XLVI. A STUDY OF SOME BIOCHEMICAL TESTS.

No. 2. THE ADAMKIEWICZ PROTEIN REAC-

TION. THE MECHANISM OF THE HOPKINS-
COLE TEST FOR TRYPTOPHAN. A NEW
COLOUR TEST FOR GLYOXYLIC ACID.
BY WILLIAM ROBERT FEARON.
Mackinnon Research Student.
From the Biochemical Laboratory, Cambridge, and the Physiological
Laboratory, Trinity College, Dublin.
(Received May lth, 1920.)
IN 1874 Adamkiewicz published [1874] a description of the colour phenomena
due to the action of strong sulphuric acid on egg-albumin. In some experi-
ments he used acetic acid as a solvent, and he observed in many such cases
that a purple colour was produced on the addition of the sulphuric acid.
He considered this colour to be due to "a special influence of the acetic acid
upon the colour already produced in the sulphuric acid protein solution."
Udranszky [1888] described the test as a furfural,reaction, which explana-
tion was generally accepted till Hopkins and Cole [1901, 1, 2] published the
first of their well-known papers, in which they showed that the colour was due
to an impurity in the acetic acid. They isolated glyoxylic acid from several
samples of acetic acid and showed that in the presence of sulphuric acid it
gave a purple colour with certain proteins, the facts thus discovered leading
later to the isolation of the amino-acid, tryptophan, which is responsible for
the reaction.
Cole [1903] investigated the various protein colour tests with aldehydes,
or bodies that gave rise to aldehydes, and proved that the majority of these
reactions depended on the presence of tryptophan. He drew attention to the
importance of the presence of an oxidising agent in the production of the
colours.
Hopkins and Cole's explanation was criticised by Rosenheim [1906], who
described a colour reaction between tryptophan and formaldehyde in sul-
phuric acid containing a trace of an oxidiser, which was similar to that
given by glyoxylic acid. Rosenheim considered that the Hopkins-Cole test
depended on the presence of formaldehyde in the acetic acid or arising from
the reagents used. He remarked upon the ease with which glyoxylic acid is
decomposed at high temperature into formaldehyde and carbon dioxide.
 THE ADAMKIEWICZ REACTION 549
Dakin [1907] investigated the tests, and found that glyoxylic acid prepared
by different methods gave the test. He thought it "hard to believe that these
uniformly positive results with different specimens of glyoxylic acid are due
to impurities." Glyoxylic acid preparations in many cases gave the test
directly, formaldehyde always required an oxidising agent. Dakin discussed
the possibilities of the production of glyoxylic acid from formaldehyde by an
aldortype of condensation followed by oxidation. Kondo (1906) described
a test for indole similar in many respects to the Rosenheim test. Formalde-
hyde with indole gives a crimson colour in presence of strong sulphuric acid.
Research was done in the following years on the general chemistry Qf
bodies related to tryptophan, which will be referred to later, as it does not
directly concern the Hopkins-Cole reaction [Blumenthal, 1909].
In 1913 Homer published the results of work on aldehyde derivatives of
tryptophan [1913]. She believed that the colour bodies were due to an inter-
action between the aldehyde and the imino-group of the tryptophan. She
concluded from the general behaviour of these bodies that the Hopkins-Cole
test was due to formaldehyde.
In the same year Mottram [1913] described a failure of the glyoxylic test
owing to an excess of oxidising agents in the sulphuric acid used. This was
confirmed by Breidahl in 1915, who stated that such acids could be made
suitable for the test by the addition of small quantities of reducing agents
[1915]. These papers emphasised the fact that a trace of an oxidising agent is
necessary in the majority of cases; if the oxidising agent be too strong or too
weak the test will not succeed.
Voisenet in 1918, from similarities in appearance and behaviour towards
reagents, concluded that the colour produced in the Hopkins-Cole and the
Rosenheim tests was the same substance, and that in each test it was due to
formaldehyde [1918].
The following research was undertaken originally with the object of finding
a satisfactory method for estimating tryptophan colorimetrically. Although
such a method has not yet been found, some information has been obtained
about the mechanism of the tryptophan colour tests. The first question investi-
gated was the nature of the aldehydes concerned in the Hopkins-Cole and
Rosenheim tests with a view to deciding the long-standing glyoxylic acid and
formaldehyde controversy.
The Hopkins-Cole Reaction in the absence of Formaldehyde. If formalde-
hyde were liberated from the glyoxylic acid during the process of the test
it should be possible to detect its presence, since there are colour tests for
formaldehyde in presence of sulphuric acid that are more delicate than the
reaction between formaldehyde and tryptophan. Should formaldehyde be
detected under the conditions of the Hopkins-Cole reaction, it probably
enters into the test. On the other hand, should the test be obtained under
conditions that are such that formaldehyde cannot be detected in the reacting
solutions, it is strong evidence that the reaction is due to glyoxylic acid.
550 W. RI FEARON
Glyoxylic acid is more stable in strong sulphuric acid than has been
generally supposed. When calcium glyoxylate was added to concentrated
sulphuric acid and the mixture warmed gently, decomposition did not begin
until the temperature rose above 500. Carbon dioxide was then evolved, and
formaldehyde was recognised in the solution by means of the thiophen test
[Fearon, 1918]. The Hopkins-Cole test may be carried out at temperatures
below that of the decomposition of the glyoxylate and in the absence of water,
by using solid tryptophan and calcium glyoxylate. The characteristic colour
develops under conditions where formaldehyde can be shown to be absent.
(In passing, it may be noted that if the sulphuric acid be contaminated with
nitrous acid an ivy-green colour may be formed on warming the solution, this
is due to an entirely different reaction which will be discussed in a later com-
munication.)
Various samples of acetic acid which gave the reaction were tested for
formaldehyde, alone and under the conditions of the test, with negative
results in nearly every case.
To detect the presence of formaldehyde in excess of sulphuric acid the
following tests were employed.
The Thiophen Test (1918). This test will show one part of formaldehyde
in half a million parts of sulphuric acid. The colour with formaldehyde is
distinctive and much more delicate than the colour with glyoxylic acid.
The delicacy of Rosenheim's test for tryptophan is about 1 in 200,000 at
most.
The Pyrogallol-aldehyde Reaction, described in the present paper.
This is an extremely sensitive test for aldehydes in presence of sulphuric
acid; it will show 1 part formaldehyde in over 1,000,000 parts water, and may
be used to show the presence of formaldehyde in mixtures containing glyoxylic
acid.
As the result of a series of experiments on the Hopkins-Cole test under
various conditions and with different preparations of glyoxylic acid, it is con-
cluded that the test can be obtained in the absence of formaldehyde.
The Rosenheim Test in the absence of Glyoxylic Acid. Although glyoxylic
acid gives many of the general aldehyde reactions, there is no simple colour
test for distinguishing it from other aldehydes in presence of sulphuric acid.
The test current in the text-books is simply the reverse of the Hopkins-Cole
reaction and open to the same criticism of uncertainty as regards the aldehyde.
This test was introduced by Eppinger [1905] and modified by Schloss [1906].

A NEW COLOUR TEST FOR GLYoXYLIc AcID.

The reagent is 1 % pyrogallol in pure concentrated sulphuric acid, which
must be free from any nitrous contamination, otherwise the solution turns
brown and is useless. The reagent will not keep for long, after a few days a
white condensation product is deposited.
 THE ADAMKIEWICZ REACTION5 551
When a trace of calcium glyoxylate, or a couple of drops of reduced oxalic
acid solution are added to the reagent, a deep blue colour develops on gentle
warming. Dilution with water changes the blue to a deep carmine which
becomes blue again on addition of excess of sulphuric acid.
The blue colour is not given by oxalic acid or any of its simple derivatives
which do not give rise to glyoxylic acid under the condition of the test.
With the aliphatic aldehydes the reagent gives various shades of red. The
presence of glyoxylic acid along with. formaldehyde can be shown by the
production of a purple due to the formation of two colours. The test is useful
for detecting aldehydes in ether, as mentioned later. Syrupy phosphoric acid
may be used instead of sulphuric acid as a condensing agent, in which case
a red colour is obtained with all the aldehydes on warming.
The delicacy of the test is greatest for formaldehyde; with glyoxylic acid
it will show less than one part in a thousand, with weaker solutions the colour
takes some time to develop.
The chemistry of the test has only been briefly investigated as yet. The
coloured condensation compounds are probably derivatives of xanthene.
When separated from the sulphuric acid they show many of the properties of
xanthydrol, such as forming typical insoluble compounds with urea in acid
solution. The colour is probably due to the formation of the pyrone ring
[Palazzo and Onorato, 1905, and references].
R*C H R*C*H R-C

HO OH HO JOH HO OH HO O
OH OH HO 0 OH HO 0 OH
Leuco-compound Pigment
Derivative of tetrahydroxyxanthene
On applying the pyrogallol-reagent to solutions of formaldehyde in sul-
phuric acid in the presence of oxidising agents, under the conditions of the
Rosenheim test, no evidence whatever could be obtained of the formation of
glyoxylic acid, though on addition to the mixture it could be detected. This
supports Rosenheim's view that the test depends on formaldehyde alone.
From a purely chemical aspect, it is not improbable that both formaldehyde
and glyoxylic acid should react with tryptophan to give distinct derivatives,
though subsequent investigation showed that the glyoxylic derivative readily
lost carbon dioxide and changed into the corresponding formaldehyde de-
rivative as might be expected. The fact that the Hopkins-Cole test is con-
sidered by many investigators to be more reliable than the Rosenheim test
would be explained by the greater tendency, as subsequently observed, of
the formaldehyde to react still further with the condensation products. A
disadvantage of both tests is the local heating at the junction of the liquids;
Bioch. xiv 36
5562 W. R. FEARON
this can be greatly reduced by employing glacial acetic acid as a solvent for
the glyoxylate or the formaldehyde as the case may be. As will be shown
later, an oxidising agent is necessary in both tests, this is usually supplied by
the sulphuric acid [Mottram, 1913].
THE MECHANISM OF THE HOPKINS-COLE REACTION.
Having concluded from the previous experiments that both glyoxylic
acid and formaldehyde gave coloured derivatives with tryptophan, which is
confirmed by subsequent research, the nature of the reaction was next in-
vestigated. Whilst it is well known that indole and its methyl derivatives
yield with different aldehydes a large number of coloured derivatives some of
which have been investigated, the aldehyde reactions of tryptophan are known
chiefly as qualitative colour tests and the exact nature of the compounds
formed never seems to have been fully investigated. This is due partly to the
difficulty of obtaining tryptophan in quantities sufficient for extended chemical
research and partly to the difficulties of isolating the coloured compounds or
their precursors in a stable form.
A number of colour reactions between tryptophan and various aldehydes
have been described [Rohde, 1905; Sasaki, 1910], but for the purposes of the
present paper it will be sufficient to consider three only: formaldehyde,
glyoxylic acid and benzaldehyde. It will also be necessary to consider the
reactions of these aldehydes with some substances related to tryptophan,
namely: indole, scatole, and carbazole.

PREPARATION AND ISOLATION OF THE ALDEHYDE CONDENSATION DERIVA-

TIVES OF INDOLE, SCATOLE, TRYPTOPHAN AND CARBAZOLE.
Condensation. Strong sulphuric acid -is quite unsuited for use as a con-
densing agent for these substances. It may contain a large number of active
impurities; it acts as an oxidising agent in a variable manner, the heat it is
apt to develop may destroy the condensation products. Fischer [1889]
employed alcohol and hydrochloric acid gas for his work on the indoles.
Homer and others used zinc chloride. Neither of these reagents was found to
be quite satisfactory, and most of the present work was done by means of
pure glacial acetic acid saturated with hydrochloric acid gas. This gave good
results as a condensing agent, and is free from the objections to sulphuric
acid. Phosphoric acid was used in a few particular cases. The -glacial acetic
acid was purified by fractionation until it no longer gave a colour with
tryptophan or any of the aldehyde reagents. When pure it will keep for a
long time if moisture and light are excluded, and the bottle be kept filled, as
the changes leading to the formation of aldehydes appear to occur in the
vapour over the liquid.
Acetic acid so prepared will be referred to in the present paper as "pure"
acetic acid.
 THE ADAMKIEWICZ REACTION 553
The hydrochloric acid gas for saturation was prepared by the action of
pure sulphuric acid on pure dry sodium chloride. If the acid be badly con-
taminated with nitrous impurities some of them may pass over into the con-
densation mixture and spoil the preparation.
The aldehydes were used in the following forms:
Glyoxylic acid as the calcium salt. (CHO .CO0)2 Ca, 2H20.
Formaldehyde as paraformaldehyde. The ordinary solutions of "formol"
are unsuitable owing to the presence of methylal and other derivatives
[Werner, 1917].
Benzaldehyde as a solution in "pure" acetic acid.
One of the first facts observed during the preparation of the colour con-
densation compounds by the above method was that an oxidising agent was
necessary in every case. The condensation products as formed are colourless
or pale yellow; on oxidation the characteristic colour is produced. This fact
escapes notice when sulphuric acid is used as a condensing agent since it acts
as an oxidiser as well.
For the oxidation a fresh 1 % solution of H202 in "pure" acetic acid was
used. In some instances the oxidation was allowed to proceed spontaneously
in the air overnight.
Notes on the Condensation. It is best to use an excess of the solvent and to
pass in the HC1 gas slowly, shaking at intervals. After the solution has been
saturated, gas is passed in for an additional quarter of an hour. A few cc. of
the solution are withdrawn in a test-tube and one drop of the 1 % H202
added, and the tube well shaken, when the colour should develop in a few
minutes. The H202 is added carefully up to the point of maximum depth of
colour, too much will discharge the colour altogether. The approximate quan-
tity of H202 is noted. Passage of HC1 into the mixture is continued until no
increase in depth of colour is seen on removing and testing the samples. It
is sometimes useful to pass a current of air into the mixture to remove excess
of HC1 before the final oxidation.
The calculated amount of 1 % H202 is then added to the mixture, which
is well shaken till oxidation is complete. It is tb4en poured into a measuring
cylinder, three times its volume of purified ether added, and the mixture
shaken. The pigment will be precipitated in small flakes. If precipitation is
not complete, another volume of ether may be added, but it is advisable not
to exceed four volumes if a pure preparation is required, as otherwise the
other substances may be precipitated.
Isolation of the Condensation Products. This proved to be one of the most
difficult parts of the investigation. It was necessary to free the products from
unchanged aldehyde, and indole derivatives, and from the condensing agent.
Fractional precipitation with pure ether was finally adopted and gave satis-
factory results. It was found that if three to four volumes of ether were
added to one volume of the acetic acid solution of the pigment and the mixture
well shaken, the latter came down as a flocculent precipitate and the aldehydes
36-2
554 W. R. FEARON
and unchanged indole derivatives were left in the ethereal solution, as was
also the calcium chloride which was formed when calcium glyoxylate was
employed.
The pwty of the ether is of great importance. It was found that merely
washing and drying it was not sufficient. It should also be freed from
aldehydes and related substances by distillation from sodium hydroxide,
after having been left in contact with it for at least 24 hours. If the ether
contains aldehydes it will react with the precipitated pigment on standing,
converting it into a greenish paste, which may resinify on drying.
For this reason the coloured compounds should be filtered from the ether
as soon as they are precipitated. The pyrogallol test is.useful for detecting
aldehydes in samples of ether.
List of Pigments isolated. The following are the colour derivatives obtained
by the condensation of indole, scatole, tryptophan and carbazole with the
three aldehydes, formaldehyde, glyoxylic acid and benzaldehyde respectively.
(The figures placed below the colour denote the number of molecules
interacting.)
ALDIEHYDE CH20 CHO. COOH C6H5. CHO
Indole Red Red Red
21+F 2I+G 2I+B
Scatole Red Red Purple
2S+F 2S+G 2S+B
Tryptophan Red Red Blue
2T+F 2T+G 2T+B
Tryptophan Blue Blue
2T + 3F 2T + 3G
Carbazole Blue Blue Blue
2C+F 2C+G 2C+B
No attempt has been made to describe the shade in the above list, anything
varying from rose to carmine has been grouped as "red" and similarly for the
other colours. The qualitative differences are quite definite in many cases.
Many other aldehyde condensation derivatives were prepared, but the
above are the most important in connection with the discussion which follows
on the chemical constitution of these bodies.
General Properties of the Condensation Derivatives. Both the leuco-compounds
and the pigments can be obtained by the ether precipitation method. The
leuco-compounds are unstable and soon oxidise to form the pigments.
As precipitated the pigments are all hydrochlorides of nearly colourless
bases. The pigments are soluble in strong mineral acids, in the lower aliphatic
acids, in esters such as ethyl chloride, in phenol to a less extent. The solu-
bility varies somewhat with the different pigments. The red and blue deriva-
tives of tryptophan are hydrolysed at once when water is added to their
solution in the HCl-acetic condensation mixture. The scatole and carbazole
pigments are much more stable, the carbazole blues being precipitated on the
addition of excess of water. Addition of alkali discharges the colour of the
 THE ADAMKIEWICZ REACTION 555
pigments, as might be expected; in most cases the base is precipitated from
the alkaline solution. The pigments were purified by repeated solution in
" pure" acetic acid and precipitation with pure ether. They were then rapidly
dried in a current of air. When pure they are amorphous deeply coloured
powders.
Melting Points. These were all high, in the region of 3000, depending on the
rate of heating. The tryptophan reds and blues tend to lose CO2 on heating.
If the pigments are dissolved in absolute alcohol and treated with dry
NH3, they are decolorised and gradually precipitated from solution as bases,
which in some cases are microcrystalline. Addition of strong acid to the bases
restores the original colour of the pigments. The bases were also obtained
during an attempt to crystallise the pigments from alcohol. The alcohol
apparently combined with the HCl of the pigment and removed it to form
ethyl chloride.
Molecular weights. The molecular weights of the bases were determined by
the freezing point method with comparatively concordant results, using
acetic acid or phenol as solvents. It has been shown that compounds of
related indigotin type do not tend to form associated molecules in these
solvents [Beckmann and Sabel, 1906].
The pigments are the salts of these bases.
Analyses. For the present preliminary'work, the only element estimated
was nitrogen. This was done by a modified Kjeldahl method, using mercury
as catalyst instead of copper, and drawing the fumes by means of a Folin
absorber into a wash-bottle, so that a control could be kept on any nitrogen
which might escape combustion. These bodies can be analysed satisfactorily
by the Kjeldahl method provided that the heating is continued for a con-
siderable time after the mixture has cleared [cf. Dakin and Dudley, 1914].
The results of the nitrogen estimations were used to check the formulae
of the condensation derivatives deduced from the molecular weights, sufficient,
tryptophan not being at my disposal for a complete analysis of the various
pigments.
CHEMICAL CONSTITUTION OF THE CONDENSATION DERIVATIVES.
Up to the present time no structural formulae appear to have been sug-
gested for the tryptophan pigments. From the preliminary examination of
these bodies it is possible to assign formulae to them which will satisfy the
requirements of the Hopkins-Cole and Rosenheim tests without attempting
the more ambitious task of giving a complete account of the chemistry of
these interesting condensations.
From determinations of the molecular weights of the colour bases, it
appears that all the pigments prepared are the salts of compounds formed by
the condensation of two molecules of the indole derivative with one of the
aldehyde, except in the case of tryptophan blue, where condensation occurs
with three molecules of the aldehyde.
-5-56 5 W. R. FEARON
In theory the aldehyde may unite with the indole derivative in three ways:
(1) by replacing hydrogen in the benzene ring,
(2) by replacing hydrogen in the pyrrole ring,
(3) by replacing hydrogen in the imino-group.
The fact that all the pigments arise from oxidation of leuco-condensation
bodies is, on the quinonoid theory of colour and constitution, strong evidence
for union in either of the ways (1) or (2). The subsequent removal of hydrogen
from the aldehyde residue and from one of the imino-groups occurs on oxida-
tion and leads to the formation of the double bonds".
HN.C80H5-CHRC8H5NH N =05H4=CR-C5H5NH
leuco-compound pigment.
Fischer [1889] gives the following formula for the aldehyde derivatives
of a-methylindole (methylketole), the union H
taking place by the aldehyde replacing the hy- H cR
drogen atoms of the ,- or 3-carbon atom of
the pyrrole groups.
Freund and Lebach have prepared and in- . X
vestigated several aldehyde derivatives of a- NH NH
methylindole (methylketole). They accept the
Fischer formula, which, as they point out [1905], accounts for the production
of a colour by oxidation.
Although Fischer's formula is satisfactory when applied to the aldehyde
condensation derivatives of indole and methylindole, it is not so applicable
to the 3-substituted indoles, such as scatole and tryptophan. Here the
3-carbon is already in union with another group and it is unlikely that under
the conditions of the condensation the aldehyde displaces these groups, and
even if it did, this would not explain the fact that the scatole-benzaldehyde
compound is deep purple while the corresponding indole-benzaldehyde com-
pound is red.
Again, if condensation did occur in the pyrrole ring in scatole-benzaldehyde
purple, it is hard to account for the existence of the leuco-precursor. Freund
and Lebach accept the formula for the aldehyde derivatives of scatole on the
lines of Fischer, though they note its inability to explain the formation of the
colour on oxidation.
If the aldehyde condenses with the carbons in the benzene ring, instead of
the pyrrole ring, all these objections can be overcome. The existence of both
leuco-compounds and pigments can be easily'explained in the case of all the
substances examined, and the general properties of the pigments can be
accounted for.
The di- and tri-phenylmethane dyes in common use -are quinonoid salts of
bases of this type. One of their disadvantages is their sensitiveness to alkalies,
1The formulae are printed so as to show that the nitrogen of the imino-group is in union with
two different carbon atoms.
 THE ADAMKIEWICZ REACTION 567
which reconvert them into the colourless bases. It will be seen that the pig-
ments described in the present paper have many properties in common with
the di-phenylmethane dyes.
The carbazole pigments are similar in many respects to the indole deriva-
tives. Here condensation with the aldehyde has almost certainly taken place
in the benzene ring, as the pigments are produced bv oxidation of leuco-
compounds. The greater stability of the carbazole blues may be explained as
being due to the absence of the mobile hydrogen atom of the 2-unsuibstituted
indoles, which is capable of changing its position from the nitrogen to the
neighbouring carbon atom, and vlice versa. The carbazole ring is stable and the
imino-hydrogen is held by the presence of the benzene rings on either side.
This tendency on the part of the indoles to undergo tautomeric change has
been studied by various investigators [Weissgerber, 1913]. In assigning
formulae to the compounds mentioned in the present paper, the hypothesis
that aldehyde condensation takes place in the benzene ring has been applied.
According to this, when the indole derivative reacts with the aldehyde in the
presence of a condensing agent, water is eliminated and the aldehyde links
up two molecules of the indole derivatives by uniting with the carbon atoms
of their benzene rings to form the leuco-compound, this on oxidation loses
one hydrogen atom from the aldehyde residue and one from an imino-group,
with the production of the quinonoid configuration in the molecule and the
formation of a colour.
These pigments, especially the benzaldehyde derivatives, and their leuco-
precursors are comparable on this hypothesis with malachite green.

EXPERIMENTAL.
The Scatole Derivatives1.
(1) Scatole formaldehyde red. NH.C8H4 (CH3).CH: CrH4 (CR): NX
M.W. of base, 272.
(2) Scatole glyoxylic red. NH.O8H4 (CH3).C (COOH): Cr.H4 (CH.): NX
M.W. of base, 316.
(3) Scatole benzaldehyde piurple. NH.C8H4 (CH3).C (C6H5): C8H4 (CH) NX
M.W. of base, 348.
Two molecular parts of scatole and one of the aldehyde are dissolved in
excess of "pure" acetic acid and condensed by means of dry hydrochloric
acid gas as described. The condensation and subsequent oxidation should be
done at a temperature not above 200, otherwise the scatole itself may be oxi-
dised with the formation of a red-brown pigment. The condensation compounds
are precipitated by ether and purified as already described. They are amor-
phous intensely-coloured powders, fairly soluble in dilute HCI and H2SO4.
l The formulae are written so as to show the union of the nitrogen atom with two different
carbon atoms. X represents the acid which combines with the pigment base to form the coloured
salt. In the present paper X represents HCO.
558 W. R FEARON
The benzaldehyde derivative is the one least hydrolysed by dilution with
water. They are all decolorised and precipitated by alkalies, boiling with
which reconverts them in part to the original scatole. As prepared by the
above method the yield is nearly quantitative.
General formula suggested for these compounds:
r--
NH.C8H4 (CH2).CHR.C8H4 n +0
(CH,). NH - NH.08H4 (CH.). CR: C8H4 (CH2): NX
NH NH N NH

HaG
C H \c H s CRi c H3
Ho'R R
Leuco compound Pigment base
The Tryptophan Derivative8 r (
(4) Tryptophan formaldehyde red. CHH 2COOH:N.HCI
M.W. of base (average), 418.
CsH4OH2OH (NH2) COOH. NH. HCI
I
~-~
(5) Tryptophan glyoxylic red. . CO
C8H4CH2CH (NH2) COOH-N.HC1
M.W. of base, above 420. HCHCH (NH2) COOH.IH.HC1

(6) Tryptophan benzaldehyde blue.

M.W. of base, 494.
(C6HE)HCH)\ C' H,CH,,CH (NH,,)
( -
COOH: HC IC
a
H.H.HC1
Two molecular parts of tryptophan are condensed with one part of the
aldehyde as previously described. The preparation must be carried out in
one stage, as the compounds may undergo decomposition if left for long in the
condensing reagent. The temperature should not rise above 150. If the alde-
hyde be added rapidly at the start in preparing tryptophan formaldehyde
red, it may interact with the NH group to form a methylene derivative
-N = CH2, but this does not affect the general formula type given above.
This methylenation of the amino-group is an intermediate stage in the forma-
tion of the blue compounds to be described later.
The tryptophan reds are the most unstable of the pigments prepared.
They readily lose C02 on heating, and for this reason it is difficult to isolate
tryptophan glyoxylic red, although there is evidence for its existence in solu-
tion. It is very difficult to obtain these derivatives in a pure state, apart
from tryptophan benzaldehyde blue. The product from the ether precipitation
appears to be a varying mixture of the true tryptophan condensation product
with condensation products from decarboxylated tryptophan and tryptophan.
in which the amino-group has been methylenated. However, there seems to
be little doubt that all these reactions are modifications of the condensation
along the lines of the formulae given.
Tryptophan benzaldehyde blue is one of the more stable pigments, and
seems to be well suited for colorimetric work as well as being a good reagent
 THE ADAMKIEWICZ REACTION 559
for qualitative tests. Cole in 1903 drew attention to the clear colour pro-
duced. As obtained by the above method, the pigment is a deep blue
powder with a faint lustre. It melts just above 3100, and rapidly decomposes
according to the rate of heating. It is soluble in the solvents mentioned, but
has not been crystallised from any. Like the previous tryptophan devivatives
it is largely hydrolysed on diluting the acetic acid solution with water.
Readings of the molecular weight of the pure base in acetic acid were on the
average below 500 and just above 490, which corresponds to the theoretical
value, 494, required by the formula.
(7) Tryptophan formaldehyde blue. C12HION2. Hal
M.W., base 1, 354 (average). CHO
M.W., base 2, 444 (average). C12H1.N2.HCl
(8) Tryptophan glyoxylic blW. Base 1 is derived from the decarboxylated tryp-
M.W., variable. Above 400. tophan, Base 2 from ordinary tryptophan.
The Tryptophan Blues. It was confirmed at an early stage in the present
research that the Hopkins-Cole test and the Rosenheim test each gave two
distinct colours with tryptophan according to the amount of aldehyde used,
a fact previously noticed by Mottram [1913] in the former test. With the
minimum amount of aldehyde and the test-tube kept cool, the colour in each
test is carmine. Using more aldehyde and allowing the temperature of the
mixture to rise, the colour is violet to blue. On studying the reaction using
"pure" acetic acid and HCI gas, it was found tha-t tryptophan is able to give
two distinct coloured derivatives when condensed with formaldehyde or with
glyoxylic acid. One class has been briefly referred to above as the tryptophan
reds.
When two parts of tryptophan are coiidensed with three parts of formalde-
hyde or of glyoxylic acid at a temperature between 500 and 700, the resulting
compound on oxidation is deep blue. Also, when tryptophan red was dis-
solved in "pure" acetic acid saturated with HCI and left for a couple of days,
it gradually changed into the blue compound. This change did not occur in
acetic acid alone. The change is accompanied by the liberation of some free
tryptophan, as was shown by extracting the mixture with ether containing
HCI.
Investigation of the molecular weight of the HCI-free base indicates that
the pigment is a di-molecular derivative of tryptophan like the other sub-
stances investigated. It was thought probable that the additional molecules
of the aldehyde reacted with the amino-group. This was shown to be so by
the preparation of some indole-hydroxypropionic acid by the action of the
calculated amount of NaNO2 on tryptophan in HCI solution and removal of
any excess of HNO2 by means of urea. This substance only gave a red aldehyde
derivative under the above conditions. From this it was concluded that for
the production of the blue -pigment interaction was necessary between the
aldehyde and the amino-group of the tryptophan. The first effect of such
reaction would be to produce a methylene derivative of the amino-group,
560 W. R. FEARON
-N= CH2 which would not pass into a methyl group unless the aldehyde
were in considerable excess or a reducing agent were present [Weiner, 1917].
Neither of these conditions obtains in the preparation as described above.
It is, however, h'ard to understand why the methylenation of the amino-
groups should change the red pigment into a blue one, and there is evidence
that tryptophan red may contain such substances without alteration of colour.
It is probable that there is an additional condensation. Bearing in mind the
fact that the corresponding aldehyde derivatives of carbazole are all deep
blue in colour, and that the H atom attached to the a- or 2-carbon of indole
is labile, it does not seem unreasonable to suppose that a condensation might
occur between the carbon of the methylene substituting group in the trypto-
phan and the carbon atom of the indole group to form a three-ring molecule
of the carbazole type, having a pyridine nucleus instead of one of the benzene
nuclei. Such a body might be formed as follows:
CH20 + H20
NH CH2 NH

H2N
COOH-C-H < COOH-CH
Tryptophan Derivative of 4-carboline
This ring should resemble carbazole in many of its properties. Perkin and
Robinson [1919] have found a similar ring in the harmine series. They have
shown that the base C12H10N2 prepared by Hopkins and Cole [1903] by the
oxidation of tryptophan with ferric chloride is harman [Fischer, 1901].
Consequently there is convincing evidence that closure of the tryptophan
side-chain can occur. Perkin and Robinson advance the theory that harmine
originates in nature by an acetaldehyde condensation with hydroxytryptophan,
a substance already identified by Abderhalden and Baumann [1908]. Addi-
tional evidence for the closure of the tryptophan side-chain has been obtained
in connection with other reactions which I hope to describe in a future com-
munication.
On this hypothesis, the formula for tryptophan blue is:
H2 NH N-HCI H2
HNK3fJ N~ilQ
iO H*
COOH HC COOH*
H2 I H2
H
The quinonoid salt of the colourless base
* Or H, if decarboxylated.
There appears to be very little difference between tryptophan blue pre-
pared from formaldehyde or from glyoxylic acid. Loss of CO2 occurs during
the preparation and the final pigment obtained is probably the same in each
case, Unless the condensation be carried out thoroughly, the pigment pre-
 THE ADAMKIEWICZ REACTION ,561
cipitated by the -ether is very impure, containing, as well as the blue derivative,
a varying amount of purple material which probably represents an earlier
stage in the condensation when the side chain of only one of the tryptoplan
molecules has closed up. The pigment should not be precipitated until it has
stood for some days, by which time most of the substances will have con-
densed and oxidised spontaneously to the blue.
The blue is precipitated more readily by ether than the red or the purple
colouring matters, but, even so, it is almost impossible to separate them com-
pletely.
Tryptophan formaldehyde blue resembles the benzaldehyde blue in some
respects and the carbazole blues in others.
When purified it is moderately stable and has been obtained in solution
in dilute HCl. Like the other tryptophan derivatives, it melts in the region of
3000, the exact temperature depending somewhat on the rate of heating.
As usually carried out, the Hopkins-Cole test results in a mixture of
tryptophan red and blue, the proportions depending on the conditions. For
this reason it is unsatisfactory as the basis of colorimetric work, as two solu-
tions of the same strength may give different shades if the temperature be at
all different. Fasal [1912] noticed this difference in tint during his work on
the estimation of tryptophan, but considered it entirely due to the amount
of colouring matter formed.
For colorimetric work an aldehyde which only reacts in one way with the
amino-acid is advisable, if not essential.
The above formula is in agreement with the experimental facts. It
accounts for the existence of the leuco-form; for the three molecules of
formaldehyde required; and for the colour and general properties of the
pigment.
Only aldehydes that react readily with the amino-group, and are also
capable of condensing with the indole nucleus, can give the carbolinine blue
compound. Tryptophan benzaldehyde and salicyl aldehyde blues are simple
condensation products and do not show any evidence of a closure of the
tryptophan side-chain.
A study of the freezing-points of the bases obtained by removal of the
HCI from the pigments indicated that the molecular weights of the formalde-
hyde and glyoxylic blues were probably the same, 352. Some preparations
had higher molecular weights of the order of 400, which would be explained.
bv the preparations not being fully decarboxylated. The pigment ultimately
obtained in each case is one in which CO2 has been lost by all the COOH
groups. It will be seen that the condensation product formed by the closure
of the tryptophan side-chain is derived from the ring nucleus:

NH
,562 W. RI FEARON
Perkin and Robinson, who -have described a similar ring, suggest the
name " carboline " for this structure, indicating an analogy both to carbazole
and quinoline. According to this nomenclature, tryptophan blue would be a
carboline blue in which the C atoms of the pyridine nucleus are fully satu-
rated, such a ring might be termed "carbolinine" by analogy to the
relation between pyrrole and pyrroline. Tryptophan blue then would be "di-
4-isocarbolininemethane," and the leuco-precursor would be, "di-4-carbo-
lininemethane."
THE CARBAZOLE CONDENSATION DERIVATIVES.
Several of these bodies were prepared: the following are of most interest
in connection with the present paper.
(9) Carbazole formaiekyde blue.
(10) Carbazole glyoaylic blue.
(11) Carbazole benzaldehyde blue.
Only a brief reference will be made to these bodies.
Their preparation and principal properties resemble those of the pigments
previously described. They are of different shades depending on the aldehyde
and the solvent. They are more stable than the other pigments and will keep
for a long time in aqueous solution. They are precipitated by alkalies. The
melting points are high, the benzaldehyde blue beginning to decompose at
280.
The general formula is probably:
R
c

NH N NH
Carbazole ClHN Pigment base
This reaction is a good general test for the presence of carbazole; benz-
aldehyde is perhaps the most convenient aldehyde to work with, but the
colours are good with all three.
A SIMPLE QUALITATIVE REAGENT FOR TRYPTOPHAN AND THE INDOLES.
The reagent is 10 % salicyl aldehyde in alcohol, which must be free from
acetone. A small quantity of the liquid under examination is mixed with
excess of strong hydrochloric acid. Four or five drops of the reagent are added
and the mixture gently warmed for a few minutes, a drop of 10 % hydrogen
peroxide is then added and the warming continued. An intense blue denotes
tryptophan, a deep purple denotes scatole, while indole gives a bright carmine.
If too much peroxiide be added the colours are destroyed, otherwise, they are
very stable and are not discharged on dilution with water. The reagent will
 THE ADAMKIEWICZ.REACTION 563
give a deep blue with less than 0.1'% tryptophan; with weak'solutions care
must be taken in the oxidation. By modifying'the strength of the reagents
the test can be made very delicate. In properties the pigments resemble
very closely the benzaldehyde derivatives, their greater stability towards
water is comparable with that of the ,malachite greens derived from sub-
stituted benzaldehyde. The fastness of these lattercdyestuffs-depends on the
presence of a sulpho-group in the ortho-position to the methane carbon, and
one is justified in assuming the hydroxyl of the salicyl aldehyde to operate
in a similar manner.
Compare the behaviour of salicyl aldehyde and benzaldehyde when con-
densed with acetone in presence of hydrochloric acid.

SUMMARY.
(1) A new colour test for glyoxylic acid is described, depending on its
interaction with pyrogallol in presence of sulphuric acid.
(2) The Hopkins-Cole test is -due to glyoxylic acid, the Rosenheim test
is due to formaldehyde. In each test at least two distinct colour derivatives
of tryptophan are formed, a red and a blue.
(3) A method is described for the preparation of the aldehyde derivatives
of indole, scatole, tryptophan and carbazole, and a number of these have
been isolated.
(4) The possible formulae of these substances is discussed, and a general
type of structure is suggested.
(5) The closure of the tryptophan side-chain is put forward as an explana-
tion of the formation of tryptophan formaldehyde blue.
(6) Salicyl aldehyde is recommended as a useful qualitative reagent for
the indole derivatives.
I should like to express my obligation to Professor F. G. Hopkins for his
kind advice and interest in the research, also to Professor E. A. Werner for
having read the proofs of this paper.
I received much help in the earlier stages of the work from Mr Robin Hill,
of Emmanuel College, Cambridge.
The expenses of the research were defrayed by the Mackinnon Research
Studentship of the Royal Society.

REFERENCES.
Abderhalden and Baumann (1908). Zeitsch. phy8iol. Chem. 55, 412.
Adamkiewicz (1874). Pflfiger'8 Arch. 9, 156.
Beckmann and Sabel (1906). Ber. 89, 3140.
Blumenthal (1909). Biochem. Zeit8ch. 19, 521.
Breidahl (1915). Biochem. J. 9, 36.
Cole (1903). J. Phy8iol. 30, 311.
Dakin (1907). J. Biol. Chem. 2, 289.
Dakin and Dudley (1914). J. Biol. Chem. 17, 275.
564 W. R. FEARON
Eppinger (1905). Beitrdge, 7, 459; 466.
Fasal (1912). Biochem. Zeit8ch. 44, 392.
Fearon (1918). Biochem. J. 12, 179.
Fischer (1889). Ber. 19, 2988.
Fischer, 0. (1901). J. Chem. Soc. Ab8tr. i, 405.
Freund and Lebach (1905) Ber. 38, 2640.
Homer (1913). Biochem. J. 7, 116
Hopkins and Cole (1901, 1). Proc. Roy. Soc. 68, 21.
(1901, 2). J. Phy8iol. 27, 418.
- (1903). J. Phy8iol. 29, 451
Kondo (1906). Zeitsch. phy8iol. Chem. 48, 185.
Mottram (1913). Biochem. J. 7, 249.
Palazzo and Onorato (1905). Gazzetta, 35, 465, 476.
Perkin and Robinson (1919). J. Chem. Soc. 115, 933, 967.
Rohde (1905). Zeit8ch. phy8iol. Chem. 44, 161.
Rosenheim (1906). Biochem. J. 1, 233.
Sasaki (1910). Biochem. Zeit8ch. 23, 402.
Schloss (1906). Beitrige, 8, 455.
Udranszky (1888) Zeitsch. phytiof. Chem. 12, 395.
Voisenet (1918). Compt. Rend. 166, 789.
Weissgerber (1913). Ber. 46, 651.
Werner (1917). J. Chem. Soc. 111, 844.