Heavy and light chain contributions to antigen binding in an anti-digoxin chain recombinant antibody produced by transfection of cloned anti-digoxin antibody genes

Moietular Immunol<>g~.Vol. 27. No. 9, PP. 901-909, 1990 Printed in Great Britain

0161-589Oj90 $3.00 + 0.00 Pergamon Press plc

HEAVY AND LIGHT CHAIN CONTRIBUTIONS TO ANTIGEN BINDING IN AN ANTI-DIGOXIN CHAIN RECOMBINANT ANTIBODY PRODUCED BY TRANSFECTION OF CLONED ANTI-DIGOXIN ANTIBODY GENES* RICHARD I. NEAR,Q

SHI CHUNG NG,$$// MEREDITHMUDGETT-HUNTER,~$

NORMAN W. HUDSON,~ MICHAELN. MARGOLIES,~** JONATHANG. SEIDMAN,?? EDGAR HABER:$ and MARLENEA JACOBSON$§~+ $The Cellular and Molecular Research Laboratory, Massachusetts General Hospital and Departments of §Medicine and **Surgery, Harvard Medical School, Boston, MA 02114. U.S.A.: FDepartment of Medicine, Indiana University School of Medicine, Indianapolis, IN 46223. U.S.A.; and ftDepartment of Genetics, Harvard Medical School, Boston, MA 02115, U.S.A. (First received 9 January 1990; accepted 27 February 1990) Abstract-We used immunoglobuhn gene transfection to study the effect that substituting an homologous light (L) chain for a parental L chain has on antigen fine specificity and affinity. High-affinity monoclonal anti-digoxin antibodies 26-10 and 40-100 were selected for study because their L chains are 92% homologous (although the H chains differ), and their binding with digoxin and digoxin analogs show very different properties. In order to generate a recombinant transfectoma, the genes encoding the 26-10 H and L chains were cloned. After the sequenced clones had been shown to contain the V gene and the transcriptional control elements, the H and L chain V region genes were subcloned into different expression vectors. Both constructs were transfected into myeloma J558L. a I1 chain producer, to verify that the genetic constructs expressed correctly. The recombined 26-10 antibody was identical to parental 26-10 antibody in fine specificity and affinity. The 26-10 L chain construct was then transfected into a cell line, CR-101, that expresses the 40-100 H chain and a II chain. The transfe~toma 1E6, secreting 40-iOO H chain and 26-10 L chain, was selected. Appropriate gene expression in 1E6 was proven by polymerase chain reaction cloning and sequencing. The fine specificity properties of the 1E6 recombinant derive from both the 40-100 and 26-10 antibodies; however, the affinity of 1E6 is 130 times less than that of the parental antibodies. We conclude that, in lE6, the H and L chains are codominant in their influence on antigen specificity and that homologous pairing of H and L chains is required for optimal affinity.

INTRODUCTION

Antibodies have an enormous repertoire of sequence variations (Kohler, 1986; Milstein, 1986; Ah et al., 1987). The genetic m~hanisms that cause this diversity have been well studied and partiaily explained. It has not been determined, however, which of the possible genetic mechanisms responsible for diversification are most likely to alter positions within the antibody protein that affect antigen binding. To contribute to an understanding of which positions affect antigen binding, we have been studying the *This work was supported by National Institutes of Health Grant HL-192.59. tAuthor to whom correspondence should be addressed at Massachusetts General Hospital. IiCurrent address: The Squibb Institute for Medical Research, P.O. Box 4000, Princeton, NJ 08.543, U.S.A. ttcurrent address: Merck Sharp and Dohme Research Laboratories, W42-300, West Point, PA 19486, U.S.A. Abbreviations: VH, heavy chain variable region; VL, light chain variable region; kb, kilobase; CDR, complementarity determining region; PCR, polymerase chain reaction.

structure-function relations of a group of antidigoxin monoclonal antibodies (Mudgett-Hunter et al., 1982, 1985). These antibodies were selected because (1) each antibody has a very high affinity for digoxin; (2) there are numerous digoxin analogs available with which to study fine specificity (Fieser and Fieser, 19.59); (3) the crystallographic structures of digoxin and several analogs show a conformationally rigid molecule large enough to occupy most of the antibody combining site (Kabat 1966; Go et ul., 1980; Haber and Margohes, 1984); and (4) among these monoclonal antibodies are subsets that possess homologous VH and/or VL sequences that differ at only a few amino acids (Mudgett-Hunter et al., 1985; Panka and Margolies, 1987; Hudson et al., 1987). Some of the monoclonal anti-digoxin antibodies have highly homologous VL regions even though their VH regions are derived from different VH families (Mudgett-Hunter et al., 1985). Specifically, antibodies 2610,40-20, 40-60,40-90 and 40-100 have VL regions that are greater than 90% homologous. With the exception of 26-10, these antibodies also have VH regions so alike that they probably are

902

RICHARD1. NEAR ef of.

derived from the same VH germline gene segment (Mudgett-Hunter er al., 1985; M. N. Margolies, unpublished~. Antibody 26 10, containing a unique VH region, also differs from the four other antibodies in its fine specificity pattern. To assess the relative contributions VH and VL regions make to the overall fine specificity and affinity of the antibodies, we performed chain recombination experiments either by reannealing isolated H and L chains in vitro or by hybridoma-hybridoma fusion (Hudson et al,, 1987). Attempts to recombine the light and heavy chains of digoxin-specific antibodies in solution have been generally unsuccessful. However, cell fusion of an H chain-loss 40-20 cell line with a cell line that secreted 26-10 VH and an irrelevant 11 L chain resulted in a recombinant antibody that exhibited specificity properties similar to those of the 26-10 VH donor, but that had antigen afiinity two orders of magnitude lower than the affinities of both donors. To determine the relative influence H and L chains have on the antigen combining site and resultant antigen specificity and a%nity, we examined what is essentially an inverse (to that described above) chain recombinant: the 400-100 H chain associated with the 26-10 L chain. Due to the advantages of being able to manipulate antibody genes by in vitro mutagenesis, we chose to clone the 26-10 antibody genes and insert them into expression vectors in order to form chain recombinants. When the 26-10 L chain gene was transfected into a cell line that produces the 40-100 H chain, the antigen binding properties of the recombinant antibody were shown to possess characteristics derived from both donor antibody chains, MATERIALS AND METHODS Cell

lines

The hybridomas producing monoclonal antidigoxin antibodies 26-10 (y2a, K), 40-100 (71, K), and 45-20 @t2a, i, 1) were generated by fusing immune A/J mouse splenocytes with SpZ/O as previously described (Mudgett-Hunter et al., I982, 1985). The fusion chain loss line CR-101 contains the 40-100 H chain and the 45-20 il IL chain and does not bind digoxin. CR- 101 was generated by fusing 40- 100 cells with a heavy chain loss variant of 45-20 and subsequently selecting for toss of the 40-100 K L chain in the fused line (N. W. Hudson, unpublished). The method used to generate these cell-recombinants has been described in detail (Hudson et al., 1987). Cell line J558L (;I 1) is a heavy chain ‘loss variant of myeloma J558 (donated by S. Morrison, Department of Microbiology, University of California at Los Angeles). All cells were maintained in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal calf serum (GIBCO, Grand Island, NY), SO~g/ml of gentamycin sulfate (GIBCO), and 0.6 mgjml L-glutamine.

Cloning of the N and L chain 26 10 immunoglobulin genes

High molecular weight genomic DNA from the 26-10 cell line was isolated and Southern blot analysis was performed as described (Near and Haber, 1989). Rearranged H chains were detected with a I .4 kb Pstl fragment that contains JH4 (Marcu et al., 1980; Near and Haber, 1989). Rearranged L chains were detected with a 2.7 kb HindIII probe that contains the entire frc region (Near and Haber, 1989). Two H chain (EcoRI) and two L chain (BarnHI) rearrangements were detected that were not present in Sp2/0 or A/J mouse (the fusion parents) DNA by Southern blot analysis. For cloning, 1 mg of digested DNA was fractionated by preparative agarose electrophoresis (Gene Machine, Hoeffer Scientific, San Francisco, CA) and fractions that hybridized with the probes were isolated. Fractions containing H chain DNA were cloned into i,gtlO and fractions containing L chain were cloned into EMBL3 (Maniatis ec al., 1982). The clone containing the functional H chain rearrangement was detected by hybridization with a If base ol~gonueleotide (5’~~CATGGAAA~A~~~TT3‘~ corresponding tu amino acids JO-45 of H chain as initially determined by partial mRNA sequencing (Clarke et al., 1985; Panka and Margolies, 1987; N. W. Hudson and M. N. Margolies, unpublished~. The clone containing the functionaf L chain rearrangement was detected by hybridization with a I.1 kb 5’ flanking region of the TECP-lOSVtc-l clone (homologous to 26-10~) of Moynet et al. (1985). The entire H and L inserts were subcloned into pBR322.

The I.7 kb Xbaf fragment containing 26-1OVH and the 2.7 kb HindIIf-XbaI fragment containing 26-1 OVK were cloned into M I3mp 18 and M 13mp 19 (Messing, 1983). The V regions and flanking areas were sequenced from both strands by the dideoxy chain-te~~nation method @anger et al., 19773, using Sequenase (United States B~ochemicals~ Cleveland, OH) and adenosine 5’-[rw-“(S)thiojtriphosphate (Amersham, Arlington Heights, IL). Oligonucleotide primers were synthesized on the Applied Biosystems 380B DNA Synthesizer.

The I 1.5 kb 26-IOVH EcoRI fragment (Fig. 1Af was cloned in the correct orientation into the vector pSV2gpt (Mulligan and Berg, 1980) that had been modified by the insertion of a polylinker at the EcoRI site (R. I. Near, unpublished~ and by the insertion into the XbaI site of the polylinker of a 5.5 kb XbaI fragment containing the germline murine 72b constant region gene (Marcu er al., 1980). The 11.2 kb 26-IOVL BamHI fragment that contains both the V and CK regions (Fig. 1B) was cloned into pBR322 along with a 2.6 kb BarnHI fragment that contains

903

Anti-digoxin transfectoma antibody binding properties

&

A EHB I1 1

X

B

X

XB

B

I

I

I I

XE

S’-VDJ-3’

- I kb

B B

E

l-i

I

I

I

EH 1 I

x I

H u

H I

B U

Ck

S’VJ-3’

Fig. I. Restriction site maps of hybrjdoma 26-10 immunoglobui~n genomic gene clones. (A) The restriction map of the functional heavy chain gene of 26-10 was derived from the cloned 11.2 kb EcoRI fragment. The location of the rearranged V gene is shown 5’ to 3’. (B) The restriction map of the functional 26-10 K chain gene was determined from the cloned 11.5kb BamHI fragment, B, BamHI; E, EcoRI; H, HindIII;

X, XbaI. the Tn5 neomycin fneo) resistance gene (Southern and Berg, 1982). The H and L constructs are denoted pSVgpt/Z&IOVH and pneo/26IOVL, respectively. DNA transfection Construct DNA was introduced into myeloma and hybridoma cells using the voltages and conditions described by Potter et al. (1984). Briefly, 3 x 10’ cells in 1.0 ml PBS at 4°C in the presence of 20 yg construct DNA were given a 2000 V pulse. The cells were distributed into 96-well plates at 3 x lo4 cells/well and incubated at 37°C for 48 hr, after which selection medium was added. For Ecogpt selection {MuIligan and Berg, 1980). 6 pg/ml my~ophenolic acid, 250 pg/ml xanthine and 1.5pg/ml hypoxanthine was used in the medium. For the selection of neo gene-containing cells, I .Omg/ml G418 (GIBCO, Gaithersburg, MD) was used. Transfected cell colonies were usually visible within two to three weeks of selection. RNA isolation and Northern analysis RNA was prepared by standard guanidinium isothiocyanate cell lysis followed by CsCl step gradient centrifugation and analyzed by formaldehyde-based Northern blotting (Maniatis et al., 1982). H chain RNA was detected with a 0,~ probe consisting of a 17 base oligonucleotide [5’GGGGCCAGTGGATAGAC3’] complementary to Cy 21 bases 3” of the V-Q junction (Tucker et al., 1979; Panka and Margolies, 1987). Expression of the 26-10 L chain RNA was determined with a 1.4 kb BglII fragment (derived from the 26-10 L chain clone) containing the 3’ portion of 26-IOVk- including the JK regions. Competition radioimmunoassay ,fbr fine speciJicity The fine specificities of anti-digoxin antibodies from the hybridoma and transfectoma supernatants were determined by measuring the competition between digoxin and its analogs in an RIA (Hudson et al., 1987; Mudgett-Hunter et al., 1985). Briefly, cell supernatants were incubated in microtiter plates coated with affinity-purified goat anti-mouse F(ab’)? (ICN Biomedicals, Lisle, IL) at 4°C overnight. The

following day, the plates were washed with distilled water and the wells were filled with 1% horse serum in PBS containing 0.01% sodium azide (PBSA) to block any remaining nonspecific protein binding sites. After the plates had been washed, cardiac glycosides (IO-” to 10m4M) in PBSA containing 1% horse serum and 5% ethanol were added to the wells. Immediately after addition of each analog, 50,000 cpm of [‘251]-digoxin-BSA (iodinated by the chloramine T method) (Hudson et al., 1987) was added to the same wells and allowed to incubate overnight at 4°C. The following day, radioactive ligand was removed and the plates were extensively washed with distilled water before each well was cut from the plate and counted (Micromedic Systems Automatic Gamma Counter, Model 4/600). Antibodies were compared by determining the amount of unlabeled analog required to achieve 50% inhibition relative to the amount of unlabeled digoxin required for 50% inhibition. Antibody isotypes were determined with the Mouse Ig Subtype Identification Kit (Boehringer Mannheim Biochemicals, Indianapolis, IN). Afinity measurement A double antibody precipitation assay was used to determine the affinity constant, Ko, as previously described (Hudson et al., 1987). Briefly, supernatant from each hybridoma or transfectoma was diluted with 1% horse serum in PBSA to a concn estimated to be approximately the Kd or lower and was incubated with varying concns of [3H]digoxin (25 pCi/ mmole, New England Nuclear, Boston, MA) overnight at 4°C. The antigen-antibody complex was precipitated with rabbit anti-mouse IgG (ICN Immunobiologicals, Inc., Lisle, IL) overnight followed by goat anti-rabbit IgG (a gift of Dr Charles Homey, Mass. General Hosp., Boston, MA) for 4 hr at 4°C. The precipitated complex was filtered onto glass fiber filters (24mm, No. 32, Schlechter and Schuel, Keene, NH) on a Millipore manifoid (Millipore, New Bedford, MA). The filters were placed in 5 ml scintillation fluid (Ultima Gold, Packard, Downers Grove, IL) and were counted in a Packard 1500 Tri-Carb Liquid Scintillation Analyzer. The data

904

RICHARDI. NEAR et al.

were analyzed with the curve fitting program LIGAND (Munson and Rodbard, 1980) which calculates K, and the antibody concn. Neither the fine specificity nor the affinity measurements appear to be affected by the presence of an extraneous >“I chain (in addition to the ti L chain) in the transfectomas studied here. We believe this is because the i I L chain, when associated with either the 40-100 or 26-10 H chain, results in an antibody that has no detectable binding to digoxin or digoxin analogs (R. Near and N. Hudson, unpublished observations; Hudson of al., 1987). Polymerase

chain reaction ampl@ation

and cloning

Transcription of the appropriate mRNAs was demonstrated by polymerase chain reaction (PCR). Briefly, whole cell RNA was isolated from the transfectoma cell line as described above and a cDNA copy was generated using constant region primers (Panka and Margolies, 1987). The cDNA was used as a template for PCR amplification with the Gene-Amp DNA Amplification Reagent Kit (according to the manufacturer’s instructions) in a DNA Thermal Cycler (Perkin Elmer Cetus, Norwalk, CT). Primers used in these experiments contained restriction sites for the subsequent cloning of the synthesized fragments into M I3 for sequencing (Scharf ef ai., 1986: S. Y, Shaw and M. N. Margolies, unpublished). The PCR primers for 26-1036 were [S’GAG~TCGAG~TCA~TGGATGGTGGGAAG3’1. _._^.___. _~~_ a CK primer containing a Sac1 site, and [5’GGTGCTGATGTTCTAGATTT 3’1, a Vti primer containing an Xbal site. The primers for 40-100 H were [S’CCAGGCATTCTAGAGTCACCGAGG3’], primer containing an Xbaf site, and a C; ~S’GCTTAAGCTTGAGGTACACCTGGTGGAAT C3’], a 40-IOOVH primer containing a Hind111 site. RESULTS

Ctoning the VH h~br~dorn~~26- 10

und

VL .finctional

genes ,fi-om

In order to generate the chain recombinants by gene transfection, it was necessary to first clone the 26-10 functional immunoglobulin genes. Southern analysis (not shown) of DNA from 26-10 showed two heavy chain EcoRI rearrangements (7.5 and 1I .5 kb) that were not present in either Sp2/0 or normal A/J mouse DNA (fusion partners for the original production of 26-10). Since either fragment could contain the functional rearranged gene, both were cloned. An oligonucleotide complementary to the partial mRNA sequence of 26-IOVH ‘(see Materials und ~e~~?~~d~) showed that the 11.5 kb clone contained the 26-IOVH. Southern analysis of 26-10 DNA showed that there were two BamHi fragments, of 4.3 and 1I.2 kb, that hybridized with the JK probe and were not found in Sp2/0 and A/J mouse DNA. Since the known amino acid sequence of 26-IOL (Novotny and Margolies, 1983) shows that Jh-I is part of the

functional gene and since the distance from Jkl to the 3’ BamHI site is known (Alt et al., 1987), the 4.3 kb fragment can be eliminated as being too small to encode the functional gene. Therefore, the 11.2 kb fragment was cloned into EMBL3 and was confirmed to contain 26-lOVL by hybridization with a 26-IOVLhomologous probe from the myeloma TEPC-IOSVL region (Materials and Methods). Figures IA and 1B show the restriction site maps of the functional 26-10 H and L chain genomic clones, respectively. The H chain clone contains the rearranged VH region and the enhancer (adjacent to the 3’ EcoRI site), but no constant region exon. The L chain clone contains the rearranged VL region, the enhancer, and the CK region. DNA seqz~ences of 26-10 VH and VL In order to confirm the identity of the cloned functional V genes and to confirm that the rearranged genes contained the sequences needed for expression, we determined the genomic DNA sequences of the 26-10 immunoglobulin gene clones. The 1.7 kb XbaI VH-containing fragment and the 2.7 kb HindIIIXbaI VLl-containing fragment were cloned into Ml3 and sequenced. Figure 2A shows the DNA sequence and its corresponding translated protein for the 26-IOVH region. The cloned V gene sequence agrees with the complete amino acid sequence of the mature protein (Huston ef al., 1988) and with the partial mRNA sequence (D. Panka, J. Novotny and M. N. Margolies, unpublished results). The 26-IOVH gene is a member of the J558VH gene family. the largest mouse VH gene family (Brodeur et al., 1988), and contains JH4. On the basis of the known J558VH sequences (Kabat et al., 1987) and the known germline JH4 sequence (Kabat et al., 1987; R. I. Near, unpublished A/J mouse JH sequences), we hypothesize that the D region encodes 6 amino acids, SSGNKW. The gene also contains the appropriate octamer, ATGCAAAT, near the TATA box that affects tissue-specific regulation (Parslow et al., 1984). Figure 2B shows the DNA sequence and the translated protein corresponding to the 26-1OVL region. The sequence shows that the clone contains the correct gene (Novotny and Margolies, 1983). The 26-IOVL protein is a member of the VK 1 group (Potter et al., 1982). From the DNA sequences of other Vti 1 antibodies and the Jx I sequences (Kabat et al., 1987), we hypothesize that the position of the VK-Jti junction is at amino acid IO1 (indicated in Fig. 2B). The gene also contains the immunoglobulin regulatory octamer, TTTTGCAT, upstream from the TATA box (Parslow it al., 1984). Figure 2B also shows the eight positions at which 26IOVL differs from the homologous light chain, 40-IOOVL (N. W. Hudson and M. N. Margolies, manuscript in preparation). In the recombinant antibody studied here, 26-IOVL is substituted for 40-IOOVL. Only three of

Anti-digoxin

transfectoma

antibody

binding

properties

905

100 200

300 10 CAACMCCGGX-2 QQSGFELVKPGASVRNSCKSSGYIFTDFYNNWVR

l-f-xc-TAcAm3ov

40 50 GGCAGRU.TATGWVIPrGAG crrGATl!AcA~TATAlTTcrGcTrl4 QSRGKSLDYIGYISPYSGVTGYNQKPKGKATLT

400

C

-7ok

80 90 n;TAGRUUIGTCCfCCAGCA TGG%CTCCGCAGCCTZACATCG%GGA~ VDXSSSTAYRELRSLTSEDSAVYYCAGSSGNKW

500

100 TfACn;rrUUiGRnr;ra;OGC4A?IARGn;C

600

I**** D REGICW ***I 110 GcpATujRcpAcn"JGGorc ANDYWGRGASVTVSS/ I***************** m4

-m=rc-crrETrAm

700

******t***tt***tt**]

(Bt 100

200

400 500 600

lTlcwrm.ITADC -mcCcAAAcrccAmcmm~ /SSSDVVRTQTPLSLPVSLGDQASISC

10

700 TCAAGCSCAF

30 A~~A~~A~~~~TAC~~~~~ RSSQSLVHSNGNTYLNWYLQKAGQSPKLLIYKVS 70 cwcmbcNRFSGVPDRPSGSGSGTDFTLKISRVSARVLGX

50

40

800 m

D 90

80

--

-lTNX-~-T

900

r 100 ~A~ffA~~~~~C~~~~~~ YFCSQTTHVPPTFGGGTKLEIR~ G

V

110 1000

L

I (*t**t***t****JR1 *t****f**t**,*+*j

ATGTC'l'AAAAA~TGTATAMA'X'R!A~ 1055 GTCTATCTCTGTCtCTTCI' Fig. 2. DNA sequences and corresponding translated protiens of the 26-10 rearranged heavy and ltght chain V regions. The leader sequences. intervening sequences and RNA splice sites are shown. The numbering is sequential from the start of the mature protein without interruption. The CDR (complementarity determining regions), TATA boxes, and immunoglobulin-specific octamers (Parslow et al., 1984) are underlined. (A) The sequence of 26-IOVH as determined by M 13-dideoxy sequencing (~~~e~~~~~ and Me/hods). The location of the D region was hypothesized by comparison with other J558VH sequences and the germhne JH4 sequence. (B) The sequence of 26lOV% as determined by Ml3-dideoxy sequencing. The location of the junction between Jx I and Vx26-10 was estimated by comparison to the JK 1 germline sequence and other Vx 1 genes. The single amino acids listed below the translated 26-IOVL sequences are the residues present in 40-IOOVL that differ from those in 26-IOVL. The complete sequences of antibody 40-100 will be published elsewhere (N. W. Hudson and M. N. Margolies).

these eight positions (amino acids 39, 96 and 101) contain nonconservative differences located within CDR regions. T~~~~~~~r~~~ of’ 26 IO E/H rrnrl 26-10 VL into J558L

If the expression vectors containing the H and L chain 26-10 genes are correctly constructed, trans-

fected carrying both vectors should express an antibody that shows properties identical to those of the original 26-10 protein. Figures 3A and 3B show the expression vectors for 26-1OVH and 26-lOVL, respectively. These vectors are the 26-10 VH and VL genes cloned into plasmids that contain the Ecogpt or neo drug selection gene and, for 26+10VH, the

RICHARD 1. NEAR er al.

906

genomic EcoRI 11.5kb fragment was with a polylinker at the EcoRI site (R. I. Near, unpublished) and by the insertion of a 5.5 kb XbaI fragment containing the germline murine 72b constant region gene (Marcu ef al., 1980) into the XbaI site of the polylinker. (B) The 26-1OVL BamHI 1 I .2 kb genomic fragment which contains both the V and CK regions (Fig. 1B) was cloned into pBR322 along with a 2.6 kb BamHI fragment that contains the Tn5 neomycin (neo) resistance gene (Southern and Berg, 1982). B, BamHI; EcoRI; X, XbaI. The figures are drawn to scale. Fig. 3. H and L chain 26-10 expression

vectors.

(A) The 26-IOVH

cloned into pSVgpt (Mulligan and Berg, 1980) that had been modified

Table Antibody:

I. Concentration Digoxin

40- IO0 26-10 H8* I E6,

1.0

I .o 1.0 I.0

of inhibitor Digoxigenin 8.8 1.4 1.7 5.1

relative

to digoxin’

yielding

50% inhibition

of [‘2’]digoxin-BSA

binding

Digitoxin

Digitoxigenin

Gitoxin

Acetylstrophanthidin

Ouabain

480 1.1 1.5 3.2

2600 1.8 I.3 I2

94 5.9 4.2 5.6

4900 1.6 I.1 35

7000 39 42 27

“The concentration of digoxin at 50% inhibition was set at 1.0. All experiments were done in duplicate and the values hsted are the average of two separate experiments. hH8 is a transfectoma expressing H and L 26-10 constructs (see Text). ’ IE6 IS a transfectoma expressing the L chain of the pneo/ZblOVL construct with natural 40-100 H chain.

genomic constant region of Cy2b (the 26-1OVL gene is on a DNA fragment that contains the genomic CK region). J558L was transfected, sequentially, with pSVgpt/26-IOVH followed by pneo/26-1OVL. The resulting transfectoma line, H8, produces an antibody whose y2b isotype is derived from the pSVgpt/26- IOVH vector. Northern analysis of RNA from H8 with H and L chain hybridization probes (Materials und Methods) showed appropriate immunoglobulin mRNA bands of 1.9 and 1.3 kb, respectively. that are not present in J558L RNA (not shown). The fine specificity of the H8 transfectoma antibody is indistinguishable from that of the 26-10 monoclonal antibody in a competition radioimmunoassay (Table I). Furthermore, the affinity constant of the H8 antibody was found to be within experimental error of that of 26-10 (Table 2). The 21 light chain present in J558L has no obvious effect on these binding assays. We conclude that the pSVgpt/26IOVH and pneo/26- JOVL vectors correctly express an antibody with functional characteristics indistinguishable from those of 26-10 (but with the y2b isotype). Chain recombination erpressiorz rector

by trumfection

of the 26-IOL

Although the VL region of antibody 40-100 is 92% homologous with that of 26-10 (N. W. Hudson

and M. N. Margohes, manuscript in preparation; Fig. 2B), the VH regions of the two antibodies are from entirely different gene families (Mudgett-Hunter et al., 1985). Antibodies 40-100 and 26-10 also have easily distinguishable specificity patterns (MudgettHunter et al., 1985). To assess the contribution of light chain to the different binding properties of 40-100 and 26-10, a recombinant antibody was generated after transfection of the 26-10 L chain gene into the cell line CR-101 (N. W. Hudson, unpublished) that contains only the 40-100 H chain and an irrelevant 11L chain (Muteriuk and Methods). The antibody produced by CR-101 itself does not bind digoxin. The plasmid pneo/26-IOVL was transfected into CR-101 by electroporation followed by G418 drug selection. Positive transfectants were confirmed by Northern analysis with a light chain probe (Materials and Methods) that detected a 1.3 kb

Table 2. Average intrinsic affinity constants” (KO) Antibody 26-10 H8 IE6 40- IO0 “KO

Affinity Constant (3.1 (3.6 (5.5 (7.4

f i * Ir

0.4) 0.5) 1.1) 1.6)

X x x x

(Mm’) IO9 IO’ IO’ IO”

is calculated according to the Algorithm LIGAND (Materials and Methods).

Anti-digoxin transfectoma

antibody binding properties

mRNA in the transfectants but not in CR-101 (not shown). One transfectant clone, lE6, was chosen for further studies.

Con~rm~tion of the structure of 1E6 by PCR In order to confirm that the appropriate light and heavy chains were expressed in 1E6 cells, mRNA was sequenced using the PCR method (Scharf et al., 1986). H and L chain constant region primers were used to synthesize cDNAs from lE6 RNA. The V regions of the cDNAs were amplified by PCR and cloned into Ml3 phage (Material.7 and Methods). The resulting sequences showed that the 1E6 transfectoma expressed mRNAs that were identical in sequence to the cloned 26- 1OVL gene (Fig. 28) and the 40-IOOVH gene contributed by the CR-lOl-transfected cell line (M. N. Margolies and N. W. Hudson, unpublished sequence).

‘Oar B so$

60-

B 2

40-

% 8

20 0

Antigen binding properties of 1E6 Fine specificity and antigen affinity were measured for the recombinant antibody lE6 and compared with those of the parental hybridoma proteins 26-10 and 40- 100 (Fig. 4 and Tables I and 2). The specificity profile of lE6 antibody shows some properties of both 26-10 and 40-100. Antibody 1E6 is sensitive to the presence of the digitoxose groups of digoxin or digitoxin, binding 4-5 times more poorly to the aglycones digoxigenin and digitoxigenin (Table 1). Antibody 40-100 also recognizes the sugar moieties, whereas 26-10 does not. Antibody 1E6, like 40- 100, does not distinguish acetylstrophanthidin from ouabain, both of which are highly substituted on the steroid A and B rings. Antibody 26-10, however, binds acetylstrophanthidin approximately 20-fold more strongly than ouabain. Antibody lE6 resembles 26-10 in that the presence of a I2-hydroxyl group on the steroid C ring that differentiates digoxin from digitoxin does not significantly alter ligand binding, whereas the absence of the 12-hydroxyl group diminishes binding by 480 fold in the case of 40- 100. Substitutions on the /J surface of steroid rings A and B characteristic of both acetylstrophanthidin and ouabain greatly inhibit binding to 40-100 but not to 26-10 or lE6 (Table 1; Mud’gett-Hunter et al., 1982. 1985; Hudson et al., 1987). Further, IE6, like 26-10, exhibits a relatively small range of binding affinities for the various compounds tested, whereas the range of affinities exhibited by 40-100 is about 7000 fold (Table 1). A comparison of panels A, B and C in Fig. 4 graphically demonstrates this difference. The affinity of 1E6 is approximately two orders of magnitude lower than that of 26-10 or 40-100 (Table 2). The diminution of affinity in lE6 may be related to its smaller range of binding a~nities. DISCUSSION

The present highly developed facility in the manipulation and expression of antibody genes

Fig. 4. Fine specificity profiles of hybridoma and transfectoma-derived anti-digoxin antibodies. Inhibition curves were generated by an RIA in which digoxin analogs compete with [‘251]digoxin for antibody binding. The antibodies assayed are (A) 40-100; (B) 26 10; and (C) 1% (40-100 H chain and 26-10 L chain). Cardiac glycoside competitors are digoxin (O-O), digoxigenin (A-A), digitoxin (+-O), digitoxigenin (El--_13), gitoxin (O--O), acetylstrophanthidin (+-•), and ouabain (A-A).

(Jones et al., 1986; Roberts and Rees, 1986; Morrison and Oi, 1988; Sharon et al., 1989) permits the ready application of these t~hniques to immunoglobulin chain recombination experiments. The gene of interest is cloned, inserted into an expression vector, and then transfected into a cell line that contains the gene for the complementary immunoglobulin chain (Jones et al., 1986; Roberts and Rees, 1986; Morrison and Oi, 1988; Sharon et ai., 1989; Wallick et a!., 1989). Other investigators have used transfection methods to study immunoglobulin mutants (Roberts and Rees, 1986; Sharon et al., 1986, 1989), antibody combining sites (Ochi et al., 1983; Morrison et al., 1984; Jones et al., 1986; Watanabe ef al., 1986), and the recombining of antibody chains (Gillies et al., 1983; Oi et al.. 1983; Wallick et al., 1989). In the present study, we used transfection to produce a specific immunoglobulin chain recombinant designed to examine the relative contributions to antigen binding by H and L chains derived from

RICHARD 1. NEAR et al.

908

antibodies that have sequence homology. This approach allows the later use of site-directed mutagenesis to test hypotheses generated from studies of the antigen binding properties of the recombined antibodies. ln the experiments described here, we have readily validated the sequence of the mRNA transcribed from the transfected genes for both immunoglobulin chains of the recombinant molecule by using PCR (Scharf ef ul.. 1986). The PCR results provide mRNA sequence data that definitively confirm correct expression of the appropriate gene. Furthermore, PCR technology has now been developed that allows rapid cloning and expression of immunoglobulin cDNAs (Orlandi et al., 1989). The application of this recent PCR method would avoid the cloning of the genomic immunoglobulin genes, which was required when the present study was initiated. Antibodies 26-10, 40-20, 40-60, 40-90 and 40-100 have VL regions that are more than 90% homologous (Mudgett-Hunter ef al., 1985; Margolies and Hudson, unpublished). The VH regions of 40-20, 40-60, 40-90 and 40-100 are also highly homologous and belong to the same VH family, while 26-10 possesses a VH region from a different family. Antibody 26-10 shows a fine specificity pattern quite different from that of the other four antibodies. The chain recombinant transfectoma described here, lE6, contains the H chain of 40-100 and the L chain of 26-10 (highly homologous to the 40-100 L chain) and shares specificity properties present in both parent antibodies. For example, lE6 resembles 40-100 in its ability to recognize the glycoside residues of the ligand, as well as its inability to distinguish acetylstrophanthidin from ouabain. lE6 resembles 26-10 in its insensitivity to the 12-hydroxyl substitution and its narrow range of binding affinities. Furthermore. the ability of iE6 and 26-10 to bind well to acetylstrophanthidin and ouabain suggests that they recognize the c( surface of the steroid A and B rings. However, the combining site of antibody 40-100, which binds poorly

to these

analogs.

/j surface (Mudgett-Hunter

probably

has contacts

at the

er ul., 1985; Hudson et al., 1987). Thus, IE6 has binding specificity properties derived from both parental immunoglobulin chain donors. Antibody lE6 has a binding affinity more than 2 orders of magnitude lower than that of either parental antibody (Table 2) and has lost much of the ability to distinguish digoxin and its analogs from each other. Thus, the homologous combination of L and H chains is needed to construct an antibody with optimal antigen affinity and with the ability to recognize structural differences among related antigens. 26-IOVL differs from 40-1OOVL at eight positions (Fig. 2B), only three of which are in CDR regions. Thus it is likely that the amino acid differences between 26-IOVL and 40-IOOVL at positions 39. 96 and 101 contribute significantly to

the differences between 1E6 and 40-100 in both affinity and specificity characteristics. A recombinant hybridoma (CR-57) that contains the 26-10 H chain and 40-20 L chain has been generated by the fusion of two hybridomas (Hudson et al., 1987). Since the 40-20 H and L chains are highly homologous with those of 40-100, recombinant CR-57 may be thought of as approximating the reciprocal of 1E6 (40- 100 H chain and 26- 10 L chain). Six of the eight amino acid sequence differences between 40-1OOVL and 26-1OVL are also present in 40-20VL. In addition, 40-IOOVL and 26-1OVL differ in positions 22 and 101. In CR-57, the H chain donor (26-IOVH) is the dominant contributor to the fine specificity profile, but the loss of binding affinity for digoxin is similar to that observed for lE6. One possibility (aside from the obvious possibility resulting from different H chains used in lE6 and CR-57) is that the differences between 40-20VL and 401OOVL at positions 22 and 101 affect the ability of the H chain to be the dominating contributor to the fine specificity properties. The positions suggested from the experiments in the present study (the amino acid differences between 26-IOVL and 40-IOOVL in the CDR segments at positions 39, 96 and 101) and the positions suggested from explorations of chain recombinants among this group of anti-digoxin antibodies (Hudson et al., 1987; Hudson et nl., manuscript in preparation) will be studied by in vitro mutagenesis. The results should yield substantial information about how immunoglobulin H and L chains contribute to ligand recognition. Acknowledgements-We thank Boutrous Bouyounes, Abbie White, Eric lida and Jeff Schuhze for their technical assistance.

REFERENCES Alt F. W., Blackwell T. K. and Yancopoulos G. D. (1987) Development of the primary antibody repertoire. Science 238, 1079-1087. Brodeur P. H., Osman G. E., Mackle J. J. and Lalor T. M. (1988) The organization of the mouse Igh-V locus: Dispersion, interspersion, and the evolution of VH gene family clusters. J. e-up. Med. 168, 2261-2278. Clarke S. H., Huppi K., Ruezinsky D., Staudt L., Gerhard W. and Weigert M. (1985) Inter- and intraclonal diversity in the antibody response to influenza hemagglutinin. J. e.yp. Med. 161, 687-704. Fieser L. G. and Fieser M. (Editors) (1959) Steroids, p. 727. Reinhold Publishing Corporation, New York. Gillies S. D., Morrison S. L., Oi V. T. and Tonegawa S. (1983) A tissue-specific transcription enhancer element is located in the maior intron of a rearranged immunoglobulin heavy chain gene. Cell 33, 717-728. Go K.. Kartha G. and Chen J. P. (1980) The structure of digoxin. Actu Cryxt. B36, 1811~1816. Haber E and Margolies M. N. (1984) Combining site specificity and idiotypy: a study of anti-digoxin and anti-arsonate antibodies. In Thi Biology of Idiotypes (Edited by Green H. and Nisonoff A.). pp. 14lL170. Plenum, New York.

Anti-digoxin

transfectoma

Hudson N. W., Mudgett-Hunter M., Panka D. J. and Margohes M. N. (1987) lmmunoglobulin chain recombination among antidigoxin antibodies by hybridomaa hybridoma fusion. J. Immunol. 139, 2715-2723. Huston J. S., Levinson D., Mudgett-Hunter M., Tai M., Novotny J., Margohes M. N., Ridge R. J., Bruccoleri R. E., Haber E., Crea R. and Oppermann H. (1988) Protein engineering of antibody binding sites: Recovery of specific activity in an anti-digoxin single-chain Fv analogue produced in Escherichia cob. Pror. Natl. Acad. Sci.

USA 85, 5879F5883.

Jones P. T., Dear P. H.. Foote J., Neuberger M. S. and Winter G. (1986) Replacing the complementaritydetermining regions in a human antibody with those from a mouse. Nature 321, 522-525. Kabat E. A. (1966) The nature of an antigenic determinant. J. Immunol.

97,

Ibl

I.

Kabat E. A., Wu T. T., Reid-Miller M., Perry H. M. and Gottesman K. S. (1987) Sequences qfProreins qflmmunological Interest (4th Edition). U.S. Department of Health and Human Services, Public Health Service, National Institutes of Health. Bethesda. Kohler G. (1986) Derivation and diversification of monoclonal antibodies. Science 233, 1281-1286. Maniatis T.. Fritsch E. F. and Sambrook J. (1982) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory. Cold Spring Harbor. N.Y. Marcu K. B.. Arnheim N.. Banerji J., Penncavage N. A.. Seperack P.. Lang R.. Miesfeld R., Harris L. and Greenberg R. (1980) Studies on the nature of germ-line stability of DNA sequences flanking the mouse immunoglobulin heavy-chain constant-region genes. Cold Spring Hat-h. Syn7p. quanr. Rio/. 45, 899991 I. Marcu K. B.. Banerji J., Penncavage N. A., Lang R. and Arnheim N. (1980) 5’ flanking region of immunoglobuhn heavy chain constant region genes displays length heterogeneity in germlines of inbred mouse strains. Cell 22, 187-196. Messing J. (1983) New Ml3 vectors for cloning. Meth. Ettqw7. 101, 20&78. Milstein C. (1986) From antibody structure to immunological diversification of immune response. Science 231, 1261-1268. Morrison S. L. and Oi V. T. (1988) Genetically engineered antibody molecules. Adr. In7munal. 44, 65593. Morrison S. L., Johnson M. J.. Herzenberg L. A. and Oi V. T. (1984) Chimeric human antibody molecules: Mouse antigen-binding domains with human constant region domains. Proc. t7clm. Acad. Sri. USA 81, 685 l--6855.

Moynet D.. MacLean S. J., Ng K. H.. Anctil D. and Gibson D. M. (1985) Polymorphism of ti variable region (Vk-I) genes in inbred mice: relationship to the Igh--Ef2 serum light chain marker. J. Immunol. 135. 7277732. Mudgett-Hunter M.. Anderson W.. Haber E. and Margolies M. N. (1985) Binding and structural diversity among high-affinity monoclonal anti-digoxin antibodies. Molec. Imn~uti. 22, 477-488. Mudgett-Hunter M.. Margolies M. N.. Ju A. and Haber E. (1982) High-atlinity monoclonal antibodies to the cardiac glycoside.-digoxin: J. Immunol. 129, 1165-l 172. Muhiean R. C. and Berg P. (1980) Exoression of a bacterial gene in mammalian dells. Science iO9, 1422-1427. Munson P. J. and Rodbard D. (1980) LIGAND: A versatile computerized approach for characterization of ligandbinding systems. Anal. Biochem. 107, 220-227. Near R. 1. and Haber E. (1989) Characterization of the heavy and light chain immunoglobulin variable region

antibody

binding

properties

909

genes

used in a set of anti-digoxin antibodies. Molec. 26, 371-382. Novotny J. and Margolies M. N. (1983) Amino acid sequence of the light chain variable region from a mouse anti-digoxin hybridoma antibody. Biochemistry 22, Immunol.

1153~1158.

Ochi A., Hawley R. G., Shulman M. J. and Hozumi N. (1983) Transfer of a cloned immunoglobuhn light-chain gene to mutant hybridoma cells restores specific antibody production. Narure 302, 340-342. Oi V. T., Morrison S. L., Herzenberg L. A. and Berg P. (1983) Immunoglobulin gene expression in transformed lymphoid cells. Proc. natn. Acad. Sri. USA 80, 825-829. Orlandi R., Giissow D. H., Jones P. T. and Winter G. (1989) Cloning immunoglobulin variable domains for expression by the polymerase chain reaction. Proc. natn. Acad. Sri. USA 86, 3833-3837.

Panka D. J. and Margolies M. N. (1987) Complete variable region sequences of five homologous high affinity antidigoxin antibodies. J. Immunol. 139, 2385-2391. Parslow T. G., Blair D. L.. Murphy W. J. and Granner D. K. (1984) Structure of the 5’ ends of immunoglobulin genes: a novel conserved sequence. Proc. natn. Acad. Sri. USA 81, 2650-2654.

Potter M., Newell J. B., Rudikoff S. and Haber E. (1982) Classification of mouse VK groups based on the partial amino acid sequence to the first invariant tryptophan: impact of 14 new sequences from IgG myeloma proteins, Molec. Immunol. 19, 1619-1630. Potter H., Weir L. and Leder P. (1984) Enhancer-dependent expression of human k immunoglobulin genes introduced into mouse pre-B lymphocytes by electroporation. Proc. natn. Acad.

Sri.

USA 81, 7161-7165.

Roberts S. and Rees A. R. (1986) The cloning and expression of an anti-peptide antibody: a system for rapid analysis of the binding properties of engineered antibodies Protein Engng 1, 59-65. Sanger F., Nicklen S. and Coulson A. R. (1977) DNA sequencing with chain-terminating inhibitors. Proc. natn. Acad.

Sci.

USA 74, 5463-5467.

Scharf S. J.. Horn G. T. and Erhch H. A. (1986) Direct cloning and sequence analysis of enzymatically amplified genomic sequences. Science 233, 107&1078. Sharon J., Gefter M. L., Manser T. and Ptashne M. (1986) Site-directed mutagenesis of an invariant amino acid residue at the variable-diversity segments junction of an antibody. Proc. natn. Acad. Sci. USA 83, 2628-2631. Sharon J., Gefter M. L., Wysocki L. J. and Margolies M. N. (1989) Recurrent somatic mutations in mouse antibodies to p-azophenylarsonate increase affinity for hapten. J. Immunol.

142, 596400.

Southern P. J. and Berg P. (1982) Transformation of mammalian cells to antibiotic resistance with a bacterial gene under control of the SV40 early region promoter. J. Mol. Appl. Genet. 1, 327-341. Tucker P. W.. Marcu K. B., Newell N., Richards J. and Blattner F. R. (1979) Sequence of the cloned gene for the constant region of murine y2b immunoglobulin heavy chain. Science 206, 1303-l 306. Wallick S. C., Kabat E. A. and Morrison S. L. (1989) The effect of isotype and the J, region on antigen binding and idiotype expression by antibodies binding r( l-6) dextran. J. Immunol. 142, 123551244. Watanabe M., Wegmann D. R., Ochi A. and Hozumi N. (1986) Antigen presentation by a B-cell line transfected with cloned immunoglobulin heavy- and light-chain genes specific for a defined hapten. Proc. natn. Acad. Sri. USA 83, 5247-5251.