THE CHEMISTRY OF NINHYDRIN

D. J. McCALDIN1
University of Nottingham, Nottingham, England

Received August SI, 1959

CONTENTS

I. Introduction...................................................................... 39
II. Synthesis and structure of ninhydrin.................................................. 39
III. General properties of ninhydrin...................................................... 40
A. Physical properties.............................................................. 40
B. Simple chemistry................................................................ 40
C. Reduction of ninhydrin.......................................................... 41
IV. Reactions of ninhydrin.............................................................. 41
Downloaded by UNIV DE SANTIAGO DE CHILE at 08:34:26:653 on May 23, 2019

A. Reaction with alkali....................................................·......... 41

B. Reaction with amino acids........................................................ 42
1. The theory of Ruhemann...................................................... 42
2. The theory of Harding and Maclean........................................... 43
3. The theory of Retinger........................................................ 43
4. The theory of Woker and Antener.............................................. 43
from https://pubs.acs.org/doi/10.1021/cr60203a004.

5. The theory of Moubasher and Schdnberg........................................ 44

6. The theory of MacFadyen..................................................... 45
7. The present theory........................................................... 45
C. Reaction with imino acids........................................................ 46
D. Reaction with amines............................................................ 47
1. Primary amines.............................................................. 47
2. Secondary amines............................................................
·
48
3. Tertiary amines.............................................................. 49
E. Reaction with pyrroles........................................................... 49
F. Reaction with ammonium salts.................................................... 49
V. Analytical uses of ninhydrin......................................................... 49
VI. References......................................................................... 50

I. Introduction hydrindene with p-nitrosodimethylaniline. He did not

obtain the expected monoanil but found that two mole-
Ninhydrin has been used for the detection of amino
acids and amines for almost fifty years, and many sug- cules of the nitroso compound had reacted to produce
the disubstituted hydrindone (II), which was hydro-
gestions have been made as to the mechanism of its
reactions. In this review it is suggested that a consistent NC,H«N(CH,)2-p O
interpretation can be placed on the majority of the
reactions of ninhydrin. The literature from 1910 to 1958
is discussed. The system of numbering the skeleton is
( NC,H,N(CH,)rp
shown below. II III

lyzed to ninhydrin (III). Ruhemann commented on the

poor yield in this experiment and unsuccessfully at-
tempted to improve it by using 2-hydrindone or 1,3-
diketohydrindene as starting material (63). Kaufmann
(34) had previously reported a synthesis of ninhydrin
Ninhydrin is also known as triketohydrindene hydrate, which involved oxidation of bis-1,3-diketoindanylidene
Ruhemann’s reagent, or more systematically as 2,2- (IV), but the compound that he obtained did not have
dihydroxy-1,3-indandione. the same properties as that isolated by Ruhemann. In
1933 Teeters and Shriner (80) published a synthesis
II. Synthesis and Structure of Ninhydrin
OO O
Ninhydrin was first prepared by Ruhemann (61) in
an attempt to oxidize 1-hydrindone (I) to 1,2-diketo-

1 Present address: Division of Pure Chemistry, National

Research Council, Ottawa, Canada. IV V
39
40 D. J. McCALDIN

involving the oxidation of 1,3-diketohydrindene (V) by in other polar solvents. When the solid is heated, it
selenium dioxide. The yield in this case was only about changes to a pink, red, or reddish-brown color at 125-
35 per cent; they attributed this to the formation of a 130°C., becomes a deep purple-red at 130-140°C., and
bimolecular product and to the recombination of the melts sharply with decomposition at 241°C. (21, 69,
starting material with the ninhydrin produced. The 89). The compound becomes red when exposed to sun-
reaction failed when other oxidizing agents, such as light and should be stored in a cool dark place. The
ceric sulfate, were used in place of the selenium dioxide. ultraviolet and visible absorption spectra of ninhydrin
Wanag and Lode (89) tried to prepare ninhydrin by and many of its derivatives have been reported by sev-
decomposing 2-nitro-l, 3-indandione under different eral authors (43, 44, 56, 58, 67).
conditions, but obtained only hydrindantin (VI), the
B. SIMPLE CHEMISTRY
0 0 0 0
V When ninhydrin is treated with thionyl chloride or
A)H
heated in vacuo, 1,2,3-indantrione (X) is obtained as
HO" dark red needles (69).
VI VII 0 0 0
reduction product of ninhydrin. However, by refluxing
2-bromo-2-nitro-l, 3-indandione in nitrobenzene, they
obtained yields of ninhydrin comparable to those ob-
tained by the previous method. In 1951, Gustowski X XI
(19) reviewed the published methods and suggested an If the blue-green solution of the trione in benzene is
improved preparation based on the work of Wanag and shaken with water, the carbonyl group in the 2-position
Lode. In the same year Khorkhlov, Shchukina, and is hydrated, the color is lost from the benzene layer, and
Shemyakin (35) showed that ninhydrin and hydrin- ninhydrin is produced. The trione is also produced in
dantin were formed when 1,4-naphthoquinone was sub- concentrated sulfuric acid solution, but if the tempera-
jected to long boiling in water at pH 7 in the presence ture is raised, bisindandione (XI) is obtained (67).
of oxygen. A 48 per cent yield of ninhydrin was claimed,
Heating ninhydrin in a current of air yields phthalic
based on the intermediate 1,2,3,4-tetraketotetralin
anhydride (69). Ninhydrin gives 2,2-dichloro-l,3-in-
(VII). Although there are a number of side reactions, a dandione (XII) with phosphorus pentachloride, and
method which requires such simple conditions might be when treated with hydrogen cyanide yields the 2-substi-
useful for large-scale preparations. tuted cyanohydrin (62). The 2-nitro compound (XIII),
On the basis of his experimental work, Ruhemann
0 O
(62) proposed 2,2-dihydroxy-1,3-indandione (III) as
the structure of ninhydrin. Since ninhydrin was color-
CXXh
0
XII XIII
which is obtained with nitric acid, forms useful deriva-
tives with most organic amines (7).
Ninhydrin shows not only ketonic properties but also
in the 1- and 3-positions must be reduces Fehling’s solution and ammoniacal silver ni-
less, the keto groups
trate. This was explained by Ruhemann (61) as being
separated. To explain the ready solubility of ninhydrin
in water and its thermal stability, Schonberg and Mou- due to the opening of the five-membered ring to give
basher (69) suggested an inner-salt or zwitter-ion for- o-carboxyphenylglyoxylic aldehyde (XIV).
mula (VIII) with the possibility of resonance between r^,COOH
VII and VIII. A structure (IX) stabilized by hydrogen lOCHOHCHO
XIV
bonding is also possible, since although the O(keto)-
O(hydroxyl) distance in ninhydrin is 3.18 A. and the Four oximes have been prepared from ninhydrin. The
normal 0—H—0 distance in hydrogen-bonded mole- 2-oxime and the 1,2,3-trioxime were first obtained
cules is 2.70 A., other five-membered hydrogen-bonded from diketohydrindene by Wislicenus (92, 93) and later
systems are known (12). by Ruhemann, who also showed that phenylhydrazine
III. General Properties of Ninhydrin and o-phenylenediamine react in the expected manner.
The reaction of phenylmagnesium bromide with nin-
a. physical properties
hydrin 2-oxime (52) and the preparation of the 1,2- and
Ninhydrin crystallizes as pale yellow prisms from 1,3-oximes have also been reported (28, 57). Reduction
ethanol; it gives a yellow aqueous solution and dissolves of ninhydrin 2-oxime with stannous chloride produces
 CHEMISTRY OF NINHYDRIN 41

the diketohydrindamine (XV) (65), which is very un- solution; if the solution is left to stand or is diluted, it
stable, but several derivatives are known, including the becomes colorless; when the solution is boiled immedi-
condensation product with benzaldehyde (XVI). Sev- ately after the addition of alkali, the yellow color
O O
changes rapidly to a deep blue and this color is not
lost on dilution. The same blue color, which is not the
color observed in the reaction with amino acids, etc.,
can be obtained by using cold concentrated alkali.
Ruhemann (62) explained these changes as follows and
the isolation of phthalidecarboxylic acid (XXIV) par-
eral authors have reported reactions between ninhydrin tially confirmed this.
and certain compounds containing the thiol group (79)
-"jCOOH yellow
and hydroxyl and carbonyl groups (20), but no general JCOCRO potassium salt
pattern can be discerned in these reactions. XXI
C. REDUCTION OF NINHYDRIN
ORP
Ninhydrin may be reduced to 2-hydroxyindandione
(XVII) with sodium amalgam (63), ascorbic acid (1), or
XX
'OH
blue
potassium salt

O
XXII
H
aCOOH
dilute HaSO«
QCHOHCOOH
XIV
XVII O

hydrogen sulfide. In the latter case the reaction pro- XXIV

ceeds further and hydrindantin (VI) is formed by com-
bination with a molecule of ninhydrin (71). This last
Hydrindantin gives a red color when dissolved in
aqueous sodium carbonate, and a blue solution in
step is reversible (27). sodium hydroxide (63). MacFadyen and Fowler (45)
The reactions of alloxan and alloxantin are strikingly
have shown that these colors are destroyed in the pres-
similar. Alloxan (XVIII) can be reduced to alloxantin
ence of oxygen. By strict control of the pH and oxygen
(XX) and, under appropriate conditions, this can dis- content of their solutions, they were able to distinguish
sociate into and be formed from alloxan and dialuric
between the color reactions due to ninhydrin and those
acid (XIX). It has been shown (51,68,70,71) that hydrin-
due to hydrindantin. This control was essential, since
it had been shown that hydrindantin can be oxidized
i°s OH
OH-HiO
to give two molecules of ninhydrin and, at certain pH
\ - +
\
/Sr
.
values, the latter can form hydrindantin by way of
OHO HO OH o-carboxyphenylglyoxal (XXI). The red and blue colors
XVIII XIX of hydrindantin were attributed to two anionic forms
HO) OH 0 of 2-hydroxy-l ,3-indandione, the monovalent ion
„n
HN NH
(XXVI) of the enolic form (XXV) being responsible
for the red color and the two possible divalent ions
/ ^J/OH
0
' H O HO
"
HO-zL-,
h0/nst\O -

(XXVII and XXVIII) for the blue color.

XX O
/H
dantin and alloxantin are correctly represented by the
pinacol structures XVII and XX, respectively, rather
than by hydrogen-bonded formulas. Tipson and
H
o
OH 0^4
XVI XXV
Cretcher (81) have confirmed these structures for (yellow)
alloxan and dialuric acid monohydrate from studies of
O 0 0-
the infrared and ultraviolet absorption spectra.

IV. Reactions of Ninhydrin

A. REACTION WITH ALKALI XXVI XXVII XXVIII
(red) (blue)
Ninhydrin exhibits many color reactions under dif-
ferent conditions. When the compound is dissolved in Under anaerobic conditions the red and blue colors are
4 N aqueous potassium hydroxide, it produces a yellow interconvertible. The effect of oxidation is irreversible
42 D. J. McCALDIN

for the blue color and is reversible for the red color only 1. Tke theory of Ruhemann
if ninhydrin is present in the solution. The most plaus- Ruhemann (61, 62, 63, 64, 65) was impressed with the
ible explanation is that the red compound (XXVI) is close similarity between ninhydrin and alloxan.
oxidized to o-carboxyphenylglyoxal (XXI) and the blue Strecker had shown that alloxan reacted with a-amino
compound (XXVII) to o-carboxymandelic acid acids to give carbon dioxide, an aldehyde, and a blue
(XXIII). compound which appeared to be murexide (8, 77). Since
0 alloxantin could be used to prepare murexide (55),
ACOCHO Ruhemann was prompted to search for the ninhydrin
U^JCOOH analog of alloxantin, which he obtained and called
hydrindantin. He found that it, like alloxantin, gave
XXVI XXI highly colored salts with alkalis and Ruhemann’s Purple
(red) with amino acids. He also showed that Ruhemann’s
0 Purple was the analog of murexide, obtaining it from
hydrindantin by following the method used to prepare
qiCHOHCOOH
kJcOOH murexide from alloxantin. This analogy was most strik-
ing and became the subject of later work (24). The
structure of murexide has been accepted as XXIX (10,
XXVII XXIII
(blue) 55, 74), and Ruhemann naturally represented its nin-

B. REACTION WITH AMINO ACIDS 0 0

NH
This reaction has been used extensively for detecting
and estimating amino acids. However, the mechanism
/"n\
ONH, OHO
0
of the reaction has given rise to a number of theories, XXIX XXX
and only recently has its nature been well understood.
Ruhemann observed that ninhydrin, like alloxan, hydrin analog as the ammonium salt of Ruhemann’s
stained the skin, and that a warm aqueous solution of Purple, the systematic name being diketohydrindyli-
an amino acid gave a deep purple-blue color when dene-diketohydrindamine (XXX). This was the sub-
treated with the reagent. He obtained the purple color stance believed to give rise to the purple color. His
with the amino acids that were available at that time, mechanism for the reaction is given below.
including j3-alanine and 0-aminopropionic acid, al-
though aromatic amino acids, such as anthranilic acid, 0
were shown to be the exception. There has been much /oh RCHNHaCQOH
confusion in later years concerning the type of amino OH (i)
acid which is reactive, and it has sometimes been stated
(66) that only -amino acids undergo this reaction. It III
has not always been appreciated, when attempts have 0
been made to explain the ninhydrin reaction, that ß-, /oh + ECHO + CO, + NH,
y-, -, and even e-amino acids do react under the appro-
priate conditions (11). This gives an indication of the
probable course of the ninhydrin reaction since, if the
reaction is as general as is stated above, it is likely that ninhydrin
(Ü)
the purple color observed is the same in each case, and
that only a fragment of the amino acid involved is con- 0 0 0 0
tained in the colored compound.
The early theories of the mechanism of this reaction
will be considered in turn. For convenience, the name
Ruhemann’s Purple will be used to denote the substance XXVII XXX
responsible for the purple color. Any theory proposed
must explain two experimental observations: (1) In the In modern terms the purple color would be associated
reaction of ninhydrin with -amino acids, ammonia, with the anion rather than the ammonium salt (see
carbon dioxide, and an aldehyde with one carbon atom below). The first stage of the reaction (i) and the last
less than the -amino acid are produced, and under (iii), in which hydrindantin reacts with two molecules
certain conditions may be obtained quantitatively. (2) of ammonia in a most unlikely manner, are not dis-
Hydrindantin is formed when ninhydrin reacts with cussed. Although Ruhemann was unaware of the fact,
-amino acids (1, 50). ninhydrin also gives a purple color with ammonium
 CHEMISTRY OF NINHYDRIN 43

salts (53) and his mechanism cannot account for this. The immediate advantage of this interpretation is
Moreover, the intensity of the purple color formed from that it explains why ammonium salts are able to pro-
ninhydrin and ammonium salts is reduced in the pres- duce a purple color. However, since ammonia is postu-
ence of hydrindantin (86). The theory cannot account lated as an intermediate, the theory does not explain
for the more rapid chromogenic reaction of a-amino why amino acids react faster with hydrindantin than
acids with ninhydrin and with hydrindantin, as com- does ammonia (45). Other criticisms of the theory are
pared with amines and ammonium salts (24, 66, 86). that a negligible amount of ammonia is evolved from
Nevertheless, his interpretation laid the foundation for amino acids in the absence of ninhydrin (42), and the
future work and his suggestion as to the origin of source of carbon dioxide, which is evolved, cannot be
Ruhemann’s Purple is substantially correct. the -keto acid. It has been shown that the evolution
of carbon dioxide from -keto acids is much slower than
Harding and MacLean
2. The theory of
from amino acids in the presence of ninhydrin (86, 87).
New experimental evidence led these workers (23, 24)
S. The theory of Retinger
to attempt to modify Ruhemann’s theory, whilst re-
taining the formula for the purple compound. They On the basis of experimental work which has re-
showed that the ultraviolet and visible absorption spec- mained unpublished, Retinger rejected the previous
tra of murexide (26) and Ruhemann’s Purple (20) were theories of the ninhydrin reaction and proposed another
similar, that pyridine greatly increased the rate of (59, 60). In this interpretation the amino acid or amine
color formation, and that ninhydrin reacted with am- yielded a monobasic colorless salt (XXXII) with hy-
monium salts, preferably in alkaline solution. The com- drindantin formed during the reaction. This salt then
plete reaction with alanine was interpreted as shown produced an unstable dibasic compound which imme-
below: diately split into two equal radicals (XXIII), which
CHaCHNHsCOOH —
NH, + CH,COCHO were supposed to account for the characteristic purple

0 color.

CH,COCHO +
(3^jp<0H

+NH2CH2C00H z
XV XVI ¿ ONH,CH,COOH
0 0 O 0 XXXIII

The theory would explain the apparent variety of colors

in the ninhydrin reaction as being due to the variable

XXXI XXX absorption spectra of the free radicals (XXXIII), de-

pending on the particular amino acid residue present.
They utilized the theory of Dakin and Dudley (9), Moreover, the fading of the color could well be due to
which postulated that -amino acids can undergo a dis- the decomposition of the free radicals. Retinger’s theory
sociation or decomposition into ammonia and the cor- is open to many objections, the most important being
responding glyoxal. Methylglyoxal is a powerful reduc- that salts of hydrindantin show no absorption in the
ing agent and possibly could reduce ninhydrin to visible range of the spectrum (44, 45), and that the
2-hydroxy-l, 3-indandione (XVI) which, with ammonia visible and ultraviolet absorption spectra of Ruhe-
obtained from the degraded amino acid, would form the mann’s Purple and hydrindantin differ greatly in alka-
amine (XV). This amine is known to condense readily line solution.
and would be expected to yield diketohydrindylidene-
diketohydrindamine (XXXI) with ninhydrin. The am- of Woker and Antener
4· The theory
monium salt of the product (XXX) would produce the These authors were strongly influenced by the
purple color on ionization. alloxan-ninhydrin analogy and represented the reac-
44 D. J. McCALDIN

tions of these two compounds with amino acids by theories proposed to explain the Strecker degradation
parallel schemes (96). They ignored the generally ac- (2, 14, 38, 91), that of Schonberg, Moubasher, and
cepted formulas for murexide and purpuric acid and Mustafa (73) is the most satisfactory. With isatin as an
without experimental evidence proposed the less likely example, the reaction is supposed to proceed as follows:
ethylenimine structures. Their suggested structures
for purpuric acid (XXXIV), murexide (XXXV), and
Ruhemann’s Purple (XXXVI) are given below: + NH2CHRCOOH ->

NCHRCOOH NCH2R

W\
0
O~f 0

n—]N=chr HaO -NH2 + RCHO

^'n\ 0 0
0 0 XXXVII

Compound XXXVII is of great significance; the anal-

ogous amine in the ninhydrin series provides a plausible
NH4 intermediate for the formation of diketohydridylidene-
XXXVI diketohydrindamine (DYDA) from ninhydrin. Any
mechanism for the Strecker degradation must explain
5. The theory of Moubasher and Schonberg
the formation of the Schiff base, the elimination of
These workers realized that the ninhydrin reaction carbon dioxide, and the hydrolysis of the rearranged
was a special case of the much more general Strecker compound. It is doubtful whether one particular step
reaction. The Strecker degradation (72, 77) refers to all determines the rate of reaction for all degradations.
degradations of -amino acids by carbonyl compounds Baddar (3, 4) interpreted the Strecker degradation elec-
and yields aldehydes or ketones containing one less tronically, but he was severely criticized (49) on the
carbon atom. They found (70) that ketones capable of grounds that he had provided insufficient experimental
effecting this reaction always contained the grouping evidence. As an example of the Strecker degradation
—CO[—C=C—]nCO—, where n was an integer or zero. and the formation of Ruhemann’s Purple, the mecha-
Examples for which = 0 are alloxan, isatin, and peri- nism of Moubasher and Ibrahim (50) is presented below.
naphthindan-1,2,3-trione hydrate. These three com- These workers consider that each of the compounds
pounds give color reactions with amino acids. Of the XXXVIII, XXXIX, XL, and XVI may be color con-
O o 0

CX^0H + RCHNH2COOH O=NCHRC00H

XX f
XJI^Jn=CRCOOH
+H»o,
-RCHO
0H -co2
A T OH

XXXVIII
o 0 o
/HH
;
OH
¡
+h2o
-NHs*
OH
2NH2CHRCOOH
—2H20
’

XL XVI VI

NCHRCOOH N=CHR
•—i OH HO, 2H20
—2RCHO
W HO
N=CHR OH NH,
XLII XLIII

O HO
-2HiO
V X
OH OH OH

XLIV XLV
 CHEMISTRY OF NINHYDRIN 45

tributors. The varied colors observed in paper chroma- This scheme is very similar to that of Ruhemann and
tography with different amino acids and ninhydrin is open to many of the same objections. The authors
could then be due to quantities of these compounds give no details of the manner in which the condensation
present along with the anion of DYDA (11). In support step is supposed to proceed, nor of the way in which
of this involved mechanism, only two new experimental the amino acid reduces ninhydrin to 2-hydroxy-1,3-
observations are offered. 2,2-Bis(l-hydroxy-3-ketoin- indandione. Ruhemann’s Purple is independent of the
dene) (XLV), a dark brown material, was obtained from nature of the cation and must therefore be due to the
the reaction of ninhydrin or hydrindantin with alanine. anion of diketohydrindylidene-diketohydrindamine
As far as the formation of hydrindantin (VI) their (XLVI). Their reaction scheme is supported by their
scheme follows the normal route observed for Strecker
degradations, but beyond this point the mechanism is
conjectural. No spectral data are offered, and in fact
there is little resemblance between the absorption curves
of 2,2-bis(l-hydroxy-3-ketoindene) (XLV) and those
of aqueous solutions of ninhydrin and amino acids after XLVI XXXI
reaction.
finding that the chromogenic reaction of -alanine with
6. The theory of MacFadyen
one molecule of hydrindantin is faster than with two
In 1944 MacFadyen stated that Ruhemann’s Purple molecules of ninhydrin, and that ammonia forms Ruhe-
was due to the anion of DYDA, since the color appeared mann’s Purple with ninhydrin in the presence of a re-
to depend on ionization (42, 44, 45). By studying the ducing agent capable of producing some hydrindantin.
behavior of both ninhydrin and hydrindantin by spec-
troradiometric methods under rigorously controlled 7. The present theory
conditions of pH, etc., he obtained results which sup-
ported none of the previous theories of the ninhydrin From the results obtained in a study of the reaction
reaction. He showed that the sodium salt of DYDA had of ninhydrin with imino acids (32), it is now suggested
the same spectrum as the ammonium salt (43) and that that the reactions of ninhydrin with amines, amino
the sodium and potassium salts had the same set of acids, and imino acids all proceed by the same mecha-
absorption coefficients. Extraction of a dilute sulfuric nism. The interpretation is based on the mechanism of
acid solution (0.001 N) of the sodium salt with benzene the Strecker degradation and explains the formation of
gave DYDA'2HsO. When this benzene solution was Ruhemann’s Purple and hydrindantin in the reactions
extracted with dilute sodium hydroxide at pH 10, with amino acids or amines (see page 46). It invokes
Ruhemann’s Purple was formed in the aqueous phase. a concerted electronic mechanism in the initial reaction
The ultraviolet and visible absorption spectra of solu- of ninhydrin with the -amino acid and it will be
tions of the sodium salt of DYDA were identical with observed that this avoids the enamine-vinylamine shift
those of reaction mixtures of ninhydrin with primary postulated by Moubasher and Ibrahim (50), which
amines, and the sodium salt was isolated from the reac- is improbable under these experimental conditions.
tion of ninhydrin with several amino acids (48). Mac- The evidence for the formation of the first intermediate
Fadyen and Fowler (45) showed that ammonium sulfate (XLVII) is strong; this reaction is the first step of the
or -alanine reacted with hydrindantin in acetate buffer Strecker degradation, and in the reaction of ninhydrin
at pH 7 and that the red color, due to hydrindantin, with cyclic bases compounds analogous to the structure
disappeared at a rate equal to the formation of Ruhe- XLVII were isolated. The zwitter ion (XLVIII) is then
mann’s Purple. It appears that the nonenolic component produced by the electronic changes shown, which in-
of Ruhemann’s Purple must be supplied by ninhydrin, volve decarboxylation and dehydration. From this
which comes from the hydrolysis of hydrindantin, since compound, the products of the reaction are formed by
only one molecule of the enolic component is used up hydrolysis or rearrangement. The amine (XL) was iso-
for each molecule of Ruhemann’s Purple formed. lated by Ruhemann, although he preferred to write it
in the diketo form. He had shown that the compound
was very reactive, and it is suggested that the three
reactions in which it is depicted as participating proceed
simultaneously. The aldehyde from the degradation
recondenses with the amine to yield XLIX, and 2-hy-
droxy-1,3-indandione (XVI) or its tautomer (XXV),
produced by further hydrolysis, combines with nin-
hydrin to furnish hydrindantin (VI). Finally, a further
molecule of ninhydrin condenses with the amine to
46 D. J. McCALDIN

39, 40, 41). The theory explains the appearance of

+ NHjCHRCOOH r\
aldehyde, carbon dioxide, ammonia, hydrindantin, and
NH—CHR Ruhemann’s Purple. It has the advantage that it sug-
w
gests a common route for the reaction of ninhydrin with
III XLVII
amines, imino acids, and amino acids, and it explains
Z why these reactions proceed more readily with com-
/
pounds which contain a carboxyl group adjacent to the
nitrogen atom. The orange, brown, gray, and green
z colors (5,11, 24) observed by various workers are almost
\tH=CHR certainly due to the presence of Schiff bases of the type
of XLIX, formed via routes (a) or (d). Indeed, Ruhe-
XL VIII mann isolated several such condensation products in-
cluding the compound XLIX (R =
phenyl), which is
hydrolysis y (a)\rearrangement orange in color.

/ \ \
C. REACTION WITH IMINO ACIDS

Whereas most of the common amino acids give purple

RCHO z colors, certain -imino acids such as proline and hy-

^ » _(d) H
=CHR droxyproline give yellow colors (16, 17). This reaction
is often used for identifications after chromatography
XL XLIX on paper. After prolonged heating, the yellow proline
spot may be transformed into purple-red. The chem-

(b)
+H,0
—NHa
\
- III
(c)
\
istry of these colored compounds was first studied by
Grassmann and Arnim (16, 17), who suggested that
\ their formation might be used for the estimation of
0 0 O 0 proline and hydroxyproline. Moore and Stein (48)
Z or
/h / HO^· realized that the cyclic -imino acids were anomalous
-NH- in the method that they developed for the photometric
OH
determination of -amino acids with ninhydrin. Other
XXV XVI authors (6, 76, 83, 90) evolved methods for estimating
all naturally occurring cyclic -imino acids, on the basis
-HsO of the absorption spectra of the colors they yield with
+m —H+
ninhydrin. In the case of proline, for example, it was
found that a yellow color (Xmax, 350 µ) was formed
with ninhydrin in acetic acid at room temperature,
whereas a reddish-purple color (Xmax, 515 µ) was
formed at 100°C. If the latter reaction was carried out
in neutral solution, the resulting compound showed
XL VI maximum absorption at 550 myu, and this was the com-
pound which was investigated chemically by Grass-
produce Ruhemann’s Purple (XLVI). Somewhat re- mann and Arnim (16,17). They suggested the structures
lated mechanisms have been discussed by Hammick LI and LII for the yellow and purple colors, respec-
(22), Hine (29), Sweeley and Horning (78), and Metzler, tively, and showed that the former was an intermediate
Ikawa, and Snell (46). At pH 1 to 2.5, the reaction pro- in the formation of the latter.
ceeds chiefly by route (b), ammonia is evolved almost
quantitatively, and no Ruhemann’s Purple is formed
(42). In solutions of pH 5, route (c) must predominate
since, under these conditions, color formation is the
basis of the analytical method of Moore and Stein (48).
The reaction with imino acids (see below) follows the LI LII
same path as far as the compound XLVIII. The reac-
tions of ninhydrin are represented here as modifications OH O
of the Strecker degradation, which is interpreted as
following route (b). This would be the path followed by
the reaction of amino acids, NH2CHRCOOH, with
other carbonyl compounds such as the isatins, etc. (38, LIII
 CHEMISTRY OF NINHYDRIN 47

These structures were accepted by later workers (20, The yellow intermediate corresponding to LV is less
54, 56), and Troll and Lindsley (85) attempted to ex- stable, although it can often be observed as a transient
plain the pigment obtained in acetic acid as being due color, and the final condensation reaction, actuated by
to an enol form (LIII) of LII. These structures have the endo double bond, furnishes the purple-red com-
been revised recently (32), and it has been shown that pound corresponding to LVI. This striking difference in
the effect of ring size on the reaction of the cyclic the behavior of the five- and six-membered cyclic
-imino acids with ninhydrin causes the formation of a -imino acids is attributed to steric factors related to
different type of stable condensation product from each the ring size. The behavior of 2-azetidinecarboxylic
of th° four-, five-, and six-membered cyclic a-imino acid (LVIII) with ninhydrin differs from either of the
acids. The reaction between ninhydrin and proline,
which is outlined below, is envisaged as following the
generalized mechanism described previously.
QcOOH <H >
LVII LVIII

previous cases. This imino acid normally gives a brown

color on paper, but if this condensation product (pre-
sumably the analog of LV) is treated with an excess of
ninhydrin and heated, the color of the spot changes to
blue. The main absorption of the new product corre-
sponds to that obtained from the reaction of ninhydrin
with a primary amino acid. 2-Azetidinecarboxylic acid
is known (13) to undergo ring fission to homoserine
under comparatively mild conditions, and it is possible
therefore that the purple condensation product is
Compound LV is the yellow product by which proline the same as that derived from ninhydrin and homo-
is generally characterized on paper. Its color, spectra, serine.
and other properties closely resemble those of the enol-
betaines described by Stafford (75). This mechanism,
which involves the -carboxylic acid group of proline, TABLE 1

Reactions of cyclic imino acids with ninhydrin*

explains why no similar intermediate is obtained in the
reaction of ninhydrin and pyrrolidine. The purple con-
Ninhydrin Yellow
densation product (LVI) can be obtained from the Product Product
intermediate LV or from the reaction of ninhydrin and
pyrrolidine, and the suggested structure is substantiated 3-CarboxypyrroIidine........................... X x

by experimental evidence and by analogy with the pro- 3-MethylproIine............................... X x

4-Methylproline............................... X X
posed structure for isatin blue (31). Under normal con- 4-Hydroxymethylproline....................... X X

ditions, however, the reaction of proline and ninhydrin l-Methyl-2-proline............................. — —

1,1-Dimethy 1-2-carboxypyrrolidine betaine........ — —

only proceeds as far as the enol-betaine (LV). Kainic acid (47)............................... X —

Structures LI and LII can be criticized on several B. 2-Carboxypiperidine (pipecolic acid)............. Transient X
5-Hydroxy-2-carboxypiperidine.................. Transient X
points. Condensations or rearrangements involving the l,2,3,6-Tetrahydro-2-carboxypyridine (baikiain)... Brown Brown
-carbon atoms of cyclic -imino acids are of limited C. 2-Azetidinecarboxylic acid...................... Brown
occurrence (30, 95), and the analogy drawn by Grass-
*
mann and Arnim (16) with migrations from nitrogen to x represents a positive reaction; a dash represents a negative
one.
carbon in the pyrrole series is clearly invalid. Further-
The groups A, B, and C correspond to acids with five-, six-,
more, the observations that V-substituted imino acids and four-membered rings, respectively.
do not give colored compounds and that characteristic
pigments may be obtained from cyclic amines substi- D. REACTION WITH AMINES
tuted at the -position make these structures no longer
tenable. The more recent interpretation follows the
1. Primary amines
suggested general mechanism for the ninhydrin reaction
and explains why cyclic -imino acids react faster than Neuberg (53) showed that many primary amines
the parent amines. give Ruhemann’s Purple with ninhydrin, Harding and
Pipecolic acid (LVII) and its derivatives are char- MacLean extended their theory for amino acids and
acterized by purple-red rather than yellow spots in suggested that in alkaline solution ninhydrin could be
paper chromatographic analysis by the ninhydrin converted to o-carboxyphenylglyoxal, which would re-
method (18, 94). duce ninhydrin to 2-hydroxyindan-l ,3-dione.
48 U. J. McCALDIN

O
methylurea give analogous products (56). In attempts
ZoH + RCH2NH2 -> to clarify their behavior the color reactions of many

ík‘ amines have been summarized (5, 24). The diamines

ornithine, cadaverine, and 1,6-diaminohexane are re-
o O O ported to yield Ruhemann’s Purple in amounts corre-
ZH'
X sponding to the reaction of only one amino group.
j^NHCHíR -X
H Ethylenediamine does not react in the same manner
O however (98).
LVI XXXI
2. Secondary amines
This would combine with the amine to form the amino
The reactions of cyclic secondary amines were orig-
compound LVI, which might condense further with
inally discussed by Grassman and Arnim (16, 17). They
ninhydrin to give diketohydrindylidene-diketohydrind- observed that pyrrolidine and piperidine formed purple-
amine (XXXI). The latter compound would give a blue
red pigments identical with those obtained from proline
color with excess of the amine by formation of a salt.
and pipecolic acid, and that colorless intermediate
Moubasher and Othman (51) suggested that the reac-
tion proceeded by the Strecker degradation, but their products could be isolated. These intermediates were
readily converted into the purple-red compounds by the
interpretation of the formation of the purple color in- action of heat or in the presence of acetic anhydride,
volved a double salt, which is unlikely. MacFadyen
and it was suggested that the intermediate compounds
proposed that amines required the presence of both were 2,2-diamino-1,3-indandiones (LXI). A similar
ninhydrin and reduced ninhydrin (43), and it was sup-
interpretation to that suggested for the reaction of an
posed that the reaction was either a simultaneous or a -imino acid with ninhydrin was put forward.
sequential condensation with ninhydrin and the enol The nature of these reactions was investigated more
form of reduced ninhydrin. The reaction of ninhydrin
recently and a variation of the general mechanism was
with primary amines can be considered as a special case
proposed for the conversion of the 2,2-diaminoindan-
of the general ninhydrin reaction outlined previously
for the reaction of amino acids. 1,3-dione into the pigment. Piperidine is taken as an
example. This conforms to the general ninhydrin reac-
O tion and implies that the ninhydrin units are not fused

to the «-positions of the piperidine nucleus. A range of

intermediate products may be obtained with different
XLVI cyclic secondary amines. Morpholine, for example, with
ninhydrin yields the compound LXIX (32), which can
Evidence is provided by the isolation of several com- be converted to a purple-red pigment (LXX).
pounds corresponding to LVIII. Thus guanidine (44,
62) and benzamidine (86) form compounds LIX and O o o
LX, respectively, and urea, monomethylurea, and di- -
Z-V
II N O
0 0 Ó ^
OH LXIX LXX
NHCNH, nhcc,h6
The spatial arrangement of atoms is thought to be
NH Ah rather congested in the latter compounds, since 2,6-
LIX LX dimethylpiperidine, hexamethylenimine, tetrahydro-
 CHEMISTRY OF NINHYDRIN 49

quinoline, and certain other cyclic secondary amines do

not form compounds of the types described above.
The secondary amino acid sarcosine is reported to be
unreactive (73), but more recent work has shown this
to be incorrect (97). Formaldehyde may be obtained
from the reaction of ninhydrin and sarcosine, ephed-
Z
rine, adrenaline, and other N-methyl amines. It is sug- z
gested that in each case the reaction proceeds through o
the formation of methylamine, which is then degraded
in the normal manner. These reactions can be explained
by the mechanism outlined earlier in this article, since
other work (33) has shown that under different condi-
tions N-methylbenzylamine yields benzaldehyde and
not formaldehyde, as reported by Yamaghishi and
Yoshida.
R - R' =
CHs; R" - COOCiHe.
3. Tertiary amines
F. REACTION WITH AMMONIUM SALTS
These do not react with ninhydrin.
All ammonium salts give a purple color with nin-
. REACTION WITH PYRROLES hydrin, provided the solution is sufficiently concen-
trated (25, 53). The color was shown to be the same as
Pyrroles react with ninhydrin to yield pigments that obtained from amino acids, and in water contain-
which were assigned similar structures to those obtained
from isatin (16, 17). It was suggested that compounds ing 0.008 per cent of ammonia was claimed (15) to give
a distinct positive reaction. Several of the earlier ana-
of structure LXXIII could be obtained from the reac-
tion of ninhydrin (or isatin) with pyrrole, pyrroline, or lytical procedures are useless, because all traces of
ammonia and ammonium salts were not initially re-
pyrrolidine. moved. Harding and Warenford (25) deduced the
following mechanism for this reaction, and the work of
MacFadyen and Fowler also supports it.
O

LXXIII

Recently Treibs, Herrmann, and Meissner (84) have

reconsidered these ideas and have shown that two pig-
ments are produced from the reaction of pyrroles with
ninhydrin. The two compounds have similar ultraviolet
spectra, but the analyses showed that two compounds
were produced for which the structures LXXIV and
LXXV were suggested. A sequence of reactions was
described by these workers. Initially the yellow com-
pound LXXVI is formed, from which "ninhydrin blue
1” can be obtained by treatment with hydrobromic
acid. Further addition of the pyrrole yields the yellow
diadduct LXXVII, and this produces “ninhydrin blue
2” (LXXV) on similar treatment with hydrobromic
acid. The formation of an ether linkage in LXXV
V. Analytical Uses of Ninhydrin
-would be unique in the chemistry of ninhydrin and
under acid conditions such a reaction is thought to be The possible use of ninhydrin for the analysis of
unlikely. amino acids was suggested by Ruhemann (62), and
since that time found wide application in identifi-
it^has
cation especially after the introduction of paper chro-
matography. The qualitative methods have seen little
refinement since 1910, but attempts were soon made to
use the ninhydrin reaction as a quantitative method of
LXXVI LXXIV analysis.
50 D. J. McCALDIN

The most satisfactory quantitative method was (6) Chinard, F. P.: J. Biol. Chem. 199, 91 (1952).
evolved by Van Slyke, Dillon, MacFadyen, and Hamil- (7) Christensen, B. E., Wang, C. H., Davies, I. W., and
Harris, D.: Anal. Chem. 21, 1573 (1949).
ton (87). This method measured the quantity of carbon (8) Copley, G. N.: Analyst 66, 492 (1941).
dioxide evolved from the amino acids at pH 1 to 5 with (9) Dakin, H. D., and Dudley, H. W.: J. Biol. Chem. 127,
great accuracy. Every known amino acid obtained from 15 (1913).

protein hydrolysis can be analyzed by this method. (10) Davidson, D.,and Epstein, E.: J. Org. Chem. 1, 305
(1936).
Since both the carboxyl and amino functions are in-
(11) Dent, C. .: Biochem. J. 43, 169 (1948).
volved in the reaction, the method has been widely used (12) Dewar, M. J. S.: Nature 155, 50, 479 (1945).
for estimating free amino acids present in protein (13) Fowden, L: Biochem. J. 64, 323 (1956).
digests. (14) Francke, W.: Biochem. Z. 258, 250 (1933).
Attempts were made to estimate amino acids in a (15) Gardner, H.: Lancet 219 (2), 525 (1930).
(16) Grassmann, W., and Arnim, K.: Ann. 609, 288 (1934).
quantitative manner by measuring other products of (17) Grassmann, W., and Arnim, K.: Ann. 619, 205 (1935).
the ninhjdrin reaction. MacFadyen (42) showed that (18) Grobbelaar, N., Pollard, J. K., and Steward, F. C.:
at pH 1 to 2.5 no Ruhemann’s Purple was formed and Nature 175, 703 (1955).
the amino group of the amino acid appeared as am- (19) Gustowski, W.: Przemysl. Chem. 30, 694 (1951).
monia. However, the yields were never more than 90 (20) Halle, N., Loewenstein, E., and Pribram, E.: Biochem.
Z. 56, 357 (1913).
per cent, and the method involves many precautions (21) Hamilton, P. B., and Orthiz, P. J.: Anal. Chem. 22, 948
which make it rather unsatisfactory. (1950).
Determination of the aldehyde produced from the (22) Hammick, D. L., Roe, A. M., Weston, F. W., and Whit-
amino acid appeared a possible method (1). Virtanen ing, K. D. E.: J. Chem. Soc. 1963, 3825.
and Laine (88) used this reaction to estimate certain (23) Harding, V. J., and MacLean, R. M : J. Biol. Chem. 20,
217 (1915).
specific amino acids which yielded aldehydes that could
(24) Harding, V. J., and MacLean, R. M.: J. Biol. Chem. 25,
be determined easily. This is not the case with all alde- 337 (1916).
hydes, however, and the -imino acids proline and (25) Harding, V. J., and Warneford, F. H. S.: J. Biol. Chem.
hydroxyproline do not form aldehydes. 24, 503 (1916).
A colorimetric method for estimating amino acids (26) Hartley, W. N.: J. Chem. Soc. 1887, 152.
(27) Hassall, C. H.: J. Chem. Soc. 1948, 50.
depending on an estimation of Ruhemann’s Purple was (28) Hensler, F., and Schieffer, H.: Ber. 32, 28 (1899).
initially investigated by Harding and Warneford (25). (29) Hiñe, J. S.: Physical Organic Chemistry, p. 288. McGraw-
These workers showed that a colorimetric method of Hill Book Company, Inc., New York (1956).
estimation seemed limited by their inability to obtain (30) Huisgen, R., and Rist, H.: Ann. 594, 161 (1955).
a reproducible color standard. Their results agreed with (31) Johnson, A. W., and McCaldin, D. J.: J. Chem. Soc.
1957, 3470.
those obtained by the Van Slyke method. Moore and
(32) Johnson, A. W., and McCaldin, D. J.: J. Chem. Soc.
Stein (48) set out to improve this method and were 1968, 817.
able by the addition of stannous chloride or hydrin- (33) Johnson, A. W., and McCaldin, D. J.: Unpublished work.
dantin to obtain a constant color yield for a given amino (34) Kaufmann, V.: Ber. 30, 387 (1897).
acid. Peptides, primary amines, and ammonia could (35) Khoklov, A. S., Shchakina, L. A., and Shemyakin, M.
M.: J. Gen. Chem. (U.S.S.R.) 21, 997 (1951).
also be estimated by this method. The procedure which
(36) Kumon, T.: Z. physiol. Chem. 231, 205 (1935).
was designed for use in conjunction with their chro-
(37) Langenbeck, W.: Ber. 60, 930 (1927).
matographic work relies on the development of Ruhe- (38) Langenbeck, W.: Ber. 61, 942 (1928)
mann’s Purple. The reaction was carried out in a citrate (39) Langenbeck, W.: Ber. 70, 367 (1937).
buffer at pH 5 and 100°C. The absorption maximum of (40) Langenbeck, W.: Ber. 70, 672 (1937).
the purple color was estimated at 570 µ. Later the (41) Langenbeck, W.: Ber. 70, 1039 (1937).
(42) MacFadyen, D. A.: J. Biol. Chem. 163, 507 (1944).
method was modified to include estimation of proline (43) MacFadyen, D. A.: Federation Proc. 6, 273 (1947).
and hydroxyproline. The accuracy of the method is (44) MacFadyen, D. A.: J. Biol. Chem. 186, 1 (1950).
approximately ±2 per cent, which is comparable with (45) MacFadyen, D. A., and Fowler, N.: J. Biol. Chem. 186,
the Van Slyke method. 13 (1950).
(46) Metzler, D. E., Ikawa, M., and Snell, E. E.: J. Am.
Ninhydrin has found certain other uses in the detec- Chem. Soc. 76, 648 (1954).
tion of compounds in biological fluids, e.g., urinary (47) Miuazaki, M.: J. Pharm. Soc. Japan 76, 695 (1955).
indican (36, 82). (48) Moore, S., and Stein, W. H.: J. Biol. Chem. 176, 367
(1948).
VI. References (49) Moubasher, R.: J. Chem. Soc. 1951, 231.
(1) Abderhalden, R.: Z. physiol Chem. 262, 81 (1938). (50) Moubasher, R., and Ibrahim, M.: J. Chem. Soc. 1949,
702.
(2) Akabori, S.: Ber. 66, 143 (1933).
(3) Baddar, F.: J. Chem. Soc. (Suppl. No. 1) 1949, 163. (51) Moubasher, R., and Othman, A. M.: J. Am. Chem. Soc.
(4) Baddar, F.: J. Chem. Soc. 1950, 231. 72, 2666 (1950).
(5) Brbmner, J. M., and Kenten, R. H.: Biochem. J. 49, (52) Mustafa, A., and Kamel, M.: J. Am. Chem. Soc. 76, 124
651 (1951). (1954).
 CHEMISTRY OF NINHYDRIN 51

(53) Neubero, C.: Biochem. Z. 56, 495 (1913). (76) Stein, W. H., and Moore, S.: J. Biol. Chem. 190, 103
(54) Piez, K. A., Irrevere, F., and Wolff, H. L.: J. Biol. (1951).
Chem. 223, 687 (1956). (77) Strecker, A.: Ann. 123, 363 (1862).
(55) Piloty, O., and Finckh, K.: Ann. 333, 22 (1904). (78) Sweeley, C. C., and Horning, E. C.: J. Am. Chem. Soc.
(56) Polonovski, M., Gounard, P., and Goltz, G.: Bull. 79, 2622 (1957).
soc. chim. 6, 1557 (1939). (79) Tayeau, F., Marqueville, J., and Marqueville-
(57) Ponzio, G., and Pichetto, A.: Gazz. chim. ital. 63, 20 Regnier, S.: Bull. soc. pharm. Bordeaux 92, 231 (1954).
(1925). (80) Teeters, W. O., and Shriner, R. L.: J. Am. Chem. Soc.
(58) Purvis, J. E.: J. Chem. Soc. 1911, 1953. 55, 3026 (1933).
(59) Retinger, J. M.: Dissertation, Leipzig (1913). (81) Tipson, R. S., and Cretcher, L. H.: J. Org. Chem. 16,
(60) Retinger, J. M.: J. Am. Chem. Soc. 39, 1059 (1917). 1091 (1951).
(61) Ruhemann, S.: J. Chem. Soc. 97, 1438 (1910). (82) Tomita, M.: J. Oriental Med. 2, 189 (1924).
(62) Ruhemann, S.: J. Chem. Soc. 97, 2025 (1910). (83) Touster, O.: J. Am. Chem. Soc. 73, 491 (1951).
(63) Ruhemann, S.: J. Chem. Soc. 99, 792 (1911). (84) Treibs, A., Herrmann, E., and Meissner, E.: Ann. 612,
(64) Ruhemann, S.: J. Chem. Soc. 99, 1306 (1911). 229 (1958).
(65) Ruhemann, S.: J. Chem. Soc. 99, 1486 (1911). (85) Troll, W., and Lindsley, J.: J. Biol. Chem. 216, 655
(66) Schmidt, C. L. A. (Editor): The Chemistry of the Amino (1955).
Acids and Proteins, p. 186. Charles C. Thomas, Spring- (86) Van Slyke, D. D., and Hamilton, P. B.: J. Biol. Chem.
field, Illinois (1938). 150, 471 (1943).
(67) SchOnbbrg, A., and Azzam, R.: J. Chem. Soc. 1939, 1428. (87) Van Slyke, D. D., Dillon, R. T., MacFadyen, D. A.,
(68) SchOnbbrg, A., Latif, N., Moubasher, R., Awad, and and Hamilton, P.: J. Biol. Chem. 141, 627 (1941).
W. I.: J. Chem. Soc. 1960, 374. (88) Virtanen, A. I., and Laine, T.: Nature 142, 754 (1938).
(69) Schonberg, A., and Moubasher, R.: J. Chem. Soc. 1943,
(89) Wanag, G., and Lode, A.: Ber. 71B, 1267 (1938).
71. (90) Weisiger, J. R.: J. Biol. Chem. 186, 591 (1950).
(91) Wieland, H., and Bergel, F.: Ann. 439, 196 (1924).
(70) SchOnbbrg, A., and Moubasher, R.: J. Chem. Soc. 1944,
366.
(92) Wislicenus, W.: Ann. 246, 347 (1888).
(93) Wislicenus, W.: Ann. 252, 72 (1889).
(71) SchOnbbrg, A., and Moubasher, R.: J. Chem. Soc. 1949,
(94) Witkop, B., and Foltz, C. M.: J. Am. Chem. Soc. 79,
212.
192 (1957).
(72) SchOnbbrg, A., and Moubasher, R.: Chem. Revs. 50, (95) Wittig, G., and Ludwig, .: Ann. 589, 58 (1954).
261 (1952).
(96) Woker, G., and Antener, I.: Helv. Chim. Acta 20, 1260
(73) Schonberg, A., Moubasher, R., and Mustafa, A.: J. (1937).
Chem. Soc. 1948, 176. (97) Yamaghishi, M., and Yoshida, T.: J. Pharm. Soc. Japan
(74) Slimmer, M., and Stieglitz, J.: J. Am. Chem. Soc. 31, 74, 1042 (1954).
661 (1904). (98) Yamaghishi, M., and Yoshida, T.: J. Pharm. Soc. Japan
(75) Stafford, W.: J. Chem. Soc. 1952, 580. 74, 1075 (1954).