PARTIAL PURIFICATION AND PROPERTIES OF TWO PHOSPHOLIPASES OFBACILLUS CEREUS

MILTON W. SLEIN AND GERALD F. LOGAN, JR.U.S. Army Chemical Corps Biological Laboratories, Fort Detrick, Frederick, Maryland

Received for publication 30 August 1962

ABSTRACT

SLEIN, MILTON W. (U.S. Army ChemicalCorps Biological Laboratories, Fort Detrick,Frederick, Md.) AND GERALD F. LOGAN, JR.Partial purification and properties of two phos-pholipases of Bacillus cereus. J. Bacteriol.85:369-381. 1963.-Culture filtrates of Bacilluscereus contain a phosphatasemia factor (PF)that markedly increases blood alkaline Vhos-phatase after intravenous injection into animals,and that releases alkaline phosphatase fromepiphyseal bone slices in vitro. Fractionation ofculture filtrates of B. cereus with N,N'-diethyl-aminoethyl cellulose results in the separation oftwo phospholipases, one that has PF activity andone that inhibits PF activity in vitro. Growth ofshaken cultures favors accumulation of theinhibitor, whereas static cultures yield more PF.Lethality for mice and hemolysin activity do notappear to be associated with the phospholipasethat inhibits PF. The relationship of the lethaland hemolysin factors to the phospholipase thatproduces phosphatasemia is not clear. Theeffects of heat, trypsin, lecithin, and antiserumon the phospholipases are reported. The intra-venous injection of relatively large amounts of thepurified PF resulted in the depletion of bonealkaline phosphatase.

Although the production of phosphatasemiafactor (PF) in culture filtrates of certain bacilliappeared to be related to that of phospholipaseC (lecithinase), previously reported evidencesupported the conclusion that the two activitieswere not identical (Slein and Logan, 1962a).[Phospholipase C is used to designate the enzymethat splits lecithin into a diglyceride and acid-soluble phosphorylcholine, according to thenomenclature given by Hayaishi (1955). Un-fortunately, the nomenclature of phospholipaseshas been variable, and the same enzyme has alsobeen called phospholipase D (Deuel, 1955) and

lipophosphodiesterase I (Schmidt and Laskowski,1961).] More recent findings show that thisinterpretation probably was confused by thepresence of more than one phospholipase, aswell as an inhibitor of PF activity (Slein andLogan, 1962b). Results reported in this paper areconcerned with the fractionation of culturefiltrates of Bacillus cereus, some properties of thefractions, and further evidence that boneepiphyses contribute to the phosphatasemia thatresults from the intravenous injection of PF intorabbits.

MIATERIALS AND 1\IETHODS

Culture filtrates and other preparations. Strains6464 and 7004 of B. cereus were originallyobtained from the American Type CultureCollection. Cell-free filtrates were prepared aftercultures had been grown statically for 24 hr at37 C, as described previously but without theaddition of charcoal (Slein and Logan, 1962a).Preparations were concentrated by saturatingthe culture filtrates with (NH4)2SO4 at 5 C. Theprecipitate was dissolved with distilled water,and was dialyzed against 0.01 M tris(hydroxy-methyl)aminomethane (tris), pH 7.5. In mostcases, the clear, slightly amber, concentratedmaterial was treated with protamine sulfate toremove nucleic acids and much of the color. Anexcess of protamine was avoided by the carefuladdition of a 2% solution of protamine sulfate,between centrifugations, to remove the bulk ofthe precipitate as it formed at room temperature(23 C). When the precipitation is carried out atabout 0 C, the protamine itself forms a copiousprecipitate as it is added to the cold solution.The precipitate of undiluted cooled protaminewill redissolve when stirred, whereas the complexof protamine with nucleic acid will not; thus, itis possible to avoid the addition of excess prota-mine, even in the cold, if this is kept in mind.The preparations were dialyzed again aftertreatment with protamine.

369

on June 25, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

SLEIN AND LOGAN

Phospholipase C assay. The assay procedurewas essentially as described by Costlow (1958).Soybean lecithin was obtained from NutritionalBiochemicals Corp., Cleveland, Ohio, and fromMann Research Laboratories, Inc. The reactionmixture (3 ml) contained 100 mg of lecithin,which had been emulsified in a Raytheon sonicoscillator (Hayaishi, 1955). The mixture, bufferedwith barbital at pH 7.6, was incubated in thepresence of 0.01 M CaCl2 for 30 min at 37 C;the reaction was stopped by the addition of 1ml of 20% trichloroacetic acid. After filtration,acid-soluble phosphorus in 0.1 ml of the filtratewas determined in a final volume of 3.3 ml at660 m,u with a 1-cm light path. One unit ofenzyme activity is defined as an increase inabsorbancy of 0.001. To conserve enzyme andsubstrate, half of the normal reagent volumeswere sometimes used, but the specific activitywas calculated for a volume of 3 ml. Propor-tionality was obtained with as much as 300units of enzyme per 3 ml of reaction mixture.Bone slice test for PF. The procedure of the

test was basically that described previously(Slein and Logan, 1962a). Although we used onlythe proximal tibial and distal femoral epiphysesof Dutch rabbits weighing about 1 kg, otherepiphyseal bone of young animals probablycould be used. The untrimmed bones may bestored packed in ice for a few days withoutapparent loss of ability to respond to PF. Justbefore slicing, the joint was freed from adherentmuscle, cartilage, and other tissues. The bonewas held firmly in the clamp of a Spencer model860 sliding microtome. The blade was kept wetwith Ringer phosphate solution, and slices werecut approximately 120 g thick. They wereimmediately placed in small beakers of coldRinger phosphate solution in an ice bath. Usually,49 slices were obtained from the knee-jointepiphyses. The slices were distributed seriallyamong seven beakers, so that each containedseven slices matched with the others. The sliceswere transferred to manometer flasks, eachcontaining 3 ml of cold Ringer phosphate solution,and the flasks were shaken at about 100 cycle/min for 15 min at 37 C. Initial samples, beforeincubation at 37 C, were not taken, because thealkaline phosphatase activity in such sampleswas consistently low. To minimize the presenceof floating particles after centrifugation, bits ofcotton wool were wound about the tapered

ground tips of I -ml pipettes used for taking 0.6-mlsamples from the flasks. The samples wereplaced in 10 X 100-mm tubes in an ice bath, and0.3 ml, or less, of PF or other test solution wasadded to each flask. The flasks were shaken foranother 15 min and were sampled again. Thesamples were chilled, centrifuged, diluted, andtested for alkaline phosphatase, as reportedpreviously (Slein and Logan, 1962a).

Since the slice test for PF is an indirect one, itis important to establish that the responsemeasured (alkaline phosphatase release) isproportional to the concentration of factoradded, and that the phosphatase activity is notdirectly inhibited or stimulated after beingreleased from the slices. We have establishedthat PF does not merely activate soluble bonealkaline phosphatase. It has been determinedthat the effect of inhibitor fractions is not due tothe inactivation of bone phosphatase, butrather to the inhibition of the PF-induced releaseof phosphatase from the slices. Bone alkalinephosphatase activity is directly proportional tothe amount of enzyme, over a wide range ofactivity. Linear proportionality was found toexist from 0 to an absorbancy of at least 1.300at 395 mA after incubation of the enzyme for 15min at 37 C, which was the time used routinelyfor the measurement of bone alkaline phos-phatase. However, measurement of the amountof alkaline phosphatase released from boneslices in 15 min at 37 C is limited to an absorbancyrange of about 0 to 0.400. The limit is determinedby the number of slices used in each flask. Withseven slices of epiphyseal bone, having a blottedmoist weight of 50 to 70 mg, the release of alkalinephosphatase, in the limited range just mentioned,is approximately proportional to the amount ofPF added. The phosphatase released by PFmust be corrected for the spontaneous leakagefrom the slices, which usually amounts to anabsorbancy of 0.080 to 0.120. One unit of PF isdefined as the amount that releases sufficientphosphatase from the bone slices in 15 min toproduce an absorbancy of 0.100 in the standardphosphatase test.

Lethality tests. The lethal effect of a preparationwas determined by injecting the material,diluted with 0.9% NaCl or Ringer phosphatesolution, into the tail veins of Swiss mice weigh-ing 15 to 20 g. The smallest amount of solutionthat caused death within 20 min was defined as a

370 J. BACTE3RIOL.

on June 25, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

PHOSPHOLIPASES OF B. CEREUS

MLD. Mice that survived for more than 20min usually did not die at all from the injectedmaterial. A MLD usually caused death within 5to 8 min.

Hemolysins. Hemolytic activity was detectedon sheep-blood agar plates, prepared as describedby Skerman (1959). Samples (0.1 ml) wereallowed to spread in a circular area on the surfaceof the agar. After 1 or 2 hr at 25 C, the plates wereincubated at 37 C, and were examined at intervalsfor comparison of the rates at which hemolysisoccurred.Paper electrophoresis. Electrophoresis was car-

ried out with 0.05 M barbital buffer (pH 8.6).The enzyme solutions were spotted on threestrips of Whatman 3MM paper (3.5 X 30 cm,excluding the submerged portions). The chamberwas cooled with circulating water at 7 C, and500 to 600 v were applied for about 4 hr. A thinsection was cut from the edge of each paperstrip, to be dried and stained with bromophenolblue for locating proteins (Kunkel and Tiselius,1951). The rest of the strips were kept at 5 Cbetween sheets of polyethylene, to minimizedrying while the control sections were beingstained. The unstained strips were then cut into0.5-in. sections at various distances from theorigin, and each section was cut into smallerpieces. These were placed in a Virtis microhomogenizer cup with 1 ml of cold distilled water,and were homogenized for about 2 min. Thepaper pulp was centrifuged, and residual liquidwas squeezed from the precipitate with a glassrod. Electrophoresis was tested at other pHvalues as low as pH 4.4, but the results only atpH 8.6 will be discussed.

Fractionation on N, N'-diethylaminoethyl(DEAE) cellulose. DEAE cellulose was obtainedfrom Eastman Kodak Co., Rochester, N.Y. Aslurry of the material in 0.25 N NaOH was usedfor preparing a column about 1.2 X 17 cm. Thecolumn was washed with NaOH and distilledwater, and was equilibrated with tris buffer (pH8.6). When elution was to be done with step-wiseincreases in buffer concentration, 0.05 M tris wasused for equilibration, and the concentration oftris in the enzyme solution was adjusted toabout 0.05 M before placing it on the column.WVhen continuous gradient elution was used, thecolumn was equilibrated with 0.01 M tris, and theenzyme solution was dialyzed against 0.01 M trisbefore it was placed on the column. Elution was

carried out at 23 C with a flow rate of about 2ml/min. Fractions of 5 or 10 ml were collected.With continuous gradient elution, the concen-tration of tris was estimated at intervals bytaking samples from the mixing chamber for themeasurement of absorbancy at 230 m,u. Afterelution from the column, the absorbancy offractions at 280 m,u was measured.

WN'hen it was necessary to concentrate fractions,they were placed in cellophane dialysis tubing,buried in Carbowax 20-M, at 5 C until the desireddecrease in volume had occurred. The tubing wasrinsed off with distilled water, and the concen-trated solution was dialyzed at 5 C against 0.01M tris (pH 7.5). Carbowax, a polyethylene glycol,was obtained from the Union Carbide Corp.,New York, N.Y.

Other procedures were described previously(Slein and Logan, 1962a). In some cases, proteinwas estimated by the method of Waddell (1956).

RESULTS

Preliminary attempts at fractionation. It wasknown that PF activity could be concentratedby saturating culture filtrates with (NH4)2S04.The addition of protamine removed much of thepigment from the solutions, as well as nucleicacids, and the specific activities of PF and phos-pholipase were increased three- to fourfold.Further attempts at purification, by means of(NH4)2SO4 or isoelectric precipitation, were notvery promising.

Electrophoresis. The first direct evidence forthe existence of more than one phospholipase inconcentrated partially purified preparations fromB. cereus 6464 was found when fractions wereeluted from paper after electrophoresis. Althoughthe recovery of protein was fair, rather poorrecovery of PF and phospholipase C activitieswas obtained, especially at pH 8.6. However,sufficient material was recovered for a comparisonof the activities from various regions of thefilter papers. In one test (Table 1), at least twophospholipases were present. Most of the PFactivity was associated with a phospholipase thatmoved toward the anode, whereas the phos-pholipase with the highest specific activitymigrated toward the cathode, and was associatedwith much less PF activity.

Since at least two phospholipase C moietieswere separated by electrophore6is at pH 8.6, we

VOL. 85, 1963 371

on June 25, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

SLEIN AND LOGAN

TABLE 1. Effect of electrophoresis at pH 8.6 on thephospholipase and phosphatasemia factor (PF)activities of protein concentrated from a culture

filtrate of Bacillus cereus 6464

Position on filter Phospholipase C PF activitypaper activity

units/Ag units/lgAnode 2 11.2 13.9Anode 1 18.4 26.6Origin 7.7 7.8Cathode 43.8 9.3

decided to try chromatography at the same pHon a DEAE cellulose anion-exchange column.

Fractionation with DEAE cellulose. Residualcolor that was present in solutions after treatmentwith (NHI)sSO4 and protamine was removed,since the pigment remained stationary at the topof the DEAE cellulose column during elutionwith tris (pH 8.6). All fractions were clear andcolorless. When elution was step-wise, startingwith 0.05 M buffer, material with the highestabsorbancy at 280 m,u came through the columnwithout being adsorbed. This material probablycorresponded to that which moved toward thecathode during electrophoresis. In one procedure,55 fractions of about 10 ml each were obtainedwith six concentrations of tris ranging from 0.05to 1.0 M. Various activities were measured in 15of the fractions after dialysis against 0.01 M tris(pH 7.5). Data are given in Table 2 for the sixprincipal fractions that had peak absorbancies at280 m,u with the six concentrations of tris used.Fraction 2, which was not adsorbed to thecolumn, had the highest phospholipase C activity,but it had no PF activity in vitro or in vivo.Fraction 23, which was eluted by 0.4 M tris,contained phospholipase C with about 40% ofthe specific activity of that in fraction 2, but ithad the highest PF activity. The separation ofat least two phospholipase activities is obviousfrom these data.The fact that hemolysin activity was not

detected in fraction 2, but was in fraction 13,indicated that it was not associated with thephospholipase C activity of fraction 2. Hemolysinwas also present in four of the other five fractionsincluded in Table 2. Since the hemolysin activityin fraction 13 appeared to be stronger than thatin fraction 23, it seems likely that the phos-pholipase C activity of fraction 23 was not

responsible for hemolysin activity. These resultsare in agreement with the conclusions of Otto-lenghi, Gollub, and Ulin (1961), that hemolysinactivity could be separated from phospholipaseactivity.The original preparation used for fractionation

in Table 2 contained approximately one MLDper 10 ,ug of protein. Although the phospholipaseC activity was increased about two- to fourfoldin fractions 23 and 2, and the PF activity wasincreased about fourfold in fraction 23, no lethalfactor activity was detected in these, nor in anyof the other 13 fractions tested. Thus, it seemsunlikely that mouse lethal activity is identicalwith either of the phospholipases of fractions 2and 23. The amounts of protein available forlethality tests with the other fractions (about 6to 18 yg per mouse) were sufficient only toindicate that lethal factor had not been highlyconcentrated in any of the fractions tested.

It was discovered that fraction 2 not only didnot have PF activity, but that it was stronglyinhibitory to added PF in the bone-slice test(Table 3). Approximately 4 ,ug of fraction 2protein had no detectable PF activity. Only 0.5Mg was required to inhibit PF activity about 70%,whereas 1 ,g inhibited the PF completely.Heating the inhibitor for 5 min at 100 C com-pletely inactivated it. Of the 15 fractions tested,1, 10, 13, and 20 also had no PF activity butonly fraction 2, which had the highest specific

TABLE 2. Fractionation of a partially purifiedphospholipase of Bacillus cereus 6464 by elutionfrom a DEAE cellulose column with step-wise

increases of tris buffer, pH 8.6

Molar- Phos- PF activityity of Fraction no. pholi- PF_activity_Hemo-tris pase C lysinactivity In vitro In vivo*

units/Mg units/Ag units/,ug

0.05 2 124.0 0 0 -

0.2 13 0.8 0 0 +0.4 23 49.5 173.0 719 +0.6 33 7.3 10.8 67 +0.8 44 5.4 8.3 30 +1.0 53 0.7 1.0 0 -

Original 28.0 45.0 149material

* Increase in blood alkaline phosphatase ac-tivity 1 hr after the intravenous injection of 15mg of protein into a rabbit.

372 J . BACTERIOL .

on June 25, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

PHOSPHOLIPASES OF B. CEREUS

phospholipase C activity, was found to inhibitPF activity.

Figure 1 gives the elution pattern obtainedwith a buffer concentration gradient thatapproached linearity; each fraction containedabout 5 ml. Only the PF and phospholipase Cactivities are presented, since the 280 m,u absor-bancy pattern of the fractions did not reveal the

TABLE 3. Inhibition of phosphatasemia factor (PF)activity by the protein infraction 2 of Table 2

AlkalineAdditions to bone slices phosphatase

released

Fraction 2 (4.0,g) 0PF (0.2 jg) 312PF (0.2,g) + fraction 2 (0.5,g) 90PF (0.2 jAg) + fraction 2 (1.0 MAg) 0PF (0.2,Mg) + heated fraction 2 (2.0,ug) 332

70*F

600k

c

1._0.

CODo 400A._;0.

#A 30C

c

_I

_F

2001.

100-

positions of the active fractions. Since the unitsof activity are arbitrary and not directly related,the similarity of the specific activities of PF andphospholipase C in fractions 52 to 80 is coinci-dental. However, the similar distribution in thesefractions indicates the identity of the twoactivities. The concentration of Ca- (0.01 M)used for the measurement of phospholipase Cactivities of the fractions in Fig. 1 was twice thatused previously. Only 0.005 M Ca+ was used forphospholipase C activity determinations withanother lecithin preparation at the time that thedata for Table 2 were obtained; this resulted inrelatively lower values for phospholipase.As with the step-wise gradient, the fractions

eluted by the lower concentrations of buffercontained phospholipase activity that wasassociated with an inhibitor for PF. The PF andphospholipase C present in fraction 62 had thehighest activities we have so far obtained. A PF

Ias

06

O4

0

._

I-

0

0I

UI I I I I I10 20 30 40 50 60 70

Fraction Number80 90 loo 110

FIG. 1. Distribution of phosphatasemia factor (PF) and phospholipase C activities in fractions obtainedby the elution of a partially purified preparation of PF of Bacillus cereus 6464 from a DEAE cellulosecolumn with increasing concentrations of tris buffer (pH 8.6). X, molarity of tris; 0, phospholipase activity;*, PF activity measured with bone slices.

_____

373VOL. 85, 1963

8-

on June 25, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

SLEIN AND LOGAN

activity of 750 units/,ug in the bone-slice testmeans that 0.005 ,ug of protein was sufficient toresult in an absorbancy of 0.375, which is a verysignificant value at the upper limits for propor-tionality with the method.To test for lethal factor and hemolysin activity

in groups of fractions obtained by the lineargradient elution, several fractions were combined,concentrated, and dialyzed (Table 4). Sample Acontained fractions 13 to 20; B contained 30, 34,35, 36, 43, 48, 49, and 50; C contained 56 to 68;and D contained 80 to 114. Although not allfractions were included, the four samples repre-sented various areas of the gradient and con-tained material from 64 of the 114 fractions. The64 fractions represented about 20% of the totalprotein that was placed on the column for elution.Only 8 Ag of protein of the original material wasrequired to kill mice, but mice survived theinjection of 20 to 51 Mug of protein of samples A,B, and C. A combination of 72 Mg of proteinprepared from the same three samples also didnot kill mice. However, approximately 15 Mg ofprotein in sample D were lethal. These resultsindicate that PF and its associated phospholipaseC are not responsible for the lethality, sincesample D contained 10 to 20 times less PF andphospholipase activities than did sample C.Although twice as much protein in sample D

TABLE 4. Activities of several fractions obtained bythe continuous gradient elution of a partiallypurified Bacillus cereus 6464 phospholipase

from DEAE cellulose

Frcinn.*PF act~ivity Phospho- MD Hemo-Fraction no. in vitro llcptasvtC MLD lHysinotactivity lsn

units/jig units/,ig ,ug

Original 66 49 8 +material

A 0 28.1 >51t -B 1.6 4.4 >29t -C 430 380 >201 -D 20.6 36.5 15 +

* Several fractions obtained by the elutionprocedure of Fig. 1 were combined and concen-trated as described in the text.

t The protein concentrations were adjusted toabout 10,ug/ml before testing on blood agar plates.

t These values merely represent the maximalamount of protein injected. The mice did not diefrom these doses.

was required for lethality as in the originalmaterial, it is possible that the lethal factor wasconcentrated in only a few of the 35 fractionsthat were combined in the preparation of sampleD.Hemolysin activity was tested in the com-

bined samples at a concentration of about 10 MAg ofprotein/ml. Significant activity was found onlyin sample D, and in the original material beforefractionation. Thus, the semiquantitative testindicated that hemolysin activity was notassociated with the PF or phospholipase Cactivities of samples A, B, and C.Although the results reported above indicate

that lethal and hemolysin activities are notassociated with either phospholipase, in anotherstep-wise fractionation procedure the lethal andhemolysin factors accompanied PF in tenfractions that were recombined and concentratedafter being eluted with 0.4 M tris. However, inagreement with our other tests, no lethality orhemolysin factor was detected in the fractionsthat contained the phospholipase that inhibitedPF activity.

Production of PF inhibitor. It appeared that,under our conditions for growth, shaken culturesproduced less PF and phospholipase than thosegrown statically (Slein and Logan, 1962a). Sincewe found that filtrates prepared from staticcultures contained both PF and an inhibitorof PF, it was decided to determine whethershaken cultures contained more of the inhibitor.Filtrates were prepared after the growth of B.cereus 6464 and 7004 in media containing phos-phate-buffered Casamino Acids or nutrientbroth (Difco) plus 0.3% yeast extract, in staticor shaken flasks (100 cycle/min). Good growthwas obtained in each case. After filtration, thesolutions were dialyzed for 23 hr against 0.01M tris (pH 7.5) at 5 C. The solutions were testedfor PF and phospholipase C activities, compari-sons being based on volumes rather than pro-tein, since nondialyzable proteins or peptidesin the media would interfere (Table 5). With oneexception (compare samples 1 and 5), phospho-lipase activity was higher in static than in shakencultures. PF activity was higher in static cul-tures in every case; in fact, no PF was detectedin shaken cultures except for relatively lowactivity in sample 8. It appeared that the phos-phate-buffered Casamino Acids medium wasbetter for the production of PF than was the

374 J. IBACTERIOL.

on June 25, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

PHOSPHOLIPASES OF B. CEREUS

TABLE 5. Effect of medium and aeration on thephospholipase, phosphatasemia factor (PF), andPF-inhibitor activities of culture filtrates of

Bacillus cereus 7004 and 6464

Culture Strain

Static 7004

6464

Shaken| 7004

6464

A

EuzU2

1

2

3

4

5

6

7

8

Medium

Nutrientbroth

CasaminoAcids

Nutrientbroth

CasaminoAcids

Nutrientbroth

CasaminoAcids

Nutrientbroth

CasaminoAcids

Phos-pholi-paseC

units/0.1 ml

38

134

59

57

74

60

5

21

PF invitro

units/0.1 ml

10

120

24

100

0

0

0

5

* Not tested because of the relatively highintrinsic PF activity.

t The over-all inhibition calculated here in-cluded the PF activity present in the inhibitoryfiltrate itself.

nutrient broth-yeast extract medium. However,the results were probably complicated by thepresence of varying amounts of PF inhibitor.The filtrates from shaken cultures all containedPF inhibitor. It was not possible to test samples1 to 4 for inhibitor activity, because of theirrelatively high PF activities. However, thefractionation with DEAE cellulose, describedabove, demonstrates that inhibitor may bepresent even in filtrates from static cultures thathave high PF activities. Dialyzed, uninoculatedmedia did not have demonstrable PF inhibitor.Hemolysin activity was found to be greater infiltrates from static than from shaken culturesof either strain of B. cereus tested.

It was reported that filtrates of cultures ofB. anthracis and certain other bacilli grownstatically, or B. cereus grown statically in thepresence of 0.25 M ethanol, did not produce muchPF (Slein and Logan, 1962a). After the discoveryof the PF inhibitor, several of the filtrates that

had relatively little PF activity were tested,but no PF-inhibitor was detected. It may beconcluded that the strains merely did not pro-duce much PF, and that ethanol inhibited theproduction of PF by a strain of B. cereus thatnormally produced it in relatively large amounts.The crude lyophilized lecithinase preparation

of Costlow (1958), which had been reported tocontain no PF activity (Slein and Logan, 1962a),has been found to be a fairly potent PF inhibitor.The protein (15 jg) inhibited PF activity about70% in the bone-slice test. The absence of PFactivity had been ascribed to the loss of PF during5 years of storage at room temperature. Sincethe material was prepared from B. cereus 7004that was grown in shaken nutrient broth, itseems probable that it never did contain anydemonstrable PF activity.A larger amount of inhibitor was prepared by

growing B. cereus 7004 in shaken cultures andconcentrating the material by saturation with(NH4)2S04 at 5 C. The precipitate was dissolved,dialyzed, and fractionated on a DEAE cellulosecolumn, with gradient steps of 0.05, 0.2, and0.4 M tris (pH 8.6). Fractions in each of the threeprincipal absorption peaks (280 m,u) were com-bined, concentrated, dialyzed, and tested forPF, phospholipase C, and PF-inhibitor activities.None of the three fractions contained PF ac-tivity, although experience with preparationsfrom static cultures led us to expect that PFactivity should be present in the fractions elutedby 0.4 M tris. All three peaks contained materialwith inhibitor activity, but most of the inhibitorwas present in the nonadsorbed fractions com-prising the first peak. Phospholipase C activitywas also highest in the nonadsorbed fractions.Thus, it appears that shaken cultures of B.cereus may be prepared that have phospholipaseC activity devoid of PF activity.

Neither the original material from the, shakencultures of B. cereus 7004, nor any of the threefractions tested, had any significant hemolysinactivity. No mice were killed by the injectionof as much as 75 Mig of protein in the (NH4)2S04precipitate obtained from the shaken cultures.When the purified inhibitor fraction was

mixed with PF in a ratio sufficient to prevent theactivity of PF with bone slices, no inhibition ofphosphatasemia occurred after intravenous in-jection of the mixture into a rabbit. This indicatesthat any complex between the two substances

VOL. 85, 1963 375

on June 25, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

SLEIN AND LOGAN

may be dissociated or destroyed in vivo, or thatthe concentration of inhibitor may be too low tocompete effectively with the PF by the time thematerials reach the foci of alkaline phosphataserelease from the tissues.The degree of inhibition of PF activity pro-

duced by an amount of inhibitor sufficient to give50% inhibition in the bone-slice test was notaffected by incubation of PF with the inhibitorfor various times up to 30 min at 37 C prior totesting. This indicated that the inhibitor wasnot progressively destroying PF activity, butrather that it was complexing with it or merelycompeting with it for a substrate in the cellmembranes of the bone slices. The competitionhypothesis is favored by the results of tests withlecithin, to be described below.

Stability of components of B. cereus culturefiltrates. A crude (NH4)2SO4 precipitate, dialyzedagainst 0.02 M phosphate buffer (pH 7.5), andcontaining about 2.7 mg of protein/ml, was foundto have rather stable PF and phospholipase Cactivities for more than 1 year at about -15 C.PF activity was stable for at least several monthsat -15 C when the crude preparation wasdiluted with Ringer phosphate to a protein con-centration of 0.1 mg/ml. However, stability waspoor when only 4 ,ug of protein/ml was presentduring storage. A protamine-treated crude prepa-ration showed no significant loss of PF or phos-pholipase C activities, during incubation ofa solution containing 0.1 mg of protein/ml for 30min at 37 C, and storage overnight at 5 C. Thesame stability was found in 0.05 M acetate or trisbuffers from pH 5 to 8.5. The crude preparationsare stable during lyophilization. Even afterstorage for more than 5 years at room tempera-ture, the lyophilized inhibitor preparation ob-tained from R. D. Costlow still had good phos-pholipase C activity. A purified inhibitor fractionwas found to be stable, when incubated at a con-centration of 20 ,ug of protein/ml of 0.005 M tris(pH 7.5) for 30 min at 37 C before testing forPF inhibition and phospholipase C activities.The activities also were not significantly de-creased when a solution containing 60 Ag ofprotein/ml of 0.01 M tris (pH 7.5) was frozen andthawed ten times within 3 hr.

Because of the minute amounts of proteinrequired in the bone-slice test for PF activity,it was necessary to determine the stability ofPF at high dilutions. A protamine-treated crude

material was stable for at least 4 hr at 0 C, at aconcentration of 0.5 Ag of protein/ml in Ringerphosphate solution. Fraction 65, from the DEAEcellulose column of Fig. 1, was used for studyingthe stability at a concentration of only 0.04Ag of protein/ml. The preparation was dilutedwith cold distilled water, Ringer phosphate,0.01 M tris (pH 7.5), an emulsion of 0.1 mg ofsoybean lecithin/ml of water, or human serumalbumin in water (0.1 mg/ml). The dilute solu-tions were stored at 0 to 5 C for 1 and 24 hrbefore testing. Control dilutions were made withRinger phosphate or albumin, in tris buffer,immediately before samples were added to theflasks with bone slices. PF activity was remark-ably stable for 1 hr in any of the diluents, andeven after 24 hr only about 50% loss occurredin distilled water. Activity was completely stablefor 24 hr when the PF was diluted with albuminas a protective protein, and only about 20% lossoccurred during storage in the tris or lecithinsolutions. Ringer phosphate appeared to bebetween water and tris in the stability tests.Hemolysin and lethal activities were markedly

decreased when solutions were heated for 5 minat 100 C. The stability of PF and phospholipaseC activities to heat was also tested, in an attemptto determine whether the two might be iden-tical. A relatively crude preparation, contain-ing 0.4 mg of protein/ml, was stable duringdialysis for 23 hr against cold distilled water,0.02 M tris, or phosphate buffer (pH 7.5). Afterdialysis, samples were heated for 5 min in a boil-ing-water bath. The material that had beendialyzed against tris or water lost almost allPF and phospholipase C activities. However,the presence of phosphate protected about 20 to30% of the activities during the heat treatment.These results indicated that phosphate might becombining with the active centers of PF andphospholipase, in the manner by which a naturalsubstrate protects an enzyme from variousinactivating agents. The effect of a phospholipasesubstrate, soybean lecithin, as a protective agentagainst heat inactivation was tested with purifiedinhibitor and PF fractions. Samples, containing18 ,g of inhibitor protein or 5 Mig of PF proteini 17 mg of emulsified lecithin/ml of water,were heated for 5 min at 60, 80, or 100 C, chilledin ice, and tested for activities along with un-heated control samples. The effect of lecithinon the inhibitor activity could not be tested with

376 J. BACTERIOL.

on June 25, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

PHOSPHOLIPASES OF B. CEREUS

bone slices, since lecithin itself was found toinhibit PF activity (see below). However, thephospholipase C activity of the inhibitor fractionwas not significantly protected by lecithin frominactivation at any of the temperatures tested.The amount of inactivation in the absence oflecithin was about the same (25, 55, and 90%O at60, 80, and 100 C, respectively) for either thePF-inhibitor or phospholipase C activity. Thisindicated that the two activities were identical.With the purified PF fraction, inactivation of

both PF and phospholipase C activities was

about 85 to 100% at 80 to 100 C in the absenceof lecithin, but this was reduced to about 30 to65% in the presence of lecithin. It was possibleto test PF activity after treatment with lecithin,since the PF activity was so great that the solu-tions could be diluted sufficiently to avoid inhi-bition by lecithin during the bone-slice tests.As with heat inactivation, lecithin was found

to inhibit the destruction by trypsin of purifiedPP and its associated phospholipase C. PF, con-

taining 22,ug of protein/ml of 0.01 M tris (pH 7.5),was incubated for 30 min at 37 C in the presence

and absence of crystalline trypsin (about 150,ug/ml), with or without lecithin (9 mg/ml). Thesolutions were chilled in ice and tested for PFand phospholipase C activities. Both were almostcompletely destroyed by trypsin, but only 10to 20% inactivation by trypsin occurred in thepresence of lecithin. Trypsin, in the concentra-tions present in the bone-slice test, did not affectbone phosphatase activity. Lecithin also hadno direct effect on PF or phosphatase activitiesin the concentrations used for the bone-slice test.Thus, lecithin protected both PF and phospholi-pase C activities in the purified PF fraction inthe manner in which a substrate may protectits enzyme from inactivation by heat or trypsin(Trayser and Colowick, 1961). This is evidencethat the PF and phospholipase C activities are

identical.Purified inhibitor, unlike PF, was relatively

resistant to attack by trypsin, even in the absenceof added lecithin. The fact that both PF-inhibi-tion and phospholipase C activities of the inhibi-tor fraction were resistant to trypsin supportsthe hypothesis that the activities are identical.

Effect of antiserum on purified fractions. Itwas reported that antisera, prepared with crudeB. anthracis lecithinase, inactivated the PFbut not the phospholipase of a B. cereus culture

filtrate (Slein and Logan, 1962a). One of the sameantisera (AS9) was used in similar tests withpurified inhibitor and PF fractions (samples Aand C of Table 4). Approximately 31 ug of pro-tein of inhibitor sample A, or 3 jig of PF sampleC, were incubated in 0.9% NaCl with 0.15 ml ofnormal rabbit serum or antiserum, in a volumeof 1.05 ml, for 30 min at 37 C. The samples werecentrifuged, and the supernatant solutions weretested for inhibitor, PF, and phospholipaseactivities (Table 6). The PF and phospholipaseactivities of the PF fraction were markedlydecreased by incubation with the antiserum.However, the inhibitor and phospholipase ac-tivities of the inhibitor fraction were not signifi-cantly reduced by treatment with the antiserum.The approximately 20% greater inhibitor ac-tivity found in the presence of antiserum wasdue to the further inactivation of PF by the0.007 ml of antiserum added with the inhibitorsample to the bone slice test system. Theseimmunochemical reactions were the basis of theprevious results that led to the erroneous con-clusion that PF and phospholipase were notidentical (Slein and Logan, 1962a). The crudeculture filtrate probably contained much morephospholipase that did not react with the anti-serum than the reactive type associated with PF.

Effect of lecithin on PF and phospholipase Cactivities. The purified soybean lecithin originallyused in our studies was an old preparation ob-

TABLE 6. Effect of antiserum on the phospholipase,PF-inhibitor, and PF activities of purifiedfractions of a culture filtrate of Bacillus cereus

6464

ActivityTreated with

Phospho- PF PEd lipase inhibitort

C Normal serum 221 346Antiserum 34 0

A Normal serum 157 - 99 (340)Antiserum 157 - 25 (340)

* The samples were the combined, concen-trated, and dialyzed preparations of Table 4.Sample A contained the inhibitor fractions 13 to20; sample C consisted of the PF fractions 56 to 68.

t The values represent alkaline phosphatasereleased from bone slices. The values in paren-theses were obtained with control samples towhich no inhibitor was added.

377VOI, . 85, 1963

on June 25, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

SLEIN AND LOGAN

tained from Nutritional Biochemicals Corp.More recent lots of lecithin contained too muchacid-soluble phosphorus for the accurate measure-ment of phospholipase C activity. Similar lecithinobtained from Mann's Research Laboratories hasbeen satisfactory, so far as the phosphorus con-tent is concerned. However, it was found that thelecithin from Mann's Research Laboratoriesinhibited the phospholipase activity of a purifiedPF-inhibitor fraction about 40%, as comparedwith the older Nutritional Biochemicals Corp.preparation. On the other hand, a purified PFfraction gave the same phospholipase activitywith either lecithin. These results are a furtherdemonstration of a difference between the twotypes of phospholipase C activity, which may beseparated on a DEAE cellulose column. Thereason for the difference in phospholipase ac-tivities obtained with the PF-inhibitor fractionis not known. An inhibitory substance may havebeen present in the lecithin from Mann's Re-Research Laboratories, or perhaps a less favorablecharge distribution existed in the complex ofphosphatides present in one lecithin as comparedwith the other (Dawson and Bangham, 1961).As was mentioned in the section on stability,

lecithin was found to inhibit PF activity inbone-slice tests. The effect of lecithin was notdue to any inhibition of bone alkaline phospha-tase. It also was not due to the destruction of PF,since activity was restored when a mixture was di-luted before testing for PF activity. The effect oflecithin appeared to result from a competition be-tween the emulsified lecithin and a natural sub-strate in the bone cell membranes for the activesite of phospholipase C in the PF fraction. Al-though the application of solution kinetics is notvalid, the effect of lecithin concentration wastreated in the manner of a competitive inhibitorand Ki values were calculated from a plot of recip-rocal reaction velocities against reciprocal lecithinconcentrations, with the molecular weight oflecithin assumed to be 750. From such data, avalue of about 1.5 X 10-4 M was obtained as theinhibition constant (Ki) for Mann's lecithincompeting with a cell membrane phosphatidefor purified PF. Under our experimental condi-tions, this means that about 0.3 mg of lecithinin 2.5 ml of Ringer phosphate solution with boneslices inhibited the PF activity about 50%O. Whenthe phospholipase C activity of the same purifiedPF fraction was measured with varying con-

centrations of Mann's lecithin, a Km. value ofabout 2.7 X 102 M wvas obtained. Thus, about200 times as much lecithin was required to givehalf-maximal rate with PF-phospholipase as wasrequired to produce a 50% inhibition of therelease of alkaline phosphatase from bone slicesby PF. The Km value calculated for the phos-pholipase C activity of a purified inhibitor frac-tion was about 1.3 X 102 M with Mann's lecithin.

Effect of PF on bone tissue in vivo. Crude orpartially purified PF was very active in producingphosphatasemia, but the presence of a lethalcomponent precluded the use of relatively largeamounts of PF for certain in vivo tests. However,with the purified combined fraction C of Table 4it was possible to obtain a rapid increase of bloodalkaline phosphatase to very high levels afterintravenous injection into young rabbits. Thisallowed us to obtain more direct support for theimmunochemical data that indicated that boneis a principal source of the phosphatasemic re-sponse (Slein and Logan, 1962a).

Rabbits were anesthetized with ether, andinitial blood samples were taken from marginalear veins into heparinized tubes. The right hindlimb was ligated above the knee, the tibia wasamputated and placed in ice. Then, 1 ml of frac-tion C, containing approximately 30,000 units ofPF, was injected into a marginal ear vein.Further blood samples were obtained at inter-vals, and then the left tibia was removed. Bloodwas diluted 1:100 with cold distilled water, and0.1-ml samples were used for the determinationof alkaline phosphatase (Slein and Logan, 1962a).The tibial epiphyses were sliced, and triplicatesamples of each were prepared for the determina-tion of the rate of release of alkaline phosphatase.The results are presented in Table 7. In rabbit 1,the blood alkaline phosphatase increased abouttenfold in only 5 min. The release of phosphatasefrom slices of the tibia, which was removed 5min after the injection of PF, was significantlyhigher than that from control slices during thefirst 15 min in vitro. However, during the second15 min of incubation, there was little differencebetween the amounts of phosphatase released.With rabbit 2, the blood phosphatase increasedalmost 100-fold by 30 min, at which time thesecond tibia was removed. The release of alkalinephosphatase from bone slices was no longermarkedly higher than the control rate at thistime; in fact, it was lower than the control

378 J. BACTERIOL.

on June 25, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

PHOSPHOLIPASES OF B. CEREUS

TABLE 7. Effect of the intravenous injection of 30,000 units of purified PF on rabbit blood and epiphysealbone alkaline phosphatasea

Blood Bone slices

Phosphatase released PhosphataseRabbit no. Alkaline spontaneously released by

Sample phosphatase Sample PF added inFirst 15 Second vitrob (secondmin 15 min 15 min)

1 Control 34 Control 351 175Post-PF (5 min) 291 Post-PF (5 min) 564 202

2 Control 16 Control 192 112Post-PF (15 min) 716Post-PF (30 min) 1,460 Post-PF (30 min) 231 70

3 Control 48 Control 282 162c 1 ,OOldPost-PF (15 min) 1,318Post-PF (30 min) 2,726 Post-PF (30 min) 277 58c 160d

a Control blood samples and tibiae were removed prior to the injection of PF. The other tibiae wereobtained after the last blood samples. Except where noted, triplicate samples of slices from each bonewere tested, and the values reported are averages.

b Approximately 80 units of a crude PF preparation were added to each flask after the first 15-minsamples were taken.

c Single samples.d Duplicate samples.

during the second 15-min interval in vitro. Simi-lar results were obtained with rabbit 3. However,inthis test, about 80 units of PF were added to twopairs of flasks after the first 15-min samples weretaken from the bone-slice incubation. The furtherrelease of alkaline phosphatase induced by thePF added in vitro was much greater from slicesof the control tibia than from that obtained 30min after the in vivo administration of PF.These results indicate that alkaline phosphataseis rapidly released from the epiphyseal bonetissue into the bloodstream after the intravenousinjection of large amounts of PF. Shortly afterinjection, the release of phosphatase from thetissue is demonstrably greater than from controlslices (rabbit 1), but at a later time the differ-ence is less, or reversed (rabbits 2 and 3). Theeffect is probably due to the depletion of thealkaline phosphatase content of bone duringthe in vivo exposure to PF, as is indicated by thelowered response to the addition of PF to slicesin vitro (rabbit 3).

DISCUSSIONThe presence of at least two enzymes with

phospholipase C activity was demonstrated inculture filtrates of B. cereus grown under staticconditions. The enzymes not only were distin-guished by being separable on a DEAE cellulose

column, but also by their different sensitivitiesto trypsin and different degrees of protection bylecithin from heat inactivation. Although theyhad similar dissociation constants with onepreparation of emulsified soybean lecithin assubstrate, they had very different activities whentested with two batches of lecithin. Furthermore,the phospholipase that was not readily adsorbedby DEAE cellulose at pH 8.6 was found to haveno PF activity and, in fact, was strongly in-hibitory to PF in the bone-slice test. The otherphospholipase, which was eluted from DEAEcellulose by about 0.4 M tris (pH 8.6), had highPF activity. The different susceptibilities toinactivation by antiserum also distinguished thetwo phospholipases.

It seems likely that the PF-inhibitor and phos-pholipase C activities of the nonadsorbed fractionare identical, since neither was very sensitiveto inactivation by trypsin, both were inactivatedto about the same extent by heat, and both wereinsensitive to an antiserum that inactivated PF.Although there was good correlation between

the production of PF and phospholipase byseveral strains of bacilli, it had been concludedthat the two activities probably were not identi-cal (Slein and Logan, 1962a). However, the pres-ent results indicate their identity and explain theapparent discrepancy in our previous publication.

379VOL. 85, 1963

on June 25, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

SLEIN ANDLOGANB

The ratio of the two activities was remarkablyconstant during elution from DEAE cellulosewith a continuous gradient from about 0.4 Mto 0.6M tris (Fig. 1). Both activities were destroyedby trypsin, and the degree of inactivation wasreduced similarly for both in the presence oflecithin. A similar protection from heat inactiva-tion by lecithin was found. The inhibition of PFactivity by lecithin supports the hypothesis thatthe release of phosphatase from bone cells iscaused by phospholipase activity on the cellmembranes. The lecithin appears to competeeffectively with a cell membrane phosphatide forthe active center of the PF. Further evidenceindicating the identity of PF and phospholipaseincludes the stability of both activities over awide pH range at 37 C, the slight but significantprotection from heat afforded to both activitiesby dilute phosphate buffer, and a proportionaldecrease of both activities during storage of adilute solution of purified PF for several weeks.The previously reported failure of phospholi-

pase to be significantly inactivated by an anti-serum, as compared with PF activity, was prob-ably confused by the presence of more than onephospholipase in the crude culture filtrate thatwas tested. It is now apparent that the phos-pholipase associated with PF-inhibitor activityis not susceptible to inactivation by that anti-serum, whereas the phospholipase and PF ac-tivities in a purified PF fraction are both inacti-vated by the antiserum.

Concerning the previously assumed differentialdecay of PF and phospholipase activities duringprolonged storage of a lyophilized preparation, itis now evident that the material, which had beenprepared from a shaken culture of B. cereus,contained only the inhibitor phospholipaseactivity and probably never had any PF activity.

It is evident that the lethal and hemolysinfactors are not related to the phospholipase thatinhibits PF activity. Conflicting data have beenobtained with regard to the identity of lethalityand hemolysin and PF activities. However, sincesome fractions that had high PF activity ap-peared to have relatively low lethality andhemolysin activities, it seems likely that thephospholipase associated with PF activity is notidentical with the lethality and hemolysin fac-tors. Molnar (1962) separated the toxin of B.cereus into two components, neither of which

had phospholipase activity when tested by theegg-yolk reaction.

Relatively large doses of purified PF given torabbits intravenously resulted in the release ofalkaline phosphatase from bone epiphyses, sothat a significant decrease in the phosphatasecontent could be demonstrated with bone slicesin vitro. This supports the immunochemicalevidence for the participation of bone in thedevelopment of phosphatasemia. It is probablethat other tissues contribute to the phosphatase-mic response.The stabilization of PF activity from heat

inactivation by phosphate, or from trypsininactivation by lecithin, might be useful in thedevelopment of further procedures for purifica-tion of the enzyme. We plan to investigate themechanism of action of phospholipase on cellmembranes. It was found that PF did not cata-lyze the release of alkaline phosphatase fromcells of Escherichia coli that had been inducedto form alkaline phosphatase. This indicatesthat PF does not attack the cell wall of the micro-organism sufficiently to allow the escape of thephosphatase that lies between the cell wall andmembrane (Malamy and Horecker, 1961).

Altenbern (1962) offered evidence that theproduction of the toxic complex, including phos-pholipase, by B. cereus is dependent upon theinduction of lysogeny, and he suggested that thephospholipase may be essential for the release ofphage from inside the bacillus.Our finding of lower phospholipase activities in

shaken cultures of B. cereus is probably related tothe inactivation of lecithinase by shaking, asrecently reported by Kushner (1962).

ACKNOWLEDGMENT

The authors are grateful to William A. Jarrettfor technical assistance.

LITERATURE CITED

ALTENBERN, R. A. 1962. Lysogeny and toxinogenyin Bacillus cereus. Biochem. Biophys. Res.Commun. 9:109-112.

COSTLOW, R. D. 1958. Lecithinase from Bacillusanthracis. J. Bacteriol. 76:317-325.

DAWSON, R. M. C., AND A. D. BANGHAM. 1961.The importance of electrokinetic potentials

380 J. BACTERIOL.

on June 25, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

PHOSPHOLIPASES OF B. CEREUS

in some phospholipase-substrate interactions.Biochem. J. 81:29P.

DEUEL, H. J., JR. 1955. The lipids, vol. 2, p. 23-26.Interscience Publishers, Inc., New York.

HAYAISHI, 0. 1955. Phospholipases, p. 660-672.In S. P. Colowick and N. 0. Kaplan [ed.],Methods in enzymology, vol. 1. AcademicPress, Inc., New York.

KUNKEL, H. G., AND A. TISELIUS. 1951. Electro-phoresis of proteins on filter paper. J. Gen.Physiol. 35:89-118.

KUSHNER, D. J. 1962. Formation and release oflecithinase activity by growing cultures ofBacillus cereus. Can. J. Microbiol. 8:673-688.

MALAMY, M., AND B. L. HORECKER. 1961. Thelocalization of alkaline phosphatase in E.coli K,2. Biochem. Biophys. Res. Commun.5:104-108.

MOLNAR, D. M. 1962. Separation of the toxin ofBacillus cereus into two components andnonidentity of the toxin with phospholipase.J. Bacteriol. 84:147-153.

OTTOLENGHI, A., S. GOLLUB, AND A. ULIN. 1961.Studies on phospholipase from Bacillus cereus.

I. Separation ot phospholipolytic and hemol-ysin activities. Bacteriol. Proc., p. 171.

SCHMIDT, G., AND M. LASKOWSKI, SR. 1961. Phos-phate ester cleavage, p. 3-35. In P. D. Boyer,H. Lardy, and K. Myrback [ed.], The enzymes,vol. 5. Academic Press, Inc., New York.

SKERMAN, V. B. D. 1959. A guide to the identifica-tion of the genera of bacteria. The Williams &Wilkins Co., Baltimore.

SLEIN, M. W., AND G. F. LOGAN, JR. 1962a. Mecha-nism of action of the toxin of Bacillusanthracis. II. Alkaline phosphatasemia pro-duced by culture filtrates of various bacilli.J. Bacteriol. 83:359-369.

SLEIN, M. W., AND G. F. LOGAN, JR. 1962b. Phos-phatasemia factor in toxic filtrates fromcultures of bacilli. Federation Proc. 21:229.

TRAYSER, K. A., AND S. P. COLOWICK. 1961.Properties of crystalline hexokinase fromyeast. III. Studies on glucose-enzyme inter-action. Arch. Biochem. Biophys. 94:169-176.

WADDELL, W. J. 1956. A simple ultraviolet spectro-photometric method for the determination ofprotein. J. Lab. Clin. Med. 48:311-314.

381VOL. 85, 1963

on June 25, 2020 by guesthttp://jb.asm

.org/D

ownloaded from