Uridine Kinase from Ehrlich Ascites Carcinoma PURIFICATION AND PROPERTIES OF HOMOGENEOUS ENZYME*

Uridine kinase from Ehrlich ascites tumor cells has been purified about 60,000-fold to apparent homoge-neity and with an overall recovery of about 40%. This purification was achieved using phosphocellulose and adenosine 5’-triphosphate-agarose affinity chroma- tography. The subunit molecular mass as judged by sodium dodecyl sulfate-polyacrylamide gel electropho- resis was 31,000 daltons. With two-dimensional electrophoresis, only one spot was observed, indicating the absence of isoenzymes. Multiple peaks of activity are routinely observed on ion exchange chromatography or gel filtration, for both crude preparations or homogeneous uridine ki- nase, in agreement with our earlier results that this enzyme as multiple interconvertible oligomeric forms R. C., T. W. J. Biol. 257,12485-12488). The purified enzyme has a specific activity of 283 pmol/min/mg of protein at 22 “C. Initial velocity studies using uridine and ATP are consistent with a se- quential mechanism. K,,, values for uridine, cytidine, and ATP are 40, 57, and 450 MM, respectively. CTP and UTP are competitive inhibitors with respect to ATP, with Ki values for CTP and UTP of 10 and 61 p ~ , respectively. The enzyme was active with several nucleoside analogs, the K , values being 69 p~ (5- fluorouridine), 200 p~ (3-deazauridine), and 340 p~ (6-azauridine).


Uridine Kinase from Ehrlich Ascites Carcinoma
PURIFICATION AND PROPERTIES OF HOMOGENEOUS ENZYME* (Received for publication, February 27, 1985) Robert C. Paynel, Nancy Cheng, and Thomas W. Traute Uridine kinase from Ehrlich ascites tumor cells has been purified about 60,000-fold to apparent homogeneity and with an overall recovery of about 40%. This purification was achieved using phosphocellulose and adenosine 5'-triphosphate-agarose affinity chromatography. The subunit molecular mass as judged by sodium dodecyl sulfate-polyacrylamide gel electrophoresis was 31,000 daltons.
With two-dimensional electrophoresis, only one spot was observed, indicating the absence of isoenzymes. Multiple peaks of activity are routinely observed on ion exchange chromatography or gel filtration, for both crude preparations or homogeneous uridine kinase, in agreement with our earlier results that this enzyme exists as multiple interconvertible oligomeric forms (Payne, R. C., and Traut, T. W. (1982) J. Biol. Chem. 257,12485-12488).
The purified enzyme has a specific activity of 283 pmol/min/mg of protein at 22 "C. Initial velocity studies using uridine and ATP are consistent with a sequential mechanism. K,,, values for uridine, cytidine, and ATP are 40, 57, and 450 MM, respectively. CTP and UTP are competitive inhibitors with respect to ATP, with Ki values for CTP and UTP of 10 and 61 p~, respectively. The enzyme was active with several nucleoside analogs, the K , values being 69 p~ (5fluorouridine), 200 p~ (3-deazauridine), and 340 p~ (6-azauridine).
The pure enzyme is very sensitive to freezing, but can be maintained at 0 "C for 8 weeks with only 20% loss of activity. For long-term storage, enzyme in 50% glycerol can be maintained at -20 "C for many months with no detectable loss of activity.
Uridine kinase (ATP:uridine 5'-phosphotransferase, EC 2.7.1.48) catalyzes the phosphorylation of uridine and cytidine to their respective monophosphates. The enzyme is the ratelimiting activity of the pyrimidine salvage pathway whereby preformed pyrimidine nucleosides are recycled for nucleic acid synthesis (Anderson, 1973). It has been shown for many tissues, normal and neoplastic, that more UMP may be synthesized via the salvage route than from de m u 0 synthesis (Weber et al., 1978;Denton et al., 1982). Two modes of regulation have been shown for the enzyme: feedback regula-* This research was supported by Grant BC-451 from the American Cancer Society, Grant PCM-8310902 from the National Science Foundation, and a Medical Faculty grant. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisernent" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ Present address: University of Virginia, Department of Chemistry, Charlottesville, VA 22901.
To whom correspondence should be sent. tion by UTP and CTP which exhibit competitive inhibition with respect to the phosphate donor (Anderson and Brockman, 1964;Liacouras et al., 1975) and regulation by changes in quaternary structure caused by orthophosphate, ATP, and CTP (Payne and Traut, 1982a).
A number of investigators have reported multiple forms of uridine kinase from normal and neoplastic tissues when crude or partially purified preparations of the enzyme are subjected to gel filtration (Krystal and Webb, 1971;Krystal and Scholefield, 1973;Keefer et al., 1975;Greenberg et al., 1977;Otal-Brun and Webb, 1979), 'to ion exchange chromatography (Skold, 1963;Fulchignoni-Lataud et al., 1976;Dubinina et al., 1982), to native isoelectric focusing (Ahmed and Welch, 1979;Ullman et al., 1979;Absil et al., 1980;Ahmed and Baker, 1980;Ahmed, 1982), and to affinity chromatography (Veseljl and Smrt, 1977). Except for Krystal and Scholefield (1973) who suggested that the two molecular weight forms of uridine kinase observed in preparations from Ehrlich ascites cells could be a monomer and tetramer of the same species, the other investigators have invariably suggested that the multiple forms of uridine kinase constitute isoenzymes. Using normal and neoplastic tissues, gel filtration studies have shown multiple uridine kinase peaks, leading to the interpretation that one or more isoenzymes of uridine kinase are preferentially associated with the neoplastic state (Krystal and Webb, 1971;Greenberg et al., 1977). Preparations of uridine kinase from fetal and postnatal rat liver also show multiple molecular weight forms when chromatographed on Sepharose 6B, leading to the interpretation that differential expression of uridine kinase isoenzymes is taking place during development (Krystal and Webb, 1971). In addition, the development of resistance to pyrimidine nucleoside analogs has been attributed to the preferential expression of one of the uridine kinase isoenzymes (Skold, 1963;Keefer et al., 1974;Keefer et al., 1975;Greenberg et al., 1977). Previous experiments in our laboratory have suggested an additional or alternative interpretation: uridine kinase can exist in a variety of different aggregation states that can be interconverted by appropriate effectors such as substrates and inhibitors (Payne and Traut, 1982a). Even highly purified uridine kinase will exhibit several different polymer forms that only slowly equilibrate with each other; the enzyme will therefore exhibit multiple peaks on a gel filtration column.
The preparation of homogeneous uridine kinase has never been previously achieved with any eukaryotic source; most studies have used enzyme preparations that were purified less than 10-fold. In the present study, Ehrlich ascites cells were used because no other cell or tissue has higher uridine kinase activity, and since it was previously reported that Ehrlich ascites cells may contain isozymes of uridine kinase (Krystal and Scholefield, 1973). EXPERIMENTAL PROCEDURES AND RESULTS' DISCUSSION Uridine kinase from Ehrlich ascites cells has been purified 60,000-fold to apparent homogeneity. The essential steps in this purification involve two columns: phosphocellulose and ATP-agarose. Adsorption of enzyme to the P-11 column was inhibited by phosphate at concentrations greater than 1 mM even though 30-80 mM potassium phosphate was necessary to elute uridine kinase from the resin. The inclusion of citrate in the TCE (50 mM Tris base and 0.01 mM EDTA titrated to pH 6.6 with 1 M citric acid) buffer stabilized enzyme activity which otherwise is labile upon dilution in Tris buffers. P-11 chromatography increased the binding capacity of the subsequent ATP-agarose column for uridine kinase by approximately 2000-fold. This large increase is presumably affected by the selective removal on the P-11 column of ATP-binding proteins which would otherwise compete for binding sites on the ATP-agarose affinity resin.
Routing the P-11 gradient eluant through two hollow fiber dialysis units facilitated the rapid equilibration of the phosphocellulose column eluant to conditions necessary for uridine kinase binding to the ATP-agarose resin. This procedure reduced the phosphate concentration from 80 mM to about 5 mM and simultaneously added to the enzyme preparation 5 mM M e necessary for adsorption to ATP-agarose. Phosphate concentrations greater than 20 mM inhibited uridine kinase adsorption to the ATP resin.
The final chromatography on ATP-agarose had several advantages: 1) purification to homogeneity, 2) 10-fold concentration of the enzyme, and 3) the activity eluted in 1.2 mM ATP which was ideal for stabilizing enzyme activity during storage.
Multiple peaks of uridine kinase were observed on both the DE52 and the P-11 columns (Fig. 2). Similar profiles have been obtained with DEAE-cellulose chromatography (Fulchignoni-Lataud et al., 1976;Dubinina et al., 1982) or with native isoelectric focusing (Ahmed and Welch, 1979;Ahmed and Baker, 1980;Ahmed, 1982;Ullman et al., 1979;Absil et al., 1980;Fulchignoni-Lataud and Roux, 1984). These authors have usually interpreted the multiple peaks as representing separate isozymes of uridine kinase. We have previously shown (Payne and Traut, 1982a) that uridine kinase exists as multiple aggregation states, containing different numbers of subunits, that are readily interconvertible. Since polymers of different sizes would vary in the number of exposed charged residues, they would readily separate in any method where migration is based on charge (Pharmacia, 1980;Scopes, 1982). Some interconversion is evident in our elution profiles: the four activity peaks on the DE52 column represent 49% of the initial enzyme activity, while the three peaks on the P-11 column and the one peak on the ATP column represent 43 and 40% of the initial enzyme activity, respectively. Since the recovery rate is so high, it follows that no peaks are lost in subsequent chromatography steps; rather, they are all finally converted to the same form. When this final homogeneous enzyme preparation was examined by two-dimensional electrophoresis, only one protein species was observed (Fig. 5). Multiple species were again evident when this enzyme preparation ( Fig. 5) was analyzed by gel filtration or ion exchange chromatography (Fig. 6, A and B ) . Also contrary to the possibility of isoenzymes and separate genes for uridine kinase is the fact that uridine kinase-deficient cell lines are easily obtained (Ullman et Whitehouse et al., 1982). All our results, as well as those of other laboratories, are entirely consistent with our finding that different molecular weight species of uridine kinase are interconvertible (Payne and Traut, 1982a) and that the different peaks seen on ion exchange chromatography can all be converted to a single form (Figs. 2, 5, and 6).
For these reasons, it is likely that previous reports of native uridine kinase isoenzymes are due to different aggregation states of a single uridine kinase gene product.  with a spectrophotometric assay that used the coupling enzymes pyruvate kinase and lactate dehydrogenase; the oxidation of NADH is then dependent on ADP, a product af uridine kinase. Assays were performed on a Beckatan Madel 25 recording spectnphometer at 22°C. For the standard assay. the decrease in absorbance at 340 nm was monitored in a 1 ml total reaction volume wntaining 50 mM HEPES (pH 8.0 B 2ZT). 50 mM KC1, 10 mM ATP (pH of s t q k ATP solution adjusted to 8.0 with 2.0 N KOH). 12 mM MgClp (sroCk Mg2 wncentration determined by EDTA titration as described by Skmg and West, 1976 Kinetic analyses were done according to Cieland (1970).

Buffer Commition
(v/v), and 10 mM magnesium chloride at a pH of 7.5. The composition of TCE buffer was 50 mM Tris-Base and 0.01 mM EDTA which had been titrated to a pH of 6.6 with 1 M citric acid.
KPHE buffer contains 5 mM potasium phosphate, 5 mM magnesium chloride and 0.W5 mM EDTA at a pH of 7.5. K P m buffers were sterilized by filtration through a or 0.22 I " Millipore filter. Buffers were stored at 4' C and used within 1 week of preparation.

Colymn Resins
Buffers designated as KPQl wntained 10 mM potassium phosphate. 1% glycerol The DE-52 cellulose was washed initially 3 times with 5 volumes of 2 M NaCl followed by 3 washes Of 5 volumes of deionized H20 before proceeding with the 0.5 N HCI and 0.5 N KOH treament. Washed DE-52 cellulose was stored a t 4OC in 5 mM EDTA and 0 . m (v/v) chlorohexidine gluconate.
DE-52 was equilibrated by a batch procedure in 10 mM potassium phosphate. pH 7.5, before packing of the column. Equilibration was considered cmplete when the pH and conductivity were identical to 10 mM potassium phosphate buffer. pH 7.5. DE-52 cellulose was recycled using the same procedures.
Before use, P-11 phosphocellulose was washed according to the fallowing schedule. The resin was iniually washed 3 times in 5 volumes of W NaCl and then suspended in 5 volumes of 0.25 N KOH and stirred gently far 10 minutes, followed by extensive washes with deionized H,O until the pH was below 8 . The resin was then twice resuspended m 5 volumes of 0.25 N HCI and stirred gently for 5 minutes followed by extensive rinsing with deionized H 2 0 until the pH was above 5.
The resin was then transferred to a large column and washed with 50 volumes of 1 mM EDTA. This is an essential Step that remves an unidentified, W absorbing compound whose presence results in significant. if not cmplete. loss of uridine kinase activity. washmg the resin on a filtration funnel is totally unsatisfactory: washing the P-11 resin on a column requires more than one week. The washed resin was stored in 5 mM EDTA at 4OC or used immediately. P-11 phosphoadiusrments to pH 6.6 with either 0.5 M Tris-Base or 1 M Citric acid. The cellulose was initially equilibrated in 5 volumes of 10 X TCE buffer. making material was then resuspended in TCE buffer several times until bath pH and conductivity were identical to TCE buffer. After use. the P-11 DhosDhocellulose was recycled by repeaung the washes with KOH and HCI Adenosine 5"tripho~phate-aga~o~e was recycled accarding to the following procedure. ATP-agarose was washed 5 times with 5 volumes of 2 N NaCl to remove any bound protein. The NaCl wash was followed by 5 x 10 volumes of demnized H,O. The reSm was then suspended in 5 volumes of 7G% ethanol and stirred gently for 10 nunUtes before a flnal H 2 0 wash and storage at -2O' C in 5% glycerol immediatelv mior to use the ATP-Aoarose was extensivelv eouilibrated m KPGM buffer. A h ; chromatography, the iTP-AgarOSe was imm'edidtely recycled and stored rn 50% glycerol a t -2O'C.
DTT. The Sample overlay solution had 1% ampholytes pH 3/10. The tube gels contamed: 3.74% acrylanude and 0.21% bis-acrylamlde, 2% ampholytes 3/10, and 0.034% ammonlum persulfate. The ends of the tubes were not covered with dlalysks tubmg. The were run 16 hrs at 450 V . The tube gels were placed on I C~ and lmsened wlth a tube gels were prefocused (at mnstdnt voltage) at 400 V before loading. and the gels gel equihberatmn contained 4 mM DTT, and the gels were equllibrated for 1 hr ~n screw 1 p1 Hamilton syrmge pnor to ejection from the tubes.
The SDS sample buffer used for cap test tubes. and prepared acwrding to L a e d i . (1970). The separating gel was 2.5 mm x 13 an The discontinuous slab gel was run usmg the Protein Slab Cell (810 Rad) x 14 cm and consisted of 12% acrylamide, 0.32% bis-acrylamide and 0.1% TEMED.
Silver Stainmg Procedure far Visualinnq Proteins in Polyacrylmde Gels following modification. Gels were fixed in sob methanol -109 acetic acid over- The silver staining method of Oakky et. al. (1980) was employed with the night and washed 1 hour with several changes of deionized HtO prior to proceeding with the glutaraldehyde step.

Gel scanning
Gels were scanned at 400 nm on a Gilford Model 2600 Micropmessor spectrophotometer eqvipped with a gel scanning apparatus and a Hewlett-Packard Model 72258 XY plotter. Peak integration was performed by an internal subroutine procedure.

Protein Assays
A dye-binding assay usmg Coomassic Blue G as described by Read and Northcote (1981)   for 45 min and centrifuged at 2 0 . W x g far 20 min. The pellet was discarded and the resulting supernatant was brought to 53% saturation, stirred for 45 mib at O T The solution was cenwifuged as before and the protein pellet dissolved 30-53% ammonium sulfate fracuon may be stomp at -20% for iater processing but in a M n m i volume of 10 mM potassium phosphate buffer. (pH 7.5). The dissolved uridine kinase activity decreases according to a first order decay with a t , of 7 weeks. For optimal recovery it is therefore advantageous to proceed dkectiy to the P-11 phosphoceliuiose column.
The 3 -5 3 ammonium sulfate fracuan Was diluted with 40 volumes of TCF buffer and applied to the P-11 col-  fraction was applied to the ATP-Agamse column. Homogeneous uridine kmase was The pmled P-11 fracuons repreamong 95% of the recovered activity were applied at a flaw rate of 10 ml/hr to the ATP-Agarose affinity calumn (1.6 on x 30 cm with a bed volume of 60 ml) equilibrated in KPGM buffer. After sample addition was campiete. the column was washed with 10 w1-volumee of KPGM buffer. This was followed bv 5 column volmes of 1M ISH ATP in YPOU hnffer Hamageneous uridine kinase-was step-eluted with-zi&&"&&"Zi .I-ZATP in KPGM ( a r m addition of ATP to the buffer, the pH was readjusted to pH 7.5 with 2.0 N KOH). Enzyme  a Protein assayed by the SDS-polyacrylamide gel-silver staining procedure ( Figure  1). b
Lanes 9 and 10 contalned 10 ng each of uridlnc kmase protein from the two peak ATP-agarose column fractions (see Figure 3 ) . eluted frm the ATP-Agarose column are shown In Flgure 4 (Lanes 9 and 10). As The SDS-polyacrylamide gel elecwophoresis of the two active fractions shown In Lane 9 the protean is apparenly homogeneous with a subunal molecular Wmqht of 31.000. Thls value IS ~n agreemen1 w t h a monomer m l e~u l a r welghl of resolve native uridine W n w into several pcnka (Fig. 6). The multiple mlccular Both the gel filtration column and the ion-exchange c o l u m n yere able to weight spccles seen with the ultmgcl col-r d l e d the mulu abuined with crude urldlns kinase (Payne and Traut I-).
The h exchange column ruolved at lea81 6 p e a k (Fig. 68): p e a t 4 kprcscnu ATP (present in the enzyme sample) as dewrmincd by thin layer chromrugraphy, the Sbaorption spectrum of the peak. and by c01p.rIson to the elution of ATP.
The other puka ell had active vrldine kinase d v i t y (data M I S h ) .

Kinetic Studiur
The purified enzyme had a specific actlvltyff 2R3 wl/min/mg enzyme at 22-C. This lea& ! n a turnover numbtr of about 150 s . Initial velcclty patrrrns with hcfnwcnmus uridine kinase wlng uridine and ATP as substrates were consistent with a sequential acehanism as oppscd to a ping-pong mechanism (data not sham). This is conaimen1 with the findings of L L m u r u and A n d e m (1975) who u.ed a 2%-fold purified uridine kinase prepmation fmurlne -1 tumor pB15. Bath CTI' and UTP showed co.lpeUtlve inhlbith wiIh m p c t lo ATP. Ki's for CTP and UTP inhibition were 10 and 61 e. respectively. The h ' s for LVldine and cyudinc were 40 and 57 ,AM, respectively (Tabla 11).
Scverai nucleoside a n a l e g~ were also good substrates for uridine kinase Table 11 (Table 1).