Caspase-3-dependent and -independent degradation of 28 S ribosomal RNA may be involved in the inhibition of protein synthesis during apoptosis initiated by death receptor engagement.

Activation of death receptors initiates intrinsic apoptosis programs in various parts of the cell. To explore the possibility that ribosomal RNA (rRNA), essential for translation in ribosomes, is a target of pro-apoptotic proteins, rRNA was analyzed by electrophoresis in two apoptosis systems: human Jurkat cells treated with anti-Fas antibody and human U937 cells treated with tumor necrosis factor-alpha. In both systems, bands in addition to those of unmodified rRNA were detected a few hours after death receptor engagement. In both systems, the primary additional band was identical and comprised the 3'-terminal region of 28 S rRNA. The degradation of 28 S rRNA was simultaneous with protein synthesis inhibition in both systems. The caspase-3 inhibitor Z-DEVD-FMK suppressed rRNA degradation and protein synthesis inhibition in Jurkat cells but not in U937 cells. Together, our data suggest that different pathways are activated in the two systems we studied, and the final steps in these pathways use very similar or identical ribonucleases to cleave 28 S rRNA. These data suggest a physiological link between rRNA degradation and inhibition of protein synthesis. In general, apoptosis execution initiated by death receptor engagement is promoted by protein synthesis inhibition. Triggered by rRNA degradation, malfunction of the protein synthesis machinery may prompt death execution.

Recently, the signal transducing molecules that are downstream from death receptors and trigger the cell's intrinsic apoptosis machinery have been studied. Two signaling pathways involving Fas have been reported: mitochondria-dependent pathway and mitochondria-independent pathway (2,3). In either pathway, downstream effector caspases such as caspase-3 are activated to initiate apoptosis. The activity of these two pathways varies between cell types and tissues. The pathway downstream of TNFR1 is more complicated than the pathway downstream of Fas because, depending on the intracellular environment, activation of this receptor can lead to apoptosis, necrosis-like cell death or a pro-inflammatory response. In addition, FAP-1 associated with death receptors modulates death signaling (4,5), and Fas-associated death domain protein (FADD) can initiate necrosis-like organized cell death (6). Analysis at the molecular level of cell death caused by death receptors improves our understanding of development, homeostasis, diseases, and cancer because apoptosis through death receptors plays a pivotal role in these phenomena (7)(8)(9).
At a particular point after an apoptotic insult, cells reach a decisive point, past which rescue from apoptosis is impossible (10). Once this point is passed, apoptosis enters the period termed the execution phase. One of the characteristics of the execution phase is irreversible degradation of cellular components, during which caspase-3-activated deoxyribonuclease (CAD) digests genomic DNA (11) and caspase-6 digests nuclear cytoskeleton lamin (12). These events contribute to the final morphologic change in the nucleus that is characteristic of apoptosis. In addition, caspase-3 digests some signaling proteins (12). Such irreversible degradations accelerate cell death. Therefore, to fully decipher the story of apoptosis, delineating the events at the onset of the execution phase is important.
Protein synthesis is a characteristic of cell viability. In fact, inhibition of protein synthesis by drugs such as actinomycin-D causes apoptosis in some cell types (13). The ribosome is a large protein-RNA complex essential for protein synthesis. In humans, the ribosome (80 S) composes a large subunit, involving 46 proteins and 28 S, 5.8 S, and 5 S RNA species and a small subunit including 33 proteins and 18 S RNA (14). Although ribosomal regulation in de novo protein synthesis has been the focus of many studies, work addressing its role in apoptosis had not been published previously. Ultrastructural analysis has shown that ribosomes dissociate from the rough endoplasmic reticulum during apoptosis of some tissues (15,16). Such morphologic studies encouraged us to pursue the molecular analysis of the ribosome during apoptosis.
Because ribosomal RNA (rRNA) is an essential component for protein synthesis (17), we have examined in the present study the role of rRNA in apoptosis caused by activation of death receptors. Apoptosis induced by cAMP analogs features the degradation of rRNA (18). Although some small drugs including cAMP analogs strongly induce apoptosis, their pleiotropic effects can complicate attempts to separate physiologic death pathways from unusual secondary pathways that are stimulated by side effects of the drugs. The molecular mechanisms and biologic significance of alterations in rRNA had not been investigated previously. Here we show that 28 S rRNA is selectively degraded by the same (or a very similar) ribonuclease (RNase) in the Jurkat and U937 systems. In addition, examination of the association between rRNA degradation and protein synthesis revealed the potential involvement of this degradation in the malfunction of protein synthesis machinery.
Apoptosis by Death Receptor Activation-In the Fas-Jurkat system, Jurkat cells were washed once with serum-free RPMI 1640 and resuspended in the same medium. To induce apoptosis, a mouse monoclonal anti-Fas antibody (CH-11; MBL, Nagoya, Japan) was added at a concentration of 100 ng/ml, and the cells were incubated for various periods of time. Control cells were incubated under the same conditions but without CH-11. Cells incubated with 100 ng/ml unrelated mouse IgM (No. 02-6800; Zymed Laboratories Inc., San Francisco, CA) were an additional control.
In the TNFR-U937 system, U937 cells were washed once with serumfree RPMI 1640 and resuspended in the same medium. To induce apoptosis, recombinant human TNF-␣ (R & D Systems, Minneapolis, MN) was added at a concentration of 30 ng/ml, and the cells were incubated for various periods of time. Control cells were incubated under the same conditions but without TNF-␣.
To examine the effect of caspase-3 inhibition on apoptosis, cells were incubated for 1.5 h at 37°C in serum-free RPMI 1640 including a cell-permeable form of Z-DEVD-FMK (50 M; caspase-3 inhibitor II; Calbiochem) that had been dissolved in dimethyl sulfoxide to create a stock solution at 50 mM, and then CH-11 or TNF-␣ was added to the cell culture. Control cells were treated with the same volume of dimethyl sulfoxide. Apoptosis was evaluated by using morphologic observation under a light microscope and the DNA fragmentation assay (see the following section).
DNA Fragmentation Assay-Fragmented DNA was extracted with Triton X-100-containing buffer (19) and subjected to 1.4% agarose gel electrophoresis followed by ethidium bromide staining.
RNA Fragmentation Assay and Northern Blotting-Total RNA was isolated from cells by using the Trizol reagent (Life Technologies, Inc.). The concentration of the solution was determined by measuring the absorbance at 260 nm.
Polyacrylamide (2.5%)-agarose (0.5%) composite gel electrophoresis was modified from the method of Peacock and Dingman (20). Distilled water (13 ml) was added to 75 mg of electrophoresis-grade agarose (Iwai Chemicals, Tokyo, Japan). After the agarose melted, the gel solution was cooled to about 40°C. The following solutions then were added to the gel solution and mixed well: 0.75 ml of 0.8 M Tris, 0.49 M acetic acid, 20 mM EDTA (20ϫ TAE buffer), 1.25 ml of 29% acrylamide, 1% N, NЈ-methylenebisacrylamide, 20 l of TEMED, and 50 l 10% ammonium persulfate. The mixture was then poured into the mold for the gel (thickness, 1.0 mm; width, 87 mm; length, 77 mm) and, after the insertion of a comb at the top of the mold, kept for 30 min at room temperature for polymerization of the polyacrylamide gel. To solidify the agarose and cool the composite gel, it then was kept at 4°C for 1 h. RNA samples (2.5 g of RNA/lane) were mixed with an equal volume of 90% (v/v) formamide, 10% (v/v) glycerol, incubated for 3 min at 65°C, cooled for 5 min at room temperature, and then loaded onto the gel. Gels were run for 30 to 40 min at 200 V in 1ϫ TAE buffer, which was precooled on ice; the gels then were stained with ethidium bromide.
For Northern blotting after composite gel electrophoresis, RNA was electroblotted onto Hybond-Nϩ membrane (Amersham Pharmacia Bio-tech) in 25 mM Tris containing 0.19 M glycine by using the Mini Transblot Cell (Bio-Rad) at 0.3 A for 1 h and fixed by using the UV Stratalinker (Stratagene, La Jolla, CA). Oligodeoxyribonucleic acid probes were synthesized by ESPEC-OLIGO Service Corp. (Tsukuba, Ibaraki, Japan) and terminally labeled by using [␥-32 P]ATP (Amersham Pharmacia Biotech) and T4 polynucleotide kinase (Takara, Otsu, Shiga, Japan). The blots were prehybridized for 15 min in QuickHyb hybridization solution (Stratagene) and hybridized for 1 h with a 32 P-labeled oligonucleotide probe. The blots were washed with 0.1% SDS, 2ϫ SSC (1ϫ SSC: 15 mM sodium citrate containing 0.15 M NaCl) followed by 0.1% SDS, 0.1ϫ SSC and subjected to autoradiography with Kodak X-Omat films.
Rapid Amplification of 5Ј cDNA Ends (5Ј-RACE)-Apoptotic RNA fragments were isolated by combining agarose gel electrophoresis with a "crush and soak" method involving phenol. 5Ј-RACE was performed with the 5Ј-RACE kit version 2.0 (Life Technologies, Inc.) according to the manufacturer's instructions. The following reverse primers recognizing the human 28 S rRNA sequence (21) were used: 5Ј-GCTCAA-CAGGGTCTTC-3Ј (antisense of nucleotides 3873-3888) for first-strand cDNA synthesis and 5Ј-CACTGGGCAGAAATCACATC-3Ј (antisense of nucleotides 3655-3674) for the subsequent polymerase chain reaction. The products were analyzed by agarose gel electrophoresis. Amplified cDNA was isolated from the gels with the QIAQuick gel extraction kit (Qiagen, Hilden, Germany) and directly sequenced by using the reverse primer for polymerase chain reaction and the Big Dye terminator cycle sequencing kit (PE Biosystems, Foster City, CA). For the confirmation of the sequence results, cDNA from apoptotic Jurkat cells was subcloned into pBluescript II KS (Stratagene) and sequenced with M13 universal primers.
Metabolic Labeling of Proteins-Cells were placed in serum-free RPMI 1640 lacking methionine (Sigma) and pulse-labeled with 110 Ci/ml L-[ 35 S]methionine (EASYTAG; NEN Life Science Products) for 20 min. Cells were then washed three times with phosphate-buffered saline, lysed for 10 min on ice with 50 mM Tris-HCl (pH 7.5) containing 1% Triton X-100, 0.15 M NaCl, 5 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride, and centrifuged for 5 min at 15,000 ϫ g. The supernatant was subjected to 10% SDS-polyacrylamide gel electrophoresis and autoradiography. Autoradiograms were scanned by using an image analyzer (Model ES-8000; Epson, Tokyo, Japan), and the intensity of the bands was evaluated with the NIH Image software (National Institutes of Health, Bethesda, MD).

Stimulation of Death Receptors Caused Degradation of rRNA in Cytoplasmic
Polysomes-We selected two apoptosis systems: the human T-cell leukemia cell line Jurkat, treated with the anti-Fas antibody CH-11 to activate Fas (the Fas-Jurkat system), and the human monoblastic leukemia cell line U937, treated with TNF-␣ to activate TNFR1 (the TNFR-U937 system). These systems including leukemia-type cell lines seem reasonable because Fas and TNFR1 have been shown to be important in the development and function of leukocytes in mice lacking these receptors (3,24).
To examine the alteration of rRNA in these apoptosis systems, total RNA was isolated by using Trizol and analyzed by polyacrylamide-agarose composite gel electrophoresis. Bands in addition to those of unmodified rRNA appeared in the apo-rRNA Degradation Caused by Death Receptor Engagement ptotic cells (Fig. 1A, lanes 1 and 3). No alteration of RNA was detected in control samples (Fig. 1A, lanes 2 and 4), and Jurkat cells treated with unrelated mouse IgM lacked any additional bands (Fig. 1A, lane 6). The electrophoresis patterns shown here are much simpler than those during apoptosis induced by cAMP analogs (18). No difference occurred in the banding patterns of apoptotic and control samples separated on 12% polyacrylamide gel electrophoresis to identify smaller RNA molecules (data not shown).
Ribosomes are generated in the nucleoli and are exported from the nucleus as mature forms (25). To examine whether the extra RNA molecules we identified occurred in cytoplasmic ribosomes, cytoplasmic polysome fractions from apoptotic Jurkat cells were collected by ultracentrifugation. RNAs collected from the fractions were analyzed by using composite gel electrophoresis. The extra bands were observed in the polysome fractions (Fig. 1B), suggesting that the extra RNA molecules were derived from ribosomes involved in protein synthesis and were not due to aberrant RNA synthesis in apoptotic nuclei. Because the extra RNA species were not concentrated by polysome isolation (compare lanes 1 and 3, Fig. 1B), this rRNA alteration was not specific to the polysome.
A Putative Cutting Site of 28 S rRNA in Variable Region Eight Is Identical in the Fas-Jurkat and TNFR-U937 Systems-Although these two systems stimulate different death receptors, the patterns of extra RNA were very similar (Fig.  1A). In particular, both systems yielded a prominent band (indicated by the arrowhead in Fig. 1A) just smaller than 18 S rRNA; the similar mobility of this band from the two cell types suggests an identical structure. We then sought the molecular identity of this RNA species. After composite gel electrophoresis, RNA from apoptotic and control Jurkat cells were electroblotted onto nylon membranes, and the blots were hybridized with probes for 28 S and 18 S rRNA. This prominent band hybridized with the 28 S rRNA probe ( Fig. 2A, left panel), but no extra bands were detected with the probe for 18 S rRNA ( Fig. 2A, right panel). This finding suggests that this RNA species derived from RNase-degraded 28 S rRNA.
To precisely map the cutting site, we performed Northern blotting analysis by using oligonucleotide probes for various partial sequences of 28 S rRNA. The prominent RNA band identified previously comprised the 3Ј-terminal region of 28 S rRNA (Table I), and the putative cleavage site presumably was somewhere between nucleotides 3251 and 3710. Next, this RNA fragment was isolated from Jurkat and U-937 cells undergoing apoptosis and assessed carefully by using 5Ј-RACE. The analysis of amplified cDNA fragments revealed a single band with identical mobility from both systems (Fig. 2B). The diffuse nature of the bands mainly is due to the various lengths of the dC-tail of the template cDNAs. Sequencing of these cDNA samples revealed that in both systems, the 5Ј-end of the RNA molecule was nucleotide 3536; by definition (26,27), the cutting site maps to variable region eight (Fig. 2C). This result suggests that although these systems stimulate different death receptors, the cutting mechanism may be very similar, possibly executed by the same RNase. This result further suggests that the extra RNA species indicated by the asterisks in Fig. 1 is the 5Ј-terminal region of 28 S rRNA (see also Fig. 2A, middle panel).
Dependence of rRNA Degradation on Caspase-3-like Activity Differed Between the Fas-Jurkat and TNFR-U937 Systems-Caspases are a family of proteases that are specifically activated by apoptotic stimuli (12). During the execution phase of apoptosis, effector caspases are activated, and the various substrates are digested to promote cell death. In particular, caspase-3 is a key player involved in degradation of many components in cells. If the execution phase leads to rRNA degradation, caspase-3 may be upstream of its degradation. To investigate this possibility, Jurkat cells were pretreated with a caspase-3 inhibitor, a cell-permeable form of Z-DEVD-FMK, then treated with the monoclonal anti-Fas antibody CH-11. RNA and DNA collected from the cells were analyzed. Z-DEVD-FMK inhibited the degradation of genomic DNA (DNA ladder formation), which is executed by caspase-3-activated CAD (Fig.  3A, lower panel). Interestingly, degradation of 28 S rRNA also was suppressed by the caspase-3 inhibitor (Fig. 3A, upper panel), suggesting that caspase-3 or a caspase-3-like protease is We next examined the involvement of caspase-3 in rRNA degradation in the TNFR-U937 system. In this system, the caspase-3 inhibitor had no effect on rRNA degradation (Fig. 3B,  upper panel). This result was not because the inhibitor failed to . The primary RNA fragment is indicated by the arrowhead. B, electrophoresis of 5Ј-ends of cDNAs from 28 S rRNA fragments. Induction of apoptosis is described under "Experimental Procedures." cDNA was amplified from the primary rRNA fragment by using reverse transcriptase and terminal deoxynucleotidyl transferase (TdT) followed by polymerase chain reaction. Controls were prepared under the same conditions but without terminal deoxynucleotidyl transferase. All samples were electrophoresed in 1.8% agarose gels and stained with ethidium bromide. Lane 1, the Fas-Jurkat system; lane 2, control for the Fas-Jurkat system; lane 3, the TNFR-U937 system; lane 4, control for the TNFR-U937 system. C, schematic representation of human 28 S rRNA and the putative cutting site in our apoptosis systems. The structure of 28 S rRNA has been defined in previous reports (26,27). Striped box, constant region; open box, variable region; S/R, ␣-sarcin/ricin loop (29, 43); V8, variable region 8; arrow, putative cutting site. nt, nucleotides.

rRNA Degradation Caused by Death Receptor Engagement
work in this cell line: Z-DEVD-FMK suppressed DNA ladder formation in U937 cells (Fig. 3B, lower panel) just as it had in Jurkat cells. In light of the identical structure of the main RNA fragment, different pathways, caspase-3-dependent and -independent pathways, likely activate the same or very similar RNases in these systems; these pathways ultimately converge in the cleaving of 28 S rRNA.

Degradation of 28 S rRNA Caused by Death Receptors Was Concomitant with Protein Synthesis
Inhibition-We then sought the role of rRNA degradation in apoptosis. Our hypothesis was that irreversible fragmentation of rRNA compromised the function of ribosomes, that is, protein synthesis. To confirm this hypothesis, the kinetics of 28 S rRNA fragmentation and protein synthesis were examined. In the Fas-Jurkat system, a prominent 28 S rRNA-derived fragment, which gradually intensified, reproducibly was detected beginning 4 h after the addition of the monoclonal anti-Fas antibody CH-11 (Fig. 4A). In the same system, protein synthesis began to decrease at the onset of rRNA fragmentation (Fig. 4B).
To examine whether this coincidence was observed in another apoptosis system caused by death receptor engagement, the kinetics analysis was performed in the TNFR-U937 system, in which RNA fragmentation began to be detected 2 h after receptor stimulation (Fig. 5A). This cell line seemed especially sensitive to the low serum environment, and this stress caused a moderate decrease in protein synthesis in the absence of apoptotic stimulation (Fig. 5B, right panel). However, apoptosis caused a more dramatic shutdown of protein synthesis (Fig. 5B,  left panel). Again, fragmentation of rRNA was concomitant with protein synthesis inhibition. These results in the two different systems suggested that fragmentation of rRNA occurs during the early execution phase of apoptosis (fragmentation occurred just a few hours after stimulation). Furthermore, rRNA fragmentation seems to be involved in the apoptosisassociated inhibition of protein synthesis in the cell.
Dependence of Protein Synthesis Inhibition on Caspase-3-like Activity Supported the Involvement of rRNA Fragmentation in Protein Synthesis Inhibition during Apoptosis-The close transient relationship between rRNA degradation and inhibition of protein synthesis (Figs. 4 and 5) suggests that this degradation is involved in the inhibition. To test this possibility, the effect of the caspase-3 inhibitor Z-DEVD-FMK on protein synthesis was examined in the two apoptosis systems. The caspase-3 inhibitor suppressed protein synthesis inhibition (Fig. 6A) and rRNA degradation (Fig. 3A) in Jurkat cells. In contrast, Z-DEVD-

rRNA Degradation Caused by Death Receptor Engagement
FMK failed to recover protein synthesis in U937 cells (Fig. 6B), in which rRNA degradation was caspase-3-independent (Fig.  3B). Therefore, degradation of rRNA may be a physiologically important cellular signal mediator for the malfunction of the protein synthesis machinery in our apoptosis systems. DISCUSSION Malfunction of the ribosome causes death of cells and organisms. Until now, various small chemicals that specifically attack the ribosome have widely been used as antibiotics (28). Furthermore some natural toxins are protein ribotoxins. For example, ricin and ␣-sarcin catalyze covalent modifications in adjacent nucleotides of 28 S rRNA, leading to malfunction of the ribosome in vitro and in vivo (29). Colicin E3 specifically cleaves a single site in 16 S rRNA in the small subunit of the bacterial ribosome, resulting in inactivation of the ribosome and death of the host bacteria (30). Here we propose the existence of a new RNase that is cryptic in normal cells but is activated to cleave 28 S rRNA in the polysome during apoptosis triggered by either of two death receptors.
The putative cleavage site was identical in the 28 S rRNAs from the Fas-Jurkat and TNFR-U937 systems (Fig. 2). According to phylogenic studies of rRNA, the cleavage site is in variable region eight, which is absent in bacteria. This finding contrasts with the observation that ␣-sarcin and ricin attack an ␣-sarcin/ricin loop in a constant region (Fig. 2C) that is involved in the binding of aminoacyl-tRNA and in GTP hydrolysis (29). The function of variable regions currently is unknown. Variable region eight is conserved in vertebrates and contains a unique double-loop structure (27). The apoptosis-associated cleavage might induce a change in the higher order structure of the rRNA and ribosome. The three-dimensional structure of the yeast ribosome has been published recently (31). Future structural analysis of eukaryotic ribosomes will provide insight into the possible function(s) of this variable region.
From the comparison of the band intensities of intact 28 S rRNA and the main RNA fragment (Fig. 1B), most ribosomes in the polysomes seem to be intact. Even though destructive digestion occurs in some ribosomes, how can degradation of so few ribosomes cause the inhibition of protein synthesis at the cellular level? This paradox was reported more than 20 years ago in studies of protein ribotoxins. In intact cells treated with ribotoxins such as ricin (32) and colicin E3 (33), protein synthesis stopped when only a fraction of the ribosomes was inactivated. Two mechanisms that are not mutually exclusive have been proposed. In the first, impaired ribosomes are a dominant negative factor in polysomes. In this situation, one or a few ribosomes with damaged rRNA occur per polysome, attach to FIG. 5. Coincidence of rRNA fragmentation and inhibition of protein synthesis in the TNFR-U937 system. Apoptosis induction in this system is described under "Experimental Procedures." See legend of Fig. 4 for a description of subsequent experimental procedures. A, kinetics of rRNA fragmentation. The main 28 S rRNA fragment is indicated by the arrowhead. B, kinetics of protein synthesis inhibition. The results shown here are representative of three independent experiments.

FIG. 6. Effect of a caspase-3 inhibitor on inhibition of protein synthesis.
A, the Fas-Jurkat system. Jurkat cells were pretreated with Z-DEVD-FMK, then treated for 8 h with anti-Fas antibody. The cells were pulse-labeled with [ 35 S]methionine and lysed with Triton X-100-containing buffer. The lysates were subjected to SDS-polyacrylamide gel electrophoresis followed by autoradiography. Incorporation of methionine was measured by using an image scanner and the NIH Image software. B, the TNFR-U937 system. U937 cells were pretreated with Z-DEVD-FMK, then treated for 3 h with TNF-␣. The subsequent procedures were the same as described for panel A. Methionine incorporation relative to the control is shown here. The data represent mean Ϯ S.D. of duplicate samples from one of three independent experiments. the mRNA, and block further elongation of the peptide chains. This block especially seems to inhibit synthesis of housekeeping proteins, which are constitutively synthesized in large quantities and are important for cell viability. In the second proposed mechanism, cleavage of rRNA generates signals for the inhibition of protein synthesis. The ricin-and ␣-sarcininduced modification of rRNA activates stress-activated protein kinase (34). Therefore, this pathway might be activated downstream of rRNA degradation in our systems. A link between the pathway of stress-activated protein kinase and inhibition of protein synthesis has been reported very recently (35).
Another candidate of the signaling molecules activated by rRNA degradation is a double-stranded RNA-activated protein kinase. Activated by double-stranded RNA, RNA-activated protein kinase autophosphorylates, phosphorylates eukaryotic translation initiation factor-2, and eventually inhibits protein synthesis (36). RNA-activated protein kinase is a ribosomebinding protein that associates with the large subunit via the ribosomal protein L18 (37). rRNA has a complex secondary structure involving many loops of double-stranded RNA (26). Cleavage of 28 S rRNA might expose part of such a structure in a large subunit, thereby stimulating RNA-activated protein kinase. RNA-activated protein kinase is activated in U937 cells that have been treated with TNF-␣ (38). Our culture cell systems may help to resolve the apparent paradox.
Determining the structure of an RNA fragment (Table I and Fig. 2) provides a key piece of information needed to identify RNase responsible for cleaving 28 S rRNA. So what kind of RNase is involved in this cleavage? Explaining the selective cleavage of 28 S rRNA in light of current knowledge of RNases is impossible. The base on the nucleotide at the 3Ј side of the phosphodiester bond preferentially broken by human pancreatic-type RNase is cytosine (39). In human cells, 2-5A-dependent RNase L (40), which is downstream of interferon receptors and also implicated in apoptosis (41), cleaves 28 S rRNA at different sites, between nucleotides 3999 and 4000 and between 4000 and 4001 (35), suggesting that the RNase in our study could be a different one. Some mammalian RNases have cytotoxicity dependent on their RNase activity (42). However, their targets are largely unknown. CAD is a well characterized nuclease that is activated during apoptosis (11). Fragmentation of rRNA occurred even when CAD activation was suppressed ( Fig. 3B), indicating that CAD is not essential for rRNA degradation in the TNFR-U937 system. Although a previous report described the correlation between rRNA fragmentation and genomic DNA fragmentation during apoptosis (18), our data suggest the presence of a nuclease other than one involved in the digestion of genomic DNA.
The functional relationship between protein synthesis and apoptosis remains enigmatic. In some systems, inhibitors of protein synthesis prevent apoptosis in the cell; in others, these inhibitors promote cell death (13). Apoptosis caused by engagement of death receptors is an example of a system in which inhibition of protein synthesis promotes cell death (1). The connection between rRNA fragmentation and inhibition of protein synthesis (Figs. 4 -6) suggests that this fragmentation is a promotive effector for the execution of cell death. Identification of the molecules responsible for this degradation likely will reveal new execution pathways of apoptosis. In addition, this identification will be noteworthy in the field of signal transduction because no intracellular RNase except 2-5A-dependent RNase L has been shown to be involved in signaling initiated from extracellular stimuli.