Studies on Animal Carotenoids: I. Carotenoids of the Brown Trout (Salmo Trutta Linn.)

This paper presents an analysis of the carotenoids of the brown trout (Salmo trutta Linn.) being the first of a series designed to throw light on the general functions of this class of pigments in animals. Apart from the role of the carotenoid derivative vitamin A as a component of visual purple, and the importance of some carotenes as vitamin A precursors, precise knowledge is almost entirely lacking of the part played by this group of substances in any metabolic process. Many of the lower Vertebrates, however, are known to accumulate large amounts of carotenoids and vitamin A in the liver and other tissues. The carotenoids are prominent also in the colour patterns of the skin, in which they are the characteristic pigments of the lipophore group of chromatophores. It seemed, therefore, that a quantitative study of their distribution in selected species of fish might prove a fruitful approach to clarifying the problems of their general metabolism. The brown trout was chosen for investigation since it is easily obtainable and possesses a characteristic colour pattern with at least two types of lipophores, the yellow or golden xanthophores and the red erythrophores. A further advantage of this species is that a considerable body of information already exists on various other aspects of its development and metabolism.

Trout were obtained from a reservoir near Edinburgh, which was stocked annually with fish of the Loch Leven type from the hatcheries of the Howietoun and Northern Fisheries, Stirling. Natural spawning also occurred in a tributary stream, so that the fish population was in part self-replacing. Samples of fish were obtained by netting, which took place at intervals in connexion with another experiment. They were brought back to the laboratory alive and, as far as possible, all extractions of carotenoids were begun immediately after killing. A few fish were caught by rod- and-line angling and used immediately on return to the laboratory. When large amounts of material were being handled it was sometimes necessary to store certain portions, usually the first crude carotenoid extracts, and these were kept in the freezing compartment of a refrigerator. In general, all operations were carried out with a minimum delay and the temperature of all solutions kept as low as possible.

The procedures used in these investigations for separating individual carotenoids were similar to those described by Zechmeister (1934), Zechmeister & Cholnoky (1941), Wald & Zussman (1938) and Baldwin & Beatty (1941). The pigments were extracted exhaustively from portions of minced tissues with acetone, methanol or petroleum ether (b.p. 40 –60°), in some cases directly, but usually after grinding with sharp sand and anhydrous sodium sulphate. The combined extracts were transferred to petroleum ether by dilution with water, dried by shaking with anhydrous sodium sulphate, and then subjected to partition with 90% methanol. The petroleum ether was then evaporated under reduced pressure and the oily residue saponified with a few ml. of 6% KOH in ethyl alcohol for 1 –2 hr. at 40 –50° C. The saponification mixture was diluted with an equal volume of water and again extracted exhaustively with petroleum ether. This extract was washed successively with 5% KOH in 50% ethyl alcohol and water, and then partitioned once more between petroleum ether and 90 % methanol. The saponification residue was acidified with a few drops of glacial acetic acid and extracted with further petroleum ether until no more pigment could be removed.

In this way the carotenoids were imperfectly separated into the following four fractions:

1. Hydroxy-carotenoids, mainly xanthophylls, which were extracted from the tissues as free alcohols and were removed by 90% methanol at the first partition.

2. Xanthophylls extracted from the tissues as esters, which were epiphasic before but hypophasic after saponification.

3. Carotenes, which were epiphasic both before and after saponification.

4. Acidic carotenoids, which remained as the potassium salt in the saponification residue after the carotenes and xanthophylls were removed by petroleum ether, and were only extracted therefrom after acidification.

These preliminary fractions were washed with water and dried with anhydrous sodium sulphate before being subjected to further fractioning and purification by chromatographic adsorption. The adsorbents most frequently used were alumina, (Savory and Moore’s special preparation for chromatography) magnesium oxide (B.D.H. preparation low in arsenic) and finely powdered calcium carbonate. For rapid and clear development of the pigment bands much use was made of 2% methanol in petroleum ether, as recommended by Baldwin & Beatty (1941). The main pigment bands were in some cases separated by division of the column before elution. Usually, however, they were washed through the column by excess of the eluent and were collected separately. Pure petroleum ether, hexane, methanol, carbon disulphide and pyridine were all used as eluents at one time or another. It was usually necessary to adsorb and re-elute each fraction two or more times, often with different adsorbents and eluents, in order to obtain solutions showing constant spectroscopic properties and sharp absorption maxima. Imperfectly purified preparations often exhibited indefinite maxima when more than one pigment was present, or the absorption rose steeply in the violet region of the spectrum owing to the presence of non-carotenoid substances, probably mainly sterols, which absorb strongly in the near ultra-violet.

Absorption spectra were measured with a Hilger-Nutting Constant Deviation Wave-length Spectrophotometer (type D 186), which was calibrated at intervals by replacing the normal light source temporarily with a sodium or mercury vapour lamp and checking the wave-length scale with the sodium line at 589 ·0 mμ and with the mercury-lines at 435 ·8, 491 ·6 and 546 ·1mμ. The carotenoid solutions, brought to suitable concentrations in pure solvents, were measured in standard type Hilger cells of 1 cm. depth. Measurements were normally made in steps of 5 ·0 mμ, with a spectral band which was about 3 mμ wide at 475 mμ.

Solutions of standard preparations of b-carotene, lutein and astacene were used for comparison with unknown pigments, both in respect of their absorption spectra and their behaviour on chromatograph columns. The preparation and purity of these standards is discussed below in the section dealing with quantitative investigations.

The pigments were extracted from chopped and ground portions of tissue, easily with acetone, pure petroleum ether or petroleum ether containing 2% methanol; slowly and with difficulty by pure methanol. They gave a clear golden solution in petroleum ether, which was entirely epiphasic before saponification. Crude extracts adsorbed on columns of magnesium oxide or alumina separated into two bands, a strongly adsorbed diffuse rose-coloured pigment at the top of the column and a more easily eluted orange-yellow pigment below. After saponification direct extraction of the diluted mixture with petroleum ether yielded a golden fraction, which exhibited the following properties when partitioned with methanol of different concentrations:

Partitioned between petroleum ether

This behaviour suggested that the fraction contained dihydroxy-xanthophylls only. It was adsorbed on magnesium oxide and developed with 2% methanol in petroleum ether. Most of the pigment was contained in a strongly defined orangeyellow band, which was eluted easily with methanol or excess of the developer. In addition a small amount of the more strongly adsorbed rose pigment appeared at the top of the column, and in some experiments one or more faintly defined pale yellow bands below the main pigment. Addition of a solution of the lutein standard reinforced the main band, while b-carotene passed through the column without being adsorbed at all. The pale yellow trace bands were not isolated in sufficient amounts for further investigation. Their adsorption properties show that they were not carotenes. They probably were either monohydroxy-carotenoids or degradation products of carotenes.

The adsorption spectrum of the main orange-yellow band in hexane is compared in Fig. 1 with that of the lutein standard. The maxima at 475 and about 445 mμ agree well with the maxima of pure lutein in this solvent given by Morton (1942). In carbon disulphide the maxima lay at 512, 475 and about 450 mμ.

The saponification residue yielded after acidification a pigment which appeared to correspond with the rose-coloured fraction of the crude extracts. It was mainly hypophasic on partition between petroleum ether and 5% KOH in 50% ethyl alcohol, but returned to the petroleum layer on reacidification. It was purified by adsorption on magnesium oxide, from which it was eluted with difficulty by acidified methanol or pyridine. Its absorption spectrum showed a single broad band with the maximum at 475 mμ in hexane and about 500 mμ in pyridine. Crude preparations showed the maximum in pyridine between 490 and 495 mμ, but after adsorption and re-elution two or three times the maximum became stable at about 500 mμ.

These properties serve to identify the second skin pigment as astacene. This was confirmed by comparing its absorption spectrum with that of that astacene standard (Fig. 2). A mixed chromatogram with the standard also gave a single intensified rose-coloured band on magnesium oxide. A few small purple crystals were obtained by partial evaporation of solutions in pyridine followed by dilution with a little water. These were insoluble in hexane, but dissolved readily in pyridine.

From their spectral properties one would expect lutein to be the pigment of the xanthophores and astacene of the erythrophores. To test this, pieces of skin were cut so as to exclude all red spots. On the chromatograph most of the pigment developed into the orange-yellow lutein band. Carefully dissected pieces of the red skin spots and the red tips of the adipose fins, on the other hand, developed intense rosecoloured astacene bands and relatively little lutein. Frozen sections were also cut of pieces of these tissues. Yellow skin apparently contained only xanthophores and melanophores fairly evenly distributed in the dermal layers, while the red spots and the tips of adipose fins showed dense aggregations of erythrophores, a few scattered melanophores, but apparently no xanthophores. These observations suggest that neither type of chromatophore contains a single pigment, but that the xanthophores contain mainly esters of lutein and the erythrophores mainly esters of astaxanthin.

Note. It is recognized that astacene or tetra-keto-6-carotene is usually regarded as a chemical artefact, being derived by oxidation from astaxanthin, which is the double a-ketol of 6-carotene and is probably the naturally occurring pigment in living organisms. Most of our knowledge of astaxanthin, however, is derived from investigation of the properties of astacene, and in this paper the latter term is used.

The main body muscles of many trout are highly pigmented, different fish showing all gradations from white, through various intensities of yellow and orange to a rich salmon pink. The amount of carotenoid extracted was roughly proportional to the intensity of pigmentation, none being obtained from fish with white muscle while well pigmented ones yielded a clear orange-pink solution in petroleum ether. Extracts of portions of pink muscle were analysed in the same way as described for skin. Some of the pigment was hypophasic between petroleum ether and 90% methanol before saponification. After saponification a further small fraction was removed by shaking the diluted mixture directly with petroleum ether, but the bulk remained in the saponification residue from which it was extracted after acidification. The acidic fraction possessed properties similar to those of the red skin pigment and the standard preparation of astacene, and after purification showed a single absorption maximum close to 500 mμ in pyridine (Fig. 3). The xanthophyll fraction when adsorbed on to magnesium oxide and developed with 2% methanol in petroleum ether separated into two closely associated bands: upper band, pink; lower band, orange-yellow. These moved down the column together and were separated with difficulty. The upper band was insufficient for further investigation, but the fewer one, which contained most of the pigment, possessed the absorption maxima at 475 and 447 mμ in hexane characteristic of lutein (Fig. 4). Addition of lutein standard solution to the chromatograph increased the intensity of the lower band. It seems, therefore, that the hypophasic fraction of muscle contains two xanthophylls, principally lutein and probably a small proportion of its stereoisomer zeaxanthin.

Some samples of muscle also yielded a minute fraction which remained epiphasic after saponification, and was therefore probably a carotene.

Frozen sections were cut at 20 –25 μ pieces of fresh muscle and others fixed in formal calcium (Baker). These appeared colourless, but on maceration or squeezing the fat which separated from the fibres had a pinkish tinge. Formal fixed sections at 10 μ, stained with Sudan IV and Sudan Black by the method of Baker (1945), showed numerous small globules of lipoid apparently lying between the muscle fibres.

Extracts of the carotenoids of unripe ovaries behaved similarly to those of muscle. The pigment was nearly all hypophasic before saponification, indicating a preponderance of free hydroxy-carotenoids. After saponification the following three fractions were separated by the usual procedure: (1) an acidic fraction, which contained the bulk of the pigment; (2) a considerable xanthophyll fraction; (3) a relatively small hydrocarbon fraction. After further purification the acidic and xanthophyll fractions each yielded a single pigment similar to the astacene and lutein respectively of skin and muscle (Figs. 3 and 4). The hydrocarbon fraction when adsorbed on alumina and developed with 2% methanol in petroleum ether partially separated into two yellow bands, which moved down the column together and could not be completely dissociated from one another. This fraction gave absorption maxima at 453 and 485 mμ in hexane, and at 485 and 518 mμ in carbon disulphide, and therefore probably consisted mainly of b-carotene (Fig. 5), together with a small proportion of another carotene.

Liver carotenoids were more difficult to free from other lipoid substances than those of any other tissue, and in many experiments the absorption maxima were obscured by steeply rising absorption at wave-lengths shorter than about 450 mμ. To obtain sufficient pigment it was necessary to analyse mixed samples obtained by combining the livers of several fish. Although the relative proportions of the different fractions varied in different samples, in general rather more than half the pigment was hypophasic before saponification. When adsorbed on magnesium oxide or calcium carbonate and developed with 2 % methanol in petroleum ether, tire hypophasic fraction usually separated into the following bands: (1) one or more trace bands at the top of the column, not obtained in sufficient amounts for further investigation; (2) an orange-yellow band; (3) a golden yellow band. Bands 2 and 3 were closely associated and difficult to separate. Band 3 possessed absorption maxima at 448 and 478 –479 mμ in hexane and at 475 and 510 mμ in carbon disulphide, which indicate lutein. Lutein standard solution added to the chromatograph intensified this band. Band 2 was only obtained in small quantities and separated imperfectly. It showed maxima at 454 mμ in hexane and at 485 and 517 mμ in carbon disulphide. The position of these maxima, and the fact that it lay above lutein on the column suggest that band 2 consisted of zeaxanthin. The trace bands at the top of the column were probably due to small amounts of xanthophylls with three or more hydroxyl groups.

The epiphasic fraction of the crude extracts yielded after saponification a further small hypophasic fraction, which could not be broken down further by adsorption. It showed indefinite maxima at 452 –454, about 470 and at 485 mμ in hexane, and at 484 –485, 510-511 and 520 mμ in carbon disulphide. These multiple absorption maxima suggest that both lutein and zeaxanthin were present, and possibly other xanthophylls also.

The hydrocarbon fraction was freed from xanthophylls by pouring on to a column of calcium carbonate, which retained the latter but allowed the carotenes to pass through. This fraction showed maxima at 425 –426, 450 and 476 mμ in hexane (Fig. 5), and at 455, 483 and 505 –510 mμ in carbon disulphide, and appeared to consist mainly of 6-carotene, though the maxima at 476 in hexane and at 505 –510 mμ in carbon disulphide correspond more closely with the accepted values for a-carotene, as fisted by Morton (1942). When adsorbed on alumina or magnesium oxide and developed with 2 % methanol in petroleum ether, partial separation into two or even three bands was usually observed, but these remained too closely associated to be eluted individually. The livers therefore probably contained both a- and b-carotene and possibly other carotenoid hydrocarbons.

The saponification residues of large samples comprising the livers of twenty or more fish were tested repeatedly for an acidic fraction but only traces, too small for further investigation, were detected.

Liver contains therefore carotenes, free and esterified xanthophylls, but no astacene.

Batches of about fifty whole eyes or enucleated posterior hemispheres were exposed to direct sunlight for a few minutes to bleach all visual purple. They were then ground with anhydrous sodium sulphate and extracted with acetone. The carotenoids were transferred to petroleum ether, in which they gave a yellow solution, by diluting the acetone to about 75% with water. The greenish yellow pigment which remained in the diluted acetone was assumed to be non-carotenoid. It fluoresced strongly green in ultra-violet illumination, and probably consisted of flavines. The carotenoid was all epiphasic before, but hypophasic after saponification. Only a single pigment, which corresponded closely with the xanthophyll of skin and muscle, was present. No carotene or astacene was detected in any experiment.

Other tissues examined, but which all yielded negligible amounts of carotenoids, were testes, brain, spleen, kidneys, stomach and intestine, the last two being washed clean of their contents. The testes, which are milky white, do not appear to contain carotenoids at any stage of the reproductive cycle.

For measuring the amounts of the different carotenoid fractions of tissues a standard separation procedure was adopted involving only partitioning before and after saponification, and omitting chromatographic methods. Weighed portions of tissue were extracted exhaustively as described above, the combined extracts transferred to petroleum ether, which was then made up to a convenient measured volume, and the absorption due to the total carotenoids was measured. The petroleum ether was then evaporated under reduced pressure and the residue saponified in the usual manner. The diluted saponification mixture was extracted exhaustively with petroleum ether, then acidified and extracted once more until no further pigment could be removed. The combined extractions before acidification were separated into carotene and xanthophyll fractions by partitioning three times between petroleum ether and 90% methanol. The hypophasic fraction was returned to petroleum ether by dilution, and all three fractions were made up to convenient volumes for estimation in this solvent. This procedure was varied slightly to estimate the relative amounts of free and esterified xanthophylls extracted from any tissue. In these cases the free xanthophyll fraction was separated from the rest by partitioning three times with 90 % methanol before saponification. The hypophasic fraction was then returned to petroleum ether and estimated separately.

It was appreciated that fractionation of carotenoids by partitioning alone is incomplete, since small amounts of other types can usually be detected on chromatograph columns even after repeated partitioning. It was considered preferable, however, to tolerate a small error from this source rather than to subject the fractions to be estimated quantitatively to further purification by chromatography, since the proportion of pigment recovered from the latter cannot be simply standardized. Moreover, since the absorption maxima of the carotenoids which are the subject of this investigation all lie close together at about 450 and 475 mμ in petroleum ether, it was not practicable to estimate their relative proportions in a single solution.

Estimations were made with a Unicam photoelectric colorimeter, type S21, using filters which transmitted only a single narrow wave-length band with maxima close to 450 and 475 mμ respectively. These were made up as follows: filter 450 mμ maximum, Ilford gelatine filters nos. 305, 803 and Q805; filter 475 mμ maximum, Ilford gelatine filters nos. 303, 803 and Q805. The gelatine strips were held between glass sheets cut from thin microscope slides, which were selected for optical clarity. The ultra-violet cut-out filter Q805 was included to reduce errors due to the presence of colourless impurities with strong absorption in the near ultra-violet, which were often apparent in the absorption spectra of liver and ovary extracts in particular. The inclusion of an infra-red absorbing filter, no. 803, was considered advisable since photocells of the type used in this instrument are highly sensitive in the near infra-red, and small amounts of substances absorbing in this region might cause serious errors.

Calibration curves were constructed from serial dilutions of petroleum ether solutions of standard preparations of carotene, lutein and astacene, and the concentrations of the appropriate fractions extracted from different tissues are expressed in terms of these standards. The standard preparations were obtained from the following sources :

b-carotene: B.D.H. pharmaceutical preparation. This proved in fact to be a mixture of carotenes, of which b-carotene was the most prominent. It contained also a small insoluble non-carotenoid residue.

Lutein: obtained from Prof. L. Zechmeister and prepared at the California Institute of Technology. Not guaranteed to be better than 90% pure. There was no insoluble residue.

Astacene: crystals prepared in the laboratory from fresh shells of lobsters. The preparation was free from other carotenoids, but contained a small proportion of impurity which was insoluble in organic solvents.

Probably none of the standards was in fact better than 90% pure, and the values obtained for the carotenoids of trout should be accepted with this reservation.

The carotenoid contents of the various tissues investigated are summarized in Table 1. The figures presented are the average values of several determinations, except in the case of muscle. As mentioned above, white muscle yielded no carotenoid, and the figures given were those obtained from a highly pigmented specimen, which probably approximated to the maximum normally attained by this species. Yellow or orange-coloured muscle yielded rather lower values, 15 –20 γ/g. for astacene and 1 –2 γ/g. for xanthophyll. By far the largest concentrations of both pigments were found in the red skin spots and the red tips of the adipose fins. These were dissected out carefully, being separated as far as possible from the bony scales and underlying muscle, so that only small portions of the bright red dermis and epidermis were estimated.

It is of interest that the skins of fish of the Loch Leven type yielded the same concentrations of lutein and astacene as those of ordinary brown trout. Loch Leven trout are distinguishable by their silvery appearance, most marked on the sides and belly, compared with the golden or yellow colour of the ordinary race. This silvery surface appears to be a structural type of pigmentation in the scales or outer layers of the skin, since sections showed that the xanthophores, which lie mostly below the scales in the deeper dermal layers, are as numerous and as well pigmented in Loch Leven as in normal brown trout.

It is well known that trout and other species of fish when kept in aquaria tend to lose the brilliant red and yellow pigmentation characteristic of the wild type. Xanthophores and erythrophores become few in number and those remaining contain little pigment. The fish can be kept indefinitely, will continue to grow and appear to suffer no other disability. If the loss of colour is due to lack of carotenoids in the usual aquarium diets, addition of specific pigments might be expected to show whether the fish are capable of converting one type of carotenoid into another; whether astacene, for example, is obtained direct from the food or is synthesized by the trout from one of the carotenes.

Yearling fish of about 8 cm. length and 5 –9 g. were fed on horseflesh varied occasionally by chopped earthworms. The carotenoid content of this diet is negligible. After nine months on this régime very little red and yellow pigment was apparent. Microscopic examination revealed pale yellow xanthophores generally distributed throughout the skin and fins, and a few orange-pink erythrophores mainly at the tip of the adipose fins. Erythrophores had disappeared from the red spots along the lateral line. Estimations of the whole skin and fins of such fish showed that the total carotenoids, estimated as lutein, were reduced to about 10% of the value normally found in wild fish (Table 2). Moreover, all the pigment remaining was lutein, whereas the skin of wild fish yielded lutein and astacene in about equal amounts.

• Three groups, each consisting of three depigmented fish, were placed in separate tanks and their diet supplemented with the standard carotenoid preparations, one group receiving b-carotene, one lutein and one astacene. These were dissolved in pure arachis oil and injected into the body cavity of earthworms in which the septa between the segments had been destroyed by running a long needle through them. As much oil solution as possible was introduced, the wound closed with a clamp and the worms placed in the refrigerator for a few minutes until the oil was fairly hard and did not immediately flow out on removing the clamp. They were then fed to hungry fish, which habitually took them at once. No attempt was made to measure the exact amount of carotenoid each fish received. The total amounts given to each group and the duration of the experiments are recorded in Table 2. Allowing for differences in size and appetite, each fish probably received at least ten times its normal content of lutein and astacene and a comparable amount of b-carotene. No fish of any group, however, showed any increase of visible pigmentation, and the carotenoids of their skin and fins at the end of the experiment were as low as those of a control group which received no supplement. It seemed possible that the carotenoids, though ingested, were not absorbed in the gut of the trout. To test this several samples of faeces were collected and estimated, but these yielded only negligible amounts of pigment.

• Another group of eight depigmented yearling fish were fed about 5 g. of fresh salmon ova daily for 35 days. Their pigmentation increased visibly, and by the end of this period five had developed red flank spots and red adipose fins, and two of these even approached normal wild trout in brilliance. The background colour of their skin was rather more pink than in normal fish, and many of the xanthophores had a pinkish tinge. Four of this group were killed and the carotenoids of their skin and fins estimated. The astacene level was almost normal and the lutein about four times above that of the depigmented controls (Table 2). The remaining four were allowed to revert to the horseflesh diet. Their pigmentation faded once more and after three months none of them was distinguishable from the controls. Salmon ova contain lutein and astacene in about the same concentrations as trout ova. Each fish of this group received therefore about five times its normal content of lutein and about fifty times its normal content of astacene. The disproportion between the lutein and astacene content of salmon ova may well account for the abnormally pink colour developed by these fish, indicating that in the absence of sufficient lutein the xanthophores are able to lay down astacene esters.

• Another group of two fish was reared from ova in the laboratory and fed for over a year on ‘natural’ foods only. They were kept in a ‘balance’ tank with a permanent flora and fauna of Diatoms, filamentous green Algae of several species, Elodea canadensis, snails, Crustacea and insect larvae, to which rich infusions of Entomostraca, mainly Daphnia, Simocephalus and Cyclops species, and Corethra larvae were added from time to time. They fed exclusively on these, being given no horseflesh or earthworms. They grew well, attaining a size comparable with that of the control yearlings, and developed normal wild-type pigmentation. The carotenoid content of their skin and fins was found to be of the same order as in wild fish. Estimations of the carotenoids of the Entomostraca infusions leave little doubt that these provided the main source of the pigments during the first year of growth following the larval period. An infusion which consisted of about 85% Daphnia longispina (Müller), the rest being mainly Cyclops species, yielded 6520 γ of astacene, 1385 γ of xanthophylls and 120 γ of carotenes per g. dry weight.

The absorption spectra of single xanthophores and erythrophores have been measured with a microspectrophotometer developed by Dr K. Pätau (late of the Kaiser Wilhelm Institut für Biologie, Berlin), details of which will be published elsewhere. Xanthophores show a sharp maximum at 487 mμ, a broad one in the region of 450 mμ, and a small but definite one at 425 mμ (Fig. 6). The maximum at 425 mμ. is interesting since the standard specimen of lutein in hexane shows one at the same wave-length, though it is absent or only just discernible in most spectra of tissue extracts. It is also of interest that though the maximum at 487 mμ is about 10 –12 mμ towards the red from its position in hexane, the one near 450 mμ is less displaced and that at 425 mμ not at all.

The erythrophore shows a single main maximum close to 490 mμ (Fig. 6), and closely resembles the general shape of the spectrum of pure astacene. The small secondary maximum in the region 435 –440 mμ is probably not due to a carotenoid.

A method described by Pantin (1923) for measuring pH inside the cell body of Amoeba was adapted for estimating the lutein and astacene concentrations of single chromatophores. Pantin compared the tint of neutral red indicator in the animal seen through the microscope with a series of standard buffered solutions containing the same indicator and placed so that they were reflected in the mirror of the microscope and focused by an achromatic condenser in the plane of the object. The two images were juxtaposed and compared directly. In the present experiments the optical density of single fully contracted xanthophores and erythrophores was compared with solutions containing known amounts of lutein and astacene respectively. Since it was desired to compare differences in optical density rather than differences of tint, it was possible to match the chromatophores by varying the thickness of the layer of standard solution. If it be assumed that the solutions conformed to Beer’s Law, then the concentration of pigment in the cell should be proportional to the length of the column of standard solution required to match it ; and knowing the concentration of the standard, the thickness and area of the chromatophore, the amount of pigment each cell contains can be calculated.

It was first necessary to prepare standard solutions the absorption spectra of which were as nearly as possible identical with those of the cell pigments. The spectra of carotenoids are affected by the solvent, the maxima being at shortest wave-lengths in hexane or petroleum, but displaced progressively towards the red end of the spectrum in solvents of higher refractive index. In carbon disulphide, the extreme case, the maxima lie about 30 mμ towards the red from their position in hexane. According to Wald (1943) carotenoid spectra in ordinary plant oils are displaced about 12 mμ, while in living tissues they are ordinarily about 15 –20 mμ towards the red from their position in hexane. The standards were dissolved therefore in pure colourless castor oil. By adding carbon disulphide in various proportions to this solution it was possible to adjust the absorption maxima to any position within the desired range. These solutions, including that in pure castor oil, were used for matching the chromatophores.

The direct measurements of the absorption spectra of living chromatophores made with the microspectrophotometer developed by Dr K. Pätau indicate that the solutions in pure castor oil correspond best with the spectra of the living cells, and the data presented below are therefore those obtained by matching cells with the standards dissolved in this solvent without the addition of any carbon disulphide.

The solutions were sucked into lengths of capillary tubing, the bore of which was selected to give an image in the field of the microscope approximately the same size as the chromatophore. The tube was mounted horizontally and its position adjusted SO that the solution was viewed accurately through the length of the column. Solution was then added or removed, care being taken to avoid introducing any air bubbles, until a match was obtained with the chromatophore. The length of the column was then measured to the nearest millimetre.

The chromatophores were examined in thin slices and frozen sections of fresh skin and fin tissues, mounted in frog Ringer’s solution. They were first contracted by immersing for a few minutes in adrenalin 1/100,000 solution or in solutions containing a little barium chloride. When completely contracted the pigmented part of the cells appear more or less spherical. Their diameters, measured with a micrometer eyepiece, varied between 6 ·5 and ioμ. Well-rounded cells of 8p diameter were used for matching, and it was assumed that they were spherical so that the maximum thickness of pigment through which the light passed was about the same as the cell diameter measured in the plane of the section. No allowance was made for scattered light, the possible lens effect of the cell, or the numerical aperture of the objectives used, since none of these factors could be evaluated simply. The microscope was focused just above the equator of the cell and the match made through the centre, i.e. through the maximum depth of pigment.

The data obtained from matching xanthophores and erythrophores with standard solutions of lutein and astacene in castor oil are summarized in Table 3. Xanthophores were matched by columns from 2 to 5 mm. in length of a solution containing 166 γ/ml. of lutein. Assuming the pigment layer to be 8μ thick, the calculated concentration within the cells was from 0 ·000011 to 0 ·000028 γ of lutein per xanthophore. Erythrophores were matched by lengths from 68 to 115 mm. of astacene solution containing 90 γ/ml., which correspond to values from 0 ·00020 to 0 ·00034/ Per cell-For both types the hue and saturation of the standards were very similar to the cells, which rendered subjective matching of their optical density relatively simple.

Previous investigators have established that carotenoids are distributed widely in fish, but few attempts have been made to identify specific pigments in definite types of cells, and quantitative information is very scanty. Carotenes, xanthophylls and carotenoid acids are all known to occur together or separately in different species. Lönnberg (1932, 1934, 1936), in a series of short papers, reported carotenes and xanthophylls in the skin of a large number of species. The methods he used did not, however, distinguish between free and esterified xanthophylls, so that the identification of carotenes is suspect (at least the carotenes could not be distinguished from xanthophyll esters). Lederer (1938) identified astacene in the skin of two marine and one fresh-water species. Among the Salmonidae, however, the pigments of muscle have received most attention. Emmerie, van Eekelen, Josephy & Wolff (1934) extracted a red lipochrome from the muscle of salmon (Salmo salar), which they named salmic acid, but its absorption spectrum and chemical properties show it to have been astacene. Sorensen & Stene (1938) identified the same pigment in the muscles of sea and fresh-water races of Salmo trutta as free unesterified astaxanthin, and obtained some quantitative data. Trout from Norwegian mountain lakes (’ Gebirgsforelle’) held about double the concentration found in fresh run sea trout. Bailey (1937) found astacene in the muscles of the North Pacific species Oncorhynchus nerka and Salmo gairdneri, and I have found it also in the char (Salvelinus sp.) of Scottish lochs (unpublished observations). It is obviously a widespread if not a general characteristic of the Salmonidae to store large amounts of this carotenoid in the muscles, probably as free astaxanthin.

Few investigators have paid much attention to carotenoids present in smaller concentrations, or to tissues other than muscle. In a little-known paper, however, Newbigin (1900) showed that the muscles and ova of Atlantic salmon contain both a pink and a yellow lipochrome, which were ‘very similar to those of Crustacea and perhaps identical with them’, and considered the pink one to be the same as crustaceorubin and tetroerythrin, which are the names under which astacene appears in the earlier literature. Examination of her data shows that the pink pigment almost certainly was astacene and the yellow, which was present in much smaller amounts, either lutein or a mixture of xanthophylls. She noted that the yellow pigment occurred also in the liver, though the methods she used would not of course distinguish between carotenes and xanthophylls.

My investigations on brown trout emphasize that the distribution of the different carotenoid types among the various tissues and even cells is highly specific. Lutein and astacene, but no carotene, occur in the skin entirely as esters. Their relative proportions in the yellow and red regions and the direct measurements of the absorption spectra of single cells show that they are the characteristic pigments of the xanthophores and erythrophores respectively. It seems that under natural conditions the two pigments do not occur together in the same cell to any great extent, although the experiments in which depigmented fish were fed on fresh salmon ova suggest that the xanthophores may take up astacene if the latter is available in great excess. Lutein and astacene occur also in the muscles and ova as free hydroxy-carotenoids. The liver contains both free and esterified xanthophylls, but no astacene. The largest concentrations of carotene, however, were found in the liver, though measurable amounts occur also in the ova and traces in the muscles. It seems likely that the carotenes, being the only group of these carotenoids known to act as precursors of the vitamins A, are mostly utilized in this way. Moreover, if the liver is the site of synthesis of vitamin A in fish, as it is believed to be in mammals, it is not surprising that most of the carotene occurs there.

Although the experiments in which depigmented fish were given supplements of pure carotenoids failed to show whether trout can convert or synthesize any of these substances, they support the view that the pigments of the lipophores, unlike those of the melanophores, are derived solely from the diet. According to McCay (1937) dried ova of salmon, goldfish, Daphnia, carrots and other foods rich in carotenoids have all been claimed to be effective in restoring or maintaining the wild-type coloration of trout kept in aquaria. Incidentally, goldfish (Carassius auratus) was one of the species in which Lederer (1938) identified astacene. Sumner & Fox (1935) found that the xanthophylls of Gtrella nigricans were progressively reduced when the fish were kept in aquaria. After 5 months the carotenoid level was about 15% of normal, a reduction comparable with that found in my trout. Moreover, they found that supplements of carotene and xanthophyll added to the basal diet had no effect on the rate of depletion. The carotenoid supplements were added to the food by evaporation from petroleum ether in a vacuum desiccator, and are open to the same objection as in my experiments, namely that they may be altered in some way which renders them incapable of being absorbed or utilized. Sumner & Fox did obtain some evidence that their supplements passed through the alimentary canal of the fish and were voided with the faeces. It does not seem to matter from the point of view of utilization whether the pigments are ingested as esters or free hydroxy-carotenoids, as there is evidence that the former are acted upon by lipases in the alimentary canal and are absorbed as alcohols. Young & Fox (1936) found that the xanthophyll esters of shrimps fed to Pacific surf perches of the genus Cymatogaster could be recovered as free alcohols in the cells of the intestinal wall and the faeces. It is more probable that many of the pure carotenoid preparations are really chemical artefacts, as astacene is known to be a reaction product of astaxanthin. If this is so there is less cause for surprise that animals may fail to absorb the purified preparation although the corresponding naturally occurring pigment is utilized.

Other experiments, which will be described in a separate communication, support the view that trout cannot synthesize carotenoids, but utilize them efficiently if available in a suitable form. Xanthophores and erythrophores mostly develop during the latter part of the larval period, when the yolk sac is being rapidly absorbed, and if nearly all the carotenoids of the yolk are removed by operation soon after hatching the larvae develop very few and pale chromatophores.

Obviously the question of synthesis and utilization of carotenoids by fish requires careful investigation, and generalizations from experiments with one or two species would not be justified. Sumner & Fox (1933) considered there was presumptive evidence that Fundulus could convert carotenes to xanthophylls, and Lonnberg was struck by the fact that while xanthophylls were the predominating fraction in most fish, the Invertebrates on which they feed contained mostly carotenes. Speculations of this type are of little value unless based on quantitative information, for, as suggested above, fish may convert carotenes to vitamin A while progressively accumulating xanthophylls and other carotenoids. In Girella and the trout at least, two species which have now been searchingly investigated, there is no evidence that the fish can convert one type of carotenoid to another.