Vol. 55

The Sulphatase of Ox Liver2. THE PURIFICATION AND PROPERTIES OF SULPHATASE A

By A. B. ROYDepartment of Biochemistry, University of Edinburgh

(Received 16 April 1953)

The first paper of this series (Roy, 1953) reportedthe presence of two distinct fractions exhibitingsulphatase activity in an aqueous extract of anacetone powder of ox liver, and gave preliminarydata on the properties of the two crude enzymes,which were separated by fractional precipitationwith acetone. The present paper describes thefurther purification and properties of fraction A,now referred to as sulphatase A.

EXPERIMENTAL

Preparation of sulphatase AThe starting material was an acetone powder of ox liverprepared as described previously (Rov, 1953). Cold acetonefractionation was carried out in an apparatus similar to thatdescribed by Askonas (1951) using thin aluminium vessels toensure rapid heat transfer. Precipitation with (NH4)2SO4was effected by adding to the solution, with constantstirring, the calculated amount of solid (NH4)2SO4; afterthe salt had dissolved the mixture was kept 4-6 hr. at roomtemperature before the precipitated protein was removed bycentrifuging. Enzyme solutions were dialysed in Viskingcellophan tubing against running tap water at room temper-ature unless otherwise specified.

Unfractionated extract. 60 g. of the acetone powder wereincubated for 1 hr. at 370 with 400 ml. water and the in-soluble material removed by centrifuging. The debris waswashed with a further 150 ml. water and the combinedsupernatants were clarified by centrifuging for 30 min.at 8000g.

Stage A. To 400 ml. unfractionated extract was added40 ml. 0-2 M-phosphate buffer, pH 7 0 (final pH 6.8) and thevolume made up to 450 ml. After bringing to 00, 340 ml.cold acetone were slowly added, the temperature beinglowered to - 9' during the process. After equilibration at- 9° for 30 min., the precipitate of fraction B (Roy, 1953)was centrifuged off at the same temperature and discarded.To the supernatant (650 ml.) kept at - 9° were added 15 ml.of the phosphate buffer, followed by a further 280 ml. ofacetone. The mixture was equilibrated as before and theprecipitate of crude sulphatase A centrifuged off at - 90,dissolved in 75 ml. water and dialysed overnight to giveapproximately 140ml. ofa clear red solution ofsulphataseA.

Stage A-1. To 140 ml. A were added 10 ml. 0-5M-sodiumacetate, pH 6-5 (final pH 7 0) and 1-5 ml. 0 3m-CaCl2. Thesolution was then precipitated with 115 ml. acetone at - 9°as described above, and the inactive precipitate discarded;the supernatant, kept at - 90, was treated with a further115 ml. acetone, the precipitate centrifuged off as before,

Table 1. The cour8e of a typical preparationof 8ulphtase A

StageUnfractionatedextractAA-1A-2A-3A-4

Total activity(s.u.)

126 000

150 000120 00090 00040 00029 000

Activity(s.u./mg. N)

38

3303 000

6 70040 000

dissolved in 25 ml. water, and dialysed overnight to give40 ml. of a solution of sulphatase A-1.

Stage A-2. Sufficient solid (NH4)2SO4 was added to A-1to make the solution 30% saturated (NH4)2SO4. Afterstanding 5 hr., the active precipitate was centrifuged off,dissolved in 10 ml. water and dialysed overnight.

Stage A-3. A-2 was 20% saturated with (NH4)2SO4, andthe inactive precipitate centrifuged off. The supernatantwas then 40% saturated with (NH4)2S04, the precipitateseparated, dissolved in 5 ml. water and dialysed overnight,giving 8 ml. of a clear, faintly straw-coloured liquid.

Stage A-4. The clear solution of A-3 was dialysed at 00against twelve or more changes of distilled water for 48 hr.and the white, flocculent precipitate of the enzyme wascentrifuged off. This was extracted for 24 hr. at 00 with4 ml. 0.1 m-NaCl, during which time the bulk of the enzymepassed into solution. The insoluble residue was centrifugedoff, washed with 0 1 m-NaCl, and the combined supernatantsstored at -10°.The concentrated enzyme solutions so obtained were

perfectly stable, even at 00, but when diluted to a concen-tration suitable for assay as described below, the enzymerapidly lost its activity. When so diluted, about 50%inactivation occurred on standing 24 hr. at 00 in the case ofsulphatase A-4. Less highly purified preparations wereconsiderably more stable.

Table 1 shows the course of a typical preparation ofsulphatase A. One sulphatase A unit (s.u.) is defined as theamount of enzyme which, under the standard conditionsdescribed below, liberates (1 pg.)I nitrocatechol. The reasonfor raising the amount of nitrocatechol liberated to thepower of I is described below.

Preparation of suphuric acid estersThe majority ofthe sulphuric acid esters were prepared by

the method of Burkhardt & Lapworth (1926) using thechlorosulphonic acid-dimethylaniline complex in CHC13solution as the sulphating agent. Dipotassium 2-hydroxy-5-nitrophenol sulphate (nitrocatechol sulphate) was prepared

653

as described previously (Roy, 1953). Steroid sulphates wereprepared by sulphation with pyridine: sulphur trioxide(private communication from Dr J. Y. F. Paterson) as werethe monosaccharide sulphates (Duff, 1949).

Estimation of enzymic activity

Routine estimations were carried out using the methodpreviously described (Roy, 1953) with nitrocatecholsulphate as substrate. The total volume of the reactionmixture was 0-8 ml. containing 0-2 ml. enzyme solution,02 ml. 0-5m-acetate buffer, pH 4-9, and 0-4 ml. 0-006M-nitrocatechol sulphate adjusted to pH 4-9 with 0.1 N-HCI,giving a final substrate concentration of 0 003 m-nitro-catechol sulphate in 0 13M-acetate buffer. After incubatingfor 1 hr. at 370, the solution was deproteinized with 3 ml.2% phosphotungstic acid in 01 N-HCI and the liberatednitrocatechol estimated colorimetrically by means ofthe redcolour developed in an alkaline quinol solution. Assayswere run in duplicate along with appropriate enzyme andsubstrate blanks. During typical assays approximately50 ,ug. nitrocatechol were liberated, corresponding to a 10%hydrolysis ofthe substrate. Under the above conditions thedegree of hydrolysis was linearly related to the time ofincubation for periods of at least 2 hr. in the case of enzymesolutions purified to stage A-3. Some preparations purifiedto stage A-4 gave a decreasing rate of hydrolysis after 1 hr.incubation.

Use of 8ub8trates other than nitrocatechol eulphate. In allcases the general technique of the enzyme assay was similarto that described above, the principal difference being in themethods of estimating the liberated phenols. In the case ofsimple phenolic compounds (phenol, cresols, naphthols,2-phenanthrol) the phenol was estimated colorimetricallyusing Folin & Ciocalteu reagent as described by Kerr,Graham & Levvy (1948) for the estimation of ,B-glucuroni-dase. To the reaction mixture at the end of incubation wasadded 3 ml. Folin & Ciocalteu reagent diluted 1 in 5 withwater. Any precipitated protein was centrifuged off and3 ml. of the supernatant pipetted into 5 ml. N-Na2CO3 forcolour development. After standing for 15 min. at 370, theblue colour was read in the Spekker absorptiometer usingflford filter no. 608 (700 m,u.). With p-nitrophenyl sulphate,the liberated p-nitrophenol was estimated colorimetricallyby means ofthe yellow colour developed in alkaline solution.The assay was similar to that described for nitrocatecholsulphate except that 3 ml. of the supernatant from thephosphotungstic acid precipitation was pipetted into 5 ml.N-Na2CO3 for colour development. The yellow colour wasread in the Spekker absorptiometer using Ilford filterno. 601 (425 my&.).

Study of inhibitory 8ubatances

The substance under consideration was dissolved in the0 5M-acetate buffer used in the routine assay describedabove to give any required concentration, after which thepH was checked (glass electrode) and if necessary adjustedto pH 4.9 with 0 1 N-HCI or NaOH. The buffered inhibitorsolutions so obtained were used in place of the normalbuffer in the routine assay, so that in all cases theenzyme was added to the previously mixed substrate andinhibitor.

I953

RESULTS

In all the following experiments the enzyme usedhad been purified as far as stage A-3, unless other-wise specified in the text.

Effect of enzyme concentration on the reactionvelocity. In the previous communication it wasreported that in the case of crude sulphatase A(stage A) the rate of hydrolysis of nitrocatecholsulphate was not directly proportional to theenzyme concentration, whereas the relation wasa linear one in the case of fraction B. It was at thattime suggested that this anomaly might be due tothe crude nature of the fraction, but it has now beenfound that highly purified preparations of sul-phatase A also exhibit this effect, as is shown inFig. 1. The effect is not due to the presence of anactivator or coenzyme, as the addition of boiledenzyme solution to the assays makes no differenceto the response to changes in enzyme concentration.

75

50 -

25.0

_ 50

.Z0z

. .o0 0-2 0W4 0Q6 0.8 1.0

Enzyme concn.

Fig. 1. Effect ofenzyme concentration on reaction velocity.Final volume of reaction mixture: 0-8 ml., containing0-2 ml. enzyme solution of varying concentration, 0 4 ml.0 006M-nitrocatechol sulphate and 0-2 ml. 0 5 M-acetatebuffer, pH 4*9. Incubated 1 hr. at 37°.

Fig. 2 shows that although the reaction velocity isnot proportional to the enzyme concentration it is,at least approximately, so related to the enzymeconcentration raised to the power of i over therange ofenzyme concentration usually studied. Thisrelationship appears to break down at higherenzyme concentrations, where the reaction velocityis even higher than would be expected from therelationship v= kE1 (v is reaction velocity, E theenzyme concentration, and k a constant). Table 2illustrates this point.

654 A. B. ROY

SULPHATASE OF OX LIVER

b-a

0-o

O 0-1 0-2 0-3 04 0-5 0-6 0-7 0-8 0-9 1-0E32

Fig. 2. Relationship between reaction velocity and theenzyme concentration raised to the power 3/2 (El).Conditions as in Fig. 1. The figuresi on the fines indicatethe state of purification of the enzyme.

suiphatase A has a well-defined optimum at pH 4 9in acetate buffers. With different preparations oftheenzyme, the position ofthe optimum varied betweenpH 4 9 and 5 0. Fig. 3 also shows the effect ofcitrate buffers on the pH optimum of sulphatase A.In 0. 13M-citrate the pH optimum is displaced toapproximately pH 5 9, and the reaction velocity atlowerpH values is considerably depressed, comparedwith the velocity in acetate buffers of correspondingpH values. It would appear that the optimum atpH 5.9 in citrate is only an apparent one broughtabout by the very strong inhibition ofthe enzyme.bycitrate below pH 6.

1.0

0l8.

tU

4-1(U

0-61

0-4

0-2Table 2. Effect of enzyme concentration

on reaction velocity

E El VobE.Vo,L if0X100X200X250-300-400-500-751X02-03-04-0

0-0320-0890-1250-1640-2530 3530-6501*002-835-208.00

1-33-54-36.08-5

13-025-539.5

134-0243-0370 0

40X639.434-436-633-636-839.339.547.4*46-8*46-3*

0

Vic.t

1-23.34.76-29.5

13-824-537-5

106*0195-0300 0

* Omitted from calculation of mean value of v.El

t ="ckE, where k is the mean value ofE

It follows from the above relationship that in thecase of enzyme assays at normal enzyme concen-

trations (liberating some 50 ,ug. nitrocatechol underthe standard conditions) it is necessary to raise theamount of nitrocatechol liberated to the power of ito obtain an approximate measure of the amount ofenzyme present. One sulphatase unit is thereforedefined as the amount of enzyme liberating (1 ,ug.)*of nitrocatechol under the standard conditionsdescribed above.

Effect of changes in pH. The effect of varying pHon the activity of the enzyme is shown in Fig. 3, theassays being carried out in 0-13M-acetate buffers ata final substrate concentration of 0 003M-nitro-catechol sulphate. It is obvious from the figure that

4 5 6pH

Fig. 3. Effect of pH on reaction velocity. Incubated for1 hr. at 370 in 0 15m-acetate or citrate buffers of varyingpH. Substrate concentration 0-003x-nitrocatechol sul-phate. Final volume of reaction mixture 0-8 ml. con-taining 0x2 ml. enzyme solution. 0-0, acetate buffer,x- x citrate buffer.

Although several early preparations ofsulphataseA gave pH-activity curves of the above type, morerecent preparations have given curves of a verydifferent nature, typical examples of which areshown in Fig. 4. This change is not due to anydetectable alteration in the experimental procedure,which has been kept constant throughout theinvestigations, and it has not so far proved possibleto repeat the preparations of specimens of sul-phatase A with a simple pH curve. The complexityof the problem is indicated in Fig. 4, which showsthat the shape ofthepH curve varies markedly withchanges in enzyme concentration. At high enzymeconcentrations the enzyme activity is greatest atpH 5-2-5-3 with a subsidiary optimum at pH 4-8-4 9: as the enzyme concentration is lowered thislatter optimum becomes more prominent and theoptimum at pH 5-2 becomes subsidiary. Finally, inlow enzyme concentrations the optimum at pH 5-2disappears, that at pH 4-8 becomes less prominent,and the major optimum appears at pH 4-5-4-6.That is, as the enzyme concentration is lowered, theoptimum shifts from pH 5-2 through pH 4-8 to 4-5.

Vol.- 55 655

%f %p

...O,X/X-11, ,X---X-

A. B. ROY

140

g 120

.0

-_ 100

n 0to

Lu

.

60

4.0

II

0

~40

20

0

I..

IFI._

0 *

-. `e e E=050

.o0 * --e..OE 02I I I p ".

IF

4 S 6pH

Fig. 4. Effect of varying enzyme concentration on the pH-activity curves. Conditions as in Fig. 3 except that theconcentration of the enzyme was varied, and onlyacetate buffers were used. The figures on the curvesindicate the relative enzyme concentrations.

This effect is quite constant, and ten entirely inde-pendent preparations at varying stages of purityhave given similar results. Unfortunately, samplesof the earlier preparations were not available forinvestigation of this problem.

The inhibition of sulphatase A by citrate sug-gested that the enzyme might require magnesiumions as an activator. As Smith (1950) showed thatthe electrophoretic behaviour of leucine amino-peptidase was altered in the presence of its acti-vator, manganese, it seemed possible that thepresence ofmagpesium might alter the shape of thepH-activity curve of sulphatase A, as any change inelectrophoretic behaviour is likely to be accom-panied by a corresponding shift in the position ofthepH optimum. The effect of magnesium ions on thepH-activity curve of the enzyme was therefore in-vestigated. No significant difference could 'bedetected between the pH curve of the normalenzyme and enzyme which had been treated with0-01M-magnesium chloride for 30 min. before assay.

Effect of change8 in 8ubstrate concentration. Fig. 5shows the effect of varying substrate concentrationon the reaction velocity. As previously reported forcrude fraction A, the reaction velocity reacheda maximum at a substrate concentration of 0 003M-nitrocatechol sulphate. Fig. 5 also gives the resultsof a typical experiment plotted according to themethod of Lineweaver & Burke (1934) and allows

bo 60

D 50

a@ 40

.0= 30

, 20(UU0,, 10z

0 0001 0002Substrate concn. (M)

0-003

Fig. 5. Effect of varying substrate concentration on thereaction velocity. Incubated for 1 hr. at 370 with varyingconcentrations of nitrocatechol sulphate in 015M-acetate buffer, pH 4*9. Final volume of reaction mixture0-8 ml. containing 0-2 ml. enzyme solution. 0-0, plotof reaction velocity against substrate concentration;x- x, the above data plotted according to theequation derived by Lineweaver & Burke (1934)

8 1 KmvV V

where 8 is the substrate concentration, v and V theobserved and maximum velocities respectively, and Kmthe Michaelis constant.

0 0001 0002 0003Substrate concn. (M)

Fig. 6. Inhibition by Na2SO4 and Na2SO3. Conditions as inFig. 5. Inhibitor concentration as indicated on the figure.Plotted according to the equation derived by Hunter &Downs (1945)

oc K.I =Ki + 8,

1-oc Km

where I and 8 are the inhibitor and substrate concentra-tions, a the fractional activity, Km the Michaelis constant,and Ki the dissociation constant of the enzyme-inhibitorcomplex.

656 I953

Vol. 55 SULPHATASE

the calculation of the Michaelis constant, KM. Inthe experiment shown the value ofKm is 7 9 x 10-4M-nitrocatechol sulphate; a further two experimentsgave values of 8-2 x 10-4M- and 7-6 x 10-4M-nitro-catechol sulphate.As the above determinations were carried out

with one of the early preparations, the experimentswere repeated using a more recent enzyme prepara-tion having multiple pH optima. The effect of sub-strate concentration was investigated with high andlow concentrations of enzyme, at pH 5-2 and 4-6respectively. As in both cases the results were verysimilar to those shown in Fig. 5 they are not re-ported here in detail. The optimum substrate con-centration was 0-003M-nitrocatechol sulphate withboth enzyme concentrations, but the values of Kmappeared to be significantly higher than thosereported above, being 12 x 10-4M-nitrocatecholsulphate in both cases.

Effect of inhibitory sub8tance8. The followingstudies were carried out with an early enzyme pre-paration having-a pH optimum of 5-0. Some of theresults of the study of a large number of substancesas possible inhibitors of sulphatase A are shown inTables 3 and 4. Table 3 shows the effect of a numberofcompounds on the enzymic activity. In confirma-tion of the results of Tanaka (1938) for molluscanphenol sulphatase, sodium sulphate was found to bea powerful inhibitor of sulphatase A, which there-fore differs in this respect from the phenol sul-phatase of Taka diastase (Taka sulphatase) as thelatter isnot inhibited bysodium sulphate (Robinson,Smith, Spencer & Williams, 1952). Like Takasulphatase, however, sulphatase A is stronglyinhibited by sodium sulphite. Fig. 6 shows that theinhibition by both these substances is competitiveand allows the calculation ofthe respective values ofKi by the method of Hunter & Downs (1945). Thesevalues are 7-5 x 10-4M-sodium sulphate and2-0 x 10-6M-sodium sulphite respectively.

( )F OX LIVER 657

The other results listed in Table 3 show little ofinterest, except that 0-05M-potassium cyanide doesnot inhibit sulphatase A, whereas Taka sulphatase iscompletely inhibited by this concentration ofcyanide (Robinson et at. 1952). Like Taka sulpha-tase, sulphatase A is strongly inhibited by phos-phate. Again, the activity of sulphatase A is un-influenced by sodium chloride, thus distinguishingit from the phenol sulphatase of the limpet (Patellavulgata) which is activated by this substance(Dodgson & Spencer, 1952). The activation of rat-liver sulphatase by potassium chloride reported byDodgson, Spencer & Thomas (1953) is probably nota true activation, but rather an apparent onebrought about by a change in the state of dis-persion of the enzyme studied by these authors.Potassium chloride is without effect onsulphatase A.

Table 3. Effect of various compounds on theactivity of suiphatase A

(Final volume of reaction mixture 0-8 ml. containing0-2 ml. 0-5M-acetate buffer, pH 4-9, 0-4 ml. 0-006m-nitrocatechol sulphate and 0-2 ml. enzyme solution. Theinhibitor was dissolved in the acetate buffer to give therequired concentration.)

Compound

Na2SO4

Na2SO3

NaClKCIMgCl2BaCl2KCNNaFKH2PO4C6H5PO4Na2C6H5SO3KC6H5SO2K

Finalconcentration

(M){ 0-025l0-00250-25 x 10-30-25 x 10-30-0025 x 10-30-050-05{0.1

0-0250-010-010-0250-0250-050-05

Inhibition(%)854010090500150

402

9510080100

Table 4. Effect of various sulphuric acid esters on the activity of sulphatase A

(Experimental conditions as in Table 3.)

CompoundPotassium methyl sulphatePotassium benzyl sulphatePotassium cyclohexyl sulphatePotassium phenyl sulphatePotassium m-tolyl sulphatePotassium 1-naphthyl sulphatePotassium 2-naphthyl sulphatePotassium 2-phenanthryl sulphatePotassium androsterone sulphateSodium glucose 6-sulphateSodium glucose 3-sulphateSodium chondroitin sulphateHeparin

Biochem. 1953, 55

Concentration Inhibition(M) (%)0-05 00-05 100-050-050-050-010-010-0020-0050-0250-025

10 mg./ml.10 mg./ml.

31030

183030300000

K,

3 x 10-11 x 10-L

4 x 10-23 x 10-24 x 10-31 x 10-2

42

A. B. ROY

Table 4 shows the effect of several sulphuric acidesters on the activity of sulphatase A. It is im-mediately obvious that only the esters of the morecomplex compounds exert any significant inhibitoryaction, there being a definite decrease in the value ofKi and therefore an increase in the affinity of theenzyme for the inhibitor with an increase in com-plexity towards a polycyclic alkyl or aryl sulphate.None of the carbohydrate sulphates so far studiedexhibited any inhibitory effect on the enzyme.

1*1

0-1

.4

.4-V.0

0-i

0.1

0-

5 6pH

Fig. 7. Effect of pH on the hydrolysis of m-tolyl sulphate.General conditions as in Fig. 3 except that the substrateconcentration was 0-1 M m-tolyl sulphate.

2-4

201

1-6 1

0 1-2,_

0L3:2 33 34

1/T x104

Fig. 8. Plot of log reaction rate (log v) against reciprocal ofabsolute temperature (1/T) for varying concentrations ofsulphatase A. Final volume of reaction mixture 0-8 ml.containing: 0-2 ml. enzyme solution, 0-2 ml. acetatebuffer, pH 5-2 and 4-6 in the high and low enzyme con-

centrations respectively; and 0-4 ml. 0 006M-nitro-catechol sulphate. Incubated for 1 hr. *-*, relativeenzyme concentration 6; x- x, relative enzyme con-

centration 1.

The degree ofinhibition is independent ofthe con-centration of the substrate for any given inhibitorconcentration, at least over the relatively smallrange of substrate concentration (0.00075M to0 003M-nitrocatechol sulphate) so far studied. Thismay be taken as an indication that the inhibition bythese sulphuric acid esters is non-competitive. Thisinhibition cannot, however, be completely non-competitive because, as described below, at leastsome of these aryl sulphates are slowly hydrolysedby the enzyme.

Hydroly8i8 of aryt &utlphata. This problem has notyet been studied exhaustively, but the resultsindicate that sulphatase A hydrolyses the simpleraryl sulphates so far studied at a much lower ratethan it does nitrocatechol sulphate. Also theaffinity of the enzyme for these sulphates is low:in the case of p-niitrophenyl sulphate and m-tolylsulphate the values of Km were 0 04 and 0-2Mrespectively, compared with a value of 0 0008M fornitrocatechol sulphate. Again, the optimum pH forthe hydrolysis of these sulphates is at approxi-mately pH 5-7 in 0-13M-acetate buffer, comparedwith the optimum at pH 4 9 for the hydrolysis ofnitrocatechol sulphate. The pH-activity curve forthe hydrolysis of m-tolyl sulphate is shown inFig. 7.As with nitrocatechol sulphate, early enzyme

preparations gave a simple pH curve for thehydrolysis of the above sulphates, while morerecent preparations have given multiple optima, themajor optimum, however, being at pH 5*7. Re-lative rates of hydrolysis have not been accuratelydetermined, but it appears that both m-tolyl andp-nitrophenyl sulphates are hydrolysed from 10 to20 times more slowly than nitrocatechol sulphate bysulphatase A. Pilot experiments have indicated thatthe rates of hydrolysis of 1- and 2-naphthylsulphates and 2-phenanthryl sulphate are of thesame order as that for the hydrolysis of m-tolylsulphate.Energy of activation of &ulphata8e A. The energy of

activation was determined at several enzyme con-centrations, in each case the enzyme activity beingassayed at the optimum pH for the appropriateenzyme concentration. Assays were carried out in0*003M-nitrocatechol sulphate at various temper-atures between 15 and 370. Fig. 8 shows the resultsof a typical experiment plotted according to theArrhenius equation (Arrhenius, 1889)

k2 A 1 1In-=----_

k, R\T T2Jwhere k is the reaction rate at absolute temperatureT, R is the gas constant and A the energy of activa-tion. From the data of Fig. 8 the energy of activa-tionmay be calculated, and in the experiment shownthe values were 6500 cal./mole and 14 000 cal./mole

o0 -

0

0

18 _ *

16 -

O2

o

I I I

658 I953

0-8

(M .

SULPHATASE OF OX LIVERat high and low enzyme concentrations respectively(relative enzyme concentrations 6 and 1, corre-sponding to the upper and lower curves of Fig. 4).

DISCUSSION

The reported method of purification gives a veryconsiderable concentration of the enzyme, as isindicated in Table 1; the final yield at stage A-4 isapproximately 20% of the enzyme present at thestage of the unfractionated extract and represents1000-fold concentration of the enzyme with respectto protein nitrogen. Even at stage A-4, however,the enzyme is not pure, as the electrophoreticpattern of a typical preparation determined atpH 6 5 in the Hilger Tiselius apparatus indicates thepresence of at least four components. In order todetermine which of the components representedsulphatase A, electrophoresis was carried out onfilter-paper strips, using an apparatus essentiallysimilar to that described by Latner (1952). Thebulk of the enzymic activity did not appear to beassociated with the major component. It may there-fore be concluded that, although a considerabledegree of purification has been attained, stage A-4does not represent a pure enzyme.

It should be noted that the method ofpreparationshows that sulphatase A is a globulin, as it is almostquantitatively precipitated by 30% saturatedammonium sulphate and is insoluble in distilledwater. Sulphatase A therefore differs markedlyfrom Taka sulphatase which Dzialoszynski (1951)has shown to be an albumin.The anomalous relationship between enzyme

concentration and reaction velocity, shown in Fig. 1and Table 3, is very unusual. It is difficult to explainthis effect but the following mechanism seems aprobable one, the more so as it has a bearing on themultiple nature of the pH curve. The basic postu-lates are two in number:

(1) That the enzyme molecules (E) are capable ofassociating reversibly with one another to forma dimeric enzyme (E2)

2E = E2The dissociation constant ofthis reaction is given by

K=n22 (1)

where n, and n2 are the concentrations ofE and E2respectively.

(2) That the dimeric form ofthe enzyme is a moreefficient catalyst than is the monomeric form, i.e.k2> k1 in the following reactions:

k]lS +E -- E + products

k2S +E2--> E2+ products.

Ifthe above postulates are valid, then the velocity ofthe enzyme reaction is given by

dn-d8=(11n1+k2n2) n,

= (k1n,+ k2Kn?) n8 (2)

where n8 is the concentration ofthe substrate. Or, interms of a total enzyme concentration, n, reckonedas monomeric units, and a, the fraction of the totalamount of enzyme which is present in the dimericform n1=(1-cx)n, n2=Jocn.Substituting the above in (2)

-dns= [kl(l-a) n+ k2K(1-x)2 n2] n8 (3)dt2where the value of a is given by the expression

4-Kn t[( 4Kn) (4)

obtained by substituting the above expressions forn, and n2 in terms of n and a into equation (1) andsolving the quadratic so obtained for m.

It is obvious from equations (3) and (4) that undersuch conditions the reaction velocity would beproportional not to the enzyme concentration, butto some power, intermediate between the first andsecond, ofthe enzyme concentration. Further, iftheenzyme were capable of polymerizing in severalstages -E, E2, E3, ..., En-the relationship shownin equation (3) would be of the same general form,provided that the second of the above postulatesremained valid.

It must be stressed that direct evidence for theabove theory will be difficult to obtain until theenzyme is obtained in a pure state, and the aboveexplanation must be regarded as tentative, themore so as no explanation can be offered as to whythe original preparations gave a simple pH curvewith a single optimnum at pH 5, while more recentpreparations have shown multiple pH optima.The first of the above postulates, however, is by

no means unlikely as many examples of protein-protein interactions have now been described. Themost studied example is that of insulin which hasdefinitely been shown (Gutfreund, 1948, 1952) toexist in various polymerized forms of a simple sub-unit; likewise Pedersen (1950) has described similarphenomena in bovine CO-haemoglobin.The second postulate is at first sight less possible,

as polymerization of the simple molecules mustdecrease the absolute number of enzymically activeparticles present, which must in turn decrease thechance of combination between the enzyme and itssubstrate, and so might be expected to reduce theenzymic activity, rather than to increase it. Thatthis is apparently not the case is indicated by the

42-2

VoI. 55 659

A. B. ROYresults shown in Fig. 8 and the values of the activa-tion energies calculated therefrom. At high enzymeconcentrations the activation energy is around7000 cal./mole whereas at low concentrations theactivation energy rises to 14 000 cal./mole. Thisimplies that at high concentrations the enzyme ismore catalytically active than at low concentra-tions, and so supports the second of the above twopostulates. It must be stressed that the abovevalues of the activation energy have little absolutesignificance as intermediate values may be ob-tained, presumably due to the fact that it is im-possible to obtain an enzyme solution completelyhomogeneous with respect to any given polymericform and therefore the values refer to solutionscontaining two or more molecular species ofenzyme.

Further evidence for the probable truth of theabove hypothesis is given by a consideration of thepH-activity curves shown in Fig. 4 which indicatea definite shift in the pH optimum with changes inenzyme concentration. If the enzyme can exist ina monomeric form and one or more polymerizedforms then it is not unlikely that each of the formswould have a definite pH optimum. From theresults shown in Fig. 4 it seems justifiable to con-clude that the pH optimum of the simple, un-polymerized form is approximatelypH 4-6 and withincreasing degrees of polymerization of the enzymethe pH optimum moves to pH 4-9 and then topH 5-2.From the inhibitory action of citrate and fluoride

on the enzyme it seems probable that sulphatase Arequires magnesium ions for its complete activity.This conclusion would not be expected from thefindings of Hommerberg (1931) and of Tanaka(1938), both of whom claimed that magnesiumchloride inhibited sulphatases from a number ofanimal tissues: this inhibition is, however, con-firmed by the results shown in Table 3 whichindicate that 0-1 M-magnesium chloride inhibitssulphatase A slightly. Lower concentrations werewithout effect, although some preparations wereactivated to a slight extent (not more than 10 %) by0-005M-magnesium chloride. This effect was un-fortunately variable. On the other hand, dialysis of

stage A-3 enzyme for 48 hr. against several changesof glass-distilled water did not significantly depressthe enzymic activity, although the enzyme wasprecipitated during the dialysis. In this connexion,Seligman, Chauncey & Nachlas (1951) claimed thatrat-liver sulphatase was activated by magnesiumchloride: it seems, however, that these claims are ofdoubtful validity. The enzyme used by theseworkers was a -'homogenate' of formalin-fixedtissues, and the so-called activation was obtained bysteeping the blocks of tissue in a solution of mag-nesium chloride before grinding: it seems probablethat the effect can be explained by the steeping inthe aqueous solution removing traces of formalin.This latter view is supported by the fact thatSeligman et al. (1951) also claimed that Fe3(SO4)2,(NH4)2SO4 was an activator of the enzyme.Only preliminary studies of the specificity of

sulphatase A have as yet been carried out, but it isobvious from the data of Table 4 that the enzymehas very little affinity for the simpler aryl sulphates,and apparently none for the few carbohydratesulphates so far studied. Also,, the simpler arylsulphates are hydrolysed much more slowly than isnitrocatechol sulphate. The considerable increase inthe affinity of the enzyme for sulphates containinga polycyclic ring system is of interest as it suggestsa possible role for the enzyme in steroid metabolism.Unfortunately, it has not been possible to investi-gate the action of the enzyme on oestrone sulphate,as no pure specimens of this substance were avail-able, all samples being badly contaminated withinorganic sulphate, nor was it possible to synthesizeoestrone sulphate in a sufficient state of purity bythe methods of Butenandt & Hofstetter (1939) orGrant & Glen (1949). In this connexion, it is ofinterest that oestrone sulphate administered to ratsis apparently very rapidly hydrolysed to oestrone(Hanahan & Everett, 1950), whereas similarlyadministered phenyl sulphate is excreted unchanged(Garton & Williams, 1949). Although sulphatase Ais strongly inhibited by androsterone sulphate, andhas a high affinity for that substance (Table 4) nohydrolysis of androsterone sulphate could bedetected, even after 18 hr. incubation with a con-

Table 5. Comparison of Taka sulphatase and aulphatase A(In both cases the substrate used was nitrocatechol sulphate.)

Type of proteinpH optimum (acetate)Optimum substrate conen.KmEffect of 0-02M-KCNEffect of 0-02M-Na2SO4Relative rate of hydrolysis ofp-nitrophenol sulphate

Sulphatase AGlobulin4-90-003M8 x 10-4MNo effect85% Inhibition1/20

* Dzialoszynski (1951).

Taka sulphataseAlbumin*5-90-003m3-5 x 10-4m100% InhibitionNo effect2

660 I953

Vol. 55 SULPHATASE OF OX LIVER 661

centrated enzyme solution. This is in accordancewith the view (Fromageot, 1938) that the sulpha-tases so far described are aryl sulphatases, specific-ally hydrolysing phenolic sulphates and having noaction on alcoholic sulphates, and contrasts sharplywith the recent claim of Henry, Thevenet & Jarrige(1952) that a sulphatase in the gut of Helix ponatiareadily hydrolyses dehydroepiandrosterone sul-phate.The above results stress the great differences

between sulphatase A and any previously describedsulphatase, especially Taka sulphatase. Table 5summarizes the principal differences between sul-phatase A and Taka sulphatase, the data for thelatter being those of Robinson et al. (1952).

SUMMARY

1. A method is described for the purification ofox-liver sulphatase A. A 1000-fold purification isachieved.

2. Differences between the original and recentenzyme preparations are described. The originalpreparations had apH optimum of 4-9-5-0 in 0-13M-acetate buffer; more recent preparations haveshown multiple optima at pH 4-6, 4-9 and 5-2.

3. In the recent preparations the pH optimumvaries with the enzyme concentration. In highenzyme concentrations the major optimum is atpH 5-2 and in low enzyme concentrations atpH 4-6.

4. The enzymic activity is not directly pro-portional to the enzyme concentration, but to the

enzyme concentration raised to some power inter-mediate between the fint and second (approxi-mately the power i). It is suggested that this effectcan be explained by polymerization of the enzymemolecules to give complexes which are more activeenzymically than are the unpolymerized molecules.

5. The variation in pH optimum with enzymeconcentration can also be explained on the abovebasis.

6. The optimum substrate concentration is0003M-nitrocatechol sulphate and the Km is8 x 10-4 M-nitrocatechol sulphate.

7. Sulphatase A is competitively inhibited bysodiumsulphateandsodiumsulphite (Ki= 7 x 10-4Mand 2 x 10- M respectively).

8. Sulphatase A hydrolyses the simpler arylsulphates at a much lower rate than it does nitro-catechol sulphate.

9. Sulphatase A is distinguished from the arylsulphatase of Taka diastase.

The author wishes to express his thanks to Dr E. A.Moelwyn-Hughes for his help in suggesting the interpreta-tion of some of the data given in this paper. He also wishesto thankProf. G. F. Marrian, F.R.S., for samples of anumberofsteroids, and Dr R. B. Dufffor a number ofpolysaccharidesulphates. The electrophoretic analysis in the Tiseliusapparatus was kindly performed by Mr H. J. Cruft. Thanksare due to Mr D. Love for carrying out the protein nitrogenestimations and to Mr R. Watt for skilled technicalassistance. The author is also deeply indebted to MessrsFrigidaire for their assistance in designing and manu-facturing the low-temperature bath used in these in-vestigations.

REFERENCES

Arrhenius, S. (1889). Z. phy8. Chem. 4, 226.Askonas, B. A. (1951). Biochem. J. 48, 42.Burkhardt, G. N. & Lapworth, A. (1926). J. chem. Soc.

p. 684.Butenandt, A. & Hofstetter, H. (1939). Hoppe-Seyl. Z. 259,

222.Dodgson, K. S. & Spencer, B. (1952). Biochem. J. 51, xliii.Dodgson, K. S. Spencer, B. & Thomas, J. (1953). Biochem.

J. 53, xxxvi.Duff, R. (1949). J. chem. Soc. p. 1597.Dzialoszynski, L. M. (1951). Bull. Soc. Ami8 Sci. PoznaA,

i1 B, 87.Fromageot, C. (1938). Ergebn. Enzymforech. 7, 50.Garton, G. A. & WiLliams, R. T. (1949). Biochem. J. 45,158.Grant, G. A. & Glen, W. L. (1949). J. Amer. chem. Soc. 71,

2255.Gutfreund, H. (1948). Biochem. J. 42, 156, 544.Gutfreund, H. (1952). Biochem. J. 50, 564.Hanahan, D. J. & Everett, N. B. (1950). J. biol. Chem. 185,

919.

Henry, R., Thevenet, M. & Jarrige, P. (1952). Bull. Soc.Chim. biol., Pari8, 34, 887.

Hommerberg, C. (1931). Hoppe-Seyl. Z. 200, 69.Hunter, A. & Downs, C. E. (1945). J. biol. Chem. 157,

427.Kerr, L. M. H., Graham, A. F. & Levvy, G. A. (1948).

Biochem. J. 42, 191.Latner, A. L. (1952). Biochem. J. 51, xii.Lineweaver, H. & Burke, D. (1934). J. Amer. chem. Soc. 56,

658.Pedersen, K. 0. (1950). Cold Spr. Harb. Symp. quant. Biol.

14, 140.Robinson, D., Smith, J. N., Spencer, B. & Williams, R. T.

(1952). Biochem. J. 51, 202.Roy, A. B. (1953). Biochem. J. 53, 12.Seligman, A. M., Chauncey, H. H. & Nachlas, M. M. (1951).

Stain Tech. 26, 19.Smith, E. L. (1950). Cold Spr. Harb. Symp. quant. Biol. 14,

148.Tanaka, S. (1938). J. Biochem., Tokyo, 28, 119.