Because the luciferin is almost completely precipitated by saturation with (NH4)2SO4, we may conclude that it occurs in water in the colloidal state. This excludes it from belonging to one of the numerous groups of biochemical compounds occurring in true solution and places it among the known groups of colloidal substances, the soaps, proteins, polysaccharides, phospholipins, galactolipins (cerebrosides), tannins or saponins. It is not a polysaccharide because nearly completely precipitated by phosphotungstic acid, nor a soap because not precipitated by calcium salts, nor a phospho- or galactolipin because insoluble in benzine, hot or cold. It gives no tannin or saponin tests. Only the protein group remains, and of the eighteen protein classes recognized by the American Society of Biochemists, the general properties of luciferin indicate that it should be placed among the natural proteoses, somewhere on the borderland between the proteoses and peptones. The fact that luciferin will dialyze, although almost completely salted out by (NH4)2SO4, is strong evidence in favor of placing it in such a position.
On the other hand, luciferin has two properties which to say the least are unusual for proteins. I refer to its
solubility in alcohols, acetone, esters, etc., and non-digestibility by trypsin or erepsin, which have almost universal proteolytic power.
The best known class of proteins soluble in alcohol is the prolamines of plants, but the prolamines are insoluble in water and in absolute alcohol. Zein, the prolamine of corn, is soluble in 90 per cent. ethyl, methyl, and propyl alcohols, in glycerol heated to 150° C., and in glacial acetic acid. Recently Osborne and Wakeman (1918) have described a protein from milk having solubilities similar to those of gliadin, the prolamine of wheat. Welker (1912) has described a substance, obtained from Witte's peptone, giving the biuret, Millon, and Hopkins-Cole tests, which is soluble in water and absolute alcohol but not in ether, and it is possible that others of the peptones are soluble in absolute alcohol. On the other hand, some proteins in the absence of salts form colloidal solutions in strong alcohol from which they may be precipitated by an appropriate salt. As the absolute alcohol extract of Cypridinæ was made from dry material containing the salts of sea water, some salt was present, but there is always the possibility of sol formation.
If we extract dried Cypridinæ, which have previously been thoroughly extracted with benzine or ether, with 800 c.c. of boiling absolute alcohol for an hour, filter the alcohol extract through blotting paper and hardened filter paper, quickly evaporate the filtrate to dryness on the water bath, and dissolve the residue in a small quantity of water saturated with CO2,[9] we obtain a yellow opalescent solution which gives a bright light with luciferase. This solution contains some protein
or protein derivatives as it gives a very faint Millon reaction, a good positive ninhydrin test, reddish blue in color, but no biuret reaction. It precipitates with tannic and phosphotungstic acids but not with picric, acetic, trichloracetic, or chromic acids. The extract gives a faint Molisch reaction for carbohydrates. As the evidence points to the presence of some protein products in the absolute alcohol extract of Cypridinæ, it is possible that this protein is luciferin. It should be emphasized, however, that the Millon reaction was very faint, although the ninhydrin was quite marked and the biuret negative.
[9] To make the solution slightly acid and prevent oxidation of the luciferin.
Although luciferin is not digested by trypsin, even after five days at 38° C., it does hydrolyze with mineral acids after about 16 hours' boiling. Some proteins, the albuminoids and racemized proteins, resist tryptic digestion but yield to acid hydrolysis. We know also that some NH-CO linkages of proteins are broken down with great difficulty by trypsin as it is difficult to obtain a tryptic digest of protein which does not give the biuret reaction, and the work of Fischer and Abderhalden has shown that certain artificial polypeptides are not digested by pure activated pancreatic juice.
We have, then, three possibilities: Luciferin is (1) either a natural proteose not attacked by trypsin, or (2) if attacked by trypsin its decomposition products (presumably amino-acids) still contain the group oxidizable with light production, or (3) it is not protein at all. I have been unable to oxidize with light production various mixtures of amino-acids (from tryptic digestion of beef and casein, or the acid hydrolysis products of luciferin itself) by means of luciferase, and consequently am led to believe that Cypridina luciferin is either a new