COMMUNICATIONS TO THE EDITOR
1010
Vol. 75
71% of the original activity. Only partial re- phosphate (formed from fructose- 1,6-diphosphate activation occurred in the absence of Mg++ ions. by aldolase) as acceptor aldehyde, the decarboxylaThe spinach enzyme also catalyzes the reaction tion of hydroxypyruvate led to the formation of of L-erythrulose and ~-glyceraldehyde-3-phosphate ribulose-%phosphate. The pentose phosphate was to form a mixture of pentose phosphate and hep- isolated as an alcohol insoluble barium salt and tulose phosphate. Neither product is formed in the determined by two independent tests as shown in absence of T h P P (Table 11). Table I . Similar results were obtained when DLWhile the mechanism of sedoheptulose phosphate glyceraldehyde-&phosphate (Concord Laboraformation is not yet known, the reactivity o f tories) was used instead cf fructose-1,6-diphosphate erythrulose in this system suggests that it may be and aldolase. formed by the reactions TABLE I 2 n-ribulose-5-phosphate L-erythrulose 2 ~-glyceraldehyde-3-phosph:lte ( 1 ) L-erythrulose ~-glycerdldehyde-3-phosphate-+ sedoheptulose-7-phosphate ( 2 i
+
+
An alternative mechanism, supported by the requirement for ThPP for sedoheptulose phosphate synthesis from erythrulose, would be a condensation of ribose phosphate with an active two-carbon fragment. Pentose phosphate isomerase is still present in the enzyme preparation and the participation of ribose phosphate has not been excluded. In either case an activated form of glycolaldehyde, formed in the cleavage of pentose phosphate, wculd undergo an acyloin condensation. The synthesis of acetoin in such reactions is known to require ThPP.g~ioThe name transketolwe, suggested by Racker, de la Haba and Lederil is consistent with this formulation. In the presence of spinach enzyme pentose formation from erythrulose is observed with other aldehydes, such as D-glyceraldehyde and L-glyceraldehyde-3-phosphate. (9) M. Silverman and C. H. Werkman. J . Biol. Chrin., 138, 35 (1941).
(IO) D. E. Green, W. W. Westerfeld, B. Vennesland and W. E. Knox, J . B i d . Chem , 145, 69 (1942). (11) E. Racker. G . de la Haba and I . G. Leder, 1010 (1953).
ENZYMATIC FORMATION OF RIBULOSE-5-PHOSPHATE F R O % HYDROXYPYRUVATE AND TRIOSEPHOSPHATE 0.5 mg. of purified yeast transketoldse (22,000 units per mg. protein) was used in these experiments. Carbon dioxide was measured manometrically, In Expt. 1, the reaction mixture (2 ml.) contained 100 micromoles of potassium phosphate (pH 6.5), 5 micromoles of fructose-l,6-diphosphate, 20 micrograms of aldolase, 12 micromoles of MgCl?, 20 micrograms of ThPP and about 30 micromoles of sodium hydroxypyruvate. In Expt. 2, 100 micromoles of tris(hydroxymethy1)-aminomethane(PH 6.9) was used instead of potassium phosphate and the concentration of fructose1,6-diphosphate was increased to 10 micromoles. The vessels were incubated a t 37” for 75 minutes in Expt. 1 and 175 minutes in Expt. 2. Deproteinization with 5% trichloroacetic acid was followed by the isolation of an alcoholinsoluble barium salt which was analyzed colorimetrically ds well as spectrophotometrically. In the latter test trans. ketolase free of pentose isomerase was used and triose phosphate formation was measured with either glycolaldehyde, glyceraldehyde, or ribose-5-phosphate as “acceptor aldehvdes.” Rxpt.
CO? liberation, micromoles
1 2
Iso!ated ribulose-5-phosphate, micromoles Orcinol reaction Spectrophotometric
4 .R
1.8 3.1
4 0
1.6 2.9
TABLE I1 THIAMINE PYROPHOSPHATE REQUIREMENT OF TRASSKETOLASE
THISJ O I J R N A L , 76,
The enzyme preparation was dialyzed against 1000 volumes of 0.6% Versene in 0.02 M potassium phosphate of pH 7.4 for 20 hours aqd then against 1000 volumes of 0.6% Versene in 0.9% RC1 for another 20 hours. The enzyme B. L. HORECKER was assayed by measuring triose phosphate formation from P. Z. SMYRNIOTIS ribulose-5-phosphate in the presence of ribose-5-phosphate as “acceptor aldehyde.”
NATIONAL INSTITUTE O F ARTHRITIS AND METAB:>LIC DISEASES INSTITUTESOF HEALTH KATIOSAL PUBLICHEALTH SERVICE FEDERAL SECURITY AGENCY RETHESDA 14, MARYLAND RECEIVED JANUARY 30, 1953
THIAMINE PYROPHOSPHATE, A COENZYME OF TRANSKETOLASE
Enzyme preparation
1-ndialyzed Dialyzed for 40 hotirs
Additions t o test system
Activity (units per ml.)
.......
58,000
.......
2 , on0 5, 000
3 phf. MgC12 50 pg. ThPP and 3 911.MgC12
.%: 43 ,000 In a note by Horecker and Smyrniotis’ previous work on enzymes concerned in the breakdown of Dialyzed for 40 hours ....... 5no pentose phosphate is quoted. We have isolated then left in icehox for 3 WM.MgC12 I ,on0 from baker’s yeast a crystalline enzyme which 24 hours 59 pg. of ThPP 7 ,noo catalyses the cleavage of ribulose-5-phosphate with 50 pg. of ThPP 40,000 the formation of D-glyceraldehyde-3-phosphate, and 3 p M . of identified by means of glyceraldehyde-3-phosphate nilgcir dehydrogenase free of triose isomerase. The Since the formation of ribulose-5-phosphate cleavage of ribulose-5-phosphate2 occurs only on the addition of an “acceptor aldehyde” such as represents a keto1 condensation, and no free glycolribose-5-phosphate, glycolaldehyde, or glyceralde- aldehyde is formed, one must assume the formation hyde. The enzyme was also found to decar- of an “active glycolaldehyde” which condenses boxylate hydroxypyruvate in the presence of an with the “acceptor aldehyde” to form a ketosugar. “acceptor aldehyde.” With o-glyceraldehyde-3- The enzyme may therefore be termed a transketolase. ( 1 ) B. L Horecker and P Z. Smymiotis, THISJOURNAL, 76, 1009 t 1953).
( 2 ) We wish to thank Ilr. B I, Horecker for a gift of rihulose-5phosphate.
(3) A similar reaction catalyzed by rabbit muscle mince has been described b y s. Akabori, Kihachiro Usbara and I. Muramatsu, Proc. J a p a n Academy, 28, 39 (1952).
Feb. 20, 1953
COMMUNICATIONS TO THE EDITOR
The activity of a partially purified transketolase from E . coli was doubled by the addition of thiamine pyrophosphate (ThPP). No requirement for T h P P was found with a twice recrystallized preparation of the yeast enzyme, but extensive dialysis against a Versene-KC1 solution caused nearly complete inactivation. Addition of magnesium chloride and of T h P P to the dialyzed enzyme restored the activity, as shown in Table IT. The crystalline yeast enzyme shows no aldolase, triose isomerase or pentose isomerase activity.
1011
and decompose in a few minutes. An unstable dark brown crystalline picrate is obtained by mixing ether solutions of bis-cyclopentadienylnickel (11) and picric acid in presence of air (Anal. Calcd. for ClaH12N30,Ni: Ni, 14.1. Found Ni, 14.0). On the above view, the nickel atom in bis-cyclopentadienylnickel(I1) should have two electrons in excess of the krypton structure, which would be expected t o occupy the 5s orbital. Magnetic susceptibility measurements show, however, that bis-cyclopentadienylnickel(I1) has two unpaired = +3440 X c.g.s.u. corrected DEPARTMENT OF PHYSIOLOGICAL CHEMISTRY E. RACKER electrons (~2:~. YALEUNIVERSITY G. DE LA HABA for diamagnetic contribution, peff = 2.88 B.M.). NEWHAVEN,CONNECTICUT I. G. LEDER In our view, this fact may be best accommodated RECEIVED JANUARY 30, 1953 by assuming that in these bis-cyclopentadienyl compounds, the metal ion utilizes three of the elecBIS-CYCLOPENTADIENYL DERIVATIVES OF SOME trons from each of the two cyclopentadienyl anions, forming bonds involving the s and two d orbitals of TRANSITION ELEMENTS Sir: the metal. In the case of Ni(II), the formation of The structure suggested' for bis-cyclopenta- CloHloNi would involve promotion of two electrons dienyliron(11)2 (ferrocene), in which the iron atom to the 4p orbitals, which must be singly occupied. is symmetrically placed between two cyclopenta- Further, we have been unable to obtain a bisdienyl rings, has been confirmed by X-ray crystal cyclopentadienyl derivative of Cu(11)) which structure measurements. l b v 3 The original proposals' has only one d orbital available, and cyclopentaof this formulation were coupled with the suggestion dienyl derivatives of the zinc group show properties that the electronic structure of the iron atom attains of typical organometallic compounds. We have, an inert gas configuration, and this idea could also however, been able to prepare bis-cyclopentadienyl be extended to the ruthenium analog C1oHloRu4 compounds of titanium, zirconium and vanadium, and to the bis-cyclopentadienylcobalt(II1) (co- where sufficient electrons are not available for balticinium) ion [C1oHloCo]+,l b s 6 which is isoelec- completion of an inert gas configuration of the metal atom. tronic with ferrocene. Bis-cyclopentadienyltitanium(1V)dibromide has In addition to the objection that a high negative charge would be placed on the central metal atom, been prepared by the reaction of excess cyclopentathe aromatic propertiesa of ferrocene make i t seem dienylmagnesium bromide with titanium tetramost unlikely that all the T electrons of the cyclo- chloride in toluene solution. It forms dark red pentadienyl rings can be involved in the filling of crystals, m.p. 240-243') from toluene (Anal. the orbitals of the metal atom. More definite Calcd. for CloHloTiBrz: C, 35.6; H, 3.0; Ti, 14.2; Br, 47.3. Found: C, 36.0; H, 3.1; Ti, 14.3; evidence against this view has now been obtained. Bis-cyclopentadienylnickel(I1) has been pre- Br, 47.3) and is diamagnetic (xzzl = - 145 X pared by the reaction of cyclopentadienylmag- lop6 c.g.s.u.). It is to some extent hydrolyzed by nesium bromide with nickel(I1) acetylacetonate. water, giving a yellow solution, which gives pre(Anal. Calcd. for CloHloNi: C, 63.6; H, 5.3; Ni, cipitation reactions similar to those of other bis31.0. Found: C, 63.7; H, 5.5; Ni, 31.0). I t cyclopentadienyl metal ions. A crystalline picrate forms dark green crystals from ligroin which de- (m.p. 139-141') explodes) has been isolated (Anal. compose slowly even in absence of air and light; Calcd. for C22H14N6014Ti: Ti, 7.55. Found: Ti, it sublimes above 130' but decomposes below the 7.48). Aqueous perchlorate solutions show a melting point. Solutions in an alcoholic supporting polarographic cathodic wave a t - 0.44 volt versus electrolyte show a polarographic anodic wave a t the saturated calomel electrode. Controlled po- 0.08 volt versus the saturated calomel electrode. tential reduction, or reduction using a Jones reThe yellow solutions obtained by oxidation contain ductor, produces a green solution containing the the bis-cyclopentadienylnickel(II1) ion, [Clo. bis-cyclopentadienyltitanium(II1) ion which shows HloNi]+, and give precipitates with silicotungstic a polarographic anodic wave a t -0.44 volt. acid, Reinecke's salt, potassium triiodide, etc., as The almost colorless bis-cyclopentadienylzirdo the ferricinium, r u t h e n i c i n i ~ mand ~ cobalti- conium(1V) dibromide (Anal. Calcd. for CloH'ocinium5 ions. Aqueous solutions of the bis-cyclo- ZrBr2,C, 31.5; H, 2.7; Zr, 24.0; Br,41.9. Found: pentadienylnickel(II1) ion are rather unstable, C, 31.4; H , 2.7; Zr, 23.9; Br, 42.0) (m.p. 260' C. dec.) was prepared from zirconium tetrachloride ( l ) ( a ) G . Wilkinson, M. Rosenblum, M. C Whiting and R . B. and cyclopentadienylmagnesium bromide. Aqueous Woodward, THISJOURNAL, 74, 2125 (1952); (b) E. 0. Fischer and W. Pfab, 2. Naturforsckung, 78, 377 (1962). solutions of this compound, show no polarographic @)(a) T. J. Kealy and P. L.Pauson, Nntztrc, 168, 1039 (1951); (b) reduction wave. S. A. Miller, J. A. Tebboth and J. F. Tremayne, J . Chcm. SOL.,632 Vanadium tetrachloride reacts similarly, forming a (1952). (3) (a) P. F. Eiland and R . Pepinsky. THISJOURNAL,74, 4971 dark green, rather unstable, ligroin soluble, bromide (l952), (b) J. D. Dunitz and L. E. Orgel. Nature, 171, 121 (1953). (Anal. Calcd. for CloHloVBrz: V, 14.9; Br, 46.9. (4) G. Wilkinson, THISJOURNAL, 74,6146 (1952). Found: V, 14.8; Br, 47.1) and a pale green, ligroin (&5) G Wilkinson, ibid., 74, 6148 (1952). insoluble chloride (Anal. Calcd. for CloHloVClz: (6) R B . Woodward, M. Rosenblum and M. C . Whiting, i b i d . , 74, C, 47.6; H, 4.0; V, 20.2; C1, 28.2. Found: 3458 (1952).
