FRUCTOKINASE FROM RAT LIVER I. PURIFICATION AND PROPERTIES

SUMMARY Fructokinase (ATP:D-fructose-r-phosphate transferase, EC 2.7.1.3) from rat liver has been purified 4oo-fold. The purification procedure involves an acid treatment, a heat step at 65”. (NH,),SO, fractionation, chromatography on Sephadex G-100 and finally (NH&SO4 extraction. The enzyme appears nearly homogeneous by density gradient centrifugation but gives a single peak in sedimentation velocity analysis. Purified liver fructokinase has a K, of 0.46-0.80 mM for fructose and 1.56-1.33 mM for MgATP at a K+ concentration of 0.4 and 0.1 M, respectively. The enzyme also phosphorylates L-sorbose and D-tagatOSe. No difference could be found in the phosphorylation of the pyranose and furanose forms of fructose. The enzyme is inhibited by P-chloromercuribenzoate and is stable up to 50-55”. fructokinase catalyzes the phosphorylation of L-sorboselOJa. Also we have shown that the enzyme is active towards D-tagatose. The purity of both sugars was ascertained by paper chromatography and shown to be free of fructose. Moreover, the change of enzyme activity with temperature is very similar for D-fructose, L-sorbose, and D-tagatose. Measurement of the equilibrium of aqueous fructose solution indicates 80% pyranose, 20% furanose and considerably less than 0.1% of the open form33-35. Which of those forms is the substrate for fructokinase? hexo-kinase3s enzyme phosphorylates @-D-fructofuranose. Our more question remains open-chain form preferred.


INTRODUCTION
Numerous biochemical and physiological investigations1-24 clearly indicate that fructose r-phosphate is the sole initial product of the hepatic utilization of fructose. The phosphorylation of fructose is catalyzed by fructokinase (ATP:Dfructose r-phosphate transferase, EC 2.7.1.3). This enzyme presents an interesting behaviour that distinguishes it from other hexokinases as it shows an absolute requirement for K+ (ref. 7). It is also strongly and noncompetitively inhibited by one of the reaction products, ADP7. The present study was undertaken to obtain information concerning these problems. Furthermore, although fructokinase has been purified by various methods5T7J0J8, little is known about the chemical properties of the enzyme. This paper reports a method for the purification of fructokinase from rat liver to near homogeneity as well as some chemical properties of the protein.

Chemicals
Fructose, ATP (disodium salt), and creatine kinase were purchased from Sigma Chemical Company; lysozyme 3 times crystallized from Mann Research Laboratories; and Sephadex G-100 from Pharmacia Fine Chemicals. (NH&SO, was recrystallized from EDTA according to the method of SUTHERLAND AND WOLISAT~~.

Analytical techniques
Fructose was determined by the thiobarbituric acid method according to the method of PERCHERON~~ in a total volume of 1.2 ml. Protein was estimated according to the procedures LOWRY et aL2' and WARBURG AND CHRISTIAN~~ (in connection with gel filtration). Measurement of protein in fractions of the sucrose gradient was determined also by the method of MURPHY AND KIES~~.

Enzyme assays
Assay a. A calorimetric assay was based on the measurement of the fructose disappearance with a similar test as the one described by HERS~. The standard incubation mixture was as follows: 0.25 pmole of fructose, 0.3 pmole of ATP, 5 pmoles of Tris buffer (pH 7.4), 25 pmoles of KCl, 0.3pmole of MgCl,, 0.375 ,umole of creatine phosphate, I ,ug of creatine kinase, and enzyme in a total volume of 0.05 ml. Incubation was carried out for 15 min at 37' and the reaction was stopped by adding 50 ,ul of 0.15 M ZnSO, and 50 ,ul of 0.15 M Ba(OH),. After centrifugation, aliquots of the supernatant were used for the thiobarbituric acid reaction. An incubation mixture to which ATP was added after stopping the reaction by adding ZnSO,, was used as a blank.
Fructokinase activity was also measured by coupling the formation of ADP with phosphoenolpyruvate and pyruvate kinase. The pyruvate formed was estimated either calorimetrically with 2,4_dinitrophenylhydrazine or spectrophotometrically with lactate dehydrogenase.

Assay b.
A standard incubation mixture was as follows: 0.25 pmole of fructose, 0.3 pmole of ATP, 5 pmoles of Tris buffer (pH 7.4), 25 pmoles of KCl, 0.3 pmole of M&l,, 0.25 pmole of phosphoenolpyruvate, 3 ,ul of pyruvate kinase, and enzyme in a total volume of 0.05 ml. Incubation was carried out for 15 min at 37' and the reaction was stopped by heating the tubes for I min at 100'. After cooling, IO ,ul of a mixture of NADH (0.1 pmole) and lactate dehydrogenase (I pg) were added, followed by a second incubation for 5 min at 37'. Water was added to 0.5 ml and the change in absorbance at 340 nm was followed in Beckman DU spectrophotometer.
A blank without fructose was used as reference.
This assay is not suitable to measure enzyme activity in the first fractions of the purification procedure, as the addition of ATP alone causes a rapid ADP formation (presumably due either to ATPase or adenylate kinase which might be inactivated during the heat treatment).
Assay c. The standard incubation mixture was as for Assay b, but the reaction was stopped by adding 80 ,~l of 10% trichloroacetic acid, followed by 0.15 ml of 2,4dinitrophenylhydrazine (0.1% in 2 M HCl). After 5 min at 37", 0.4 ml of 2.5 M NaOH was added, and the tubes were mixed and centrifuged. The absorbance of the supernatant at 520 nm was measured.
unit One unit of fructokinase is defined as the amount of enzyme that will catalyze the disappearance of I ,umole of fructose or the formation of I ,umole of ADP or fructose r-phosphate per min at 37".

Gel jiltration
Sephadex G-100 was prepared by swelling in distilled water on a boiling-water bath for I h. The gel suspension was allowed to cool down and settle. The supernatant was poured out, removing small particles. The cleaning cycle was repeated until the supernatant was cleared. Usually it was sufficient to perform the operation twice.

Sucrose gradient centrifugation
Sucrose gradients were 5-20% (w/v) in 0.05 M Tris buffer (pH 7.4) prepared by the method of MARTIN AND AMES 30. The gradients had a volume of 4.6 ml. A 30-,ul sample (200 pug protein) was layered on top. After running for 5 h at 40 000 rev./min in a SW-50 rotor in a Spinco Model L, the rotor was allowed to coast to a stop. Fractions of 150 ~1 were collected. As sedimentation control lysozyme was used. Lysozyme activity was measured in a mixture containing 3 mg of Bacillus stearothermophilus 1503-4R (ref. 31), 5 ~1 of 0.1 M EDTA (pH 7), 60 ,~l of 0.01 M Tris buffer (pH 7.5), and 15 (~1 of each gradient fraction. Incubation was carried out for 15 min at 20°, and the absorbance at 520 nm was measured.

Disc gel electrophoresis
Electrophoresis of the enzyme was carried out in polyacrylamide gel (pH 8.9) by the method of ORNSTEIN AND DAVIS 32*3Q. The degree of polymerization of the gel was 7.5%; it was stained for proteins with 1% Amido Schwarz in 7% acetic acid for 30 min and washed continously with 7% acetic acid for 24-36 h or run through electrophoresis. Runs were performed at 4' for 2 h with 5 mA per tube. Fructokinase activity in the gel was measured after electrophoresis by cutting the gel cylinder longitudinally. One-half was used for staining. The other half was divided in 2-mm thick sections. Enzyme activity was tested with each section according to Assay a.

Analytical ultracentrif%gation
Ultracentrifuge studies were done in a Spinco Model E analytical ultracentrifuge. Analyses were performed with the enzyme dissolved in 0.01 M Tris buffer (pH 7.5). Runs were for 214 min at 42000 rev./min and at 7".
The sedimentation constant of the protein was calculated according to the procedure of SCHACHMAN~~.

Conformation studies
Differently prepared solutions of fructose were used in these studies. Solution a was a 50 mM fructose solution prepared 24 h before use; this solution corresponds to an equilibrium mixture of the pyranose and furanose forms. Solution b was prepared by dissolving, just before incubation started, the amount of crystalline fructose B&him.
(pyranose) necessary to make a 50 mM solution. The time lapse between the addition of water to dissolve the fructose and the start of incubation was approx. 15 sec.
Activity determinations with both fructose solutions were carried out under the condition of Assay b except that the incubation times were as follows: 0.5, I, 1.5, 2, 4, and 6 min (Fig. 7). The ADP formed in the reaction was estimated as described in Assay b after stopping the reaction by heating the tubes for I min at 100'.

PuriJication
of fmctokinase Wistar rats weighing 150-200 g were decapitated and allowed to bleed. Livers were removed, sliced and chilled in ice. The slices were freed of fat tissue and repeatedly washed in ice-water to remove blood. Liver slices were then weighed and homo- Biochim. Biophys. Acta, 227 (1971) 67-78 at 80 nm After 30 min stirring, the precipitate was collected by centrifugation for 15 min at 20000 x g and dissolved in ice-cold water. The solution was dialyzed for 3 h against 500 vol. of 0.01 M Tris buffer (pH 7.5). The dialysis step could be omitted without modifying the purification obtained by gel filtration. I ml of enzyme solution (50 mg protein) was applied to a 2 cm x 80 cm column of Sephadex G-100 previously equilibrated overnight with 0.01 M Tris buffer (pH 7.5). The column was eluted with the same buffer, collecting r-ml fractions at a flow rate of 0.2 ml/min. The fractions containing fructokinase activity (Fig. I) were pooled and freeze-dried to about one-hundredth volume of the homogenate. The concentrated preparation was then dialyzed overnight against 0.01 M Tris buffer (pH 7.5).
(NH&SO, solution (saturated at room temperature, pH 6.8) was added to the dialysate to bring the enzyme solution to 70% saturation. After standing for 30 min, the precipitate was collected by centrifugation for 20 min at 20 ooo x g. The precipitate was extracted twice with 0.4 ml of 70,50, and 30% saturated (NH&SO,, successively. In each case the remaining precipitate was centrifuged for 20 min at 20 ooo x g. The last two extractions with 30% (NH&SO, contained the fructokinase activity. The enzyme was purified 25o-4oo-fold over the initial homogenate with an overall yield of 30%. Typical data of a purification are shown in Table I. The profile of the enzyme is depicted in Fig. 2. A single absorption peak calculated by the difference between the absorbances at 215 and 225 nm (ref. 29), closely fitting the fructokinase activity, was observed. The profile or the position of the peak was not affected by 0.25 M KC1 and 0.04 M mercaptoethanol. The sedimentation coefficient with lysozyme as a standard was found to be 2.85 (ref. 30).
Ultracentrifkgal pattern. The sedimentation pattern of the enzyme preparation is shown in Fig. 3. The boundary of the purified enzyme moved as a single symmetrical peak with a sedimentation constant of s~,,~ = 2.95 S.

Disc gel electrophoresis.
When the purified enzyme preparation was subjected to electrophoresis in polyacrylamide gel at pH 8.9, two bands of protein migrating towards the cathode were observed. Both bands, as can be seen in Fig. 4, overlap with the enzyme activity detected in gel slices as described under MATERIALS AND METHODS.   Profierties of enzyme Stability. No loss of activity occurs upon storage of the purified enzyme in 30% saturated (NH&SO, or in 0.01 M Tris buffer (pH 7.4) at -15' for at least 7 months. When fructokinase activity was tested as a function of temperature, the enzyme was stable up to 50-55' and then decreased. A similar behaviour was found out with other ketoses also phosphorylated by fructokinase, as can be seen in Fig. 5. From these data the energy of activation for the phosphorylation of fructose was calculated as II kcal per mole and a Qr,, of about 2. A similar value was obtained for sorbose.

K&et&.
The broad pH optimum already reported5 was also found for the   As shown in Fig. 6, proportionality between fructose transformed and enzyme concentration or time only held when the reaction was allowed to proceed to a small extent in the absence of an ATP-regenerating system. This behaviour has been shown to be caused by inhibition of the ADP formed in the reaction'.
The Michaelis constants determined for fructose and MgATP are presented in Table II for different K+ concentrations.
It should be noticed that an increase in K+ concentration brings a diminution of the Km for fructose.
Substrate specz&ity: structure. The purified enzyme phosphorylates also Lsorbose and D-tag&OS& Moreover, K + stimulation of enzymatic activity towards these sugars can be seen in Table III phosphate, fructose 1,6-diphosphate, glucose r-phosphate, dihydroxyacetone phosphate, and glyceraldehyde have no effect on the activity of fructokinase. Substrate specijicity: conformation.
It can be seen in Fig. 7 that fructokinase phosphorylates at the same rate a newly prepared solution of fructose (pyranose form) and a solution of fructose where the pyranose and furanose forms were equilibrated.
Efects of thiols. It has been shown that fructokinase can be inhibited by p-chloromercuribenzoategJo~l*, suggesting the presence of essential sulthydryl groups for enzyme activity. When the purified enzyme was treated with fi-chloromercuribenzoate, it was found that 50% inhibition was attained at a concentration of 3.2 * IO-~ M at 37" (see Fig. 8). However, the same p-chloromercuribenzoate concentration caused 100% inhibition at 55". The concentration of p-chloromercuribenzoate which causes 100% inhibition, was found for the phosphorylation of sorbose to be 2.5 *IO-~ M and for tagatose 1.75. IO-~ M. Molecular weight. The molecular weight determined by density gradient centrifugation using lysozyme as standard, according to the method of MARTIN AND AMES~~, averaged 28 ooo.
Absorption spectra. In Fig. 9 are shown the ultraviolet spectra of the purified protein at both neutral and alkaline pH. The A,,, nm/A260 nm ratio was 1.2. The characteristic absorption maximum at 278 nm contributed by aromatic amino acid residues was seen at neutral pH.

DISCUSSION
The purification procedure reported here for fructokinase from rat liver permits the isolation of the enzyme with a yield of about 35% by a rather small number of purification steps.
The enzyme preparation appears to be homogeneous by sedimentation velocity, but only near-homogeneous by density gradient centrifugation. By disc gel electrophoresis two protein bands were obtained. However, both proteins showed enzyme activity. The significance of this result is not clear, but it could indicate the existence of two enzyme forms. These results are approximate and require further study, but nevertheless they might have some significance.
The fructokinase is very stable under a variety of conditions, such as aging, heat inactivation and acid denaturation. The enzyme possesses sulthydryl groups which are related to the activity, but contradictory results regarding its number are obtained with the p-chloromercuribenzoate experiments. Effectively, assuming a molecular weight of 30 ooo for the protein, the number of sulfbydryl groups calculated according to the inhibition data would be about I at 55", which is not in agreement with the number obtained at 37' which is about 2. A possible explanation for the difference observed could be a change in conformation of the enzyme with increased temperature.
The K, values reported here for fructose (0.46-0.8 mM) and for MgATP (1.56-1.33 mM) at K+ concentrations of 0.4 and 0.14 M, respectively, are in good agreement with those previously found by ADELMAN et aL1* and HERS~. The value for MgATP obtained by PARKS et al.? is about IO times lower than any other xeported.
We have confirmed that fructokinase catalyzes the phosphorylation of L-sorboselOJa. Also we have shown that the enzyme is active towards D-tagatose. The purity of both sugars was ascertained by paper chromatography and shown to be free of fructose. Moreover, the change of enzyme activity with temperature is very similar for D-fructose, L-sorbose, and D-tagatose. Measurement of the equilibrium of aqueous fructose solution indicates 80% pyranose, 20% furanose and considerably less than B&him. Biophys. Ada. 227 (1971) 67-78