UNCLASSIFIED

AD NUMBER

AD843466

NEW LIMITATION CHANGE

TOApproved for public release, distributionunlimited

FROMDistribution authorized to U.S. Gov't.agencies and their contractors;Administrative/Operational Use; SEP 1968.Other requests shall be referred to theArmy Biological Laboratory, Attn:Technical Releases Branch, Fort Detrick,MD 21701.

AUTHORITY

BDRL, per d/a ltr dtd 29 Sep 1971

THIS PAGE IS UNCLASSIFIED

":. AD::::::. TECHNICAL MANUSCRIPT 487

FURTHER INVESTIGATIONS

ON THE MOLECULAR WEIGHT

::::::::: OF STAPHYLOCOCCAL ENTEROTOXIN B

Roer L. SineJr.

Edar J. Schantz

SETeBE 1968

Fr drid Mrlan

RoetL SnOr

*.:.: .*.*

::::::: SPEBR16*OBO*BBB*

* *.........BR b r io J

*. . .. ..:::... :::Ewad J Shn

.:: : :

. ..:::"::::::::

DEARMET F HEARY IIFort Detric

Freerik, aryan

eooeeeo-

Reproduction of this publication in whole or inpart is prohibited except with permission of theConsanding Officer, Fort Detrick, ATTN: TechnicalReleases Branch, Technical Information Division,Fort Detrick, Frederick, Maryland, 21701. However,DDC is authorized to reproduce the publication forUnited States Government purposes.

DDC AVAILABILITY NOTICES

Qualified requesters may obtain copies of .hispublication from DDC.

Foreign announcement and dissemination of thispublication by DDC is not authorized.

Release or announcement to the public is notauthorized.

DISPOSITION INSTRUCTIONS

Destroy this publication when it is no longerneeded. Do not return it to the originator.

The findings in this publication are not to beconstrued as an official Department of the Armyposition, unless so designated by other authorized

documents.

.. .. .. ... ..

31 RLMII

013 1evil.

DEPARTMENT OF THE AM~Fort Detrick

Frederick, Maryland 21701

TECNNICAL M&NUSCRIT 487

FURTHER INVESTIGATIONS ON THE NOCLR EGUOF STAPHYLOCOCCAL EIITERUJCOMM'

David TrkUla

Robert L. sine; Jr.

Edward J. Schanltz

STATEMENT #2 UNCLASSIFIED

This document is subject to special export

controls and each transmittal to foreigngovernments or foreign nationals may be madaonly with prior approval of Dept. of Army,Fort Detrick, ATTN: Technical Release Branch/

TID, Frederick, Maryland 21701

physic&l Science Divi~iOn

BIOLOGICAL SCIENCES LAB3ATO"Y

project 1156303D164 etbr1.8

2

ACKNOWLEDGMENTS

This investigation was facilitated through the kind help of the follow-ing: Marc S. Lewis with the equilibrium sedimentation; Anne Buzzell withthe diffusion experiments; Max A. Lauffer and Severo Paglini with theosmoutry. George Anderson, William Henneberry, Joseph Gorman, PeterNelson, and Stanley R. Lieber provided skillful technical assistance atvarious times. Mrs. Marian Jones performed some of the least squaresanalyses. We thank J.B. Bateman, M.B. Sokolowski, M.S. Lewis, and L. Sperofor useful suggestions and criticism.

ABSTRACT

The molecular weight of staphylococcal entcrotoxin B was determinedindependently and by methods not previously applied to the protein; theresults agree well with those obtained in two earlier studies. In orderto check the previous determinations, analyses were performed in 0.05 Msodium phosphate buffer, pH 6.8, and in 0.1 ionic strength sodium Veronalbuffer, pH 8.6, the isoelectric point of the protein. The sedimentationand diffusion method yielded 33,000±1,400 daltons in Veronal buffer and31,800-1,000 daltons in phosphate buffer. The Yphantis meniscus-depletionsedimentation equilibrium method yielded average values of 31,000±1,300daltons in Veronal and 30,400±2,000 daltons in phosphate. Osmometryyielded an average value of 26,800±300 daltons in Veronal buffer.

3

CON~TENTS

Acknowledgments ........... ............ .. 2

Abstract .. .... ........... ........... .. 2

I. INTRODUCTION. .. ....... ............ ...... 5

II. MATERIALS AND METHODS .. ........ .............. 5A. The Protein .. ......... ........... .... 5B. Concentration .. ........ ........... .... 6C. Electrophoresis .. ......... ........... .. 6D. Sedimentation .. ........ ........... .... 6E. Diffusion .. ........ ............ ..... 6F. Buoyancy Factor .. ......... .......... ... 7G. Calculation of Molecular Weight by the Svedberg Equation . .. 10H. Equilibrium Sedimentation .. ......... ........ 10I. Osmometry. .... ........... ........... 12

III. RESULTS .. ........ ........... ......... 15

IV. DISCUSSION. .. ........ ........... ....... 24

Literature Cited .. .... ........... ......... 29Distribution List. .... ............ ........ 31DD Form 1473 .. .... ........... ........... 33

FIGURES

1. Plot of Mass of Pycnometer Contents (Normalized to 1 ml) AgainstWeight Fraction, w2, of Staphylococcal Enterotoxin B in Distilled

2. Plot of Mass of Pycnometer Contents (Normalized to I ml) AgainstWeight Fraction, V2, of Staphylococcal Enterotoxin B in 0.1 IonicStrength Veronal Buffer, pH 8.6 .. .... ....... . . . . . .9

3. Densities of Staphylococcal Enterotoucin B Protein in DialysatePhosphate Buffer at 20 C . . . . . . . . . .* . . . ...... 9

4. Plots of Specific Osmotic Pressure (,n/c) versus c for TvoOsmomtry Experiments on Staphylococcal Enterotoxin B Suspendedin 0.1 Ionic Strength, pH 8.6 Veronal Buffar . . . . . . . . . . .14

5. Photograph of Schlieren Pattern from IScaphylococcal Enterotoucin BProtein in 0.1 Ionic Strength Veronal, p11 8.6 .. .... ..... 16

6. Photograph of Schlieren Pattern Corresponding to StaphylococcalEnterotoxin B Protein at a Concentration of 7.15 r4/ml in 0.1Ionic Strength Veronal Buffer, pH1 8.6 .. .... ..... . .. 16

14

7. Sedimentation, at 20 C, of Staphylococcal Enterotoxin B at aConcentration of 7.15 mg/ml in 0.1 Ionic Strength Veronal, pH 8.6 . 17

8. Sedimentation Coefficients, in Svedbergs (10- 1 seconds), at 20.0 C,an a Function of Concentration of Two Lots of StaphylococcalEnterotaxin B in the Indicated Buffers . .. .. .. .. .. .. .. 17

9. Diffusion, at 4 C, of Staphylococcal Enterotoxin B Protein at aConcentration of 1.2 mg/ml in 0.05 M Phosphate Buffer, pH 6.8 . . . 18

10. Diffusion of Staphylococcal Enterotoxin B at a Concentration of1.2 mg/ml in 0.05 M Phosphate Buffer, pH 6.8 .... ............ 18

11. Sedimentation Equilibrium of Staphylococcal Enterotoxin B .. ..... 2212. Plot of Loge Fringe Displacements (100 Times the Fringe Displacement

in Millimeters on the Photographic Plate) versus Square of RadialPosition (in Centimeters) During Sedimentation Equilibrium ofStaphylococcal Enterotcoin B ....... .................... 22

TABLES

1. Sedimentation Coefficients of Staphylococcal Enterotoxin B atInfinite Dilution in Different Media. . . . . . . . . . . . . . . . 19"

2. Buoyancy Factors and Partial Specific Volumes of StaphylococcalEnterotoin B Protein in Different Media at 20.0 C . . . . . . . .. 19

3. Comparison of Parameters Used to Calculate Molecular Weight by theSvedberg Equation ....... ......................... 21

4. Molecular Weights of Staphylococcal Enterotoxin B ProteinDetermined by the Yphantis High-Speed Equilibrium Method ...... 23

5

I. INTRODUCTION

The molecular weight of purified staphylococcal enterotoxin B has beenexperimentally determined twice Previously. The first determination wasby Wagman, Edwards., and Schantz, who reported a value of 35,300 by sedi-mentation and diffusion (free boundary diffusion in an analytical ultra-centrifuge) and a value of 35,100 by the Archibald approach-to-equilibriummethod; these analyses were done in 0.05 M potassium phosphate buffer atpH 6.8. The second determination was by Bergdoll et al., who performedtheir experiments in 0.1 H Veronal buffer at pH 8.6, the isoelectric pointof the protein. They reported a value of 30,777 by sedimentation anddiffusion; the diffusion analysis was carried out in agar by the methoddescribed by Schantz and Lauffer.3

The purpose of the present investigation was to attempt to resolvethe discrepancy in the molecular weights reported in the two precedinginvestigations. To this end, the molecular weights were determined bysedimentation and diffusion, sedimentation equilibrium, and osmometricmethods. Experiments were conducted in sodium phosphate buffer, pH 6.8,and in sodium Veronal buffer at the isoelectric point, pH 8.6. Thus,it was possible to check the previous determinations by repeating measure-ments in the two media and by making other, different measurements.

II. MATERIALS AND =HODS

A. THE PROTEIN

Solutions of staphylococcal enterotoxin B were prepared by dissolving10- or 100-mg amounts of lyophilized protein contained in vials in buffer.The Iyophilised protein had been purified by chromatography on CC-50resin and CM-cellulose, as described by Schantz et &l.4 The protein solu-tions were, in most cases, filtered through Millipore filters (0.4 5-,Apores) prior to concentration determinations and experimental operations.The buffers and water used as solvents were routinely filtered through0.22-g or 0.45-p Millipore filters prior to use. Lot C-40 protein wasused only in a series of sedimentation experiments as a comparison wA.,lot C-41, used in all the other experiments reported hare. The proteinsolutions were routinely stored in the refrigerator at about 5 C.

!6B. CONCENTRATION

Concentration of protein was determined by WV absorption at 277 No and/or

by KJeldahl nitrogen analysis. The extinction coefficients used in the UVanalyses were 14 in 0.05 H phosphate and 14.5 in 0.1 ionic strength sodiumVeronal, for a 1% protein solution and a '-cm path length. The factor usedfor conversion of nitrogen to protein concentration was 6.21, based on anitrogen content of 16.1%.6

C. ELECTROPHDRESIS

Electrophoretic analyses were conducted in a Perkin-Elmer Model 238Electrophoresis Apparatus using schlieren optics and standard 15-M cells.The solutions were analyzed at a concentration of 5 u./ml, followingdialysis in three changes of the suspending buffers (0.1 ionic strengthsodium Veronal and 0.1 ionic strength sodium borate, both at pH 8.6).The boundaries were formed against the final dialysate. The analyseswere carried out at about 5 C. The current of 4 milliamps was maintainedfor as long as 8 hours. Photographs were taken on the Polaroid camerafurnished with the instrument.

D. SEDIIMTATION

Velocity sedimentation runs were performed in a Spinco Model E analyticalultracentrifuge using the standard schlieren optical system equipped with aphase plate. The analyses were done in 0.05 M sodium phosphate and 0.1 ionicstrength sodium Veronal buffers at pH1 6.8 and pH 8.6, respectively, at pro-tein concentrations between I and 10 mg/ml. All runs were conducted at59,780 rpm using a rubber valve-type, alvuinum synthetic boundary cell,with 4 degree sector and 12 me thickness, to facilitate formation of theboundary.

Sedimentation coefficients were calculated from displacements ofmaximum ordinates of the gradient curves as functions of time, accordingto the method described by Schachman. The reference distances wereobtained by measurement of distances of the counterbalance referenceedge* from the rotor axis on a Gaertner Hicrocomparator. Referencedistances were corrected for rotor stretch at 59,780 rpm according tothe method described by Schachmsn.

g. DIUFUSION

Diffusion analyses were done accordin# to the method described bySchumeker and Schachman and by Pedersen. UV absorption optics wereemployed in the Spinco Model E analytical ultracentrifuge with recording

7I

ton Kodak Cmnnrcial film. One diffusion run was performed in 0.05 M sodiumphosphate buffer, pH 6.8, at 4 C. The other diffusion experiment wascarried out in 0.1 ionic strength sodium borate buffer, pH 8.6, at 20.0 C.This experiment would have been done in Veronal buffer, except that thebuffer absorbs at 254 qi, the principal wavelength of the UV lamp on thecentrifuge. The concentration of enterotcin was 1.2 mg/ml in bothexperiments; the protein had been dialyzed three times against the suspend-ing buffer. The salt content of the protein solution was increased by0.025 M with KCI to prevent boundary disturbance when the overlay bufferwas deposited in the valve-type synthetic boundary cell. The overlaybuffer was the third dialysate in both cases. Both experiments were doneat 21,740 rpm. A double-beam recording microdensitometer* was used toconvert the film images into traces of optical density as a function ofposition in the centrifuge cell.

Diffusion coefficients were calculated by dividing the slope of thePlot of un2 (I - w St) versus t by 3.64. The radial distance between the25% and 75% optical density points in the boundary region is un, t is timein seconds, and (1 - w2St) is the correction for rotor speed, where w isangular velocity of the rotor, and S is the sedimentation coefficient ofthe protein under the conditions of the diffusion experiment. The calcu-lated diffusion coefficient at 4 C was corrected to 20 C by use of therelationship

D - D' TLT'TI

where D' is the calculated diffusion coefficient at Kelvin temperatureT' (277.16 K), D is the diffusion coefficient corrected to 20 C (T a

293.16 K), %' is the viscosity of water at 4 C, and TLw is the viscosityof water at 20 C.

F. BUOYANCY FACTOR

Buoyancy factors (I - ip) were determined by pycnometry in 10- and20-ul Lipkin pyenometers** at protein concentrations ranging between 2and 18 w/.l in water, sodium phosphate buffer (0.05 11, pH 6.8) andsodium Veronal buffer (0.1 ionic strength, pH 8.6) at 20.0±0.1 C. Proteinconcentrations were determined by Kjeldahl nitrogen analyses. The nitro-Sen values for protein in Veronal were corrected by subtracting the valuefor nitrogen determined on a Veronal blank.

* 311, Joyce, LOe and Co., Ltd.Ace Class Co., Vineland, N.J.

I

8

The pycnometers were calibrated with water. Weighing was on anAinsworth Type SCN balance or a Voland Universal Analytical Balance, Model860D. Buoyancy corrections were applied to all weights. Weights ofpycnometer contents were converted to a pycnometer capacity of 1 ml forthe calculations.

In the detecmination of (I - VP) in water, the protein was preparedby dialysis against distilled water, and the successive concentrations ofprotein were prepared by dilution of the dialyzed protein stock withfiltered distilled water; filtered distilled water was used for the zeroprotein concentration point (Fig. I). For the determination in Veronalbuffer, the stock protein was prepared by dialysis against stock Veronalbuffer, and the successive concentrations of protein were prepared bydilution of the dialyzed protein stock with stock Veronal buffer; thedensity of stock Veronal buffer was used as the zero protein concentrationpoint (Fig. 2). For the (I - VP) determination in phosphate buffer, theprotein solutions were prepared in two ways. In one experiment, stockprotein solution was prepared by dialysis against stock phosphate buffer,and successive dilutions were made in stock phosphate buffer; the densityof stock phosphate buffer was used as the zero concentration point. Inthe other experiment, stock protein solution was prepared by dialysisagainst stock phosphate buffer. Successive concentrations of proteinwere, however, prepared by dilution of the dialyzed non-isoelectric proteinstock solution with dialysate;8 the density of the dialysate buffer solu-tion was used as the zero concentration point in the graphical estimationof (1 -vp) (Fig. 3).

1.0020

1.0010

S

* 1.0000I

0.9990

0.9960p p I II

2 4 6 10 12

w 2 x103

Figure 1. Plot of Mass of Pycnometer Contents (Norualized toI ml) Against Weight Fraction, V2, of StaphylococcalEnterotoxin B in Distilled Water.

%w4J . 4

0 44

As! f 0.4 as

m 0d

~~~4 86 4 4

o 4 0

a g.

.44 .

0 0 44 AJ

4 .J-

G0~I I I I El 140*

0 4 Oci .4 '4

0 p-

.44

-4

0

4

04.

r0 14 D

0

15 .

4.' -0 .J

P4

10

Buoyancy factors were calculated according to the procedure describedby KayR and by Kraemer.10 Weight of pycnometer contents, m, is plottedversus weight fraction, w2, of protein solution. If a straight line isobtained (as was the case in all our experiments, Figures 1 through 3),the buoyancy factor for sedimentation, (1 - VP), is evaluated by multiply-ing the slope of the straight line by the reciprocal of the mass of thepycnometer contents at zero protein concentration (buffer or water). Thisprocedure uses the relationship

(1 - VP) 1 - w2 dm

m dw2

When the slope dm/dw2 is evaluated at zero protein concentration, w2 equalszero, and m equals the mass of water or buffer in the pycnometer. Thepartial specific volume, V, is calculated from (1 - Vp) by taking p asthe density of the solvent-the solution used to obtain the zero concentra-tion point. The straight line relating m and w2 was calculated by themethod of least squares, considering the zero concentration point to haveno variation.

G. CALCULATION OF MOLECULAR WEIGHT BY THE SVEDBERG EQUATION

The Svedberg equation

RTSoB

DP (1 - VP)20,B

was used to calculate the anhydrous molecular weights of the prote6n fromthe sedimentation and diffusion coefficients. In this equation, S 2 0 andD 0 represent the sedimentation and diffusion coefficients, respectIvely,o T e protein at infinite dilution in the buffer, B, at 20 C. Thebuoyancy factor, (1 - VP), has the value corresponding to the same buffermedium (and at the same temperature) employed in the determinations of Sand D.

H. EQUILIBRIUM SEDIMENTATION

Equilibrium sedimentation analyses were performed according to themeniscus depletion method of Yphantise in 0.05.M sodium phosphate bufferat pH 6.8 and 0.1 ionic strength sodium Veronal buffer at pH 8.6; theprotein had been dialyzel three times against the suspending buffer. Theruns were done in 12-mm thick, Yphantis, six-channel, Kel-F centerpiecesat protein concentrations ranging between 0.18 and 0.60 mg/ml. The thirddialysate was used as the reference buffer in each case. The Spinco Model E

pt

II

analytical ultracentrifuge used in these studies was equipped with anelectronic speed control. Concentration distributions were recorded withthe Rayleigh optical system, utilizing a symmetrical, medium slit-widthlens mask. Rotor speeds were 29,500 rpm for phosphate buffer and 30,000for Veronal buffer. The temperature of the runs was 20.0 C. KodakMetallographic plates or ll-G plates were used for photographic recording.Sapphire windows were employed to minimize distortion of the interferencepattern. A photograph taken after 16 hours of centrifugation was measuredfor the run in phosphate; photographs taken at 19 hours and at 22 hourswere measured for the Veronal run.

The photographs of the fringe patterns were measured on a NikonShadowgraph Model 6C comparator, by both the fringe-location and thefringe-trace methods. In the fringe-location method, the fringe positionis determined by measuring radial coordinates at half-fringe increments,as fringes are crossed in the radial traversal of the fringe pattern fromthe protein-depleted solvent region through the protein solution region.In the fringe-trace method, the vertical displacement of a single fringefrom its meniscus position, or the average displacement of several fringes,is measured at a number of radial positions as the fringe is traversedfrom the region of protein-depleted solvent through the protein solutionregion. In most cases, when the fringe-trace method was used, three blankfringes in the central region of the diffraction envelope were measuredat each radial position. The fringe distortion correction had no signifi-cant effect because the maximum deviation observed in blank runs was 1/10fringe, which in these experiments amounts to about 30 g. Because theuncertainty in the measurement of the fringes is of the order of 10 4, thecorrection for blank shift was considered to be within reasonable experi-mental error. Moreover, the application of the correction d d not signifi-cantly alter the values of the slopes obtained for d(lnc)/dri .

In the fringe-trace method, the fringe displacement was measured atradial increments ranging from 500 to 50 p. on the photographic plate,depending on the vertical displacement; the interval was selected togive approximately equal intervals in vertical displacement. Measurementswere made to the point where the fringes became blurred at the bottom andit was impossible to measure their positions with precision. Fringe dis-placements below 100 p on the photographic plate were not included in thestatistical analyses because of the large variation in measurement."

The fringe displacement is proportional to refractive index increment(solution to solvent), which is proportional to protein concentration indilute solutions. Thus the fringe displacement is proportional to proteinconcentration, and the concentration distribution at equilibrium in thecentrifuge cell can be obtained from the fringe shifts at a number oflevels in the cell.

-. . . ... .. . . . . . .. .. . . .. ........... .....-- - - - -. ....-

12

Molecular weights are calculated with the formulai"

2RT d(lnc)M w'(l - Vp) dr2

In practice, In (10F) or in (100Y) was plotted versus r2, where F is thenumber of intercepted fringes at half-fringe increments and Y is displace-ment (as measured on the photographic plate) of a fringe relative to thesolvent base line in the protein-depleted region of the c' 1. The slopesdin (IOF)/dr2 and din (IOOY)/dr2 are equivalent to d(lnc)/dr2, where c isprotein concentration and r is distance of the radial position in thecell from the axis of rotation. V is the partial specific volume of theprotein as determined in the solvent used in the equilibrium experiment;P is solvent density, w is angular velocity of the rotor in radians persecond, R is the gas constant, and T is the absolute temperature.

Slopes of the straight lines obtained in the plots of the naturallogarithm of fringe displacement (10F or 100Y) versus square of radialposition were evaluated by the method of least squares. Molecular weightsand their standard deviations (as evaluated from standard deviations ofslopes) were calculated for each set of measurements.

I. OSMOMETRY

Osmotic pressure measurements were performed in the Mechrolab High-Speed Membrane Osmometer, using Schleicher and Schuell Co. Bac-T-FlexMembrane Filters. Type B-19 membranes were used for osmometry with 0.05 M,pH 6.8 sodium phosphate buffer as the medium. Type B-20 membranes withsmiller pore size (O.OOV) were used for osmometry with 0.1 ionic strength,pH 8.6 sodium Veronal as the medium. The third dialysate buffer, whichwas used in all the measurements, was de-aerated at room temperature. Themembranes, which had been stored in alcohol, were rinsed, soaked andde-aerated in this dialysate buffer and then mounted on the osmometerchamber. Under the conditions of the measurements, at least 3 hourswere required for equilibration of the membranes against third dialyeatebuffer. Solutions, in which the protein concentrations (determined byOD measurements) ranged between 1 and 11 mg/ml in the dialysate buffer,were introduced into the osmometer, and the membranes were allowed toequilibrate at least 10 minutes. The equilibrium osmotic pressures wereread in centimeters of dialyeate buffer. All measurements were made at20.0 C. Several experiments were performed with phosphate buffer as themedium, using a different membrane in each experiment, and two differentexperiments were performed with Veronal buffer as the medium.

There was evidence for leaky membranes in the experiments in phosphatebuffer, using B-19 membranes. Therefore, the osmometry results in phosphatewill not be considered, except to state that molecular weight values in theneighborhood of 30,000 to 40,000 daltons were obtained.

13

The fact that horizontal straight lines represent the data in theosmometry experiments when it/c is plotted versus c (Fig. 4) shows thatthe solutions were ideal and made possible the calculation of M from thesimple van't Hoff relationship

H _ RT

If/C

M is molecular weight of the protein, R is the gas constant, T is theabsolute temperature, and n/c is specific osmotic pressure in centimetersof Veronal buffer per gram per liter of protein. To calculate M as justdescribed, it is necessary to convert the gas constant, R, from units ofliter-atmospheres per mole per degree (Kelvin) to units of liter-cm bufferper mole per degree as follows:

liter-atm 76.0 cm Hg 13.5- x 1

mole-degree atm cmHg'cm2 PB

where PB is density of the buffer employed. For the experiments in Veronalbuffer, PB - 1.0033 g/ml. Thus for the experiments in Veronal buffer at20.0 C, the molecular weight was calculated from the formula

,(liter-cm buffer)

M 24688.4 MOLeIt/c

t

14

1.000

xx -X

< 0.900 0 0 0

0x

0.800

I I I I I I I I0 1 2 3 4 5 6 7 8

C

Figure 4. Plots of Specific Osmotic Pressure (x/c) versus c for TvoOsuotry kperiets an Staphylococcal Enterotoxin 3Suspended in 0.1 Ionic Strength, p1 8.6 Veronal Buffer.The horLstontal, 'traight, least squares lines vere dramthrough the five points representing the finsl muasurementsmade after equilibretion of the mmbranes. Units: w incentimeters Veronal buffer, c in grm per liter.

15

III. RESULTS

Only one component was observed in the sedimentation, diffusion, andelectrophoresis experiments. Typical schlieren patterns obtained inelectrophoresis and sedimentation experiments are shown in Figures 5 and6, respectively.

Measurement of the maximum ordinate of the schlieren pattern obtainedfollowing 7 hours under free-boundary electrophoresis conditions in Veronalbuffer indicated zero displacement and thus zero electrophoretic mobilityat pH 8.6. A measurement on the photograph obtained following 8 hoursunder free-boundary electrophoresis conditions in borate buffer indicateda possible alight displacement, from which an electrophoretic mobility ofabout 3 x 10 . 6 cm2/volt-sec was calculated. This displacement was con-sidered to be insignificant.

Plots of natural logarithm of displacement of the boundary from theaxis of rotation versus time of sedimentation of the protein were linearin all cases. A typical plot is shown in Figure 7.

Sedimentation coefficients plotted as functions of concentration ofenterotoxin B protein are shown graphically in Figure 8 for runs in phos-phate and Veronal buffers. The sedimentation results for both lot C-40and lot C-41 are represented. The linear least squares intercepts,corresponding to sedimentation coefficients at infinite dilution, andtheir standard deviations are shown in Table 1.

Figure 9 is a set of photographs of toxin protein diffusing inphosphate buffer in the centrifuge cell. The quantity u' (1 - w2St),representing the squared spread (corrected for rotor speed) of theconcentration distribution between the 0.25 and 0.75 co levels in theboundary, is plotted as a function of time of diffusion in phosphatebuffer in Figure 10. The diffusion coefficient and its standard devia-tion, calculated from the least squares slope of this plot, are 8.42±0.25 x 10. 7 cm2/sec. For the diffusion ,xpelisent in borate buffer,the calculations yielded 7.96±0.34 x 10 cm/sec for the diffusioncoefficient and its standard deviation. In the calculations of molecularweight, these values for the diffusion coefficient were considered to beequal to the infinite dilution values because the initial concentrationsof 1.2 is/ml were quite low.

The densities of enterotoxin B protein solutions are plotted as func-tions of concentration of protein in Figure 1 for the experiment in distilledwater, in Figure 2 for the experiment in Veronal buffer, and in Figure 3 forthe experiment in phosphate buffer in which the dialysate was used for dilu-tion and comparison. The values (I - TP) and their standard deviations,calculated from the linear least squares slopes, are presented in Table 2.The corresponding values for V are also shown.

16

-0 -.

I"I

. I 'o.A

*ao SO&.0w w

En a:toI4 Id 2

ud OY)Iow'Id

C..-

~~00

* tj

17

00C-, ,6 0

'. a

ooa. a 01 ,

OC; ".3 o V,

o " 0 A

o ++. .o os.. 60iI I I- o 0

m c v** 0 6

WO .0 .*.4

0 r . Ido + "g'4

r 0 eI Ci 0, con- Km in

I

o .

0

OOhl

, I o 1 .,

i j" I .13~ ,l

lo a# 0--, I 10. .dip

I IImIu III

, +i !iii ,, ii

* w I

.41

" m %A...

| • m • • • im Sol

see 41 ze

t4 a e4:,

o0

00

0 434

'Wi 2

o to

19

TABLE 1. SEDIMENTATION COEFFICIENTS OFSTAPHYLOCOCCAL ENTEROTOXIN B

AT INFINITE DILUTIONIN DIFFERENT MEDIA

Lot No. edi / b/ 10-13 sec20,B

C-41 Phosphate 2.88±0.02

C-41 Veronal 2.83±0.03

C-40 Phosphate 2.78±0.04

a. 0.05 M sodium phosphate at pH 6.8 and 0.1 ionicstrength sodium Veronal at pH 8.6.

b. The values represent the least squares interceptsand their standard deviations, derived from plotsof sedimentation coefficients versus concentrationsof enterotoxin protein. SO B stands for sedimen-tation coefficient at infinie dilution in buffer,B, at 20 C.

TABLE 2. BUOYANCY FACTORS AND PARTIAL SPEIIC VOLUMESOf STAPHYLOCOCCAL ENTERCtOIN B PROTEIN

IN DIFFERENT MEDIA AT 20.0 C

-diu&V (I - p)I , sml l/

Phosphate 0.2624±0.0039 1.0045 0.734

Veronal 0.2623±0.0009 1.0033 0.735

DistilledWater 0.2813±0.0009 0.99823 0.720

a. 0.1 ionic strength sodium Veronal at pH 8.6 and 0.S Ksodium phosphate at pH 6.8.

b. The values were calculated from least squares slopesof plots of densities versus concentrations of entero-toin protein; the standard deviations listed werecalculated from standard deviations of the leastsquares slopes.

20

A value of 31,800±1,000 daltons was calculated for the anhydrousmolecular weight of the protein, using the values of the sedimentationand diffusion coefficients and the buoyancy factor for protein in phosphatebuffer. From the sedimentation coefficient and buoyancy factor in Veronaland the diffusion coefficient in borate, a value of 33,000-1,400 daltonswas calculated for the protein at the isoelectric point, pH 8.6. Thestandard deviations of these molecular weights were derived from thestandard deviations of the sedimentation and diffusion coefficients andof the buoyancy factors, as applied in the Svedberg equation. A sumaryof the experimental parameters used to calculate molecular weights withthe Svedberg equation is given in Table 3; molecular weight results aswell as comparative values from Wagman et al.' and Bergdoll et al., arealso given there.

Figure 11 is a photograph of the interference fringe patterns obtainedin the experiment in Veronal buffer. Straight lines were obtained in allthe plots of the natural logarithm of fringe displacement as a function ofradius squared in the sedimentation equilibrium experiments at positionscorresponding to fringe displacements greater than 100 g. A typical plotis shown in Figure 12. This plot represents data obtained by measurementof fringe displacements corresponding to sedimentation equilibrium of theprotein, following centrifugation for 19 hours in Veronal buffer; theinitial concentration of protein was 7/12 mg/ml.

The molecular weights calculated from data obtained through a numberof independent mea--irements of the Rayleigh interference patterns forVeronal and phosph-.te buffer media are sumnarized in Table 4. The thirdsolution in the run in phosphate was not included in the results becausea shift of less than one fringe was observed at the concentration employed(-2/11 mg/ml), leading to unreliable data. The standard deviations of themolecular weights are derived from the standard deviations of the correspond-ing buoyancy factors (1 - Vp) and from standard deviations of slopes ofleast squares plots of In fringe displacement versus r

Molecular weights of 27,400±400 daltons and 26,200±200 daltons werecalculated Lrom the osmotic pressure data for the experiments in 0.1 ionicstrength Veronal buffer, pH 8.6. The osmotic pressure data for the twoexperiments are shown in Figure 4. The ratio n/c is plotted as a functionof c; v is the osmotic pressure in centimeters of Veronal buffer as readon the osmometer, and c is concentration of protein in grams per liter.The points corresponding to the lowest concentrations used in each experi-ment were omitted in the calculations of molecular weight because the leastconcentrated solutions were the first introduced into the osmometer ineach experiment and the osmometer was not considered to be equilibrateduntil the second solution was intrcduced. A molecular weight and itsstandard deviation of 26,800*300 daltons were calculated from the tenselected points pooled from the two experiments.

_A-

21

,4 I-

f a C! a4 9cIt 0 0 co

+1 C-4 -4 0-W o c 0 0n 1.. c0

04K 00 0 0 c i00 000

0 0

14 4 v4Jr--. .' 0m * en 0 P0

0 0ft 01.'

044 5.40> )

0 0

0 00

0 0 1. 0 1

C0 0 en -4

1-4 0 0 0d r.UA~ +1 1 040 fnO

s-I en 10 %C4 NNc >

P -4 0O 0% N* 0..f5w 4 0c.40n -4 44 0 4J0

0 ~ 44 44 J 0 00 >I

44 0d -4 U 0 co..4N n4 00 0O 0 144'04b

N . .0 W0j .10P4 4 - c0.0 U

0 m n 4 00J45440 UUN -0 >

co ~ ~ ~ ~ ~ . 0.0 0l * * 0tiw- 0

5. -N S"41 -4 c

044 0 44.0"40

E-4 u m40 0 .e

:0 -4 114JJ-4 444W

ON4 >-

22

0 9 w Qd.CV M .

ga 00u

(44>

OId - C-

(0 0 "S4 .4 @ - 0

0 8~ .44k4.S

in in

.,

(A Q0

-44

*.4 a~4I

U41 'A 0 -

~I'1JBj

to

4

id

23

4 -4 4 N NI 0

a

za a

w4 C; 1

OJJ4J'

u~i 0 ou s0N s 0 44 I ,

14 U* ad

0a - 0-

0 4 I

Ioy w'" II '1 w l

44 1 e! ' i :I'4 of&a

'4aC 4 ) 1'I

* .- 4. *.4i...,.II' * i O/

I0~~S it -11 a

km i 1

la 88 as-88 41 Z, - ,'

1'1 GO 4.1 wuo -.*eu mm 't.8=

VO I 0NI

In'1 ,., in, N ft,,., nV%,

164~

* tt ,'jj lJ ,ii"

oI41 i l l

A d, i 1

4 .6 ,** U ,~

24

IV. DISCUSSION

The fact that there was either no, or at least no significant, displace-ment of the toxin under the electrophoresis conditions in either Veronal orborate buffer indicates that the toxin protein is isoelectric in Veronal andborate buffers, 0.1 ionic strength at pH 8.6. The presence of a single,sharp schlieren peak during prolonged electrophoresis in each buffer isevidence for the electrical homogeneity of the protein.

Values of the experimental parameters used by us, by Bergdoll et al.,3

and by Wagman et al. 1 for calculation of M by the Svedberg equation areshown for comparison in Table 3.

The infinite dilution value of the sedimentation coefficient at 20.0 C(s20) seems to be somewhat less for lot C-40 (2.78 Svedbergs) than forlot C-41 (2.88 Svedb rgs) toxin protein in 0.05 M phosphate buffer, pH6.8. The value of SIO for lot C-41 toxin in 0.1 ionic strength Veronal,pH 8.6, is about the same (2.83 Svedbergs) as the value in phosphate.These values of si are not significantly different from the valueobtained by Bergdoll et al. of 2.78 Svedbergs for sOO w derived fromvelocity sedimentation runs in 0.1 ionic strength Vergal, RH 8.6, andthe value of 2.89 Svedbergs obtained by Waguman et al. for SI0 w derivedfrom experiments in 0.05 M phosphate buffer, pH 6.8. Thus, there is noapparent gifference in S80 among the various lots of toxin employed, orbetween 820 in phosphate and in Veronal.

The apparent greater dependence of S on c in Veronal compared withphosphate shown in Figure 8 is not considered significant in view of thefact that Wagman et al. show a greater dependence of S on c in pH 6.8phosphate than do Bergdoll et al. for experiments in pH 8.6 Veronal.

Because there is no significant difference between the S0 values weobtained for sedimentation in Veronal (sero net charge on the protein)and pH 6.8 phosphate buffers, it can be concluded that the salt concentra-tion of the buffer is sufficient to eliminate charg4 effects.

The diffusion.goefficients in phosphate (8.42 x 10, 7 cz 2 /sec) andborate (7.96 x 10' cm'/sac) buffers are not considered to be tignificantlydifferent from each other nor from the value of 8.22 x 10 m /secobtained by Bergdoll et al. for diffusion of the toxin in 0.1 ionicstlength, pH 8.6 Veronal buffer in agar gel. The value of 7.72 x 10' 7

c /seac obtained by Wagmn et al. for diffusion in 0.05 N phosphatebuffer, pH 6.8, differs considerably from our value of 8.42 z 107 CS2 /secin phosphate buffer; but this difference is still considered to be withinthe range of experimental error. The fact that the diffusion coefficientsare the same in phosphate and Veronal indicates that there is no chargeeffect for the diffusion in 0.05 K phosphate buffer.

25

The values 0.2623 and 0.2624 derived for buoyancy factors for theprotein at 20 C in Veronal and phosphate, respectively, are considered toagree with each other within experimental error. The value 0.2813 derivedfor the buoyancy factor of the protein at 20 C in distilled water isprobably significantly different from the other two.

When the buoyancy factors, (I - VP), are converted to partial specificvolumes, V, the values 0.735 and 0.734 ml/g for 7 in Veronal and phosphate,respectively, do not differ significantly from each other (Table 2).Neither do they differ significantly from the values of 0.731 and 0.726ml/g calculated by Spero* and by Bergdoll et al., respectively, from theamino acid composition of the protein. Also, the value 0.743 ml/I inwater obtained by Wagman et al. agrees within experimental error' withour values in Veronal and phosphate.

There does not seem to be a pH effect on V, observed by Charlwood," 6

according to which % increases consid.erably as the pH is increased ordecreased from the isoelectric pH. The lack of a pH effect on %T isconsistent with the constancy of S between Veronal and phosphate media.Charlwood and Ens1s noted a decrease in S for bovine serum albumin at pHvalues less than the isoelectric point. The fact that the values weobtained for 7 in Veronal and phosphate, 0.735 and 0.734 ml/g, respectively,are almost the same indicates that there are no significant Donnan effectsin the phosphate buffer in which the protein is not isoelectric. Thisagrees with the results obtained by Casaesa and Eisenberg.8 Also it isinteresting to note that the value (7 - 0.727 ml/g) in the determinationin phosphate in which dialyzed protein was diluted and compared with stockphosphate buffer does not differ significantly from the value (V - 0.734ml/g) obtained in the experiment in which dialyzed protein was dilutedwith dialysate and the dialysate was used in obtaining the comparisondensity.

Our value for V of 0.720 ml/g at 20 C in distilled water is consideredto differ significantly from our values in phosphate and Veronal. We donot know how to explain this result. The value (V- 0.693 ml/&) obtainedin an experiment in which protein was dialyzed against distilled water,then made 0.05 X in phosphate, and then diluted and compared with 0.05 Kphosphate buffer, pH 6.8 at 20 C is considered unreliable, because therewas more scatter of the points about the regression line than in any ofthe other experiments.

A value (7 0.716 al/&) obtained for the protein at 25 C in phosphateby dilution of the dialyzed protein with dialyeate and comparison withdialysate also is probably significantly different frca 7 - 0.734 al/&for the protein at 20 C in phosphate. We cannot explain the lesser valuefor V at 25 C than at 20 CO in view of the fact that Kxaf reported apositive temperature dependence of 7 for several muscle proteins.

Spero, L., 1968, personal coeuinication.

26

The agreement between our H values (as calculated by the Svedbergequation) in Veronal and phosphate confirms that there is no significantcharge effect and no significantly different ion binding effects in thetwo solvents. The results just discussed and the data in Table 3 showthat the discrepancies between the molecular weights we calculated fromthe Svedberg equation and those calculated by Bergdoll et al. and by Wagmanet al. can be explained in terms of experimental errors. The value inVernal of 30,777 reported by Bergdoll et al. is lower than our value of33,000 because of nonsignificant differences in S%0, D%, and V; the value20.1ain phosphate of 35,300 reported by Wagman et al. Is greater theR our valueof 31,800, mainly because of the nonsignificant difference in Dj0. Thus,we conclude that the four values for M derived by the method of sedimenta-tion and diffusion agree within experimental error.

The straight lines obtained in all the lnc versus r2 plots are con-sistent with the interpretation that the protein solutions are monodisperseand ideal. If the lines were concave upwards, heterogeneity would beindicated; if concave downwards, non-ideality would be indicated."Straight lines were obtained in these plots at all the different loadingconcentrations over the same range of fringe displacements. This is alsoconsistent with molecular homogeneity of the protein."

We do not consider our values for M obtained by the Yphantis high-speed equilibrium mthod to be significantly different from each otherIn the phosphate series or the Veronal series. Values are listed inTable 4 for different loading concentrations of tmin protein, differenttime of centrifugation at the equilibrium speed, and different methodsof measure-t. The mean value for the experiment in phosphate bufferis 30,400.2,000 daltons; the mean value for the experiment In Veronalbuffer is 31,000±1,300 daltons.

It should be pointed out that Waguan et al. obtained higher values(34,400 average) for H of the toxin protein in phosphate by the Archibaldapproach-to-equilibrium method than we did by the Yphantis method(30,400 average). Because this difference cannot be attributed to the(1 - VP) factor, the results of measurements using the Archibald methodas compared with the Yphantis method appear to differ.

Again, because the values for molecular weight by the sedimentationequilibrium method in phosphate and Veronal agree so well, there is noevidence for charge effect in phosphate buffer or for significantlydifferent ion-binding effects in phosphate as compared with Veronalbuffers. The agreement between molecular Weight values at the severalloading concentrations employed is evidence that the solutions areessentially thermodynamically ideal at these concentrations.

The agreement between the molecular weight values obtained by themethods of sedimentation equilibrium and sedimentation and diffusion canbe taken as further confirmation of homogeneity.

27

The horizontal straight lines obtained in plotting n/c versus c in theosmometry experiments are consistent with solution ideality over the rangeof concentrations employed (no dependence of n/c on c). This agrees withthe criteria of linear Inc versus r2 plots and equal molecular weights atseveral loading concentrations it the equilibrium experiments.

The mean value of molecular weight for the toxin protein in Veronalobtained by osmometry in two experiments is 26,800+-300 daltons. Thisvalue is lower than the mean values obtained for the toxin by the methodsof sedimentation and diffusion and sedimentation equilibrium. The lowermolecular weight of the toxin by osmometry might be explained either onthe basis of experimental error or on the basis of heterogeneity of theprotein. However, the rather considerable heterogeneity required for thediscrepancy between M. and Mn ought to be detectable by sedimentationand sedimentation equilibrium analysis. Therefore, it seems more reason-able to attribute the difference to experimental error.

It is not possible to estimate, from our experimental data, the effectsof ion binding on the molecular weight of the protein. It has been deter-mined that about six phosphate groups are bound by a molecule of the toxinprotein (exhaustively dialyzed against distilled water, from 0.03 M phos-phate buffer at pH 6.5 to 6.7).* This would add about 1,000 daltons tothe molecular weight of the toxin protein, under the conditions of ourexperiments. This would not contribute significantly to the molecularweight. Because the molecular weight of the protein in Veronal is thesame, within experimental error, as in phosphate, no significant differ-ence is expected in the binding of the barbiturate and the phosphateions. It is possible to test for significant binding by determining Sin KVl, for example, where little binding is expected.**

The average value for molecular weight of the toxin protein in Veronalbuffer, 31,000±_1,300 daltons, obtained by the sedimentation equilibriummethod, is considered to be the most reliable value we obtained. Thisis because the method does not involve as many measurements, and is thusnot subject to as many sources of inaccuracy, as the method of sediments-tion and diffusion a"' because it is not as sensitive as osmometry topossible low molecular weight components. Furthermore, the value inVeronal is for art isoelectric protein, obviating the necessity for anycorrections due to charge effects.

This value for the molecular weight of 31,000±1,300 daltons is con-sistent, within experimental error, with all our other determinations andwith the values obtained by Vagman at &l. and Bergdoll at &l. Moreover,it is consistent with the value of 29,600 obtained by viscometry of reducedand ilkylated toxin in 6 M 8uanidine.***

* Spero, L., 1968, personal conication.** Saroff, H.A., 1968, personal coinztcation.

* Spero, L., 1968, personal canznication.

28

Evidence for the homogeneity of the toxin protein used in this investi-Sation is the single, Gaussian schlieren diagrams obtained in both electro-phoresis and sedimentation analyses and the straight lines obtained in theInc versus r2 plots in the sedimentation equilibrium experiments. More-over, straight lines were obtained in these plots for different loadingconcentrations, two in the phosphate experiment and three in the Veronalexperiment, over the same range of fringe displacements. This latterevidence for homogeneity is discussed by Yphantis. 2 The agreement betweenthe molecular weight values obtained by sedimentation equilibrium and bysedimentation and diffusion is also evidence for homogeneity. We cannotdiscount the possibility of some molecular heterogeneity of the protein,in view of the discrepancy between the osmometry values and the valuesobtained by sedimentation methods. But we believe most of the discrepancyis due to experimental error because the large degree of heterogeneityotherwise required should have been detectable by the various experimentalmethods used.

29

LITERATURE CITED

1. Wagman, J.; Edwards, R.C.; Schantz, E.J. 1965. Molecular sizehomogeneity, and hydrodynamic properties of purified staphylococcalenterotoxin B. Biochemistry 4:1017-1023.

2. Bergdoll, M.S.; Chu, F.S.; Huang, I.-Y.; Rove, C.; Shih, T.1965. Staphylococcal enterotoxin B: III. The physicochemicalproperties and the N- and C-terminal amino acid sequences. Arch.Biochem. Biophys. 112:104-110.

3. Schantz, E.J.; Lauffer, M.A. 1962. Diffusion measurements in agargel. Biochemistry 1:658-663.

4. Schantz, E.J.; Roessler, W.G.; Wagman, J.; Spero, L.; Durnery, D.A.;Bergdoll, M.S. 1965. Purification of staphylococcal enterotoxin B.Biochemistry 4:1011-1016.

5. Schachman, H.K. 1957. Ultracentrifugation, diffusion, and viscometry,p. 52-56. In S.P. Colowick and N.O. Kaplan (ed.) Methods inenaymology, Vol. IV. Academic Press, Now York, N.Y.

6. Schumaker, V.N.; Schachman, H.K. 1957. Ultracentrifugal analysisof dilute solutions. Biochim. Biophys. Acts 23:628-639.

7. Pedersen, K.O. 1940. Calculation and recording of results for mono-and paucidisperse system, p. 273-287. In T. Svedberg and .O.Pedersen (ed.) The ultracentrifuge. Clarendon Press, Oxford, England.

8. Casssa, E.F.; Eisenberg, H. 1961. Partial specific volumes andrefractive index increments in wnlticomponent systems. J. Phys. Chem.65:427-433.

9. Kay, C.?.. 1960. The partial specific volume of muscle proteins.Biochim. Biophys. Acta 38:420-427.

10. Kraemr, E.O. 1940. The partial specific volume factor, p. 57-66.In T. Svedberg and K.O. Pedersen (ed.) The ultracentrifuge.Clarendon Press, OKford, Enl8and.

11. Yphantis, D.A. 1964. Equilibrium ultraciutrifugation of dilutesolutions. Biochemistry 3:297-317.

12. Schachnan, H.i. 1959. Chapter VI, Sedimentation equilibrium, p.201-214; Chapter VII, Interpretation ef sedimentation data, p. 215-247. 11 H.K. Schchmea, Ultracentrifugation in biochemistry.Academic Press, New York, N.Y.

/4

30

13. Pedersen, K.O. 1963. Sedimentation, diffusion and partial specificvolume, p. 211-213. In M. Florkin and E.H. Stota (ed.) Comprehensivebiochemistry, Vol. 7. Ilsevier Publishing Company, New York, N.Y.

14. Charlvood, P.A. 1957. Partial specific volumes of proteins inrelation to composition and environment. J. Auer. Chem. Soc. 79:

716-781.

15. Charlvood, P.A.; Ins, A. 1957. Effect of pH on the sedimentationof serum albumins and ovalbunin. Can. J. Chem. 35:99-101.

p-0• I~

Unclassified 33

DOCUIANT CONTROL DATA - R & D(S,..-wU, a..fteat_ of tieg.* of 064ftetO kad hm ,m..ewc fbe wen m* . OOM a.. _

OUIOINAYINO ACtIVITY (Ces9.mh 0~) k. mPONt SECURIV CLASIFICATION

DeprtentoftheAry Uclssiie

MCUPOB? TITLE

FURTER INVESTIGATIONS ON THE MOLECULAR WEIGHT OF STAPUTOOCCAL EKEROlTIN B

SCAUPNVoI No(Pha e o a" Ew*6iha es)

David (NI) TrkulaRobert L. Sine, Jr..dva1 d J. Schantz

E ORT AT T& TOTAL 0O. OP PAS 7 Ise. morP CP

September 1968 33 15f CONTMACY OO *NANT NO. S0. ONROMATOW9 NPONT NUalRM

6. PROJeCT mO. 13563603D164 Technical Manuscript 487

0.~~JkN~m PNOWMG (AMw *,. O O 6

to. ISTMOSUTNN STATUM"?

Qualified requesters may obtaits copies of this publication from DDC.Foreign announcement and dissemination of this publication by DDC is not authorized.Release or announcement to the public is not authorized.|I, 5UeIHbl45llMTNYANV N * ii. SPONt0N nIt.iTANV ACTIVITV

Department of the Army

Fort Detrick, Frederick, Maryland 21701ms, semvuagY

- The molecular weight of staphylococcal enterotozin 3 was determined independentlyand by methods not previously applied to the protein; the results agree well withthose obtained in two earlier studies. In order to check the previwus determinations,analyses were performd t 0.05 N sodium phosphate buffer, pH 6.8, and in 0.1 ionicstrength sodium Veronal buffer, p11 8.6, the isoelectric point of the protein. Thesedimentation and diffusion method yielded 33,000t1,400 daltons In Veronal buffer and31.800*1,000 dalitons in phosphate buffer. The Yphantis meniscue-depletion sedimenta-tion equilibrium method yielded average values of 31,000*1,300 dalton. in Veronal and30,400M2,000 dalton. in phosphate. Ossometry yielded at average value of 26,800*00daltoes in Varonal buffer.

14. Key Words

Staphylococcal enterotain 5olecular weight

SedimentationDiffuaionSodium phosphateSodium Veromalbuffere

W47 laaeifte