Diadenosine Tetraphosphatase from Human Leukemia Cells PURIFICATION TO HOMOGENEITY AND PARTIAL CHARACTERIZATION*

Diadenosine tetraphosphatase, an enzyme splitting diadenosine tetraphosphate to AMP and ATP, has been purified to apparent homogeneity from a permanent cell line derived from a leukemic child. The purification procedure consisted of fractionation by ammonium sul- fate precipitation, followed by Sephacryl 200 and DEAE-cellulose chromatography, and finally a differential membrane filtration. The enzyme is a single poly- peptide chain of M, = 17,500 as determined by gel electrophoresis in the presence of sodium dodecyl sulfate. The apparent molecular weight of the native en- zyme was calculated as 20,000 from gel filtration data. The apparent K,,, for ApIA was 0.5 PM as determined by two independent kinetic assays. None of the following compounds were substrates of the enzyme: diadenosine triphosphate, NAD, nucleoside 5'-phosphates (AMP, ATP, GDP, GTP, and W). The enzyme had optimal activity in the presence of 1 m~ MgZ+, showing no activity in the presence of EDTA. Several lines of evidence support the hypothesis of diadenosine 5',5"'-P',P4-tetraphosphate to play an essential role in controlling growth and cell division. The intracellular concentration of the unusual dinucleotide fluctuates rapidly and is to the proliferative activity of those cells (1). initi-ation

Several lines of evidence support the hypothesis of diadenosine 5',5"'-P',P4-tetraphosphate to play an essential role in controlling growth and cell division. The intracellular concentration of the unusual dinucleotide fluctuates rapidly and is directly related to the proliferative activity of those cells (1). In resting permeabiliied baby hamster kidney cells, the initiation of DNA replication could be stimulated by Ap4A' (2).
The possible molecular function of Ap4A was demonstrated with preparations of DNA polymerase-a from calf thymus showing specific noncovdent binding of labeled Ap4A (3) and with DNA polymerase-a from HeLa cells catalyzing DNA synthesis with utilizing Ap4A 8s a primer (4).
The synthesis of diadenosine tetraphosphate was discovered by Zamecnik et al. with a reaction mixture containing lysyl-tRNA synthetase, ATP, Mg2+, and lysine, and seems to be a special property of some aminoacyl-tRNA synthetases with zinc as a specific trigger (5-7). A diadenosine tetraphosphate degrading enzyme (diadenosine tetraphosphatase) has been described and partially purified from rat liver (8), Artemia salina (9), Physarum polycephalum (lo), and mouse ascites tumor cells (11). Here, we describe tbe homogeneous preparation of the enzyme from human cells and some of its characteristics.
* This work was supported by Grant F3/SFB 118 from the Deutsche Forschungsgemeinschaft. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Protein Determination
Protein was measured with the Bio-Rad protein determination kit (14) and bovine serum albumin as a standard.

Assay for Diadenosine Tetraphosphatase
Coupled Luminescence Assay-The assay is based on luciferinluciferase producing light with the ATP that is generated from Ap4A after diadenosine tetraphosphatase action. The assay has been described in detail (11). For determining enzyme activities in fractions of the purification steps, the assay (final volume 0. After having dried the plates, a second run with 1 M LiCl was performed resulting in a clear separation of the nucleotides. The nucleotides were marked under ultraviolet light and cut off. The thin layer pieces were overlayed with the scintillation mixture (toluene; 0.5% 2,5-diphenyloxazole, 0.03% 1,4-bis-(4-methyl-5-phenyloxazol-2-y1)benzene) and counted. Counting efficiency of tritium was 6%.

RESULTS
Purification of Diadenosine Tetraphosphatase-All steps were carried out at 2-4 "C unless otherwise indicated.
Preparation of Crude Extract-About 5 ml of packed, washed cells from the human leukemic cell line RU-3 (13) were suspended in 25 ml of phosphate-buffered saline. Disruption of cells was performed by homogenizing with a Potter-Elvehjem homogenizer equipped with a Teflon pestle for 10 min. The homogenate was first centrifuged for 15 min at 800 x g. The supernatant was again centrifuged for 1 h at 4 8 , The homogenization of the pellet was repeated two times. To the combined supernatants, magnesium chloride (2 m~) , Na2EDTA (0.2 m~) , and 2-mercaptoethanol (10 m~) were added. The enzyme was kept frozen at -25 "C. Ammonium Sulfate Fractionation-Ammonium sulfate was added to the supernatant to 30% saturation. After being stirred for 1 h, the suspension was centrifuged for 15 min at 11, OOO x g. To give 80% saturation, the supernatant of the previous step was again stirred with crystalline ammonium sulfate. After centrifugation, the precipitate was dissolved in 5 ml of buffer (50 m~ Tris-HC1, pH 7.7, 2 m~ MgC12, 10 mM 2-mercaptoethanol, 0.2 m~ NaZEDTA).
Sephacryl 200 Chromatography-The enzyme obtained from the previous step (22.5 mg) was applied to a Sephacryl 200 column (2.5 X 85 cm) equilibrated with buffer A. The enzyme was eluted from the column with a constant flux (15 ml/h) of buffer A. Fractions of 2.5 ml were collected and tested for enzyme activity with the coupled luminescence assay (Fig. 1). The peak of the enzyme approximately coincided with the peak of myoglobin as measured by a separate chromatography, suggesting a low molecular weight of the enzyme.
DEAE-cellulose Chromatography-The fractions of the Sephacryl chromatography with enzyme activity were pooled and directly applied to a DEAE-cellulose column (2.5 X 50 cm) equilibrated with buffer A. The enzyme was eluted (12 ml/h) from the column with a linear gradient of 0-0.2 M NaCl (in buffer A) for 16 h, and the fractions (3 ml) were again tested for enzyme activity and pooled (Fig. 2). This purification step seemed to be rather ineffective, yielding only a small increase in specific activity ( Table I). No explanation for the lability of the enzyme during this procedure could be found. On the other hand, no further efforts have been made to stabilize the enzyme by additives, for instance albumin, etc.
UZtrafiZtration-Using an Amicon concentrating system (8 was applied to a DEAE-cellulose column and eluted by a NaCl gradient as described in the text. Aliquota (10 p l ) of each fraction were assayed for diadenosine tetraphosphatase activity. Enzyme activity is expressed in arbitrary units (  * Activity is the increase in luminescence in the standard assay at 25 "C expressed in counts/min X specific e Activity is measured by following the degradation of [3H]Ap4A at 25 "C.
e The activity of the S-48 fraction is taken as 100%. The increase in total,activity upon differential precipitation with ammonium sulfate is thought to be due to the removal of inhibitors of the coupled enzymatic assay, Le. other ATP-consuming proteins.
by guest on March 23, 2020 http://www.jbc.org/ Downloaded from MC), the pooled fractions of the DEAE-cellulose chromatography with enzymatic activity were filtered under N B pressure through a Diaflo membrane PM-10. About 50% of enzyme activity passed through the membrane and was obtained from the eluate. The residual concentrate was diluted with 10 ml of buffer and again filtered. The combined eluates were concentrated in the same apparatus equipped with a Diaflo membrane UM-2. The eluate was free of enzymatic activity. Part of the concentrated enzyme was frozen at -25 "C, another part was diluted with an equal volume of pure glycerol and also kept at -25 "C. This preparation step was very effective, achieving about a 20-fold purification. When kept frozen, the enzyme was stable for several weeks. To characterize the enzyme, all of the following experiments were performed with this material. The complete purification procedure is summarized in Table I.
Determination of Molecular Weight and Purity-when examined with SDS-polyacrylamide gel electrophoresis on gradient gels as outlined under "Experimental Procedures," diadenosine tetraphosphatase showed a single band after protein staining with Coomassie blue (Fig. 3) and was judged essentially homogeneous. By comparison of this band with the migration of proteins of known molecular weight, that of denatured diadenosine tetraphosphatase (Fig. 4) appeared to be 17,500 which is the mean of several determinations.
The enzyme was also chromatographed together with molecular weight standards on Sephadex G-75 in the absence of SDS. Under these conditions, diadenosine tetraphosphatase was found to have a molecular weight of about 20,000 (Fig. 5).
Products of the Reaction of Diadenosine Tetraphosphatase-The hydrolysis of ['HlApA by diadenosine tetraphosphatase was followed by product analysis on polyethyleneimine-cellulose thin layers with authentic markers as described under "Experimental Procedures." A complete degradation of Ap4A to ATP plus AMP was obtained (Fig. 6). Under these conditions, no indication for a symmetric cleavage or a further degradation of ATP to AMP was obtained. The difference in the counting efficiency of the two products ( Fig. 6) could be explained by the more expanded spot of AMP on the thin layer plate relative to that of ATP which is caused by the much longer migration distance of AMP. Enzyme Kinetics-The initial velocity of enzymatic hydrolysis of Ap4A was measured by the coupled luminescence assay as well as by thin layer analysis of the products as described under "Experimental Procedures." The double reciprocal plots of the initial velocity uersus Ap4A concentration  5 (center). Molecular weight determination of native diadenosine tetraphosphatase by Sephadex G-75 column chromatography. Human diadenosine tetraphosphatase or standard proteins were successively applied to a Sephadex G-75 column and chromatographed as described under "Experimental Procedures." T h e abbreviations used for the molecular weight standards are given in Fig. 4 The activity of the enzyme was measured following the degradation of [3H]Ap,A on thin layer plates as described under "Experimental Procedures." Activity in the absence of magnesium was measured with NaZEDTA at, 0.3 nm in excess over Mg2+, which was added with the stored enzyme. (Fig. 7, A and B ) were linear suggesting a simple Michaelis-Menten condition. The calculated K,,, for Ap4A was 0.5 p~.
Magnesium Requirement of Diadenosine Tetraphosphatase-Dialysis of the enzyme against buffer containing NazEDTA resulted in a completely irreversible loss of enzyme activity. When NazEDTA was present at 2-fold excess over Mg2+, no activity could be measured in the enzymatic assay, suggesting an absolute requirement for Mg+. Enzyme activity was also measured at various concentrations of MgClz (Fig.  8). At 0.2 mM MgC12, the activity was about 80% of its maximal value, which is reached at 1 mM.
Substrate Specificity-The following radioactively labeled compounds were tested as substrates: diadenosine triphosphate, NAD, ATP, AMP, GTP, GDP, and UTP. In every case the hydrolysis of the potential substrates was followed for 3 h under standard conditions. Aliquots of the assays were chromatographed on polyethyleneimine-cellulose thin layers with No hydrolysis of the tested compounds could be measured during the incubation time.

DISCUSSION
This report represents the fist homogeneous purification of a human diadenosine tetraphosphatase. The apparent molecular weight as determined by SDS-polyacrylamide gel electrophoresis was 17,500. This value is in good agreement with those for the partially purified enzymes from rat liver (21,000) or from A. salina (17,500) as estimated by Sephadex G-75 fitration (9). From mouse liver a diadenosine tetraphosphate hydrolase has been described with a molecular weight of 64,000 (16).
The discrepancy in the molecular weights of the human enzyme obtained from SDS-gel electrophoresis and from gel fdtration as well as the unexpected feature of the enzyme to pass an ultrafiltration membrane with a nominal exclusion limit of 10,000 suggests a molecular size of the enzyme that is nonspherical.
In accordance with the rat liver and A. salina enzyme, diadenosine tetraphosphatase from human cells splits ApA asymmetrically to ATP and AMP. The enzyme shows a high specificity for diadenosine tetraphosphate. Diadenosine triphosphate is not degraded, in accordance with reports on analogous enzymes from other sources (9,16). For diadenosine triphosphate, a specific hydrolase has been isolated from rat liver (17). This enzyme has a molecular weight of 29,800 and is different from the diadenosine tetraphosphatase isolated from the same source (8). Other dinucleotides as NAD or nucleoside 5"phosphates (AMP, ATP, GDP, GTP, and UTP) are also not substrates of the enzyme.
The human enzyme revealed a Michaelis constant for Ap4A of 0.5 PM as determined by two independent kinetic assays. The corresponding values reported for the enzyme from rat liver (2 p~) , A. salina (2 p~) , or for the mouse tumor enzyme (2.8 PM) were severalfold higher (9, 11). The requirement for Mg2+ of the human enzyme is common with the rat liver enzyme. The human enzyme reaches maximal activity with 1 m~ of M F , whereas the rat liver enzyme needs about 5 mM for optimal action (8).
The intracellular concentration of diadenosine tetraphosphate of proliferating cells was found to be in the range of 0.1-1 p~ (1). Changes in metabolic conditions or inhibition of protein synthesis triggers an immediate degradation of Ap4A that is more rapid and drastic than the drop of ATP, suggesting an essential role of an Ap4A consuming activity in regu-

109
lating the intracellular concentration of the nucleotide. Considering the high cytoplasmic activity of diadenosine tetraphosphatase as well as its apparent Michaelis constant of 0.5 p~, the enzyme probably plays that essential role in human cells. Whether the steady state level of Ap4A could be controlled by the rate of synthesis or by modulating the activity of diadenosine tetraphosphatase is an important question to be investigated.