Volume 9, Number Mary Ann Leibert,

5, 1990 Inc., Publishers

Specific Amplification of Rearranged Immunoglobulin Variable Region Genes from Mouse Hybridoma Cells JORGE V. GAVILONDO-COWLEY,' MARIA J. COLOMA,2 JAVIER VAZQUEZ,' MARTA AYALA,' AMPARO KIRK E. FRY,2 and JAMES W. LARRICK2

MACÍAS,3

of Hybridomas and Animal Models,

Center for Genetic Engineering and Biotechnology, P.O. Box 6162, La Habana-6, Cuba 2Genelabs Incorporated, 505 Penobscot Drive, Redwood City, CA 94063 3Department of Biotechnology, Institute of Oncology and Radiobiology, Ey 29, Vedado, La Habana-4, Cuba

'Division

ABSTRACT In this article we show how the polymerase chain reaction (PCR) and primers designed for conserved sequences of leader (L), framework one (FR1) and constant (CONST) regions of immunoglobulin light and heavy chain genes can be used for the cloning and sequencing of rearranged antibody variable regions from mouse hybridoma cells. RNA was extracted from the mouse hybridoma cells secreting MAbs: IOR-T3a (anti-CD3), C6 (anti-P1 of N. meningitidis B385), IOR-T1 (anti-CD6), CB-CEA.1 (anticarcinoembryonic antigen), and CB-Fib.1 (anti-human fibrin). First strand cDNA was synthesized and amplified using PCR. The newly designed primers are superior to others reported recently in the literature. Isolated PCR DNA fragments of C6 and IORT3a were sequenced after asymmetric amplification, or M13 cloning. The FR1/CONST primer combinations selectively amplified mouse lights chain of groups kappa II, V, and VI, and heavy chains of groups Ha and He. The L/CONST primers for light chains amplified light chains from all four hybridomas. These methods greatly facilitate structural and functional studies of antibodies by reducing the efforts required to clone and sequence their variable regions.

INTRODUCTION The use of monoclonal antibodies (MAbs) in human therapeutics (1, 2) is hindered because: (a) murine MAbs being used or tested jn vivo induce a human antimouse antibody response (HAMA) that can limit the effectiveness and safety of these reagents (3, 4) and (b) human MAbs (HuMAbs) are difficult to obtain with an adequate combination of specificity and effector function. This is despite recent developments using improved human hybridoma fusion partners and advances in jn vitro immunization

techniques (5).

407

Recombinant DNA technology can now be used to engineer antibodies with a number of defined characteristics such as isotype, size, domain structure, carbohydrate addition sites, etc. Hence efforts have been made to construct chimeric mouse/human MAbs in the hope that replacement of all but the murine variable regions or complementarity determining regions (CDRs) with human sequences will reduce the HAMA response (6, 7). Several recûmbinant antibodies retain the binding capacity and affinity of the murine MAb both in vitro and in vivo (8, 9, 10, 11, 12, 13), and a recent clinical trial showed that a 'reshaped" rat anti-human T cell MAb was active in vivo against lymphoma cells, and contrary to the parental murine MAb, elicited no antiantibody response in the patients (14). Despite these advances, determination of the variable region sequences has been a limiting step in the rapid construction of recombinant antibody molecules. Recently, Larrick et al. (15) and Orlandi et al. (16) have used the polymerase chain reaction (PCR) (see review in Oste 1988, 17) for the rapid direct cloning and sequencing of human and mouse immunoglobulin variable region genes, respectively. Other laboratories have adapted these strategies for the construction of recombinant libraries in phage lambda (18) and expression in E. coli (19, 20, 21). In this paper we present a general method to clone and sequence the rearranged variable region light and heavy chain genes from mouse hybridoma cells, and discuss the importance of the PCR primer design for cloning and elucidation of antibody sequences. MATERIAL AND METHODS

Cell Lines: The following mouse hybridoma cell lines were used in the experiments: IORT3a (anti-CD3; lgG2a kappa; obtained from the Institute of Oncology and Radiobiology of Havana), C6 (anti-P1 outer membrane protein of N. meningitidis, strain B385; lgG2a kappa; obtained from the National Center for the Antimeningococcal Vaccine of Havana),

(anti-carcinoembryonic antigen; lgG1 kappa; 22, IOR-T1 (anti-CD6; lgG2a kappa; 23, CB-Fib.1 (anti-human fibrin; lgG1; obtained from the Center for Genetic Engineering and Biotechnology of Havana).

CB-CEA.1

All cells were grown in Iscove's modified Dulbecco's medium with 10% fetal bovine serum supplemented with 2mM glutamine, 50 fjM 2-mercaptoethanol, 0.48 mM sodium pyruvate, 0.17 uM bovine insulin and 1.3 mM cis-oxaloacetic acid (GIBCO, Grand

Island, NY).

Oligonucleotide primer design: To test a general method for cloning the variable regions

have designed oligonucleotide PCR primers using the database of Kabat et al.(24), and sequences available from GenBank The 5' primers were constructed from information available on the conserved sequences of the leader (light chains) or first framework (FR1) regions (light and heavy chains) (see designations, "EcoRI/L." or "EcoRI/FR1..." in Table 1a). The 3' primers were designed for annealing within the constant regions of the light mouse (C kappa) chains, or heavy (CH1 gamma) chains, (see designations, "CONST..." in Table 1a). The 5' end primers contained Eco R1 sites and the 3' end primers contained Hind III sites. Two extra bases (Gs) were added outside the restriction site to improve enzyme digestion and avoid exonuclease activity of Taq polymerase. The primer series VK1 BACK/FOR and VH1 BACK/FOR of Orlandi et al. (16) for FR1 and framework four (FR4) of light and heavy variable regions were also used (Table 1b). Oligonucleotides were made on an Applied Biosystems 380B DNA synthesizer (Foster City, CA), or a Gene Assembler Plus DNA synthesizer (Pharmacia-LKB, Uppsala). we

.

408

TABLE 1 : DESIGN OF THE 5'N TERMINUS AND 3* END SYNTHETIC PRIMERS FOR PCR a)

Now

primer design.-

MOUSE UGHr KAPPA CHAINS: 5':

EcoRl/FR1-ML(kappa) (amino adds 1-8):

5'-GGGAATTCGA(CT)ATTGTG(AC)T(AG)AC(AC)CA(AG) (GT) (AC)TCAA-3' or

5':

EcoRI/Lead-ML(kappa) (loader poptide):

5'-GGGMTTCATGGGC(AT)TCM(GA)ATG(GA)A(GAUAT)C(AT)CXr--3'

Hindlll/ML(kappa) CONST.: S'-GGAAGÇTTACTGGATGGTGGGAAGATGGA—3'

3':

MOUSE HEAVY GAMMA CHAINS: 5': EcoFÜ/FRI-MH

(amino acids 1-7):

5'-GGGAAJTÇ(GC)AGGT(CGKAC)A(AG)CTGCAG(CG)AGTCT_3' 3':

Hindlll/MH(gamma)-CONST.:

5'-GGMGCTTArtC)CTCCACACA(^GG(AGHAG)CC^GTGGATAGAC---3'

b) Oligonucleotide primers of Orlandi et al. (1989).MOUSE LIGHT KAPPA CHAINS:

5,:VK1BACK(FR1)

:

5'-GACATTCAGCTGACCCAGTCTCCA-—3'

3':VK1FOR(FR4)

:

S'-GTTAGATCTCCAGCTTGGTCCC—3' MOUSE HEAVY GAMMA CHAINS:

5':VH1BACK(FR1) : 5•-AGGT(CG)(AC)A(AG)CTGCAG(CG)AGTC(Ä^)GG-3•

3':VH1FOR(FR4)

:

5'-TGAGGAGACGGTGACCGTGGTCCCTTGGCCCCAG—3'

Notes: Bases in parentheses are substitutions at a given position; i.e. (AT) A and T were present in equimolar amounts during the synthesis at that position. EcoRI (5' end) and Hindlll (3' end) sites are underlined. =

No purification was of priming sites.

performed prior to

use.

See

Figure 1 for a graphic representation

Preparation of RNA: RNA was extracted either with a microadaptation of the guanidinium thiocyanate/cesium chloride (GuSCN/CsCI) procedure (15) or the NP-40/SDS technique suggested by Gough (25); 105 hybridoma cells were used as starting material. First strand snthesis: First strand DNA synthesis was performed using the BoehringerMannheim (Chicago, IL) cDNA kit; briefly, total RNA samples (approximately 0.5 /-g; derived from 105 hybridoma cells) were heated at 65CC for 1-5 minutes, and incubated with a mixture of RNase inhibitor, deoxynucleotides, oligo (dT);5 as primer, and AMV reverse transcriptase, for 60 minutes at 42°C.

409

LEADER

VARIABLE

CONSTANT

FR1 CDR1 FR2 CDR2 FR3 CDR3 ^-llllllll.I.Ill llllllllll

FR4

A) ECORI/LEAD-ML

_^___ HINDI I I /M(L

Sj ECOR1/FR1-M(L

or

or

H)CONST

H)

B)

V(K

or

H)l BACK

¥(K

or

H) I FORWARD

FIGURE 1: STRATEGIES FOR PCR CLONING OF MOUSE REARRANGED IMMUNOGLOBULIN VARIABLE LIGHT AND HEAVY CHAIN REGIONS. A: Priming sites for the EcoRI/FR1 or Leader, and Hindlll/CONST oligonucleotides; B: Priming sHesforthe FR1 (V[KH]1BACK) and FR4 (V([KH]1FOR) primers.

Polymerase chain reaction: Eighty fl of PCR mix was added to the 10 u\ of first strand cDNA. The PCR mix was made following the instructions of the Perkin-Elmer Cetus (Norwalk, CT) PCR kit. Five ß of each primer was added to give a final primer concentration of 1 /vM and the mixture was subjected to PCR amplification using the Perkin-Elmer or Hybaid (UK) thermal cycler sets, for 30 cycles. The temperatures and times used for PCR were: melting at 94°C, 1 minute; primer annealing at 55°C, 1 minute; primer extension at 72°C, 1 minute. Normally one minute ramp times were used between these temperatures. Ethidium bromide stained 2% agarose (NuSieve) gels were used to vizualize PCR fragments.

Asymmetric Amplification: Twenty jul of amplified cDNA were run in low melting temperature agarose (NuSieve) gels. The relevant bands were isolated under long wave UV illumination, and melted at 65°C. Ten u\ samples were asymmetrically amplified, using the conditions above mentioned, and the final primer concentrations: 0.2 uU of the 5' end "FR1"

and 0.01 uM of the 3' end "CONST" or "FR4" primers. After and chloroform extractions, the samples were spin-dialyzed in phenol/chloroform Centricon-30 microconcentrators (Amicon, Danvers MA), as suggested (26). The precipitated DNA was then subjected to sequencing reactions, using 0.5 /iM of the 3' end "CONST primer for the annealing reaction.

primer,

M13 cloning: Gel purified PCR products (NA45 paper, Schleicher & Schuell, Keene, NH) digested with EcoRI and HindiII restriction enzymes and ligated into M13mp18/19

were

sequencing vectors (27,28). DNA Sequencing: Dideoxynucleotide chain termination sequencing was carried out using the Sequenase 2.0 kit from United States Biochemical Corp. (Cleveland, OH) according to the manufacturer's protocol, with 35S-aATP (Amersham). Asymmetric amplifications of two independent PCR samples and/or at least two independent M13 clones were used to validate each sequence, the latter generally in opposite orientations.

RESULTS

Figure 2 shows the major PCR amplification products for the IOR-T3a (lanes 3-6) and C6 (lanes 7-10) light and heavy chain variable regions, using the VK1/VH1 (lanes 3, 4, 7 and 8) and FR1/CONST (lanes 5, 6, 9, and 10) primer combinations. For the IOR410

12345 6 789 10

FIGURE 2. IMMUNOGLOBULIN LIGHT AND HEAVY CHAIN PCR PRODUCTS DERIVED FROM FIRST-STRAND cDNA. Ten mioroliters from a 100 microliter PCR reaction were applied to a 2% agarose gel for electrophoresis in TBE buffer. The major PCR products obtained after ethidium bromide staining are shown. LANE 1:1 kb marker ladder; LANE 2: Perkin Elmer Cetus kit control (500 bp); cDNA IOR-T3a: LANES 3-6, C6: LANES 710; Primers LANES 3,7: VK1 BACK and VK1 FOR, LANES 4,8: VH1 BACK and VH1 FOR, UNES 5,9: EcoRI/FR1 ML(kappa) and Hindlll/ML(kappa)-CONST., LANES 6,10: EcoRI/FR1-MH and Hindlll/MH(gamma)-CONST. =

»

-

T3a light and heavy chains, and C6 heavy chain, the shift in size seen when comparing lanes 3 and 5 (330 versus 390 bp), lanes 4 and 6 (370 versus 420 bp), and lanes 8 and 10 (340 versus 400 bp), can be explained by the differences in 3' end priming (FR4 gives a shorter fragment than the constant region), and the presence of attached restriction sites in our oligonucleotides. This is not the case for the C6 light chain, where the combination of VK1BACKA/K1 FOR primers gave a predominant band of only some 220 bp (lane 7), while a 390 bp (lane 9) cDNA fragment was amplified with the FR1/CONST primer combination. The two bands were isolated in low-melting agarose, asymmetrically amplified, and sequenced. The 220 bp fragment base structure had a similar sequence to the larger C6 light chain fragment, except that: (a) the bases coding for the first 36 amino acids were absent and, (b) the new N terminus was identical to the VK1 BACK primer. This result indicated that the VK1BACK primer had hybridized within the FR2 region of C6 light chain, binding the bases coding for amino acids 37 to 44. Figure 3 shows a graphic representation of annealing possibilities for the VK1 BACK oligonucleotide with the actual C6 light chain FR2 base sequence (amino acids 37 to 44), and with a "test" base sequence constructed so as to agree with the first "invariant" and most frequent "variant" amino acids of the mouse kappa group II FR1. The VK1BACK primer has a greater homology with the FR2 of C6 than with the "test" FR1 region. It can also be seen that the C and G complementary pairings are more frequent for FR2 than for FR1. The same Figure shows that the opposite occurs with the EcoRI/FR1-ML(kappa) primer, due to the different base composition (including degeneracy) and the attached restriction site. IOR-T3a light and heavy chain variable regions were sequenced after asymmetric amplification and M13 cloning (Tables 2 and 3). Comparison of the sequences obtained with both methods shows a >99% homology and reveals that they could be asigned to kappa group VI and gamma group Ha, respectively (20). Base changes in sequences coming from asymmetric amplification and M13 cloning were detected at position 48 for IOR-T3a light chain variable region and in position 61 for the heavy chain. The combination of Leader/CONST primers for light chains amplified cDNA fragments of expected size (=420 bp) for the hybridomas CB-CEA.1, and C6. These are shown in lanes 4 and 6, respectively, of Figure 4, next to the fragments derived with the 411

A) ECORI/FRI

ML:

T CGC GTC GGGAATTCGACATTGTGATAACACAAGATCAA •

•

•

••••••••

12/32 (37.5%)

•

C6 FR2

CTGCAGAAGCCAGGCCAGTCTCCA

VK1BACK:

CAGATTCAGCTGACCCAGTCTCCA

15/24 (62.5%)

B) ECORI/FRI FRl

kappa

ML:

T CGC GTC GGGAATTCGACATTGTGATAACACAAGATCAA

19/32 (59.4%)

T C ACG GACATTGTNATGACNCAGGATCAA

II

VK1BACK:

CAGATTCAGCTGACCCAGTCTCCA

10/24 (41.7%)

FIGURE 3.- COMPARISON OF ECORI/FR1 -ML(KAPPA) AND VK1 BACK PRIMING ON: (A) FR2 of C6 (amino acids 37 to 44) and (B) FR1 of a test" kappa II light chain region (amino acids 1 to 8). For B) the base combinations resulting in the 'invariant* and most frequent "variant* amino acids (database of Kabat et al. 1987) were considered (two or four [N] bases per position). In case of alternative bases for one same amino acid, an arbitrary NO priming was established for both primers (unless coincident with the degenerate positions of the EcoRI/FR1-ML(kappa) primer. Percentages reflect the ratio of paired bases/total bases. Homologies between the EcoRI site of the EcoRI/FR1-ML(kappa) primer and the C6 FR2 (two additional pairings in bases 106 and 108) or the test* FR1 sequences (three possible pairings in bases -4, -5, and -6 of a 'consensus' leader region) are not shown.

TABLE 2: COMPARISON OF SEQUENCES FOR IOR-T3a LIGHT CHAIN VARIABLE REGION OBTAINED BY ASYMMETRIC AMPLIFICATION (a) AND M13 CLONING (b:). COMPLEMENTARY DETERMINING REGIONS (CDR's) ARE UNDERLINED AND FRAMEWORK REGIONS ARE LABELLED. CONSERVED AMINO ACIDS FROM KABATS DATABASE (LIGHT CHAIN KAPPA GROUP VI) ARE CAPITALIZED. FRl

10

primer

at-AGTCTCAA

bl

ALA ILE met ser ALA SER PRO GLY glu lys VAL THR met GCA ATC ATG TCT GCA TCT CAA GGG GAG AAG GTC ACC ATG GCA ATC ATG TCT GCA TCT CAA GGG GAG AAG GTC ACC ATG

CDRl a:

b:

27

29

THR CYS ser ALA SER SER SER VAL ser TYR met asn TRP tyr GLN GLN ACC TGC AGC GCC AGC TCA AGT GTA AGT TAC ATG AAC TGG TAC CAG CAG ACC TGC AGT GCC AGC TCA AGT GTA AGT TAC ATG AAC TGG TAC CAG CAG

CDR2

FR2 a:

bl

LYS ser gly thr SER PRO LYS arg trp ILE TYR asp thr SER lys LEU AAG TCA GGC ACC TCC CCC AAA AGA TGG ATT TAT GAC ACA TCC AAA CTG AAG TCA GGC ACC TCC CCC AAA AGA TGG ATT TAT GAC ACA TCC AAA CTG

54 ala

as

bs

FR3

GLY VAL PRO ala his PEE arg GLY SER GLY ser GLY thr ser GCT TCT GGA GTC CCT GCT CAC TTC AGG GGC AGT GGG TCT GGG ACC TCT GCT TCT GGA GTC CCT GCT CAC TTC AGG GGC AGT GGG TCT GGG ACC TCT ser

FR3

70 as

bs

gly

met GLU ala GLU ASP ala ALA thr TYR TAC TCT CTC ACA ATC AGC GGC ATG GAG GCT GAA GAT GCT GCC ACT TAT TAC TCT CTC ACA ATC AGC GGC ATG GAG GCT GAA GAT GCT GCC ACT TAT

tyr

ser LEU

86

thr ILE

ser

CDR3

gln

FR4

100

bs

GLN trp ser ser asn pro phe THR PEE GLY ser GLY THR TAC TGC CAG CAG TGG AGT AGT AAC CCA TTC ACG TTC GGC TCG GGG ACA TAC TGC CAG CAG TGG AGT AGT AAC CCA TTC ACG TTC GGC TCG GGG ACA

bs

LYS LEU GLU He LYS-> ARG ala asp thr ala pro thr val AAG TTG GAA ATA AAA -> CGG GCT GAT ACT GCA CCA ACT GTA

TYR CYS

as

FR4

107

CHl

412

TABLE 3: COMPARISON OF lOR T3a HEAVY CHAIN VARIABLE REGION SEQUENCES OBTAINED BY ASYMMETRIC AMPLIFICATION (a) AND M13 CLONING (b:). COMPLEMENTARY DETERMINING REGIONS (CDR'S) ARE UNDERLINED AND FRAMEWORK REGIONS ARE LABELLED. CONSERVED AMINO ACIDS FROM KABATS DATABASE (HEAVY CHAIN GAMMA GROUP II A) ARE CAPITALIZED. GLY ala glu primer AGGTCAAGCTGCAGGAGTCT GGG GCT GAA

bs FRl

13 bs

LYS ALA ILE PRO GLY ala SER VAL lys met SER CYS LYS ALA CTG GCA ATA CCT GGG GCC TCA GTG AAG ATG TCC TGC AAG GCT

36

CDRl as

bs

50 CDR2

FR2 as

bs

GLN arg pro gly gin gly LEU GLU TRP ILE GLY tyr ILE asn CAG AGG CCT GGA CAG GGT CTG GAA TGG ATT GGA TAC ATT AAT CAG AGG CCT GGA CAG GGT CTG GAA TGC ATT GGA TAC ATT AAT

CDR2

52a as

bs

PRO ser arg gly tyr thr asn TYR asn gin LYS PHE lys asp CCT AGC CGT GGT TAT ACT AAT TAC AAT CAG AAG TTC AAG GAC CCT AGC CGT GGT TAT ACT AAT TAC AAT CAG AAG TTC AAG GAC

66 as

bs

FR3

LYS ALA THR LEU THR thr asp lys SER SER ser THR ala TYR met AAG GCC ACA TTG ACT ACA GAC AAA TCC TCC AGC ACA GCC TAC ATG AAG GCC ACA TTG ACT ACA GAC AAA TCC TCC AGC ACA GCC TAC ATG

81

82a 82b 82c

glu

as

bs

as

CDR3

as

100a

b

103

ALA ARG tyr tyr asp asp his tyr cys leu asp tyr TRP GLY gln GCA AGA TAT TAT GAT GAT CAT TAC TGC CTT GAC TAC TGG GGC CAA GCA AGA TAT TAT GAT GAT CAT TAC TGC CTT GAC TAC TGG GCC CAA

FR4 bs

FR3

LEU ser ser LEU THR SER GLU ASP ser ALA val TYR tyr CYS CAA CTG AGC AGC CTG ACA TCT GAG GAC TCT GCA GTC TAT TAC TGT CAA CTG AGC AGC CTG ACA TCT GAG GAC TCT GCA GTC TAT TAC TGT

93

bs

FR2

SER GLY TYR thr PHE thr arg tyr thr met his TRP VAL lys TCT GGC TAC TCC TTA ACT AGG TAC ACG ATG CAC TGG GTA AAA TCT GGC TAC ACC TTT ACT AGG TAC ACG ATG CAC TGG GTA AAA

113

CHI

GLY THR thr leu THR VAL SER ser ala lys thr thr ala pro GGC ACC ACT CTC ACA GTC TCC TCA —> GCC AAA ACA ACA GCC CCA GGC ACC ACT CTC ACA GTC TCC TCA —> GCC AAA ACA ACA GCC CCA

FR1/CONST primer combination for the same hybridomas. This primer combination gave somewhat different amplification pattern for hybridomas I0R-T1 and CB-Fib.1, with double bands in the 400 bp size range (lanes 10 and 12, respectively). The well defined a

fragments obtained with FR1/CONST primer combination for the same hybridomas are shown in lanes 9 and 11.

DISCUSSION PCR is an attractive method for the rapid cloning of variable regions of rearranged immunoglobulin genes, a step that is often time- and effort-consuming by more conventional techniques. The design of "universal" primers for this task is very critical, and has been discussed in some recent publications (15, 16, 19, 20). In 413

12 3 4

5 6 7 8 9 10 11 12

FIGURE 4: IMMUNOGLOBULIN LIGHT CHAIN PCR PRODUCTS DERIVED FROM FIRST STRAND cDNA. Ten microliters from a 100 microliter PCR reaction was applied to a 2% agarose gel for electrophoresis in TBE buffer. The major PCR products obtained after ethidium bromide staining are shown. LANES 1 and 7: pBR322 Alul marker ladder; LANES 2 and 8: Parkin Elmer Cetus kit control (500 bp); cDNA: LANES 3-4: C6, LANES 5-6:CBCEA.1, LANES 9-10: IOR-T1, LANES 11-12: CB-Fib.1; Primers = LANES 3, 5, 9 and 11: EcoRI/FR1-ML(kappa) and Hindlll/ML(kappa)-CONST.; LANES 4, 6,10, and 12: EcoRI/Lead-ML(kappa) and Hindlll/ML(kappa)-CONST.

principle, one must consider that variability in sequence exists among antibodies, even for the most conserved regions. Figure 1 shows the strategy for design of PCR primers for mouse hybridoma cells. We have designed a series of 5' end primers for leader (light chain) and FR1 (light and heavy chain), and 3' end primers for C kappa and CH1, using conserved sequences available in databases. Our 5' end FR1 primers hybridize with the first 7 and 8 amino acids of the mouse heavy and light chain variable regions, respectively. Our heavy chain FR1 primer is very similar to that reported by Orlandi et al. (1989), while the light chain EcoRI/FR1 -ML(kappa) oligonucleotide has important differences due to base composition, and degeneracy (Table 1a). We decided that the use of 3' end primers that hybridize in the constant domains, instead of the FR4 site used by Orlandi et al. (16), would be advantageous both because of the high degree of conservation found in these regions, and because the oligonucleotide structure would not interfere with the actual sequence of the variable region. This avoids potential changes in the spatial projection of the relevant complementarity determining region three (CDR3). This strategy using FR1/CONST primer design resulted in specific amplifications of mouse light chains of the kappa group II (C6), group V (CB-CEA.1, not shown in detail) and group VI (IOR-T3a), and heavy chain gamma gene groups He (C6) and Ha (IORT-3a), as defined by Kabat et al. (20). It was necessary to include degenerate base positions in four of our primers (see Table la). Our results show that for mouse cDNA, degenerate oligonucleotide primers will nevertheless prime the PCR reaction; this fact had already been pointed out before by Larrick et al. (15) in the case of human immunoglobulin region amplification, and suggests strongly that PCR can tolerate a number of mismatches and still initiate consistent priming of the polymerase extension reaction. Regarding the functioning of oligonucleotides reported by other authors, we have demonstrated that the particular design of the VK1 BACK primer of Orlandi et al. (16) can produce an undesirable artifact: it anneals with the FR2 of mouse kappa group II light chain genes. Three experimental results support this finding: (a) The sequencing experiments showed hybridization of the VK1 BACK primer within the FR2 region of C6 light chain, corresponding to and replacing amino acids 37 to 44. (b) The poor priming obtained with VK1BACK when amplifying cDNA from the IOR-T3a hybridoma. The FR1/FR2 structure of this MAb is characteristic of a different kappa gene group (VI). (c) Correct amplifications of MAbs from both the kappa II and VI groups using the

414

EcoRI/FR1-ML(kappa) oligonucleotide

that

incorporated important changes

in base

structure were successful.

In the particular case of C6, the preferential hybridization of the VK1 BACK primer in FR2, rather than in FR1, may have been favoured by the specific G and C content of the target sequences. In addition our use of an annealing temperature of 55°C in the PCR (vs 30°C) used by Orlandi et al.(16) may have facilitated priming fidelity. These findings indicate how conditions of the PCR reaction can modify yield of products. We demonstrate that heavy and light immunoglobulin variable region sequences can be obtained directly from asymmetrically amplified cDNA, reducing the effort required to obtain sequences via conventional cloning techniques. Comparing asymmetrically amplified sequences with sequences obtained after cloning we found a base change rate of about 1 base/500 sequenced. One base change was identified in both heavy and light chain variable regions of IOR-T3a. Other authors have noted problems with the fidelity of PCR (29). Sequences obtained from other mouse and human hybridomas ,(not shown here) have given similar results. The impact that such findings will have in the overall use of PCR-cloned immunoglobulin products for expression is not clear. We normally confirm sequences on several independent PCR samples. Finally, our preliminary data based on amplified fragment size in agarose gels, suggest that kappa light chain 5' end consensus primers that hybridize with the mouse leader regions can be designed. In fact, Larrick et al. (15) have previously used leader and constant region primers and succeeded to amplify specifically and sequence light and heavy chain gene regions coding for several human IgG (kappa and lambda) and IgM monoclonal antibodies. The use of leader region primers may provide a method to clone a 'complete' native leader/variable region insert for direct insertion into expression vectors, and this could greatly reduce the time and effort to produce recombinant antibodies and to obtain V regions of B cells involved in cancer, autoimmunity, antibody responses etc. Nevertheless, the mouse leader primers have been less consistent than the ones prepared for human V regions, and recent unpublished results from our laboratories indicate that a similar phenomenom may exist with mouse heavy chain leader primers. Whether this reflects an increased heterogeneity for mouse leader sequences, or derives as an artifactual effect of the low number of reported data that can be found for this region in available databases, has still to be elucidated.

ACKNOWLEDGEMENTS We thank Cameron Hoover and Victor Jiménez for help with oligonucleotide synthesis, and Marta Guerra for her advice in sequence reading. J.V.G.C. was partially supported by a grant from the National Swedish Agency for Research Cooperation (SAREC).

REFERENCES 1. Larrick J.W. (1990) Potential of monoclonal antibodies as pharmacologcial agents. Pharm Rev. 41:539-557. 2. Larrick, J.W., J.M. Bourla (1986) Prospects for the therapeutic use of monoclonal antibodies. J. Biol. Resp. Mod. 5:379-387. 3. R.M. Bartholomew, LM. Smith, R.D. Dillman (1985) Human immune response to multiple injections of murine monoclonal IgG. J. Immunol. 135:1530-1535. 4. Schroff, R., K. Foon, S. Beatty, R. Oldham, A. Morgan Jr. (1985) Human anti-murine immunoglobulin responses in patients receiving monoclonal antibody therapy. Cancer Res. 45:879-885.

415

5. Borrebaeck, C.A.K. (1988) Human MAbs produced by primary in vitro immunization. Immunol. Today 9:355-359. 6. Williams, G. (1988) Novel antibody reagents: production and potential. Trends in Biotech. 6:36-42. 7. Morrison, S.L, V.T. Oi (1989) Genetically Engineered Antibody Molecules. Adv. Immunol. 44:65-92. 8. Liu, A.Y., R.R. Robinson, K.E. Hellstrom, E.D. Murray Jr., C.P. Chang, I. Hellstrom (1987) Chimaeric mouse-human lgG1 antibody that can mediate lysis of cancer cells. Proc. Nati. Acad. Sei. USA 84:3439-3444. 9. Steplewski, Z., L Sun, C. Shearman, J. Ghrayeb, P. Daddona, H. Koprowski (1988) Biological activity of human-mouse lgG1, lgG2, lgG3, and lgG4 chimaeric monoclonal antibodies with antitumor especificity. Proc. Nati. Acad. Sei. USA 85:4852-4856. 10. Sun, L.K., Curtis, P., Rakowicz-Szulczynska, J. Ghrayeb, N. Chang, S. L Morrison, H. Koprowski (1987) Chimaeric antibody with human constant regions and mouse variable regions directed against carcinoma-associated antigen 17-1 A. Immunology 84:214-218. 11. Beidler, C.B., J.R. Ludwig, J. Cardenas, J. Phelps, CG. Papworth, E. Melcher, M. Sierzega, LJ. Myers, B.W. Unger, M. Fisher, G.S. David, M.J. Johnson (1988) Cloning and high level expression of a chimaeric antibody with specificity for human carcinoembryonic antigen. J. Immunol. 141: 4053-4060. 12. Riechmann, L, M. Clark, H. Waldmann.G. Winter (1988) Reshaping human antibodies for therapy. Nature 332:323-327. 13. LoBuglio, A.F., R.H. Wheeler, J. Trang, A. Haynes, K. Rogers, E.B. Harvey, L Sun, J. Ghrayeb, M.B. Khazaeli. Mouse/human chimaeric monoclonal antibody in man: Kinetics and immune response (1989) Proc. Nati. Acad. Sei. USA 86:4220-4224. 14. Hale, G. M.R. Clark, R. Marcus, G. Winter, M.J.S. Dyer, J.M. Phillips, L Riechmann, H. Waldmann (1988) Remission induction in non-Hodgkin lymphoma with reshaped human monoclonal antibody Campath-1H. Lancet, 1394-1400. 15. Larrick, J.W., L Danielsson, C.A. Brenner, E.F. Wallace, M. Abrahamson, K. Fry, C.A.K. Borrebaeck (1989) Polymerase chain reaction using mixed primers: cloning of human monoclonal antibody variable region genes from single hybridoma cells. Bio/Technology 7:934-938. 16. Orlandi, R., D.H. Gussow, P.T. Jones, G. Winter (1989) Cloning immunoglobulin variable domains for expression by the polymerase chain reaction. Proc. Nati. Acad. Sei. USA 86:3833-3837. 17. Oste, C. (1988) Polymerase chain reaction. Biotechniques 6:162-167. 18. Huse, D.W., L Sastry, S.A. Iverson, A. S. Kang, M. Alting-Mees, D. R. Burton, S. J. Benkovic, R. A. Lerner. (1989) Generation of a large combinatorial library of the immunoglobulin repertoire in phage lambda. Science 246:1275-1281. 19. Sastry, L, M. Alting-Mees, W.D. Huse, J.M. Short, J.A. Sorge, B.N. Hay, K.D. Janda, S.J. Benkovic, R.A. Lerner (1989) Cloning of the immunological repertoire in Escherichia coli for generation of monoclonal catalytic antibodies: Construction of a heavy chain variable region-specific cDNA library. Proc. Nati. Acad. Sei. USA 86: 5728-5732. 20. Ward, E.S., D. Gussow, A.D. Griffiths, P.T. Jones, G. Winter (1989) Binding activities of a repertoire of single immunoglobulin variable domains secreted from Escherichia coli. Nature 341: 544-546. 21. Chaudhary, V.K., J.K. Batra, M.G. Gallo, M. C. Willingham, D. J. Fitzgerald, I. Pastan. (1990) A rapid method of cloning functional variable-region antibody genes in Escherichia coli as single chain immunotoxins. Proc. Nati. Acad. Sei. (USA) 87:10661070. 22. Gavilondo, J.V., A.M. Vázquez, S. Fong, E. Rengifo, A Fernández, C.A. Garcia, A. Velandia, A. Pavlenko, S. Hernández, N. Ruisánchez, U. Sundín, M.E. Faxas (1987)

416

Monoclonal antibodies against carcinoembryonic antigen in ¡mmunohistochemical and immunochemical assays. Interferon y Biotecnología 4: 143-156. 23. García, CA., J.V. Gavllondo, J.F. Amador, A.M. Vázquez, A. Fernández (1984) Obtention of mouse hybridomas producing monoclonal antibodies that recognize human cells of T origin. II. Characterization of monoclonal antibodies IOR-T1 and IOR-T2. Interferon y Biotecnología 1: 29-38. 24. Kabat, E.A., T.T. Wu, M. Reid-Miller, H.M. Peery, K.S. Gottesman (1987) Sequences of proteins of immunological interest. 4th. Edition. U.S. Dept. Health and Human Services. 25. Gough, N.M. (1988) Rapid and quantitative preparation of cytoplasmic RNA from small number of cells. Analyt. Biochem. 173: 93-95. 26. Mihovilovic, M., J.E. Lee (1989) An efficient method for sequencing PCR amplified

BioTechniques 7:14-16. Messing, J., B. Groneborn, B. Muller-Hill, P.H. Hofschneider (1977) Filamentous coliphage M13 as a cloning vehicle: insertion of a Hindi 11 fragment of the lac regulatory region in M13 replicative form in vitro. Proc. Nati. Acad. Sei. USA 74:3642-3647. 28. Vieira, J., J. Messing (1982) The pUC plasmids, a M13mp7-derived system for insertion mutagenesis and sequencing with synthetic universal primers. Gene 19:259-264. 29. Eckert, K.A. and T. A. Kunkel. (1990) The fidelity of Taq DNA polymerase. Nucleic Acids Res. (in press). DNA. 27.

Address

reprint requests

Dr. James W. Larrick

Genelabs Incorporated 505 Penobscot Drive Redwood City, CA 94063

Received for publication May 2, 1990 Accepted after revision June 26, 1990

417

to: