mmbr.asm.org · KADIS, MONTIE, ANDAJL phenomenon is also involved in the inhibitory effect ofthe toxin on mitochondrial respiration. Thisis correlatedwiththetoxin'sabilitytoinduce

BACEROLOGICAL REVEWS, Mar., 1966 Vol. 30, No. 1Copyright © 1966 American Society for Microbiology Printed in U.S.A.

The Murine Toxin of Pasteurella pestis: A Study inIts Development

SOLOMON KADIS, THOMAS C. MONTIE, AND SAMUEL J. AJLResearch Laboratories, Department ofBiochemistry, Albert Einstein Medical Center, Philadelphia, Pennsylvania

I NTRODUCTION................................................................ 177ISOLATION AND PURIFICATION OF THETON.....................................T.178IMMUNOLOGICAL PROPERTES .................................................... 179MECHANIM OF ACTION OF THE ToXIN .......................................... 180

Effect of the Toxin on Microbial Extracts and Crude Tissue Homogenates........... 180Action of the Toxin on Mammalian Mitochondria ......... ....................... 180

Effect on respiration....................................................... 180Ability of the toxin to induce mitochondrial swelling............................ 182Effect on the electron transport system....................................... 183Effect on mitochondrial ion accumulation...................................... 185

DISTRIBUTION AND SYNTHSIS OF PLAGUE MuRm ToxIN.......................... 185Location of the Toxin in the P. pestis Cell...................................... 185Resolution and Isolation of Two Toxic Components............................... 186Selective Inhibition of Murine Toxin Synthesis by Tryptophan Analogues............ 186

Antagonism ofanalogue bacteriostatic action and the selective inhibition of toxin... . 187Toxin protein content in extracts from analogue-treated cells..................... 187Inhibitory effect of analogues on spheroplast toxin content...................... 188

SummARY.................................................................... 188LrrERATURECrnu. ............................................................ 188

INTRODUCrION

Plague, for which the organism Pasteurellapestis is responsible, has been a serious problemfor mankind through each epoch of his existence.Today, despite great knowledge and technology,many world areas still suffer seriously from thisdisease. In the late 1940's and early 1950's, a seriesof investigations on plague toxin was initiated byAjl and co-workers, and continued to the pres-ent. These studies, which provided a renewed ap-proach to the pathogenesis of plague, form thebasis of this review.

Certain aspects of the symptomatology andpathology of the disease prompted Dieudonneand Otto (16) in 1928 to suggest that toxic sub-stances of considerable potency are liberated byP. pestis during infection. Additional interest inthe role of toxin in the disease was generated bythe findings of McCrumb et al. (53), in Mada-gascar, that antibiotics administered 36 to 48 hrafter the onset of the disease failed to save pa-tients despite the fact that, on autopsy, the bloodand organs were sterile.The plague bacillus is known to contain several

distinct and readily separable antigenic compo-nents. Our interests, however, have remained inthe toxin produced by P. pestis, which is specifi-cally responsible for death of mice and rats andis known, therefore, as the "murine" toxin.

It is not the purpose of this review to covercompletely the available literature on plague

toxin. We shall trace here the development of thework in our laboratory with the murine toxin ofP. pestis and refer to the investigations of othersonly as they relate to and amplify our studies.The primary objectives from the start of our

work were isolation and purification of the toxinand study of the mechanism of its action at anenzymatic level. Recently, we have concerned our-selves with how this toxin is synthesized by thebacterium, the regulation of this synthesis, andthe distribution of this protein in the P. pestis cell.The murine toxin of P. pestis has been ob-

tained in a highly purified form. However, wefound, as have others working with certain exo-toxins (for example, diphtheria toxin), that thetoxic activity of highly purified preparations isassociated with more than one component. Ourpurest preparations are composed of two distinctmolecular species, each exhibiting essentiallyidentical toxic activity.

Studies on the mode of action of the toxin in-volve primarily the effects of the toxin on mam-malian mitochondria. An association between theability of the toxin to inhibit mitochondrial respi-ration and its action in vivo has been established.The site of inhibition of mitochondrial respira-tion resides in the electron transfer chain in theregion between reduced nicotinamide adeninedinucleotide (NADH2) or succinate and cyto-chrome b and, more specifically, at the NADH2-coenzyme Q reductase complex. A permeability

177

on October 10, 2020 by guest

http://mm

br.asm.org/

Dow

nloaded from

http://mmbr.asm.org/

KADIS, MONTIE, AND AJL

phenomenon is also involved in the inhibitoryeffect of the toxin on mitochondrial respiration.This is correlated with the toxin's ability to inducemitochondrial swelling and to interfere withrespiration-dependent mitochondrial ion accumu-lation.

Concerning the mode of synthesis of this toxinby P. pestis, it was found that tryptophan ana-logues selectively inhibit toxin biosynthesis, andthat this inhibitory effect can be reversed by inter-mediates of tryptophan biosynthesis in micro-organisms. This finding suggests a tryptophanrequirement for toxin formation.

ISOLATION AND PURIFICATION OF THE ToxiN

Studies on the biological action and antigenicbehavior of plague toxin were handicapped bythe lack of adequately pure material. It wasknown that the toxin of P. pestis is associatedwith the cells and must be liberated from them.Numerous procedures, including cell lysis, sonicvibration, and chemical extraction, were em-ployed to liberate the toxin from cells. Markl(51, 52) used filtrates of old broth cultures. Girard(21) employed a freezing and thawing system, andJawetz and Meyer (31) held agar-grown suspen-sions at 37 C for 48 hr and then at 4 C for 24 hr.

Sonic vibration was used successfully by Smithet al. (68) to obtain highly toxic material fromP. pestis. Extraction of dried cell powders orwhole cells with relatively simple compounds wasemployed by Baker et al. (7, 8) with sodiumchloride, by Lustig and Galeotti (48) with 1%potassium hydroxide, and by Goodner et al. (22)with sodium deoxycholate.These toxin preparations obtained were crude,

for they contained numerous antigenic compo-nents in addition to the toxin. Initial attempts topurify these toxic materials consisted of precipi-tating them with either ammonium sulfate (8)or calcium chloride (79) and freeing them fromthe nonprecipitable fractions. Baker et al. (8)treated toxin, prepared by extracting acetone-dried cells of P. pestis with 2.5% sodium chloride,with 30 and 40% saturated ammonium sulfate toremove antigens other than the toxin as the resi-dues. The toxin was also precipitated directly bythe addition of 55 to 67% saturated ammoniumsulfate. The best samples, having an LD50 for 20-gmice of 0.6 to 0.8 Mug, represented a 10-fold con-centration of the toxin. However, all the toxicpreparations were heavily contaminated withatoxic soluble antigens, and attempts at furtherpurification resulted in a 50 to 75% loss in tox-icity.

Englesberg and Levy (18) obtained highly toxicfractions by precipitating the crude toxin ob-

tained from autolysates of P. pestis grown at 30 Cin semisynthetic medium (casein hydrolysate-mineral-glucose medium) with saturated am-monium sulfate, followed by dialysis and lyophili-zation. Though the relative purity of this materialwas not given, this method, compared with previ-ous attempts, provided substantially greateryields of toxin.

Ajl et al. (1) undertook the first extensive studydesigned to obtain preparations of plague toxinwhich were pure by all of the known standardsemployed to ascertain protein purity. The initialphase of this work considered purification of thetoxin by chemical means, involving extraction ofthe toxin from acetone-dried cells of the avirulent"Tjiwidej" (TJW) strain of P. pestis with 2.5%sodium chloride, followed by ammonium sulfateand isoelectric precipitations for partial separa-tion of the toxin from the envelope substance.This preparation was then treated with manganesechloride for removal of nucleic acids, with metha-nol precipitation to concentrate protein andremove extraneous materials and calcium phos-phate gel absorption with elution to separatefurther the toxin from the envelope substance.Lipoid materials were removed by chloroformextraction. The final material, having an intra-peritoneal LDrO of 2.6 jig for 16- to 18-g mice,exhibited a sevenfold increase in toxicity. How-ever, it contained one major and frequently oneor more minor components when observed in theanalytical ultracentrifuge.The high resolving power for fractionating pro-

tein mixtures afforded by continuous-flow paperelectrophoresis led to its utilization (1) for fur-ther purification of the toxin. Material obtainedfrom the final stage of the chemical purificationprocedure was passed twice through a paperelectrophoresis cell described by Durrum (17).Electrophoretic patterns showed that this materialwas considerably more pure than that obtainedby chemical procedures alone. Purified toxin withan isoelectric point of 4.7 behaved as a homoge-nous protein in the ultracentrifuge and Tiseliuselectrophoresis cells and was free from carbohy-drates, nucleic acids, and capsular antigen. Sedi-mentation and diffusion data indicated a molecu-lar weight for the toxin in the order of 74,000.

This toxin was subjected (3) to the very sensi-tive gel diffusion precipitation reactions of Oudin(60) and Ouchterlony (61), and at least two andfrequently three or more individual zones of pre-cipitation were found. Since the main use for thistoxin was to be the determination of the mecha-nism of its action at an enzymatic level, it wasimperative to achieve an even greater degree ofpurity. As considerable trauma to the toxin mole-

178 BACTERIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from


MURINE TOXIN OF PASTEURELLA PESTIS

cule was involved in chemical extraction proce-dures, the method was simplified considerably.Crude toxin was fractionated with ammoniumsulfate, and the dialyzed fraction between 35 and70% saturated ammonium sulfate was passedseveral times through the continuous-flow, hang-ing-curtain electrophoresis apparatus developedby Karler (39). The first two passes were inVeronal buffer, 0.01 ionic strength (pH 8.6), withsubsequent passes in maleate buffer, 0.01 ionicstrength (pH 6.0). After final passage, a 17-foldpurification of the toxin was achieved. This finalmaterial exhibited only one band against cruderabbit antisera in the Oudin reaction and had anintraperitoneal LD5o for 14- to 18-g Swiss albinomice of 0.7 jg and an intravenous LD50 of lessthan 0.2 Mig. Similar results were obtained withtoxin from virulent strains of P. pestis by Spivackand Karler (69). Their material had an intra-venous LD50 for mice of 0.1 ug.

Extremely pure preparations of toxin were ob-tained with the Karler electrophoresis apparatus.This procedure, however, is characterized byslow rates of separation of protein components,entailing hours or days needed for fractionationof large volumes of dilute protein solutions.There also may be losses of sample due to ad-sorption on paper, the supporting medium.A transparent methyl methacrylate cell packed

with fine glass beads was developed by King,Jensen, and Stubbings (32, 40) for preparativeelectrophoresis, and was modified and refined forgeneral electrophoretic separatory procedures.Since large quantities of toxin are required forstudies involving the mechanism of toxin action,the continuous-flow electrophoresis apparatusemploying glass microbeads, fractionating largevolumes of material in short periods with insig-nificant adsorption of sample upon the glass beadmatrix, was chosen for purifying plague murinetoxin. Crude toxin, obtained from autolyzed P.pestis cells and fractionated with ammonium sul-fate between 35 and 70% saturation, was passedthrough the electrophoresis apparatus accordingto the procedure reported by Winsten et al. (77)using serum proteins. The toxin so obtained hadan intraperitoneal LD5o for 16- to 18-g Swissalbino mice of approximately 2 Mg of protein, andit contained 95% protein.To determine whether plague toxin possessed

any unusual components which could accountfor its high toxicity, detailed elemental and aminoacid analyses were performed (9) on the mosthighly purified samples of toxin available. Eight-een amino acids and a number of elements wereidentified. On a dry-weight basis, over 98% of thetoxin molecule was accounted for by organic

analysis, including ammonia and ash content.There was nothing unusual about the toxin mole-cule aside from the high proportion of acidicamino acids, which verified the previously ob-served isoelectric point of the toxin of 4.7.

IMMUNOLOGICAL PROPERTIES

Before purified P. pestis toxin was available, itsrole in immunity against plague was largely con-jectural. The purified preparations of Ajl et al.(1) provided an excellent source of antigen to beused for detailed investigations of the immuno-logical properties of plague toxin. Warren et aL(76) utilized the most purified preparations ofAjl et al. (1) obtained from the TJW strain ofP. pestis to produce antitoxin in rabbits. Theantiserum obtained was able to flocculate, hemag-glutinate, and fix complement with toxin andtoxoid.

Additional investigations revealed that thespecific TJW antitoxin neutralizes toxins ob-tained from different strains of P. pestis. Thisfinding and the observation that purified TJWtoxin reacts with the antisera prepared from avariety of avirulent and virulent strains of theplague bacillus suggest strongly that all thesetoxins have similar antigenic structures.To determine the degree of animal immuniza-

tion against toxin (2), mice were challenged intra-peritoneally with formalin-treated toxin andintravenously with toxin-antitoxin mixtures. Inthe former case, mice were protected against 60 to80 LD5o doses of the toxin, and in the latter onlya few LD5o doses of the toxin were neutralized.The role of toxin in plague infection has been

studied by Meyer (unpublished data), who foundthat rats and guinea pigs died in shock when in-jected with soluble toxins or with killed and driedor living P. pestis cells. The liver and spleenserved as filters which removed P. pestis organ-isms from the blood after intravenous injection.Bacilli were destroyed in the liver and releasedtoxic materials responsible for the intoxication.The total amount of bacillary somatic antigenwas related directly to the speed with which thesymptoms of intoxication were noted.

It has been tacitly assumed that a definabletoxic material is present in the circulating blood inthe course of plague infection. However, its rolein the pathophysiology of the disease has notbeen elucidated. In a fatal case of plague, manydifferent events, all related, proceed so rapidlythat it is difficult to ascertain the dominant factor.As it is difficult to determine the order of thesephysiological events and to identify factors re-sponsible for these events, the role of toxin in

VOL. 30, 1966 179


http://mm

br.asm.org/

Dow

nloaded from


KADIS, MONTIE, AND AML

plague infection is certainly not completely under-stood.

MECHANISM OF ACTION OF THE ToxmN

Effect of the Toxin on Microbial Extracts andCrude Tissue Homogenates

The first serious attempt to study the mecha-nism of action of toxin on an enzymatic levelinvolved the effect of toxin on oxidation of a

variety of substrates by cell-free microbial ex-tracts and crude mouse liver homogenates. It wasfound (5) that the toxin specifically inhibitedoxidation of a-keto acids. For example, whereasthe toxin inhibited oxidation of a-ketoglutaricand pyruvic acids, it exhibited a significantlylesser effect on such acids as succinic and citric.

Similar results were obtained with heat-inac-tivated and formalin-treated toxin. These findingscan be understood only when compared withresults of investigations concerning the effect ofinactivated toxin on mitochondrial respiration.This discussion follows in a subsequent section.The inhibitory effect of toxin on the oxidation

of a-keto acids by cell-free microbial extracts andcrude mouse liver homogenates was reversed bythe addition of an excess of nicotinamide adeninedinucleotide (NAD) but not by nicotinamideadenine dinucleotide phosphate (NADP). Thisfinding suggests that the toxin possesses nicotin-amide adenine dinucleotidase activity and therebyinterferes with certain NAD-dependent reactionsby depriving them of the required coenzyme.Thus, when this toxin was incubated with NADin the presence of phosphate and the reactionproducts were analyzed (4), three compoundswere recovered, namely, nicotinamide mononu-cleotide (NMN), adenylic acid, and a compoundsimilar to, but not identical with, the classicaladenosine diphosphate (ADP). In the absence ofphosphate very little NAD was broken down.This was to be expected, since the formation ofNMN and a compound similar to ADP per moleof NAD cleaved requires the incorporation of anadditional phosphate group from the reactionmedium. The requirement for phosphate classi-fied this reaction as an "unusual" type of nicotin-amide adenine dinucleotidase reaction, for neitherof the following classical ways by which NAD iscleaved enzymatically is phosphate-dependent:at the pyrophosphate linkage to yield adenylicacid and NMN (41) or at the nicotinamide riboselinkage to form nicotinamide (24) and adenosinediphosphoribose (38).Toxin preparations used to study nicotinamide

adenine dinucleotidase activity and inhibitoryeffects on a-keto acid oxidation were purified bychemical and electrophoretic procedures of Ajl

et al. (1). When serologically homogenous toxinwas obtained (3), each of these properties wasreinvestigated and it was found that the activityon NAD disappeared, whereas the characteristica-keto acid inhibition remained. The nicotin-amide adenine dinucleotidase activity was re-covered in a different protein fraction. The elec-trophoretic mobility of this protein was verysimilar to that of the toxin.

Although nicotinamide adenine dinucleotidase(NADase) activity does not appear to be asso-ciated with mode of action of plague murinetoxin, an interesting NADase has been un-covered. Further investigations relative to themechanism of action of this enzyme are war-ranted.

Action of the Toxin on Mammalian MitochondriaEffect on respiration. It was important to de-

termine toxin action at a higher level of organiza-tion from the standpoint of enzymatic structureand closer to that expected to occur in the animalwhile still maintaining an in vitro system. Towardthis goal investigations were undertaken to studytoxin effect on mitochondria, which are known tobe the active sites of respiration in animal tissues.A critical observation, providing an insight tounderstand the susceptibility and resistance ofanimal species to biological poisons, was in theassociation present between the ability of toxin toinhibit mitochondrial respiration from certainspecies and the susceptibility of these animals toin vivo action of the toxin.

TABLE 1. Effect ofplague toxin, Vi and 0 antigens,and bovine serum albumin on the respiration of

heart mitochondria*

PerSource of Additions Og con- cent

mitochondria sumedt inhibi-tion

Rat heart None 0.72Boiled plague 0.66 8.3

toxin (2.5 mg)Bovine serum al- 0.72 0.0bumin (2.5 mg)

o antigen (2.5 mg) 0.65 9.7Plague toxin (1.0 0.24 66.7mg)

Rabbit heart None 0.41Vi antigen (2.5 mg) 0.41 0.0o antigen (2.5 mg) 0.50 0.0Plague toxin (2.5 0.44 0.0mg)

* This table was compiled from results given inreference 62.

t Expressed as micromoles per liter per second.

180 BACTERIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from



It is known that the toxin is lethal for themouse and rat but not for the rabbit, chimpanzee,dog, or monkey. Therefore, we were interested tolearn that toxin inhibited the respiration of heartmitochondria from the toxin-susceptible rat andmouse, but had no effect on similar preparationsfrom the toxin-resistant rabbit (Table 1). Addi-tional experiments revealed that the toxin wasunable to inhibit respiration of heart mitochon-dria from chimpanzee, dog, and monkey (66).Only the exogenous mitochondrial respiration oftoxin-susceptible animals was inhibited. Theendogenous respiration remained unaffected.Likewise, oxidative phosphorylation was in noway altered by the toxin. The inhibitory effect onmitochondrial respiration was specific for thetoxin investigated, since bovine serum albumin,representing another protein, and the Vi and 0lipopolysaccharide antigens had no effect onmitochondrial respiration.Toxin action varied not only with respect to the

species yielding the mitochondria but also withrespect to the specific organ from which theywere isolated. It was found that toxin had noeffect on the respiration of brain mitochondria oftoxin-susceptible rats. Studies on the effect oftoxin on liver mitochondria revealed that thetoxin inhibited the respiration of liver mitochon-dria from the rat and rabbit to the same extent asrat heart mitochondria (35). This inhibition ofrabbit liver mitochondrial respiration is of inter-est, since other studies indicate that 10 mg oftoxin, injected intraperitoneally, is not lethal for2-kg rabbits.

Although physiological concentrations of toxinwere found to inhibit mitochondrial respiration,toxin treated in any manner that made it atoxicfor the animal resulted in the concomitant loss ofability to inhibit mitochondrial respiration (seeTable 1). This contrasted directly with results ob-tained with cell-free bacterial extracts and crudetissue homogenates. This inconsistent behavior ofinactivated toxin can be explained as follows.One basic characteristic of this toxin, which is itsability to inhibit the oxidation of certain com-pounds, remains with the molecule even whentoxicity is lost. When homogenates are employed,the inhibition of keto acid oxidation reflectschance interaction of the toxin or toxoid withdispersed enzyme molecules. The active portionof the toxin molecule for this effect remains afterheating or formalin treatment of the molecule.However, the mitochondrion retains a relativelyhigh degree of organization. In this case the ef-fects depend upon critical spatial relations, and,to exert these effects, the toxin molecule mustretain not only that portion reacting with theenzyme, but also the portion permitting effective

orientation within the structure of the mitochon-drion.The inability of inactivated toxin to inhibit

mitochondrial respiration demonstrated a cor-relation between toxicity of the toxin in vivo andits in vitro action on mitochondria and led to thehypothesis that inhibition of mitochondrial res-piration may explain the action of this toxin invivo.As this toxin inhibited heart mitochondrial

respiration of toxin-susceptible species, the hy-pothesis above would indicate heart malfunctionas an early sign of toxin action in susceptiblespecies. Heart malfunction is detected easily byelectrocardiographic measurements. Indeed, al-terations did occur in the S-T segment of theelectrocardiogram of the rat within 60 min afterinjection of lethal or sublethal doses of toxin andprior to any changes in hematocrits or bloodpressures. In animals surviving sublethal doses ofthe toxin, electrocardiographic changes observedinitially were no longer evident after 24 to 48 hror after the animal had recovered completely.Similar electrocardiographic changes did notoccur in rats dying from hemorrhagic shock,hypoxia, glucose intoxication, or in toxic deathsfrom Escherichia coli endotoxin. Toxin-resistantrabbits exhibited no alterations in their electro-cardiographic tracings when injected with toxindoses up to 50 mg. Thus, whenever the toxin in-hibited heart mitochondrial respiration, corre-sponding changes in electrocardiographic patternsobtained from intoxicated anils were found.Conversely, failure to observe such in vivo effectswas associated with unaltered mitochondrial res-piration.

If the inability of the toxin to inhibit the respi-ration of heart mitochondria derived from thetoxin-resistant rabbit were due to the exclusionof toxin by the rabbit heart mitochondrial mem-brane, disruption of these mitochondria shouldresult in inhibition of their respiration with addi-tion of toxin comparable to that obtained withintact heart mitochondria from the toxin-sus-ceptible rat. Experimentally, mitochondria weredisrupted chemically with deoxycholate andphysically by means of sonic vibration, and it wasfound that, although toxin had little or no effecton respiration of unaltered rabbit heart mito-chondria, respiration of disrupted mitochondriawas inhibited to a significant extent (35).

Toxin-susceptible animals can be immunizedagainst toxin. When mitochondria were isolatedfrom the hearts and livers of such immunizedanimals and incubated with the toxin, it wasdemonstrated that the inhibitory effect of toxin onheart mitochondrial respiration was reversed,whereas the respiration of liver mitochondria of

VOL. 30, 1966 181


http://mm

br.asm.org/

Dow

nloaded from



TABLE 2. Effect of toxin on swelling of heart, liver,and brain mitochondria*

Net change inoptical density

Source of mitochondria Toxin concn at 520 mp of ex-perinmental minus

controlt

mg

Rat heart 2.0 0.3250.5 0.210

Rabbit heart 2.0 0.0250.5 0.020

Rat liver 2.0 0.2430.5 0.120

Rabbit liver 2.0 0.2350.5 0.170

Rat brain 2.0 0.020

* This table was compiled from results given inreferences 34 and 35.

t Change in optical density at 520 mAs recordedafter 30 min of incubation.

0.7

E0CMNI')

I-

Fn-zwa

0

a-0

0.6

0.5

04

0.3

0.2

0.1

o CONTROL& CONTROL+ 0.1 mMNoCN° 2.0mg TOXIN* 2.Omg TOXIN + O.1mM NaCN* 0.5mg TOXIN* 0.5mg TOXIN + 0.1 mM NaCN* 0.1 mg TOXINv 0.1mg TOXIN + O.ImMNaCN

0 3 6 9 12 15 18 21 24 27 30

MINUTESFIG. 1. Effect of NaCN on toxin-induced swelling

of rat heart mitochondria. From Kadis and Ail (34).

immunized and nonimmunized animals was in-hibited to the same extent (35).

Ability of the toxin to induce mitochondrialswelling. Mitochondria, primary sites of oxidativemetabolism, also possess the following character-istic properties: ability to cause transport andaccumulation of certain electrolytes and abilityto take up water and swell and to extrude waterand contract. These transport processes are de-pendent upon respiratory energy and are asso-

ciated with electron carriers and coupling enzymesof the respiratory chains, present in and consti-tuting a large portion of the protein layer of themitochondrial membrane.A wide variety of agents are known to cause

mitochondrial swelling (44). One type, known aselectron transport-dependent swelling, is thought(30) to depend upon increased membrane perme-ability which stops short of osmotic rupture ofthe mitochondria. The toxin under considerationdoes not inhibit the respiration of intact rabbitheart mitochondria as the membranes of thesemitochondria apparently exclude the toxin. Thissuggests the involvement of a permeability phe-nomenon. The studies by Kadis and Ajl (34)concerning the effect of toxin on mitochondrialswelling and the relationship between its respira-tion and swelling effects revealed that toxin in-duced rat heart and rat and rabbit liver mito-chondria to swell but had no such effect on rabbitheart mitochondria (Table 2). Likewise, brainmitochondria exhibited very little spontaneousswelling and this was not affected by the additionof toxin (35). Thus, in every case where the toxininhibits mitochondrial respiration, it inducesswelling, and when it is incapable of inhibitingrespiration, it is, likewise, unable to promoteswelling.

Additional experiments (34) showed that,when the toxin is heat-inactivated or neutralizedwith antitoxin, it no longer induced swelling.This indicates that only toxin, active in vivo, canexert the in vitro swelling effect.The mechanism of toxin-induced mitochondrial

0.7LE I

° 0.6

0.5

I-n 0.4zw0

-i0.301 0.2a-0

0.1

o CONTROLA CONTROL + O.OmM NaN3o 2.0 mg TOXIN* 2.0mg TOXIN + IO.OmM NaN3A 0.5mg TOXIN* 0.5mg TOXIN + O.OmM NaN3* 0.1mg TOXINv O.1mg TOXIN + IO.OmM NaN3

.aa .a a . . a a I

0 3 6 9 12 15 18 21 24 27 30

MINUTESFIG. 2. Effect of NaN3 on toxin-induced swelling of

rat heart mitochondria. From Kadis and Ail (34).

182 BACmwFoL. REV.

- F


http://mm

br.asm.org/

Dow

nloaded from



E 0.70N

I)

1 0.6

cnzw

0.4

-J

I-a-

0

0.3

0.2

0.1

o CONTROL

& CONTROL + 0.1mM DNPa 2.0mg TOXIN* 2.0mg TOXIN + 0.lmM DNP* 0.5mg TOXIN* 0.5mg TOXIN + 0.lmM DNP* O.mg TOXINv O.lmg TOXIN + O.lmM DNP

0 3 6 9 12 15 18 21 24 27 30

MINUTESFIG. 3. Effect of 2,4-dinitrophenol on toxin-induced

swelling of rat heart mitochondria. From Kadis and

Ajl (34).

swelling was investigated by determining the effectof certain inhibitors of electron transport or highenergy intermediate-supported swelling, such ascyanide (45), azide (30), and 2,4-dinitrophenol(70), on toxin-induced swelling. Each of thesecompounds prevented the swelling of rat heartmitochondria promoted by the toxin (Fig. 1 to 3).When rat heart mitochondria, for example, areincubated in an appropriate medium, a small butconsistent degree of spontaneous swelling is ob-served. This spontaneous swelling is characterizedby an initial, short lag period. The swelling curvesin some of the toxin-induced swelling experimentstogether with the inhibitor studies showed that theabove-mentioned lag periods were eliminated bythe toxin and restored by cyanide, azide, and2, 4-dinitrophenol. This elimination of the lagperiods by a swelling agent and their restorationby electron transport inhibitors is characteristicof electron transport-dependent swelling. Theswelling caused by low concentrations of toxinwas prevented completely by each of the inhibitorsinvestigated. The partial prevention of swellingexhibited in the presence of high concentrationsof toxin could be due to the damaging effect oflarge concentrations of toxin on the mitochon-drial membrane, resulting in the elimination ofelectron transport.

Lehninger (43) observed that, whereas a largevariety of chemical agents promoted swelling,

only one agent, adenosine triphosphate (ATP),together with Mg+, was responsible for mito-chondrial contraction. In agreement with thesefindings, toxin-induced swelling was reversed byATP plus Mg++. Although ATP alone producedsome initial reversal, this reversal effect rapidlydecreased with time.

Effect on the electron transport system. At thetime when it became apparent that a permeabilityphenomenon was involved in the action of thistoxin, studies on the inhibitory effect of this toxinon mitochondrial respiration were reinitiatedwith the view towards pinpointing as precisely aspossible the site of the inhibition. Since the toxininhibited mitochondrial respiration but did notinterfere with oxidative phosphorylation, it ap-peared logical that the toxin might exert its effecton the electron transport system. Kadis et al. (36)reported that, although the toxin inhibited theoxidation of a-ketoglutarate, succinate, malate,and fl-hydroxybutyrate in the presence of ADPas phosphate acceptor, the percentage inhibitionwas, in no case, altered upon the addition of2,4-dinitrophenol. In this respect the toxin didnot behave, for example, like oligomycin, a classi-cal inhibitor of oxidative phosphorylation. Lardyet al. (42) described the properties of oligomycinand other inhibitors of phosphorylating oxidationand reported that the inhibition of oligomycinwas relieved by 2,4-dinitrophenol.The cytochromes are among the chief compo-

nents of the electron transport system, and eachmust remain in a reduced state for electrons to betransferred from reduced nicotinamide adeninedinucleotide (NADH2) or succinate to oxygen,the terminal electron acceptor. A general indica-tion can be obtained as to the site of action ofthe compound under investigation if, upon theexamination of the absorption spectrum of amitochondrial suspension to which has beenadded a compound whose mode of action is tobe determined, an alteration can be found in oneor more of the absorption peaks correspondingto specific cytochrome components.

Absolute and difference spectra of the cyto-chrome components of rat heart and liver mito-chondria incubated in the absence and in thepresence of toxin revealed that the addition oftoxin caused the oxidation of cytochromes a, b,and c. This suggested that the toxin exerts itsinhibitory effect on mitochondrial respiration byacting on the electron transport system in theregion between NADH2 or succinate and cyto-chrome b.

Confirmatory evidence on this point was ob-tained. The toxin had no effect on the oxidation ofascorbate by rat heart or liver mitochondria inthe presence of tetramethylphenylenediamine

VoL.30,1966 183

I


http://mm

br.asm.org/

Dow

nloaded from



(TMPD) or cytochrome c as electron carriers(36). Since TMPD serves as a mobile electroncarrier between ascorbate and members of therespiratory chain and acts between cytochrome c

and oxygen (20), it appears that the toxin has noeffect on the area of the electron transport sys-

tem between cytochrome c and oxygen.

The next step in locating the precise site ofaction of the toxin on the electron transport sys-

tem involved investigations on the effect of thetoxin on the activity of enzymes known to occur

between NADH2 or succinate and cytochrome b.One of these enzymes is NADH, dehydrogenase.The specific activity of this enzyme, as assayed bythe reduction of ferricyanide (54), in rat heartand liver mitochondria, as well as in electrontransport particles prepared from rat heart, wasnot altered by the addition of toxin. This findingsuggested that the toxin might not act on indi-vidual enzymes but on complexes of two or morerespiratory carriers, representing limited segmentsof the electron transfer chain.Four such complexes were isolated from beef

heart mitochondria, and each one was obtainedin highly purified form. Complex I corresponds tothe NADH2-coenzyme Q reductase of Hatefi et al.(26), which catalyzes the reduction of CoQ byNADH2. Complex II refers to the succinic-coen-zyme Q reductase of Ziegler and Doeg (80), whichcatalyzes the reduction of CoQ2 and, to a con-siderably lesser extent, CoQo0 by succinate. Com-plex III is the reduced coenzyme Q-cytochromec reductase of Hatefi et al. (27) that catalyzes thereduction of cytochrome c by reduced CoQ.Complex IV is the cytochrome oxidase system

(19, 23) that catalyzes the oxidation of reducedcytochrome c by molecular oxygen. It should benoted that Hatefi et al. (25) discovered that com-plexes I to IV can be used as building blocks toreconstitute all or part of the electron transferchain.When purified NADH2-cytochrome c reduc-

tase, from which NADH2-coenzyme Q reductaseand reduced coenzyme Q-cytochrome c reductaseare derived, was incubated with toxin, its activitywas inhibited to a significant extent (unpublisheddata). NADH2-cytochrome c reductase activityof electron transport particles obtained fromheavy beef heart mitochondria (ETPH) was,likewise, inhibited by the toxin.

Similar results were obtained with NADH2-coenzymeQ reductase. SinceNADH2 dehydrogen-ase is a major component of complex I, NADH2-coenzyme Q reductase is capable of catalyzingthe rapid reduction of ferricyanide by an amytal-insensitive reaction, a characteristic property ofmitochondrial NADH2 dehydrogenase. NADH2-ferricyanide reductase activity of the purifiedcomplex I, as well as that from ETPH, was notinhibited by the murine toxin of P. pestis. Thisindicates that the toxin has no effect on the ac-tivity of NADH2 dehydrogenase and confirms theresults of previous investigations (36) with intactmitochondrial suspensions. Additional evidenceon this point stems from the fact that differencespectra of toxin-treated complex I revealed thatthe flavoprotein of this enzyme complex was asreadily reduced in the presence of toxin as in itsabsence. Moreover, electron paramagnetic res-onance spectroscopic studies on toxin-treated

TABLE 3. Inhibition by toxin of Ca+ and Pi uptake by rat heart mitochondria and reversal by EDTAG

Per cent inhibitionComponents present Ca++ taken upb Pi taken Upb

Ca++ taken up Pi taken up

Complete systeme............................... 896 562 -

Plus 2.0mg of toxin.......................... 473 292 47.2 48.0Complete system (without substrate) plus 15 mm

ATP....................................... 578 338 -

Plus 2.0 mg of toxin .......................... 261 172 54.8 49.1Complete system including ascorbate and

TMPDd..................................... 1028 602Plus 2.0 mg of toxin .......................... 410 265 61.5 55.9

Complete system plus EDTA.................... 1065 668Plus 2.0 mg of toxin .......................... 879 550 17.5 17.6

Complete system including ascorbate and TMPDplus EDTA................................... 1365 803 - -Plus 2.0 mg oftoxin .......................... 1130 697 17.2 13.2

a This table was compiled from results given in reference 37.b Expressed as millimicromoles per milligram of protein.c Succinate was present as substrate.d Reaction mixture included ascorbate plus TMPD instead of succinate.

184 BACTERIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from



complex I showed that the toxin in no way alteredthe nonheme iron content of the purified NADH2-coenzyme Q reductase. Although the precisemanner in which the toxin inhibits NADH2-coenzyme Q reductase activity has not as yet beenelucidated, the investigations on the effect of thetoxin on the electron transport complexes suggestthat the toxin exerts its inhibitory effect on theelectron transport system by interfering with theenzymatic activity of NADH2-coenzyme Q re-ductase, thus preventing coenzyme Q from beingreduced.

Effect on mitochondrial ion accumulation. It hasbeen shown that isolated mitochondria bind andaccumulate Ki (20), Mg+,+ (10, 12,13), and Ca++(11, 46, 64, 74, 75). Data suggested (13, 14, 20,74) that alterations in the integrity of mitochon-dria may influence their ability to retain accumu-lated ions. Since plague murine toxin inducesswelling, studies were initiated concerning theeffect of the toxin on the accumulation of ions bymitochondria and the relationship between thiseffect and the ability of the toxin to induce mito-chondrial swelling (37). Such knowledge shouldhelp in obtaining a better understanding of howthe toxin influences myocardial physiology andeventually results in the death of a toxin-sus-ceptible animal, because the absence or over-abundance of ions results in abnormal states andreactions of the heart. It was found that the toxininhibited the accumulation of Ca++ and inorganicphosphate (Pi) by rat heart mitochondria in thepresence of succinate (Table 3) as well as a-keto-glutarate, malate, or f3-hydroxybutyrate as sub-strate. Since it is known that the toxin can inhibitthe oxidation of each of these compounds (62), itseemed conceivable that the toxin was preventingmitochondrial ion accumulation by interferingwith the respiratory chain. That this was not thecase was suggested by the finding that the toxininhibits the ATP-supported accumulation ofCa++ and Pi by rat heart mitochondria in theabsence of a respiratory substrate. Additionalevidence on this point stemmed from the factthat, although toxin had no effect on the respira-tion of rat heart mitochondria in the presence ofascorbate and TMPD (36, 37), it inhibited theuptake of Ca++ and Pi supported by this segmentof the electron transfer chain.The relationship between the toxin's ability to

induce swelling and to inhibit mitochondrial ionuptake was examined by incubating toxin-treatedmitochondria with ethylenediaminetetraaceticacid (EDTA) at a concentration of 0.1 mm. Thisconcentration of EDTA prevents swelling (70)without inhibiting ion uptake (75). The additionof EDTA largely prevented the inhibitory effectof the toxin on the accumulation of Ca+ and Pi

by rat heart mitochondria in the presence of arespiratory substrate as well as in the presence ofascorbate and TMPD (see Table 3). These datasuggested that EDTA, by preventing toxin-in-duced swelling, allowed mitochondria to retainions that were accumulated in the mitochondriallumen. However, it was also noted that EDTA-treated controls accumulated somewhat greateramounts of ions than untreated controls, indicat-ing that EDTA could exert a general stabilizingeffect on mitochondrial membranes.

DIsTRIBuTIoN AND SYNTHESIS OF PLAGUEMuRiNE ToxIN

Location of the Toxin in the P. pestis Cell

Part of the general problem of toxin synthesiswas the determination (55) of the location oftoxin in P. pestis cells. At least 10% of total toxicactivity was associated with the membrane frac-tions of spheroplasts prepared from whole cells;the remainder resided in the cytoplasmic fractions(Table 4).Ribosomes obtained by high-speed centrifuga-

tion of the cytoplasmic fraction contained lessthan 1% of total toxin and total protein, suggest-ing that the cytoplasmic toxin of P. pestis existsas a nonconjugated, nonparticulate protein.Membranes were disrupted by various means

to examine the relationship of the toxin to themembrane of the P. pestis cell. A membrane sus-pension subjected to sonic vibration increasedsignificantly the specific toxic activity of the pro-tein compared with the original suspension (Table4). Addition of magnesium ions to these suspen-sions protected the isolated membranes against

TABLE 4. Distribution and release from membranesof toxin and total protein in Pasteurella pestis

spheroplast fractions*

Inn Total iDso ProteinSpheroplast fraction dose doses, (perSpheroplast (pgg of (toxic cent of

protein) units) total)

Cytoplasm.50 1,260 72Membrane.162 152 28Sonically treated mem-branes.79 80 45

Control.83 22 12Sonically treated mem-branes plus MgCl2 83 20 15

Control.83 19 9Trypsin-treated mem-branes.109 40 60

Control.39 12 49


185VOL. 30, 1966


http://mm

br.asm.org/

Dow

nloaded from



disruption by sonic treatment. The LD50 of re-leased protein decreased approximately threefoldbelow the initial membrane-bound protein. Thesedata suggest that the potential toxic activity ofbound toxin cannot be adequately expressed untilthe toxin is solubilized.When isolated membranes were incubated with

trypsin, most of the protein released by trypsinaction was nontoxic (Table 4). Since trypsin didnot destroy the toxin, these findings suggestedthat the toxin may be bound in some mannerwhich makes it inaccessible to trypsin action.

Resolution and Isolation of Two Toxic ComponentsGel electrophoresis has been used in studying

protein components in crude cell fractions formetabolic studies (28, 78) and in determining thepurity of isolated proteins (50). While determin-ing feasibility of this method for locating the toxinin crude cell fractions of P. pestis, toxin activitywas observed to be associated with more than oneprotein component of the partially purified ma-terial. As demonstrated by the LD5o, increase inpurity of a toxin sample was found to be relateddirectly to a reduction in the total number of pro-tein bands detected by gel electrophoresis (57).The purest samples obtained exhibited two bands,and each was found to be toxic when eluted andinjected into mice.Each protein had an LD5o for 16- to 18-g Swiss

albino mice of 0.7 to 1.5 ,.g of protein and pro-duced a single characteristic precipitation bandwhen subjected to the gel diffusion precipitationtechnique. The slower migrating toxin, desig-nated as toxin A, yielded a concave-shaped pre-cipitin band, suggesting an antigen of greatermolecular weight than the antibody (33). Themore electrophoretically mobile toxin B, foundto be identical with the earlier isolated toxin witha molecular weight of 74,000 (3),* revealed astraight or slightly convex precipitin band, sug-gesting a smaller molecular weight band than theantibody.

Protein patterns from extracts of membraneand cytoplasmic fractions 'obtained from lysedspheroplasts showed toxin A to be associatedpredominantly with the membrane and toxin Bto be associated with cytoplasmic fraction. ToxinA is located in a different part of the P. pestiscell than toxin B, and toxin A appeared to be alarger molecular weight protein. This suggestedtwo toxic proteins of different molecular species.This hypothesis was supported by data indicating

* Recent re-examination of the molecular weightof toxin B by the Sephadex method has indicated alarger molecule of approximately 12,000 molecularweight.

differences in sensitivities of individual toxins toprotein reagents such as deoxycholate, digitonin,urea, and various sulfhydryl reagents (57, 58).On the other hand, it is speculated that both

toxins are possibly located in, and are part of, thesame particulate structure, namely, membranesin the native bacterial cell. It is proposed thatduring isolation, toxin B, located on the surface,splits off into the cytoplasmic fraction, whereastoxin A remains attached more strongly. It is alsopossible that toxin B is a part of the toxin A pro-tein in the membrane; during stress of isolation,toxin A disaggregates to form an essentially "newprotein," toxin B. This phenomenon was observedwith bovine growth hormone (47) and glutamicdehydrogenase (72). The relative similarity inspecific toxic activity of the two proteins suggestsa structural relationship.

Recent evidence (unpublished data) points tosome basic similarities in the amino acid contentand ultraviolet spectra of the two toxins. Alsothere is a strong suggestion that toxin A is twicethe molecular weight of toxin B, 240,000 and120,000, respectively. This would suggest thatthe sensitivities of the two toxins to proteinreagents is reflecting primarily diferences intertiary structure, and that toxin A is a dimer oftoxin B.

Taldng these data together, if the separation ofthe two toxins by cell location is not an artifactof isolation, then one could speculate that toxinB may be a precursor "polypeptide" of toxin Awhich is cemented into the membrane enzy-matically after its dimerization. On the otherhand, the possibility that toxin A is split in vivoto give two monomers of toxin B appears to beeliminated by results from some of the tryptophananalogue experiments subsequently discussed.

Selective Inhibition of Murine Toxin Synthesis byTryptophan Analogues

In addition to studies on the location of murinetoxin in P. pestis, experiments were also designedto determine mechanisms by which toxin synthe-sis is regulated. Toxin synthesis is selectivelyinhibited during growth of P. pestis at 37 C (55),and a number of metabolic inhibitors were ex-amined in an attempt to separate toxin from totalprotein synthesis. The utilization of tryptophananalogues proved most effective for this purpose(56). A number of investigators reported theselective action of tryptophan analogues on pro-tein synthesis (49, 63, 71). With these findings itbecame increasingly clear that tryptophan ana-logues may regulate protein biosynthesis in aspecific and selective manner in addition to theirrole in end products inhibition.

186 BAcrERioL. REV.


http://mm

br.asm.org/

Dow

nloaded from



100

z

w

Q

0w

FiG. 4. Se

toxin formati

When vaicubated wit]all cases, to)to the actioprotein fornand 6-metithan werehibiting toxi

A f-rn " ;is

acid completely reversed the selective effect ofSynthesis 4-methyltryptophan on toxin synthesis but onlyToxin

partially prevented growth inhibition. The in-yTonsis ability of anthranilic acid to reverse the inhibitory

effect of 4-methyltryptophan is puzzling, be-cause, from the data reported by Trudinger andCohen (73) on the synthesis of anthranilate inmutants of E. coli, one would expect anthranilateto be a better antagonist than shikimate.When tryptophan determinations were made on

protein from analogue-treated and untreated P.pestis cells, it was found that the tryptophan con-tent of the protein varied with the effectiveness ofthe inhibitors. Protein from cells treated with 4-methyltryptophan showed an 18% or more de-crease in tryptophan. The tryptophan content of

cells incubated with 4-methyltryptophan and in-4-MT 5-MT 6-MT 7-AZA 5-FT dole was significantly higher than that of cells

lective action of tryprophan analogues on treated with 4-methyltryptophan alone. Thision. Results from reference 56. paralleled the reversal of growth and toxin inhibi-

tion by indole.

rious tryptophan analogues were in- Toxin protein content in extractsfrom analogue-treated cells. The toxin content of cell extracts

i P.ntpestis cellsitwasfuhoud ti from analogue-treated cells, cells treated withoin synthesiwnalogs mcha sensite analogue plus indole, and control cells were com-

nation The methyl analogues (4gn 5e pared to demonstrate the quantitative decrease in

nation.Themwethyl aogues (4-,cte' toxm synthesis in analogue-treated cells. Ana-iyltryptophan) were more effective .a logue addition in experiments lasting 4 to 5 hr

and5ofluorotrynthan. 4 resulted in a relative reduction in toxin A. How-m and total protemi syntnesis (Fig. 4).

. ever, extracts from analogue-treated cells re-

F1IG.L4~i.SeUUdj

zintagunism ujy anaulgue uuctertustatic actlwnand the selective inhibition of toxin. The additionof L-tryptophan, at a concentration as low as 1.0,g/ml, to rapidly growing P. pestis cells waseffective in reversing inhibition of growth andtoxin synthesis by 4-methyltryptophan. Moyedand Friedman (59), using Escherichia coli, andAmes (6), using Salnonella typhimurium, reportedpermease competition between tryptophan andits analogues.Three intermediates in the tryptophan biosyn-

thetic pathway, i.e., indole, anthranilic acid, andshikimic acid, were used as antagonists of trypto-phan analogue action (Table 5) to avoid possiblepermease competition in P. pestis, and to obtainunequivocal evidence for the hypothesis that thetryptophan synthetic pathway is the site of ana-logue inhibition. With indole the total proteinand toxin inhibitory effects exhibited by 4- and5-methyltryptophan and 5-fluorotryptophan werepartially reversed by concentrations of indolelower than the analogue level. Anthranilic acid,however, produced a synergistic action in com-bination with 4-methyltryptophan; i.e., growth,total protein, and toxin were inhibited to a greaterdegree in P. pestis cells incubated with 4-methyl-tryptophan and anthranilic acid than in cellstreated only with 4-methyltryptophan. Shikimic

TABLE 5. Indole, anthranilic acid, and shikimic acidas antagonists of the action of tryptophan ana-logues on total protein and toxin synthesis*

Sample Protein LDno (jgg offormed protein)

mgControl ........................ 11.5 404-Methyltryptophan ............. 5.3 60-804-Methyltryptophan + indole... 8.9 40Control ........................ 5.9 305-Methyltryptophan............. 4.3 605-Methyltryptophan + indole 5.7 40Control ........................ 17.1 505-Fluorotryptophan............. 15.1 705-Fluorotryptophan + indole... 15.9 60Control ........................ 11.3 <404-Methyltryptophan............. 6.7 50Anthranilic acid ................ 12.1 <404-Methyltryptophan +

anthranilic acid............... 4.1 >60Control ........................ 9.9 <404-Methyltryptophan............. 4.7 60Shikimic acid................... 8.3 404-Methyltryptophan +

shikimic acid................. 9.9 40


VOL. 30, 1966 187


http://mm

br.asm.org/

Dow

nloaded from



versed by indole exhibited a high concentrationof toxin A, coinciding with the reversal of ana-logue inhibition. These results indicated that re-duced toxin activity initiated by metabolicchanges and detected by mouse assay is a result ofthe reduced amount of toxin A formed in ana-logue-treated cells. The possibility that the toxinsare related and that toxin B is a metabolic endproduct, accumulating in the cytoplasm after re-lease from the membrane, was suggested by in-vestigations of Montie and Ajl (55) and by Csanyiet al. (15) with mammalian systems. If the mem-brane is a precursor of the cytoplasmic toxin, atthe end of a 4- to 5-hr experiment, band B wouldbe composed of the initial cytoplasmic toxin andinitial membrane toxin transferred to the cyto-plasm. The membrane toxin, according to thistheory, would be composed of de novo synthe-sized toxin. Thus, longer term treatments be-ginning with small numbers of cells should resultin the depletion of toxin B.

Dilute suspensions of cells were exposed to 5-fluorotryptophan for 8- to 11-hr periods (ap-proximately 10 generations). This analogue wasused because it caused little or no growth inhibi-tion over a number of generations. When the ex-tracts were assayed for toxic activity, it was foundthat the extracts of the analogue-treated cells ex-hibited a three- to fourfold increase in the LDiO ascompared with that of the control cells. Toxin Awas completely absent from alkaline extracts oftreated cells, whereas toxin B was formed at ratescomparable with the control cells. These resultsappeared to negate the precursor hypothesis.

Inhibitory effect of analogues on spheroplasttoxin content. It was suggested (57) that toxin Ais associated with the membrane fraction of thecell and toxin B with the cytoplasmic fraction. Toconfirm these data, intact cells were converted tospheroplasts in the presence of tryptophan ana-logues. Low concentrations of methyl analogueswere employed to allow for complete conversionof whole cells to spheroplasts. When membraneand cytoplasmic fractions obtained from lysedspheroplasts were assayed for toxic activity, it wasfound that the membrane toxin was preferentiallyinhibited by the tryptophan analogues, whereasthe cytoplasmic or soluble toxin was inhibited tothe same degree as total protein. Consequently,these results present evidence confirming theidentity of membrane-associated toxin with thetryptophan analogue-sensitive toxin A.

SUMMARYDuring the past decade and a half or more,

studies have been carried out on the murine toxinof P. pestis. A series of chemical and electro-phoretic purification procedures yielded a most

highly purified, serologically homogenous toxin.The amino acid composition and immunologicalproperties of the purified toxin have been eluci-dated. Recent investigations revealed that this ma-terial consists of two distinct protein components,each of which possesses the same toxic activityand very similar amino acid content. Differencesbetween the two toxic proteins are seen in theirtertiary structure, molecular weight, cell location,and mode of biosynthesis. However, toxin Bappears to be one-half the molecular weight oftoxin A which suggests a monomer-dimer rela-tionship between the two proteins.

Initial investigations on the mechanism of ac-tion of plague murine toxin carried out with cell-free microbial extracts and crude tissue homoge-nates showed that a-keto acid oxidation isspecifically inhibited, and that this inhibition is re-versed with excess NAD but not with NADP. Themost significant findings, however, involved theeffect of the toxin on mammalian mitochondria.A correlation has been established between theability of the toxin to inhibit heart mitochondrialrespiration of certain animal species and the sus-ceptibility of these species to the in vivo action ofthe toxin.The inhibitory effect on mitochondrial respira-

tion is exerted on a segment of the electron trans-port system between NADH2 or succinate andcytochrome b and, more specifically, at the levelof NADH2-coenzyme Q reductase.

In addition, plague murine toxin induces mito-chondria to swell and curtails the accumulationof calcium and inorganic phosphate ions by thesestructures. A direct relationship between theseeffects has been established. By virtue of its abilityto alter the integrity of intact mitochondria, thetoxin does not allow the ions that have been ac-cumulated in the mitochondrial lumen to be re-tained.

Investigations on toxin biosynthesis revealedthat this process is inhibited by tryptophan ana-logues which inhibit tryptophan biosynthesis, andthat this inhibition is reversed by intermediates ofthe tryptophan biosynthetic pathway in micro-organisms, suggesting that tryptophan is requiredfor the biosynthesis of toxin.

ACKNOWLEDGMENTS

The early part of this work was done while one ofthe authors (S.J.A.) was a member of the WalterReed Army Institute of Research in Washington,D.C. This work has been continued at the AlbertEinstein Medical Center in Philadelphia. The bulk ofthe work described was supported by grant GB-2405from the National Science Foundation and by PublicHealth Service grant AI-03866 from the NationalInstitute of Allergy and Infectious Diseases.

188 BACTERIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from



LrTERATuRE Cmr

1. Ax, S. J., J. S. REEDAL, E. L. DURRUM, AND J.WARREN. 1955. Studies on plague. I. Purifica-tion and properties of the toxin of Pasteurellapestis. J. Bacteriol. 70:158-169.

2. Am, S. J., AND J. RuST. 1960. The biochemistryand physiology of the plague murine toxin.Ann. N.Y. Acad. Sci. 88:1152-1154.

3. Am, S. J., J. Rusr, JR., D. HUNTER, J. WOEBKE,AND D. F. BEwr. 1958. Preparation of serologi-cally homogeneous plaque murine toxin andits reactions with physical, chemical and en-zymatic agents. J. Immunol. 80:435-440.

4. Am, S. J., J. RUST, JR., J. WOEBKE, AND D. H.HUNTER. 1956. Action of plague toxin ondiphosphopyridine nucleotide. Federation Proc.15:581.

5. Am, S. J., J. WOEBKE, AND J. RUST, JR. 1958.Inhibition of keto acid oxidation by plaguetoxin. J. Bacteriol. 75:449-452.

6. AMEs, G. F. 1964. Uptake of amino acids bySalmonella typhimurium. Arch. Biochem. Bio-phys. 104:1-18.

7. BAKER, E. E., H. SOMMER, L. E. FOSTER, E. MEYER,AND K. F. MEYER. 1947. Antigenic structureof Pasteurella pestis and the isolation of acrystalline antigen. Proc. Soc. Exptl. Biol.Med. 64:139-141.

8. BAKER, E. E., H. SOMMER, L. E. FOSTER, E. MEYER,AND K. F. MEYER. 1952. Studies on immuniza-

tion against plague. I. The isolation and char-acterization ofthe soluble antigen of Pasteurellapestis. J. Immunol. 68:131-145.

9. BENT, D. F., H. ROsEN, S. M. LEVENSON, R. B.LINDBERG, AND S. J. AmL. 1957. Elemental andamino acid composition of purified plaguetoxin. Proc. Soc. Exptl. Biol. Med. 95:178-181.

10. BRIERLY, G. P., E. BACHMANN, AN) D. E. GREEN.1962. Active transport of inorganic phosphateand magnesium ions by beef heart mitochon-dria. Proc. Natl. Acad. Sci. U.S. 48:1928-1935.

11. BRIERLY, G. P., E. MuRER, AND E. BACHMANN.1964. Studies on ion transport. III. The ac-cumulation of calcium and inorganic phosphateby heart mitochondria. Arch. Biochem. Bio-phys. 105:89-102.

12. BRIERLY, G., E. MURER, E. BACHMANN, ANDD. E. GREEN. 1963. Studies on ion transport.II. The accumulation of inorganic phosphateand magnesium ions by heart mitochondria. J.Biol. Chem. 238:3482-3489.

13. BEuRLY, G. P., E. MmuR, Am D. E. GREEN.1963. Participation of an intermediate of oxida-tive phosphorylation in ion accumulation bymitochondria. Science 140:60-62.

14. CHAPPELL, J. B., AND G. D. GREVILLE. 1963. Theinfluence of the composition of the suspendingmedium on the properties of mitochondria.Biochem. Soc. Symp. (Cambridge, Engl.)23:39-65.

15. CsANYi, V., M. KRAMER, AND F. B. STRAUB.1960. Purification of the ribonucleic acid in-ducing penicillinase formation in Bacillus

cereus cells. Acta Physiol. Acad. Sci. Hung.18:171-178.

16. DmEuOmNNE, A., AND R. OrTO. 1928. Pest, p.179-412. In W. Kolle, R. Kraus, and P. Uhlen-huth [ed.], Handbuch der pathogenen mikro-organismen. VEB Gustav Fischer Verlag,Jena, Germany.

17. DURRUM, E. L. 1951. Two dimensional electro-phoresis and inophoresis. J. Colloid Sci. 6:274-290.

18. ENGLESBERG, E., AND J. B. LEw. 1954. Produc-tion of Pasteurella pestis toxin. J. Bacteriol.68:57-60.

19. FOWLER, L. R., S. H. RICHARDSON, AND Y.HATEri. 1962. A rapid method for the prepara-tion of highly purified cytochrome oxidase.Biochim. Biophys. Acta 64:170-173.

20. GAMBLE, J. L., JR. 1957. Potassium binding andoxidative phosphorylation in mitochondriaand submitochondrial fragments. J. Biol.Chem. 228:955-971.

21. GIRARD, G. 1939. Recherches sur la floculationdu serum antipesteux en presence de l'endo-toxine. Arch. Inst. Pasteur Madagascar, p. 14-16.

22. GOODNER, K., L. PANNELL, P. BARTELL, AND E. L.ROTHSTEIN. 1955. Toxic end products fromPasteurella pestis. I. A comparison of lysatetoxin with that obtained from the action of bilesalts. J. Infect. Diseases 96:82-87.

23. GRIFF1THS, D. E., AND D. C. WHARTON. 1961.Studies on the electron transport system.XXXV. Purification and properties of cyto-chrome oxidase. J. Biol. Chem. 236:1850-1856.

24. HANDLER, P., AND J. R. KLEIN. 1942. The in-activation of pyridine nucleotides by animaltissues in vitro. J. Biol. Chem. 143:49-57.

25. HATEFI, Y., A. G. HAAvIc, L. R. FowLER, ANDD. E. GRITHs. 1962. Studies on the electrontransfer system. XLII. Reconstitution of theelectron transfer system. J. Biol. Chem. 237:2661-2669.

26. HATE, Y., A. G. HAAVIK, AND D. E. GRwrrHs.1962. Studies on the electron transfer system.XL. Preparation and properties of mitochon-drial DPNH-coenzyme Q reductase. J. Biol.Chem. 237:1676-1680.

27. HATEri, Y. A., G. HAAVIK, AND D. E. GRIFFITHS.1962. Studies on the electron transfer system.XLI. Reduced coenzyme Q (QH2)-cytochromec reductase. J. Biol. Chem. 237:1681-1685.

28. HENNING, U., AND G. YANOFSKY. 1963. An elec-trophoretic study of mutationally altered Aproteins of the tryptophan synthetase of Esch-erichia coli. J. Mol. Biol. 6:16-21.

29. HOWLAND, J. L. 1963. Phosphorylation coupledto the oxidation of tetramethyl-p-phenylene-diamine in rat-liver mitochondria. Biochim.Biophys. Acta 77:419-429.

30. HUNTER, F. E., JR., J. F. LEvY, J. FINK,B. ScHUTz, F. GUERRA, AND A. HuRwrrz.1959. Studies on the mechanism by whichanaerobiosis prevents swelling of mitochondriain vitro: effect of electron transport chain in-hibitors. J. Biol. Chem. 234:2176-2186.

189VOL. 30, 1966


http://mm

br.asm.org/

Dow

nloaded from



31. JAWETZ, E., AND K. F. MEYER. 1943. Avirulentstrains of Pasteurella pestis. J. Infect. Diseases73:124-143.

32. JENSEN, J. 1960. Continuous electrophoresis inglass bead media. M.S. Thesis. Lehigh Uni-versity, Bethlehem, Pa.

33. KABAT, E. A., AND M. M. MAYER. 1961. Experi-mental immunochemistry, 2nd ed., p. 86.Charles C Thomas, Publisher, Springfield, Ill.

34. KADIS, S., AND S. J. AmL. 1963. Mitochondrialswelling induced by plague murine toxin. J.Biol. Chem. 238:3472-3477.

35. KADis, S., S. J. AJL, AND J. H. RUST, JR. 1963.Action of plague murine toxin on mitochondriafrom resistant and susceptible animals. J.Bacteriol. 86:757-765.

36. KADIs, S., M. COHEN, AND S. J. AJL. 1965. Theeffect of plague murine toxin on the electrontransport system. Biochim. Biophys. Acta 96:179-186.

37. KADis, S., A. TRENCHARD, AND S. J. AJL. 1965.Effect of plague murine toxin on the uptake ofcalcium and inorganic phosphate ions by heartmitochondria. Arch. Biochem. Biophys. 109:272-278.

38. KAPLAN, N. O., S. P. COLOWICK, AND A. NASON.1951. Neurospora diphosphopyridine nucleo-tidase. J. Biol. Chem. 191:473-483.

39. KARLER, A. 1955. New horizontal curtain elec-trochromatographic apparatus for both paperstrip and continuous flow electrophoresis.Federation Proc. 14:233.

40. KING, R. M. 1959. Continuous electrophoresisin glass bead media. M.S. Thesis, Lehigh Uni-versity, Behlehem, Pa.

41. KORNBERG, A., AND W. E. PRICER, JR. 1950.Nucleotide pyrophosphatase. J. Biol. Chem.182:763-778.

42. LARDY, H. A., D. JOHNSON, AND W. C. MCMUR-RAY. 1958. Antibiotics as tools for metabolicstudies. I. A survey of toxic antibiotics inrespiratory, phosphorylative and glycolyticsystems. Arch. Biochem. Biophys. 78:587-597.

43. LEHNINGER, A. L. 1959. Reversal of various typesof mitochondrial swelling by adenosine tri-phosphate. J. Biol. Chem. 234:2465-2471.

44. LEHNINGER, A. L. 1962. Water uptake and ex-trusion by mitochondria in relation to oxida-tive phosphorylation. Physiol. Rev. 42:467-517.

45. LEHNINGER, A. L., AND B. L. RAY. 1957. Oxida-tion-reduction state of rat liver mitochondriaand the action of thyroxine. Biochim. Biophys.Acta 26:643-644.

46. LEHNINGER, A. L., C. S. Rossi, AND J. W. GREEN-AWALT. 1963. Respiration-dependent accumu-lation of inorganic phosphate and Ca++.Biochem. Biophys. Res. Commun. 10:444-448.

47. LEwis, U. J. 1962. Enzymatic transformationsof growth hormone and prolactin. J. Biol.Chem. 237:3141-3145.

48. LUSTIG, A., AND G. GALEoTri. 1897. Versuchemit Pestschutzimpfungen bei Thieren. Deut.Med. Wochschr. 23:227-230.

49. MACH, B., E. REICH, AND E. L. TATUM. 1963.

Separation of the biosynthesis of the antibioticpolypeptide tyrocidine from protein biosyn-thesis. Proc. Natl. Acad. Sci. U.S. 50:175-181.

50. MAIZEL, J. V. 1963. Evidence for multiple compo-nents in the structural protein of type 1 polio-virus. Biochem. Biophys. Res. Commun. 13:483-489.

51. MARKL, G. 1898. Beitrag zur Kenntnis der Pest-toxine. Zentr. Bakteriol. Parasitenk. 24:641-649.

52. MARKL, G. 1900. Ueber die Pesttoxine und dieGewinnung von antitoxischen Pestserum. Wien.Med. Wochschr. 50:2412-2414.

53. MCCRUMB, F. R., JR., S. MERCIER, J. ROBIC, M.BOUILLAT, J. E. SMADEL, T. E. WOODWARD,AND K. GOODNER. 1953. Chloramphenicol andterramycin in the treatment of pneumonicplague. Am. J. Med. 14:284-293.

54. MINAKAMI, S., R. L. RINGLER, AND T. P. SINGER.1962. Studies on the respiratory chain-linkeddihydrodiphosphopyridine nucleotide dehydro-genase. J. Biol. Chem. 237:569-576.

55. MoNTEE, T. C., AND S. J. Am. 1964. The anatomi-cal distribution of murine toxin in spheroplastsof Pasteurella pestis. J. Gen. Microbiol. 34-249-258.

56. MoNTiE, T. C., AND S. J. AJL. 1964. Selectiveinhibition by tryptophan analogues of murinetoxin synthesis in Pasteurella pestis. J. Bacteriol.88:1467-1475.

57. MONTIE, T. C., D. B. MONTIE, AND S. J. AJL.1964. The identification and isolation of twomouse-toxic protein components in extractsfrom Pasteurella pestis. J. Exptl. Med. 120:1201-1212.

58. MoNTIE, T. C., D. B. MONTEE, AND S. J. AJL. 1965.Studies on two mouse-toxic proteins from Pas-teurella pestis. Federation Proc. 24:419.

59. MoYED, H. S., AND M. FRIEDMAN. 1959. Alteredactive transport: a basis for resistance to anti-metabolites. Bacteriol. Proc., p. 107.

60. OUDIN, J. 1948. L'analyse immunochimiquequalitative: methode par diffusion des antigensau seinde l'immunserum precipitant gelose.Ann. Inst. Pasteur 75:30-51.

61. OUCHTERLONY, 0. 1949. In vitro method fortesting the toxin-producing capacity of diph-theria bacteria. Acta Pathol. Microbiol. Scand.26:516-524.

62. PACKER, L., J. H. RUST, JR., AND S. J. AJL. 1959.Action of plague murine toxin on mammalianmitochondrial respiration. J. Bacteriol. 78:658-663.

63. PARDEE, A. B., AND L. S. PRESTIDGE. 1958. Ef-fects of azatryptophan on bacterial enzymesand bacteriophage. Biochim. Biophys. Acta27:330-344.

64. RossI, C. S., AND A. L. LEHNINGER. 1963. Stoi-chiometric relationships between mitochondrialion accumulation and oxidative phosphoryla-tion. Biochem. Biophys. Res. Commun. 11:441-446.

65. Rossi, C. S., AND A. L. LEHNINGER. 1963. Stoichio-metric relationships between accumulation of

190 BACTERIOL. REV.


http://mm

br.asm.org/

Dow

nloaded from



ions by mitochondria and the energy-couplingsites in the respiratory chain. Biochem. Z. 338:698-713.

66. RUST, J. H., JR., D. C. CAVANAUGH, S. KADIS,AND S. J. AmL. 1963. Plague toxin: its effect invitro and in vivo. Science 142:408-409.

67. RUST, J. H., JR., A. F. GOLEY, H. J. BAKER, ANDS. J. AJL. 1959. Further studies on the in vivoand in vitro action of plague toxin. Bacteriol.Proc., p. 95.

68. SMITH, H., J. KEPPrE, E. C. COCKING, AND K.WITT. 1960. The chemical basis of the virulenceof Pasteurella pestis. I. The isolation and theaggressive properties of Past. pestis and itsproducts from infected guinea-pigs. Brit. J.Exptl. Pathol. 41:452-459.

69. SPIVACK, M. L., AND A. KARLER. 1958. Purifica-tion of the toxin of Pasteurella pestis by con-tinuous-flow paper electrophoresis. J. Immunol.80:441-445.

70. TAPLEY, D. F. 1956. The effect of thyroxine andother substances on the swelling of isolatedrat liver mitochondria. J. Biol. Chem. 222:325-339.

71. THANG, M. N., F. R. WILLIAMS, AND M. GRUN-BERG-MANAGO. 1963. Synthese in vivo de lapolynucleotide phosphorylase chez Escherichiacoli. II. Synthese de novo de la polynucleotidephosphorylase en presence de chloramphenicol.Biochim. Biophys. Acta 76 :572-588.

72. TOMPKINS, G. M., K. L. YIELDING, N. TALAL,AND J. F. CURRAN. 1963. Protein structure andbiological regulation. Cold Spring HarborSymp. Quant. Biol. 28:461-471.

73. TRUDINGER, P. A., AND G. N. COHEN. 1956. Theeffect of 4-methyltryptophan on growth andenzyme systems of Escherichia coli. Biochem.J. 62:488-491.

74. VASINGTON, F. D. 1963. Ca++ uptake in fragmentsof rat liver mitochondria and its dependenceon electron transport. J. Biol. Chem. 238:1841-1847.

75. VASINGTON, F. D., AND J. V. MURPHY. 1962.Ca++ uptake by rat kidney mitochondria andits dependence on respiration and phosphoryla-tion. J. Biol. Chem. 237:2670-2677.

76. WARREN, J., U. WALz, J. S. REEDAL, AND S. J.Ax. 1955. Studies on plague. II. Immunologi-cal properties of purified Pasteurella pestistoxin. J. Bacteriol. 70:170-176.

77. WINsTEN, S., H. FRIEDMAN, AND E. E. SCHWARTZ.1963. Large-volume continuous-flow electro-phoresis of serum proteins with glass micro-beads. Anal. Biochem. 6:404-414.

78. ZELDIN, M. H., AND J. M. WARD. 1963. Acryla-mide electrophoresis and protein pattern duringmorphogenesis in a slime mould. Nature 198:389-390.

79. ZHELTENKOV, A. I. 1946. Plague microbe toxinand the antitoxic antiplague vaccines. Zh.Mikrobiol. Epidemiol. i Immunobiol. 3:81-82.

80. ZIEGLER, D. M., AND K. A. Doun. 1962. Studieson the electron transport system. XLIII. Theisolation of a succinic-coenzyme Q reductasefrom beef heart mitochondria. Arch. Biochem.Biophys. 97:41-50.

VOL. 30, 1966 191


http://mm

br.asm.org/

Dow

nloaded from


Documents

mmbr.asm.org · KADIS, MONTIE, ANDAJL phenomenon is also involved in the inhibitory effect ofthe toxin on mitochondrial respiration. Thisis correlatedwiththetoxin'sabilitytoinduce