the enzymatic mechanisms for · t+pydrwtdr+py. reaction (1) was initially characterized by friedkin...

The Enzymatic Mechanisms for Deoxythymidine

Synthesis in Human Leukocytes

IV. COMPARISONSBETWEEN

NORMALANDLEUKEMICLEUKOCYTES

ROBERTC. GALLOand SEYMOURPERRY

From the National Cancer Institute, National Institutes of Health,Bethesda, Maryland 20014

A B S T R A C T (1) Synthesis of deoxythymidine byeither direct transfer of deoxyribosyl to thymine (py-rimidine deoxyribosyltransferase) or by a coupled deoxy-nucleoside phosphorylase mechanism is approximatelytwofold greater with normal leukocyte extracts (55 to88%o granulocytes) than with extracts prepared fromleukocytes obtained from patients with chronic mye-logenous leukemia. Activities in lymphocytes (normal orleukemic) are one-fifth the activity of normal granulo-cytes.

(2) The lower activity in chronic myelogenous leuke-mia remains at 50% of normal even when patients are inhematologic remission with a normal per cent maturegranulocytes in the peripheral blood.

(3) The leukemic enzyme could not be distinguishedfrom the normal by pH optima, thermal stability, orkinetic properties. The Km's for the deoxyribosyl ac-ceptor and deoxyribosyl donors were identical for bothenzymes. Both are subject to substrate inhibition by thy-mine and to inhibition by purine bases with similar Ki's.In addition, the transferase component of both the leu-kemic and the normal cell enzyme is activated by phos-phate and arsenate. It appears, therefore, that there isno qualitative difference between the enzyme obtainedfrom leukocytes of patients with chronic myelogenousleukemia and the enzyme obtained from normal leuko-cytes, suggesting that the difference in total cell activityis due to an actual decrease in amount of enzyme inchronic myelogenous leukemia or to a mixed cell popu-lation, one with a normal quantity of enzyme and theother with little or no active enzyme.

(4) In both the normal cell and the leukemic cell ex-tracts, transferase and phosphorylase activities could

Received for publication 14 June 1968 and in revised form15 August 1968.

not be separated. The ratio of the two activities re-mained constant over a 140- and a 230-fold purificationin normal and leukemic cell extracts, respectively. Theseand other observations indicate that transferase andphosphorylase activities are associated with the sameprotein.

(5) The metabolism of pyrimidine and purine deoxy-nucleosides is similar for normal and leukemic cells.Catabolism of all deoxynucleosides tested was by directphosphorolysis, except for deoxyadenosine which re-quired initial deamination to deoxyinosine before phos-phorolysis. In contrast to the greater rates of pyrimidinedeoxynucleoside synthesis and cleavage with normalleukocyte extracts, the rates of purine deoxynucleosidesynthesis and cleavage were approximately twofoldgreater with extracts prepared from cells of -patientswith chronic myelogenous leukemia. There was no sig-nificant difference in the rate of phosphorolytic cleavageof pyrimidine nucleosides (uridine) between the CMLand normal leukocyte extracts.

INTRODUCTION

It has been known for some time that pyrimidine prod-ucts of DNAdegradation can be reutilized for the syn-thesis of new nucleic acid molecules. These studies havegenerally been concerned with reutilization of deoxy-nucleosides and deoxynucleotides as, for example, theseveral observations on the reutilization of deoxythy-midine (1-3). Much less information is available on thereutilization of free pyrimidine bases (thymine, uracil,and cytosine), particularly in mammalian cells. Onemechanism by which this could occur would be throughthe synthesis of pyrimidine deoxynucleosides from a py-rimidine base and a deoxyribosyl donor.

The Journal of Clinical Investigation Volume 48 1969 105

The enzyme trans-N-deoxyribosylase, which catalyzesdeoxynucleoside synthesis, has been well characterized inbacterial systems (4-7). The absence of this enzyme inhuman tissues led Beck to hypothesize that human tis-sues may lack a mechanism by which the deoxyribosylmoiety can be transferred from one base (purine orpyrimidine) to another (8). However, we have recentlydemonstrated that crude extracts of human leukocytesactively catalyze transfer of deoxyribosyl to thymine andto uracil, a reaction catalyzed by deoxythymidine phos-phorylase (9). There is evidence in human spleen (10)and in normal human leukocytes (11, 12) indicatingthat the one enzyme protein, deoxythymidine phos-phorylase, catalyzes synthesis of deoxythymidine ordeoxyuridine by two distinct enzymatic mechanisms:

(1) coupled deoxynucleoside phosphorylase

(a) XdR+Pi-dR-1-P+XX(b) T+dR-1-P TdR+Pi

(2) pyrimidine deoxyribosyltransferase

T + PydR w TdR + Py.

Reaction (1) was initially characterized by Friedkin andRoberts (13).

The capacity of leukocyte extracts from patients withCML to support deoxythymidine synthesis by either thecoupled or transferase mechanism is reduced to approxi-mately one-half of normal. The decreased activity inCML leukocytes was found even though leukocyte dif-ferential counts were comparable to those in the nor-mal controls and in donors who were in clinical and he-matologic remission (14). These observations sug-gested the possibility of a defect in the enzyme in CMLand prompted an investigation of the comparative kineticproperties of the CML and normal leukocyte enzymes.The purpose of the present communication is to reportthe data obtained from these comparisons and the re-sults of the relative activities of enzymes involved in themetabolism of pyrimidine and purine deoxynucleosidesin normal and leukemic leukocytes.

METHODS

Source of leukocytesLeukocytes were obtained from the peripheral blood of

normal volunteers and from patients with CML. The cellsfrom the normal donors consisted primarily of maturegranulocytes (55 to 88% granulocytes). The CML patientshad prominent leukocytosis (peripheral leukocyte counts over

1 Abbreviations used either in the text or in the figures:XdR, any purine or pyrimidine deoxynucleoside; PydR,pyrimidine deoxynucleoside; X, any purine or pyrimidinebase; Py a pyrimidine base; TdR, deoxythymidine; UdR,deoxyuridine; AdR, deoxyadenosine; IdR; deoxyinosine;T, thymine; U, uracil; Hx, hypoxanthine; dR-1-P ordR-1-PO4, deoxyribose-1-P; CML, chronic myelogenousleukemia; CLL, chronic lymphocytic leukemia.

50,000/mm3) with a population of cells of variable maturity.All had moderate splenomegaly. None of these patients wasin a blast phase and none was receiving antileukemictherapy. Bone marrow preparations of the CML patientswere all positive for the Philadelphia chromosome. Studiesof various enzymatic activities of CMLpatients in completehematologic remission were also performed and those resultshave been reported in a preliminary communication (14).

Isolation of leukocytesLeukocytes were obtained and extracts prepared by meth-

ods previously described (9). All experiments were per-formed with cellular extracts. Lymphocytes were isolated bythe nylon wool column technique (15).

Enzyme assays(1) Deoxythymidine phosphorylase. This enzyme cata-

lyzes the reversible synthesis and phosphorolytic cleavage ofdeoxythymidine and measurements were made in both direc-tions. Deoxythymidine synthesis was measured by followingformation of deoxythymidine-14C from the reaction of thy-mine-"C with deoxyribose-1-P. Details of the procedurehave been previously described (9). Cleavage of deoxy-thymidine was measured by following thymine productionfrom deoxythymidine spectrophotometrically (16) or byfollowing conversion of deoxythymidine-'4C to thymine-14C.Assay conditions were described previously (9).

(2) Pyrimidine deoxyribosyltransferase activity was mea-sured by following deoxythymidine-"C formation from thereaction of thymine-Y4C and deoxyuridine in the presence ofhigh (0.1 M) concentrations of arsenate. Arsenate is used toprevent formation of deoxyribose-1-P so that determinationof newly synthesized deoxythymidine is a measure of deoxy-thymidine formed only by direct transfer of deoxyribosyl tothymine and not through a deoxyribose-l-P intermediatewhich may result from phosphorolysis of deoxyuridine. De-tails of the assay have been reported (9).

(3) Uridine phosphorylase was assayed by followinguracil production from uridine spectrophotometrically at 290m,u. The reaction mixture in a final volume of 0.5 mlincluded: 10 mmuridine; 0.1 M phosphate buffer, pH 7.2;0.05 M Tris-HCl buffer, pH 7.2; and f rom 0.5 to 0.7 mgprotein obtained from crude leukocyte extracts. Incubationswere at 37°C for 20 min. Labeled substrates and productswere separated by descending paper chromatography (9).

(4) Deoxyadenosine deaminase was assayed by deter-mining the rate of deoxyinosine-8-4C formation from deoxy-adenosine-8-14C. Reaction mixtures contained 10 mMdeoxy-adenosine-2-'4C (46.5 mc/mmole). The buffers and condi-tions were otherwise as described for uridine phosphorylase.

(5) Purine nucleoside phosphorylase was assayed by mea-suring hypoxanthine-8-4C production from deoxyinosine-8-'4C (4.2 mc/mmole), but otherwise experimental conditionswere identical to the assay for uridine phosphorylase.Hypoxanthine was not catabolized significantly during theperiod of assay, indicating that leukocytes contain minimalxanthine oxidase activity. Its determination could, therefore,be used for measurement of deoxyinosine cleavage. Sepa-ration of purine bases from purine deoxynucleosides was bydescending paper chromatography (17).

Purification of deoxythymidine phosphorylaseThe details of the purification of the enzyme from normal

leukocytes has been previously described (17). The pro-

106 R. C. Gallo and S. Perry

cedure used for purification of the CML leukocyte enzymeis described below.

All procedures were carried out at 40C. Approximately5 X 1010 CML leukocytes were homogenized 10 min in aVirtis homogenizer and centrifuged 30 min at 40,000 g. Thesupernatant (fraction I) was treated with 1 M acetic acidto bring the pH to 5.2 and centrifuged 30 min at 40,000 g.The supernatant (fraction II) was treated with solidammonium sulfate to 55% saturation and allowed to stand45 min, and then centrifuged at 40,000 g for 30 min. Theprecipitate (fraction III) was taken up in 5 ml of 0.01 Msodium phosphate buffer, pH 7.3, containing 4 mm di-thiothreitol (buffer A) and applied to a 2.5 X 15-cm G-25Sephadex column and eluted with buffer A. The samplesfrom the protein peak were combined and applied to a1 X 10-cm DEAEcellulose column and eluted with buffer A,containing 0.75 or 1.0 M NaCl. Elution of the enzyme wasaccomplished in a small volume (1-2 ml) that was thendialyzed overnight against buffer A. The dialyzed materialwas applied to a 1 X 60-cm DEAE cellulose column and aconvex salt gradient was initiated by permitting buffer Acontaining 1 M NaCl to pass into a 250 ml mixing chambercontaining only buffer A. The samples containing the enzymewere combined and dialyzed for 10-15 hr against 0.1 MTris-HCl buffer, pH 7.2. This fraction (fraction IV) wasa 230-fold purification over fraction I. 1 ml containedapproximately 0.00025 mg of protein and from 1 to 2 unitsof enzyme. (A unit of enzyme was defined as that amount

200

80

160

-mE 140e

o 120

F- 100

80.uM 60u .--40

20

0

DEOXYNUCLEOSIDECLEAVAGE(Normal vs. Leukemic)

TdR - T+dR IdR= Hx+dR

PER CELL PERMG

PROTEIN

PER CELL PERMG

PROTEIN

of protein that cleaved 1 /Amole of deoxythymidine in 1 hr.)Except when specifically indicated, fraction I was the sourceof enzyme. The principal objective in purification was todetermine if the CML leukocyte enzyme like the normalleukocyte enzyme (11) contained both phosphorylase andtransferase activities.

Protein was determined by the method of Lowry, Rose-brough, Farr, and Randall (18) with bovine serum albuminas the standard. Phosphate was determined by the methodof Chen, Toribara, and Warner (19).

RESULTS

A comparison of purine nucleoside phosphorylase, deoxy-thymidine phosphorylase (measured in the direction ofdeoxythymidine cleavage), and pyrimidine deoxyribosyl-transferase activities of various leukocyte extracts is il.lustrated in Fig. 1. Normal leukocytes contain approxi-mately two-fold greater activity of both deoxythymidinephosphorylase and pyrimidine deoxyribosyltransferasethan CMLleukocytes, and approximately fivefold greateractivity than normal lymphocytes or lymphocytes frompatients with CLL. In contrast, CMLleukocyte extractscontain approximately 1.5- to 2-fold more purine nucleo-side phosphorylase than normal leukocytes (see alsoFig. 9). As with the pyrimidine enzyme activities, pu-

TdR SYNTHESIS FROM T + UdR(Normol vs.Leukemic)

T +UdR==T*dR+U

PER CELL PERMG

PROTEIN

200

160--4

-4120I<np.4m

80 -"'

0a

40

0

FIGURE 1 Comparison of levels of deoxythymidine phosphorylase, purine nucleo-side phosphorylase, and pyrimidine deoxyribosyltransferase activities in normaland leukemic leukocyte extracts. Results are expressed as per cent of the normalgranulocyte activity both on a per cell (108 leukocytes) or per milligram ofprotein bases. Deoxythymidine phosphorylase activity is determined by followingcleavage of deoxythymidine (left panel); purine nucleoside phosphorylase by mea-

surement of deoxyinosine cleavage (center panel); and the transferase levels bymeasuring deoxythymidine synthesis from thymine and deoxyuridine. Details ofassay are described in Methods. Points and accompanying lines are the meanvalues and standard deviations.

Deoxythymidine Synthesis 107

| Normal granulocytesa CML

I8 |°0 Normal lymphocytesAo11+° I

T*,dR-I-P;t= T*dR+PiU*+dR- I-PO4 = tdR+Pi T*+WR = T*dR+U

FIGURE 2 Left panel: Time course of deoxy-uridine and deoxythymidine synthesis withdeoxyribose-l-P as the deoxyribosyl donorcatalyzed by deoxythymidine phosphorylase.The rates with both the normal and CMLleukocyte enzyme are illustrated. The reac-tion mixture in a final volume of 0.5 mlcontained 5 /Amoles of thymine-A4C or uracil14C; 5 Amoles of deoxyribose-l-P; 25

umoles of Tris-HCl buffer, pH 7.2; and25 jumoles of succinate buffer, pH 7.2.Right panel: Time course of deoxythymidinesynthesis with deoxyuridine as the deoxy-ribosyl donor. The normal and CMLleuko-cyte enzymes are compared. The reactionmixture contained 5 /imoles of thymine-YC;5 Iumoles of deoxyuridine; 25 Amoles of Tris-HCl buffer, pH 7.2; 25 /moles of suc-cinate buffer, pH 7.2; and 50 /moles ofsodium arsenate. Fraction I was the enzymesource in both experiments.

TIME (Minutes)

rine enzyme activity was least in lymphocytes, andthere was no significant difference in activities betweennormal and leukemic lymphocytes. The results pre-sented in Fig. 1 are the mean values and standard devia-tions of the activities in leukocytes from 40 normal vol-unteers (predominantly granulocytes), 25 patients withCML, 7 patients with CLL, and 5 normal volunteers(lymphocytes only). The lower activity of deoxythymi-dine phosphorylase in CML leukocyte extracts does notappear to be the consequence of an inhibitor present inthese cells. Experiments in which equal amounts of leu-kocyte extracts from CML patients and leukocyte ex-tracts from normal volunteers were mixed gave valuesthat approximated the calculated mean activity of eachassayed alone.

Pyrimidine deoxyribosyltransferase in CML pa-tients in remission

Since persistently low levels of deoxythymidine phos-phorylase and pyrimidine deoxyribosyltransferase ac-tivities were found in leukocytes obtained from patientswith CML (approximately 50% of the activity of nor-mal granulocytes), an investigation was carried out todetermine if the low values were simply the result ofcellular immaturity or were, in fact, associated with theleukemic process. Pyrimidine deoxyribosyltransferasewas assayed in nine patients with CML in remissionwho were not receiving chemotherapy. These patientshad approximately the same number of mature granu-locytes in their peripheral blood as the normal controls.Paired assays revealed that the enzymatic activity re-mained at approximately 50% of normal despite apparentclinical and hematologic remission. A report of thesefindings has been published (14).

Properties of the enzyme in CMLleukocytespH optima. With both the normal and CML-enzyme

the pH optimum for deoxythymidine synthesis fromthymine and dR-1-P (phosphorylase) is from 5.9 to 6.0.The optimum is 7.2 for deoxythymidine synthesis fromthymine and deoxyuridine (transferase).

Heat denaturation. The rate of enzyme denaturationwas determined by heating the enzyme preparation at60'C for various periods, cooling, and then assayingfor deoxythymidine synthesis. The ti for inactivation ofboth transferase and phosphorylase with either CMLornormal enzyme is approximately 10 min.

Time curves. Equilibration of deoxythymidine ordeoxyuridine synthesis with deoxyribose-1-P as thedeoxyribosyl donor for normal and CML cell extractswas not observed over a 90 min period (Fig. 2, leftpanel). Similarly, over a 120 min assay period, equili-bration was not obtained with deoxyuridine as the deoxy-ribosyl donor (Fig. 2, right panel). Maximum rates ofdeoxynucleoside synthesis are not obtained in these ex-periments (Fig. 2), both because of the inhibitory con-centrations (10 mM) of pyrimidine bases used in theseexperiments and the suboptimal concentrations of deoxy-uridine (right panel). Substrate inhibition by the pyrimi-dine bases is discussed below.

Substrate specificity. With deoxyribose-1-P as thedeoxyribosyl donor, the rate of deoxynucleoside synthe-sis is approximately two- to three-fold greater withuracil as the deoxyribosyl acceptor than with thymine(Fig. 2). Uracil is also the favored deoxyribosyl ac-ceptor with deoxyuridine as the deoxyribosyl donor(9). In neither case did cytosine serve as a substrate.These relationships were found with both the normaland the CMLcell extracts. The Km for deoxyribose-1-P


FIGURE 3 Lineweaver-Burk plot of the velocityof deoxythymidine synthesis catalyzed by CMLleukocyte deoxythymidine phosphorylase as a func-tion of the concentration of deoxyribose-1-P (S).The thymine concentrations are indicated in thefigure. The enzyme source in these experiments wasfraction IV.

Is

is from 2 to 4 mmfor the enzyme obtained from normalleukocytes (9, 17), and similar values were obtainedwith the CML leukocyte enzyme (see curve D of Fig.3). The Km for deoxyuridine is approximately 15 mmfor the normal leukocyte enzyme (9). A similar Km fordeoxyuridine was obtained with the CMLenzyme. Thiscan be calculated from the data shown in Fig. 4 at non-inhibitory concentrations of thymine (2 mM).

Substrate inhibition of deoxythymidine synthesisby pyrimidine basesThe determination of a Km for the pyrimidine base is

complicated by substrate inhibition. Using normal leu-

-0.05

T*+ UdR '- T*dR +U

0 0.05 0.10 0.15 0.20

kocyte extracts, Gallo, Perry, and Breitman demon-strated that both uracil and thymine inhibit deoxythy-midine synthesis markedly with deoxyribose-1-P as thedeoxyribosyl donor (9). The inhibition by thymine isgreater than the inhibition by uracil. The same degreeof inhibition was observed with the CML enzyme(Figs. 3 and 5) and, in addition, as with the normalcell enzyme, Lineweaver-Burk plots reveal that the in-hibition involves a change in both the apparent Km fordeoxyribose-1-P as well as Vm. (Fig. 3).

Deoxythymidine synthesis from thymine and deoxy-uridine is also inhibited by thymine but not by uracil.The kinetics of inhibition with the normal leukocyte

/5mM T

FIGURE 4 Lineweaver-Burk plot of the velocity2.5mM T of deoxythymidine synthesis catalyzed by CML

leukocyte pyrimidine deoxyribosyltransferase as2mMT a function of the concentration of deoxyuridine.

The thymine concentrations are indicated in thefigure. Fraction I was the source of enzyme.


U +dR I-P04 .U d R + Pi

2.01

I'lb

Zk

sj,

1.51

1.01

0.5

0 2 4 6 aURACIL (mM)

10 12 14

FIGURE 5 Substrate inhibition of CML leukocyte deoxythymidine (deoxyuridine) phosphorylase by uracil. Thereaction mixture contained 5 Amoles of deoxyribose-1-P and the indicated concentrations of uracil. Fraction I wasthe source of enzyme.

enzyme revealed only a change in Km for deoxyuridinewith no change in V..., indicating that excess thymineinterferes with the binding of deoxyuridine (9). Similarkinetics were found when the CML enzyme was used(Fig. 4).

Fig. 6 illustrates that the magnitude of substrate in-hibition by thymine of deoxythymidine synthesis fromthymine and deoxyuridine (pyrimidine deoxyribosyl-transferase) is identical with both the normal enzyme

and the CML leukocyte enzyme. As anticipated fromthe competitive kinetics shown in Fig. 4, the curves il-lustrated in Fig. 6 intercept at a common point thatrepresents the K4 for thymine (approximately 1.5 mmfor both enzymes).

Thymine also inhibits the phosphorolytic cleavage ofdeoxythymidine, in which case it is an example of prod-uct inhibition. This has been demonstrated in bacterialsystems (20) as well as in mammalian tissues (16, 17,21). The possible regulatory significance of thymine incontrolling the intracellular pool of deoxythymidine bytwo mechanisms, substrate inhibition and product in-hibition, has been the subject of previous speculation byGallo and associates (9, 17). It is of interest in this re-

gard that unlike the human leukocyte enzyme, E. colideoxythymidine phosphorylase is subject only to product

inhibition, since there is no substrate inhibition of de-oxythymidine synthesis by thymine with the E. coli en-zyme (R. Gallo and T. R. Breitman, unpublished ob-servations.)

Activation of pyrimidine deoxyribosyltransferaseby phosphate and arsenate

Previous studies with the normal leukocyte enzymerevealed anion activation of the transferase mechanismfor deoxythymidine synthesis by phosphate and arsenate(9). The activation by phosphate was not due to forma-tion of deoxyribose-1-P from deoxyuridine, since con-centrations of phosphate that were optimal for activationof deoxythymidine synthesis from thymine and deoxy-uridine markedly inhibited deoxythymidine synthesisfrom thymine and deoxyribose-1-P. Similar character-istics of the enzyme were also found with the CMLen-zyme (Fig. 7).

Purine base and purine deoxynucleoside inhibitionof deoxythymidine synthesisThe normal leukocyte enzyme has been shown to be

subject to inhibition by purines (17, 22). A comparisonof the effect of various concentrations of different purineson the CMLand normal enzymes is presented in Fig. 8.


I I I I I I I I I

I I I

5-

4-

II I I. .1~~ I~ I I I

10

8 0 5 idi

25MLdR6

THYMINE(mM) ~ ~ /0m

4

2

-KiN ~ ~ ~ 3mMd -i

-4 0 5 10 -4 0 5 10THYMINE (mM)

FIGURE 6 Deoxythymidine synthesis from thymine and deoxyuridine. Comparison ofpyrimidine base (thymine) substrate inhibition of the normal leukocyte and CMLleukocyte pyrimidine deoxyribosyltransferase. The reciprocal of the velocity of deoxy-thymidine synthesis is plotted as a function of the concentration of thymine with severalconcentrations of deoxyuridine. The Ki's for thymine inhibition of both enzymes areindicated. The enzyme source in these experiments was fraction IV.

Inhibition is observed with the 6-oxypurines and 6-mer-captopurine and the magnitude of inhibition is the samefor both enzymes. Kinetic studies of inhibition revealeda slight change in Km for deoxyribose-1-P and a changein Vrna. (17).2 This "mixed type" inhibition is similarto the inhibition by thymine, but unlike thymine, pu-rines did not inhibit the enzyme in the direction of de-oxythymidine cleavage. The kinetics were identical forthe normal cell and CMLcell enzyme.

Evidence that human leukocyte pyrimidine de-oxyribosyltransferase and deoxythymidine phos-phorylase activities are associated with the sameprotein

Zimmerman has reported that the transferase andphosphorylase activities were not separated after a3000- and a 30-fold purification of the human spleen andE. coli enzymes, respectively (10). In addition, thesepurifications did not alter the ratios of the activities,which suggested to him that both activities were a func-tion of the same protein. Wehave confirmed these find-

2 Inhibition by purines is not due to depletion of substrate(deoxyribose-1-P) by reaction with purines to form apurine deoxynucleoside, since the concentration of deoxy-ribose-1-P in these experiments is in excess (10 mM). Fur-thermore, the purified enzyme is completely free of purinenucleoside phosphorylase, the enzyme which catalyzes areaction of purines with deoxyribose-1-P, yet inhibition bypurines is still observed (17).

ings with the enzyme activities from normal leukocytes.Over a 140-fold purification transferase and phosphory-lase activities were constant. Furthermore, both enzymeactivities had identical rates of heat denaturation (11).As part of the investigation of the properties of theCML enzyme, purification was carried out along withcomparisons of transferase and phosphorylase activities.The yield of purified enzyme (fraction IV) by thismethod varied from 5 to 10% of fraction I and in thepurification presented in Table I it was 6%. As shownin Table I, over a 250-fold purification the two activitiescould not be separated and the ratio of transferase tophosphorylase remained constant. It is concluded there-fore that like the normal leukocyte and spleen protein,one CML protein probably has both transferase andphosphorylase activity.

Comparison of pyrimidine nucleoside phosphoro-lytic cleavage (uridine phosphorylase) in normaland CMLleukocyte extracts

Uridine is the most commonly used precursor to mea-sure RNAsynthesis and pool sizes of uridylate. Varia-tion in the rate of catabolism by phosphorolysis betweennormal cells and leukemic cells could lead to erroneousconclusions concerning differences observed in rates ofRNA synthesis or pool sizes when this compound isused as a precursor. Furthermore, as shown in thiscommunication, we have observed significantly lower

Deoxythymidine Synthe.sis 111

I-C

x

U-CD

0.005 0.05 0.5 5.0 50.0PHOSPHATE(mM)

levels of enzymes involved in the phosphorolysis of py-rimidine deoxynucleosides (deoxythymidine phosphory-lase) in CMLleukocyte extracts. The possibility was con-sidered that CMLleukocytes may have in general lowerlevels of enzymes involved in the metabolism of pyrimi-dine pentosides. The data in Table II demonstrate thatthis is not the case. The rate of phosphorolysis of uri-dine is not significantly different in CML leukocyte ex-tracts from patients in relapse or in remission from therate observed with normal leukocyte extracts.

Comparison of pyrimidine deoxynucleoside andpurine deoxynucleoside metabolism in normaland leukemic leukocytesAs previously discussed, pyrimidine deoxynucleoside

(deoxythymidine or deoxyuridine) synthesis and cleav-age are greater in normal cell extracts than in extractsfrom CML cells. The mechanism for cleavage appearsto be entirely phosphorolytic in both cell types and themechanisms for synthesis also appear identical in nor-mal and leukemic leukocytes. As shown in the centerpanel of Fig. 1 and the right panel of Fig. 9, cleavageof purine deoxynucleosides (deoxyinosine) is 1.5- to2-fold greater in CMLleukocyte extracts than in normal

FIGURE 7 Phosphate activation of deoxythymidinesynthesis from thymine plus deoxyuridine and inhi-bition of deoxythymidine synthesis from thymineplus deoxyribse-1-P, comparison of the normal andCML leukocyte enzymes. The leukocyte extractswere dialyzed until phosphate free, and then phos-phate was added to the concentrations indicated inthe figure. Note the absence of deoxythymidinesynthesis from thymine and deoxyuridine (deoxy-ribosyltransferase) in the absence of phosphate.That the marked increase in deoxythymidine syn-thesis with addition of phosphate is due to anactivation of the enzyme rather than a consequenceof deoxyribose-1-P production from deoxyuridineis indicated by the marked phosphate inhibition ofdeoxythymidine synthesis from thymine and deoxy-ribose-1-P at concentrations that activate the trans-ferase. Fraction I was the source of enzyme.

leukocyte extracts. Fig. 9 (right panel) also illustratesthat cleavage of deoxyinosine, like that of pyrimidinedeoxynucleosides, is by a phosphorolytic (or arsenolytic)mechanism since the rate in the presence of 0.1 M ar-senate is approximately three-fold greater than in theabsence of arsenate. The low rate of cleavage in thearsenate might be attributed to a hydrolytic mechanism.However, this is not the case. Rather, it is due to theendogenous phosphate (0.25-0.32 mM) of the leukocytehomogenate, since after prolonged dialysis of the ho-mogenate all enzymatic activity was lost but could berestored by the addition of phosphate.

All deoxynucleosides studied (deoxyguanosine, deoxy-inosine, deoxythymidine, and deoxyuridine), except fordeoxyadenosine, were cleaved by direct phosphorolysisto produce their respective pyrimidine bases and deoxy-ribose-1-P. Adenine was not isolated from the reactionmixture with deoxyadenosine as substrate, but insteaddeoxyinosine and hypoxanthine quickly appeared. There-uIts indicate that deoxyadenosine is initially deami-nated to deoxyinosine before phosphorolysis. The datapresented in the left panel of Fig. 9 illustrate that, aswith purine nucleoside phosphorylase activity, deoxy-adenosine deaminase activity is approximately threefold


T*+dR-I-P =T*dR+Pi

6 uanine

z_AL/ Deotryguonosine X

/; lhie&-o 40 p //41/~____ eDeoxoenosne Xonthine

Adenine yaDeoydenosine20

66-methyl- Adeninemercoptopurine 6-metyl-W I I~c~ merclOomine

0 1 2 3 0 2 3PURINE BASE OR DEOXYNUCLEOSIDE(mM)

FIGURE 8 Purine inhibition of deoxythymidine synthesis from thymine and deoxy-ribose-1-P; comparison of the normal and CML leukocyte enzyme. The reactionmixture in a final volume of 0.5 ml contained 2.5 &moles of thymine-j4C, 5 Amoles ofdeoxyribose-l-P, buffers as described in the legend to Fig. 2, and the indicatedconcentrations of purines. Results were expressed as per cent inhibition comparedwith the control with no purine added. Fraction I was the source of enzyme.

0 0 CMLz DEOXYADENOSINE A A NORMAL DEOXYINOSINE0 DEAMINASE 0 A 0.1 M AsOW PHOSPHORYLASE6 a ~~~oAsO4

E Z

0 303E 0TE~~~~~~~~~~

:104 0

z 6

0

0~~~~~~~

0 30 60 0 30 60TIME (min)

FIGURE 9 Comparison of the levels of deoxyadenosine deaminase and purine nucleo-side phosphorylase activities in normal and CML leukocyte extracts. Details ofassay are described in Methods.


TABLE IPurification of CMLLeukocyte Deoxythymidine Phosphorylase and Pyrimidine Deoxyribosyltransferase

Specific activity PurificationRatio of

Deoxythymidine Pyrimidine Deoxythymidine Pyrimidine PhosphorylaseFraction phosphorylase deoxyribosyltransferase phosphorylase deoxyribosyltransferase to transferase

I 1.4 1.6 0.87II 1.8 2.1 1.3 1.3 0.84

III 5.2 5.6 3.7 3.5 0.93IV 322 398 230 249 0.81

greater in CMLleukocyte extracts than in nor]kocyte extracts.

As reported by Gallo and Breitman, pyrimidin(nucleosides support deoxythymidine synthesis cby leukocyte enzyme preparations by either of twanisms: direct transfer of deoxyribosyl to thymiiinitial phosphorolytic cleavage of the pyrimidin(nucleoside to deoxyribose-1-P. The latter compacts with thymine to form deoxythymidine. Puoxynucleosides do not act as deoxyribosyl donor

TABLE I IUridine Phosphorylase A ctivity in Leukocyte Extracts

From Patients with CMLin Relapse, CMLin Remand From Normal Controls

Source of leukocyte extracts

Normal12345678

Mean

CMLin relapse1234

Mean

CMLin remission12345

Mean

Enzyme activity

umoles/mgprotein/hr

0.200.240.280.180.240.210.160.290.23

0.300.230.190.250.24

0.320.260.200.280.220.26

Maturcyte

leupre]

leu

mal leu-

e deoxy-atalyzedio mech-ne or bye deoxy-ound re-irine de-

direct transfer mechanism but function only through adeoxyribose-1-P intermediate (11). These findings withnormal leukocytes were confirmed with CMLleukocytes.No direct transfer of deoxyribosyl from deoxyinosine,deoxyadenosine, or deoxyguanosine to either hypoxan-thine, xanthine, adenine, guanine, allopurinol, or 6-mer-captopurine could be demonstrated.

DISCUSSIONs by the Studies of cellular metabolism, enzyme mechanisms, and

even biochemical approaches to the study of neoplasiacan be conveniently studied with human leukocytes. One

Obtained significant disadvantage, however, is the usual presenceission, of a mixed population of cells. This problem can be cir-

cumvented with normal leukocytes by the separation ofre granulo- granulocytes from lymphocytes by passage of the cells!sin final through a nylon column (15). Enzyme activity of theikocyte total cells and of the isolated lymphocytes can be mea-paration

sured. The difference between the two is primarily dueof total to the mature granulocyte. These experiments have beenLkocytes performed with deoxythymidine phosphorylase and with

81 pyrimidine deoxyribosyltransferase, and more than 80%84 of both activities appears to be associated with the gran-80 ulocyte. Evidence has been presented in this communi-71 cation and by Gallo and associates elsewhere (11) that77 transferase and phosphorylase activities are associated65 with the same protein. Studies with leukemic leukocytes82 revealed that CLL cells (more than 95% lympho-76 cytes) also contained low levels of both enzyme ac-

tivities. However, these values were not lower thannormal lymphocytes. Transferase and phosphorylase

150 activities of leukocytes obtained from CML patients34 in relapse with peripheral leukocytosis and containing23 the typical spectrum of cells at various stages of18 maturation were approximately one-half of the ac-

tivities found in the mature normal granulocyte66 (Fig. 1). Patients with CML in relapse usually have60 leukocytosis. Therefore adequate numbers of cells83 are readily obtained and for that reason com-86 parisons are often made between enzyme activities of

60 leukocytes from this type of CML patient with leuko-cytes from normal individuals. However, this results in


a comparison of leukocytes at various stages of dif-ferentiation (CML) with mature cells (normal) sothat the differences may only be a consequence of differ-ences in the level of maturation rather than a result ofthe leukemic process. For this reason, a study was madeof the level of pyrimidine deoxyribosyltransferase ofleukocyte extracts of 14 CML patients in complete he-matologic remission (14). The granulocytes in theperipheral blood of these patients were morphologicallymature, and the -mean per cent granulocytes was quitecomparable to the normal controls. In 16 of 17 pairedassays of the leukocyte enzyme activity of these 14 CMLpatients, the results were below the normal (mean, 50%of normal) and were not significantly different from theenzyme levels of the mixed population of cells from theCMLpatients in relapse. In nine instances the CMLpa-tients in remission were not receiving chemotherapy sothat a drug effect could reasonably be ruled out.

This demonstration of an enzyme deficiency in leuke-mia is unique with the exception of the reports of re-duced leukocyte alkaline phosphatase in CML. This hasbeen reported to remain low in some patients in remis-sion despite the presence of mature granulocytes in theperipheral blood (23-25). Although it is tempting to re-late the reduced enzyme activities to the Philadelphiachromosome, correlation between the two was not ob-served, since in the one CML patient in remission inwhich the enzyme activity did return to normal, the cellsremained positive for the Philadelphia chromosome.

There are several possible explanations for the lowerenzyme activity for the synthesis of deoxythymidine inCML leukocyte extracts as compared to normal leuko-cytes: (1) the presence of an inhibitor in the CMLcells;(2) an alteration at the catalytic site of the enzyme; and(3) an actual decrease in the number of enzyme mole-cules.

The first consideration, the presence of an inhibitor inthe CMLcell extracts, has largely been excluded by themixing experiments with normal leukocytes. As to thesecond possibility, no qualitative functional differencebetween the normal leukocyte enzyme and the CMLen-zyme was identified in the study reported in this com-munication. It appears, therefore, that the lower activityof the CMLenzyme which persists even though patientsare in complete hematologic remission, is the result ofa decrease in the number of enzyme molecules ratherthan a consequence of an altered catalytic function.Alternatively, the reduction to approximately 50% ofnormal could be explained by two populations of cells,one without enzyme activity and the other with a nor-mal enzyme level. At present there appears to be no wayto test this hypothesis.

Although the metabolic importance of the enzymaticsynthesis of pyrimidine deoxynucleosides has not beenclearlv established, these reactions allow for redistribu-

tion of deoxyribosyl from one deoxynucleoside DNAprecursor to another, as for example, it may permit theinterconversion of the salvage and de novo pathways forthymine deoxynucleotide synthesis as illustrated by thefollowing reaction sequence:

(A) U+ TdR UdR+T(B) UdR+ ATP-dUMP+ADP(C) dUMP+ N5,No0-methylenetetrahydrofolate

dTMP+ dihydrofolate.In reaction (A) the deoxyribosyl moiety of a compoundin the salvage pathway, deoxythymidine (TdR), is re-distributed to another pyrimidine deoxynucleoside, deoxy-uridine (UdR), catalyzed by pyrimidine deoxyribosyl-transferase. In the presence of ATP, deoxyuridine canbe directly phosphorylated to deoxyuridylate (dUMP),a key intermediate in the de novo pathway. This reaction(B) is catalyzed by deoxythymidine kinase. Deoxy-uridylate is then directly converted to deoxythymidylate(dTMP) as shown in reaction (C), and is catalyzedby the important enzyme of the de novo pathway, deoxy-thymidylate synthetase. Furthermore, as demonstratedby Gallo and associates (9), the leukocyte enzyme issubject to the influence of modifiers, e.g., activation byanions, substrate inhibition by thymine, product inhibi-tion by thymine (17), and inhibition by purines (17).The inhibition by purines is complex and is critically de-pendent upon the thymine concentration, higher concen-trations of thymine augmenting inhibition by purines(17). From this data and other observations, Gallo haspresented a hypothesis for the role of this enzyme incontributing to the essential DNA precursor, dTMP,under conditions in which there has been a stimulus toDNA synthesis. It was emphasized that the control ofthis enzyme intracellularly may be particularly dependenton the thymine concentration (17).

In the comparative studies on the metabolism of somepurine compounds, the levels of purine nucleoside phos-phorylase and deoxyadenosine deaminase were mea-sured and found to be significantly higher in CMLleu-kocyte extracts than in normal leukocytes. It should beemphasized that the assays of purine nucleoside phos-phorylase and deoxyinosine phosphorylase were per-formed only with leukocyte extracts obtained from CMLpatients in relapse. The metabolic significance of thefinding that CML leukocytes contain lower levels ofenzymes catalyzing the metabolism of pyrimidine deoxy-nucleosides but higher levels of enzymes catalyzing re-actions in the metabolism of purine nucleosides and de-oxynucleosides is not apparent from these studies andawaits further investigation.

ACKNOWLEDGMENT

We wish to thank Mrs. Carla Davis for her excellenttechnical assistance.


flEERENcES1. Robinson, S. H., and G. Brecher. 1963. Delayed incorpo-

ration of tritiated thymidine into DNA. Science. 142: 392.2. Feinendegen, L. E., V. P. Bond, and W. L. Hughes.

1966. Physiological thymidine reutilization in rat bonemarrow. Proc. Soc. Exp. Biol. Med. 122: 448.

3. Bryant, B. J. 1963. In vivo reutilization of the DNAthymidine of necrotized liver cells by cells of testis andintestine. Exp. Cell Res. 32: 209.

4. MacNutt, W. S. 1952. The enzymatically catalyzed trans-fer of the deoxyribosyl group from one purine or pyrimi-dine to another. Biochem. J. 50: 384.

5. Kalckar, H. M., W. S. MacNutt, and E. Hoff-Jorgensen.1962. Trans-N-glycosidase studied with radioactive ade-nine. Biochem. J. 50: 397.

6. Roush, A. H., and R. F. Betz. 1958. Purification andproperties of trans-N-deoxyribosylase. J. Biol. Chem.233: 261.

7. Beck, W. S., and M. Levin. 1963. Purification, kinetics,and repression control of bacterial trans-N-deoxyribo-sylase. J. Biol. Chem. 238: 702.

8. Beck, W. S. 1964. The metabolic basis of megaloblasticerythropoiesis. Medicine. 43: 715.

9. Gallo, R. C., S. Perry, and T. R. Breitman. 1967. Theenzymatic mechanisms for deoxythymidine synthesis inhuman leukocytes. I. Substrate inhibition by thymineand activation by phosphate or arsenate. J. Biol. Chem.242: 5059.

10. Zimmerman, M. 1964. Deoxyribosyl transfer. II. Nucleo-side:pyrimidine deoxyribosyltransferase activity of threepartially purified thymidine phosphorylases. J. Biol.Chem. 239: 2622.

11. Gallo, R. C., and T. R. Breitman. 1968. The enzymaticmechanisms for deoxythymidine synthesis in humanleukocytes. II. Comparison of deoxyribosyl donors. J.Biol. Chem. 243: 4936.

12. Gallo, R. C., and T. R. Breitman. 1968. The allostericproperties of human leukocyte deoxythymidine phos-phorylase. Fed. Proc. 27: 340. (Abstr.)

13. Friedkin, M., and D. Roberts. 1954. The enzymatic syn-thesis of nucleosides. II. Thymidine and related pyrimi-dine nucleosides. J. Biol. Chem. 207: 257.

14. Gallo, R. C., and S. Perry. 1968. An enzyme abnormalityin human leukemia. Nature. 218: 465.

15. Greenwalt, T. J., M. Gajewski, and J. L. McKenna.1962. A new method for preparing buffy coat-poorblood. Transfusion (Philadelphia). 2: 221.

16. Friedkin, M., and D. Roberts. 1954. The enzymaticsynthesis of nucleosides. I. Thymidine phosphorylase inmammalian tissue. J. Biol. Chem. 207: 245.

17. Gallo, R. C., and T. R. Breitman. 1968. The enzymaticmechanisms for deoxythymidine synthesis in human leu-kocytes. III. Inhibition of deoxythymidine phosphorylaseby purines. J. Biol. Chem. 243: 4943.

18. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J.Randall. 1951. Protein measurement with the folinphenol reagent. J. Biol. Chem. 193: 265.

19. Chen, P. S., T. Y. Toribara, and H. Warner. 1956.Microdetermination of phosphorus. Anal. Chem. 28: 1756.

20. Razzell, W. E., and P. Casshyap. 1964. Substrate speci-ficity and induction of thymidine phosphorylase inEscherichia coli. J. Biol. Chem. 239: 1789.

21. Zimmerman, M., and J. Seidenberg. 1964. Deoxyribosyltransfer. I. Thymidine phosphorylase and nucleosidedeoxyribosyltransferase in normal and malignant tissues.J. Biol. Chem. 239: 2618.

22. Gallo, R. C., S. Perry, and T. R. Breitman. 1968. Inhibi-tion of human leukocyte pyrimidine deoxynucleoside syn-thesis by allopurinol and 6-mercaptopurine. Biochen.Pharmacol. 17: 2185.

23. Valentine, W. N., J. H. Follett, D. H. Solomon, and J.Reynolds. 1957. The relationship of leukocyte alkalinephosphatase to "stress," to ACTH, and to adrenal 17-OH-corticosteroids. J. Lab. Clin. Med. 49: 723.

24. Xefteris, E., W. J. Mitus, I. B. Mednicoff, and W.Dameshek. 1961. Leukocyte alkaline phosphatase inbusulfan-induced remission of chronic granulocytic leu-kemia. Blood. 18: 202.

25. Frei, E., III, J. H. Tjio, J. Whang, and P. P. Carbone.1964. Studies of the Philadelphia chromosome in patientswith chronic myelogenous leukemia. Ann. N. Y. Acad.Sci. 113: 1073.


the enzymatic mechanisms for · t+pydrwtdr+py. reaction (1) was initially characterized by friedkin...

Documents