Internucleosomal DNA Cleavage Precedes

Vol . 264. No. 26, Issue of September 15, pp. 15261-15267, 1989 Printed in U.S.A.

CHEMISTRY THEJ O U R N A L OF BIOLOGICAL 0 1989 by The American Society for Biochemistry and Molecular Biology, Inc.

Internucleosomal DNA Cleavage PrecedesDiphtheria Toxin-induced Cytolysis EVIDENCE THAT CELLLYSIS

ISNOT

A SIMPLE CONSEQUENCE OFTRANSLATION

INHIBITION*

(Received for publication, April 21, 1989)

Michael P. Chang$$!l,John BramhallQII, Scott Graves$**, Benjamin Bonavida$ll, and BernadineJ. WisnieskiS 11 $$ From the $Departmentof Microbiology and the Molecular Biology Institute, the §Departmentof Microbiology and Immunology, and the IlJonsson Comprehensive Cancer Center, The University of California, Los Angeles, California90024

Diphtheria toxin (DTx) is an extremely potent inhib- Diphtheria toxin (DTx)’ is a potent cytotoxin for a wide itor of protein synthesis. Cell death hasbeen generally range of eukaryotic cells. The best-characterized activity of accepted as a straightforward effectof translation in- DTx is its ability to inhibit target cell protein synthesis, and hibition. Using human U937 cells, we found that DTx it has been estimated that a single molecule of DTx is suffiintoxication leads to cytolysis; indeed, release of “Cr- cient to kill a susceptible cell (1).DTx is secreted by Coryand ‘%e-labeled proteins could be detectedwithin 7 h. nebacterium diphtheriae as a single polypeptide chain (Mr = However, little orno cell lysis was observed over a 2060,525); subsequent cleavage generates two subunits that re50-h period when human U937 cells were exposed to cycloheximide, amino acid-deficient medium, or met- main connected by a disulfide bond (2). The A subunit (Mr = abolic poisons even though protein synthesis was rap- 20,500) inhibits protein synthesis by catalyzing the transfer idly inhibited to levels observed withDTx. Likewise, of the ADP-ribose moiety of NAD to elongation factor 2 (EFinvestigations with human K562 cells revealed full 2). The B subunit (Mr = 40,000) is the receptor-binding resistance to the cytolytic action of DTx over a 50-h domain of the toxin and hasno known enzymic activity (3). period despite a severe reduction in translation activAlthough proteinsynthesisinhibitioncan be monitored ity. These observations establish that inhibition of pro- precisely and with complete objectivity, target cell death is tein synthesis per se is not sufficient to provoke cell much more problematical and can be variously monitored as lysis. A characterization of DTx-induced cytolysis revealed a longlag period (6-7 h) which could be short- cessation of cell replication, leakage of cytoplasmic solutes, enedconsiderablyby a shortexposureto low pH. loss of the ability to exclude certain water-soluble dyes, loss NH,Cl and metabolic poisons blocked the cytolytic ac- of the ability to take up nutrients, or, in the case of specific tion of DTx, indicating that endocytic uptake of toxin effector cells, a decline of function. It is not at all clear that is required for lytic activity. Surprisingly, DTx also these parameters are equivalent to each other or even muinduced extensive internucleosomal degradation of cel- tually consistent. In studies of immune cytolysis, cell death is lular DNA, a characteristic feature of apoptosis or usually operationally defined as loss of the ability to retain programmed cell death. DNA-fragmentation preceded cytosolic marker solutes suchas those radiolabeled with 51Cr. cell lysis and did not occur in DTx-treated K562cells In a previous report (4)we noted that DTx provokes W r or in U937 cells that were treated with the other protein synthesis inhibitors. From these observations, release from target cells and that thetime course of killing is we conclude that DTx-mediated cytolysis is not a sim- similar to thatobserved with tumor necrosis factor-a (TNF), ple consequence of translation inhibition and that in- an agent which does not diminish translation activity (5). The currentstudy was undertaken toaddress the question ternucleosomal DNA fragmentation is a newly identified and relatively early step in the cytolytic pathway of whether the observed cytolytic response to DTx is simply of DTx. or an ipso facto hallmark of proteinsynthesisinhibition, whether aTNF-like effect is occurring, characterized by target cell DNA damage independent of translation inhibition. By using a cell line in which DTx provokes translation inhibition but not lysis we were able to demonstrate that the cytolytic * This work was supported in part by Grants CA-43121 (to J. B.), activity of DTx is strictly correlatedwith extensive (prelytic) observed during“proCA-35791 (to B. B.), and GM-22240 (to B. J. W.) from the National DNA fragmentation akintothat Institutes of Health and by grants from the California Cancer Re- grammed cell death” or “apoptosis.” This autolytic pathway search Committee and the UCLAAcademic Senate. The costs of has been shown to be induced by a number of unrelated publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduer- agents, including T N F (6, 7), lymphotoxin (8), cytotoxic Ttisernent” in accordance with 18 U.S.C. Section 1734 solelyto indicate lymphocytes (9), natural killer cells (lo), ionizing radiation this fact. (111,and glucocorticoids (12). Our findings lead us toconclude ?l Recipient of United States Public Health Service/National Re- that the ability of DTx to provoke target cell lysis involves a search Service Award Grant CA-09056. Present address: Dept. of Pharmacy, University of California at SanFrancisco, 513 Parnassus new killing activity. Ave., San Francisco, CA 94143. **Recipient of National Cancer InstituteTraining Grant CA09120. $$ TOwhom correspondence should be addressed Dept. of Microbiology, University of California, Los Angeles, CA 90024.

The abbreviations used are: DTx, diphtheria toxin; TNF, tumor necrosis factor-a; RPMI, tissue culture medium; BCS, bovine calf serum; EF-2, elongation factor 2; NaN3/2-deoxyglucose,sodium azide in combination with 2-deoxyglucose;NH4Cl, ammonium chloride.

15261

15262

The Cytolytic Pathway of Diphtheria Toxin EXPERIMENTALPROCEDURES

Celk"U937, RAJI, YAG-1, and K562 target cells (obtained from ATCC) were all maintained in suspension culture (37 "C, 5% CO,, 95% air) in RPMI-1640 medium supplemented with nonessential amino acids, sodium pyruvate, L-glutamine, sodium bicarbonate, and 10% bovine calf serum (RPMI + 10% BCS); PA-1 was grown in the same medium as an adherent culture. U937, RAJI, K562, and PA-1 are all human cell lines; YAC-1 is murine. K562 cells obtained from ATCC in 1985 were susceptible to DTx-mediated cytolysis in contrast to the K562 cells obtained in January 1989 and used in this study. Cells and/or karyotype data are available upon request. Reagents-DTxwas supplied by List Biological Laboratories (Campbell, CA) as a lyophilized powder; after reconstitution it was stored at a protein concentration of 1 mg/ml at -80 "C. As supplied, the preparation was 2 kDa) acid-precipitable 75Selabeled cell proteins (data not shown). The lag period preceding cytolysis could be shortened by 35 h by maintaining acidic conditions during the initial exposure of target cells to DTx (pH5.3; 15 min). Moreover, acidpulsed cells showed enhanced susceptibility to DTx-triggered lysis throughout a 20-h assay (data notshown). Cell Lysis Is Not aSimple Consequence of Protein Synthesis Inhibition-We have observed that theK562 cell line is fully resistant to the cytolytic action of DTx. The basis for this resistance is not explained by a block in the intoxication process, since protein synthesis is severely inhibited by DTx.

CytOlYSiS %"

14.2 f 0.1 5.0 f 1.4 14.8 f 1.5 77.5 f 5.1

~

Net release of 51Cr measured after a 20-h incubation at 37 "C. RPMI medium deficient in leucine, lysine, methionine, and glutamine.

conditions demonstrated that cell lysis was not an automatic consequence of inhibition of protein synthesis activity. Table I1 shows that exposure of U937 cells to cycloheximide (250 ng/ml) at aconcentration that caused greater than 85% decrease in protein synthesis activity (measured by the incorporation of [3H]leucineinto acid-precipitable material) caused less than 10% release of trapped 5'Cr during the 20-h incubation period. Cytolysis was observed, however, when DTx was added to these cycloheximide-treated cells, suggesting that cycloheximide treatment did not decrease the intrinsic ability of target cells to be lysed (data not shown). Similarly, when U937 cells were maintained in medium deficient in several essential aminoacids (leucine, lysine, methionine, and glutamine), protein synthesis was inhibited by greater than 97% whereas only 14% cytolysis was recorded (Table 11). Even when cells were poisoned metabolically with levels of sodium azide/2-deoxyglucose that led to 95% protein synthesis inhibition, only 5% cytolysis was observed. In sharp contrast, Table I1 shows that DTx (75 ng/ml) caused 85% decrease in protein synthesis and 78% cytolysis during the same standard 20-h incubation period. Theseresults raised the possibility that the cytolytic activity of DTx may not be a direct consequence of inhibition of protein synthesis. Kinetics of Cell Lysis and Translation Inhibition-Parallel experiments were performed to analyze the kinetics of DTxmediated cytolysis and intoxication. Fig. 2, A and B shows that DTx-mediated cytolysis occurred much later than inhibition of protein synthesis, with half-maximal points at 10 and 3.5 h, respectively. A long lag period (6 h) preceded the sudden onset of marker release whereas inhibition of translation was apparent even at the earliest time points, demonstrating that intoxication began almost immediately after

z 20

0

0

10

30

20

TIME

40

o

50

(nr)

6

L 0

5

""-

b = 10

-

-

"

"

/ 15

20

TIME (hr)

FIG. 2. Kinetics of target cell lysis ( A ) or protein synthesis inhibition (B).A, 51Cr-labeledU937 cells were distributed in Falcon tubes (1 X lo6 cells/tube in 20 ml of medium) and exposed at time zero to DTx (300 ng/ml; filled circles), NaN3/2-deoxyglucose (NaN3/ 2-DOG) (2.5/5 mM; filled squares) or cycloheximide (1 pg/ml; open circles). 'lCr-Labeled K562 cells were also examined with an identical concentration of DTx (open squares). Percent cell lysis was determined at specified intervals over a 50-h period as described under "Experimental Procedures." Data represent mean and standarderror of triplicate determinations. B, time course of protein synthesis inhibition in U937 cells exposed to DTx (filled circles), NaN3/2deoxyglucose (filled squares) or cycloheximide (open circles). DTxtreated K562 cells (open squares)were monitored in parallel. Reagents were added at concentrations specified in panel A , and translation activity was determined as described under "Experimental Procedures."

The Cytolytic Pathway

15264

Even at extremely high concentrations of toxin, where translation activity was completely inhibited (>99%), no cytolysis was observed over a 20-h period (Fig. 3). To test whether resistance to lysiscould be attributed to a slower rate of intoxication, we monitored the kinetics of DTx-induced protein synthesis inhibition in both K562 and U937 cells. Fig. 2B shows that translation inhibition occurred significantly later in K562 cells as compared with U937 cells, with halfmaximal points at8 and 3.5 h, respectively. However, continuous exposure of K562 cells to a high dose of DTx (300 ng/ ml) for 50 h failed to elicitacytolyticresponse (Fig. 2 A ) despite the severe inhibition of translation activity observed throughoutthis period (datanot shown). Of considerable interest is the fact that DTx-lysis-resistant K562 cells were found to be cross-resistant to the cytolytic action of TNF, a cytokine with no effect on protein synthesis activity (5).This cell line, however, can be readily lysed by natural killer cells (15) and antibody and complement (16). Our findings with K562 target cells establish a striking lack in correlation between the two activities of DTx; cells that were fully sensitive to DTx intoxication completely resisted cytolysis (Fig. 3) in distinct contrast to the pattern observed with U937 target cells (Fig. 1).What clearly emerges is that protein synthesis inhibition alone is not sufficient to trigger cell lysis. Hence the cytotoxic pathway of DTx appears to include a new but as yetundefined lysis triggering activity. Internalization of Cytolytically Active DTx Molecules Is Required for Killing Activity-The conventional route of entry to the cytoplasm for DTx, namely, receptor-mediated endocytosis followed by escape from the endosomal compartment, shows an absolute requirementfor acidic endosomes (17-20) and canbe inhibited by depletion of cellular ATP levels (21). Fig. 4 shows that DTx-induced cytolysis was only observed under conditions that allowed for exposure of the toxin to a low pH environment. In the presence of 10 mM NH4C1, a potent inhibitor of endosomal acidification (22), U937 target cells were completely resistant to thelytic effectsof DTx, and this resistance was readily overcome when the pH of the external medium was reduced to 5.3. Fig. 5 shows that DTxinduced cytolysis was alsosharply inhibited by the metabolic poisons NaNs/2-deoxyglucose; pretreatment of 51Cr-labeled U937 cells with a combination of the two energypoisons resulted in >75% inhibition of DTx-mediated cytolysis. This finding strengthens earlierproposals that cellular ATP is needed in order toeffect DTx delivery to thecytosol (21,231. I

I

'

I 0

.75

7.5

75

750 7500

TOXIN ( n g / r n l )

FIG. 3. DTx inhibits translation activity in K562 cells but does not induce lysis. DTx-mediated cytolysis andtranslation inhibition were monitoredin K562 cells after a20-hincubation. Methods were identical to Fig. 1.

of Diphtheria Toxin

2ol

- 0

I

8

40

I

I

80

60

100

TOXIN (og/rnl) ?NH,CI tDTx I

ApH

pH7.4

I

I

o Z O m5 imni n

ASSAY I 20 h

"

TIME FIG. 4. Inhibition of DTx-mediated cytolysis by lysosomotropic agents can be overcome by exposure to low pH. "CrLoaded U937 cells were exposed to DTx at pH 7.4 either in the presence (filled triangles) or the absence (open triangles) of 10 mM NH4C1. After 5 min at 37 "C, the pH of the medium was lowered to 5.3 in some of the samples, either with (filled circles)or without(open circles) NH4Cl.After 15 min,allsamples were washedwith fresh medium and incubated at 37 "C in RPMI + 2.5% BCS, pH 7.4, either with or without NH4C1for 20 h.

cn 40

3

3 20 w 0

0

A 0

C D

FIG. 5. DTx-mediated cytolysis of U937 cells is energy dependent. 51Cr-Labeled U937 cells were incubated in RPMI + 10% BCS containing2.5 mM NaN3 in combination with 5 mM 2-deoxyglucose ( B and D)or in medium alone ( A and C) for 30 min prior to the addition of DTx to 40 ng/ml ( A and B ) or 75 ng/ml (C and D). Incubation was continued for 20 h at 37 "C, and the extent of "Cr release from target cells was measured.

Taken together, thesetwo observations stronglysuggest that the cytolytic activity of DTx, much like its translation inhibition activity, is normally dependent on internalization of toxin molecules and thatcell-surface-bound toxin itself is not sufficient to signal the cytolytic response. DNA Fragmentation Is a Specific Consequence of DTx Exposure-When the integrity of cellular DNA was monitored 18 h after DTx treatment, significant levels of DNA degradation were observed. Fig. 6 shows the fragmentation profile size band being has a laddered appearance with the minimum -200 base pairs. Cells treated with cycloheximide (Fig. 6) and other agents/conditions thatled to inhibition of protein synthesis failed to display significant levels of DNA degradation over thesame 18-h time period. When we examinedthe integrity of cellular DNA from K562 cells (resistant to DTxmediated lysis), no detectable DNA fragmentation was ob-

The Cytolytic Pathwayof Diphtheria Toxin

15265

1 2 3 4

t

I ‘“ 401

692 521 284 257 226 /-

--

_J

A W

-

0

20 -

2

6

FIG.6. DTx-induced DNA fragmentation. HumanU937 cells were treated with (lane 2) or without DTx (lane I), or incubated with cycloheximide (lune 3) or NaN3/2-DOG (lane 4 ) for 18 h at p H 7.4, 37 “C. Toxin was added a t 75 ng/ml (200 pl/well), cycloheximide a t 250 ng/ml, and NaN3/2-deoxyglucose a t 2.5/5 mM. Molecular weight markers are indicated in basepairs.

served in the presence of toxin (data not shown). We next determined the onsetof DNA damage by examining theDNA profile of U937 target cells a t hourly intervalsafterDTx treatment. Fig. 7 shows that DNA fragmentation begins at -6 h. Parallel experiments were performed to determine the kinetics of DTx-mediated lysis under conditions identical to those employed in Fig. 7. Thus, Fig. 8, a time course of 51Cr release from human U937 cells, shows that cell lysis began -8-9 h after addition of toxin and reached a maximum a t 16 h. Essentially identical results were obtained when cell death wasassayedby trypan blue staining (data not shown). In order toprove that theDNA fragmentation observed a t early time points did not arise from a small fraction of cells that may have undergone cytolysis, DTx-treated cells were separated into intact and leaky populations a t hourly intervals, and the intact (trypan-blue excluding) cell population was examined for DNA damage. Fig. 9 shows that the onset time of DNA fragmentation in cells harvested by this procedure closely matched the onset timeobserved in Fig. 7, supporting the conclusion that DNA breakdown has occurred in nonlysed cells and is not the consequence of cell death. These results establish thatDNA fragmentation is an“early” event in DTx-mediated cytolysis, preceding lysis by 2-3 h. Hence,

14

18

FIG. 8. KineticsofDTx-mediatedcytolysis. 51Cr-Labeled U937 cells were exposed to DTx (75 ng/ml) for 18 h a t 37 “C under conditions identical to those employed in Fig. 7. At hourly intervals, aliquots were removed and assayed for“Cr release. Spontaneous release from toxin-freesamples was measuredat hourly intervals and subtracted as reported under “Experimental Procedures.”

S C

FIG. 7. Time course of DTx-mediated DNA fragmentation. The first three lanes contain the following molecular weight markers: PstI-digested X-phage DNA;Alu-digested pUC8 DNA; HueIII-digested pUC8 DNA. Lanes 1-18 represent DNA extracted from DTxtreated cells (75 ng/ml toxin) after 1-18 h of incubation as described under “Experimental Procedures.” The sample shown in lune C is from untreated cells.

10

TIME (hours)

45678910

FIG.9. Kinetics of DTX-triggered DNA fragmentation in nonlysed targ. cells. U937 cells exposed to DTx (75 ng/ml) were separated on F , Yypaque a t hourly intervals into lysed and non~ n e s4-20 represent DNA extractedfromnonlysed populat’ lysed cells r* ’ . : 0 h of incubationwith DTx. Lanes S and C represent m ’ . . . weight markers(HaeIII-digestedpUC8DNA) espectively. and untreatt .! ’

we conclud,. identified m

’

DNA fragmentation represents a newly in the cytolytic pathway of DTx. DISCUSSION

Averylarge body of researchexperience supportsthe concept that the cytotoxic activity of DTx results from the capacity of the toxin todeliver an enzymically active subunit (A) into the target cell cytoplasm and then to effect inhibition of ribosomal activity by ADP-ribosylation of EF-2 (24).It has been generally assumed that this scenario fully explains the killing mechanism of DTx. Thedatapresentedinthisreport allow two surprising conclusions to be drawn. First, DTx-triggered cytolysis is not a simple or straightforward consequence of translation inhibition. Second, the loss of cellular integrity is preceded by internucleosomal DNA fragmentation, a discovery that raises the possibility that DTx destroys target cells by a novel action. As shown in Table I, DTx triggers the lysis of a variety of target cells. This lysis, measured by the release of ”Cr or ”Selabeled peptides, is not particularly rapid (half-time of several hours). It has been well documented (17-20, 25, 26) that in

15266

The Cytolytic Pathway of Diphtheria Toxin

order for DTx to inhibit translation activity in susceptible matin breakdown or any of the morphological changes assocells, the toxin must be exposed to a low pH environment, ciated with programmed cell death. Thus, a mechanism for presumablysuppliedby the endosomal compartment. Our DTx-mediated lysis based solely on direct pore formation results show that DTx-mediated cytolysis is also dependent would notaccount for thecriticalchangesintarget cell on acid conditions since NH&l had an inhibitory effect on integrity observed in this report. These facts coupled with our the lytic activity of DTx, and this inhibition could be over- recent findings with CRM197, amutant formof DTx inwhich come by briefly lowering the external pH(Fig. 4). a single amino acid substitution in the A domain renders it In direct contrast to results obtained with DTx-treated inactive as an ADPr-transferase yet fully active in membrane cells, rapid and complete inhibition of target cell protein channel assays,’ lead us to favor a mechanismfortoxinsynthesis by treatment withcycloheximide, metabolic poison- induced cytolysisthat involves a relatively slow accumulation ing with NaNa/2-deoxyglucose, or incubation in media defi- of membrane damage causedby as yet tobe defined intracelcientinessentialamino acidsfailed to evoke comparable lular events, rather than the direct formation of toxin lesions levels of cytolysis (Table 11). Clearly, cells whose protein within the targetcell boundary membrane. synthesis has been irreversibly arrested will ultimately lyse, The killing mechanism we currently favor is the existence but our observations establish that the onset time of lysis of a second messenger substrate which is tightly coupled to under these conditions is considerablylonger than 6-7 h. the activation of an endogenous endonuclease. DTx-induced Indeed, little lysis is seen before 20-50 h (Fig. 2 4 ) . The data modification (e.g. ADP-ribosylation) of such a substrate could we have obtained with the K562 cell line further demonstrates trigger the cell suicide response. In speculation, this pathway severe inhibition of translation activity without the inductionmight involve the direct activationof a constitutive endonuof cell lysis (Fig. 2, A and B ) . Although the reason is unclear clease, the increased transport of critical ions such as Ca2+ as to why the K562 cell line is resistant to thelytic action of into thenucleus, or themobilization of essential cofactors for DTx, the explanation maybe the failure to initiate DNA endonucleaseactivity. Of interest is the fact that protein breakdown or the presence of a highly efficient DNA repair synthesis isclearly not required for DTx-triggered DNA fragsystem. mentation and cell lysis, resembling the cytotoxic properties Of considerable significance is our finding that DTx proexhibited by cytotoxic T-lymphocyte effector cells (9). This voked extensive DNA fragmentation in target cells and that observationdiscountstheinvolvement of a “death gene” this process was initiated a t least 2-3 h before the loss of product as proposed in models for glucocorticoid- (39) and membrane integrity. The fragmentation of host DNA is a radiation-induced (11) killing of lymphocytes. well-characterizedbiochemical feature of programmed cell EF-2 is the only protein known to be ADP-ribosylated by death (27). According to this model, cell death inextricably physiological concentrations of DTx, but specificity for EF-2 commences when DNA is degraded to oligonucleosome-size is not absolute and additional proteins are ADP-ribosylated fragments. This typeof DNA fragmentation has been shown with low efficiency (40). In addition, a broad range of intrato precede cytolysis in various cell systems, including glucocellular proteins have been identified that are modified by corticoid-mediated killing of thymocytes (28) and cytotoxic T-lymphocyte (CTL)killing of targets (9). Many investigatorscytosolic and chromatin-bound(41) ADP-ribosyltransferases have demonstrated thatDNA fragmentation is notmerely the (EC 2.4.2.30) with a catalytic mechanism similar to DTx. of the result of cell death since killing of target cells by heating, These enzymes are responsible forthesynthesis freeze-thawing, or lysis with antibody and complement does mono-, oligo-, and poly-ADP-ribosylated proteins that have been detected in practicallyevery major compartment of the not induce DNA degradation (8, 9). Extensivemembrane cell and ina variety of eukaryotic tissues(41). Although little blebbing is another morphological marker for programmed precise identity andbiological role of these cell death (27) that we consistently observed in DTx-treated is known about the modified proteins, EF-2 was recently identified as a target target cells (data not shown). Since DTx-triggered cytolysis is not, in fact, an automatic substrate (42). Thus, cellular components proven to be ADPor straightforward consequence of translation inhibition, it ribosylated by endogenous transferases under physiological must result either from (i) toxin-induced disruption of the conditions may likewise be susceptible to modification and cell boundary membrane(s) or alternatively from (ii) toxin- hence regulation byDTx. We proposethat a connection exists mediated perturbation of some other vital cell system. Asso- between substrate modification, induced directly or indirectly of anendonuclease-mediated ciation of DTx (or toxin fragments) under acidic conditions by DTx,andtheactivation with a variety of model membranes results in the formation death program. The DNA fragmentation profile of cells treated with DTx of lesions (pores/channels) capable of being penetrated by a is identical to thatof cells treated with TNF anddexamethnumber of water-soluble solutes and ions (29-32). It is con,~ which do not inhibit translation activity. ceivable that the long (>6 h) lag period between the initial a ~ o n e molecules exposure of the cell to DTx and the first signs of cytolysis The fact thata phage-encoded bacterial product elicitsDNA results from the time takenlesions for introduced intospecific fragmentation of a sort that resembles profiles observed in endosomes to recycle back to thecell surface. However, even response to several completelyunrelated agents trulysuggests when DTx was driven directly into the plasma membraneby the activation of a common cell death pathway. On the basis an acid pulse, a condition calculated to lead to toxin translo- of data presented in this report,we propose that the convencation to thecytoplasm within 20 s (33), therewas still a time tional view of DTx-mediated killing must be expanded to lag of 4-5 h before cytolysis occurred.This is slow in compar- include activation of a cytolytic pathway that displays the ison with lysis triggered by melittin orStuphylococcus a-toxin characteristic featuresof programmed cell death. (34); indeed, even the complex sequence of events leading to Acknowledgment-We thank Dr. David A. Campbell for his genlysis of nucleated cells by serum complement can be completed erous assistance. within 30 min (35). Another argument against direct membrane damage is derived fromstudiesdemonstratingthat * M. P. Chang, B. Kagan, B. J. Wisnieski, and J. Bramhall, subtarget cell lysis induced by isolated perforin granules (36) or mit,ted for publication. M. P. Chang andB. J. Wisnieski, unpublished results. activated complement (37, 38) occurs in the absence of chro-

T h e Cytolytic Pathway

of Diphtheria Toxin

REFERENCES 1. Yamaizumi, M., Mekada, E., Uchida, T., and Okada, Y. (1978) Cell 15,245-250 2. Pappenheimer, A. M. (1977) Annu. Reu. Biochem. 46,69-94 3. Proia, R. L., Wray, S. K., Hart, D.A., and Eidels, L. (1980) J. Biol. Chem. 255, 12025-12033 4. Baldwin, R. L., Chang, M. P., Bramhall, J., Graves, S., Bonavida, B., and Wisnieski, B. J. (1988) J. Immunol. 1 4 1 , 2352-2357 5. Ruff, M. R., and Gifford, G. E. (1981) in Lymphokines, (Pick, E., ed) Vol. 2, pp. 235-272, Academic Press, Orlando 6. Dealtry, G. B., Naylor, M. S., Fiers, W., and Balkwill, F. R. (1987) Eur. J. Immunol. 1 7 , 689-693 7. Schmid, D. S., Hornung, R., McGrath, K. M., Paul, N., and Ruddle, N. H. (1987) Lymphokine Res. 6 , 195-202 8. Schmid, D. S., Tite, J. P., and Ruddle, N. H. (1986) Proc. Natl. Acad. Sci. U. S. A. 8 3 , 1881-1885 9. Duke, R. C., Chervenak, R., and Cohen, J. J. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, 6361-6365 10. Duke, R. C., Cohen, J. J., and Chervenak, R. (1986) J. Immunol. 137,1442-1447 11. Sellins, K. S., and Cohen, J. J. (1987) J. Immunol. 139, 31993206 12. Wyllie, A. H. (1980) Nature 284, 555-556 13. Maniatis, T., Fritsch, E.F., and Sambrook, J. (1982) in Molecular Cloning; A Laboratory Manual, pp. 466-467, Cold Spring Harbor Laboratory, Cold Spring Harbor, New York 14. Mishell, B. B., and Shiigi, S. M. (1980) Selected Methods in Cellular Immunology, pp. 24-25, W. H. Freeman and Co., New York 15. Trinchieri, G., London, L., Koboyashi, M., and Perussia, B. (1987) in Mechanisms of Lymphocyte Actiuation and Immune Regulation, (Gupta, S., Paul, W. E., and Fauci, A. S., eds) pp. 285-298, Plenum Press Publishing Co., New York 16. Yanovich, S., Hall, R. E., and Weinert, C. (1986) Cancer Res. 46, 4511-4515 17. Marnell, M. H.,Shia, S. P., Stookey, M., andDraper, R. R. (1984) Inject. Immun. 44, 145-150 18. Sandvig, K., and Olsnes, S. (1981) J. Biol. Chem. 256,9068-9076 19. Didsbury, J. R., Moehring, J. H., and Moehring, T. J. (1983) Mol. Cell. Biol. 3 , 1283-1294 20. Merion, M., Schlessinger, P., Brooks, R. M., Moehring, J. M.,

21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.

15267

Moehring, T . J., and Sly, W. S. (1983) Proc. Natl. Acad. Sci. U. S. A. 80,5315-5319 Sandvig, K., and Olsnes, S. (1982) J. Bid. Chem. 257,7504-7513 Maxfield, F. R. (1982) J. Cell Bid. 9 5 , 676-681 Hudson, T. H., Scharff, J., Kimak, A. G., and Neville, D. M., Jr. (1988) J. Bid. Chem. 263,4773-4781 Honjo, T., Nishizuka, Y., Hayaishi, O., and Kato, I. (1968) J . Biol. Chem. 243,3553-3555 Zalman, L. S., and Wisnieski, B. J. (1984) Proc. Natl. Acad. Sci. U. S. A. 81,3341-3345 Gonzalez, J. E., and Wisnieski, B. J . (1988) J . Biol. Chem. 263, 15257-15259 Wyllie, A. H., Kerr, J. F. R., and Currie, A. R. (1980) Intl. Reu. Cytol. 68, 251-306 Cohen, J. J., and Duke, R. C. (1984) J. Immunol. 132,38-42 Kagan, B. J., Finkelstein, A., and Columbini, M. (1981) Proc. Natl. Acad. Sci. U. S. A. 78, 4950-4954 Misler, S. (1984) Biophys. J. 45, 107-109 Donovan, J. J.,Simon, M. I., Draper, R. K., and Montal, M. (1981) Proc. Natl. Acad. Sci. U. S. A. 78, 172-176 Donovan, J. J., Simon, M. I., and Montal, M. (1982) Nature 298, 669-672 Sandvig, K., and Olsnes, S. (1988) in Immunotoxins (Frankel, A. E., ed) pp. 39-73, Academic Press, Orlando Ahnert-Hilger, G., Bhakdi, S., and Gratzl, M. (1985) J. Biol. Chem. 2 6 0 , 12730-12734 Morgan, B. P., Imagawa, D. K., Dankert, J. R., and Ramm, L. E. (1986) J . Immunol. 136, 3402-3406 Gromkowski, S. H., Brown, T . C., Masson, D., and Tschopp, J. (1988) J. Immunol. 1 4 1 , 774-778 Podack, E. R. (1986) J . Cell. Biochem. 30, 133-170 Young, J. D.-E., and Cohn, Z. A. (1987) in Advances in Immunology. (Dixon, F. J., ed) Vol. 41, pp. 269-332, Academic Press, Orlando Bourgeois, S., and Gasson, J . C. (1985) in Biochemical Actions of Hormones (Litwack, G., ed) Vol. 12, pp. 311-351, Academic Press, Orlando Collier, R. J . (1975) Bacterid. Reu. 39, 54-85 Ueda, K., and Hayaishi, 0. (1985) in Annu. Reu. Biochem. (Richardson, C . C., Boyer, P. D., Dawid, I. B., and Meister, A., eds) Vol. 54, pp. 73-100, Annual Reviews, Palo Alto, CA Lee, H., and Iglewski, W. J. (1984) Proc. Natl. Acad. Sci. U. S. A. 81,2703-2707