Ann, Bot. 40,615-624, 1976

Carotenoid Pigment Changes in Ripening Momordica charantia Fruits* DELIA B. RODRIGUEZ, 1 L. C. RAYMUNDO,2 TUNG-CHING LEE,1 1 1 K. L. SIMPSON and C. O. CHICHESTER 1

2

Department of Food and Resource Chemistry, University of Rhode Island, Kingston, Rhode Island02881, U.S.A. Department of Food Science & Technology, University of the Philippines, Los Bonos Units, College, Laguna Received: 7 March 1975

INTRODUCTION

Numerous investigations have been devoted to the characterization and quantitative measurement of the carotenoids of ripe fruits. Studies on the carotenoid changes during development and ripening have been comparatively limited. Some workers have dealt solely with the gross change in carotenoid content (Gortner, 1965; Miller, Winston and Schomer, 1941; Reid, Lee Pratt and Chichester, 1970; Stamberg, 1945). The compositional changes of the individual carotenoids with maturation and ripening have been increasingly explored in recent years. Aside from the tomato, which has been the most commonly studied fruit (Edward and Reuter, 1967; Goodwin and Jamikorn, 1952; Meredith and Purcell, 1966; Raymundo, Griffiths and Simpson, 1967, 1970; Zscheile, 1966), pepper (Cholnoky, Gyorgyfy, Nagy and Panczel, 1956; Davies, Matthews and Kirk, 1970), lemon (Yokoyama and Vandercook, 1967), mango (John, Subbarayan and Cama, 1970), rowan berries (Valadon and Mummery, 1972), peaches and apricots (Katayama, Nakayama, Lee and Chichester, 1971) have now been examined at two or more stages of maturity. The tropical fruit Momordica charantia, balsam pear or bitter melon, may constitute a most instructive model for the study of carotenogenesis as a function of ripening. Unlike most fruits which contain colourless or pale-coloured seeds, Momordica charantia has intensely red seeds which contrast strikingly with the orange pericarp. When fully • Contribution No. 1569, Agricultural Experiment Station, University of Rhode Island, Kingston, R.I.

Downloaded from http://aob.oxfordjournals.org/ at UQ Library on September 6, 2016

ABSTRACT The carotenoid composition of Momordica charantiafruit(pericarp) at four levels of maturity was extensively investigated. The number of carotenoids isolated increased from five in the immature fruit to six at the mature-green and 14 at the partly-ripe and ripe stages. Cryptoxanthin, which could not be isolated at the immature and mature stages, accumulated rapidly at the onset of ripening to become the principal pigment of the ripe fruit. Moderate increases were seen in /J-carotene, zeaxanthin and lycopene concentrations as ripening progressed. The reverse trend was observed with lutein and a-carotene which were the major pigments of the immature fruit. Prior to the colour break, only the hydroxy derivatives of a-carotene (zeinoxanthin and lutein) could be detected; the ^-hydroxy compounds (cryptoxanthin and zeaxanthin) appeared and predominated thereafter. The hydroxy carotenoids of the ripe fruit were almost entirely esterified in contrast to those of the unripe fruit which were mainly uncsterified. Traces of flavochrome, 5,6-monoepoxy-^-carotene, mutatochrome, C-carotene, <5-carotene, y-carotene and rubixanthin were detected in the partly-ripe and ripe fruits but not in the immature and mature-green samples. Phytofiuene was observed in trace levels at all stages.

616

Rodriguez, Raymundo, Lee, Simpson and Chichester

ripe, the fruit (pericarp) splits into several valves revealing the red seeds whose colour is probably nature's device to attract birds, the agents of seed dispersal. In Asian countries such as the Philippines, Thailand and China, the mature-green fruit is a popular vegetable preferred for its distinctive bitterness. The only reported investigation on the carotenoids of Momordica charantia was made in 1913 by Duggar who detected lycopene and ^-carotene (Goodwin and Goad, 1970). Obviously, the carotenoid composition should be more complex than this early analysis revealed. We have re-examined the carotenoid pattern at different stages of development and found that the fruit has indeed a complex and rather unusual carotenoid biosynthetic pattern. MATERIALS AND METHODS

Extraction and separation of carotenoids Analysis of the sample was started immediately after harvest. The seeds were removed and analysed separately. The results of this concurrent study is presented in a separate paper. The fruit (pericarp) was weighed, blended with cooled acetone and light petroleum (1:1) in a Waring blendor and exhaustively extracted with acetone. The carotenoids were transferred to light petroleum by the addition of water, washed free of acetone and saponified overnight at room temperature with 10 per cent methanolic KOH. The alkali wasremovedby thorough washing; the pigment solution was dried with sodium sulphate and concentrated in vacuo. Preparative chromatography was accomplished on a MgO: HyfloSupercel (1:2) column developed successively with 100 ml each of 1, 2 and 5 per cent ethyl ether and 2, 5 and 8 per cent acetone in light petroleum. Sufficient 10 per cent acetone in light petroleum (c. 150 ml) was then used to elute fraction 3 from the column and to resolve fractions 5 and 6 of the ripe and partly ripe samples. The column was allowed to run dry; the remaining bands were cut and eluted with acetone or a combination of acetone and methanol for the more polar pigments. The eluates were transferred to light petroleum and concentrated for further chromatography. The pigments from the ripe and partially-ripe fruits formed seven fractions on the MgO: HyfloSupercel column. The seventh fraction occasionally separated into two or three bands; these bands were shown to consist of the same pigments and were thus taken as one. Fractions 2 to 7 werere-chromatographedon smaller columns of neutral alumina (activity grade III) using 1-5 per cent ether in light petroleum for development. Flavo-

Downloaded from http://aob.oxfordjournals.org/ at UQ Library on September 6, 2016

Materials Momordica charantia was grown in the greenhouse of the University of Rhode Island. The fruits were harvested at four stages of maturity: (1) immature, light green fruits; (2) mature, dark green fruits; (3) partially ripe fruits, partly green and partly orange in colour; and (4) fully ripe, uniformly orange fruits. Fruit maturity was judged visually. The fruits were collected over a period of 2 months. Authentic samples of ^-carotene, zeaxanthin, lutein and isocryptoxanthin were provided by F. Hoffman-La Roche Company. Other standard carotenoids were extracted and purified from tomatoes (lycopene, a-carotene, (-carotene, /?-zeacarotene, y-carotene, phytofluene), carrots (a-carotene, <5-carotene, y-carotene, (-carotene, phytofluene) and lemon (cryptoxanthin). Isozeaxanthin was obtained by the sodium borohydride reduction of canthaxanthin from Hoffman-La Roche Company. Analytical grade solvents and reagents were used. Petroleum ether and acetone were distilled prior to use. Ethyl ether was purified on a column of basic alumina.

CarotenoidPigments in Momordica charantia Fruits

6l7

Identification of the carotenoids The pigments were identified by the following parameters: UV or visible absorption spectra, position on the column, TLC Rf values, co-chromatography with authentic carotenoids and chemical reactions. The absorption spectra were determined in a Cary-14 recording spectrophotometer. The Rf values were the averages of several runs. Precoated silica gel sheets (Eastman Chromagram 13179, formerly 6060) and the pre-coated aluminium oxide plates described earlier were employed for identification purposes. The solvent systems for the development of the silica gel sheets were 0-5 per cent methanol in light petroleum, 3 per cent methanol in benzene and 20 per cent ethyl acetate in methylene chloride. The aluminium oxide plates were developed with 1 per cent acetone in light petroleum and 5 per cent ether in light petroleum. Acetylation with acetic anhydride, methylation with acidified methanol and dehydration with acidic chloroform were accomplished as described elsewhere (Rodriguez, Simpson, and Chichester, 1973). The possible existence of epoxy groups was tested by exposing the chromatograms to HC1 gas for a few minutes (Gross, Gabai, and Lifshitz, 1971) and by adding a few drops of 0-1 N H O to an ethanolic solution of the pigment and recording the absorption spectrum after 3 min. Iodine catalysed cis-trans isomerization was undertaken by dissolving a few crystals of iodine in light petroleum and adding a drop of this solution to the pigment dissolved in light petroleum. The spectrum was taken after 1-5 min exposure to light.

Quantitative determination The concentration of the various carotenoids was determined spectrophotometrically as described by Davies (1965). At least five fruits were analysed individually for each maturity stage. The presence and identification of the trace compounds were confirmed in larger batches of samples. The extinction coefficient of a-carotene was used in larger batches of samples. The extinction coefficient of a-carotene was used for lutein; hence the values for lutein could be considered only as approximations.

Downloaded from http://aob.oxfordjournals.org/ at UQ Library on September 6, 2016

chrome, 5,6-monoepoxy-/J-carotene, mutatochrome, C-carotene, <5-carotene and ycarotene were purified on precoated thin layers of aluminum oxide (E. Merckag. Darmstadt TLC plates type E, F 234 ) with 5 per cent ether or 1 per cent acetone in light petroleum as solvent system. Phytofluene was purified on another column of MgO: HyfloSupercel with 1-2 per cent ether in light petroleum as solvent. The upper band formed when fraction 7 was rechromatographed on an alumina column was further separated on precoated silica gel plates (Quanta Gram TLC plates QIF) developed with 20 per cent ethyl acetate in methylene chloride. The pigments coming from the mature and immature samples separated into five fractions on the MgO: HyfloSupercel column. The five fractions of the immature fruits were shown to consist of single carotenoids; hence further chromatography was not required. The first four fractions of the mature fruits also contained single carotenoids and therefore needed no re-chromatography. The fifth fraction contained lutein and traces of lycopene. These carotenoids were resolved on an alumina column developed with 1-2 per cent ether in light petroleum. The colourless eluate preceding phytofluene in all sample's was collected and tested for the possible presence of phytoene using the method described for phytoene purification by Raymundo et al. (1967).

618

Rodriguez, Raymundo, Lee, Simpson and Chichesier RESULTS

Identification of the carotenoids of ripe and partly ripe fruits

Downloaded from http://aob.oxfordjournals.org/ at UQ Library on September 6, 2016

Fifteen carotenoids were detected in the ripe and partially ripe pericarps. The chromatographic, spectral and chemical properties of these pigments are summarized in Table 1. Fraction 1 exhibited the characteristic green fluorescence and absorption spectrum of phytofluene. It was identical to phytofluene isolated from carrots and tomatoes. Re-chromatography of fraction 2 on the alumina column yielded two bands. The lower yellow band (fraction 2a) showed the same visible absorption spectrum and co-chromatographed with a-carotene isolated from tomatoes and carrots. A small hypsochromic shift was seen on iodine isomerization, confirming that it was not a cis form of ^-carotene. The absorption maxima in light petroleum of the upper yellow band (fraction 2b) were at 399, 421 and 448 nm. Exposure of the chromatograms to HCl gas turned the yellow spot to intense blue. The absorbance remained unaltered on addition of dilute HCl to an ethanolic solution of the pigment. Although the Rf values on the aluminum oxide plates were comparable to mutatochrome, it ran as a spot distinct from mutatochrome on the silica gel sheets. Iodine catalysed isomerization shifted the maxima to shorter wavelengths, showing that it was not a cis isomer of mutatochrome. It was therefore tentatively identified as flavochrome (5,8-epoxy-a-carotene). Fraction 3 was resolved into three bands: a main orange and two much fainter yellow bands. The lower orange band (fraction 3a) exhibited the same absorption curve and ran as a single spot with authentic ^-carotene on all the thin layers employed. The maximum absorbances of the middle, faint yellow band (fraction 3b) in light petroleum were at 421, 441 and 470 nm. Iodine isomerization, which resulted in a hypsochromic shift, showed the pigment was in trans form. Its epoxy nature was first revealed by exposure of the chromatograms to HCl gas which changed the yellow colour to blue-green. Addition of 0-1 N HCl to the pigment dissolved in ethanol readily shifted the maxima to shorter wavelengths by 20 nm, consistent with the presence of a single 5,6-epoxy group. The resultant spectrum resembled that of mutatochrome. Fraction 3b was thus identified as 5,6-monoepoxy-yS-carotene. The topmost faint yellow band (fraction 3c) absorbed maximally at 403, 424 and 449 nm in light petroleum. The pigment, which appeared as a yellow spot on the TLC, turned blue when exposed to HCl gas. No change in absorbance was seen on addition of dilute HCl to an ethanolic solution of the pigment. It co-chromatographed with the product formed from fraction 3b treated with dilute HCl. Fraction 3c was clearly mutatochrome (5,8-epoxy-/?-carotene). The almost colourless interzone between fractions 3 and 5 was extracted and designated as fraction 4. The purified main constituent of this fraction showed the visible absorption spectrum typical of (-carotene. It co-chromatographed with (-carotene purified from tomatoes and carrots on the silica gel sheets and aluminium oxide plates developed with different solvent systems. Two bands were formed on rechromatography of fraction 5 on the alumina column. The lower faint band (fraction 5a) absorbed maximally at 430, 456 and 485 nm in light petroleum. It was inseparable from <5-carotene isolated from carrots on all the thin layer systems, while being distinctly separable from y-carotene. The possibility that it could b e a r a isomer of y-carotene was eliminated when isomerization with iodine shifted the maxima to shorter wavelengths. The absorption curve of the more polar, yellow pigment (fraction 5b) resembled that of a-carotene but the polarity was indicative of a monohydroxy compound. The pigment was readily acetylated, producing a monoacetate with an Rt of 0-93 on the silica gel sheets developed with 3 per cent methanol in benzene. Treatment with acidic chloroform repeatedly failed to dehydrate the pigment, showing that the hydroxyl function was not in allylic position to the polyene chain and thus ruling out 4-hydroxy-a-carotene as the structure. Repeated attempts to methylate the pigment

435, 458,488

(425), 448, 476

a-Carotene

Flavochrome* ^-Carotene

5,6-Monoepoxy^-Carotenc Mutatochrome C-Carotene

<5-Carotene Zeinoxanthin

y-Carotene

Cryptoxanthin

Lycopene Rubixanthin

Lutein

Zeaxanthin

2a

2b 3a

3b

5a 5b

6a

6b

7a 7bl

7b2

7b3

(423), 445,473

419,442,471

443,469,501 435,457, 486

430,456,485 421,443,472

403,423, 449 377, 399,424

421,441,470

002

0-92 0-31 0-88

0-66 0-32 0-92 0-82 000

064

0-59

0-44

0-84 0-40 0-21

019

0

0-06 0 0

0

0-32 0

0

0-21 0 0

0

0-94

0-20

0-67

0-79

0-82

0-94 0-56

0-23 000

0-69 000

0-72

0-87 0-95 0-84

5

0-04 0-35

0-95

0-93 0-97

0-17 0-56

0-32 0-93

399,421,448 (425), 448,476

0-31 0-92

0-93

421,443,472

0-96

4

0-67

Silica gel 3

+ furanoid test Co-chromatography {-carotene Co-chromatography + acetylation — methylation — dehydration Co-chromatography >>-carotene Co-chromatography + acetylation — methylation — dehydration Co-chromatography + acetylation — methylation — dehydration Co-chromatography lutein + acetylation + methylation — dehydration Co-chromatography zeaxanthin + acetylation — methylation — dehydration

with Hoffman-La Roche

with Hoffman-La Roche

with tomato lycopene

with lemon cryptoxanthin

with carrot and tomato

with carrot (5-carotcnc

with carrot and tomato

Co-chromatography with carrot and tomato phytofluene Co-chromatography with carrot and tomato a-carotene + furanoid test Co-chromatography with Hoffman-La Roche ^-carotene + epoxide test

Other criteria for identification

* Tentative identification. The TLC plates were developed as follows: (1) 5 per cent EtjO in petroleum ether; (2) 1 per cent MejCO in petroleum ether; (3) 0-5 per cent MeOH in petroleum ether; (4) 3 per cent MeOH in QsH*; (5) 20 per cent EtOAc.

3c 4

330, 345, 365

Phytofluene

1 0-92

Absorption Rf values spectra in light petroleum ether (nm) Alum, oxide 1 2

1. Carotenoids ofripe and partly ripe Momordica charantia fruits

Fraction Identification

TABLE

oaded from http://aob.oxfordjournals.org/ at UQ Library on September 6, 2016

620

Rodriguez, Raymundo, Lee, Simpson and Chichester

Downloaded from http://aob.oxfordjournals.org/ at UQ Library on September 6, 2016

with acidified methanol also failed, excluding the allylic 3'-position of the a-ring as the possible location of the hydroxyl substituent. Fraction 5b was therefore identified as zeinoxanthin (3-hydroxy-a-carotene). The pigment occurred naturally as an ester; the Rf of the unsaponified form being 0-98 on the silica gel sheets developed with 3 per cent methanol in benzene. Further separation on the alumina column resolved fraction 6 into two zones: a faint orange band (fraction 6a) and an intense red-orange band (fraction 6b). The absorption spectrum of fraction 6a was identical to that of y-carotene from tomatoes and carrots. It was indistinguishable from authentic y-carotene on the thin layers. Fraction 6b in light petroleum absorbed maximally at 448 and 476 nm with a shoulder at 425 nm. The Rf on the silica gel sheets developed with 3 per cent methanol in benzene was 0-97 before saponification; the value was reduced to 0-44 after saponification. This was reflective of an esterified, monohydroxy carotenoid. Acetylation was easily accomplished, producing a single compound with an Rf of 0-92 on the silica gel sheets with 3 per cent methanol in benzene as solvent. Methylation and dehydration reactions were negative, eliminating the 4- or 4'-position as the location of the hydroxy substituent. Finally, fraction 6b ran as a single spot with cryptoxanthin purified from lemon and was clearly separated from isocryptoxanthin. It was therefore unambiguously identified as cryptoxanthin (3-hydroxy/?-carotene) present in the fruit as an ester. Traces of unesterified cryptoxanthin could be detected on the chromatograms of the unsaponified samples. The topmost fraction on the MgO:HyfloSupercel column separated into a yelloworange band (fraction 7b) and a lowerreddishband (fraction 7a) on the alumina column. The visible absorption spectrum of fraction 7a consisted of three peaks with the maxima at 435, 469 and 501 nm in light petroleum. It was inseparable from tomato lycopene on the different thin layer systems. Fraction 7b was subjected to further separation on a silica gel plate developed with 20 per cent ethyl acetate in methylene chloride, forming a red-orange band at the solvent front and two closely associated yellow bands at the midpoint of the chromatogram. The red-orange band (fraction 7b 1) was identified as rubixanthin (3-hydroxy-y-carotene). The absorption spectrum was identical to y-carotene but the polarity was that of a monohydroxy compound. Acetylation produced a monoacetate with an Rf of 0-90 on the silica gel sheets developed with 3 per cent methanol in benzene, thus establishing the presence of a single primary or secondary hydroxyl group. The pigment could not be dehydrated with acidic chloroform, discounting the possibility that the hydroxyl group could be located at the 4- or 4'-position. The negative response to methylation was consistent with the non-allylic nature of the hydroxyl function. The absorption maxima of the second band suggested an a-carotene chromophore; the adsorption affinity was that of a dihydroxy carotenoid. The possibility that it could be a cis isomer of zeaxanthin was first eliminated by iodine isomerization which shifted the maxima to lower wavelengths. Two products were formed on acetylation with Rf values of 0-61 (monoacetate) and 0-83 (diacetate) on the silica gel sheets with 3 per cent methanol in benzene as solvent. Methylation produced a single compound with an Rf of 0-48; thus one of the hydroxyl groups was in allylic position. The pigment could not be dehydrated, however, showing that the allylic hydroxyl substituent was not in conjugation with the polyene chain and was therefore located at the 3'-position. Fraction 7b 2 also co-chromatographed with authentic lutein while being clearly separated from authentic zeaxanthin on the silica gel sheets developed with 20 per cent ethyl acetate in methylene chloride. This fraction was therefore identified unequivocally as lutein (3,3'-dihydroxy-acarotene). It existed as an ester in the fruit; the Rf of the unsaponified pigment was 0-97 on the silica gel sheets developed with 3 per cent methanol in benzene. The slightly lower yellow band (fraction 7b 3) absorbed maximally at 445 and 473 nm with an inflection at 423 nm. It co-chromatographed with authentic zeaxanthin but not with isozeaxanthin. It was also separable from authentic lutein on the silica gel sheets developed with 20

Carotenoid Pigments in Momordica charantia Fruits

621 per cent ethyl acetate in methylene chloride. The presence of the two hydroxyl groups was confirmed with acetylation which produced two compounds with Rf values of 0-61 and 0.83 on the silica gel sheets with 3 per cent methanol in benzene as solvent. Failure to undergo both methylation and dehydration showed that both hydroxyl groups were not allylic. These properties were entirely consistent with the identification of fraction 7b 3 as zeaxanthin (3,3'-dihydroxy-/?-carotene). As with the other hydroxy carotenoids, zeaxanthin occurred naturally as an ester. The Rf of the unsaponified pigment was 0-97 on the silica gel sheets. Traces of free zeaxanthin were seen on the thin layer chromatograms of the unsaponified sample. Identification of the carotenoids of immature and mature fruits

Quantitative changes during ripening

The quantitative distribution of carotenoids at four levels of maturity is given in Table 2. Since the maturity of the fruits was not rigorously determined (i.e. counting the days after petal fall) and because of the inherent variability of fruit samples, even when the chronological age is known, there were variations in the concentration of the individual carotenoids. Definite patterns and trends could still be confidently deduced from the results. The total carotenoid content rose from an average of 6-3 /ig/g in the immature fruit to an average of 22 /ig/g in the ripe fruit on a fresh weight basis. The corresponding values TABLE

2. Quantitative distribution of carotenoids in developing Momordica charantia fruits

Carotenoid

Phytofluene a-Carotene Flavochrome /?-Carotene 5,6-MonocpoxyyS-Carotene Mutatochrome {-Carotene
Immature-green

Mature-green

Partly ripe

G^gg)-1*

(%)

fogg)"1*

Oigg)"1*

34-9

Trace 1-5 ±0-3

Trace 2-2 ±1-1 l-5±O8 —

238

— —

2-3 ±0-8 —

(%) 28-3 43-4

— —

0-2 ±0.1 — — —

3-2

2-4 ±1-2 — 6-3

381

0-5 ± 0-3 — — Trace

9-4

1-0 ±0-4 — 5-3

18-9

Trace 0-3 ± 0 1 Trace 2-5 ± 0-3 Trace Trace Trace Trace 0-3 ±0-1 Trace 4-7 ± 1 -7 0-2 ± 0 1 Trace 0-1 0-4 8-5

Fully ripe (%) 3-5 29-4

3-5 55-3 2-4 1-2 4-7

C^gg)-1* Trace 0-4 ±0-1 Trace 5-3 ±2-3 Trace Trace Trace Trace 0-6 ± 0-2 Trace 13-7 ±5-4 0-5 ± 0 1 Trace Trace l-5±0-l 22-0

(%) 18 24-1

2-7 62-3 2-3 6-8

• Each value is the mean of five samples. The average fresh weight of the fruits was 38-2 g at the immature stage, 77-1 g at the mature-green stage, 61 -5 g at the partially ripe stage and 64-4 g at the ripe stage. 22*

Downloaded from http://aob.oxfordjournals.org/ at UQ Library on September 6, 2016

Onlyfivecarotenoids were found in the immature and six in the mature fruits. Fractions 1 to 4 were identified as in the preceding section as phytofluene, a-carotene, ^-carotene and zeinoxanthin. Fraction 5 contained only lutein in the immature fruits; traces of lycopene were detected along with lutein in the mature fruits. Unlike the hydroxy carotenoids of the ripe fruits which were almost entirely esterified, zeinoxanthin and lutein were essentially unesterified in the unripe fruit.

622

Rodriguez, Raymmdo, Lee, Simpson and Chichester

DISCUSSION Momordica charantia, like other fruits, synthesizes a large number of carotenoids during the ripening process. With only five carotenoids at the immature and six at the maturegreen stage, the fruit yielded 15 pigments when ripe or partly ripe. This diversification and the accelerated rate of synthesis account for at least a threefold increase in the total carotenoid concentration. Cryptoxanthin is clearly responsible for the orange colour of the ripe fruit since the appearance of colour coincides distinctly with the detection of cryptoxanthin and the intensity parallels its concentration. It is generally agreed that xanthophylls are formed from parent carotenes at a late stage in the biosynthetic sequence. In many xanthophyll-containing fruits, ripening is accompanied by a decrease in carotenes simultaneous with a rapid increase in the xanthophylls. This trend was not seen in Momordica charantia where the carotenes, except a-carotene and ^-carotene, were detected only during the period of active carotenogenesis, i.e. the partly ripe and ripe stages. This type of pattern is also seen in mango (John et al., 1970) where the carotenes are formed preferentially at the partially ripe stage. Only the hydroxy derivatives of a-carotene, zeinoxanthin and lutein, were detectable at the pre-ripening period. As soon as the colour changed, the hydroxy derivatives of ^-carotene, cryptoxanthin and zeaxanthin, appeared and accumulated. The disappearance of lutein with ripening has also been reported in pepper (Cholnoky et al., 1956; Davies et al., 1970), lemon (Yokoyama and Vandercook, 1967) and mango (John et al, 1970) and the appearance of zeaxanthin has also been reported in mango and pepper. However, the distinct shift from the a-carotene to the ^-carotene hydroxy derivatives as seen in Momordica charantia has not been demonstrated in other fruits. The increase in /?-carotene, cryptoxanthin and zeaxanthin is of a magnitude severalfold higher than the decrease in lutein and a-carotene; thus the changes which accompany ripening are not accounted by an interconversion of a- and /^-compounds. In any case, this type of transformation is generally considered unlikely. The fate of lutein on ripening then presents an intriguing question. In accordance with the generally accepted scheme for carotenoid biosynthesis (Goodwin and Goad, 1970; Goodwin, 1971) a plausible interpretation of our results is presented in Fig. 1. Prior to ripening, the pathway is diverted to the a-carotene sequence (left-hand

Downloaded from http://aob.oxfordjournals.org/ at UQ Library on September 6, 2016

per fruit were 243 and 1460 \ig. The slightly lower concentration per gram at the maturegreen stage did not reflect a true decline since the carotenoid content per fruit (426 fig) increased at this stage but the gain in size and weight was marked. The substantial increase in carotenoid concentration of the ripened fruit could be attributed directly to the accelerated synthesis of cryptoxanthin, /?-carotene and zeaxanthin. The most dramatic change was seen in cryptoxanthin which was absent in the immature and mature fruits but became the principal pigment of the ripe fruit, comprising about 62-3 per cent of the total carotenoid content. A similar trend, though to a much lower extent, was observed with zeaxanthin. Although cryptoxanthin and zeaxanthin could be presumed to be derived from ^-carotene, an overall increase was seen in ficarotene during ripening. A slight increase in lycopene was also discernible. The reverse pattern was seen with a-carotene and lutein which were the major carotenoids of the immature fruit. The change in zeinoxanthin was not clear-cut. It appeared to increase with maturity at the preripening stages. The concentration fell off at the partially ripe stage, but increased again as the ripening continued. Flavochrome, 5,6-monoepoxy-/?-carotene, mutatochrome, C-carotene, ^-carotene, y-carotene and rubixanthin were found at trace levels in the partly ripe and ripe fruits, but were not detected before the colour break. Phytoene was not isolated at any of the stages examined.

Carotenoid Pigments in Momordica charantia Fruits

623

Phyloene Phytolluene

1

{ - Carotene Neurosporene

a-Zeacarotene /

\

8- Carotene

/3-Zeacarolone y-Carotene —-Rubixonthm /9-Carotene —•• 5,6 -Monoepoxy-/9- carotene — Mutatochrome

Esterlfied-i Zeinoxanthin

Zeinoxanthin I

Cryptoxonthin — Esterif led I Cryptoxanthln

Esterifiea1-' Lutein

Lutein

Zeaxanthin

Active pathway at prenpening period

-Exterified Zeaxanthin

Active pathway at ripening period

FIG. 1. Possible carotenoid pathways in developing Momordica charantia fruits.

side of pathway). However, at the onset of ripening, the pathway is diverted to the ficarotene sequence (right-hand side of pathway). The control mechanism responsible for this shift is not known. As in other fruits, esterification occurs with the ripening process. The detection of <5-carotene at the ripening stages seems surprising since a-carotene and lutein levels are declining at this point. However, since a-carotene is no longer formed, the precursor ((5-carotene) can now accumulate at detectable levels. In the same manner, zeinoxanthin, which is no longer transformed to lutein, increases slightly in concentration. Epoxidation in Momordica charantia evidently occurs only at the a-carotene and ^-carotene levels, not at the xanthophyll levels. Violaxanthin and neoxanthin, which have been reported in a wide variety of fruits and are presumably formed from zeaxanthin, are conspicuously absent. Likewise, the epoxides of cryptoxanthin and lutein were not detected. The natural occurrence of furanoids, mutatochrome and flavochrome in this case, is still an open question since epoxide-furanoid rearrangement could easily occur during the usual isolation procedure (Liaaen-Jensen, 1971). ACKNOWLEDGEMENTS The authors acknowledge with gratitude the gift of authentic carotenoidsfrom F. HoffmanLa Roche Ltd., Basle, Switzerland, and Nutley, New Jersey. This work was supported by U.S./H.E.W. Public Health Services Grant 5R01-FD0O433 to C. O. Chichester. LITERATURE CITED CHOLNOKY, L., GYORGYFY, C , NAGY, E., and PANCZEL, M., 1956. Function of carotenoids in chlorophyll-

containing organs. Nature, Lond. 178,410-11. DA VIES, B. H., 1965. Analysis of carotenoid pigments. In Chemistry and Biochemistry of Plant Pigments, ed. T. W. Goodwin. Academic Press, London and New York. MATTHEWS, S. and KIRK, J. T. O., 1970. The nature and biosynthesis of the carotenoids of different colour varieties of Capsicum annum. Phytochem. 9, 797-805. EDWARDS, R. A. and REUTER, F. H., 1967. Pigment changes during the maturation of tomato fruit. Food Technol. Aust. 19, 352-5.

Downloaded from http://aob.oxfordjournals.org/ at UQ Library on September 6, 2016

Flavochrome-"— a - C a r o t e n e * ^

624

Rodriguez, Raymwdo, Lee, Simpson and Chichester

GOODWIN, T. W., 1971. Biosynthesis. In Carotenoids, ed. O. Isler. Birkhauser Verlag, Basel. and GOAD, L. J., 1970. Carotenoids and triterpenoids. In The Biochemistry of Fruits and Their Products, vol. 1, ed. A. C. Hulme. Academic Press, London and New York. and JAMIKORN, M., 1952. Biosynthesis of carotenes in ripening tomatoes. Nature, Lond. 170,104-5. GORTNER, W. A., 1965. Chemical and physical development of the pineapple fruit. 4. Plant pigment constituents. / . FdSci. 30, 30-2. GROSS, J., GABAI, M., and LIFSHTTZ, A., 1971. Carotenoids in juice of Shamouti orange. Ibid. 36,466-73. JOHN, J., SUBBARAYAN, C., and CAMA, H., 1970. Carotenoids in 3 stages of ripening of mango. Ibid. 35, 262-5. KATAYAMA, T., NAKAYAMA, T. O. M., LEE, T. H., and CHICHESTER, C. O., 1971. Carotenoid transformations

in ripening apricots and peaches. Ibid. 36, 804-6. LIAAEN-JENSEN, S., 1971. Isolation, reactions. In Carotenoids, ed. O. Isler. Birkhauser Verlag, Basel. MEREDITH, F. I. and PURCELL, A. E., 1966. Changes in the concentration of carotenes of ripening homestead tomatoes. Proc. Am. Soc. hort. Sci. 89, 544-8. MILLER, E. V., WINSTON, J. R., and SCHOMER, H. A., 1941. Physiological studies of plastid pigments in

rinds of maturing oranges. / . agr. Res. 60,259-67. RAYMUNDO, L. C , GRIFFITHS, A. E., and SIMPSON, K. L., 1967. Effect of dimethyl sulfoxide (DMSO) REID, M. S., LEE, T. H., PRATT, H. K. and CHICHESTER, C. O., 1970. Chlorophyll and carotenoid changes

in developing muskmelons. / . Am. Soc. hort. Sci. 95, 814-5. RODRIGUEZ, D. B., SIMPSON, K. L., and CHICHESTER, C. O., 1973. The biosynthesis of astaxanthin. 17.

Intermediates in the conversion of /(-carotene. Int. J. Biochem. 4,213-22. STAMBERO, O. E., 1945. Vitamin C and carotene in rose hips and products. FdRes. 10,392-6. VALADON, L. R. G. and MUMMERY, R. S., 1972. Carotenoids of rowan berries. Ann. Rot. 36,471-4. YOKOYAMA, H. and VANDERCOOK, C. E., 1967. Citrus carotenoids, 1. Comparison of carotenoids of mature-green and yellow lemons. / . Fd Sci. 32, 42-8. ZSCHEILE, F. P., JR., 1966. Correlation of spectroscopic observations on polyenes with ripening of tomato fruits. Archs. Biochem. Biophys. 113, 554-68.

Downloaded from http://aob.oxfordjournals.org/ at UQ Library on September 6, 2016

on the biosynthesis of carotenoids in detached tomatoes. Phytochem. 6,1527-32. 1970. Biosynthesis of carotenoids in the tomato fruit. Ibid. 9,1239-45.