Bioconjugate Chem. 1990, 1, 331-336
331
Immunotoxin Construction with a Ribosome-Inactivating Protein from Barley Ray F. Ebert’ and Lucinda A. Sprynt Akzo Pharma/Organon Teknika Corporation, Biotechnology Research Institute, 1330-A Piccard Drive, Rockville, Maryland 20850-4373. Received June 22, 1990
The aim of this study was to determine the suitability of a ribosome-inactivating protein (RIP) from barley endosperm for use as an immunotoxin. This barley RIP is identical with the 30-kDa protein first reported by Coleman and Roberts [(1982) Biochim. Biophys. Acta 696, 2391 and sequenced by Asano and co-workers [(1986) Carlsberg Res. Commun. 51, 751. Use of the terms barley toxin I, 11, and I11 is proposed to describe the three isoforms resolved by cation-exchange chromatography. An improved procedure for isolating the protein involving the steps of aqueous extraction, ammonium sulfate precipitation, and cation-exchange HPLC is described. Barley toxin I1 retained activity after exposure to ca. 40% acetonitrile and 0.1 % trifluoroacetic acid or lyophilization. In a comparative study using the rabbit reticulocyte lysate assay, the protein was about 68% and 30% as potent as gelonin and ricin A-chain (RTA), respectively. Introduction of SH groups with 2-iminothiolaneresulted in a substantial loss of activity as the number of thiol groups approached four. Therefore, it was necessary to limit thiolation to an average of one to two SH groups per toxin molecule. Anti-transferrin receptor-based immunotoxins constructed with RTA, gelonin, and barley toxin I1 exhibited comparable cytotoxicity against a human colon tumor cell line. We conclude that the availability of raw material, ease of purification, and stability of barley toxin I1 to lyophilization and denaturing conditions render it a suitable protein for the construction of immunotoxins.
Type 1ribosome-inactivating proteins (RIP1) are a class of plant-derived polypeptides, possibly involved in disease resistance, that lack cell-binding (B-chain) domains and function by inactivating eukaryotic ribosomes ( I ) . A translation inhibitor with these characterisitics was first isolated from pearled (dehusked) barley (hordeum uulgare) by Coleman and Roberts in 1982 (2). Like other RIP’S from the family Gramineae, the barley protein synthesis inhibitor was found to be a basic (PI>lo), 30kDa protein that was relatively nontoxic to intact cells. A subsequent report (3) disclosed that fungal ribosomes are approximately 10 times more sensitive than eukaryotic ribosomes to the barley RIP. Independently, Asano and co-workers isolated and characterized three isoforms of a 30-kDa protein synthesis inhibitor from barley seeds ( 4 , 5 ) . They also determined the sequence for the most prevalent of these isoforms, “barley protein synthesis inhibitor 11” (6). A direct comparison of this protein with RTA disclosed a 21% homology overall, with an extraordinary conservation of residues in the putative activesite region (7). Comparative evaluation of RTA and this “barley toxin” demonstrated that both proteins inactivate ribosomes via the same mechanism: enzymatic hydrolysis of the N-glycosidicbond at A43Z4 of 28s rRNA (8).
* To whom requests for reprints should be addressed.
Portions of this work were submitted t o The American University in partial fulfillment of the requirements for the Master of Science degree. 1 Abbreviations: DTNB, 5,5’-dithiobis( 2-nitrobenzoic acid); IDw, concentration at which protein synthesis is inhibited by 50%; PBS, Dulbecco’s phosphate-bufferedsaline (pH 7.3): NaZHP04 (10 mM), KHzPOl(l.8mM), KCl(3.4 mM), and NaCl(l71 mM); RIP, ribosome-inactivatingprotein; RTA, ricin A-chain; SDSPAGE, sodium dodecyl sulfatepolyacrylamide gel electrophoresis; SPDP, N-succinimidyl 3-(Ppyridyldithio)propionate. t
1043-1802/90/2901-0331$02.50/0
The available evidence indicates that the protein synthesis inhibitor/30-kDa antifungal protein described by Roberts and co-workers (2, 3), the protein synthesis inhibitor sequenced by Asano and colleagues ( 4 4 3 , and the toxin evaluated by Endo et al. (8)represent the same enzyme. In this report we shall refer to this protein as “barley toxin” and, where appropriate, use the descriptors I, 11, and I11 to denote the three isoenzyme peaks eluting from cation-exchange columns. The aim of the present study was to determine the suitability of barley toxin I1 for use as an immunotoxin. We describe an improved, three-step purification procedure for isolating the protein and an optimized method for conjugating it to an anti-transferrin receptor monoclonal antibody. A comparison between this immunotoxin and similarly constructed conjugates with RTA and gelonin disclosed comparable in vitro cytotoxicity against a colon tumor cell line. EXPERIMENTAL PROCEDURES
Materials. RTA was from Vector Laboratories, Burlingame, CA. Gelonin was kindly provided by Dr. Walter Blattler, ImmunoGen, Inc., Cambridge, MA. Ammonium sulfate (enzyme grade) was from Bethesda Research Laboratories, Gaithersburg, MD. SPDP and 2-iminothiolane were from Pierce Chemical Co., Rockford, IL. DTNB (Ellman’s reagent) was from Aldrich Chemical Co., Milwaukee, WI. Medium, pearled barley was a product of The Quaker Oats Co., Chicago, IL, and was purchased at a local grocery store. Tritiated leucine was purchased from Amersham Corp., Arlington Heights, IL. SDSPAGE molecular weight standards were purchased from Bio-Rad, Inc., Rockville Centre, NY. Isolation of Barley Toxin. Barley RIPSwere isolated essentially as described by Roberts and Selitrennikoff (3), 0 1990 American Chemical Society
332 Bioconjugate Chem., Vol. 1, No. 5, 1990
except that the conventional carboxymethyl-Sepharose cation-exchangestep was replaced with a high-performance cation-exchange procedure. This modification yielded homogeneous barley toxin and eliminated the need for sizeexclusion chromatography. To summarize, 3-5 kg of pearled barley was ground into a coarse dry powder in a blender and stirred for 1-2 h at 4 "C in 7.5 L of a 10 mM sodium phosphate buffer (pH 7.4) containing 130 mM NaC1. All remaining procedures were performed at 4 "C, and all centrifugation steps were at lO00Og for 10 min. The extraction mixture was filtered through cheesecloth and clarified by centrifugation. The supernatant was brought to 55% (w/v) saturation by slow addition of solid (NH4)2SO4. After stirring for 2 h, the precipitate was removed by centrifugation and the supernatant was filtered through Whatman # 1paper. The filtrate was brought to 80% (w/ v) saturation with solid (NH&S04, stirred overnight, and centrifuged. This 55-80 % (NH4)2S04precipitate, which contains barley toxins I, 11, and 111, was resuspended in 200-400 mL of 25 mM sodium phosphate (pH 7.0) containing 50 mM NaCl (eluent A) and dialyzed against 4 L of eluent A with four changes at 7-14-h intervals. The dialysate was clarified by centrifugation and stored at -70 "C. Fractions containing approximately0.8-1.0 g of protein were thawed and applied to a 0.94 X 20 cm cationexchange HPLC column (PolyCAT-A, PolyLC Inc., Columbia, MD) that had been equilibrated with eluent A. After sufficient washing with eluent A to remove nonadsorbed protein, the column was developed with a gradient from 50 to 300 mM NaCl (in 25 mM sodium phosphate, pH 7.0) over a period of 60 min at 2 mL/min. Between runs, the column was regenerated with 2-4 volumes of 0.5 M NaCl in 25 mM sodium phosphate, pH 7.0. Approximately four runs were needed to process the 5 5 4 0 % (NHJ2S04 precipitate resulting from an extract of 3-5 kg of pearled barley. Anti-Human Transferrin Receptor Monoclonal Antibody. An hybridoma secreting monoclonal antibody 5E9Cll (anti-human transferrin receptor; ref 9) was obtained from the American Type Culture Collection (ATCC HB-21), Rockville, MD. The antibody was generated in pristane-primed BALB/c mice and purified from ascites fluid by protein A affinity chromatography or by ion-exchange HPLC on a Bakerbond ABx column (0.775 X 10 cm; Baker Chemical Co., Phillipsburg, NJ) according to the manufacturer's instructions. The antibody preparation used to construct immunotoxins was judged to be >95% pure on the basis of SDS-PAGE. Thiolation of Barley Toxin and Gelonin. Traut's reagent (2-iminothiolane) was used to introduce sulfhydryl groups into barley toxin and gelonin, both of which lack cysteine. The toxins were dialyzed against a 100 mM potassium phosphate buffer (pH 8.0) containing 1 mM EDTA and adjusted to a final concentration of 2-4 mg/ mL. After prewarming to 30 OC, a sufficient amount of 0.4 M 2-iminothiolane was added to yield a 50-fold molar excess over the toxin. After incubation for 10-30 min at 30 OC, the solutions were dialyzed against a 10 mM potassium phosphate buffer (pH 7.0) containing 1 mM EDTA (three changes, 8-12 h each) to remove low molecular weight components. This procedure resulted in the addition of one or two sulfhydryl groups per toxin molecule, as estimated with the Ellman method (IO). Higher substitution ratios were obtained by incubating the reaction mixture as above for 30-60 min. Activation and Conjugation of the Anti-Transferrin Receptor Monoclonal Antibody. Antibody HB21 ( M R180 000) was dialyzed into a buffer containing 50
Ebert and Spryn
mM potassium phosphate (pH 7.5) and 200 mM NaCl (coupling buffer) and incubated for 30 min a t room temperature with a 3-fold molar excess of SPDP. Excess reactants and non-protein products were removed by dialysis at 4 OC against coupling buffer. Under these conditions, approximately two pyridinedithiol groups were incorporated per monoclonal antibody. The activated monoclonal antibody was incubated overnight at room temperature with thiolated toxin at a concentration 5 times higher than that of the antibody-linked 2-pyridinedithiol. The immunoconjugatewas separated from unreacted toxin by gel-permeation HPLC using a Zorbax GF-250 column (2.12 X 25 cm, E. I. du Pont de Nemours & Co., Wilmington, DE) equilibrated with PBS. Reticulocyte Lysate Assay. Inhibition of protein synthesis in a cell-free system was used to measure toxin activity. Reagents provided in a translation kit (Promega Biotec, Madison, WI) were employed with a modified assay procedure (Note: conjugates were not reduced prior to assay, nor were reducing agents present in the reaction mixture). A solution of kit reagents consisting of one part 1mM amino acid mixture minus leucine, one part 0.5 pg/ pL Brome mosaic virus mRNA, one part glass-distilled water, and five parts [3H]leucine (5 pCi/mL) was made. Each assay tube contained 8 pL of this solution, 35 pL of lysate, and 10 pL of sample. All dilutions were made with a 40 mM sodium phosphate buffer (pH 7.4) containing 150 mM NaCl (sample buffer), and molar concentrations of the immunotoxins were based on an assumed molecular weight of 240 000 (antibody plus two toxin molecules). The solutions were mixed and centrifuged in a microfuge apparatus for 5 s to ensure that all liquid was at the bottom of the tubes. After 1 h at 37 "C, 50 pL of assay mixture was transferred to a microtiter plate (Immulon 11, Dynatech Laboratories, Chantilly, VA) that had been blocked with 1%bovine serum albumin in sample buffer. The translated protein was transferred to glass-fiber filters by means of a cell harvester (PhD, Cambridge Technology, Inc., Cambridge, MA). The filters were then dried and radioactivity was determined in a liquid-scintillation counter. The percent inhibition of protein synthesis was calculated relative to a control to which no toxin had been added. One unit of toxin activity is defined as the sample dilution that inhibits protein synthesis by 50 76. Cytotoxicity Assay. A human colon adenocarcinoma cell line, HT-29, was obtained from the American Type Culture Collection (Rockville, MD) and used to evaluate the cytotoxicity of immunoconjugates. Microtiter plates (96 well, Immulon 11, Dynatech Laboratories, Chantilly, VA) were inoculated (day 1)with 200 pL/well of cells (2.5 X 105/mL) suspended in Eagle's MEM (GIBCO,Grand Island, NY) plus 10% fetal calf serum plus 0.1% gentamicin. All incubations were in a constant humidity chamber set at 37 OC and 94% 0 2 / 6 % COz. On day 2, the medium was replaced (100 pL/well), and test samples (50 p L in sterile PBS) were added. All samples were assayed in quadruplicate. On day 4 the cells were washed with PBS and 100 pL of Eagle's MEM leucinefree medium was added to each well. After a 1-3-h incubation at 37 "C, 50 pL of [3H]leucine (0.04 pCi/pL) in PBS was added. After 3-4 h at 37 "C, the medium was aspirated from the wells; cells were released by addition of 0.1 % trypsin (100 pL/well), harvested on glass-fiber filters as above, and evaluated for radioactivity by liquid scintillation counting. The molar concentrations of the immunotoxins were based on an assumed molecular weight of 240 000 (antibody plus two toxin molecules). Reversed-Phase (RP) HPLC of Barley Toxin. RP-
Bioconjugate Chem., Vol. 1, No. 5, 1990 333
Immunotoxin Construction with Barley Toxin II
Table I. Purification of Barley Toxin mg of total units sp act., step protein (1OG)o units/pg 10100 140 14 aqueous extract 3570 318 89 5540% (NH&S04 ppt 174 107 618 cation-exchange HPLC ~
~
~
5% yield
~~
100 227 76
n
H
0 One unit of toxin activity is defined as the sample dilution that inhibits protein synthesis by 50% in a reticulocyte lysate assay system. Total units are obtained by multiplying this dilution factor by the quotient of the total sample volume (in mL) divided by 0.01 mL (the volume of sample in the reticulocyte lysate assay).
HPLC of purified barley toxin was performed with a Vydac TP-C4 (butylsilane) HPLC column (1.0 X 25 cm, The Separations Group, Hesperia, CA). Eluents A and B were water and acetonitrile, respectively, each containing 0.1 % trifluoroacetic acid. The column was equilibrated with eluent A and the protein was eluted by a gradient (1.5% / min) from 30% to 50% B. Protein was detected by absorbance a t 280 nm. Protein-containing peaks were collected manually, evaporated to dryness in a SpeedVac concentrator (Savant Instrument Co., Farmingdale, NY), and redissolved in a 40 mM potassium phosphate buffer (pH 7.4) containing 150 mM NaC1. Electrophoresis. SDS-PAGE was carried out in the buffer system of Laemmli (1I) using a 5 % stacking gel and a gradient of 5-15 % polyacrylamide (0.8% bisacrylamide) in the separating gel. The gels were stained with Coomassie Brilliant Blue R-250. Other Asdays. All protein concentrations were determined with the method of Lowry et al. (12) using bovine serum albumin as the protein standard. RESULTS
Isolation and Characterization of Barley Toxin. The three-step purification procedure comprised aqueous extraction of triturated barley endosperm, differential ammonium sulfate precipitation, and cation-exchange HPLC. The results of a representative lot are summarized in Table I: an aqueous extract of 3.17 kg of pearled barley yielded 10.1 g of crude protein, from which 174 mg of barley toxin was isolated. The recovery of toxin activity in the 55-80 % ammonium sulfate fraction consistently exceeded 100%. The possibility that this was due to removal of an inhibitor was not explored further. The introduction of a highly selective cation-exchange HPLC step (Figure 1) obviated the need to remove contaminantsby gel filtration, as has been required following conventional cation-exchangechromatography ( 3 , 4 , 13). The major peak of RNA N-glycosidase activity (peak 11) corresponded to a major A280 peak a t about 43 min (Figure 1). Ribosome-inactivating activity also was detected in protein eluting at approximately 25-30 min and at approximately 55 min. These were presumed to represent barley toxins I and 111, respectively (4). The isolated protein migrated as a single 30-kDa protein band on SDS-PAGE (Figure 2), from which purity was estimated to be >99 % . However, when evaluated by RPHPLC, barley toxin I1 eluted as two protein peaks (Figure 3). Each protein peak was collected, lyophilized, redissolved in PBS, and compared in the reticulocyte lysate assay with cation-exchange-purified material that had not been lyophilized. The results (Figure 4) disclosed virtually no difference between the two RP-HPLC peaks, nor was there any substantial loss of activity upon exposure to the reversed-phase solvents (0.1% TFA and acetonitrile).Thus, barley toxin I1 appears to be stable to acid, organic solvent, and lyophilization.
0
0
1
0
2
0
3
0
4
0
5
0
6
0
TIME (MIN)
Figure 1. Cation-exchange HPLC of a 55430% (NH&SOr precipitate prepared from an aqueous extract of pearled barley. See Experimental Procedures for chromatographic conditions. The NaCl gradient is indicated by the dashed line. The bars denote regions in which toxin activity was detected in a reticulocyte lysate (cell-free) protein synthesis assay. 92,000 -
-
67,000 45,000
-
21,000 -
30,000
1
2
3
4
Figure 2. Comparative SDS-PAGE of RIP’S. A 5-15% polyacrylamidegradient gel under nonreducing conditions was used lane 1,ricin A-chain; lane 2, gelonin; lane 3, barley toxin 11; lane 4, Pseudomonas exotoxin.
For the preparation described in Table I, barley toxin I1 comprised about 1.3% of the protein in the aqueous extract. In general, the toxin represented 1-2% of the protein in these extracts. On average (N = 3), 55 mg of toxin were obtained per kg of pearled barley. Specific activities of the isolated protein ranged from 400 to 1600 units/pg (mean = 980 units/pg), and the concentration at which cell-free protein synthesis was inhibited by 50% ranged from 0.52 to 2.1 nM (mean = 1.1nM; N = 4). In agreement with the observations of Asano and coworkers (4), carbohydrate was not detected on the barley toxin. Also in agreement with a previous report (6),the N-terminus of the toxin was blocked. Sequencing of CNBrderived peptides (not shown) confirmed the identity of barley toxin with the protein synthesis inhibitor described by Asano et al. (4-6). In a comparative study using the rabbit reticulocyte lysate assay, barley toxin (ID50 = 0.47 nM) was found to be about 75% and 30% as potent as gelonin (ID50 = 0.32 nM) and RTA (ID50 = 0.14 nM), respectively (Figure 5). Effect of Thiolation on Barley Toxin I1 Activation. Prior to conjugation with SPDP-activated antibody, sulfhydryl groups were introduced via 2-iminothiolane. We found that as the average number of thiol groups approached four, RIP activity was severely inhibited (Figure
Bioconjugate Chem., Vol. 1, No. 5, 1990
334
Ebert and Spryn 100
0.6
t
m
E
1 5 0
z
ae
0.4 W
25
Y 2a
5: 2
0
0.2
I
-10
-a
-9
-7
LOG CONCENTRATION (mol I L)
~
0
20
10
~
_
_
Figure 6. Effect of thiolation on barley toxin I1 activity in the reticulocyte lysate assay: control (unmodified) barley toxin I1 (e), thiolated barley toxin I1 a t 1.4 (m) and 4.4 (A)SH groups per molecule. See ExperimentalProcedures for additional details.
30
TIME (mln)
Figure 3. Reversed-phase (butylsilane) HPLC. Isolated barley toxin I1 (200pg) was applied to a butylsilane column and eluted with a gradient of acetonitrile (3040%;1.5%/min) as described in Experimental Procedures. Peaks I and I1 were collected manually, lyophilized,and evaluated for activity as described in Results.
z
loo
0
-
92,000 -
I
1
200,000
116,000
75-
k
m
z
'
5025-
45.000
-
01 -11
-10
-9
-8
-7
-6
LOG CONCENTRATION(mol /
Figure 4. Activity of barley toxin I1 preparations in the reticulocyte lysate assay: barley toxin I1 purified via cationexchange HPLC ( 0 ) ;isolated barley toxin I1 after reversedphase HPLC (see Figure 3),peak I @), peak I1 (A). 100
g
75
E 8 5 0
E 25 0 ~~__________
-12
-11
-10
-9
21,000 -
30,000
L)
-8
LOG CONCENTRATION(mol I L)
Figure 5. Comparison of ribosome-inactivating activity in the reticulocyte lysate assay of barley toxin I1 (O), gelonin (A),and ricin A-chain (m). Details of the assay are described in Experimental Procedures. ID50 values were 0.47,0.32,and 0.14 nM, respectively, for barley toxin 11, gelonin, and RTA.
6). Accordingly, it was necessary to limit thiolation to an average of one to two SH groups per toxin molecule. Evaluation of Immunotoxins. SDS-PAGE was used to determine the approximate number of toxin molecules linked to the anti-transferrin receptor monoclonal antibody. In this system the unconjugated antibody migrated at M R 180 000 (Figure 7, lanes 3 and 6). The results for barley toxin I1 and gelonin conjugates (Figure 7, lanes 2 and 5) disclosed a minor protein band with the same electrophoretic mobility as unconjugated antibody, major bands at M R210 000,240 000, and 270 000, and higher molecular weight multimers. From this we estimate that these im-
1
2
3
4
5
6
Figure 7. SDS-PAGE of immunotoxins constructedwith an antitransferrin receptor (aTFR) monoclonal antibody and gelonin or barley toxin 11. All samples were run under nonreducing conditions: lane 1, barley toxin 11; lane 2, barley toxin IIaTFR monoclonal antibody conjugate; lanes 3 and 6, unconjugated (control) aTFR monoclonal antibody; lane 4,gelonin;lane 5,gelonin-aTFR monoclonal antibody conjugate.
munotoxin preparations consisted of less than 10% unconjugated antibody, with a major fraction comprised of one to three toxin molecules per antibody. Similar results were obtained for the RTA conjugate (not shown). Anti-transferrin receptor-based immunotoxins constructed with RTA, gelonin, and barley toxin were evaluated for ribosome-inactivating activity in the rabbit reticulocyte lysate assay. In this cell-free system the barley toxin conjugate exhibited comparable activity to the gelonin conjugate (ID50values of 4.5 and 3.3 nM, respectively); however, it was only about one-tenth as active as the RTA conjugate (ID50 = 0.32 nM). When tested in vitro against a colon tumor cell line (Figure 8),all three immunotoxins exhibited comparable, cytotoxicity, with the RTA conjugate remaining the most potent, followed by the gelonin and barley toxin I1 conjugates. From these data ID50 values of 4.2,7.6, and 10 nM, respectively, were calculated for the RTA, gelonin, and barley toxin I1 containing immunotoxins. DISCUSSION
Our interest in evaluating barley toxin for suitability as an immunotoxin stems from several considerations: (i) it
Immunotoxin Construction with Barley Toxin II
-10
-9
-8
BiOCOtl}Upt8
-7
LOG CONCENTRATION (mol / L)
Figure 8. In vitro cytotoxicity of conjugates constructed with the anti-transferrin receptor monoclonal antibody and barley toxin I1 ( O ) , gelonin (A),and ricin A-chain (m). Details of the assay are described in Experimental Procedures. IDm values of 10,7.6, and 4.2 nM were calculated for barley toxin 11, gelonin, and ricin A-chain, respectively.
is an abundant, easily-isolated,and well-characterized RIP; (ii) since it is a type 1 RIP, purified barley toxin is free of B-chain contamination and therefore will be relatively nontoxic to intact cells; (iii) barley toxin is not glycosylated ( 4 ) ; therefore, the plasma half-life of immunotoxins constructed therewith will not be reduced via the asialoglycoprotein receptor pathway; and (iv) since it is now recognized that patients undergoing RTA-immunotoxin therapy may develop (in addition to a human antimouse antibody response) an anti-RTA immune response (14-16), immunoconjugates with barley toxin I1 may represent an alternative approach for a follow-up course of treatment. An improved purification procedure for isolating barley toxin I1 yielded an electrophoretically homogeneous preparation that nevertheless was resolvable into two isoforms by reversed-phase HPLC. The possibility that this heterogeneity represents a failure to resolve barley toxin I1 from barley toxins I and I11 has not been ruled out, but appears to be unlikely because activity corresponding to the latter two proteins was detected in two other regions of the cation-exchange elution profile (Figure 1). Our average recovery of 55 mg of barley toxin I1 per kg of seeds is substantially different from the estimate by Coleman and Roberts (2)of 1365 mg/kg but is similar to the yields of barley toxin I1 reported by Asano et al. (37.5 mg/kg; 4 ) , and Barbieri et al. (95 mg/kg; 13). The observation that our recoveries of toxin activity in the 558070 ammonium sulfate precipitate exceeded 100% raises the intriguing possibility that a specific toxin inhibitor is present in the barley endosperm. Barley toxin I1 is a remarkably stable protein, unaffected by exposure to 0.1 % TFA, 40% acetonitrile, or lyophilization (Figure 4). In our hands, RTA did not tolerate these conditions. However, we observed a significant loss of barley toxin I1 activity in the thiolation step as the number of alkylated amino groups approached four (Figure 6). This characteristic is in apparent contrast to the effect of thiolation on the activity of gelonin: Lambert et al. (17) observed only a slight loss of activity in 2-iminothiolanetreated gelonin preparations containing 2.93 sulfhydryl groups per molecule. Evaluation of RTA, gelonin, and barley toxin I1 in a cell free (reticulocyte) protein synthesizing system disclosed that the native type I RIP’s exhibited about one-third the potency of RTA (Figure 5). However, when conjugated to a monoclonal antibody, the type I RIP’s were only about one-tenth as potent as the corresponding RTA conjugate. This indicates that the active sites of gelonin and barley toxin I1 are less exposed than that of RTA after the
Chem., VOl. 1, No. 5, 1990 335
conjugation step. Other factors that may account for these discrepancies between the type I and type I1 toxins include (i) fundamental differences in substrate specificity and (ii) reduction in the activity of the type I RIP’s as a result of the covalent modification required to introduce SH groups. Interestingly, the major differences between immunotoxins containing type I vs type I1 RIPSwere considerably attenuated when they were assessed in vitro against a colon tumor cell line that expresses the transferrin receptor. In this case the immunotoxins constructed with barley toxin I1 and gelonin were about equivalent in potency and 4276% as effective as the RTA immunotoxin. We may speculate that efficient translocation of the conjugates via receptor-mediated endocytosis partially compensated for the differences in (cell-free) toxin activity of the immunotoxins. Notwithstanding this effect, the general trend in potency [RTA >> gelonin = barley toxin 111 persisted. In conclusion, among the type I and type I1 plant RIP’s evaluated in this report, RTA remains the toxin of choice, due to its potency in both cell-free and intact-cell inhibition assays. Gelonin and barley toxin I1 were found to be about equivalent in their activity, but the latter apparently is more sensitive to covalent modifications through amino groups. Nevertheless, we conclude that the availability of raw material, ease of purification, and stability of barley toxin I1to denaturing conditions render it a suitable protein for the construction of immunotoxins. LITERATURE CITED (1) Stirpe, F., and Barbieri, L. (1986) Ribosome-inactivating
proteins up to date. FEBS Lett. 195, 1-8. (2) Coleman, W. H., and Roberts, W. K. (1982) Inhibitors of animal cell-free protein synthesis from grains. Biochim. Biophys. Acta 696, 239-244. (3) Roberts, W. K., and Selitrennikoff, C. P. (1986) Isolation and partial characterizationof two antifungal proteins from barley. Biochim. Biophys. Acta 880, 161-170. (4) Asano, K., Svensson, B., and Poulsen, F. M. (1984) Isolation and characterization of inhibitors of animal cell-free protein synthesisfrom barley seeds. CarlsbergRes. Commun. 49,619626. (5) Asano, K., Svensson, B., Poulaen, F. M., Nygard, O., and Nilsson, L. (1986) Influence of a protein synthesis inhibitor from barley seeds upon different steps of animal cell-free protein synthesis. Carlsberg Res. Commun. 51, 75-81. (6) Asano, K., Svensson, B., Svendsen, I., and Poulsen, F. M. (1986) The complete primary structure of protein synthesis inhibitor I1 from barley seed. Carlsberg Res. Commun. 51,129141. (7) Ready, M. P., Katzin, B. J., and Robertus, J. D. (1988) Ribosome-inhibiting proteins, retroviral reverse transcriptase, and RNase H share common structural elements. Proteins 3,5359. (8) Endo, Y., Tsurugi, K., and Ebert, R. F. (1988) The mechanism of action of barley toxin: a type 1ribosome-inactivating protein with RNA N-glycosidase activity. Biochim. Biophys. Acta 954, 224-226. (9) Haynes, B. F., Hemler, M., Cotner, T., Mann, D. L., Eisenbarth, G. S., Strominger, J. L., and Fauci, A. S. (1981) Characterization of a monoclonal antibody (5E9) that defines a human cell surface antigen of cell activation. J . Immunol. 127,347-351. (10) Ellman, G. L. (1959) Tissue sulfhydryl groups. Arch. Biochem. Biophys. 82,70-77. (11) Laemmli, U. K. (1970)Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-685. (12) Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951) Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193, 265-275. (13) Barbieri, L., Stoppa, C., and Bolognesi, A. (1987) Large scale chromatographic purification of ribosome-inactivatingproteins. J. Chromatogr. 408,235-243.
336 Bioconjugete Chem., Vol. 1, No. 5, 1990 (14) Hertler, A. A., Schlossman, D. M., Borowitz, M. J., Poplack, D. G., and Frankel, A. E. (1988) An immunotoxin for the treatment of T-acute lymphoblastic leukemic meningitis: Studies in rhesus monkeys. Cancer Immunol. Immunother. 28,59-66. (15) Kernan, N . A., Byers, V., Scannon, P. J., Mischak, R. P., Brochstein, J., Flomenberg, N., Dupont, B., and O’Reilly, R. J. (1988) Treatment of steroid-resistant acute graft-vs-host disease by in vivo administration of an anti-T ricin A chain immunotoxin. J . Am. Med. Assoc. 259, 3154-3157.
Ebert and Spryn
(16) Mischak, R. P., Foxall, C., Rosendorf, L. L., Knebel, K., Scannon, P. J., and Spitler, L. E. (1990) Human antibody responses to components of the monoclonal antimelanoma antibody ricin A chain immunotoxin XomaZyme-MEL. Mol. Biother. 2, 104-109. (17) Lambert, J. L., Blatter, W. A,; mCIntyre, G. D., Goldmacher, V. S., and Scott, C. F. (1988) Immunotoxins containing single chain ribosome-inactivating proteins. In Immunotoxins (A. E. Frankel, Ed.) pp 175-209, Kluwer Academic Publishers, Boston.
