128
Bioconjugate Chem. 1994, 5, 126-132
Cysteine Analogs of Recombinant Barley Ribosome Inactivating Protein Form Antibody Conjugates with Enhanced Stability and Potency in Vitro Susan L. Bernhard,*J Marc Better,* Dianne M. Fishwild,t Julie A. Lane,t Ann E. Orme,+ Darryl A. Garrison,t Cynthia A. Birr,+ Shau-Ping Lei,$ and Stephen F. Carrollt XOMA Corporation, 2910 Seventh Street, Berkeley, California 94710, and 1545 17th Street, Santa Monica, California 90404. Received August 6, 1993@
Antibody immunoconjugates were made with native and recombinant forms of the type-I ribosome inactivating protein from barley (BRIP) and with three recombinant BRIP (rBRIP) analogs engineered to contain a unique cysteine residue near the C terminus (at amino acid 256,270, or 277). rBRIP and all three cysteine analogs (rBRIPC256,rBRIPC270, and rBRIPc277) were produced in E. coli, with yields of soluble protein as high as 1 g/L, and were as active as native BRIP in inhibiting protein synthesis in vitro. Interestingly, the position of the engineered cysteine influenced not only the efficiency of conjugation to antibody but also the efficacy and disulfide bond stability of the immunoconjugates. AntLCD6 antibody conjugates prepared with native and rBRIP were relatively inactive against antigenpositive target cells, while the conjugate made with rBRIPC277 was 5-fold more cytotoxic. AntLCD7 antibody conjugates made with rBRIPC277 or rBRIPc270 also exhibited improved potency and stability compared to the conjugate with native BRIP. These results indicate that engineering a cysteine residue into selected positions near the C-terminus of a type-I RIP such as BRIP can improve immunoconjugate yield, disulfide bond stability, and potency.
INTRODUCTION Plant leaves and seeds often contain a variety of proteins for defense against invasion by pathogens. One class of such proteins is the ribosome-inactivating proteins (RIP)' which enzymatically inhibit protein synthesis by hydrolyzing a single N-glycosidic bond in the 28 S rRNA of the eukaryotic ribosome (1,2). Type-I RIP contain a single catalytic subunit (A chain). In addition to an A chain, type-I1RIP contain a cell-binding lectin (B chain). There is considerable homology among RIP, especially for residues believed to be involved in catalysis (3). Best characterized of all RIP is the type-I1 protein ricin, and the X-ray crystal structure of ricin A chain (RTA) has recently been described (3). These naturally occurring cytotoxic agents can be chemically coupled to antibodies or other cell-targeting agents to generate immunoconjugates capable of selectively killing antigen- or ligand-positive cells. Many antibodyRIP immunoconjugates have been evaluated i n vitro and in vivo as potential therapeutic agents for cancer and autoimmune disease in man (4-7). Typically, such im-
* To whom correspondence and reprint requests should be addressed. + Berkeley, CA. Santa Monica, CA. Abstract published in Advance ACS Abstracts, January 15, 1994.
Abbreviations: RIP, ribosomeinactivating protein(@;BRIP,
the barley ribosome inactivating protein; nBRIP, native BRIP; rBRIP, recombinant BRIP; PCR, polymerase chain reaction;
PBMC, human peripheral blood mononuclear cell(s); mAb, monoclonal antibody; RTA, ricin toxin A chain; RTAao, M , 30 kDa glycoform of RTA;MZIT, 5-methyl-2-iminothiolane; DTNB, 5,5'-dithiobis(2-nitrobenzoicacid);TNB, 5-thio-2-nitrobenzoate anion; 2-ME, 2-mercaptoethanol;SPDP, 3-((N-succinimidyl-2pyridy1)dithio)propionate; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis; RLA, rabbit reticulocyte assay; GSH, glutathione; PHA, phytohemagglutinin; DTT, dithiothreitol. 1043-1802/94/2905-0126$04.50/0
munoconjugates have been constructed with the A chains of type-I1 RIP (such as RTA), which can be specifically targeted to cells only if the B chains are removed or inactivated. Type-I RIP from cereal grains such as barley (BRIP) are attractive alternatives to the A chains of type-I1 RIP for the construction of targeted cytotoxic molecules. For example, BRIP is particularly nontoxic to intact cells (8) and is not glycosylated (9). This latter property may lead to reduced uptake and clearance by mannose receptors in the liver, resulting in prolonged serum residence time in vivo and, as a consequence,increased time to reach cellular targets. Moreover, BRIP conjugates of an IgG recognizing the transferrin receptor (10) and an IgG recognizing a melanoma antigen (11) are cytotoxic to antigen-positive human cell lines. An important consideration for immunoconjugate assembly is the nature of the linkage between antibody and RIP. A disulfide linkage is usually thought to be essential for maximal cytotoxicity (12). RTA contains a single available cysteine,which can directly form a disulfide bond with an activated antibody thiol via a disulfide-exchange reaction. Type-I RIP, including BRIP, typically lack an unpaired cysteine, so disulfide-linked conjugates require modification of both antibody and RIP with heterobifunctional reagents. Unfortunately, the biologicalactivity of many RIP (10, 13, 14) is diminished or destroyed by modification with the most commonly employed reagents. To address this issue, we have cloned and expressed BRIP and engineered several analogs of rBRIP to contain unique free cysteine residues for direct conjugation to antibody. Three residues near the C-terminus of BRIP were selected as sites for amino acid substitution with cysteine. Comparison of the amino acid sequence of BRIP with the known primary and tertiary structure of RTA suggested that these sites should be on the surface of the molecule, and either a t or proximal to the position of cysteine 259 in RTA. The enzymatic activity of each BRIP analog 0 1994 American Chemical Society
Bioconjugate Chem., Vol. 5, No. 2, 1994
Cysteine in rBRIP Improves Conjugate Potency
with its uniquely placed cysteine was examined, as was the effect of the cysteine position on conjugation yield, in uitro disulfide bond stability, and immunoconjugate potency. EXPERIMENTAL PROCEDURES
Bacterial and Mammalian Cells. The E. coli host for production of BRIP is an Ara- derivative of W3110 (American Type Culture Collection no. 27325). HSB2 is a CD5+,CD7+ human T cell line (American Type Culture Collection CCL 120.1). PBMC were separated from whole blood of normal human donors by centrifugation through a discontinuous gradient (25). Antibodies H65 and 4A2. H65 is a murine IgGl/k mAb that recognizes the human CD5 antigen (15). 4A2 is a murine IgG2a/k mAb that recongizes the human CD7 antigen on T cells (16). Concentrations of antibodies were determined for 1 mg/mL solutions by absorbance at 280 nm, using extinction coefficients of 1.25 for 4A2 and 1.30 for H65. For calculations, an approximate molecular weight of 150 000 was used for each mAb. Purification of nBRIP from Barley Seeds. nBRIP (isoform 11) was purified from pearled barley flour by a modification of the protocol of Asano et al. (9). Four kg of barley flour was extracted with 16 L of 10 mM NaP04, 25 mM NaC1, pH 7.2 (extraction buffer), for 20 h at 4 OC. The sediment was removed by centrifugation, and 200 mL of packed S-Sepharose (Pharmacia, NJ) was added to adsorb the nBRIP. After 20 h at 4 “C with mixing, the resin was allowed to settle, rinsed several times with extraction buffer, and packed into a 2.6- X 40-cm column. The column was washed until the absorbance of the effluent approached zero. nBRIP was then eluted with a linear gradient of 0.025 to 0.3 M NaCl in extraction buffer. The nBRIP-containing fractions were pooled, concentrated, and chromatographed on a 2.6- X 100-cm column of Sephacryl S-2OOHR (Pharmacia, NJ) equilibrated in 10 mM NaP04,125 mM NaC1, pH 7.4 at 10 mL/h. The pure nBRIP was concentrated and stored at -70 OC. Cloning and Expression of rBRIP. A cDNA expression library prepared from germinated barley seeds in X ZAP11 was purchased from Stratagene, La Jolla, CA. Approximately 700 000 phage plaques were screened with rabbit antibody raised against nBRIP, and six immunoreactive plaques were identified. DNA from one clone was isolated and sequenced with Sequenase (USB, Cleveland, OH). The cDNA contained therein was excised and subcloned into pUC18 (Pharmacia, NJ). The pUC 18 clone containing the BRIP cDNA, pBS1, was used to construct a bacterial secretion vector. This process involved (1) introduction of a unique XhoI restriction site downstream of the BRIP termination codon, (2) addition of the pelB secretion signal to the 5’end of the gene (18) and (3) positioning the BRIP gene under the control of the inducible araB promoter (19). The resulting plasmid was named pING3322. Details of vector construction and a schematic view of the expression vector are shown in Figure 1. Construction of rBRIP Analogs with a Free Cys Residue. rBRIP analogswith introduced Cys codonswere assembled by PCR amplification of the BRIP gene with mutagenic oligonucleotides followed by gene reassembly. A plasmid capable of expressing a BRIP analog with a Cys in place of Leu at position 256 (rBRIPCZ56)was constructed using PCR to amplify the 3’-end of the gene while introducing an amino acid substitution. Plasmid pING3322 was amplified with 5’-TGTCTGTTCGTGGAGGTGCCG-3’ and 5’-CGTTAGCAATTTAACTGT-
127
GAT-3’. The PCR product was treated with T4 polymerase and cut with XhoI, and the 87-bp fragment was purified. Plasmid pING3322 was also cut with BamHI, treated with T4 polymerase, and cut with EcoRI, and the 891-bp fragment containing the BRIP gene was purified. These two gene fragments were then assembled into pING3322 to generate pING3801. A plasmid for expression of a BRIP analog with Cys substituted for Ala at position 270 (rBRIPc270) was also generated by using PCR to amplify a new 3’-end segment with an altered residue. pING3322 was the template for PCR amplification with 5’-CCAAGTGTCTGGAGCTGTTCCATGCGA-3’ and 5’-CGTTAGCAATTTAACTGTGAT-3’. The PCR product was treated with T4 polymerase and cut with XhoI, and the 51-bp fragment was purified. This segment was cloned into pING3322 with the internal BRIP restriction fragment from SstII to MscI (151 bp) and the vector piece cut with SstII and XhoI. The resulting plasmid was named pING3802. A plasmid capable of expressing a BRIP analog in which the Ser at position 277 was converted to a Cys residue (rBRIPc27,) was constructed by substituting the 3’-end of the BRIP gene in pING3322 with a DNA segment conferring this change. Oligonucleotides 5’-GCATTACATCCATGGCGGC-3’ and 5’-GATATCTCGAG?TAAC-
TATTTCCCACCACACGCATGGAACAGCTCCAGCGCCTTGGCCACCGTC-3’wereused to PCR amplify the BRIP gene. An 82-bp BamHI to XhoI fragment containing the 3’-end of BRIP with the altered amino acid was purified on a 5 % polyacrylamidegel, and this fragment was substituted into pING3322 in a three-piece ligation with the 891-bp EcoRI to BamHI fragment containing the BRIP gene and the EcoRI and XhoI cut vector. This generated pING3803. Purification of rBRIP and rBRIP Cysteine Analogs. rBRIP and rBRIP analogs were purified from concentrated bacterial fermentation broths. For rBRIP, concentrated broth from a 10-L fermentation batch was exchanged into 10 mM Tris-HCl,20 mM NaC1, pH 7.5, applied to an S-Sepharose column, and eluted with a 20500 mM NaCl linear gradient. Pooled rBRIP was further purified by affinity chromatography on Blue Toyopearl (Supelco, PA), equilibrated in 20 mM NaC1, and eluted with a 20-500 mM NaCl gradient in 10 mM Tris-HC1, pH 7.5. For the Cys analogs, concentrated fermentation broths were applied to a column of CM52 (Whatman) in 10 mM Nap04 buffer, pH 7.5, and eluted with a 0.0-0.3 M NaCl linear gradient. Further purification on Blue Toyopearl was as described above for rBRIP. Construction of BRIP Immunoconjugates. For conjugation to mAb, nBRIP or rBRIP (3 mg/mL) was first modified with 0.5 mM 5-methyl-24minothiolane (M2IT) (20) and 1 mM DTNB in 25 mM triethanolamine-HC1, 150 mM NaC1, pH 8.0, for 3 h at 25 “C. The derivatized BRIP-(M2IT)-TNB was then desalted on a column of GF-05LS (IBF Biotechnics), and the number of thiol groups introduced was quantitated by measuring absorbance at 412 nm after the addition of 0.1 mM DTT. On average, each BRIP molecule contained 0.7 SH/mol. 4A2 or H65 antibody (4 mg/mL) in the same triethanolamine buffer was similarly incubated with M21T (0.3 mM) and DTNB (1mM) for 3 h at 25 “C. The modified antibody (for example, H65-(M2IT)-TNB was then desalted, and the TNB/antibody ratio was determined. Typically, the number of linkers per antibody was in the range of 1.7 to 2.
To assemble the conjugate: (1)activated BRIP-(M2IT)TNB was first reduced to BRIP-(MBIT)-SH with 5 mM
128
Bernhard et al.
Bioconjugate Chem., Vol. 5, No. 2, 1994
DTT for 1.5 h at 25 "C; (2) the reduced BRIP-(M2IT)SH was desalted on GF-05LS to remove excess reducing agent and TNB leaving group, and (3) a 5-fold molar excess of BRIP-(M2IT)-SH was mixed with activated antibody(M2IT)-TNB. After 3 h at 25 "C and an additional 18 h at 4 "C, an equimolar amount of mercaptoethylamine was added for 15min at 25 "C to quench any unreacted linkers. The quenched conjugate was promptly applied to an Ultrogel AcA44 (IBF Biotechnics) equilibrated in 10 mM Tris-HC1, 100 mM NaC1, pH 7. The partially purified material was subsequently diluted with Tris-HC1, pH 7, and applied to a Blue Toyopearl column equilibrated in 10 mM Tris-HCl,20 mM NaC1, pH 7, and eluted with 10 mM Tris-HC1,0.5 M NaCl, pH 8. Conjugates prepared in this manner contain two M2IT linkers and are denoted MM, for example, 4A2-MM-nBRIP. Prior to conjugation of the BRIP Cys analogs rBRIPc256, rBRIPc270, or rBRIPc277, each was treated with 50 mM DTT (2 h at 23 "C or 18 h at 4 "C) to expose the single sulfhydryl group and then desalted. Free sulfhydryl content was verified by reaction with DTNB as described above. To assemble each conjugate, M2IT-derivatized mAb was incubated with a 5-fold molar excess of freshly reduced rBRIP analog at 23 "C for 3 h and then 16 h at 4 "C. After being quenched with mercaptoethylamine, each conjugate was purified by sequential size exclusion and affinity chromatography as described above. Samples of all immunoconjugates were examined by 5 % SDS-PAGE under nonreducing conditions. Gels were stained with Coomassie R250 and scanned with a Shimadzu laser densitometer to quantitate the number of RIP molecules per mAb. The RIP to mAb ratio was 1.5 for 4A2-M-rBRIPc256, 0.93 for 4A2-M-rBRIPc270, 1.77 for 4A2-M-rBRIPc277, and 2 for 4A2-MM-nBRIP. The RIP to mAb ratio was 1.5 for H65-M-rBRIPc256,1.43 for H65M - ~ B R I P c ~ 1.9 ~ o for , H65-M-rBRIPc277, and 1.1for H65MM-rBRIP. Preparationof H65-RTA and 4A2-M-RTAm. H65RTA is an immunoconjugatecomposedof H65 mAb linked via SPDP to RTA (21,221. The RIP to mAb ratio was 1.9 for H65-RTA. There was less than 5% free mAb in the H65-RTA preparation and no free RTA. 4A2-M-RTA30 is an immunoconjugate composed of 4A2 mAb linked via M2IT to RTA30, the 30 kDa glycoform of RTA (16, 23). The RIP to mAb ratio was 1.1for 4A2-M-RTA30. There was no detectable free mAb or RTA30 in the 4A2-MRTAso. Activities of RIP and Immunoconjugates. The ability of RIP to inhibit protein synthesis in a cell-free system was examined in the rabbit reticulocyte lysate assay (RLA;23,24). The incorporation of 3H-Leu was measured as a function of RIP concentration, and the concentration of RIP (in pM) which inhibited protein synthesis by 50% (the IC50) relative to untreated controls was calculated. All samples were tested in triplicate. The cytotoxicity of immunoconjugates or RIP for HSB2 cells or PHAactivated human peripheral blood mononuclear cells (PBMC) were examined as described (25,261. Cells were incubated with increasing concentrations of immunoconjugates for a total of 24 h (HSB2 cells) or 90 h (PBMC), and inhibition of macromolecularsynthesis was quantified. By comparison with untreated controls, the concentration of immunoconjugate (or RIP) that inhibited protein synthesis by 50% (the IC50) was calculated. Results are expressed in pM RIP, after multiplying the IC50 (in pM) by the RIP to mAb ratio. HPLC Disulfide Bond Stability Assay. Immunoconjugates (ca. 0.5 mg/mL) in 0.1 M NaP04,0.15 M NaCl,
Table 1. Inhibition of Protein Synthesis in Vitro by BRIP and rBRIP Analogs
RLAa ICsoc (PM) RTA3o nBRIP rBRIP rBRIPczM rBRIpcz70 rBRIPcz77
3.1 15 18
23 20 24
cellular cytotoxicityb Ic50 (PM) 20 000 170 000 170 000
260 000 300 OOO >500 000
Inhibition of protein synthesis in the cell-free rabbit reticulocyte lysate assay. Inhibition of protein synthesis in intact HSB2 cells by RIP. The IC50 is the concentration of RIP (in pM) that inhibits protein synthesis by 50%.
1.5 mM EDTA, pH 7.5, were incubated at 37 "C for 1 h with increasing amounts of reduced GSH, (0.1-50 mM). A control reaction was incubated in the absence of reductant, and a 100% RIP release control sample was prepared by incubation in 60 mM 2-ME. The reduction was terminated by addition of iodoacetamide (final 100 mM) and incubation at 37 "C for 20 min. Samples were filtered (0.45pm) and analyzed on a Waters HPLC system equipped with a TSK-G2000 (30- X 0.78-cm)size-exclusion column and a 280-nm detector. The column was eluted with 0.1 M Na~S04,0.02 M Na2P04, pH 6.8, at 0.7 mL/ min, and the absorbance of the effluent at 280 nm was recorded. The amount of RIP released from each conjugate after treatment with GSH was quantified by area integration and compared to a fully reduced (2-ME) sample. The percent of RIP released was plotted as a function of the log mM GSH concentration, and an RCm for each conjugate (the GSH concentration required to release 50% of the RIP) was calculated. The recovery of total RIP was consistently high (usually >80 % ) compared to the theoretical maximum calculated from the RIP to mAb ratios for each conjugate. RESULTS
Characterization of nBRIP. nBRIP was purified from barley seeds to assess (i) its ability to inhibit protein synthesis in uitro and (ii) its properties when conjugated to mAbs that recognize surface antigens on human T cells. As previously demonstrated (8-10,27), nBRIP is a basic protein (PI = 9.7) with an apparent M , of 30 kDa, based upon its mobility when analyzed by SDS-PAGE (data not shown). Consistent with earlier reports ( 2 3 , the N-terminus of nBRIP was blocked. However, the amino acid sequence of BRIP fragments generated by limited proteolysis was consistent with that previously published (27) for isoform I1 from barley seeds. The enzymatic activity of nBRIP relative to RTA30 was examined in the RLA assay as shown in Table 1. In repeated assays nBRIP was about 5-fold less potent at inhibiting protein synthesis in uitro than was RTA30. Also shown in Table 1 is the nonspecific cytotoxicity of both nBRIP and RTA30, measured against intact human T cells. nBRIP was about 8-fold less cytotoxic to the HSB2 cell line than was RTABo. Cloning, Expression, and Activity of rBRIP. Although barley flour is readily available for the purification of nBRIP, we were interested in cloning the BRIP gene so that recombinant production and protein engineering of BRIP were possible. A BRIP cDN,Aclonewas identified directly from a barley cDNA library in a phage X vector. The DNA sequence of the BRIP cDNA we obtained was identical to the barley RIP isoform I1 cDNA ( I 7;Genebank Accession no. M36990) and encodes a protein of 280 amino acids.
Bioconjugate Chem., Vol. 5, No. 2, 1994
Cysteine in rBRIP Improves Conjugate Potency
MW std. (KD)
1
2
3
4
5
129
6
180 150
istl Mscl
/
BamHl
1 Sstll
Figure 1. Assembly of the rBRIP expression vector. A schematic view of the rBRIP expression vector pING3322 is shown. The relative positions of the tetracycline resistance gene (tet), the araC regulatory gene, and the pelB leader are indicated. Also shown are the positions of primers used for PCR amplification of the BRIP 3’-end and the restriction sites used for gene assembly: XhoI, NcoI, BamHI, EcoRI, SstI, SstII, and MscI. For vector assembly, the cDNA clone containing BRIP (pBS1) was cut with NcoI, treated with mung bean nuclease (MBN), and cut withBamH1, and the 760bp fragment containing the majority of the BRIP coding sequence was then purified. Concurrently, the 3’ end of the BRIP gene was amplified by PCR with the vector primer reverse (5’-AACAGCTATGACCATG-3’) and the specific primer BRIP 3’xho (5’-TGAACTCGAGGAAAACTACCTATTTCCCAC-3’). The reaction product was digested with BanHI and XhoI, and the 87bp fragment containing the 3’-end of the BRIP gene was purified on an 5% polyacrylamide gel. The two purified fragments were ligated adjacent to the peZB leader sequence from Erwinia carotauora (18)in pING1500, which had been cut with SstI, treated with T4 polymerase, cut with XhoI, and purified. The finalpeZB::BRIP expressionvector, pING3322, was then assembled by placing the gene under the control of the inducible araB promoter (19).
Once identified, the BRIP gene was expressed in E. coli under control of the aruB promoter by induction with arabinose. A schematic view of the recombinant BRIP expressionvector is shown in Figure 1. The mature BRIP coding sequence was linked to a bacterial signal sequence, and the recombinant product could be purified directly from the E. coli culture supernatant. Induction of E. coli containing the plasmid pING3322 in a fermenter resulted in the production of soluble rBRIP at a yield of about 1 g/L. N-terminal sequence analysis of the purified protein demonstrated that the leader peptide was properly removed and that the N-terminus was not blocked (data not shown). E. coli-produced rBRIP exhibited the same molecular weight as nBRIP when analyzed by SDS-PAGE and reacted identically by ELISA and on Western blots with rabbit antisera raised against nBRIP (data not shown). nand rBRIP also had comparable enzymatic activities in the RLA and the same low level of nonspecificcytotoxicity against HSB2 cells (Table 1). Cloning, Expression, and Activity of rBRIP Cysteine Analogs. To explore the possibility of site-specific conjugationwith rBRIP (whichcontains no Cys residues), three analogs were generated which contained a single Cys residue near the C-terminus, in place of Leu 256, Ala 270, or Ser 277. These positions were chosen by comparison of the amino acid sequence of BRIP with the known tertiary structure of RTA (3),which suggested that these three positions would be at the surface of the molecule and therefore available for conjugation to antibody.
Figure 2. Position of the introduced cysteine residue influences rBRIP conjugation yield. Activated H65 antibody (H65-M2ITTNB) and BRIP-SH were incubated for 3 h a t 25 “C and an additional 18 h a t 4 “C. Aliquots of reaction mixtures were then analyzed by SDS-PAGE on 7.5 % gels without reduction. Shown are activated H65 antibody alone (lanes 1,3, and 5) or reaction mixtures containing activated H65 antibody and rBRIP-M2IT (lane 2), rBRIPc277 (lane 4), or rBRIPc270 (lane 6).
The rBRIP Cys analogs were produced in the same expression system shown in Figure 1. They exhibited the same molecular weight as rBRIP when analyzed by SDSPAGE and showed some tendency to form dimers in the absence of reducing agents (data not shown). Following reduction with DTT, the 1:l molar Cys content of each analog was confirmed by DTNB analysis (28). The ability of the rBRIP Cys analogs to inhibit protein synthesis in the RLA is shown in Table 1. Each analog was approximately as effective at inhibiting protein synthesis as rBRIP. These data suggest that Cys substitution at positions 256,270, or 277 did not adversely affect rBRIP synthesis, structure, or catalytic activity. Cysteine Position Alters rBRIP Analog Conjugation Yield. Antibody immunoconjugates were prepared with rBRIP and the three Cys analogs. Each conjugation reaction mixture was analyzed by densitometry of Coomassie stained SDS polyacrylamide gels, in order to quantitate the conversion of free antibody to immunoconjugateproduct. Figure 2 showsthe unpurified mixtures of H65 conjugated to rBRIP-M2IT, rBRIPc270, or rBRIPc277, in which the average number of M2IT-TNB linkers on H65 was between 1.7 and 2. Effects of the BRIP cysteine position on conjugation yield are evident. While the conversion of H65 mAb to immunoconjugatewas about 30% with linker-modified rBRIP (lanes 1 and 2), conversion was enhanced with rBRIPc277 (78% lanes 3 and 4) and diminished with rBRIPc270 (lo%,lanes 5 and 6). Not shown is the reaction with H65 and rBRIPc256, which yielded about 10% conversion. There is a correlationbetween conversion efficiency and formation of multimeric conjugates. Thus, the reaction of H65 and rBRIP-M2IT (in lane 2) produced both monoand diconjugates (one or two rBRIP per MAb). The reaction with rBRIPc270 (in lane 6) which was less efficient at promoting conjugation with H65 mAb made essentially monomer (one rBRIP per mAb). In contrast, the reaction of H65 conjugated with rBRIPc277 (in lane 4) produced a mixture of mono-, di-, tri-, and tetraconjugates (one to four rBRIP per mAb). Stabilityof BRIP Immunoconjugates to Reduction. The relative stability of the disulfide bond linking mAb and BRIP moieties was determined for a series of conjugates made with nBRIP, rBRIP, and the recombinant Cys analogs. The results are shown in Table 2; again, the position of the engineered BRIP Cys had a noticeable effect. Whereas 4A2 and H65 conjugates made with RTA, nBRIP, or rBRIP were all of similar stability (RC50s 3-7 mM), 4A2-M-rBRIPc270, 4A2-M-rBRIPc277, and H65-
Bernhard et al.
130 Bioconjugate Chem., Vol. 5, No. 2, 1994
Table 2. Disulfide Bond Stability of BRIP Conjugatesa conjugate RCm (mM GSH)
conjugate H65-RTA H65-MM-rBRIP
RCm (mM GSH)
7.0 2.8 4A2-MM-nBRIP 53.0 H65-M-rBRIPc277 196.0 4A2-M-rBRIPc270 187.0 4A2-M-rBRIPc277 0 The relative disulfide bond stability of a series of BRIP conjugates was determined by measuring the amount of RIP released after 1 h of incubation with increasing amounts of GSH. The RCm for each conjugate is the concentration of reducing agent (GSH)needed to release 50% of RIP from a given conjugate (see also ref 29). A higher RCm indicates increased disulfide bond stability. 4.4 3.3
4A2-M-RTA30
-2
"
100-
C
0
ij
80-
u)
60-
I
+.....
2-
5
5.
4 0 - -t- H65-MM-rBRIP
v)
-
'
-E
20-
2
'
--t
C
a
0
!
'
'
.1
'
H65-M-rBRIPC270 H65-M-rBRIPC277 H65-RTA '
""'I
'
'
"'"'I
10
1
'
" ' " ' I
100
'
'
"""I
'
'
' ' 1
1000
10000
Conjugate Concentration (ng/ml)
= 2 "
-
100-
C
0 r
80 -
0
UI
4A2-M-rBRIPC270
--C
4AZ-M-rBRIPC277
--O-
4A2-M-RTA30
60-
DISCUSSION
........
u)
5
5
4A2-MM-nBRIP
--t-
40-
v)
-" C
Q
2
20-
a
0 .1
positive T cells. The IC50 values and extend of kill are summarized in Table 3. Although 4A2-MM-nBRIPinhibited protein synthesis in HSB2 cells with an IC50 of 122 pM, it was about 10-fold less effective than 4A2-MRTAso. 4A2 conjugates made with rBRIPc270 or rBRIPc277 were 2- to 3-fold more potent than 4A2-MM-nBRIP. In contrast, the conjugate made with rBRIPcm was less active than the native BRIP immunoconjugate. The H65-MM-rBRIP conjugate was virtually noncytotoxic to HSBB cells. At the highest concentration tested (1000 ng/mL), inhibition of protein synthesis was less than 50 % . Two conjugates, H65-M-rBRIPc270 and H65-MrBRIPc256, were likewise not very effective at inhibiting translation in HSBB cells. In contrast, H65-M-rBRIPc277 showed enhanced potency toward the target cells, with an IC50only 8-fold less than that of the H65-RTA control, and approached 70% inhibition, compared to 78% for H65-RTA. These conjugates were tested in similar experiments against human PBMC. As shown in Table 3, there was a trend toward enhanced cytotoxicity with some of the rBRIP Cys analog conjugates. 4A2-M-rBRIPc277 and 4A2-M-rBRIPc270 were, respectively, 10 and three times more cytotoxic than 4A2-MM-nBRIP. Both H65-MrBRIPc270 and H65-M-rBRIPc277 were five times more cytotoxic than H65-MM-rBRIP.
1
10
100
1000
10000
Conjugate Concentration (nglml)
Figure 3. Cytotoxicityof BRIP immunoconjugates. The ability of several immunoconjugates to inhibit protein synthesis (incorporation of SH-Leu)in HSB2 cells was measured as described in the ExperimentalProcedures. A representative assay is shown. The H65 conjugatesare shown in panel A and the 4A2 conjugates in panel B. The H65-M-rBR1p~~~ and 4A2-M-rBRIPc2s conjugates were not tested in this assay. M-rBRIPc277 were 10-50 times more stable to reduction (RC~OS 50-200 mM). It is likely that small variations in RIP to mAb ratios would have little impact in this assay, since different lots of H65-RTA containing RTA to H65 ratios between 1.0 and 3.0 gave identical RC50 values (data not shown). In Vitro Cytotoxicity of BRIP Conjugates. The 4A2- and H65-BRIP conjugates were tested for specific cytotoxicity against the human T cell line HSB2, which expresses both CD5 and CD7 antigens. Dose-response titration curves are shown in Figure 3A and B, where BRIP immunoconjugates are compared to H65-RTA (15,251or 4A2-M-RTA30 (16) which are both cytotoxic to antigen
Nonspecific modification of proteins with cross-linking reagents can adversely affect their subsequent activity. Nevertheless, most immunoconjugatesbetween antibodies and cytotoxic plant proteins such as the type-I RIP rely on chemical modification of both antibody and RIP lysine residues. Retention of biologic activity depends upon both the degree of modification and, unpredictably, the reactivity of active site or important structural amino acids. For example, modification of BRIP, the type-I RIP from barley, with 2-iminothiolane greatly inhibits its activity. An average of 1.4 thiols per BRIP was reported to reduce activity about 5-fold, while 4.0 thiols per molecule nearly abolished activity (10). Similar results have been observed with other cross-!.inking reagents and other type-I RIP including Bryodin, Gelonin, Momordin, Pokeweed antiviral protein, Saporin, and Trichosanthin (14). Immunoconjugates that result from such randomly derivitized RIP are heterogeneous with regard to the physical orientation of the two linked proteins. Thus, random derivitization of RIP can lead to decreased enzymatic activity and increased conjugate heterogeneity. The complex nature of the immunoconjugate pool can be simplified if a single amino acid in the cytotoxic subunit is linked to antibody. For example, disulfide-linked RTA conjugates can be assembled using the available Cys residue a t position 259. In previous experiments (28), we found that by genetic manipulation of the fungal protein mitogillin to expose an accessible cysteine immunoconjugates could also be generated without initial derivatization by nonspecific cross-linking agents. Like many
Cysteine in rBRIP Improves Conjugate Potency
Bioconjugate Chem., Vol. 5,
Table 3. Cytotoxicity of BRIP Immunoconjugates for Human T CellsP HSB2 cells conjugate IC50 (pM RIP) % inhibition 4A2-M-RTA3o 15f3 95 4A2-MM-nBRIP 122 f 6 85 4A2-M-rBR1Pcz~ 464 f 92 72 4A2-M-rBRIPc270 46 f 3 85 4A2-M-rBRIPc277 57 f 7 85
No. 2, 1994 131
PBMC IC50 (pM RIP) 34 (20-48) 744 ND 226 (130-321) 60 (45-74)
H65-RTA H65-MM-rBRIP H65-M-rBRIPc256 H65-M-rBRIPc270 H65-M-rBRIPcz77
% inhibition
89 80
ND 80
79
150 f 40 78 429 (46-1170) 70 >5000 5000 5000
