroten-2-ol, [3,\3-carotene-2,2’-diol and 2-hydroxyechinenone. J. Chromatogr. 478, 379-386.

Mathews-Roth, M. M. (1989). [Xarotene: clinical aspects. In New Protective Roles for Selected Nutrients, pp. 17-38. A. R. Liss Inc., New York.

Mayonado, D. J., Hatzios, K. K., Orcutt, D. M. and Wilson, H. P. (1989). Evaluation of the mechanism of action of the bleach- ing herbicide SC-0051 by HPLC analysis. Pestic. Biochem.

609-612.

BRAMLEY

Physiol. 35, 138-145. Milborrow, B. V. (1983). Pathways to and from abscisic acid. In

Abscisic Acid (Addicott, F. T., ed.), pp. 79-86. Prager Scientific, New York.

Moore, T. A., Gust, D. and Moore, A. L. (1990). The function of carotenoid pigments in photosynthesis and their possible involvement in the evolution of higher plants. In Carotenoids: Chemistry and Biology (Krinsky, N. I., Mathews-Roth, M. M. and Taylor, R. F., eds.), pp. 223-228. Plenum Press, New York.

Nelis, H. J. C. F. and DeLeenheer, A. P. (1983). lsocratic non- aqueous reversed-phase liquid chromatography of carotenoids. Anal. Chem. 55, 270-275.

Nelis, H. J. C. F. and De Leenheer, A. P. (1989). Profiling and quantitation of bacterial carotenoids by liquid chromatogra- phy and photodiode-array detection. Appl. Environ.

Nelis, H. J. C. F., Lavens, P., Moens, L., Sorgeloos, P., Jonckheere. J. A., Criel, G. R. and De Leenheer, A. P. (1984). Cis-canthaxanthins. Unusual carotenoids in the eggs and the reproductive system of female brine shrimp Artemia. J. Biol. Chem. 259, 6063-6066.

Ng, J. H. and Tan, B. (1988). Analysis of palm oil carotenoids by HPLC with diode-array detection. J. Chromatogr. Sci. 26,

Ruddat, M. and Will, 0. H. (1985). HPLC of carotenoids. Methods Enzymol. 11 1, 189-200.

Ruedi, P. (1985). HPLC-a powerful tool in carotenoid research. Pure Appl. Chem. 57, 793-800.

Ruttimann, A., Schiedt, K. and Vecchi, M. (1983). Separation of (3R,3’R)-, (BR.3’S;meso)- and (3S,3’S)-zeaxanthin, (3R,3’R,6‘R)-, (3R,3‘S,6‘S)- and (3Sr3’S,6’S)-lutein via the dicarbamates of (S)- ( + )-a-(1-naphthy1)ethyI isocyanate. J. High Res. Chromatogr., Chromatogr. Commun. 6, 61 2-61 6.

Schiedt, K. (1990). New aspects of carotenoid metabolism in animals. In Carotenoids: Chemistry and Biology (Krinsky, N. I., Mathews-Roth, M. M. and Taylor, R. F., eds.), pp. 247- 268. Plenum Press, New York.

Straub, 0. (1987). Key to Carotenoids. Birkhauser-Verlag, Easel. Stuart, P. S., Bailey, P. A. and Bevan, J. L. (1983). HPLC

separation of carotenoids in tobacco and their characteriza- tion with the aid of a microcomputer. J. Chromatogr. 282,

Szabolcs, J. (1990). Plant carotenoids. In Carotenoids: Chemistry and Biology (Krinsky, N. I.,’Mathews-Roth, M. M. and Taylor, R. F., eds.), pp. 39-58. Plenum Press, New York.

Taylor, R. F. (1983). Chromatography of carotenoids and retin- oids. In Advances in Chromatography (Giddings, J. C., Grushka, E., Cazes, J. and Brown, P. R., eds.), Vol. 22, pp. 157-213. Marcel Dekker, New York.

Taylor, R. F. (19901. Carotenoids: Products, applications and markets. In Spectrum Food lndustry, pp. 12.1-12.11. Decision Resources, Massachusetts.

Taylor, R. F. and Ikawa, M. (1980). Gas chromatography, gas chromatography-mass spectrometry and HPLC of caroten- oids and retinoids. Methods Enzymol. 67, 233-261.

Taylor, R. F., Farrow, P. E., Yeale, L. M., Harris, J. C. and Marenchic, I. G. (1990). Advances in HPLC and HPLC-MS of carotenoids and retinoids. In Carotenoids: Chemistry and Biology (Krinsky, N. I., Mathews-Roth, M. M. and Taylor, R. F., eds.), pp. 105-123. Plenum Press, New York.

Vecchi, M. and Muller, R. K. (1979). Separation of (3S,3’S)-, (3R,3‘R)- and (3S,3’R)-astaxanthin via ( - )-carnphanic acid esters. J. High Res. Chromatogr., Chromatogr. Commun. 2, 195.

Weedon, B. C. L. (1971). In Carotenoids (lsler, O., ed.), pp. 29-59. Birkhauser-Verlag, Easel.

Young, A. J., Britton, G. and Musker, D. (1989). A rapid method for the analysis of the mode of action of bleaching herbi- cides. Pestic. Biochem. Physiol. 35, 244-250.

Zonta, F., Stancher, B. and Marletta, G. P. (1987). Simultaneous HPLC analysis of free carotenoids and carotenoid esters. J. Chromatogr. 403, 207-215.

Microbial. 55, 3065-3071.

463-469.

589-594.