Isolation of Chinese Hamster Ovary Ribosomal Mutants Differentially ...

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1990 by The American Society for Biochemistry

Isolation Resistant

Vol. 265, No. 1, Issue of January 5, pp. 582-5&3,1990

and Molecular Biology, Inc.

Printed in U. S. A.

of Chinese Hamster Ovary Ribosomal to Ricin, Abrin, and Modeccin*

Mutants

Differentially

(Received

Sandra From

Sallustio

the Department

and Pamela

Stanley

of Cell Biology,

Albert

Einstein

College

The molecular action of ricin A chain involves cleavage of the N-glycosidic bond between ribose and the adenine 4324 nucleotides from the 5’ end of mammalian 28 S rRNA (Endo, Y., and Tsurugi, K. (1987) J. Biol. Chem. 262, 8128-8130). In this paper, four ritin- and abrin-resistant Chinese hamster ovary cell mutants that possess ribosomes resistant to this Nglycosidase action are described. Three of the mutant phenotypes, Lec26, Lec27, and Lec28, were recessive in somatic cell hybrids and define at least two new lectin-resistant complementation groups. The most extensively characterized mutant type, LEC17, was dominant in such hybrids. None of the mutants were crossresistant to modeccin. Post-mitochondrial supernatants from each of the four mutants were resistant to inhibition of cell-free protein synthesis by ricin, ricin A chain, and abrin. In addition, polysomes isolated from mutant cells were resistant to cleavage of the adenine-ribose N-glycosidic bond by ricin A chain or abrin, as assayed by the release of an -470-nucleotide fragment following aniline treatment of ribosomal RNA extracted from toxin-treated polysomes. The unique lectin-resistance properties of the different mutants suggests that the accessibility of adenine 4324 to each toxin differs. It seems likely that the recessive Chinese hamster ovary ribosomal mutants reflect structural changes in different ribosomal proteins while the dominant phenotype may be due to the modification of protein(s) or rRNA involved in toxin-ribosome interaction. Further analysis of these cell lines should provide new insights into the structure/function relationships of eukaryotic ribosomes.

of Medicine,

Bronx,

New

York

for publication,

June

29, 1989)

10461

structural details of toxin-ribosome interactions, have yet to be elucidated. However, the importance of higher order structure of the ribosomal domain recognized by ricin A chain is apparent from the finding that intact ribosomes are nearly lO,OOO-fold more sensitive to depurination of 28 S rRNA than naked rRNA (5, 6). It is of considerable interest, therefore, to identify the ribosomal proteins in the ricin-binding domain and to characterize the role of these proteins in toxin sensitivity. In this paper, we describe four new CHO’ cell mutants that should prove valuable in elucidating the structure-function relationships of ribosomal proteins located near the adenine cleaved by the N-glycosidase activity of ricin, abrin, and modeccin. The mutants were isolated in four independent selections for toxin resistance. Three behave recessively in somatic cell hybrids with parental cells whereas the fourth acts dominantly in hybrids. Protein synthesis in post-mitochondrial (SlO) supernatants of all four mutant lines was inhibited to a lesser extent than in SlO extracts from parent cells when incubated with ricin, ricin A chain, or abrin. Similarly, intact polysomes from mutant lines were less sensitive than those from parent cells to depurination of 28 S rRNA at a site located -470 nucleotides from the 3’ end of the RNA when treated, in uitro, with ricin or abrin. Because of their unique patterns of cross-resistance to ricin and abrin, but not to other carbohydrate-binding proteins, it seems likely that these mutants have identified four distinct gene products affecting the accessibility of adenine 4324 in 28 S rRNA to the toxins. In addition, these mutants provide evidence that the three highly homologous A chains of ricin, abrin, and modeccin interact with different microenvironments of the 60 S ribosomal subunit to exert their N-glycosidase activity. EXPERIMENTAL

Ricin and its structural and functional analogues abrin and modeccin are potent cytotoxins that consist of two, distinct, disulfide-linked peptides (A and B), each of -30 kDa (1, 2). The B chain binds to Gal or GalNAc residues present on glycoconjugates at the surface of eukaryotic cells (1) and facilitates the internalization and transport of the heterodimer (3). Within the cytoplasm, the A chain catalytically inactivates the 60 S ribosomal subunit by cleaving the Nglycosidic bond between the ribose and adenine 4324 nucleotides from the 5’ end of 28 S rRNA (4). The molecular events that follow the depurination of the rRNA and functionally inactivate the ribosome for protein synthesis, as well as the

PROCEDURES

Materials-Sodium [‘251]iodide (100 mCi/ml), L-[4,5-3H]leucine (120-190 Ci/mmol, 1 mCi/ml), human placental ribonuclease inhibitor were purchased from Amersham Corp. BSA (fraction V), the lectins from Ptilota plumosa, Robinia pseudoaccacia, Viscum album, and Triticum vulgar+ (WGA), were obtained from Sigma; ricin, ricin A chain, the lectins from Erythrina cristagalli, LCA, L-PHA, E-PHA, Ricinus communis agglutinin-agarose and ricin-agarose were from Vector (Burlingame, CA). EMS was obtained from Eastman Kodak (Rochester, NY); Pronase and RNA molecular weight markers (III) were from Boehringer Mannheim; ConA-Sepharose and Staphylococcus aureus micrococcal nuclease were from Pharmacia LKB Biotechnologies Inc.; rabbit P-globin mRNA was from Bethesda Research Laboratories; abrin from EY Laboratories (San Mateo, CA.); aniline was from Aldrich and monensin was from Behring Diagnostics. Alpha

* This work was supported by MSTP Training Grant T32GM7288 from the National In&utes 0; Health (to S. S.,, Grants ROl 30645 and ROl 35434 (to P. S.) from the National Cancer Institute, an Irma T. Hirsch1 Faculty Award (to P. S.), and Cancer Core Grant PO1 13330. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1 The abbreviations used are: CHO, Chinese hamster ovary; BSA, bovine serum albumin; FCS, fetal calf serum; L-PHA, leukophytohemagglutinin; E-PHA, erythrophytohemagglutinin; WGA, wheat germ agglutinin; LCA, Lens culinaris agglutinin; ConA, concanavalin A; EMS; ethylmethanesulfonate; Hepeii 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; PBS, phosphate-buffered saline; EGTA, [ethylenebis(oxyethylenenitrilo)]tetra&etic acid.

582

Ricin-resistant

Ribosomal Mutants

(01) medium was purchased from GIBCO; horse serum was obtained from Hazleton Biologics (Lenexa, KS), and FCS was from FLOW Laboratories (McLean, VA). Pseudomonas toxin and modeccin were generously donated by Dr. April Robbins (National Institutes of Health, Bethesda, MD); pactamycin was a gift from Upjohn. Cell Lines-The derivations of parent CHO cell lines Pro-5, Gat-2, and the glycosylation mutant Pro-LEC10.3C (LECLO) were previously reported (7, 8). Cells were routinely cultured in suspension or on loo-mm tissue culture dishes in an atmosphere of 5% CO, in air at 37 “C in a medium containing 10% (v/v) FCS (alO%FCS) or 10% (v/v) horse serum with 1% (v/v) FCS. No mycoplasma contamination was detected by Hoechst staining of indicator 3T6 cells grown in the presence of test cells for 23 days (9). Mutagenesis-Suspension cultures of 2-2.4 x lo7 Gat-2 and Pro-5 cells, at a density of 2 x 10m5 cells/ml, were mutagenized with 100 @g/ml EMS at 34 “C for 18 h. Cells were washed free of mutagen and cultured for an expression period of 2-4 days in alO%FCS. The relative plating efficiency of the two cell lines immediately following EMS treatment was -100%. The efficacy of the mutagen treatment was estimated by comparing the number of surviving colonies in plates containing 104, 105, and lo6 cells incubated at 37 “C for 8-10 days in alO%FCS with or without 10 Kg/ml WGA. EMS-treated Pro-5 cells had -20-fold more WGA-resistant colonies than unmutagenized Pro-5 cells. Selection of Z&in-resistant (Ri?) Mutants-Four new Rick mutants termed LEC17, Lec26, Lec27, and Lec28 were isolated from different selections. The LEC17 mutant type was isolated serendipitously following an attempt to transfect the LEClO lectin-resistance (LecR) phenotype into parent CHO cells, Pro-5 cells at a density of 2 X 10’ cells/ml were cultured in suspension in alO%FCS containing 4 Kg/ ml E-PHA for -72 h. Cells were subsequently plated at lo6 cells/lOOmm tissue culture dish in alO%FCS and incubated at 37 “C for 8 h, when -40 pg of LEClO genomic DNA (in the form of a calcium phosphate coprecipitate) was added to each plate. Cells were exposed to DNA for 24 h; medium was then aspirated from plates and replaced with fresh otlO%FCS. Following a 36-h incubation period, the medium was aspirated and replaced with alO%FCS containing 12 pg/ml ConA and 50 rig/ml ricin. After incubation for 7 days at 37 “C in a humidified 5% CO* atmopshere, the medium was replaced with fresh lectincontaining medium and the incubation continued at 37 “C for 14 days. At this time, surviving colonies were picked, transferred to 17 X loo-mm sterile polystyrene tubes in alO%FCS containing 12 pg/ ml ConA and 50 rig/ml ricin, and incubated at 37 “C. The selection schemes to obtain Lec26, Lec27, and Lec28 were designed to specifically prevent the survival of frequently arising glycosylation mutants such as Lecl and LEClO (a), and to enhance the survival of rare, ricin-resistant mutants. The Lec26 mutant type was obtained when EMS-mutagenized Pro-5 cells at a density of lo5 cells/ml were grown in suspension in alO%FCS containing 7.5 pg/ml ConA for 36 h, centrifuged out of ConA and subsequently plated at 2 X lo5 cells/lOO-mm tissue culture dish in ollO%FCS containing 20 rig/ml ricin and 5 rig/ml abrin. Following an incubation period of 14 days at 37 “C, selective media was aspirated from the plates and replaced wit,h alO%FCS containing 6 pg/ml E-PHA. After an additional 3-day incubation period, surviving colonies were picked, transferred to sterile tubes in culO%FCS, and maintained at 37 “C. The Lec27 mutant type was isolated following plating of 5 x lo5 EMS-mutagenized Gatt2 cells/lOO-mm tissue culture dish in otlO%FCS containing 7.5 pg/ml ConA and 75 rig/ml ricin. After incubation at 37 “C for 14 days, surviving colonies were picked into elO%FCS as described above. The Lec28 mutant type was obtained when EMS-mutagenized Gat-2 cells, at an initial density of 5 x lo4 cells/ml, were cultured in suspension for 4 days in nlO%FCS containing 5 pg/ml ConA, centrifuged out of ConA, and plated at 10fi cells/lOO-mm tissue culture dish in alO%FCS containing 20 rig/ml abrin. After incubation at 37 “C for 6 days, the selective medium was aspirated from the plates and replaced with nlO%FCS containing 10 rg/ml E-PHA. After 2 days at 37 “C, surviving colonies were picked, transferred to sterile tubes in nlO%FCS containing 7.5 pg/ml ConA and cultured at 37 “C. Colonies picked from selection plates into 17 x loo-mm sterile tubes were transferred after -5 days in 10 ml of culO%FCS or 010% horse serum l%FCS to 16 x 25-mm tubes for suspension culture and subsequently cloned by limiting dilution in 96.well microtiter dishes (8). The four clones characterized in these studies were ProLEC17.4A, Pro-Lec26.7C, Pro-Lec27.1B, and Gat-Lec28.1A. Genetic Characterization of Mutants-The concentration of LPHA, WGA, abrin, ricin, modeccin, ConA, LCA, or E-PHA required

of CHO Cells

583

to reduce survival of cells by 90% (termed the Die value) was determined by titration in 96-well microtiter dishes (P-test) as previously described (8). Briefly, 2 x lo3 cells were plated in 200 ~1 of alO%FCS with increasing concentrations of a given lectin and incubated at 37 “C until cells in the culO%FCS control well became confluent. The proportion of surviving cells in lectin-containing wells relative to survivors in control wells was determined after staining of dishes with 0.2% (w/v) methylene blue in 50% methanol. Somatic cell hybrids between each of the four new mutants and Gat-2 or Pro-5 parental CHO cells and between Lec26 and Lec27 as well as Lec26 and Lec28 cell lines were made using the polyethylene glycol/dimethyl sulfoxide method (8). Hybrids were selected in (Y medium lacking glycine, adenosine, thymidine, and proline and supplemented with 10% dialyzed FCS (hybrid medium). Control plates were included in each experiment to determine the frequency of reversion to prototrophy and the frequency of spontaneous hybrid formation. Hybrid colonies were picked from tissue culture plates, maintained in hybrid medium in sterile tubes at 37 “C, and grown in monolayer or suspension in alO%FCS or hybrid medium. The Lecn phenotypes of hybrids were determined from P-tests (8). Modal chromosome numbers were counted from spreads of hybrids treated with colcemid, as described previously (8). Binding and Internalization of ‘Z5Z-Ricin by Intact Cells-Ricin was iodinated to a specific activity of 6.2 X lo6 cpm/pg protein using the chloramine-T method as described previously (10). The amount of iz51-ricin bound by parent, LEC17, and LEClO cells after 1 h at 4 “C was determined by centrifugation through BSA (10). The cpm recovered in the cell pellet divided by the specific activity of iz51-ricin and by the number of cells in the final reaction were plotted against the total amount of ricin (cpm in the pellet and cpm in the supernatant divided by the specific activity of Y-ricin) to determine binding. The amount of cell-associated (cell-surface-bound plus internalized) ‘?ricin was determined as follows: LEC17 and Pro-5 cells (6 x lo6 cells each) were incubated at 37 “C for 1 h in a final volume of 7 ml of phosphate-buffered saline containing 2% (w/v) BSA (PBS/ BSA) and 14.4 rig/ml iz51-ricin (specific activity of 1.1 x lo7 cpm/pg). Aliquots (6 X 1 ml) of the reaction mixtures for each cell line were centrifuged for 5 min at 575 x g and 4 “C to pellet the cells. Each cell pellet was washed once at 4 “C with 1.0 ml of PBS/BSA, and the cellassociated iz51-ricin determined by counting duplicate samples in an LKB gamma counter. The remaining samples were resuspended in 0.5 ml of PBS/BSA in the presence or absence (in duplicate) of 0.1 pg/ml Pronase and 0.1 M Gal (to remove bound, but not internalized, ““I-ricin). Following a 15.min incubation at 4 “C, cells were pelleted by centrifugation at 575 x g for 5 min at 4 “C. Each cell pellet was washed once with 0.5 ml of PBS/BSA and the cpm associated with the cell pellet determined. These cpm represented the amount of iz51ricin internalized by the cells during the original incubation at 37 “C. Cell-free in Vitro Protein Synthesis-Post-mitochondrial (SlO) supernatants were prepared from exponentially growing CHO cells as described (11, 12). The extracts were assayed for their ability to incorporate [3H]leucine into trichloroacetic acid-precipitable cpm in the absence or presence of ricin, ricin A chain, abrin, or modeccin. The in vitro translation mixtures consisted of -6 AzeO units of an SlO supernatant, 25 mM K+Hepes, pH 7.6,llO mM KCl, 1 mM magnesium acetate, 1 mM ATP, 0.25 mM GTP, 0.4 mM spermidine, 4% (v/v) glycerol, 5 mM creatine phosphate, 180 pg/ml creatine kinase, 6 mM /Y-mercaptoethanol, and 5 &i of [3H]leucine and water or toxin in a final reaction volume of 100 ~1. After incubation for various times at 30 “C, 10.~1 aliquots were removed in duplicate at different times, added to 1.0 ml of 3% (w/v) casamino acids in water and mixed with 1.0 ml 20% (w/v) trichloroacetic acid. The trichloroacetic acid precipitates were collected on presoaked GF-C glass fiber filters, dried, and counted in a toluene-based liquid scintillant. The effect of ricin A chain on the elongation step of protein synthesis was determined by carrying out in vitro translation reactions in the presence of 3 pM pactamycin, an inhibitor of initiation of eukaryotic translation (13). SlO supernatants (-20Agw units) from parent CHO cells were brought to 1 mM CaC& and incubated with 15 units of S. nurem micrococcal nuclease at 20 “C for 12 min. The Ca’+dependent nuclease was inactivated by the addition of 0.1 M EGTA to a final concentration of 2 mM, and nuclease-treated extracts were assayed for their ability to translate exogenously added, rabbit globin mRNA (7.5 pg/ml). At this concentration of pactamycin, only background levels of [“Hlleucine were incorporated into trichloroacetic acid-precipitable cpm when exogenous mRNA was added to the translation mixture. RNA Isolation-Total RNA was isolated from 1.3 x 10R exponen-

584

Ricin-resistant

Ribosomal Mutants

tially growing CHO cells by the guanidinium method exactly as described (14) and stored at -20 “C as an aqueous solution. Preparation of Polysomes-Polysomes were prepared at 4 “C from mutant and parent CHO cells by a modification of a previously described procedure (5). Exponentially growing cells (3-4 x 109) were harvested by centrifugation at 370 X g for 10 min and washed with 0.15 M NaCl. Following centrifugation at 1600 x g for 10 min, the cell pellet from 10’ cells was resuspended by vortexing in 75 ~1 of 2.5% (v/v) Nonidet P-40 in 0.2 M sucrose in buffer A (20 mM TrisCl, pH 7.6, 5 mM magnesium acetate, 100 mM NH4Cl, 1.0 mM dithiothreitol). The mixture was centrifuged at 2300 x g for 10 min, and the upper 2/3 of the supernatant (PMS) collected. Sodium deoxycholate (13% w/v) at l/9 the volume of the PMS was added dropwise to this supernatant which was incubated for an additional 30 min in an ice bath with stirring. The deoxycholate-treated PMS was layered on a discontinuous sucrose gradient of 7 ml of 0.7 M sucrose in buffer A over 7 ml of 2.0 M sucrose in buffer A in Beckman SW-28 polyallomer tubes and centrifuged at 105,000 x g for 24 h. The polysome pellet was washed with an equal volume of distilled, deionized HgO, resuspended in 1.0 ml of 0.2 M sucrose in buffer A, and stored in loo-150 al-aliquots at -70 “C. Sedimentation profiles of this preparation on continuous gradients of 15-33% (w/w) sucrose in 0.15 M NaCl, 10 mM Tris-Cl, pH 7.4, 1.5 mM MgCl, centrifuged at 25,000 rpm in a Beckman SW-27 rotor for 90 min, confirmed that most of the preparation consisted of polysomes. Treatment of Polysomes with Toxins-Treatment of polysomes with ricin or abrin was carried out essentially as described (15). Polysomes (-3.0 A260 units) were incubated for 15 min at 37 ‘C in a final reaction volume of 50 pl of buffer B (113 mM KCl, 10 mM MgCl,, 6 mM @mercaptoethanol, 25 units of human placental ribonuclease inhibitor) lacking or supplemented with various concentrations of toxins. The reactions were terminated by the addition of 550 ~1 of 0.5% (w/v) sodium dodecyl sulfate in 50 mM Tris-Cl, pH 7.6, followed by- 600 ~1 of buffered phenol. The mixture was vortexed, and the organic and aqueous phases separated by centrifugation at 1600 X g for 10 min. The RNA in the aqueous phase was recovered following precipitation with two volumes of absolute ethanol at -20 “C. Removal of Ethanol-soluble Ribosomal Proteins from PolysomesAcidic ribosomal proteins were extracted from polysomes by a modification of a previously described procedure (16). All steps in the extraction protocol were carried out at 4 “C. Aliquots of 150 ~1 thawed polysomes were mixed with 64.3 ~1 of 0.7 M sucrose in buffer A containing 5.8 pl of 7.0 M KCl, 1.3 ~1 of 0.5 M P-mercaptoethanol, and 221.4 ~1 of absolute ethanol and stirred in an ice water bath for 30 min. The mixture was centrifuged at 15,000 rpm for 10 min in an SS-34 rotor, and the pellet resuspended in 150 ~1 of 0.2 M sucrose in buffer A. Buffer A, containing 0.7 M sucrose, 2.0 M KCl, 0.5 M pmercaptoethanol and absolute ethanol (in the volumes given above), was added to the resuspended pellet and the mixture incubated with stirring in an ice/Hz0 bath for an additional 30 min. The extracted polysomes were centrifuged as above and the pellet resuspended in 110 ~1 of 0.2 M sucrose in buffer A for immediate treatment with ricin as described above. Aniline Treatment and Gel Electrophoresis of Ribosomal RNARNA extracted from polysomes or cytoplasm (-6.6 rg) was incubated on ice for 10 min with 13.2 ~1 of 1.0 M aniline buffer, pH 5.1 (prepared freshly by the addition of 0.5 ml of glacial acetic acid, 7.0 ml of distilled deionized water, 1.0 ml of purified aniline stored under nitrogen, and 2.5 ml of 1.0 M acetic acid, in that order). The reaction was terminated with the addition of 200 ~1 of distilled deionized water, and the aniline was removed by evaporation of the sample to dryness in vacua. Aniline-treated RNA was resuspended in 15 ~1 of gel loading buffer (100 mg/ml sucrose, 7 M urea, 0.4 pg/ml bromphenol blue, 89 mM Tris, 89 mM boric acid, 2.5 mM EDTA, pH 8.3), incubated at 90 “C for 30 s, immediately chilled on ice, and electrophoresed for 3 h at 30 mA constant current in gels of 5% (4.85% (w/v) acrylamide monomer, 0.15% (w/v) bisacrylamide) acrylamide containing 7 M urea in 89 mM Tris, 89 mM boric acid, 2.5 mM EDTA, pH 8.3. Negatives of photographs of gels stained for 30 min with 1 @g/ml ethidium bromide were scanned with an LKB 2222-010 Ultro-Scan XL densitometer to approximate the relative amount of -470 nucleotide fragment obtained compared with an internal standard (5 S rRNA). RESULTS

From a selection CHO cells, a colony

designed to with a novel

obtain highly lectin-resistance

ricin-resistant phenotype

of CHO Cells

was identified. Clones of the isolate were -35-fold more ricinresistant and -4-fold more abrin-resistant than CHO parent cells, but were not altered in their resistance to other carbohydrate-binding proteins such as L-PHA, ConA, WGA, and LCA (Table I). Somatic cell hybrids between a clone of the new isolate (termed LEC17) and parent CHO cells were as

lectin-resistant as the mutant (Table II), thereby identifying LEC17 as a mutant with a novel, dominant LecR phenotype (8,17). Because LEC17 was cross-resistant to ricin and abrin, two lectins that recognize and bind Gal/GalNAc-terminating carbohydrates (l), and because of the known properties of CHO glycosylation mutants (17), it seemed likely that LEC17 cells might express altered carbohydrates on cell-surface glycoconjugates. However, both LEC17 and parent cell lines bound essentially the same amount of ‘251-ricin at 4 “C over a toxin concentration range of 0.1-100 Kg/ml (data not shown). By contrast, LEClO, a ricin-resistant glycosylation mutant with

altered

cell-surface

carbohydrates TABLE

(10)

bound

signifi-

I

Lec?phenotypes of four new ricin-resistant CHO mutants The concentration (in rg/ml) of each lectin that reduced survival of parent CHO cells to 10% was determined by P-test (see “Experimental Procedures”) and is listed in parentheses under the appropriate lectin. Fold-differences in D,,, values for the four mutant cell lines compared with parent cells were averaged from ~3 experiments. R = resistant; S = sensitive; - = D,, equivalent to that of parent cells; ( ) = 30 < 50 pg/ml), E. cristugalli (Dlo -200 pg/ml), P. plumosa (D10 > 175 < 200 Mg/ml), and V. album (0~ -100 rg/ml) were also essentially indistinguishable (data not shown). Furthermore, neither the elution profiles nor the relative amounts of different glycopeptide species isolated from vesicular stomatitis virus cultured in mutant and parent cells in the presence of [3H]glucosamine differed significantly on lectin-affinity columns. Thus, WV glycopeptides from the two cell lines whether untreated or following neuraminidase digestion, yielded essentially identical elution profiles on ConA-Sepharose, R. communis agglutinin-agarose or ricinagarose columns (data not shown). Therefore, in contrast to previously described ricin-resistant glycosylation mutants (17), LECl7 cells did not appear to synthesize N-linked carbohydrates with altered branching patterns or terminal galactosylation. Resistance to ricin may also be acquired because of an inability to endocytose the toxin effectively (18-22). A characteristic of two ricin internalization mutants (20) is that resistance to ricin is not altered by cultivation in the presence of the ionophore nigericin (23). In contrast, both nigericin and monensin increase the sensitivity of parental CHO cells to ricin, presumably by facilitating the internalization of receptor-bound toxin (23). Another characteristic of ricin internalization mutants is cross-resistance to Pseudomonas uerugirwsa toxin (20), a protein with a different receptor specificity and intracellular site of action than ricin (24). As shown in Table III, LEC17 and parent CHO cells were equally sensitive to P. aeruginosu toxin (D10 > 20 < 25 rig/ml) and exhibited similar fold increases in sensitivity to ricin, abrin, and P. ueruginosa toxin (-lo-, -lo-, and -20-fold, respectively) when cultured in the presence of 50 nM monensin. A more direct assay of the internalization of bound lz51ricin by parent and LECl7 cells was performed by comparing the amount of cell-associated (bound and internalized) lZ51ricin at 37 “C with the amount of cell surface-bound ‘?-ricin. The latter could be selectively removed from the cell surface by a combined Gal-Pronase treatment as described under

of CHO Cells

585

“Experimental Procedures.” Both cell lines internalized the same percentage of cell-associated toxin (17.8 and 16.5% for parent and LEC17 cells, respectively) in this assay. The possibility that the LEC17 mutant might be altered in the intracellular target of ricin, the 60 S ribosomal subunit, was explored in a cell-free in vitro translation system. Postmitochondrial (SlO) supernatants were prepared from LEC17 and parent CHO cells and assayed for their ability to incorporate [3H]leucine into trichloroacetic acid-precipitable cpm at 30 “C in the presence or absence of various concentrations of ricin, ricin A chain, abrin, or modeccin. As shown in Fig. 1, protein synthesis proceeded essentially linearly as a function of time for up to 20 min after transfer of translation mixtures to 30 “C. However, protein synthesis in parent and LEC17 cell extracts was inhibited to different extents in the presence of ricin, ricin A chain, or abrin. As expected from their lack of cross-resistance to modeccin (Table I), LEC17 and parent extracts were equally sensitive to this toxin in vitro. Plots of log toxin concentrations versus % trichloroacetic acid-precipitable cpm obtained after 40 min of incubation at 30 “C were linear, and therefore were used to compare the concentration of toxin necessary to reduce incorporation of trichloroacetic acid-precipitable cpm by 50% (I& in LEC17 and parent cell extracts (Table IV). It is apparent that LEC17 extracts were significantly more resistant to the inhibitory action of ricin, abrin, and ricin A chain on protein synthesis compared with parental CHO SlO extracts (22-, 9-, and 56fold, respectively) yet were as sensitive as the latter to inhibition by modeccin. As ricin has been shown to inhibit the elongation step of

I

Ric

Ric A

Abr 1

TABLE III Effect

Mutants

of monensin

on the toxicity of ricin, and I’. aeruginosa toxin

abrin,

The D10 values of the different toxins for parent and LEC17 cells incubated with or without monensin (50 nM) were determined (see “Experimental Procedures”). The fold-enhancement of cytotoxicities of ricin, abrin, and P. aeruginosa toxin in monensin-treated versus untreated cells for each cell line is indicated in the -Monensin/ +Monensin column.

no Toxin

-Monensin

+Monensin

-Monensin/ +monensin

w/ml P. aeruginosa

toxin Parent LEC17

20 25

Ricin Parent LEC17

>5 < 10 200

Abrin Parent LEC17

>l c 2.5 10

z-1 c 2.5 1.0

>8 < 20 25

0.5 20

>lO < 20 10

>O.l < 0.25 1.0

-10 10

FIG. 1. Post-mitochondrial (SlO) supernatants from parent (0, 0) or LEC17 (0, H) cells were incubated at 30 “C in the absence (open symbols) or presence (closed symbols), respectively, of 3 pg/ml ricin (WC), 50 rig/ml ricin A chain (Ric A), 100 rig/ml abrin (Abr), or 300 rig/ml modeccin (Mod) as described under “Experimental Procedures.” The percent trichloroacetic acid-precipitable cpm in duplicate lo-p1 aliquots of reaction mixtures taken at the indicated incubation times are plotted relative to the cpm obtained after 40 min of incubation in mixtures lacking toxins. The trichloroacetic acid-precipitable cpm in lo-r1 aliquots of the latter samples ranged from 10,800 to 23,700. Each panel shows results obtained from one experiment.

586

Ricin-resistant

Ribosomal Mutants

TABLE IV Inhibition of cell-free in vitro protein synthesis by various toxins Post-mitochondrial SlO supernatants were assayed for their ability to carry out protein synthesis in the presence or absence of ricin, ricin A chain, abrin, or modeccin as described in the legend to Fig. 1. The toxin concentrations (Iso) that resulted in a 50% reduction in the incorporation of [3H]leucine into trichloroacetic acid-precipitable cpm after a 20 min incubation of translation mixtures at 30 “C is indicated for mutant and parent cells. ND = not determined. Is0 Ricin A

Ricin

Abrin

Modeccin

Parent LEC17

14.2 800

110 2450

150 1350

-300 -300

Parent Lec26 Lec27 Lec28

5.3 59 34.5 11.8

ND ND ND ND

76 340 121 310

ND ND ND ND

w/ml

1l)l-J

Parent

80 t

/

0

5

IO

A I 20

5

IO

20

FIG. 2. Effect of pactamycin on protein synthesis. SlO extracts prepared from parent or LEC17 CHO cells were incubated at 30 “C in 3 fiM pactamycin without ricin (0) or in the presence of 100 rig/ml (0) or 1,000 rig/ml (A) ricin. The percent relative incorporation was determined as described in the legend to Fig. 2, except that 100% relative incorporation is taken as the trichloroacetic acid-precipitable cpm in duplicate lo-p1 aliquots from assay tubes lacking toxin after 20 min of incubation at 30 “C, which ranged from 10,500 to 16,500. In the absence of pactamycin, 2- and 1.6-fold more trichloroacetic acid-precipitable cpm were obtained after 20 min of incubation with parent and LEC17 SlO extracts, respectively, indicating that -2 rounds of reinitiation had occurred. Each panel shows results obtained from one experiment. protein synthesis, in part by interfering with the EF-2-dependent hydrolysis of GTP (reviewed in Refs. 1 and 2), it was important to determine whether elongation of protein synthesis in LEC17 cells was differentially sensitive to ricin compared with parent CHO cells. Cell-free in uitro translation assays were therefore carried out in the presence of 3 ~.LM pactamycin, an inhibitor of initiation of protein synthesis (13). At a pactamycin concentration that completely inhibited translation of exogenous rabbit globin mRNA in micrococcal nuclease-treated parent SlO extracts, translation of endogenous mRNA in extracts from LEC17 cells was resistant to inhibition, and therefore, to arrest of elongation of protein synthesis by ricin (Fig. 2). Thus, in the presence of 100 ng/ ml ricin and 3 pM pactamycin, parent cell extracts incorporated nearly 40% less [3H]leucine into trichloroacetic acidprecipitable cpm than parent extracts incubated with pactamycin alone, whereas under the same set of conditions, LEC17 extracts were essentially unaffected in their protein synthetic ability. Similar results were obtained at a lo-fold higher ricin

of CHO Cells

concentration (data not shown). LEC17, therefore, appeared to be a likely candidate for a novel class of ricin-resistant mutant, a dominant phenotype altered in the intracellular site of ricin action. Three other ricin- and abrin-resistant mutants of CHO cells (Lec26, Lec27, and Lec28) were isolated from independent selections and resembled LEC17 in not being significantly cross-resistant to lectins such as L-PHA, WGA, or LCA (Table I). All three mutants were less ricin-resistant but at least as abrin-resistant as LEC17 cells. Cell-free SlO extracts from Lec26, Lec27, and Lec28 cells were also more resistant than parent extracts to inhibition of protein synthesis by ricin A chain or abrin (Table IV). The three mutant phenotypes were recessive in somatic cell hybrids with parent cells and Lec27 and Lec28 exhibited complementation in hybrids with Lec26 (Table II). Therefore, at least two new complementation groups were represented by Lec26 and Lec27 or Lec28. Because of their unique LecR phenotypes, Lec27 and Lec28 are also likely to be genetically distinct. However, additional markers are required to select for Lec27 X Lec28 hybrids in order to test directly for complementation. The elucidation of the molecular site of action of ricin A chain by Endo et al. (4, 5) permitted a more detailed investigation of the nature of the mutations in LEC17, Lec26, Lec27, and Lec28 cells. The A chains of ricin, abrin, and modeccin, as well as a number of ribosome-inactivating proteins including gelonin, saporin, and the family of pokeweed antiviral proteins (4, 25-30), are enzymes that specifically hydrolyze the N-glycosidic bond of adenine located in an evolutionarily conserved region 4324 nucleotides from the 5’ end of 28 S rRNA of rat liver and rabbit reticulocytes. Removal of the adenine from the sugar-phosphate backbone of the nucleic acid renders the phosphodiester bonds located on either side of the adenine-associated ribose susceptible to cleavage (via a @-elimination reaction) following treatment with aniline at acidic pH. Aniline treatment of depurinated 28 S rRNA results in the release of a fragment consisting of the 3’ most 470 nucleotides of this RNA (5). To determine whether CHO ribosomes were similarly acted on by ricin, total cellular RNA was isolated from exponentially growing parent and LEC17 cells that had been cultured for 3 h in the absence or presence of 100 rig/ml or 300 rig/ml ricin or 25 rig/ml abrin. The generation of an -470 nucleotide fragment was observed for each cell population (Fig. 3). A significantly greater amount of fragment was released from parent RNAs compared with RNAs from LEC17 cells under all conditions examined. Densitometer scans of the negative of the photograph in Fig. 3 revealed a 4.1-7.2-fold difference in amount of fragment released by ricin and a 1.4-2.7-fold difference in the amount of fragment released by abrin for parent compared with LEC17 RNAs, consistent with the different sensitivity of LEC17 cells to the two toxins (Table I). To determine whether each of the mutants was resistant to the action of ricin and/or abrin on 28 S rRNA in vitro, polysomes were isolated from mutant and parent CHO cells and incubated in the absence or presence of different concentrations of ricin or abrin at 37 “C for 15 min. Ribosomal RNAs extracted from the treated polysomes were incubated with aniline and electrophoresed on acrylamide/urea gels (Fig. 4). At a given ricin or abrin concentration, the amount of fragment obtained in each case was greater for rRNAs derived from parent compared with mutant polysomes. The differences in fragment levels were considerably smaller than expected from lectin-resistance patterns (Table II) and Is0 values (Table IV). This may be due to non-linear relationships in either ethidium bromide staining or negative exposures or to

Ricin-resistant

Ribosomal Mutants

Parent ,I LECl7

Ric Abr I oln gem O~mNO-mN

FIG. 3. Cleavage of 28 growing parent and LEC17 the absence or presence of rig/ml abrin (Ah). Aliquots from these cells were treated phoresed as described under points to the -470.nucleotide aniline treatment of rRNAs

1

Ric Abr oln

B C D E-F

I Parent

'A

S rRNA in intact cells. Exponentially cells were incubated for 3 h at 37 “C in 100 rig/ml or 300 rig/ml ricin (Ric) or 25 (6.7 Kg) of total cellular RNAs prepared with aniline at acidic pH and electro“Experimental Procedures.” The arrow fragment that was released following derived from toxin-treated cells.

G H I

J'

Let 27

587

B C D ENF

LECI7 G H I

J'

,oO,

,ent

id. S'A

of CHO Cells

II S'ABCDEFGHIJ

Parent

Lec26

Let 28

B~'FGHIJ'

prepared from parent and mutant cells were inucbated at 37 “C for 15 min in the absence and presence of ricin or abrin. RNA was extracted, treated with aniline, and electrophoresed on acrylamidel urea gels. In panels n-c, polysomes were incubated without (A and F) or with 40 rig/ml (R and Cl, 160 rig/ml (C and H), 300 rig/ml (D and I), or 1000 rig/ml (E and J) ricin. In panel d, polysomes were incubated without (A and F) or with 0.1 @g/ml (B and G), 0.2 pg/ml (C and HI, 1.0 pg/ml (D and I), or 2.0 rg/ml (E and J) abrin. Lanes labeled S in each panel correspond to the parent RNA sample from Fig. 3 (300 rig/ml Ric), included as a control for the efficacy of the aniline treatment. RNA molecular weight markers of 0.3-0.6 kb are shown in panel o (Std ).

FIG. 5. Cleavage of 28 S rRNA in ethanol-extracted polysomes. Polysomes prepared from parent and LEC17 CHO cells were extracted with ethanol as described under “Experimental Procedures” and incubated with ricin for 15 min at 37 “C. Extracted RNA was treated with aniline and electrophoresed on acrylamide/urea gels. Lanes A and F, RNAs extracted from intact parent and LECl7 polysomes, respectively, incubated with 10 &ml ricin. Lanes B-E and G-J; RNAs extracted from ethanol-treated parent and LEClT polysomes, incubated wit.hout ricin (B and G) or in the presence of 0.2 pg/ml (C and H), 2.0 pg/ml (D and I), or 10 pg/ml (E and J) ricin. The arrow marks the 470 nucleotide fragment. the fact activity theless,

that the lowest concentrations or parental polysomes were

of lectin with not determined.

maximal Never-

the data in Fig. 4 show that all four mutants possess polysomes that are resistant to the action of abrin or ricin compared with parental polysomes. Thus, all four mutants appear to have acquired resistance to abrin and ricin by virtue of an alteration in a structural component of the ribosome. A preliminary attempt to localize the altered structural component in LECI7 to a specific subset of ribosomal proteins was made by comparing the ricin sensitivity of ethanolextracted polysomes from parent and LEC17 cells. The subset of ribosomal proteins that have acidic pIs are believed to be present in >l molar equivalent/ribosome (31) as would be required if an altered form is to have a dominant effect. They are also involved in ricin-sensitive steps of protein synthesis such as EF-l-dependent binding of aminoacyl-tRNA (32) and EF-a-dependent GTPase activity (33). In addition, most of the acidic ribosomal proteins appear to be associated with the 60 S subunit and may be situated in close proximity to the ricin-binding site (32). Therefore, the possibility that one of these proteins might be altered in LEC17 was explored. As demonstrated by McConnell and Kaplan (16), treatment with 50% ethanol results in the selective removal of the acidic proteins from the large subunit of the 80 S ribosome of rat liver. When ethanol-treated polysomes from parent and LEC17 cells were incubated with ricin as in Fig. 4, a reduced amount of fragment was obtained from parent ribosomes and extensive degradation of 18 S rRNA was observed (Fig. 5). In addition, the amount of -470-nucleotide fragment obtained from ethanol-treated LEC17 polysomes suggests that extraction of acidic ribosomal proteins does not abrogate the LECl7 phenotype. It is apparent from Fig. 5 that ethanol-treated polysomes are significantly more resistant to the action of ricin than intact polysomes (compare lanes E and J with lanes A and F, respectively). This result concurs with previous reports on the resistance of naked rRNAs to ricin A (5, 15) and further supports the idea that ribosomal integrity plays a crucial role in facilitating toxin activity. DISCUSSION

Four new mutants of CHO cells that are differentially resistant to ricin, abrin, and modeccin have been isolated and partially characterized. Each of the mutants is resistant to

588

Ricin-resistant

Ribosomal Mutants

the inhibition of cell-free protein synthesis by ricin A chain and abrin and each possesses polysomes resistant to the Nglycosidase activity of ricin or abrin in uitro. Thus, all four cell lines, LEC17, Lec26, Lec27, and Lec28, belong to a class of RicR mutants that exhibit alterations in the eukaryotic ribosome. Despite extensive searching for toxin-resistant ribosomal mutants in mammalian cells (l), only one has been previously reported. Ono et al. (34) obtained RicR CHO cells that possessed 60 S ribosomal subunits more resistant than parental CHO ribosomes to inhibition of polyphenylalanine synthesis by ricin. With the isolation of the mutants described in this paper, it would seem possible to isolate a series of CHO ribosomal mutants. Although the frequency of the dominant mutant, LEC17, was low in CHO populations (