THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 268, No. 14, Issue of May 15, pp. 10686-10693,1993 Printed in U. S. A .

A Cytotoxic Ribonuclease STUDY OF THE MECHANISM OF ONCONASE CYTOTOXICITY*

(Received for publication, October 9, 1992, and in revised form, January 4, 1993)

YouNeng WuS, Stanislaw M. Mikulskiq, Wojciech Ardeltll, Susanna M. Rybakt, and Richard J. Youlet From the $Biochemistry Section, Surgical Neurology Branch, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland 20892 and the nAlfacell Corporation, Bloomfield, New Jersey 07003

Onconase, or P-30, is a protein initially purified from extracts of Rana pipiens oocytes and early embryos based upon its anticancer activity both in vitro and in vivo. It is a basic single-chain protein with an apparent molecular mass of 12,000 daltons and is homologous to RNase A. In cultured 9L glioma cells, onconase inhibits protein synthesis with an ICao of about M. The inhibition of protein synthesis correlates with cell death determined by clonogenic assays. ‘2aI-Labeled onconase binds to specific sites on cultured 9L glioma cells. Scatchard analysis of the binding data shows that onconase appears to bind to cells with two different affinities, one with a K a of 6.2 X lo-* and another of 2.5 X M. Each cell could bind about 3 X lo5 molecules of onconase at each of the two affinity sites. The low affinity K d is similar to the IC5,, for onconase toxicity. Onconase also demonstrates a saturability of cytotoxicity at a concentration that would saturate the low affinity binding site. Incubation at 4 “C increased the binding of onconase to cells relative to 37 “C bind- ing and also increased the sensitivity of cells to oncon- ase toxicity, indicating that receptor binding may be an initial step in cell toxicity. Onconase cytotoxicity can be blocked by metabolic inhibitors, NaN3 and 2- deoxyglucose, and cytotoxicity is potentiated 10-fold by monensin. Ribonuclease activity appears necessary for onconase toxicity because alkylated onconase, which only retains 2% of the ribonuclease activity, was at least 100-fold less potent in inhibiting protein syn- thesis in cells. Onconase inhibition of protein synthesis in 9L cells coincides with the degradation of cellular 28 S and 18 S rRNA. In contrast to RNase A, onconase is resistant to two RNase inhibitors, placental ribonu- clease inhibitor and Inhibit-AceTM. Northern hybridi- zation with placental ribonuclease inhibitor cDNA probe indicates that 9L glioma cells contain endoge- nous placental ribonuclease inhibitor mRNA. Based on these results, we propose that onconase toxicity results from onconase binding to cell surface receptors, inter- nalization to the cell cytosol where it degrades ribo- somal RNA, inhibiting protein synthesis and causing cell death.

Onconase in isolated from Rana pipiens eggs and early embryos based upon its anticancer activity both i n vitro and i n vivo (1-3). Phase I clinical trials of onconase therapy for various carcinomas have been completed with encouraging

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertkement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

results,’ and phase 11 clinical trials are currently ongoing. The mechanism of the onconase cytotoxic and/or cytostatic

activity is unknown. Sequencing of onconase revealed that it was homologous to pancreatic RNase A (4). Onconase also exhibits ribonuclease activity, at about one one-hundredth the specific activity of RNase A (4). Other members of the RNase A superfamily homologous to onconase include two frog lectins found in the eggs of Rana japonica and Rana catesbeiana (5-10). These proteins are about 53% identical to onconase in amino acid sequence and also possess ribonucle- ase activity (10). They bind to cells and preferentially agglu- tinate tumor cells (8, 9). A receptor for the R. japonica lectin has been purified to homogeneity and is high in sialic acid content (8). These results suggest that, by analogy, onconase may bind to cells as an initial step in its anticancer activity.

In this study we present evidence for a model of the mech- anism of action of onconase. Onconase can bind to the cell surface and appears to enter the cell cytosol where it degrades ribosomal RNA which, in turn, results in protein synthesis inhibition and cell death. These results not only illuminate the mechanism of onconase toxicity but also suggests how other cytotoxic ribonucleases and RNase chimeras may kill cells.

EXPERIMENTAL PROCEDURES

Materials-Onconase was obtained from Alfacell Corp. (Bloom- field, NJ); bovine pancreatic RNase A was purchased from Calbi- ochem; human placental ribonuclease inhibitor (PRI),* from Promega Biotech (Campbell, CA); and Inhibit-AceTM, from 5’-3’ (Paoli, PA); NaIz5I (17 Ci/mg iodine) was from ICN Biomedicals; total RNA extraction kit was from Pharmacia LKB Biotechnology Inc.; brefeldin A was from Sigma; calf liver 28 S and 18 S RNA, from Boehringer Mannheim Biochemicals.

Cell Line-9L (glioma) was grown in Dulbecco’s modified Eagle’s medium containing 10% fetal calf serum, 2 mM glutamine, 1 mM sodium pyruvate, 0.1 mM nonessential amino acids, and 10 pg/ml gentamycin.

Alkylation of Onconase-Alkylation of onconase was performed according to the modified method of Crestfield et al. (11). 1.0 mM protein was incubated with a 50-fold molar excess of iodoacetic acid, sodium salt in 0.1 M sodium acetate buffer, pH 5.5, for 18 h at 23 ”C. The samples were desalted on a Bio-Gel P-2 (Bio-Rad) column in 5% (v/v) formic acid, and lyophilized. Onconase alkylated with a 50-fold molar excess of iodoacetate retained 2% of the original enzymatic activity assayed with ribosomal yeast RNA as a substrate by a modification of the precipitation method of Anfinsen et al. (12). The enzyme assay contained 0.56 mg/ml yeast ribosomal RNA, 20 pg/ml human serum albumin, 0.2 mM MES buffer, pH 6.0, and 5-20 nM enzyme or 20-70 nM alkylated enzyme. The mixtures were incubated for 15 min at 37 “C, and the reaction was terminated with 2 volumes of 3.4% (v/v) ice-cold perchloric acid (13). The samples were kept in

S. M. Mikulski et al., unpublished results. The abbreviations used are: PRI, placental ribonuclease inhibitor;

DTT, dithiothreitol; MES, 2-(N-morpholino)ethanesulfonic acid.

10686

Mechanism of Cytotoxic Ribonuclease 10687

an ice-bath for 10 min, and the precipitate formed was centrifuged off. Absorbance at 260 nm was then measured in the supernatants.

Protein Synthesis Assay-Protein synthesis inhibition by oncon- ase, alkylated onconase, RNase A, and ricin was as described previ- ously (14). Briefly, cells were plated at concentrations of 1 X 10 cells/ml in 96-well microtiter plates in leucine-free RPMI 1640 me- dium without fetal calf serum in a volume of 100 pl. Sample or control additions were added in a volume of 10 gl, and the plates were incubated at 37 or 4 "C for the times indicated for each experiment. Phosphate-buffered saline containing 0.1 pCi of ["C]leucine (20 pl) was added, and the cells were harvested onto glass fiber filters using a PHD cell harvester, washed with water, dried with ethanol, and counted. The results were expressed as a percentage of ['4C]leucine incorporation in the mock-treated control cells. All cytotoxicity as- says were performed 2-5 times. Values given represent the average of duplicate or triplicate samples with less than 10% standard error.

Microautorodiography-9L cells were plated on coverslips in six- well plates in leucine-free RPMI 1640 medium without fetal calf serum in a volume of 2.0 ml. Sample or control additions were added in a volume of 200 pl, and the plates were incubated at 37 "C for 24 h. Phosphate-buffered saline containing 2.0 pCi of ["C]leucine (400 pl) was added for 2.0 h, and the cells were washed 3 times with leucine-free RPMI 1640 medium without fetal calf serum. Cells were fixed with 1:3 (acetic acid:methanol) for 15 min, washed three times with 80% methanol, and air dried. The emulsion gel (Kodak NTB2) was diluted 1:l in H20, and the coverslips were dipped, air dried, and exposed desiccated for 16 h. The coverslips were developed, fixed, and mounted on glass slides.

Clonogenic Cell Assay-The number of clonogenic cells surviving treatment with onconase or control additions was determined by using a limiting dilution assay as described previously (15). Cells were treated with additions in a 1-ml volume in 24-well plates for 24 h under the same culture conditions described above. The cells then were washed, trypsinized, and resuspended in complete growth me- dium and six serial 10-fold dilutions were made. Ten aliquots of each dilution were plated into 96-well microtiter plates, and the plates were incubated for 14 days at 37 "C in a humidified atmosphere. Wells with growing colonies were scored by examination under an inverted phase microscope. The number of clonogenic cells remaining from the original number treated was calculated using a Spearman-Karber estimator (16).

Binding Assay of 1261-Labeled Onconase to 9L ~ells-'~~I-Iodination of onconase was performed with lactoperoxidase and glucose oxidase EnzymobeadTM from Bio-Rad. Specific activity of onconase was 0.5 mCi/mg protein. 1.0 p~ of lZ5I-labeled onconase is about 3,000,000 cpm/2OO pl. Cells were trypsinized from culture flasks, washed, and resuspended at 1 X 10 cells/ml in 200 pl/microcentrifuge tube that had been coated overnight with 5 mg/ml bovine serum albumin in Hanks' balanced salt solution. Cells were incubated with '251-onconase in leucine-free RPMI 1640 medium containing 5 mg/ml bovine serum albumin for 2 h at 4 "C. Cells were centrifuged and washed three times with Hanks' balanced salt solution containing 1.0 mg/ml bovine serum albumin, and then the cell pellets were counted. Nonspecific binding was determined by using a 300-fold excess of unlabeled onconase.

Total RNA Extraction-Control cells or sample cells that have been treated with onconase, alkylated onconase, RNase A, or ricin, were washed then lysed with "Extraction Buffer" (RNA extraction kit, Pharmacia) containing guanidinium thiocyanate and N-lauryl- sarcosine. The lysates were loaded onto a self-generating cesium trifluoroacetate gradient and centrifuged for at least 16 h at 125,000 X g. The total RNAs were recovered in TE buffer (10 mM Tris-HC1, pH 7.5, 1 mM EDTA).

Inhibit-AceTM and PRI Inhibition of Ribonuclease Activity-3.0 pg of calf liver rRNA (28 S + 18 S) were incubated with onconase or RNase A in the presence or absence of either Inhibit-AceTM(l.O unit) or PRI (60 units) for 15 min at 37 "C in 20 pl. For PRI assays, 10 mM DTT was added to both the sample and control reaction mixtures. RNA electrophoresis was performed with 1.2% agarose gels.

Northern Hybridization of PRI mRNA in 9L Glioma Cells with PRI cDNA Probe-Total RNAs were extracted from 9L cells as described above, and 20 pg of the total RNAs were run on a 1.2% agarose gel and transferred to nitrocellulose. After prehybridization at 42 "C for 4 h, the nitrocellulose membrane was hybridized overnight in 10 ml of hybridization buffer containing 32P-labeled PRI cDNA (6 X 10' cpm/50 ng), washed, and autoradiographed.

RESULTS

Onconase Toxicity to Cultured 9L Glioma Cells-Onconase incubated with 9L cells for 24 h inhibited cellular protein synthesis as measured by [14C]leucine incorporation into pro- tein. Increasing concentrations of onconase inhibited protein synthesis down to 10% of control values, with an ICSo of about

M (Fig. 1, panel B ) . There were no significant changes in the total cell number when the cells were treated with the same concentration of onconase as used for the protein syn- thesis assay (Fig. 1, panel A ) . Cells remain intact after treat- ment with increasing concentrations of onconase as measured by trypan blue exclusion. Bovine pancreatic RNase A, at concentrations 1000 times greater than the ICso of onconase, only inhibited incorporation of the [14C]leucine into 9L cells by about 10%. Protein synthesis is inhibited to the same extend by 1 mM RNase A uersus 1 p M onconase. Native onconase showed the same cytotoxicity to 9L cells even in the presence of 100-fold excess RNase A.

Microautoradiography of the control and onconase-treated cells also demonstrates that onconase has a direct cytotoxic effect on cells and the protein synthesis inhibition does not result from inhibiting cell division through a cytostatic activ- ity. Onconase at M markedly decreased [14C]leucine in- corporation in individual cells, and no significant effect on

B Concentration (M)

9 9" 9 - 1 ,

RNase A

Alkylalid Onconara \

k \

b - 2

Concentration (M)

FIG. 1. Panel A, total cell counts and trypan blue excluding cell counts of 9L cells treated with onconase. 9L cells (1 X lo5 cells/ml) were plated into 96-well microtiter plates and treated with varying concentrations of onconase as described under "Experimental Pro- cedures." After 24 h in the presence of onconase or control additions, cells were counted in a hemocytometer. Total counts, total number of cells; trypan blue, number of trypan blue excluding cells. After 24 h in the presence of onconase or control additions cells were counted in a hemocytometer. Panel B, inhibition of protein synthesis in 9L cells by onconase versus alkylated onconase and RNase A. Onconase, onconase-treated cells; RNase A + oncome, onconase + 10" M RNase A-treated cells; RNase A, RNase A-treated cells; alkylated onconase, alkylated onconase-treated cells. After 24 h in the presence of additions, protein synthesis was determined as described under "Experimental Procedures." The data points are determined from the mean of duplicate incubations. The standard error was less than 10%.

10688 Mechanism of Cytotoxic Ribonuclease

cell density was found (Fig. 2). Thus onconase is cytotoxic in addition to the cytostatic activity previously described (1, 3).

Assessment of Cell Killing by Onconuse in a Clonogenic Assay-Since protein synthesis inhibition could reflect tran- sient nonlethal mRNA degradation or interference in meta- bolic labeling, clonogenic assays were performed to quantitate the extent to which onconase kills cells. After 24 h in the presence of onconase, 9L cells were washed, resuspended in Dulbecco's modified Eagle's complete medium, and diluted serially into 96-well microtiter plates. The wells that con- tained surviving cells were scored after 2 weeks, and the results are presented in Fig. 3. Onconase at M killed between 1 and 2 logs of cells, whereas a concentration of M killed more than 4 logs of cells. Thus onconase is a cytotoxic protein.

Kinetics of Protein Synthesis Inhibition Caused by Onconuse

A B

FIG. 2. Microautoradiography of control 9L cells uemwon- conase-treated 9L cells. 9L cells were cultured on coverslips and incubated in the absence (panel A ) or presence (panel R) of 1 X lo-' M onconase for 24 h. After pulsing for 2 h with ["C]leucine, cells were washed, fixed, and autoradiographed as described under "Ex- perimental Procedures."

1 2 3

Treatment FIG. 3. Clonogenic growth assay of 9L cells treated with

onconase. 9L cells (1 X IO6 cells/ml) were treated as described under "Experimental Procedures" with control buffer ( l a n e 3, 0.15 M NaCl and 10 mM NaPO,, pH 7.4) or buffer containing onconase (lanes I and 2, respectively, 1 X lo" or 1 X 10" M). After a 24-h treatment the cells were washed, diluted into growth medium, and plated into 96-well microtiter plates (10 wells/dilution) as described under "Ex- perimental Procedures." After 14 days the fraction of surviving cells was calculated using the Spearman-Karber estimator. In this assay a 1-log difference or more is considered statistically significant.

in 9L Cells-The cytotoxicity of plant and bacterial toxins is a time-dependent process, and the kinetics of this process have been carefully examined to probe the mechanism of cell killing. Ricin and diphtheria toxin both display a lag time before protein synthesis inhibition begins, then inhibit pro- tein synthesis by first order kinetics (17, 18). The time course for onconase a t five different concentrations is shown in Fig. 4. Protein synthesis decreases according to a pseudo first- order process with a log linear decrease in leucine uptake with time. However, in contrast to plant and bacterial toxins, onconase does not show any lag time before protein synthesis inhibition begins (Fig. 4). 1 PM onconase was saturating since 10 p~ onconase did not increase the steepness of the slope.

Onconuse Cytotoxicity to 9L Cells Is Mediated by Specific Cell Surface Receptors-Onconase shares 53% amino acid identity with two lectins from the eggs of R. japonica and R. catesbeiana. These two lectins specifically bind to cell surface glycoprotein receptors found on several tumor cell lines (8, 9). The homology between onconase and these two lectins suggests that cell surface receptor binding may be an initial step in the mechanism of onconase toxicity.

We examined whether radiolabeled onconase would bind to 9L cells. Different concentrations of '2sII-labeled onconase were incubated with 9L cells for 2 h a t 4 "C in the presence or absence of a 300-fold excess of unlabeled onconase. As shown in Fig. 5, the binding curve shows a saturable binding of onconase to 9L cells. Scatchard analysis with a versatile computer program for characterization of ligand-binding sys- tems (19) of the data shows that onconase appears to bind to cells with two different affinities ( p < 0.02). one with a Kd of 6.2 X lo-', and another of 2.5 X 10" M, respectively (Fig. 5, irwet). Each cell could bind about 3 x IO5 molecules of onconase at the two affinity sites. The Kd for the low affinity binding is close to the concentration of onconase required to kill cells. The concentration of onconase that yields the max- imal rate of cytotoxicity, 10" M (Fig. 4), correlates well with the concentration of onconase that saturates binding (Fig. 5).

- C Q) L

n e

l ! I 1 -I 0 1 0 20 30

Time (h) FIG. 4. Time course of protein synthesis inhibition caused

by onconase in 9L cells. 9L cells (1 X 10 ' cells/ml) were treated as described in the legend to Fig. 1 and under "Experimental Proce- dures." 0, lo-' M onconase; 0, M onconase; 0, IO-' M onconaw; m, lod M onconase; A, M onconase.

Mechanism of Cytotoxic Ribonuclease 10689

FIG. 5. Binding of radiolabeled onconase to 9L cells. Increasing con- centrations of radiolabeled onconase were incubated with 9L cells at 4 "C for 2 h and the cell-associated radioactivity was determined by subtracting nonspe- cific binding from total radioactivity bound. Inset, Scatchard analysis of the binding data. B, bound; F, free.

n

0

x r

E 0

1

0 2 0 0 400 600 800 I 0 0 0 6

B (tmolasll0 calls)

0 I I I I I

The time course of binding of onconase to cells was exam- ined at 4 and 37 "C (Fig. 6, panel A) . The time needed for maximal binding varies between 120 and 150 min in different experiments at 4 "C. To test whether cell surface binding may be involved in the cytotoxicity of onconase, 9L cells were incubated with onconase or controls for 2 h at 4 "C. After such treatment almost no protein synthesis inhibition of 9L cells by onconase has occurred (Fig. 4). The cells were washed and the incubation continued for 24 h either with or without onconase. The results in Fig. 6, panel B, show that exposure of cells to onconase for only 2 h allows enough binding of onconase to the cells to mediate toxicity 24 h later. Cells were killed by the prebound onconase at a concentration only 4- fold higher than when onconase was continuously incubated with cells, consistent with the model that binding of onconase to the cell surface is an initial step in cytotoxicity. Comparing a 2-h preincubation of onconase with cells at 37 "C, where the binding to cells is significantly weaker than at 4 "C (Fig. 6, panel A ) , with preincubation of onconase with cells at 4 "C shows that subsequent cytotoxicity to cells was greater when the prebinding was a t 4 "C. This further correlates the cell surface binding of onconase with the inhibition of protein synthesis by onconase.

The Effect of NH4C1, Monensin, Metabolic Inhibitors, and Brefeldin A on the Action of Onconme-Endocytosis requires ATP and is blocked by inhibition of both glycolysis and oxidative phosphorylation (20, 21). To test whether endocy- tosis may be a step involved in onconase toxicity, cells were incubated with the glycolysis inhibitor 2-deoxyglucose and the inhibitor of oxidative phosphorylation, NaN3. As shown in Fig. 7, pretreatment of 9L cells with 10 mM NaN3 and 50 mM 2-deoxyglucose blocked onconase toxicity. Thus protein synthesis inhibition by onconase appears to require ATP, consistent with the suggestion that endocytosis may be a step in onconase toxicity. NH4Cl and monensin raise endosomal pH in cells, reduce the proton gradient across the membrane and have powerful effects on the toxicity of protein toxins, such as diphtheria toxin and ricin (21-24). These lysosomo- tropic agents block diphtheria toxin entry into cells and potentiate, or increase ricin entry into cells. Incubation of 9L cells with 25 mM NH4Cl has little effect on the inhibition of protein synthesis by onconase, however, incubation with mo- nensin potentiated the activity of onconase 10-fold. Brefeldin

2 0 0 4 0 0 6 0 0 8 0 0 1 0 0 0 1 2 0 0

Concentration (nM)

A, a fungal metabolite, blocks endoplasmic reticulum to Golgi transport and blocks ricin toxicity (25). Incubation of 9L cells with up to 5.0 hg/ml brefeldin A has no significant effect on the activity of onconase. Thus onconase displays a unique pattern of sensitivity to inhibitors of intracellular vesicular traffic compared with plant and bacterial protein toxins (Fig. 7).

Role of Ribonuclease Activity in Cytotoxicity-Onconase ex- hibits ribonuclease activity suggesting that this activity may play a role in cell toxicity. RNase A has 2 active histidine residues that are susceptible to iodoacetate alkylation, and are both conserved in the onconase sequence (4). Iodoacetate treatment of onconase decreases its ribonuclease activity. We examined the capacity of alkylated onconase to inhibit 9L cell protein synthesis. Alkylated onconase, which only retains 2% of the ribonuclease activity, showed no toxicity a t concentra- tions up to 20-fold higher than the ICso of native onconase (Fig. 1, panel B ) . This is consistent with the model that onconase kills cells by ribonucleolytic action.

Degradation of 9L Cell R N A by Onconase-We examined the ribosomal RNA in both control and onconase-treated cells. There was no detectable degradation of 28 S and 18 S rRNA from untreated, RNase A-treated, or alkylated oncon- ase-treated 9L cells (Fig. 8, lanes 2 and 7-14). However, 28 S and 18 S rRNA are both degraded in onconase-treated cells (Fig. 8, lanes 3-6). Protein synthesis in native onconase- treated cells decreases (Fig. 1, panel B ) in close parallel to 28 S and 18 S rRNA degradation (Fig. 8). There is no detectable degradation of 5.8 S and 5 S ribosomal RNA in onconase- treated cells, and there may be a decrease in transfer RNA in onconase-treated cells (data not shown).

RNA degradation in onconase-treated cells may result from cellular RNases activated by cell death and lysis. However, trypan blue exclusion showed that the 9L cells were intact and viable after treatment for 24 h with onconase a t concen- trations that showed dramatic rRNA cleavage (Fig. 1, panel A ) . We also incubated 9L cells with ricin to abolish protein synthesis and found no degradation of rRNA resulting from activation of intracellular RNases upon protein synthesis inhibition (Fig. 9). Therefore, the rRNA degradation by on- conase appears to be caused directly by onconase that has entered the cytosol compartment.

Sensitivity of Onconase to PRI and Inhibit-AceTM and the

Mechanism of Cytotoxic Ribonuclease

Time (min)

120 -

100 - 80 - 60 -

40 - 20 -

0 - 1 O - l 0 1 0 ~ ~ 1 0 . 8 1 o - ~ 1 o - ~ 1 o- '

Concentration (MI

* ..

FIG. 6. Panel A, time course of '=I-onconase binding to 9L cells. 'mI-Onconase (final concentration, 250 nM) was added to each tube in binding buffer (200 PI) . At the indicated time interval, the reaction was stopped and the cell-associated radioactivity was determined. Panel E , inhibition of protein synthesis in 9L cells by a brief incu- bation with onconase at either 4 or 37 "C. 9L cells (1 X 10 r, cells/ml) were plated and treated as described in the legend to Fig. 1. Different concentrations of onconase were added and cells were incubated either at 37 or 4 "C. After 2 h cells were washed and further incubated a t 37 "C for 24 h in the absence of onconase followed by a 1-h pulse with ["C]leucine. Control cells were incubated continually with on- conase for 24 h a t 37 "C. The data points were determined as described in the legend to Fig. 1.

Demonstration of PRI mRNA in 9L Cells-Specific protein inhibitors of RNase A exist in most tissues in the body (26). These extremely high affinity inhibitors may function to protect cells from the toxicity of members of the RNase A superfamily. The levels of RNase and the inhibitor vary in cells and regulators of cell proliferation can dramatically alter the ratio of these proteins (27,28). PRI has been purified and cloned (29-31), reviewed in Ref. 32, whereas little is known about Inhibit-AceTM, a commercially available protein inhib- itor of RNase A. PRI is about 50,000 daltons in molecular mass and forms a 1:1 complex with several members of the RNase A superfamily. As shown in Fig. 10, RNase A is very sensitive to both PRI and Inhibit-Acem. Onconase is less sensitive than RNase A to both of the two inhibitors tested in vitro (Fig. 10). The concentration of PRI and Inhibit-AceTM sufficient to inhibit activity of 1000 pmol of RNase A has no effect on the activity of 10 pmol of onconase (data not shown). So onconase is a t least 100 times less sensitive than RNase A

Control (16 h) NaN + 2dwoxyglucose (6 h) 3

C

0 0

c 0 80 - ar Y

60 - Q) 0

C

v)

C 0

f * 40 - - - 20 - 2 9.

Monensln + K Concentrrtion (M)

FIG. 7. The effect of NH4CI, monensin, NaNS + 2-deoxyglu- cose, and brefeldin A on the toxicity of onconase. 9L cells were plated in 96-well microtiter plates and incubated in assay medium for 40 min with 25 mM NH4CI (NH.CI), 10 pM monensin (rnonennin), 5.0 pg/ml brefeldin A (brefeMin A ), or without the above additions (control, 16 h ) before onconase was added. Increasing concentrations of onconase were added and further incubated for 16 h. followed by a 1-h pulse with ["Cjleucine. In the case of NaNJ + 2-deoxyglucose, cells were incubated with 10 mM NaNa + SO mM 2-deoxyglucose (NaNa + 2-deoxyglucose) or control buffer (control. 6 h ) for 40 min. then lod M onconase was added and incubated for 6 h followed hy a 2-h pulse with ["Cjleucine.

1 2 3 4 5 6 7 8 91011121314

FIG. 8. RNA degradat ion in 9L cells by onconaee uereud a lky la t ed onconase a n d RNase A. 9L cells were cultured in 75 cmz flasks and treated with varying concentrations of samples as described under "Experimental Procedures." After 24 h cells were washed twice with phosphate-buffered solution and lysed with "Ex- traction Buffer" (RNA extraction kit, Pharmacia). Total RNAs were recovered by centrifuging overnight on a self-generating cesium tri- fluoroacetic acid gradient. Aliquots of the recovered RNAs were loaded on a 1.2% agarose gel. Lane I. calf liver 28 S and 18 S rRNA; lone 2, total RNAs from untreated cells; lones 3-6, total RNAs from 10-9.-8.-7.-6 M onconase-treated cells; lanes 7-10, t o t a l RNAs from 10-9.-8.-7." M alkylated onconase-treated cells; lanes 11-14, total RNAs from 10-9.-R.-'.-fi M RNase A-treated cells.

to both inhibitors tested in vitro. Also in rabbit reticulocyte lysates PRI and Inhibit-AceTM do not inhibit onconase tox- icity (data not shown). RNase A is more active enzymatically than onconase in vitro, but onconase is more cytotoxic than RNase A in cultured 9L cells (Fig. 1). One possible explana- tion for this is the presence of endogenous RNase inhibitors

Mechanism of Cytotoxic Ribonuclease 10691

A

n 1 2 3 4

FIG. 9. Panel A, inhibition of protein synthesis in 9L cells by ricin. 9L cells (1 X 10’ cells/ml) were plated into 96-well microtiter plates and treated with varying concentrations of ricin. After 24 h protein synthesis was determined as described under “Experimental Proce- dures.” Lane 1, control; lanes 2-4, 10”2*-”.”o M ricin. Panel B, RNAs extracted from control and ricin-treated cells. 9L cells (1 X 10 ‘cells/ ml) were plated in 75 cm2 flasks and treated with varying concentra- tions of ricin. After 24 h cells were washed with phosphate-buffered solution twice, and total RNAs were extracted as described under “Experimental Procedures.” Lane 1, calf liver 28 S and 18 S rRNA; lane 2, RNAs from untreated cells; lanes 3-5, RNAs from M ricin-treated cells.

1 2 3 4 5 S 2 3 l S C

FIG. 10. Effeet of Inhibit-AceTM and PRI on onconase and RNase A enzymatic activity. The reaction conditions were as described under “Experimental Procedures.” The substrate used was calf liver 28 S and 18 S rRNA. Panel A: lane I , control; lanes 2 and 3, 1 and 10 pmol of onconase; lanes 4-6, 0.1, 1.0, and 10 pmol of RNase A; lanes 2’ and 3’, 1.0 and 10 pmol of onconase + 2 units of Inhibit-AceTM; lanes 4’-6’, 0.1, 1.0, and 10 pmol of RNase A + 2 units of Inhibit-AceTM. Panel B: lane 1, control rRNA + 10 mM D m , lanes 2 and 3, 1 and 10 pmol of onconase + 10 mM DTT; lanes 4 5 , and 6, 0.1, 1.0, and 10 pmol of RNase A + 10 mM D m , lane 2’ and 3’, 1 and 10 pmol of onconase + 60 units of PRI and 10 mM DTT; lanes 4’, 5’, and 6’, 0.1, 1.0, and 10 pmol of RNase A + 60 units PRI and 10 mM DTT.

that block RNase A cytotoxicity and not onconase cytotox- icity. Northern hybridization with PRI cDNA gene demon- strates that PRI mRNA is present among the total RNA

FIG. 11. Northern hybridization of PRI mRNA with PRI cDNA gene probe. 20 fig of total RNAs extracted from 9L cells wan used for the hybridization with ‘’‘1’ radiolabeled PRI cDNA as de- scribed under “Experimental Procedures.”

extracted from 9L cells consistent with this suggestion (Fig. 11).

DISCUSSION

Native onconase causes protein synthesis inhibition and cell death with an ICW of about 10” M. Microautoradiography also demonstrates that protein synthesis of individual cells is inhibited by onconase treatment without a significant change in cell density. Thus onconase is directly cytotoxic to cells in addition to its cytostatic activity after a 7-day exposure as previously described (1, 3).

The high amino acid identity of onconase with two lectins from eggs of R. japonica and R. catesbeiam suggests a possible step for the mechanism of onconase action. These two lectins are 80% identical in sequence to one another and showed 26- 28% sequence identity to RNase A and exhibit ribonuclease activity with a 10-fold lower specific activity than RNase A for high molecular weight substrates (5-7, 10). Onconase is about 53% identical in sequence to these lectins with all four disulfide loops conserved. These two lectins were originally purified because they preferentially agglutinate some tumor cells. Receptors for these two lectins exist on some tumor cells and have been purified (8, 9). Receptors have also been reported for bovine seminal RNase, unpublished data cited in Ref. 33, and angiogenin (34), which share 29 and 27% se- quence identity with onconase, respectively, reviewed in Ref. 32. These similarities suggest that onconase may also bind to cell surface receptors. The presence of specific receptors on 9L cells was demonstrated by the direct binding assay of labeled onconase to cultured cells. We not only find that onconase binds to the cell surface, but that the optimal conditions of cell surface binding correlate with optimal con- ditions for protein synthesis inhibition. It therefore appears that cell surface receptor binding is the first step in onconase cytotoxicity. Nothing is known about the nature of onconase receptors on mammalian cell surfaces. Onconase may bind to cell surface carbohydrates as in the case of ricin, or it may bind to receptors originally developed for physiologically im- portant molecules like polypeptide hormones, such as growth factors as proposed previously (35). Indeed, the receptors identified on the outer membrane of susceptible bacteria for colicin E3 and cloacin DF1.7, which are also ribonucleases, are the receptors for vitamin B12 and for ferrichrome aero- bactin, respectively (36, 37).

Alkylated onconase, which only retains 2% of the ribonu- clease activity, exhibits less than 5% of native onconase cytotoxicity, suggesting that RNase activity is necessary for onconase toxicity. Direct examination of RNA in onconase- treated cells shows that the 28 S and 18 S ribosomal RNA are degraded, and protein synthesis in onconase-treated cells decreases in parallel to the major rRNA degradation. This

10692 Mechanism of Cytotoxic Ribonuclease

RNA degradation could result from cellular ribonucleases that may be activated upon cell death and lysis as a result of the inhibition of protein synthesis. However, two experiments argue against such a possibility. First, RNA degradation of onconase-treated cells occurred before the cells had become permeable to the vital dye, trypan blue, and second, ricin causes complete inhibition of protein synthesis without show- ing any degradation of rRNA which may result from the activation of endogenous ribonuclease because of cell death and lysis. Therefore, onconase appears to reach the cell cy- tosol and cleave cellular RNA. This cellular lesion can, in turn, inhibit protein synthesis and lead to cell death. Although there is no detectable degradation of 5.8 S and 5 S ribosomal RNA, there appears to be some decrease of transfer RNA in onconase-treated cells. In the reticulocyte lysate, tRNA ap- pears to be more sensitive to onconase than rRNA.3 tRNA may also be more sensitive in cells, however, this may not be rate-limiting for protein synthesis. tRNA may be continually transcribed in response to degradation by onconase and only at doses of onconase that cleave rRNA is translation blocked leading to cell death. This may account for the fact that protein synthesis inhibition in cells by onconase correlates with the dose that results in rRNA degradation.

The presence of specific cell surface receptors and the degradation of rRNA in onconase-treated cells indicates that onconase must cross the plasma membrane or an endocytic vesicle membrane to reach the cytosol compartment. Plant and bacterial protein toxins, such as ricin, diphtheria toxin, and Pseudomonas exotoxin A, are enzymes that inhibit pro- tein synthesis by specifically cleaving a single N-glycosidic bond of the 28 S rRNA (39), or by ADP-ribosylation of elongation factor 2 (21,40,41). These proteins traverse intra- cellular membranes following receptor-mediated endocytosis to reach the cytosol. Onconase is also an enzyme that inhibits protein synthesis. It is not known if onconase can cross the plasma membrane directly or if it enters cells initially via endocytosis. The inhibition of onconase cytotoxicity by NaN3 and 2-deoxyglucose, inhibitors of ATP synthesis and endo- cytosis, may indicate that receptor-mediated endocytosis is a requisite step in the pathway. Monensin, a carboxylic iono- phore that alters intracellular trafficking potentiates oncon- ase cytotoxicity 10-fold, again consistent with the model that endocytosis is involved in onconase toxicity. However, the overall pattern of sensitivity of onconase endocytosis to in- hibitors differs from that of diphtheria toxin, Pseudomonas exotoxin A, and ricin.

RNase A is less toxic than onconase although RNase A is enzymatically more active. One significant difference in the structure between onconase and RNase A is that the disulfide loop CS5J& in RNase A is missing in onconase and frog lectins, reviewed in Ref. 32. Angiogenin also lacks the same disulfide loop. A new disulfide loop at the C-terminal exists in onconase and frog lectins. Introducing a region of RNase A containing the CG54& disulfide bond into angiogenin by regional mutagenesis makes angiogenin become a more potent degradative enzyme (42). This indicates that this disulfide loop present in RNase A is important for its high enzymatic activity and may explain why onconase and frog lectins are less active enzymatically than RNase A. Why does RNase A demonstrate such a low toxicity? Many factors may cause the differences of cytotoxicity between onconase and RNase A. We have examined the potential role that endogenous ribo- nuclease inhibitors may play in cell susceptibility to ribonu- cleases. PRI and Inhibit-AceTM are two known protein RNase

S. M. Rybak, D. L. Newton, S. M. Mikulski, W. Ardelt, and R. J. Youle, manuscript submitted for publication.

inhibitors. PRI is a 50-kDa protein with a high leucine and cysteine content. It forms a 1:l complex with several members of the RNase A superfamily. It displays a very low K, for angiogenin- and eosinophil-derived neurotoxin (also a ribo- nuclease) (43), and a slightly higher Ki for RNase A of 3 X 10"' M (29, 30). PRI was originally purified from human placenta and is widely distributed in other tissues suggesting it may function to protect cells from the toxicity of members of the RNase A superfamily. Little is known about Inhibit- AceTM, another protein inhibitor of the RNase A superfamily. We have found that onconase is resistant to both PRI and Inhibit-AceTM relative to RNase A. In addition, 9L cells express the PRI mRNA. This may be one of the key factors for the ability of onconase to kill mammalian cells. RNase A, although more active enzymatically than onconase, needs a much higher concentration to exert its cytotoxicity to mam- malian cells possibly because of the presence of the endoge- nous protein inhibitors. The PRI-RNase interaction appears to exhibit an interesting species specificity. Chicken and frog ribonuclease inhibitors appear not to inhibit mammalian RNase A but do inhibit their own species ribonuclease (44, 45).

Another level of resistance to RNase A may exist at the cell surface. RNase A cannot compete for onconase toxicity indicating that RNase A does not bind to the same receptor (Fig. 1, panel B ) . As reported previously, pancreatic RNase A conjugated with human transferrin or transferrin receptor antibodies can inhibit protein synthesis in cultured cell lines at much lower concentrations than native RNase A (38, 46). Transferrin can mediate cell surface binding and RNase A endocytosis resulting in intracellular accumulation of suffi- cient amounts of RNase A that apparently exceed the level of endogenous RNase inhibitors. Other possibilities are that conjugation of RNase A with other proteins blocks the RNase A binding domain for ribonuclease inhibitors, making RNase A more resistant to these inhibitors, or the cell line (K562 cell) used to assay RNase A conjugates has low level of PRI. Northern hybridization with PRI cDNA probe demonstrated a positive PRI mRNA signal in 9L cells, however, K562 cells have not been examined to date.

Taken together these results suggest a model for the mech- anism of onconase action. Onconase specifically binds to cell surface receptors, is internalized into the cell cytosol where it degrades RNA, inhibiting protein synthesis and killing the cell.

Acknowledgments-We thank Drs. Massimo Gadina and Ping Chen for performing Northern hybridization of PRI mRNA and for their critical reading and discussions of this manuscript; Dr. Cynthia Sung for her help with Scatchard analysis; Dr. Katherine Wood for critically reading this manuscript; Patricia Johnson for her technical help and other members of the Biochemistry Section of National Institute of Neurological Disorders and Stroke, Surgical Neurology Branch, for useful discussions.

REFERENCES 1. Darzynkiewicz, Z., Carter, S. P., Mikulski, S. M., Ardelt, W. J., and Shogen,

2. Mikulski. S. M.. Ardelt. W. J.. and Shoaen, K. (1990) J. Natl. Cancer Inst. K. (1988) Cell. Twsue Kinet. 21,169-182

8 2 , 1 5 i - m ' - .

3. Mikulski, S. M., Viera? A., Ardelt, W. J., Menduke, H., and Shogen, K.

4. Ardelt, W., Mikulski, S. M., and Shogen, K. (1991) J. Biol. Chern. 2 6 6 , (1990) Cell. Tissue Ktnet. 23,237-246

5. Tltani, K., Takio, K., Kuwada, M., Nitta, K., Sakakibara,, F., Kawauchi, 245-251

H., Takayanagi, G., and Hakomon, S. I. (1987) Bmhernwtry 26,2189- 2194

6. Kamiya, Y., Oyama, F., Oyama, R., Sakakibara, F., Nitta, K., Kawauchi, H., Takayanagi, Y., and Titani, K. (1990) J. Biochern. (Tokyo) 108,139-

7. Lewis, M. T., Hunt, L. T., and Barker, W. C . (1989) Protein Seq. Data 143

8. Sakakibara, F., Kawauchi, H., Takayanagi, G., and Ise, H. (1979) Cancer Anal. 2,101-105

Res. 39,1347-1352

Mechanism of Cytotoxic Ribonuclease 10693 9. Nitta, K., Takayanagi, G., Kawauchi, H., and Hakomori, S. (1987) Cancer

Res. 47,4877-4883 10. Okabe, Y., Katayama, N., Iwama, M., Watanabe, H., Ohgi, K., hie, M.,

Nitta, K., Kawauchi, H., Takayanagi, Y., Oyama, F., Titani, K., Abe, Y., Okazaki, T., Inokuchi, N., and Koyama, T. (1991) J. Biochem. (Tokyo) 109,786-790

11. Crestfield, A. M., Stein, W. H., and Moore, S. (1963) J. Biol. Chem. 238, 2413-2420

12. Anfinsen, C. B., Redfield, R. R., Choate, W. L., Page, J., and Carrol, W. R. (1954) J. Biol. Chem. 207,201-210

13. Blank, A., and Dekker, C. A. (1981) Biochemistry 20, 2261-2267 14. Johnson. V.. Wilson. D., Greenfield, L., and Youle, R. J. (1988) J. Biol.

15. Bregni, M., De Fabritiis, P., Raso, V., Greenberger, J., Lipton, J., Nadler, Chem. 263, 1295-1300

L.. Rothstein. L.. Ritz. J.. and Bast. R. C.. Jr. (1986) Cancer Res. 46. , , . , 1208-1213

16. Johnson, E., and Brown, B. (1961) Biometrics 17,79-88 17. Olsnes, S., Sandvig, K., Refsnes, K., and Pihl, A. (1976) J. Biol. Chem.

251.3985-3992 18. Uchida, T., Pappenheimer, A. M., Jr., and Harper, A. A. (1973) J. Biol.

20. Sandvig, K., and Olsnes, S. (1981) J. Biol. Chem. 256,9068-9076 19. Munson, P. J., and Rodbard D. (1980) A d . Biochem. 107,220-239

21. Johnson, V., and Youle, R. J. (1991) Intracellular Trafficking of Proteins, pp. 183-216, Cambridge University Press, Cambridge

22. Kim, K., and Groman, N. (1965) J. Bacteriol. 90, 1557-1562 23. Sandvig, K., Sudan, A., and Olsnes, S. (1984) J. Cell Biol. 98,963-970 24. Sandvig, K., and Olsnes, S. (1982) J . Biol. Chem. 267,7504-7513 25. Hudson, T. H., and Grillo, F. G. (1991) J. Biol. Chem. 266, 18586-18592 26. Kawanomoto, M., Motojima, K., Sasaki, M., Hattori, H., and Goto, S.

Chem. 248,3845-3852

(1992) Btochrm. Biophys. Acta 1129,335-338

27. Kyner, D., Christman, J. K., and Acs, G . (1979) Eur. J. Biochem. 99,395-

28. Kraft, N., and Shortman, K. (1970) Biochim. Bio h s Acta 217, 164-175 29. Blackburn, P., Wllson, G., and Moore, S. (1977) if. &ol. Chem. 252,5904-

399

591 n 30.

31.

32.

33.

34.

35. 36. 37.

38.

39.

40.

41.

42. 43. 44. 45.

46.

Lei;kI S., Fox, E. A,, Zhou, H. M., Strydom, D. J., and Vallee, B. L. (1988)

Schneider, R., Schneider-Schemer, E., Thurnher, M., Auer, B., and

Youle, R. J. Newton, D., Wu, Y. N., Gadina, M., and Rybak, S. M. (1993)

Laccetti, P., Portella, G., Mastronicola, M. R., Russo, A., Piccoli, R.,

Hu G.-F. Chan S I , Riordan, J. F., and Vallee, B. L. (1991) Proc. Natl.

Olsnes, S., Refsnes K., and Pihl, A. (1974) Nature 249,627-631 Watson, D. H., and Sherratt, D. J. (1979) Nature 278, 278-280 Van Tiel-Menkveld, G. J., Ment'ox Vervuurt, J. M., Oudega, B., and de

Newton, D., Ilercd, O., Laske, D. W., Oldfield, E., Rybak, S. M., and Youle,

Endo, Y., Mitsui, K., Motizuki, M., and Tsurugi, K. (1987) J. Bid. Chem.

Colher, R. (1988) Immunotoxins, pp. 25-35, Kluwer Academic Publishers,

Biochemlstry 27,8545-8553

Schweiger, M. (1988) EMBOJ. 7,4151-4156

Crit. Reu.'Ther. Drug Carrier Syst., in press

D'Alessio, G., and Vecchio G. (1992) Cancer Res. 52, 4582-4586

.&ad. dci. U. 5: A:-&, 2227-2231

Graaf, F. (1982) J. Bacteriol. 180;490-497

R. J. (1992) J. Biol. Chem. 267, 19572-19578

262, 5908-5912 Rnatnn

Honjo, T., Nishizuka, Y., Hayaishi, O., and Kato, I. (1968) J. Biol. Chem.

Harper, J. W. and Vallee, B. L. (1989) Biochemist 28, 1875-1884 Sha iro R and Vallee, B. L. (1991) Biochemistry YO, 2246-2255

Dijkstra, J., Touw, J., Halsema, J., Gruber, M., and Ab, G. (1978) Biochim. Rot!, J:S.'i1962) Biochim. Biophys. Acta 61,903-915

Rybaf, J..M., Saxena, S. K., Ackerman, E. J., and Youle, R. J. (1991) J.

l"u"v.. 243,3553-3555

Bio h s Acta 621,363-373

Biol. Chem. 266,21202-21207