Proc. Nat. Acad. Sci. USAVol. 69, No. 7, pp. 1761-1765, July 1972

Identification of Lactate Dehydrogenase Isoenzymes by Rapid Kinetics(stopped-flow technique/serum enzymes/ternary complex)

MICHAEL J. BISHOP, JOHANNES EVERSE, AND NATHAN 0. KAPLAN

Department of Chemistry, University of California, San Diego, La Jolla, Calif. 92037

Contributed by Nathan 0. Kaplan, April 17, 1972

ABSTRACT Stopped-flow kinetics indicate that humanand chicken heart-type4 lactate dehydrogenases (LDH) be-come inhibited as DPNH is oxidized in the presence ofhigh concentrations of pyruvate. This inhibition is muchless marked with the human and chicken muscle-type4enzymes. The initial rates and the difference in inhibitionbetween the two types of enzyme have made it possible todetermine the amount of, as well as the ratio between, thetwo types of LDH that are present in a given sample by asingle kinetic assay. The stopped-flow kinetic method hasbeen used to analyze amounts of LDH isoenzyme in dif-ferent tissues, as well as in serum.

Significant kinetic differences exist between chicken heart (H)and chicken muscle (M) lactate dehydrogenase (LDH)(1- 3). At high pyruvate concentrations, or after previousincubation with DPN+, the heart enzyme is inhibited to agreater extent than the muscle enzyme (3). Similar differ-ences have been found between the enzymes from humanheart and human liver (4).The kinetic behavior of the various isoenzymes of LDH is

governed by their subunit composition. Hence, the H2M2isoenzyme will exhibit kinetics that are almost identical withthose of a mixture of equal amounts of H4 and M4 LDH (5).

Concentrations of LDH and isoenzyme patterns are impor-tant in clinical diagnosis. Vessell and Bearn noted that theisoenzyme pattern of serum changes after myocardial infarc-tion (6). Changes in the isoenzyme pattern are observed asearly as 6- 12 hr after the initial attack of pain and have beenshown, in dogs, to be a more sensitive indicator of myo-cardial necrosis than total serum activity (7). Changes inLDH isoenzyme patterns are a constant feature in diseaseswhere liver-cell damage occurs (8). Malignancies, hemo-lytic anemias, and muscular dystrophy will also alter LDHconcentrations and isoenzyme patterns of serum (9). Thechanges are usually evaluated by electrophoretic methods,which are considerably time-consuming. We are reporting amethod involving a single assay taking less than a minutethat determines the total amount of LDH, as well as therelative concentration of H and M subunits. This method isbased on the difference in inhibition of the two different typesof LDH at high concentrations of pyruvate.

MATERIALS AND METHODS

Materials. Pyruvate was purchased from Calbiochem.DPN+ and DPNH were purchased from P-L Biochemicals.

DE-1 1 is a product of Reeves-Angel. Chicken tissues were ob-tained from commercial sources. Human tissues were pro-vided from autopsy material of the University Hospital ofSan Diego County. Normal human serum was purchasedfrom the San Diego Blood Bank. All other reagents werepurchased from commercial sources.

Enzyme Preparations. Chicken H4 and M4 and human H4enzymes were prepared by the method of Pesce et al. (10, 11).Human hybrid enzymes H3M1, H2M2, and H1M3 were sepa-rated from human heart extracts by ion-exchange chroma-tography with DE-11 cellulose (12). Human M4 enzyme wascrystallized from human liver extract, after the removal ofcontaminating proteins with DE-11 cellulose and ammoniumsulfate fractionation. All enzymes migrated as single bands ofactivity on starch-gel electrophoresis at pH 7.5.

Tissue Extracts. Human tissue extracts were prepared in0.25 M sucrose with a Waring Blendor. Tissue extracts andhuman sera were stored at -20° before assay, and werequickly frozen and rapidly thawed.

Assay Techniques. Chicken enzymes were assayed spectro-photometrically by the method of Pesce et al. (10).An Aminco-Morrow Stopped-Flow Spectrophotometer was

used to initiate and follow the reactions of the human en-zymes. The spectrophotometer was interfaced with a Fabri-Tek model 1074 Instrument computer that was triggeredsimultaneously with the initiation of mixing in the observa-tion chamber. The computer was used to digitize and averagethe changes in absorbance at 340 nm with time obtained fromfour repetitions of each assay. * One of the stopped-flowsyringes contained 0.1 M potassium phosphate buffer(pH 7.5) that was 40 mM in pyruvate and 0.26 mM inDPNH. The other syringe contained either an appropriatelydiluted enzyme solution in the same buffer or undilutedserum. Standard H4 and M4 enzymes were diluted to withinthe range of activity of normal serum. The concentrations ofthe reagents in the observation chamber were half of those inthe syringes, since equal volumes of the two solutions aremixed. All observations were made at 21°.

Activity is expressed in arbitrary units/ml. One arbitraryunit catalyzes the oxidation of 2.6 pmol of DPNH/ml per sec.

Abbreviations: LDH, lactate dehydrogenase; H, Heart-typesubunit; M, Muscle-type subunit; (AcPy)DPN +, 3-acetyl-pyridine DPN +.

1761

* An instrument that digitizes the reaction rates to an accuracyof 2% has been developed by Nguyen-Huu Xuong, RichardEntenmann, and Michael J. Bishop. A description of this instru-ment will be published elsewhere.

1762 Biochemistry: Bishop et al.

TABLE 1. Effect of pyruvate concentration on theenzymatic activity of human LDHs, as determined

by the method of Pesce et at. (10)

Isoenzyme 200 370

Human heart 0.45 0.55Human muscle 0.90 1.09

The data are presented as the ratio of the activity obtained in10 mM pyruvate to that obtained with 0.33 mM pyruvate.

RESULTSInhibition by high concentrations of pyruvate

When extracts of human heart tissue and liver tissue wereassayed in the Stopped-Flow Spectrophotometer withvarious amounts of pyruvate, a significant inhibition of theheart enzyme occurred at high pyruvate concentrations,whereas the inhibition was much less pronounced with theenzyme from liver (Table 1). The data are presented as theratio of activity obtained in the presence of 10 mM pyruvate(pyrH), with respect to that obtained with 0.33 mM pyruvate(pyrL). The data indicate that at pH 7.5, very little inhibitionby high concentrations of pyruvate is observed with themuscle isoenzyme, whereas the heart enzyme shows almost50% inhibition.These results are similar to those obtained with the lactate

dehydrogenases from chicken (3), and confirm the resultsthat were obtained by other workers with human LDH (4).

Elevation of the temperature from 200 to 370 results in anoverall decrease of the inhibition by pyruvate, as indicatedby an increase in pyrT/pyrL ratios; however, the relativeinhibition of the heart enzyme as compared to that of theliver enzyme appears not to be significantly affected by thechange in temperature (see Table 1).

Addition of oxidized coenzymes

The inhibition of chicken LDH by high concentrations of py-ruvate appears to be the result of the formation of an abortiveternary complex of LDH, DPN+, and pyruvate (3, 13). Theformation of such a complex occurs much more readily withthe H-type LDH than with the M-type enzyme. Conse-quently, the addition of oxidized coenzymes to the reactionmixture generally results in an increased inhibition, and theincrease in inhibition is more pronounced when the enzymeis incubated for a few minutes with DPN+ and pyruvatebefore the initiation of the reaction by the addition of DPNH(3).

TABLE 2. Effect of acetyl pyridine DPN+ onthe rates of human lactate dehydrogenases

% Inhibition

(AcPy)DPN+ Heart Muscle

1.4 X 10-6M 87 711.4 X 10-6M 84 601.4 X 10-7 M 48 181.4X10-8M 12 5

Pyruvate concentration is 0.33 mM.LDH was assayed spectrophotometrically by the method of

Pe3ce et at. (10).

L.

FIG. 1. Curves of time against absorbance at 340 nm forchicken (A and C) and human (B and D) 114 and MI LDHs showthe relative inhibition of each at high pyruvate concentrations.The chicken M4 enzyme (C) is uninhibited, while there is a bendwithin the first 2 see in curveD as the human MA enzyme becomesinhibited. Time scale of the curves is 2 sec/division. Vertical scalesare 0.01 absorbance units/division for the chicken enzymes and0.02 absorbance units/division for the human enzymes.

Similar observations have been made with the humanenzymes. Incubation of the human heart extract with 0.1 mMDPN+ and 0.01 M pyruvate for 10 min resulted in an inhibi-tion of more than 80% of the LDH activity, whereas theinhibition in the absence of DPN+ was about 45%. Thisphenomenon is even more pronounced when the DPN+ isreplaced by its acetylpyridine analog [(AcPy)DPN+].Table 2 summarizes the percent inhibition that is obtainedwith the human heart and liver extracts with various con-centrations of (AcPy)DPN+. The data indicate that asignificant inhibition is observed when as little as 1.4 X10-7 M (AcPy)DPN+ is present in the reaction mixture.Inhibition of enzymatic activity was also evident whenpurified human enzymes were used. Equal activities of theH4 and M4 enzyme were both inhibited upon previous incu-bation with oxidized coenzyme, but to different extents.After 5 min of incubation in 1 mM pyruvate and 28 AMDPN+, the M4 enzyme retained 75% of its initial activity,while the H4 enzyme exhibited only 23% of its initialactivity. These results dramatize the effects of low concen-trations of oxidized coenzyme on the two types of humanLDH.

TABLE 3. Inhibition of human and chicken H4 andM4 lactate dehydrogenases at high pyruvate concentrations

Human Chicken

4 MA 1H4 M4

Ratio of steady-state velocityto initial velocity 0.39 0.60 0.33 0.98

Proc. Nat. Acad. Sci. USA 69 (1972)

Lactate Dehydrogenase Isoenzyme Kinetics 1763

TABLE 4. Velocities in units/ml of reactions at highand low concentrations of pyruvate for pure

human H4 and M4 isozymes by stopped-flow measurements*

0.2 mM Pyruvate 20 mM Pyruvate

M4 H1 M4Hi

Initial velocity 620 620 720 580Steady-state velocity 620 620 450 180

* Definition of units is given in Methods.

Time-course of LDH assays

When chicken 114 LDH is assayed in the presence of highpyruvate concentrations, no substrate inhibition is observedduring the initial part of the reaction; the phenomenonbecomes observable only after a certain amount of DPN+has been formed (14). Thus, in the presence of high pyruvateconcentrations, the activity of the H4 enzyme is maximalduring the first second or less, the time being dependent onthe amount of enzyme present in the assay. During theprogress of the reaction, the activity of the enzyme decreasesas a result of the formation of the abortive complex until asteady state is obtained and the reaction proceeds more orless linearly with time (Fig. 1 and Table 3). The chicken M4enzyme remains uninhibited, whereas the chicken H4 enzymeis inhibited 66%. Similar observations were made withpurified human H4 and M4 LDH. The human H4, like thechicken H4, has a steady-state velocity about one-third of theinitial velocity. However, human M4 LDH, unlike chickenM4 LDH, also shows significant inhibition, retaining in thesteady state only slightly over half of its initial activity.The inhibitory effects on the human enzyme are also

demonstrated in Fig. 2, in which the three stages of the re-action-the initial stage, the stage of complex formation, andthe steady state-are clearly observable. The curves for timeagainst absorbance of equal activities of the two enzymesmeasured at 0.2 mM pyruvate are also shown. Both areessentially identical and linear, showing no significant in-hibition.The degree of inhibition of the enzymes is a function of the

concentration of pyruvate that is present in the reactionmixture, as shown in Table 1. When the pyruvate concen-tration is increased, first the 14 enzyme is inhibited, and athigher concentrations, the M4 enzyme is inhibited as well.We found that the greatest difference in the steady-staterates of equal activities of the two enzymes, when done in the

TABLE 5. Activities and subunit compositions of artificialmiaxtures of human H4 and M4 LDHs by

stopped-flow measurements

Predicted Observed Ratio of Ratio ofinitial initial H/M H/M

Samples velocity* velocity* predicted observed

1 698 704 22/78 19/812 682 716 45/55 38/623 673 677 59/41 58/424 661 750 73/27 78/22

* Definition of units of initial velocities is given in Methods.

FIG. 2. Inhibition of human LDH at high pyruvate concen-trations and differential inhibition of the H4 and M4 isozymes.The uppermost line shows time against Am0 for both 114 and M4enzymes at 0.2 mM pyruvate. The curve is actually composed ofthe two curves superimposed. At high pyruvate concentration(20 mM), the steady-state rate of the M4 enzyme is somewhat in-hibited, whereas the H4 activity has declined to less than 40% ofits initial rate. The initial velocity is nearly identical for all fourcurves for over 2 sec. Total time of reaction is 25 sec. The hori-zontal lines are artifacts.

stopped-flow apparatus (see Methods), is obtained with 0.02 Mpyruvate.The difference in degree of inhibition of the two human

enzymes in the presence of 0.02 M pyruvate is seen in Fig. 2.The initial activities are within 15% of each other and areclose to the activities at low concentrations of pyruvate(Table 4). However, in the steady state the M4 isozymeshows 62% of its initial activity, whereas the H4 enzymedisplays only 31% of its initial activity (Table 4).We exploited this difference in steady-state velocities in

order to determine, in a single assay, both the isoenzymecomposition and the total activity of a sample. The totalactivity of the sample is ascertained from its initial velocity.The steady-state velocity is determined from the same curveonce this stage is reached. From this data the relative pro-portions of H and M subunits can be calculated if the initialand steady-state velocities of pure H4 and M4 LDH areknown. (See Addendum.)

100

80-o

° 60

a 40 O

20 - 920

0 20 40 60 80 100% H-TYPE LDH EXPECTED

FIG 3. Subunit composition as determined by stopped-flowmethods are compared with their expected values. Mixtures ofH1 and M4 LDH of known composition were assayed in a stopped-flow apparatus, and the ratios of H/M subunits were determined.The plot shows the experimental values (points) that were ob-tained as compared to the values that were expected from thecomposition (line). Pure H1 and M4 enzymes were used as stand-ards and, therefore, define the 0% H and 100% H positions.

Proc. Nat. Acad. Sci. USA 69 (1972)

1764 Biochemistry: Bishop eta.at.

*~~~~~~

'.su;W b A.: t *1~v

'AMP > |f :. <AS

b4fe io ! rg.

human LDH isoenzymes. Time against A~u0 recordings for, from

left to right, H1M3, HUMS, and H3M1 hybrids and H4 enzyme are

shown. While the initial parts of the curves are equal, the steady-

state rate varies inversely with the percentages of H subunits.

Total time of the reactions is 25 sec.

Four solutions of different composition were made from M4

and H4 LDH solutions whose activities were previously

established. Assays in the stopped-flow spectrophotometer

gave initial velocities and subunit compositions close to the

expected values (Table 5 and Fig. 3).

In vivo, LDH exists in all five possible tet~rameric com-

binations of the H and M subunits; H4, HUM1, H2M2, H1MS,and M4. We isolated the intermediate "hybrid" isoenzymes

from human heart tissue. Assay of the hybrid fractions by the

stopped-flow technique showed heart/muscle subunit com-

positions of 73/27, 51/49, and 18/82, respectively, for the

isolated H3M1, H2M2, and HM3 enzymes (Table 6). Fig. 4

shows the~reaction rates of equal activities of the three hy-

brids, as well as for the pure H4 enzyme. The initial velocities

are all identical, but the steady-state velocities decrease in

proportion to the percentage of heart-type subunits present.

Artificial mixtures and natural hybrids with the same sub-

unit content yield the same results, even if different pro-

portions of the five LDH tetramers are present, since the

kinetic behavior appears to be determined by the individual

subunits (5).

FIG. 5. Serum LDH changes in a patient with liver disease.The curve with the steepest steady-state rate shows time againstAso for serum of a patient with acute hepatitis. Two weeks laterthe LDH activity shows more inhibition, indicating an increase inthe percentage of H subunits. The curve showing the most inhibi-tion is from the serum of a normal blood bank donor. All threecurves are adjusted to equal total activities to demonstrate thedifferences in inhibition. Full time-scale of the curves is 25 sec.

TAiBSx 6. Subunit composition of purited humanLDH hybrids assayed by the stopped-flow technique

HM1 HM2 HiM,Ratio of H/M expected 75/25 50/50 25/75Ratio of H/M observed 73/27 51/49 18/82

Tissue extractsHuman tissues have characteristic LDH isoenzyme patterns.During tissue breakdown, the enzyme is released into theserum, causing an increased concentration of the enzyme inserum, in a pattern comparable to--that seen in the tissue.Knowledge of the isoenzyme patterns in normal tissue canthus be used to help determine the source of the increasedconcentrations of LDH in serum.We measured the relative H and M activities in crude

extracts of human kidney, liver, lung, skeletal muscle, andspleen by the stopped-flow method. Subunit compositionranged from 96% M-type protein in the liver to 18% M-typein thed'kidney; these values are compatible with the 88% M and21% M-type, respectively, that have been determined byelectrophoretic methods (15). Skeletal muscle, lung, and spleenwere intermediate in composition, showing 69, 58, and 58% Msubunits, respectively, comparable to the 59, 53, and 46% Msubunits estimated from electrophoresis patterns (15). TheLDH isoenzyme pattern in skeletal muscles varies with thetype of muscle, the more aerobic muscles having a higherH/M ratio than the predominantly anaerobic muscles.A sample of blood bank serum yielded an average value of

71% H-type subunits by the stopped-flow method; this valueis within the normal range observed with electrophoreticmethods (17).

Changes in serum LDH activity during hepatitis

We observed changes in LDH activity and subunit com-position in the serum of a patient who was hospitalized forhepatitis. During the acute phase of the disease, 2 days afteradmission to the hospital, the patient's serum LDH activitywas 1512 arbitrary units (Table 7). Two weeks later thelevel had dropped to 931, and 3 weeks later it had decreasedto 826. The normal blood bank serum activity of LDH byour method gave an average value of 686 units. The pro-portion of M-type LDH in the serum dropped from 70 to53% during the 3-week period, and was thus still somewhatelevated. In Fig. 5, the reaction rates of sera from 2 daysafter admission and from 2 weeks later are compared withthat of normal serum. All three sera were diluted to the sametotal LDH activity so that the changing degree of inhibitioncould be observed.

TABLE 7. Serial LDH activities and subunit determinationfrom a patient with hepatitis by the stopped-flo technique

Total LDH Ratio of H/MDate in units/ml* subunits

10/27 1512 30/7011/10 931 38/6211/17 826 47/53

* Definition of units is given in Methods.

Proc. Nat. Acad. Sci. USA 69 (1972)

Lactate Dehydrogenase Isoenzyme Kinetics 1765

DISCUSSION

The findings reported here indicate that kinetic differencesexist between human H4 and M4 LDH that closely parallelthose found in many other species. The differential inhibitionof the heart and muscle enzymes also occurs with LDH fromfish, birds, and mammals, and its mechanism has beenextensively studied in the chicken (13).An abortive ternary complex has been proposed as the

inhibitory moiety producing the marked inhibition of the Henzyme at high pyruvate concentrations (3, 13, 18, 19). Thiscomplex consists of DPN+, pyruvate, and enzyme. Our dataindicate that the human enzymes, especially the H-type, willbe inhibited by a high concentration of pyruvate, as well as byprevious incubation with DPN+ in the presence of pyruvate.As with the chicken enzyme, this inhibition can be enhancedby substitution of (AcPy)DPN+ for the natural coenzyme(3, 13).Lack of inhibition for the first several seconds in the ab-

sence of DPN+, even in the presence of high concentrations ofpyruvate, demonstrates the requirement of DPN+ for com-plex formation. As the reaction proceeds and the oxidizedcoenzyme is produced, inhibition is initiated.The difference-in the degree of inhibition between the two

types of LDH permits the rapid determination of the ratio ofthe H and M types, as well as the total enzymatic activity ofa sample. The activity in the initial uninhibited part of thereaction correlates with the total activity as determined atlaw pyruvate concentration. The ratios of initial activity tosteady-state velocity for pure M4 and H4 enzymes can then beused to calculate the percentages of each subunit from theratio given by the sample.Experiments with the hybrid enzymes and with mixtures of

known composition of the H4 and M4 LDH demonstrate theaccuracy that is attained by this method (Tables 5 and 6).We believe that this method determines LDH subunit com-position more accurately than electrophoretic procedures.However, as with electrophoretic procedures, the accuracy ofthe determination of the ratio of H to M subunits is greatestwhen the two are present in equal concentrations and dimin-ishes as one type of subunit becomes predominant.The applicability of this method of subunit determination

to human sera should make possible its use as a diagnostic toolin clinical medicine. Our data from a patient with hepatitisdemonstrates its possible application as a monitoring devicein patients with liver disease. LDH isoenzyme concentrationsare widely used as a diagnostic indicator in myocardial in-farction. The speed of stopped-flow isoenzyme assays shouldmake the method valuable as an early indicator of myocardialdamage and should supply a mechanism for monitoringfurther infarction.We thank Dr. Nguyen-Huu Xuong for his valuable suggestions.

We also thank Drs. Lawrence Bernstein, Michael Goldhammer,and Richard Terry for their assistance in obtaining sera andautopsy specimens. This work was supported in part by grantsfrom the National Institutes of Health (CA 11683-02) and theAmerican Cancer Society (BC-60N), as well as by an Institu-tional Grant from the American Cancer Society (IN-93A).

ADDENDUMWe determined the initial velocity (Vi) by measuring the differ-ence between the absorbances at 340 nm that were obtained at

0.5 and 1.5 sec after the reaction was initiated. In order tomeasure the degree of inhibition, we again recorded the changein absorbance between 12.5 and 13.5 sec (Vs). Standards of purehuman M4 and pure H4 isozymes were assayed daily to determinethe constants used in the calculations.The velocity of the reaction at 13 sec is a linear function of its

initial velocity. That is, kiVI(s) = Vs(I) and k2VI(u) = V8(M),where VI is the initial velocity measured at 1 sec, and Vs repre-sents the steady-state velocity measured at 13 sec. By themeasurement of VI and Vs for the standards, k1 and k2 may beestablished.

Since any sample of LDH consists of H and M subunits thatare catalytically independent of each other, its rate curve can beregarded as the sum of the curves for its H and M components.The initial activity of the sample, as well as its steady-stateactivity (VI(s) and Vs(s)), are measured from the total curve.

Then, VI(s) = VI(M) + VI(a)[1]

and Vs(s) = V8(M) + Vs(H)

where VI(m),() and Vs(m),(s) represent the M and H componentsof the total curve.

But, since VB(M) and Vs(a) are functions of their respective VI's

Vs(s) = k1Vs(E) + k2VI(M). [2]

By solving (1) and [2] for VI(H), we obtain:

VI(H) = (k2VI(B) - Vs(8))/(k2 - ki)

% H subunits in the initial sample = VI(s)/VI(s) and % M sub-units = 100 - % H subunits.

1. Cahn, R. D., Kaplan, N. O., Levine, L. & Zwilling, E.(1962) Science 136, 962-969.

2. Kaplan, N. 0. (1962) Proc. Int. Congr. Biochem. 6th 3, 97-101.

3. Kaplan, N. O., Everse, J. & Admiraal, J. (1968) Ann. N.Y.Acad. Sci. 151, 400-412.

4. Latner, A. L., Siddiqui, S. A. & Skillen, A. W. (1966)Science 154, 527-529.

5. Kaplan, N. 0. (1964) Brookhaten Symp. Biol. 17, 131-153.6. Vessell, E. S. & Bearn, A. G. (1958) J. Clin. Invest. 37,

672-677.7. Wilkinson, J. H. (1970) in Isoenzymes J. B. Lippincott Co.,

(Philadelphia), pp. 43-84.8. Wieme, R. J. & Van Maercke, Y. (1961) Anti. N.Y. Acad.

Sci. 94, 898-911.9. Wilkinson, J. H. (1970) Clin. Chem. 16, 733-739.

10. Pesce, A. J., McKay, R. H., Stolzenbach, F. E., Cahn,R. D. & Kaplan, N. 0. (1964) J. Biol. Chem. 239, 1755-1761.

11. Pesce, A., Fondy, T. P., Stolzenbach, F., Castillo, F. &Kaplan, N. 0. (1967) J. Biol. Chem. 242, 2151-2167.

12. Fondy, T. P., Pesce, A., Freedberg, I., Stolzenbach, F. &Kaplan, N. 0. (1964) Biochemistry 3, 522-530.

13 Everse, J, Barnett, R. E., Thorne, C. J. R. & Kaplan, N. 0.(1971) Arch. Biochem. Biophys. 143, 444-460.

14. Everse, J., Berger, R. L. & Kaplan, N. 0. (1970) Science168, 1236-1238.

15. Wroblewski, F. & Gregory, K. F. (1961) Ann. N.Y. Acad.Sci. 94, 912-932.

16. Dawson, D. M., Goodfriend, T. L. & Kaplan, N. 0. (1964)Science 143, 929-933.

17. Cohen, L., Djordjevich, J. & Ormiste, V. (1964) J. Lab.Clin. Med. 64, 355-374.

18. Fromm, H. J. (1963) J. Biol. Chem. 238, 2938-2944.19. Gutfreund, H., Cantwell, R., McMurray, C. H., Criddle,

R. S. & Hathaway, G. (1968) Biochem. J. 106, 683-687.

Proc. Nat. Acad. Sci. USA 69 (1972)