THE DETECTION AND ESTIMATION

OF 2-KETOHEXONIC ACIDS*
BY MARY C. LANNING AND SEYMOUR S. COHEN
(From the Children’s Hospital of Philadelphia (Department of Pediatrics), and the
Department of Physiological Chemistry, School of Medicine, University of
Pennsylvania, Philadelphia, Pennsylvania)
(Received for publication, October 17, 1950)

The pathway of conversion of 6-phosphogluconate to ribose-5-phosphate

is being studied in this laboratory (1, 2). It was suggested by Dickens
(3) that the first product of this enzymatic oxidation was 2-keto-6-phos-
phogluconate. Although we have obtained evidence for small amounts of
substances behaving like a 2-ketohexonic acid among the enzymatic deg-
radation products of 6-phosphogluconate, it is not possible to state that
these substances were generated in the reaction or were in fact 2-keto-
hexonate. A specific, sensitive analytical method for these substanceswas
not available. Stubbs et al. (4) utilized the ability of 2- and 5-ketogluconic
acids to reduce alkaline copper solutions and the optical rotations of these
substances to estimate their formation from glucose by cultures of Aceto-
batter subozyduns. Gijrlich and Ocensek (5) also estimated these com-
pounds by copper reduction methods, using empirical tables to relate re-
ducing values at 55” and 100” to a concentration of 2-ketogluconate in the
presenceof glucose. Militzer (6) stated that 5-ketogluconic acid could be
measured in the presence of 2-ketogluconic acid by reduction of Benedict’s
solution at 25”. Gorlich and Liebster (7) have published a method for
the determination of ketohexonic acids based on their conversion to dienol
hexonic acids by hydrochloric acid and subsequent measurement of the
time required to decolorize methylene blue. This method requires a min-
imum of 10 mg. of the ketohexonic acid and fails to distinguish between
the 2-keto and the 5-keto acid. Since none of these methods give the
sensitivity or specificity required for the analysis of our reaction mixtures,
we have developed a method for the determination of 2-ketohexonic acids
based on the formation of 2-hydroxyquinoxalines. These derivatives have
been used by Ohle (8) in the identification of compounds of this class.
o-Phenylenediamine will condense with hexonic acids to form benzimid-
azoles (I), or with 2-ketohexonic acids to form 2-hydroxyquinoxalines (II).
Benzimidazole formation has been investigated by Moore and Link (9,
* The work described in this paper was conducted in part under a grant-in-aid
from the Commonwealth Fund, and in part under the Office of Naval Research con-
tract N6ori-188, Task Order 1, NR 135455.
109

This is an Open Access article under the CC BY license.

110 ESTIMATION OF %KETOHEXONATE

10). Strongly acid reaction conditions and temperatures of 3.20-150’ are

required. In contrast, quinoxaline formation can occur at room tempera-
tures with 1 equivalent of acid per amino group.

(II)

EXPERIMENTAL

Absorption spectra of representative quinoxalines are shown in Fig. 1,

together with that of o-phenylenediamine. 2-Hydroxy-3-methylquinoxa-
line was prepared by the condensation of pyruvic acid with o-phenylene-
diamine according to the procedure of Hinsberg (11). 2-Hydroxy-S-((n-
arabo)tetrahydroxybutyl)quinoxaline was prepared according to the
procedure of Ohle (8). The ultraviolet absorption of these compounds is
seen to have a characteristic maximum at about 335 rnp and is indepen-
dent of the presence of the 4-carbon carbohydrate residue at the 3 position.
Similar findings have been reported by Kuhn and Bar (12). Since meas-
urement of the absorption at 330 my is possible in the presence of excess
o-phenylenediamine, the isolation of the quinoxaline is unnecessary for its
estimation.
The dependence of the reaction on the concentration of the o-phenylene-
diamine is shown in Fig. 2. A high molar ratio of 280 of reagent to keto
acid (100 7) is necessary to give a maximum yield. Further increase in
concentration of the amine improves the yield only slightly.
In Fig. 3 the rate of the reaction at 100” is shown; a 30 minute heating
period is seen to give satisfactory quinoxaline formation.
Fig. 4 shows that the reaction of o-phenylenediamine with 2-keto acids
is maximum when 2 equivalents of acid per mole of diamine are present,
and is not affected by any further increase in acidity.
The following conditions were selected for the estimation of 2-ketohex-
onic acids: The reagent is a freshly prepared 2.5 per cent aqueous solu-
tion of o-phenylenediamine dihydrochloride or a solution containing 15
mg. of free amine per ml. of 0.25 N HCl. To 2 ml. of a neutral solution
containing 10 to 100 y of 2-ketohexonic acid is added 1 ml. of reagent.
The. reaction tube is heated in a boiling water bath for 30 minutes and
cooled to room temperature. Optical density is measured at 330 and 360
mp. The ratio of the optical density at 330 to that at 360 rnp is 1.51 Z!Z
0.07 for reaction mixtures containing a 2-ketohexonic acid. The optical
 1%. C. LANNING AND S. S. COHEN 111

density-concentration relationship is linear over the range given with a

precision of it 1.6 per cent in triplicate determinations. Table I shows

2.8’

24

20

1.6-
t
%2-
X’
2.8.
0
k 4.
0.

0’ \,
220 240 260 280 300 320 340 360 380 403mk.t
WAVE LENGTH
FIG. 1. Quinoxaline spectra. Curve A, 1 X 10-x M o-phenylenediamine dihydro-
chloride. Curve B, 2.8 X 10-b M 2-hydroxy-3-methylquinoxaline. Curve C, 1.8 X
lo+ M 2-hydroxy-3-tetrahydroxybutylquinoxaline.

V
0 50 100 I.50 200
MICRO MOLES REAGENT
FIG. 2. Reaction of 0.215 FM of potassium D-2-ketogluconate, heated at 100” for
10 minutes with varying amounts of o-phenylenediamine. Total volume, 3 ml.

the relative reactivity of various 2-keto acids to which the procedure has
been applied. We are indebted to Dr. Peter Regna of Charles Pfizer and
Company for generous gifts of salts of the 2-keto acids.
112 ESTIMATION OF 2-KETOHEXONATE

The reactivity of various interfering substances is included in Table I.

The direct condensation of the neutral sugars with o-phenylenediamine

.6

0 I5 30 4
MINUTES
FIG. 3. Reaction rate of 0.215 pM of n-2-ketogluconate with 139 PM of o-phenyl-
enediamine dihydrochloride at 100’; total volume, 3 ml.

0 2 4
m EQ HCI F:R ?nE! BAS:o

FIG. 4. Effect of acidity on reaction of 139 PM of o-phenylenediamine with (a)

0.215 PM of n-2-ketogluconate; (b) 0.667 PM of n-ribose; (c) 0.114 PMof n-dehydro-
isoascorbate.

was shown by Ohle et al. (13, 14) to occur to a significant extent only un-
der conditions favoring an Amadori type of rearrangement. We have
particularly investigated the reaction with ribose. In contrast to the re-
 M. C. LANNING AND S. S. COHEK 113

action with 2-keto acids, the reaction with ribose is sensitive to the acidity
of the medium, as shown in Fig. 4. By using no more than 2 equivalents
of acid per mole of diamine, the reaction of ribose is minimized, giving
only 10 per cent of the optical density at 330 mp given by an equivalent
amount of 2-ketogluconate.
The presence of interfering sugars is readily recognized by observing the
ratio of optical densities at 330 and 360 mp. This ratio is raised from the
1.5 of a pure 2-keto acid solution to 2.5 for an equimolecular mixture of
2-ketogluconate and ribose. In such a case an independent determination
of ribose permits correction of the results.

TABLE I
Reactivitiesof Carbohydrate in Quinoxaline Formation
The values are given in millimoles X 104/(optical density (330 m
-
A. 2-Keto acids 2-Keto-n-ghrconate 4.02
2-Keto-n-galactonate 4.36
2-Keto-n-gulonate 3.32
2-Keto-n-glucoheptonate 8.63
2-Keto-n-galactoheptonate 3.96
Pyruvate 17.0
B. Other substances 5-Keto-n-ghrconate 67.3
n-Glucose 66.0
n-Galactose 40.0
n-Ribose 51.3
n-Arabinose 66.7
n-Fructose 117.0
D-Gluconate 2000
Oxalate 122.0
Dehydroisoascorbate 2.77
Dihydroxyacetone 21.9
Acetylacetone 135.0

The reaction of compounds of the ascorbate type with o-phenylenedi-

amine has also been studied by Ohle and Erlbach (15). They have shown
that condensation occurs between dehydroascorbate and o-phenylenedi-
amine but does not occur with the enediol form. Application of our re-
action to 2-keto acid solutions when the presence of such compounds is
suspected therefore involves precautions against oxidation before and dur-
ing the reaction period. The reaction between dehydroascorbate and
o-phenylenediamine differs from that of the 2-ketohexonic acids with the
reagent in that it proceeds at a measurable rate at room temperature.
This has been useful in indicating the presence of such compounds in mix-
tures .
In applying this 2-keto acid estimation to products of enzymatic reac-
114 ESTIMATION OF %KETOHEXONATE

tions it should be noted that trichloroacetic acid solutions may contain

traces of interfering substances. Deproteinization with perchloric acid,
followed by neutralization, permits quinoxaline formation.
o-Phenylenediamine has also proved to be a useful reagent for detection
of 2-ketohexonic acids on paper chromatograms. For this purpose we use
a 2 per cent solution of the dihydrochloride in 80 per cent ethanol. 50
y of 2-ketogluconic acid will give a marked reaction when heated with this
reagent for several minutes at loo”, the background color being unobtru-
sive. Although some aldoses also give a visible reaction, the color of such
spots is different from the greenish gray keto acid spots. The keto acid
spots are fluorescent, and 10 to 20 y quantities, which do not give a visible
spot, may be detected under an ultraviolet beam.

SUMMARY

o-Phenylenediamine may be used as a quantitative reagent for the de-

termination of 2-ketohexonic acids in amounts of 10 to 100 y, quinoxaline
formation being recognized by the ultraviolet absorption at 330 rnp. Re-
actions with other types of substances are minimized by control of acidity
and recognized when present by the ratio of absorption at 2 wave-lengths,
330 and 360 rnp.
o-Phenylenediamine has also been used as a spraying agent for paper
chromatograms, readily revealing the presence of 2-ketohexonic acids.

BIBLIOGRAPHY

1. Cohen, S. S., and Scott, D. B. M., Science, 111,543(1950).

2. Scott, D. B. M., and Cohen, S. S., J. Biol. Chem., 188, 509 (1951).
3, Dickens, F., Biochem. J., 32, 1626 (1938).
4. Stubbs, J. J., Lockwood, L. B., Roe, E. T., Tabenkin, B., and Ward, G. E.,
Ind. and Eng. Chem., 32, 1626 (1940).
5. Giirlich, B., and Ocensek, F., Collect. Czechoslov. Chem. Communicat., 13, 448
(1948).
6. Militzer, W. E., J. Biol. Chem., 164,325(1944).
7. Gorlich, B., and Liebster, J., Collect. Czechoslov. Chem. Communicat., 13, 616
(1948).
8. Ohle, H., Ber. them. Ges., 67, 155 (1934).
9. Moore, S., and Link, K. P., J. Org. Chem., 6, 637 (1940).
10. Moore, S., and Link, K. P., J. Biol. Chem., 133,293 (1940).
11. Hinsberg, O., Ann. Chem., 237, 340 (1887).
12. Kuhn, R., and Bar, F., Ber. them. Ges., 67 B, 898 (1934).
13. Ohle, H., and Heilscher, M., Ber. them. Ges., 74, 13 (1941).
14. Ohle, H., and Kruyff, J. J., Ber. them. Ges., 77 B, 507 (1944).
15. Ohle, H., and Erlbach, H., Ber. them. Ges., 67, 55 (1934).