Keto-Enol Tautomerism of Pyruvate Ion Studied

rated calomel electrode with a large surface which was connected to the test solution by an agar bridge, saturated with potassium chloride. The...

0 downloads 123 Views 3MB Size


formation of [email protected]) hydroxy compounds even when the reduction is done in the presence of

Vol. 61

mineral acids. STATE




The Keto-Enol Tautomerism of Pyruvate Ion Studied Polarographically* BY OTTOH. MULLERAND J. PERCY BAUMBERGER Introduction Pyruvate ion holds a key position in practically all schemes of carbohydrate metabolism in the intermediary steps of which the enol form is often postulated. The only evidence for this enol form has been spectroscopic‘ as the usual methods for its determination have failed in this particular case. In this paper, we present data regarding the keto/enol enolate ratio and polymerization of pyruvic acid and of pyruvate ion in relation to PH, and determine the apparent reduction potential of both forms of pyruvate ion a t different PH. These determinations are made electrochemically, using the polarographic method.


Method Materials.-The pyruvate used in these experiments was kindly given to one of us (J. P. B.) by Dr. H. Borsook of Pasadena in the form of solid lithium pyruvate of great purity. All other chemicals were c. P. products. The McIlvane series of buffers was used for pH 2.2-8.0 and Clark and Lubs HC1-KCl for PH 1.0-2.2. The pH was checked by means of the glass electrode. The temperature a t which these experiments were carried out was 25 * 0.1” maintained in a water thermostat. General Description of Polarographic Method.-The method makes use of a dropping mercury electrode as halfcell connected to a calomel half-cell by an agar bridge saturated with potassium chloride. A battery, slide wire potentiometer, and galvanometer are also in the circuit so that the voltage applied t o the system may be varied and the resulting current measured. The procedure depends upon the fact that many compounds may be reduced at the surface of mercury dropping slowly from a capillary when the voltage applied is sufficiently negative. The resulting current is a function of the concentration of reducible substance, while the potential a t which the reduction occurs is characteristic of the compound reduced. This principle has been incorporated into a very convenient apparatus, the polarograph, in which the current and voltage are automatically graphically coordinated on photographic paper to give a current-voltage curve, called a polarogram. J. Heyrovskjr and his schoolz are responsible (*) Original manuscript received November 26, 1937. (1) V. Henri and C. Fromageot, Bull. soc. chim., [4137, 846 (1925).

(2) J. Heyrovskq, “Polarographie,” in W. Bbttger, “Physikalische Methoden der analytischen Chemie,” Akademische Verlagsgesellschaft, Leipzig, 1936,Vol. 11, pp. 260-322.

for this development. On such a polarogram, the voltage is indicated by abscissas, and changes in current appear as waves. Apparatus.-Our experiments were carried out with a Nejedlg polarograph Model VI11 and a Nejedljl galvanometer with a sensitivity of 2.3 X 10-0 amp./mm./m. A special shunt permitted the modification of this sensitivity to any desired fraction, while critical damping was maintained. The apparatus was so adjusted that the distance between two abscissas on the polarogram corresponded to 200 mv. The dropping mercury electrode was made according to instructions given by Heyrovskjrz; its drop time in distilled water was six seconds when the mercury reservoir was 60 cm. above the capillary. As oxygen is also reducible at the dropping mercury electrode, hydrogen gas was bubbled through all solutions before each experiment to remove the dissolved oxygen. Instead of using a large layer of mercury a t the bottom of the electrolysis vessel as non-polarizable anode, as is customary in polarographic work, we made use of a saturated calomel electrode with a large surface which was connected to the test solution by an agar bridge, saturated with potassium chloride. The potential of this separate electrode is constant*and its value is known, thus eliminating the troublesome measurement of the anode potential. All polarograms reproduced in this paper, therefore, show automatically the applied e. m. f. referred to the saturated calomel electrode. For the plotting of graphs, liquid junction potentials have been neglected, while corrections for the iR dropa have been made in calculations of the half-wave potentials4 which are all referred to the normal hydrogen electrode as zero. The accuracy of these values is *20 mv., which seems sufficient for the present treatment. As we have shown8 polarographic potentials may be read with an accuracy of * 1 mv. if proper precautions are taken.

Keto-Enol Tautomerism This attempt to determine the concentration of keto and enol pyruvate ion was the direct outgrowth of the work on the coupled oxidation-reduction potential of pyruvate lactate, carried out in this Laborato~$.~ (3) 0.H . Muller and J. P. Baumberger, Trans. Electrochem. Soc., 71, 181 (1937). (4) J. Heyrovskg and D. IlkoviP, Coll. Cscchosloo. Chcm. Comm., 7, 198 (1935). (5) J. P. Baumberger, J. J. Jlirgensen, and K. Bardwell, J . Cen. Physiol., 16,961 (1933).


March, 1939 PI



1 s


. % . b/l






-.-, A 0.4





















/ / , ,







-,’..l’>--L-~< - --..’ 0.8






_._ i : ~_,’


Volts. Fig. 1.-Polarograms of pyruvate reduction at different PH. The initial applied voltage for each curve is indicated: abscissas are 0.2 v. apart and . promessivelv more negative. The applied voltage at the half-wave io indicated . in each case

The reduction of pyruvic acid a t the dropping waves must be that two distinctly different submercury electrode seemed to present a satisfactory stances are reduced which are in equilibrium with method for analytical determination, especially each other. This equilibrium changes with as several authors had mentioned such a reac- pH. The most logical cause for such an equilibtion!,’ However, closer investigation showed rium in the case of pyruvate ion is the keto-enol that the polarogram of pyruvate really involves tautomerism which has been mentioned frethree waves and that the voltage a t which each quently in the literature but for which no defioccurs is shifted by pH. This has not been rec- nite and conclusive evidence has so far been ognized previously as no systematic study has presented. It is our opinion that the first wave been made. We therefore carried out a series of which we obtain on the more acid side is due to the experiments in buffered solutions of different pH. reduction of the keto form of pyruvic acid or pyLithium pyruvate was used when i t was found to ruvate ion while the second wave is due to the give identical results with freshly distilled py- “enol” form. ruvic acid. The basis of our opinion is the well established We found it most convenient to use solutions of fact that only those acids and their ions are relow concentration, because the iR correction ducible a t the dropping mercury electrode which would be small and there would be a minimum have conjugated double bonds. The pyruvate of polymerization. Figure 1 is a reproduction of ion (I) has two -0 double bonds which are conthree polarograms pieced together showing curves jugated. The enol (11) and the enolate (111) due to the reduction of 0.001 M pyruvate in buf- forms of the pyruvate ion have one C=C double fers of different PH. We see that as the pH in- bond and a C=O group which are also conjucreases, the single wave a t pH 4.14 becomes smal- gated. We may therefore expect them to be ler and another wave appears. In the rest of this also reducible, although most likely a t a different paper, these two waves are called keto and “enol” potential. respectively, “enol” referring to the sum of enol (1) (11) (111) and enolate pyruvate. It may be noticed that CHrC-C-0H%C-C--O-0-C-C-Othe sum total of the two waves (keto “enol“) is II II II II /I 0 0 HE 0 HC a constant, proportional to the concentration of pyruvate present. A t pH 8 the keto wave has The need for conjugation of double bonds elimipractically vanished and the “enol” wave is now nates the possibility that the two waves may repas big as the original keto wave. The reduction resent reductions of different hydrates of the pyof pyruvic acid is known to require two electrons ruvate. per molecule6; from our curves we must conclude As associated and dissociated acids are indisthat this holds a t any pH, i. e., for pyruvate ion E. Laqueur. Rec. Iran chim.. 66, 959 (193611. There with s also and that the reduction goes to completion change in OHaoofhtr wave appears which, however, never hecomes than one-half the orisinn1 WPYL This meam that et a ceneio in a single step.8 The reason for the two different larger *€I the reduction goell on in two steps. each requiting one electron per


( 6 ) L. Schweer. Coll. Cuchoslm. Chrm. Comm.. 7.328 (1935). 171 . . A. Winkel sod G. Proskc. Bcr.. 69. 1917 11938). . . (8) In this fact we have P fundamental difference fmm the i n s mmiblr redvetions of benzaldehyde [M. Tokuoks. Coll. Cuchorla. Chrm. Conm., 7, 382 (193511 and bcnzophcoonc [H. Schwcitrer and


. .

molreule, while et another pH the reduction goes to completion in one rfeo. .. each molecule addine two eleeUons at once. The rnrrriblr two-step oxidations and reduction+ are mother entirely dinermt case of electro-reduction at the dropping mercury


tiuguishable froin each other polarographically, we cannot segregate the enol (11)from the enolate (111). We have therefore labelled our second wave “enol” to indicate that it represents the reduction of the sum of the enol and enolate forms of pyruvate ion.

6 7 8 pH. Fig. 3.-Relative concentration of pyruvate in the kclu and “enol” form a t different pH’s as indicated by the electro-reduction current of each form.






- 1.300 -€no/


- 1.400


4 6 8 PH. *gig. 2.-Apparent reduction potentials of pyruvic acid in its different forms a t different pH’s calculated from polarograms similar t o and including those in Fig. 1 after making proper corrections. 2

In Fig. 2 we have plotted the changes of potential of keto and “enol” reductions against the changes in pH. We have only included the pH range from 1 to 8, as the “enol” wave loses its smooth “S” shape, which makes the potential measurements unreliable. The curves represent the values obtained when the half-wave method4 tsf measuring potentials was used. For the case of reversible organic electro-reductions we have established the significant value of this half-wave p ~ t e n t i a l . ~Although we no doubt deal with a fairly irreversible process we have preferred this method of measurement to others because of its relationship to the ‘(apparent reduction potential” (A.R.P.) of Conantg as will be shown later on. Figure 3 shows a comparison of the heights of the keto and “enol” waves a t different pH values; it may be seen that at pH 5.8 they are equal. (9) J. 8 . Coaaat, Chrm. Rev., 9, 1 (1926),

This value, of course, would be valid only if the rate of enolization, respectively, ketone formation, is infinitely slow compared to the rate of reduction. We may assume that both keto and “enol” forms diffuse to the electrode a t about the same rate.1° If keto is reduced first and the cathode surface depleted of keto molecules, no doubt some of the remaining “enol” molecules will change into keto and also be reduced. The following experiments show that the rate of ketone formation is too great to be measured by our procedure. Solutions of lithium pyruvate in a fairly alkaline medium were shown polarographically to be practically free from the keto form. Then a constant potential of 1.2 v. was applied a t which keto is reduced in acid solution whne “enol” is not; the corresponding current was of course very small. When oxygen-free acid was added to the solution in a stream of hydrogen, the current reached the full value which it would have if only keto were present, after a time interval of only one-half minute. This was the minimum time necessary to ensure complete mixing of the solution by bubbling hydrogen gas through it. Even when the solution was kept in an ice-bath, the ketone formation had taken place before the mixing was considered adequate. It would seem, therefore, that the wave for keto may be somewhat larger because of conversion of “enol” to keto locally a t the surface of the electrode. However, this effect must be slight judging by the fact that the keto wave practically vanishes a t pH 7.5.

The ratio of keto/“enol” becomes unity a t

pH 5.8 and decreases with increase in pH and vice versa; e. g., it is approximately 10 a t PH 4.8 and 0.1 a t pH 6.8. This relation of the equilibrium of the two forms of pyruvate to pH brings up the question as to whether or not the enol pyruvate ion (11) is dissociated to give an enolate ion (111). The polarographic method cannot be used to answer this question as it does not distinguish between associated and dissociated acids. A preliminary titration of pyruvic acid seemed (IO) The rate of diffusion of the reducible substance is one of the important factors governing the height of polarographic waves [D. IlkoviE, Colt. Ceechoslov. Chen. Comm., 6, 498 (1934)l.

.i I

Fig. 4.-The

polymerization of pyruvate


affected by pH and concentration


Three families of

curves, all starting at zero applied voltage, with members at similar approximate concentrations a n l galvanometer sensitivities identically arranged as follows: (a) 0.001 M ; (b) 0.002 M; (c) 0.005 M; (dr) 0.1 11; (e,) 0.2 M; all at 1/5M) sensitivity. (a,) 0.1 M; (e,) 0.2 M; each at 1/50 sensitivity.

to indicate a second dissociation constant around

pH 5.8 but speculations as to the steps in the enolization (either enolization and subsequent dissociation, or dissociation from the methyl groups with induced enolization) have to be postponed until more accurate measurements are made. Polymerization Our next study concerned the wave a t -0.2 v. from the normal calomel zero which Schwaer‘ had reported as due to the undissociated, molecular form of pyruvic acid. As we had not seen it in our dilute solutions, we increased the concentration of lithium pyruvate. Fimre 4 is a reproduction of 3 polarograms which are representative of these studies. All are carried out analogously, the starting solutions (curves 4a) were from the experiments which are presented in Fig. 1. To these solutions increasing amounts of lithium pyruvate were added until in the last curve on each polarogram the concentration of lithium pyruvate was about 0.2 M. The first group of curves in Fig. 4 is a t pH 4.1 and represents typical ketowaves. The second group is a t pH 5.9 and shows both keto and “enol” reductions. It should be noticed that while the keto reduction even a t the higher concentration shows no maxirnum,l the “enol” reduction is marked by a very pronounced maximum. This can be seen again in the third group, which represents typical “enol” reductions, in a pH 7.7 buffer.“ In the first and second groups of curves (Fig. 4) we furthermorc notice a little wave indicated by an arrow which is absent in the third group and only appears with higher sensitivities of the gal(11) The significance of this maximum which only DECUII in the reduction i9 still uneerfpin. We believe i t indicates s differ. e m e in the nature of the eieetro-reduction of the two tautomeric forms. “Fool”

vanometer in the more concentrated solir’;~:: This wave changes to more positive poter.:r.l and becomes greater with decreasing pH. In i 1 N solution of hydrochloric acid, it is quite PIC. nounced and is almost at -0,200 v. We then: fore feel that this is the same wave which Schwaw reported as due to the associated, molecular form of pyruvic acid. However, we believe i t to b: caused by a polymerization product of pytuvir acid, because we observe that this wave g ~ m v . . with age; e. E., the two waves of pH 4.1 and 5.3 ir: creased from 3.5 to 11mm. and from 1.5 to 4 nm.. respectively, in twenty days, while pH 7.7 never shows this wave. As is well known, pyruvic acid easily polymerizes upon standing: in most work, therefore, only freshly vacuum distilled pyruvic acid has been used. Figure 5 represents polarograms of 0.01 M soktions of lithium pyruvate which have been left standing for a week after their preparation; the galvanometer sensitivity and the height of the waves are indicated in the figure. The highest wave is obtained a t pH 8.0 where the pyruvate is all in the unpolymerized “enol” form. Other relative concentrations of keto, “enol,” or polymer are calculated by referring the heights of their waves to this standard, assuming that the height of polarographic waves is a linear function of concentration.12 Thus from a comparison of the diminution in heights of the keto wave and the correspondingly appearing polymer wave we can calculate It for the polymer on the assumption (12) While the strict promionaiity of heisht of wave t o eoncan. trstion he9 been pmvcn in polarographic work for mnoy systems, there is io generd no re-n to believe that thin ia true io nil e-. crpeeidlr if the conditions are ehnnged. From our experience rith pyruvic add in b v K m ofdinereof 9 H Pod at varying concentratioas (seer. I.. Figs. 1 and 4 ) we find, however. that the eoncentration/ heiaht-of-wave rdationnhip i s sufficiently eoortaat to give signifiC B ~ C Cto the deductions from our experiments wbich we make ig thiv wer.






... _ .. .. .


Vol. 61

. . .. . .... .








J 1

Fig. 5.-The relative concentrations of polymerized and unpolymerized pyruvate at different pH's are shown as heights of polarographic waves. The conditions for each curve at a specified pH are identical except that the galvanometer sensitivity is different.

that the undeterminable percentage of prepared pyruvate is fully represented in these polymer waves. We then get 2.2


4.0 2.9



6.0 2.9

8.0 ~opalymer

A polymer of pyruvic acid with 1z = 3 was found by M a s ~ o l using '~ cryoscopic methods. 15



PYRUWC A C/O ~oe//&t&


o M~/W/OPAAwqim

mmpmd w/fh Po/orqqhoh/ca/% Defermhub/e %ofPmmdConcPntrrd/M









PH. Fig. 6.-The molecular absorption coefficient ofpyruvic acid (associated and dissociated) at different PH's plotted from Henri and Fromageat's data' is shown by Open circles. For comparison the percentage of total pyruvic acid which is determinable as keto and "enol" pyruvate is shown as solid circular dots.

namely 3300 A,,is really the absorption maximum for C=O and that < 2300 A. would have been proper for "enol" absorption. Therefore "enol" was probably not measured by Henri and Fromageot. Our observation on polymerization and the spectrophotometric work of Henri and Fromageot' show an interesting analogy. This is illustrated in Fig. G where the molecular absorption coefficient, e, and the polarographically determinable percentage of prepared concentration of pyruvate vary similarly with p H . I t seems reasonable to conclude that the change in light absorption measured was the result of polymerization of pyruvate rather than enolization. The supposition of the above polymer might also explain the later work of Fromageot and co-workers" on substituted pyruvic acids. Bromine titrations in which ferric chloride was used as indicator showed an increase of the reacting compound toward more acid solutions: e. g., 0% at PH 5.0 and 0.5% a t PH 1. Fromageot therefore altered his previous viewPointi and concluded that in the acid region the enol form predominated, while the increase in absorption of the ultraviolet toward the alkaline range should be attributed to a compound of the structure

Henri and Fromageot' presented spectrophotometric evidence for the enolization of pyruvate ion in alkali. We made a similar study" on R,c=c-c(oH buffer solutions of 0.001. 0.005 and 0.01 M \d OH lithium pyruvate. Only the o.ol pyruvate If ferric chloride and bromine are specific reashows marked differences in light absorption gents for a double bond,?= is assumed for the with pH and allows the calculation of values which titrations, then the last formula must be with their findings. It however? wrong, because it, too, has a double bond, and from Grossmann's" work that the wave length therefore should react, Second, as z c e n o l s ~are ~ of light chosen to study the C - C concentration, generally known to form toward the &a. (13) M.G . Masrol. Bull. IDC. chin.. 131 3s.335 (1905). line side than the acid side; it is un(14) We are very much indebted to Dr. B. L. Crowford. Jr.. oi t h e Chemistry Department of this University this aprtmphotomctric study.

for helping

us carry out

(15) P. Gro%%m.nn.2. phyrik. Chrm.. 109.305 (1924).

(16) C. Promageat and S. Pcrraud. Biochmi. 2.. 92%. 213 (1830);

C. Fmmagwt, M. Pelletis nod P. Ehrcortcin. Ruil. roc. chim.. I41 81, 1283 (1932).


March, 1939


If the irreversible step in the above reactions is likely that Fromageot titrated “enols” in his experiments. Third, the bromine reaction and that sufficiently slow, compared to the electron transof ferric chloride is known to fail in a number of fer itself in the reversible process, the product of cases where the existence of “enols” is fairly well the reversible reduction (L”) would remain a t the established. Hantzschl’ and Dieckmann18 came electrode surface long enough to impress upon it to the conclusion that the formation of colored a potential governed by the ratio of its concenferric compounds between “enols” and ferric tration to that of the remaining, still unreduced, chloride may be expected only if inner complex oxidant. As the dropping mercury electrode is salts or “conjugated enol-salts” with a secondary unique in presenting a constantly renewed, fresh valence linkage between metal and carbonyl surface, we are able to pick out the reversible oxygen are possible, as may well be the case in steps in such irreversible reactions and treat them separately. polymerized pyruvate. Of the compounds studied by Conant only three The Apparent Reduction Potential of Keto and have been investigated polarographically. We “Enol” Pyruvate and of Pyruvic Acid.-Conantg describes an important characteristic of irrevers- have estimated their half-wave potentials from ibly reducible compounds which he calls the the published data and find good agreement “apparent reduction potential” ( A . R . P . ) . This is with Conant’s determination of the A . R . P . ~ as determined by observing the rate a t which re- shown below.20 I n all these cases the solvent versible oxidation-reduction systems (B) are was hydrochloric acid and the potential is exoxidized by the irreversible system (A). He pressed as Eh. gives evidence that if the rate of oxidation of B NitroMaleic 1,3 Dinitroacid benzene benzene is related to the E‘o of B and &e A.R.P. of A, there $0.16 +0.06 A. R . P. Conant,Q v. -0.25 must be a reversible step in the reduction of A, Polarographic halffollowed by an irreversible step wave, v. -0.28’l +0.16’2 +0.07**



RedA* +RedA

We have shown3 that, by means of the polarograph, potentials of reversible oxidation-reduction systems can be measured, even if only oxidant or reductant alone is present in solution. The half-wave potential of a polarographic curve corresponds with the E‘o of the particular system.

It would appear that the polarographic method has the advantage (in cases where it can be used) that the reversible step P z L * is measured rather than the irreversible L* +L step, which Conant’s method employs, i. e., we measure E,‘ of P s L * directly whereas he obtains it indirectly. It also has the advantage that the A . R . P . of more than one irreversible compound can be determined in the same solution; this would be very difficult by any other method we know of. Figure 2 shows the A.R.P. of keto-pyruvic acid a t different pH. The general slope of this curve might be expected to be 60 mv./pH as in general for E’Q of organic compounds, and to be modified by dissociation of the carboxyl groups of pyruvic acid and lactic acid. The curves show marked changes in slope a t pH’s close to the pKa of pyruvic acid (2.45) and lactic acid (3.85). In our experiments with reversible oxidationreduction systems we were able to oxidize the reductant a t the dropping mercury e l e ~ t r o d e . ~ The half-wave potentials thus obtained from

(17) A. Hantzsch, Ann. Chem., S92, 286 (1912). (18) W.Dieckmann, Ber., 60, 1373 (1917). (19) In this paper lactic acid is stated to be the end-product of the reduction of pyruvic acid at the dropping mercury electrode. This assumption is made for convenience but also has support in the fact that lactic acid is formed when pyruvic acid is reduced in acid ~. solution by zinc or by electrochemical means.

(20) M.Shikata [ M e n . CoElege Agr., Kyoto Imp. Uniw., No. 4, 7 (1927)I recognized an analogy between his polarographically determined electrode reduction potential and Conant’s A.R.P. (21) E.Vopifka, Coll. CecchosZoo. Ckem. Comm., 8,349 (1936). (22) M. Shikata and N . Horaki, Mem. College Agr., Kyoto Imp. Univ.. No. 17. 1 (1931). (23j M.Shikata, T7ans. Faradoy SOC.,21,42 (1926).

Thus considering the reversible stage in the reduction of A, the relation may be expressed

Therefore i t follows that A . R . P . = E’oA. I n the polarographic method a similar situation exists a t the half-wave potential3 which therefore should correspond to A . R . P . For the case of the reduction of pyruvic acid (P) to an electrornotively active form of lactic acid (L*) and then to the stable form of lactic acid (L),l9 we have P-L*+L




curves due to the anodic oxidation of reductant mere identical with those due to the cathodic reduction of the oxidant of the same system. ThB fact serves as a very convenient criterion to ~:kt~errriine the reversibility of an oxidation-rediiclion system. For instance we were unable to oxidize lactic acid in a solution of nitric acid at the dropping mercury electrode, using the most positive potential available a t this electrode, ~G;Z.~ 2$, =: 4-0.65 v. This supports the view that the reversible step P L* is followed by an irreversible process I,* L. Using Conant’s method, BarmoreZ4 found that Imctic. acid could be oxidized in acid solution by potassiu:m permanganate which corresponded t o an apparent oxidation potential (A. 0. P.) of Ej, 4-1.4v. However, when he tried to deteimine the A. R. P. of pyruvic acid by this method he failed, because the mast negative reagent he csiAcl ~rse,i. e., titanitup ,trichloride (Eh = -0.8 a.t +EI 7.2) was not negative enough to bring about a red.uction. For such systems, therefore, the ,polarographic method and Conant’s method supplement each other.


I -

(!!fit) M. A. Barmore, “Electrometric Studies on Pyruvic Acid, i m t i c Acid, and Glyceric Aldehyde” (uupublished thesis), Stanford <. rtiwrsity Library, 1929.

Vol. 61

This work was aided by a grant from the Rockefeller Foundation.

Summary 1. The application of the polarographic method to the study of keto-enol tautomerism, polymerization and apparent reduction potential is described. 2. Evidence is presented in support of the following facts: (a) The ratio of keto/enol enolate forms of pyruvate ion becomes unity at pW 5.8 and decreases with increase in flH and vice versa. (b) Pyruvic acid polymerizes even in fairly dilute solutions and this polymerization increases with increase in acidity and with time. (c) The “polarographic half-wave potential” is equivalent to the “apparent reduction potential” of Conant and represents the E’o of the reversible step in an irreversible reaction. (d) The apparent reduction potentials of the various forms of pyruvic acid vary with PH. 3. An interpretation of the conflicting data in the literature concerning the keto-enol tautomerism of pyruvate ion is given on the basis of some new facts presented in this paper.




Yhc Synthesis from pNaphthohydroquinone of a Tautomer of .l-Benzyl-l,Z-naphthoquinone BY LOUISF. FIESER AND MARYFIESER i n the course of an investigation with Bradsher’ it was observed in preliminary experiments that 4-benzyl-l,2-naphthoquinone (I) can be converted by treatment with concentrated sulfuric acid or with alkali into a yellow substance which, as shown in the present work, is a monomer isomeric with the starting material. In undertaking 0


structure of the acidic tautomer, 11, even though there is no analogy in quinone chemistry for the independent existence of such a compound. It was thought that if this were the structure the substance might be obtainable more readily than through the rather inaccessible quinone I by application of a reaction employed successfully by Fieser and Hartwel12for the preparation of a similar compound, 2-hydroxy-1,4-naphthofuchsone-l (111). This substance, or its tautomer, was ob0







ri further study of the problem, the possibility was considered that the yellow substance has the

(1) Fieser and Bradsher, THISJOURNAL, 61, 417 (1939).

111 (2) Fieser and Hartwell, ibid., 67, 1484 (1935).