CA2138009A1 - Production of monoclonal recombinant antibodies without the use of hybridomas by iin vitro spleen fragment culture combined with isothermal self-sustained sequence replication of rna - Google Patents

Abstract

The invention is directed to a method for producing a recombinant protein having a preselected binding affinity and specificity for a given antigen. Generally, the method combines spleen fragment culture for the in vitro somatic mutation of B cells in response to antigen, followed by preferential amplification of RNA
encoding heavy and light chains of desired antibodies produced by those B cells. The amplified RNA is then converted to DNA and incorporated into expression vectors; transfected host cells then express the antibody chains or a single chain antibody having preselected characteristics. Also disclosed are generic degenerate primer pools for the efficient amplification of RNA encoding mouse IgG.

Description

" ` WO94J1~507 PCT/US93/11295 PRODUCTION OF MONOCLONAL RECOMBINANT ANTIBODIES WITHOUT
THE USE OF HYBRIDOMAS BY IN VITRO SPLEEN FRAGMENT CULTURE
COMBINED WITH ISOTHERMAL ~ELF-SUSTAINED SEQUENCE
REPLICATION OF RNA

This is a continuation of US patent application serial number 07/978,835, filed November 19, 1992.

Portions of the research described in this application were supported by a grant from the National Institutes of Health, and the U.S. government may have certain rights in this invention.

Field of the Invention ;

The present invention relates to methods for the ;;
production of recomhinant monoclonal antibodies. The ~;
methods of the instant invention are based on in vitE

2~0~ spl~een~fragment culture of B cells which produce mo~oclonal~antibodies combined ~ith the technique known as~`isothermal self-sustained sequence replication (3SR), which~amplifies RNA~encoding the antibodies of interest.
The invention also~relates to generic degenerate primer `25~ pools~for use in replication of RNA encoding monoclonal antibodies of interest.

ound of the invention 30~ MonocIonal antibodies~are used in therapy, diagnostics,~ "
and~basic research. For in vivo therapy, monoclonal antibodies are directed against a toxin or a tumor ; surfacel antigè~ to mo~nt the patiient's immune response against the targeted toxin or pathogenic cell.
35 ~ Alternatively, a monoalonal antibody may be conjugated with~a toxin, thereby directing the toxic action to the tumor~cell. Ex vivo, or extracorporeal immuno-therapy "~ , . ..

, ~, , .

;;

W094/ 1507 PCT/~S93tll295' ~
~ ~3 2 involves the removal of the patient's bone marrow or blood and isolation of a selected cell population by means of binding a predetermined antigen on the cells' surface with a monoclonal antibody against that cell surface antigen. The monoclonal antibody may be~
chemically bound to magnetic beads which facilitate the removal of the cells bound to the antibody. Depending on the type of therapy being administered, the isolated cells may be discarded or they may be returned to the patient at a later date. Isolation of hemopoietic cells also offers the possibility of genetic manipulation of the patient's ceIls to correct a genetic defect or to expand a desired population of hemopoietic cells, which are thereafter returned to the patient. Immunodiagnostic tests are based on the generation of a signal proportional to the extent of binding of a monoclonal antibody to a specific substance (antigen) in a patient's ~ blood or tissue.
: ' Immunotherapy and reproducible immunodiagnostic tests ~require antibodies of standardized specificity and `~
aff~inity for~the antigen of interest. A significant step ~ toward this goal was achieved with the development of '~,2~ hybridoma technology that allows the growth of clonal ~populàtions of cells secreting monoclonal antibodies with `~
a defined specificity (Kohler, G. and Milstein, C., 1975, Nature 256:495-497~. In this technique an antibody-secre~ing~B-cell, isolated from an immunized animal, is fused with an immortalized myeloma cell. The products of ~, . .
this fusion are called hybridoma cells and are the source -~
of monoclonal antibodieq which~currently provides the most reproducible source of antibodies for immuno-therapy and immunodiagnostics.

Production of hybridoma cell lines producing-monoclonal , : .
m antibodies of a desired specificity requires multiple ': ' -~

~ , ~
' ~ W094/l1507 2 1 3 8 0 0 9 PCT/US93/1129~ ~

experimental steps which may extend over a significant time period. The immunization, screening and hykridoma production stages all pose problem~tic challenges in - which costs increase in proportion to the amount of time and effort needed. Additional difficulties are a~t times encountered even after a monoclonal hybridoma cell line is identified and clonally selected. An expanded population of hybridoma cells can lose its ability to produce a specific monoclonal antibody after prolonged growth in culture. This propensity to lose antibody expression is also observed in hybridoma cell lines that have been frozen and stored after clonal selection.
Thus, the longevity of valuable hybridoma cell lines is ~ not predictable.

One approach to address this longevity problem has been to clone the light and heavy chains for a specific monoclonal antibody from the hybridoma mRNA pool and then to employ the polymerase chain reaction ~PCR) to clone ~ ~and~expréss the variabIe domains of light and heavy chalns of monoclonal antibodies in bacterial or mammalian ce11 lines (orlandi, R., et al., 1989, Proc. Natl. ~cad.
Sci. U.S.A. 86:3833-3837; Sastry, L., et al., 1989, Proc.
Natl. Acad. Sci. U.S.A. 86:5728-5732; Larrick, J.W., et a~ (1989) Bio/TechnoloqY 7:934-938; Ward, E.S., et al., (1989)~Nature 341:544-546). In so doing, a stable supply of~a specific monoclonal antibody might be obtained if ,~, :
-~ the host cell clone is sufficiently stable.

However, several factors suggest that this PCR/cloning ~- approach may not always prove satisfactory. First, a hybridoma cell line must be identified and establis~ed in order to serve as a target for the PCR/cloning steps. As indicated above, this step may prove to be problematic.
-~ 35 Second, the use of reverse transcriptase (RT)/PCR
technique (Kawasaki, E.S., 1990, In: PCR Protocols. Eds.

, ~, ~:

WO94/11~07 PCT/US93/ll2~5( `
213800!~ 4 - M.A. Innis, D.H. Gelfand, J.J. Sninsky, T.J. White pp.
21-27) to clone light and heavy chain mRNA from a hybridoma cell line producing a mono~lonal antibody of interest has frequently proven to be problematical. This - 5 RT/PCR failure has been attributed to the complication of non-productive immunoglobulin rearrangements which increase the complexity of the amplification-generated cDNA library. In turn, this significantly increases the likelihood of missing clones containing the productive immunoglobulin rearrangements.

In an effort to recover the originally paired VH and VL
genes expressing the desired antibody, as well as to test novel synthetic V gene combinations, a bateriopllage display approach has been used. Antibodies engineered as a single chain or as heterodimeric Fab constructs have - been displayed on the surface of a fila~entous phage as a fusion of the phage pIII protein (McCafferty, J., et al., Nature 348:552-S54). Fusions of the phage pVIII protein a~lso~have~been used to display antibody fragments on the surf~ace~of~ the filamentous phage (Chang, C.N., et al., (1991~ J ImmunoloqY 147:3610-3614; Huse, W.E~, (1991) In:
Borrebaeck, D.A.K. (ed.), ntibodv Enqineerinq. A
Practical Ap~roach., W.H. Freeman & Company, N.Y., p.103;
Kang, A.S., et al., (1991) PNAS USA, 88:11120-11123).
Binding of the displayed antibody on the surface of the phagé with antigen can be used as a way to recover -~ antibodies composed of the original or novel VH and VL
~i~ gene pairings. However, because each phage contains only - ~ 30 one such VH and VL pairing, the frequency of the occurrence of any single VH and VL pairing is less than l/108. Consequently, this approach requires the construction of phage display libraries of 10l1 copies to ~-~ provide sufficient copies of any specific VH and VL
-~ 35 combination for detection. In order to recover antibodies of especially high affinity or unique , ; , , ,~ , ~:

~ ~ WO94/11507 2 i ~ 8 0 0 9 PCT/US93/11295 specificity, the size of a phage display library may have to be even greater than lO copies.

How does an animal's immune system overcome these innate challenges to develop antibodies of the necessar~y specificity and high affinity? The differentiation of B
cells in vivo and in vitro is described in Cell 59:1049-1~59, 1989, Linton, et al. In vivo, immunization (exposure to foreign antigen) induces both a primary antibody response mounted by primary B cells and generation of T and secondary B cells which increase in ~- specificity for the antigen and enable the organism to mount a more vigorous response upon additional exposure to the same antigen.
In order for the immune system to respond to the enormous - spectrum of different antigens in the environment, it must~be capable of producing antibodies with tremendous diversity. This problem is solved in part by the ability 20~ of~immunoglobulin genes to rearrange the multiple forms of~the four gene regions, V (variable), D (diversity), J
(joining), and C (constant), prior to transcription and translation. Additional diversity is attained by the aasociation of the heavy chain with one of many types of 25 ~ ght chains. A primary antibody is composed of a light chain~and a heavy chain, which are connected by a disulfide bridge to form a light/heavy chain complex;
this~complex is associated with other light/heavy chain complexes in either dimeric or pentameric forms.

,~ , Secondary B cells differ from primary B cells with respect to the variablè regions which dominate a response. For each isotype of immunoglobulin (e.g., IgM, ~-~ IgG, IgD), there are families of variable gene cassettes7i: ~ 35 from which a single variable gene is selected in response , ~
~ to an antigen. Cells which carry an appropriate gene , , ~, ~ .
''~ ~: ' .

W094/11507 PCT/US93/11295 -~
21 3 8 0~9 6 cassette are known as "responsive B cells" for that particular antigen, even if they have not been previously exposed to that particular antigen, i.e. the responsive B
c.ells have been pre-programmed to respond to the general three-dimensional configuration and physico-chemical attributes of the antigen of inter~st. Secondary B cells differ from primary B cells in their propensity to accumulate somatic mutations in the variable region to produce antibodies with substantially higher affinities for the antigen than those produced by primary B cells.

Clearly, it would be advantageous to be able to (1) isolate an~ identify a secondary B cell producing an antibody of desired affinity and specificity, in order to - 15 (2) isolate and clone the genetic sequence encoding that antibody. Once the genetic sequence is cloned, it is possible to use genetic engineering to produce the ; antibody of~ interest in the form of a recombinant protein. Considerable progress has been made towards 20 ~ achieving~;~the~first step via in vitro spleen fragment culture ~Lintonj et al., supra). However, it remains problematic to isolate and clone the immunoglobulin gene of~interest because the identified B cells are expected to~contain non-productive immunoglobulin gene ~," ~as~ rearrangements~as well as the gene of interest.
Therefore, attempts to amplify the gene af interest dir~àtly from DNA by PCR or RT/PCR, using primers encoding portions of the gene of interest, are complicated by the concomitant amplification of other 30 ~ productive and non-productive immunoglobulin gene rearrangements. The resulting amplified DNA would be expected to contain a complex mixture of the gene of interest and other genes. This situation would -~ - necessitate extensive æcreening of proteins expressed --~ -35 from the amplified DNA in order to identify a clone ~ expressing an antibody of the desired specificity and '''~
.~
,.....
,' ! ~--~ WO94/11507 2 1 3 8 0 0 9 PCT/US93/11295 affinity. Moreover, there would be a low expectation of success from this approach.

What i~ needed is a means to selectively amplify nucleotide sequences encoding only the desired H~and L
chains.

Summary of the Invention The invention is directed to methods for obtaining an amplified RNA encoding a predetermined chain of a desired antibody against a specific antigen. Briefly, the methods involve the steps of ~l) isolating B cells from one animal; optionally, JllD
cells may be selected, (2) injecting the B cells into a second animal, (3) allowing the B cells to colonize the spleen of the second animal, (4) culturing spleen fragments from the second animal, (5) stimulating the spleen fragments with antigen, (6) allowing B cells to proliferate and produce antibody in the cultures, (7) assaying the cultures for a desired antibody, (8) extracting total RNA from the fragment in an identified positive culture, and (g) amplifying target RNA encoding the heavy or light chain of the desired antibody using 3SR in conjunction with a generic degenerate primer pool.
-,., The~methods of the invention can be extended to produce a desired recombinant monoclonal antibody by converting RNA
encoding the heavy chain and RNA encoding the light chain - to cDNA. The cDNA is converted into double-stranded DNA
.~ product by primer extension or RT/PCR and is then incorporated into one or two expression vectors, which -;~ 35~ are transfected into host cells. The host cells synthesize the recombinant monoclonal antibody possessing ~, .:

~ .

WO94/11507 PCT/US93/11~95 ~3~0~ ` 8 ~ the desired specificity and affinity for the antigen as identified in previous assays of the spleen cultures.

The invention is also directed to a method for constructing a generic degenerate primer pool for~the efficient amplification of RNA encoding an immunoglobulin heavy or light chain using 3SR.

The invention is also directed to degenerate primer pools generic for mouse immunoglobulin light and heavy chains.

Definltions For the purpose of construing the intended meaning herein, the following terms are defined as follows:
, The term "B cell" encompasses the entire cateyory of , cells which lead to antibody production. Thus, "B cell"
refers~to a mammalian cell of hemopoietic origin whose ~20~ ;progèny~produce antibodies. The term "B cell" also includes stem cells or precursor cells which have the capacity to proliferate to produce progeny which ~
- ultimately produce antibodies. Herein, the term "B cell"
also includes "memory B cells" or "secondary B cells"
25~ that~have been generated as the result of a previous immunization.
,~
Herein~ the term "antigen" refers to a molecule which elicits an immune response in the form of antibodies produced by B cells. The antigenic molecule may be a protein or other molecjule j of biological or environmental origin. For instance, antibodies may be produced against --~ a protein, a nucleic acid, or a steroidal hormone when cells are stimulated by the specific molecule.
The term "responsive B cell" refers to a B cell which ,, , . .

,~
,~. .

;`; WO94/11507 2 i 3 8 0 0 9 PCT/US93/11295 contains an immunoglobulin gene rearrangement which enables the B cell to produce an antibody against a specific antigen. A responsive B cell may be a precursor cell which contains a gene cassette encoding an antibody of low specificity for the general three dimensio~al configuration and physico-chemical properties of the specific antigen. Later in differentiation, a responsive B cell may contain an immunoglobulin gene rearrangement which encodes an antibody of greater specificity and affinity for the specific antigen. The above defined term "B cell" is understood to include "responsive B
cells".

The term "antibody" minimally refers to an immunoglobulin lS having at Ieast one light and one heavy chain (Fab), or - at least two light chains and two heavy chains, each chain having a variable and possibly a canstant region.
The ant~body may be an IgG, an IgA, an IgE, an IgM, or an IgD.
20~ ~
The term "monoclonal antibodies" refers to antibodies which are identical to each other because they have been produced by cells arising from a single cell clone expressing defined H and L chain genes encoding the 25 ~ antibodies.
~; ~
The term "recombinant antibody" refers to an antibody encoded on an expression vector contained in host cells and expressed as a protein by the host cells. A
recombinant antibody may be in the form of heavy chains and li~ht chains bound by di-sulfide bonds, just as a l~ naturally occurring antibody. Alternatively, a j~;`; recombinant antibody may be a single-chain antibody ~ construct which has the capability to bind a specific i~ 35 antigen.
'~

1.`~":
~, ., ,~
~ ~.
~::
:~

WO94/11507 PCT/US93/1129~i 21380~ lo - The term "heterogeneous RNA" refers to RNA which encodes many different cellular proteins. Thus the term "heterogeneous RNA" encompasses total RNA extracted from a culture of cells, as well as poly~A)~ RNA isolated from total RNA extract.

The term "target RNA" refers to RNA which encodes a single protein.

The acronym "3SR" refers to isothermal self-sustained sequence replication. The term "isothermal self-sustained sequence replication" refers to a process by which a target RNA is amplified in a process which essentially mimics the way in which retroviral RNA is amplified in nature. Alternative terms are 3SR and TAS.
- Generally the process makes use of three different enzyme activities: reverse transcriptase, RNA polymerase, and ~ ~ RNaseH. The word 'iisothermal" refers to the fact that -- ~ the process is carried out at~a single or narrow 2Q~ temperature range below about 95C so that double-s stranded DNA is not denatured, and therefore is not amplified. This feature makes th`e process specific for the amplification of RNA rathér than double-stranded or genomic DNA. During the process, the temperature may be varied over any range at which all the necessary enzymes function. Thus, the word "isothermal" is not intended to abrogate variat~ions in the temperature range below about ~` 95~C.

The term "degenerate primer pooll' refers to a collection of primer ~sequences ~in which many or all possible variations in the target sequence are represented.

` ~ The term "generic degenerate primer pool" refers to a degenerate primer pool in which the variations are ~; ~ limited to those sequence substitutions represented in a '~

;;. WO94/11~07 2 1 3 8 0 0 9 PCT/US93/11~95 catalogue of sequences known for a specified immunoglobulin chain. Thus, for instance, a degenerate primer pool generic for mouse IgG light chain would contain sequences encoding a 3' constant region as well as all possible sequences known for 5' variable regions of all known mouse IgG families such as lambda and kappa.

The term "consisting essentially of", in reference to an amplified target RNA, means that the product contains the amplified 3SR product RNA in overwhelming propor~ion to any other amplified RNA or DNA se~uences derived from non-targe~ RNA which may have been present in the starting material.

Brief_Descri~tion of Drawinqs .~ Figure l depicts the-major steps in producing a monoclonal recombinant antibody using the methods of the invention.
Figure 2~ defines the sequences of oligonucleotide primers : referred to herein by the number in the far left column of Figure 2.

25: ~ ~ igurc ~ 3how_ Northe~n (~) and outhcrn (~
analyses of 3SR and PCR amplification pro ~ (A) 3SR
RNA products using heavy chain lead ~ C-L) and light ahain variable (LC-V) ~'end p ~ s. (B) The 3SR RNA
products~were converte ~ -DNA. LC=light chain;
30 ~ HC=heavy chain; ~ e markers of ~s~I digested pBR3~2 fra~ments ~ ti-PC specific 5'end primers; Gv=generic var ~ region 5'ènd primers; GL=generic leader region ~ Y~nd primcrs ."~
Figure ~ i5 a scheme for the construction of plasmid pTl5-llO which encodes a single chain (sc) antibody form : RE~IFIED SHEE~ (RULE gl) ~ ~ ISA/EP
:

~380 0 9 of the anti-PC antibody. MBS=metal binding site; TAG=c-myc tag; LC-light chain; HC=heavy chain.

- rigurc 5 ohow~ nucleotide a~d amino ~cid ~qucnco~ of four heavy (4A) and light t48) chain groups (G ~ ed from spleen fragment culture D~A6 (anti- ~ The CDR
re.gions are noted in the heavy a ~ t chain groups by - overlining the appropriate ~ ences. The heavy chain sequences include ~ nd variable region primers (Figure 2) a ~ light chain sequences include only the last t ~ nucleotides of the 5'end variable region ~6 (Figur~ 2)-Detailed Descri~tion The general concept of the invention is outlined in Figure l as follows:

B cells are isolated from a donor mammal that has been previously immunized with a specific antigen.
[Optionally, a non-immunogenized mammal may be used as ~- the donor. Optionally, B cells may be further selected for precursor status by their lack of memory B cell markers.]
~- 25 Isolated B cel1s are in~ected into a recipient mammal without viable endogenous immune system cells.
Optionally~ the recipient may have been previously immunized with all or part of the antigen to promote T
cell helper function. The injected B cells are allowed to colonize the recipient's spleen.

Spleen of the reci~ient is removed and dissected, each fragment being placed in a separate culture well.
'' ~ S~leen fraoments are maintained in culture and stimulated . ~

- RECrIFIED SHEET (RULE91) ~ ISA/EP
:~ ~

' W094/11507 2 i 3 8 0 0 9 PCT/US93/11295 l3 with antigen. Certain fraqments contain responsive B
cells which will proliferate to produce progeny which synthisize antibody against the antigen.

Each culture is assayed for an antibody of interest, based on desired affinity and specificity for the antigen. A culture containing a positive fragment is identified.

The positive fraqment is extracted for total RNA.

RNA from the positive fraqment is subjected to 3SR
amplification of the RNA encoding the antibody of ; interest. Generic degenerate primer pools are provided herein for the amplification of RNA encoding mouse immunoglobulin light and heavy chains.

RNA encodinq the antibodY of interest is converted into a double-stranded cDNA.
~-~ 20 ~
Double strand_d cDNA encodina the antibodY of interest is incorporated into an expression vector.

The exore-sion vector is tran3~ected into a host cell.

The desired antibodY is expressed by the host cell and isolated.
~ . " ~ ~
Preferably, a donor animal such as a mouse is immunized with an antigen of interest such as estradiol. Injection ;of antigen will stimulate the proliferation and di~ferentiation of B cells which produce a primary antibody against the antigen as well as memory B cells.
Alternatively, a naive animal, i.e. one which has not 35~ been injected with antigen, may be a donor of B cells since it is expected that certain precursor B cells . ~ , ..~, -, .

i : :

WO94/11~07 PCT/US93/1129 2,~38009 possess gene rearrangements encoding antibodies of low specificity for a large repertoire of antigenic characteristics, including those of the antigen of interest.
After a period of 2 - 16 weeks, in the case of the immunized donor, B cells are isolated from the donor animal's bone marrow, peripheral blood, or spleen.
Alternatively, B cells may be isolated immediately from a naive donor. The isolated B cells may be subjected to a further selection process whereby cells which exhibit low binding capacity for JllD monoclonal antibody [JllDI
cells] are selected for injection (Litton, et al., Cell 59:1049-lOS9, 1989). The JllD cell population is considered to be enriched for secondary B cells, which are expected to produce antibodies of higher affinity and specificity for the antigen of interest;

Isolated B cells are injected into a recipient animal which does not possess viable endogenous immune system cells. ~The recipient animal may have been "lethally rradiated" sUch that essentially all its immune system cells~have been destroyed. The term "lethally irradiated" refers to the killing of immune system cells and does not refer to killing of the animal itself.
However, without sterile isolation or immune system replenishment the lethally irradiated animal would soon 5''~^'~ suceum~to infection. Alternatively, the recipient animal, for instance an SCID mouse, may be genetically ^, --, deprived of an immune system. In either case, the spleen of the recipient animall is denuded of endogenous B cells and is open for colonization by injected B cells.
~" ~
The concentration of injected B cells is calculated so 35~ that, on average, only one responsive B cell will - colonize a predefined segment volume of the recipient , ' ' ',~ ,:
.~
, ^ -~.'' ~n~a; ~
- W O 94~11507 21 3 8 0 0 9 PC~r/US93/11295 . 15 animal's spleen. In the case of a mouse recipient, the concentration calculation is described in Linton, et al, (su~ra).

The recipient animal is maintained for sufficient time to allow the injected B cells to colonize the spleen.
Preferably, the recipient animal is also immunized with - antigen to stimulate helper T-cell function, which facilitates B cell proliferation and differentiation. If . .
an immune deficient recipient is used, the transfer of antigen stimulated helper T cells is needed to facilitate B cell proliferation and differentiation~
;::
~ .
The spleen is then removed from the recipient animal and - 15 dissected into fragments, the size of the fragments being - - chosen so that each fragment contains, on the average,only one responsive B cell. In the mouse system described in Linton, et al., (su~ra), the optimum ~ragment size is about l mm3.

Each spleen fragment is placed in a separate culture well, typically in 96 well plates. The fragments are maintained in culture as described in Example 1 below.

~o eaoh culture well, an aliquot of the antigen of - interest is added. In any gi~en spleen fragment where there is a responsive B cell for the antigen, the responsive B cell is expected to proliferate and dif~ferentiate to produce multiple cells which synthesize antibodies of increased specificity and affinity for the ~.~
antigen of inte,rest.~

The medium of each culture well is assayed for affinity and specificity for the antigen of interest.

A positive well is identified which contains a spleen , ~, ~:

.` ,_.
W094/11507 ' PCT/USg3/11295'~
~oO9 l6 ~ fragment having B cells producing the antibody of ,', interest. The spleen fragment technique for obtaining ~, monoclonal antibodies has several advantages over :- hybridoma technology. The time and expense involved in - 5 production and screening of spleen fragments is expected ~,` to be less than the time and expense involved in ,~ producing hybridomas as described above. Additionally, ,,' the spectrum of antibody products may'be much greater using the spleen fragment technique.
~1 0 From the identified spleen fragment, total RNA is extracted. Optionally, poly(A)+ RNA may be isolated from - the total RNA. In either case, the resulting heterogeneous RNA encodes many proteins synthesized by ., .
the B cells. The target RNA species, which encode light and heavy chains for the antibody of interest, are expected to represent only a portion of 'the total RNA or the poly(A)+ RNA extracted therefrom.

~;. '~ 20 The next task is to amplify the RNA which encodes the , ~ heavy and light chains of the antibody of interest. This task is made,difficult by the inevitable presence of non-desired extraneous nucleotide sequences which have certain homologies to the RNA of interest but which, if amplified, will yield a-non-productive genetic sequence.
,;'~ That is, amplified extraneous nucleotides would probably not encode a functioning antibody, and they certainly would not encode the previously identified antibody of interest.

The problem with ampli,fication of non-productive gene ~- ~,' rearrangements was recognized in early attempts by the '-- ~ present inventors to amplify the immunoglobulin gene of '~- interest using polymerase chain reaction (PCR~. PCR is '- ~35 based on cycles of denaturing doubled stranded DNA by temperatures above about 95C to form single stranded ::

WO 94/11507 2 1 3 8 0 0 9 Pcr/usg3/l!2g5 DNA, lowering the temperature to activate DNA polymerase which copies the DNA sequences complementary to the single strands, followed by again raising the temperature above about 95 C to separate the amplified double stranded DNA. It was recognized that double-stranded DNA
should be left unamp~ified in order to increase the chances for amplification of the transcribed gene of interest without concomltant amplification of non-productive gene rearrangements. This suggested that statistically the chances of amplifying the gene of interest might be better out of the RNA pool rather than the DNA pool. However it was known that non-productive gene rearrangements also existed in the RNA pool, making the result~ from RNA amplification also unpredictable.
Fortuitously, it was found that amplification of RNA
using 3SR in conjunction with generic degenerate primer pools could work to produce amplified RNA encoding the gene of interest. The basis for 3SR is described in PNAS
`~ 20 87:1874-1878, 1990, Guatelli, et al; J. Infect. Dis ; 164:1075-81, 1991, Richman, et al; J. Infect. Dis.
16~:1066-74, 1991, Gingeras, et al. For reviews, see also Fahy, et al, In: PCR Methods and ApPlications pp.
2S-33, Cold Spring Harbor Laboratory Press, 1991;
Gingeras and Kwoh, In: Praxis der Biotechnoloaie. In Vitro Am~lification Techniques, pp. 403-429, 1992, Publ:
Carl Hanser; Gingeras, et al~, Ann. Biol. clin. 48:498-~-~ 501, 1990.

Briefly, in 3SR, a continuous series of reverse ~- transcription and RNA transcription steps replicates a nucleic acid target by means of cDNA intermediates. Two or three different enzymes work together in this system;
typically a reverse transcriptase, an RNaseH, and an RNA
polymerase are required. Alternatively, exogenous RNaseH
(i.e. from E. coli) may be omitted when reaction 1. - .

'~ ~
,~

WO94/11507 PCTtUS93/11295 ~3 a conditions are optimized (Fahy, et al., su~ra). 3SR can be conducted in a single pot at a single temperature, below 95OC, which allows the activity of the enzymes in the reaction. The temperature is chosen to optimize the activity of the specific enzymes. Currently, 3SR enzymes are optimally active in the temperature range of 37C to 42C. However, new enzymes are in development which may be optimally active at substantially higher temperatures, although the 3SR reaction will routinely be conducted below about 95C.

The RNA of interest is targeted by a primer set including a specific 3' primer and a 5' primer pool. The 3'primer encodes a T7 promoter as well as antisense for the RNA of interest. Since the constant region of each type of immunoglobulin generally has a well defined single sequence, the 3' primer can be clearly defined. Primer sequences for the 3' region of mouse immunoglobulin heavy and~l~ight chains are listed in Figure 2 (91-267*, 9l-268*, 920-004, 92-006, 92-002*).

Devising primers for the 5' end, however, is complicated - by the great variety of sequences possible for the variable region. It is not predictable which ; 25 immunoglobulin variable chain family is present in the antibody of interest. Therefore, a degenerate 5' primer pool, generic for all variable chain families of the animal species of interest, is designed to enable efficient amplification of the RNA of interest.
~`~ 30 For illustration of the basic principles, the following desoribes the desilgn a'nd synthesis of a degenerate 5' primer pool generic for mouse variable region families.
Degenerate primers were designed using the data base of Kabat et al. (In: Sequences of Proteins of Immunological ` Interest, 4th Edition (1987) U.S. Dept. Health and Human " ~ , ..,' ~ :
"
~-'~' :~

: WO94/11507 ~ 1 3 8 0 0 9 PCT/US93/11295 Ser~ices), taking into account codon degeneracies for each amino acid in the conserved sequences of the leader and FR1 regions. The degenerate pool generic for the light chain is depicted in Figure 2 (se~uences 92-099 to 92-102), for the heavy chain starting with the leader sequence in Figure 2 (sequences 92-107 and 92-108), and for the heavy chain starting with the coding region for the mature protein in Figure 2 (sequences 92-095 to 92-098 and 92-109 to 92-110).
Oligonucleotide primers were synthesized using an Applied Biosystems Incorporated DNA synthesizer using phosphoamidite chemistry. Parentheses in Figure 2 indicate a single degenerate site. Wherever nucleotide bases are indicated in parentheses in Figure 2 r those bases should be added in equal proportion at that site to j~ obtain a degenerate pool which is truly representative of all the possible variations at each site. It was recognized that, in order to increase the chances of amplification of the RNA of interest, it would be impor~ant to~have the correct 5' primer present in the pool in proportion to the 3i primer.

Howey-r, ie was also recognized that, in synthesizing 25 ~ degenerate sequences, often a base substitution is underrepresented at a certain site in spite of adding ~ equal proportions of the nucleotides to the synthesizer -~ pool at the appropriate point in the synthesis. To i ~ address this problem, equal representation of all possible substitutions at selected 3' sites of the 5' primers was forced according to the following method.

Each of the degenerate sequences was synthesized in a separate sub-pool. The following description of the synthesis of the degenerate pool for the light chain Figure 2, sequences 92-099, 92-100, 92-101, and 92-102]

, ~ .
; "
.

WO94/~1~07 ' PCT/US93/11295l' '~

3~009 illustrates the principle for synthesis of the pool.
Briefly, the oligonucleotide synthesis of sub-pool g2-099 was begun by the addition starting at the 3'end of A, then C, followed by C. Then A was specifically added at the degenerate site four nucleotides from the 3' end.
Oligonucleotide synthesis was continued with the addition of C followed by T. Then A and G were added to the reaction mix in l:l proportion for addition at the heterogeneous site 7 nucleotides from the 3'end.
' Synthesis was continued with the specific addition of A
followed by C. Then G,C, and A were added to the reaction mix in the proportion of l:l:l for addition at the heterogeneous site lO nucleotides from the end.
Oligonucleotide synthesis was continued in this,manner , until degenerate sub-pool 92-099 was complete. Each degenerate sub-pool was synthesized in this manner, resulting in four separate primer sub-pools. Then the sub-pools~were combined in equal proportion to form the ,~ light chain 5'degenerate primer pool generic for mouse light chain immunoglobulin. By this method, it was assured that the total pool would have an'equal representation of A, T, C, and G at the position four .~ .
nucleotides from the 3' end.
The rationale for forcing equal representation at the 3' ~' 25 end of the 5' primer ensures that exponential, not ~,, linear, increases are observed throughout the "' ampli~ication period of the 3~R reaction.
-, . .
Thus the degenerate primer pool generic for the mouse immunoglobulin light chain comprises a 5' primer pool '' having an equal proportion of I, II, III, and IV, wherein I is:
~ GAX1 ATX2 GTX3 CTX4 ACX5 CAX6 TCA CCA ' ', !: 35 II is:
',~ GAX1 ATX2 GTX3 CTX4 ACX5 CAX6 TCT CCA

1~: .
I
1' . WO 94/1 1507 2 1 ~ 8 0 0 9 PCr/US93/1 1295 III is:
GAXl ATX2 GTX3 CTX4 ACXs CAX6 TCC CCA
IV is:

wherein X1=T or C, X2=T or C, X3=G or C, X4=G, C, or T, X5=
A, C, or G, and X6=A or G.

:~ The above described degenerate 5' primer pool must be combined with a 3' primer in order to allow 3SR
amplification of the RNA of interest. Since the most : common type of secondary antibody is IgG, it is preferred to use a 3' primer specific for the constant region of mouse IgG. Preferably, the degenerate primer pool for the mouse IgG light chain further comprises a 3' primer ~:~ having the sequence 92-002* or-92-006 in Figure 2, said degenerate primer pool comprising one part of said 3' prLmer and~ four parts~of said 5' primer pool. This proportion~:of 1.4,~-3':pool:5' pool, facilitates the 20: ~ efficlènt~amplification~of the RNA of interest because it increases the::chances that the appropriate 5' sequence will be present in~proportion to the 3' sequence.

A~degenerate 5' primer pool generic for the mouse 25~ mmunoglobulin heavy chain, coding from the leader sequence,~may be constructed to contain~an equal proportion of I and~II, wherein A~G X1AX2 TTX3 X4GG:Xs~X6 AX7C TX8G X9TT

: ATG X1oAA TGX11 AXl2C TGG GTX13 X14TX15 CTC

wherein X1 = G or A, X2 = G or C, X3 = G or C, X4 = T or G, X5 = T or C, X6 = A or C, X7 = A or G, X8 = G or T, X9 = G
35~ or:A, X10 - G or A, Xll = G or C, Xl2 = G or C, X13 = C or T, Xl4 = T or A, X15 = T or C.

,,'` ~
.,,,,~ ~ .

2~ 3 8 n 9 For 3SR amplification of mouse IgG heavy chain, the above 5' primer pool for the heavy chain leader sequence is combined with a 3' primer specific for mouse IgG heavy chain constant region. The 3' primer has the sequence 91-267*, 9l-268*, or 92~004 in Figure 2. The total primer pool comprises one part of the 3' primer and two parts of the 5' primer pool.

:;, ' ' A degenerate 5' pri~er pool generic for the mouse immunoglobulin heavy chain, coding from the mature amino acid sequence, may be constructed to contain an equal - proportion of I, II, III, and IV, wherein I is:
CAG GTXl CAA CTX2 CAG CAX3 TCA GG
II is:

is:

~ IV is:
CAG ~GTX1 CA~ CTX2 CAG CAX3 TCG GG
wherein X1 = G, C, or A, X2 = G, C, or A, X3 = A or G.

For 3SR amplification of mouse IgG heavy chain, the above 5' primer pool for the heavy chain mature sequence is combi~ed with a 3' primer specific for mouse IgG heavy chain constant region. The 3' primer has the sequence 1-267*, 91-268*, or 92-004 in Figure 2. The total primer ~5, '~ pool comprises one part of the 3' primer and four parts .~
of the 5' primer pool.

Using the above described primer pools, target RNA's encoding the light chain and the heavy chain are amplified in two separate rounds or a single round of 3SR
~ 35~ reaction.
.'~
'--,~"~: ~
":
~,-WO94/11507 ' PCT/USg3/11295 .,; , . 23 Given the above description, it will be apparent to one of skill in the art that other techniques besides 3SR can be used to obtain amplified RNA. 'For instance, the technique known as RT/PCR may soon evolve to utilize S temperatures which do not denature double-stranded DNA, and thus RT/PCR could be substituted for 3SR to preferentially amplify RNA. Another technique known as "Q Beta Replication" is based on a phage enzyme which replicates negative strand RNA into positive strand RNA, thus also providing preferential amplification of RNA.

- For each amplfied product RNA, complementary double stranded DNA is made using primer extension or RT/PCR.
.
Once the cDNA for each chain of the antibody of interest is obtained, there are several routes to choose for making the recombinant antibody.
:. ~
Each~double stranded DNA of the 35R products can be inserted~;into a separate expression vector, both vectors can~be transfected into a host ceIl such'as E. coli, and the host cell will synthesize both immunoglobulin chains - and cause them to associate properly via di-sulfide bonds to~form a functioning antibody. Alternatively, a plasmid 25 ~ containing~both light and heavy chain genes under the regulation of separate promotors may achieve expression of~each of the antibody chains. In either'case, with a single vector or multiple vectors, the two Ig chains can then be expressed by a single host cell. Alternatively, ~ each vector can~be transfected into a separate host cell,'' each of' which produces a light or a heavy chain separately. The ~wo immunoglobul!in chains can then be ; isolated from the host cells and caused to form appropriate di-sulfide bonds by manipulating the pH
and/or urea and salt content of the medium surrounding ~ the proteins as described in U.S. Patent No 4,~16,397.

,,, ~ .

WO94tl1~07 PCT/US93/11295 ~3~oo9 In yet another alternative, portions of the sequences of the light and heavy chains can be linked to encode a single-chain antibody which is then expressed by the transfected host cell as described in Example 3 below.
~
Thus, practicing the methods of the invention, and without the use of hybridoma technology, it is possible - to obtain a recombinant monoclonal antibody of desired specificity and affinity for a predetermined antigen.
Unlike cell suspension cultùres, spleen fragment cultures retain sufficient germinal center architecture to enable memory B cell generation and somatic mutation in response to antigen. The invention is advantageous over hybridoma . ~
technology and phage dispLay technology because the ~ 15 invention allows for greatly reduced scree~ing.
i1 Moreover, the invention makes possible the preselection for: l) specific isotypes, 2) antibodies with binding affinities of desired levels,~ and 3) defined specificity as~determined by competitive analog screening. Using 20 - the methods of the invention, screening is performed on a relatively small library of antibodies which were r~ecombined-and paired by the innate processes of the responsive B cells in the germinal centers of the spleen -~ fragments. Thereafter, the nucleotide sequences encoding 25 ~ predominantly the desired antibody chains arè selectively amplified, cloned, and expressed.
,~
The following experiments are described by way of example $~ to~illustrate the methods of the invention and are not to 30~ be construed as limiting the scope of the invention.
~:

Desi~ ic for the anti-PC
; monoclonal antibodY.
35~ ~ Tl5 refers to an epitope of certain antibodies specific for the Pneumococccus coat protein (PC) which is found on ,~
~ , , :~: ' , i ., ~ ~

,, - WO94/11507 ~ l 3 8 0 0 9 PCT/US93/11295 two hybridoma cell lines ~R2-26 and R2-09) that produce anti-PC monoclonal antibodies. Nucleotide sequences for both light and heavy chains of Tl5 antibodies have been determined pre~iously. The Tl5-specific primers were selected from published sequences for the VkTl5(SlO7) light chain rearrangement and VHS107 Tl5 heavy chain. The ,~
sequences of the specific primer pair are listed in Figure 2 (heavy-chain:92-OOl*/91-267*; light chain: 92-003*/92-002*). Oligonucleotide primers were synthesized ' with an Applied Biosystems DNA Synthesizer, Model 394, using phosphoamidite chemistry.

Desian and svnthesis of deaenerate 5' ~rimer POO15 ~eneric for mouse IqG light and heavy chains. ~' Degenerate primers were designed such that members of each set would hybridize to representatives of several gene families. This approach was undertaken because ~ ' antibodies of future interest may belong to any of ~' ,~ 20 several gene families.

Sets of degenerate primers synthesized from the relatively conserved sequences of the heavy chain leader (VH~leader), 5' end of the heavy chain gene variable region (VH), or 5' end of the light chain gene variable region ~VL~ were analyzed for amplification of the Sl07 family members from total RNA isolated from the PC-specific mouse hybridoma line R'2-09. The 3'-end constant region of the light (CL) and heavy chain (CH1) genes have reiatively little sequence variation. The most common type of antibody light'chain is k and the predominant isotype for heavy'chain from secondary cells is IgGJ
Thus, the 3' end primers used for the amplification of the kappa light chain and IgGl heavy chain of anti-PC
with Tl5-specific primers in Section V.C.l were used in combination wieh the 5'-end degenerate primers. ¦~

~' WO94~ 07 PCT/US93/11295~
- 0. ., 213~9 The 5'-end heavy and light chain variable region degenerate primers were designed to contain a SalI
restriction enzyme site at their 5' ends to aid in the cloning of the genes; however, no restriction enzyme sites are present in the degenerate VH-leader primers.

Degenerate primers were designed using the data base of Kabat et al. (In: Sequences of Proteins of Immunological Interest, 4th Editio~n (1987)~U.S. Dept. Health and Human - Services), taking into account codon degeneracies for each amino acid in the conserved sequences of the leader and FRl-regions. The~rationale for the design of this degenerate pool is described above in the Description section. The degenerate pool generic for the light chain is depicted in Figure 2 (92-099 to 92-102), for the heavy chain starting with the leader sequence in Figure 2 (92-07~to~92-108), and fori~the heavy~chain starting with coding~;~region for the mature protein in Figure 2 (92-095 ~20~ to~92-098~and 92~-109;to 92-ll0).

Eà`ch~of~the~degenerate sequences was synthesized in a separate sub-pool. The following description of the synthes~is of the dogenerate pool for the light chain Figure~ a~, ~sequencès 92-099, 92-lO0, 92-lOl, and 92-102]
illùstrates the principle~ f or synthesis of degenerate pools.~ Brief1y, the oligonucleotide synthesis of sub-pool;92-O99~was begun by the addition starting at the 3'~end;~of~A~ then C~,~followed by C. Then A was 30~ spec~ifically àddèd at the deqenerate site four nucleotides from the 3' end. Oligonucleotide synthesis was continued with the add'i~tion of C foll~owed ~y T.~ Then A and ~G were added to the reaction mix in l:l proportion for addition at~the heterogeneous site 7 nucleotides from 35~ thei~3'~end.~ Synthesis~was continued with the specific addition~of A followed by C. Then G,C, and A were added ~"~

~- WO94/11~07 2 1 3 8 0 0 9 PCT/US93/1129~

to the reaction mix in the proportion of 1:1:1 for addition at the heterogeneous site 10 nucleotides from the end. Oligonucleotide synthesis was continued in this manner until degenerate sub-pool 92-099 was complete.
Each degenerate sub-pool was synthesized in this manner, resulting in four separate primer sub-pools. Then the sub-pools were combined in equal proportion to form the light chain 5'degenerate primer pool generic for mouse light chain immunoglobulin. By this method, it was ~
assured that the total pool would have an equal `
representation of A, T, C, and G at the position four nucleotides from the 3' end.
.~ , 3SR Am~lification of RNA Encodinq Anti-PC Monoclonal Antibodv. and Synthesis of Anti-PC Sinqle Chain Antibodv.
This experiment demonstrated that the degenerate primer pool described in E~ample 2 yields the same results in 3SR as primers specific for the known antibody sequence (Example 1). Moreover, an anti-PC single chain antibody was produced using the 3SR amplification products.
, Preparation of RNA from hybridoma extracts:
25~ Total RNA from two PC-specific hybridomas, R2-26 and R2-O9,~was extracted by an acid guanidinium isothocyanatelphenol-choroform extraction protocol (Chomczynski and Sacchi, Anal. Biochem., 162:156-159, 1987). Purification of poly(A) RNA from total RNA
~30 isolated from R2-26 was performed with the Dynal~
Dynabeads~ mRNA Purification Kit according to manufacturer lnstrùctions.
' , 3SR amplification:
35 ~ The basic technique of 3SR amplification is described in PNAS 87:18~4-1878, 1990, Guatelli, et al. Briefly, RNA

~. . . ~:
,;,,~

~3 was denatured in reaction mix containing primers at 65Cfor 1 minute. Two separate 3SR amplifications for the heavy and light chain genes were performed, the first using the specific primers described in Example 1, and the second using the degenerate primer pool described in Example 2.

Annealing was conducted at 42C for 1 minute, after which enzyme mix was added. Enzymes used were (1) AMV Reverse Transcriptase (Life Sciences~ (2) T7 RNA polymerase (Bartels) and ~3) E. coli RNaseH (Bartels). The mixture was incubated at 42C for 1 hour, then frozen on dry ice a~d stored at -70C.

Standard 3SR amplifications of 0.1 pmole of HIV-1 RNA
with primers 88-211 and 88-347 for the envelope region were carried out with each amplification of antibody as a positive control for the amplification reaction. cDNA
was generated with the original 3SR primers 92-(107,108), ~which hybridize to the 5' end of the heavy chain leader ~-region. The concentration of each primer in the reaction was 0.375 ~M. Following the reverse transcription reaction, pr~imer 92-004, which hybridizes to the 5' end of the heavy chain constant region~(CHl), was added to 25 ~ complete the primer pair for the PCR reaction. The concentration of primer 92-004 was 0.15 ~M.

The final con~entration of each of the T15-specific . ~ . , .
primers in the 3SR, cDNA, and PCR reactions used in the succes3ful isolation of heavy and light chain genes was 0.1 ~M, 1.0 ~M, and 0.2 ~M, respectively. The 3SR
reaction was modified ~or the degeneratè primers as follows: the concentration of each primer in the reaction was either 0.025 ~M or 0.25 ~M, and the reaction was alIowed to proceed for 1.5 hours. The PCR reaction was modified by increasing the concentration of the 5'--~: .

,~:

_ ~IL V ~J ~J V J
W O 94/11507 PC~r/US93/11295 end degenerate primers for the cDNA reaction as described below.
;~
Heavy chain, leader region: Aliquots of approximately o.9 ~g of total RNA from PC-specific hybridoma line R2-09 - were amplified by 3SR with primers 92-(107,108)/91-268 (Figure 2). Sequences for the degenerate primers 92-107 and 92-108 were obtained from Coloma et al.
(BioTechniaues, 11:152-156, 1991) and hybridize to the 5' end of the leader sequence of multiple heavy chain gene families. Specific primer 91-268 hybridizes to the s' end of the heavy chain variable re~ion (CH1)~

Heavy chain, variable region: Aliquots of approximately 0.9 ~g of total RNA from PC-specific hybridoma line R2-09 were amplified by 3SR with primers 92-(95-98)/91-268 (Figure 2). Sequences for degenerate primers 92-(95-~- 98) hybridize to the 5l end of the variable region of multiple heavy chain gene families and contain a SalI
restriction enzyme site at their 5' ends. Primer 91-268 hybridizes to the 5' end of the he~vy chain constant region ( CH1).
-~ Primers 92-109 and 92-110 hybridize to the heavy chain - variable region.
Light chain variable region: Aliquots of approximately 0.9 ~g of total RNA from PC-specific hybridoma line R2-09 were amplified by 3SR with primers 92-(99-102)/92-002 (~igure 2). Sequences for the degenerate primers 92-(99-102) hybridize to the 5' end of the variable region of multiple light chain gene families and contain a SalI
restriction enzyme site at their 5' ends~ Primer 92-002 hybridizes to the 5' end of the light chain constant region (CL) ~5 Detection of 3SR products, slot blot analysis: 3SR

;::

WO94/11507 ' P~T/US93/1129 ~3~0~9 products were detected by immobilizing three serial dilutions of each reaction on a nylon filter (Bio-Rad Zeta-Probe~) utilizing a slot blot apparatus (Schleicher and Schuell). Aliquots corresponding to 2 ~l, 0.2 ~l and 0.02 ~l of the reaction were immobolized on the blot.
Each blot included 3SR amplifications of HIV-l as a ' pcsitive control. The blots were probed with 32P-labeled -- oligonucleotides specific for the RNA of interest.
Duplicate filters were probed separately with oligonucleotides 92-008 and 92-009 (Figure 2), which are anti-sense and sense probes, respectively, homologous for a relatively conserved sequence in the J_CH1 region, specifically JH1, JH2, and JH4. Primers 92-012 and 92-013 hybridize to the 5' end of the constant region of the , light chain.

Northern analyses: Northern blot analyses were performed -~ on a NuPAGE 8% RNA Gel Kit (Novex) according to the manufacturer's instructions. Aliquots of 3SR
0~ amp~lification products were denatured by diluting each sample~ in 2xNuPAGE Urea Sample Buffer (Novex) and ~, ~
heating~for~2 min at 85C before separation. Following ,~ electrophoresis, transfer of the nucleic acids onto a nylon filter (Bio-Rad Zeta-Probe~) was carried out for 45 25, ~ - mi~n at~0.4 amps in lxNuPAGE Running Buffer (Novex) on a transfer apparatus (Hoefer, TE22). The DN~ was cross-linked to the filter by' W irradiation at 0'.125 joulesJcm. Each blot included a 3SR amplification of HIV-l as a positive control . The blots were probed with ~- 30 32P-labeled oligonucleotides specific for `' the gene of interest. Either a MsPI digest of pBR322 DNA
(New''Engla`nd Biolabs)ithat had bèen 32P-labeled or a~0.16-- l.77 Kb RNA Ladder (Gibco BRL) was included on the blot ' as molecular weight markers.The filter was probed with ~ oligonucleotides 92-009 and 92-013, which hybridize to ' the 3' end of the heavy chain variable region (JH1~ JH2, ~:
: ~ :

:

W094/1l507 - PCT/VS93/11295 J~4-CH1) and the 5' end of the light chain constant region (C~), respectively.

Further amplification by ~CR: For the T-15 specific primers, aliquots of 2 ~l of 3SR amplification reactions of hybridoma RNA extracts were further amplified by RT/PCR essentially according to the manufacturer's specifications (Perkin Elmer Cetus GeneAmp~ RNA PCR Kit).
The thermal cycle parameters with T-15 specific primers were: 1 cycle for 2 min at 94Cî 3 cycles of 1 min at 94C for denaturation, 1.5 min at 42C for primer annealing, and 2 min at 72C for elongation; 30 cycles of 1 min at 94C, 1 min at 64C, and 2 min at 72C; followed by 10 min at 72C and a 4C soak. The thermal cycle parameters for the degenerate primers were: 2 cycles of 1 min at 94C for denaturation, 1 min at 42C for primer annealing, and 2 min at 72C for elongation; 30 cycles of l min at 94C, 1 min 60C, and 2~ min at 72OC; followed by ~; 10 min at 72C and a 4 soak.

SQUthern analyses: Southern blot analyses were performed on~-5-~1 àliquots of PCR amplification products separated on~pre-cast 6% TBE polyacrylamide gels (Novex) as described (Engelhorn and Raab, BioTechni~ues, 11:594-596, 19~91). Following electrophoresis, the gel was soaked for 10 min in 0.05 M NaOH and 5 min in lxTBE. Transfer of the nucleic acids onto a nylon filter (Bio-Rad Zeta-ProbeO) was carried out for 45 min at 0.4 amps in lxTBE
buffer on a transfer apparatus (Hoefer, TE22). After the transfer, the nylon membrane was washed for 5 min each in 0.1 M NaOH and H20. The DNA was~cross-linked to the filter by W irradiation at 0.15 joules/dm . The bl!ots ~
were probed with 32P-labeled oligonuoleotides specific for the gene of interest. The blots included an Ms~I digest of pBR322 DNA (New England Biolabs) that had been P-labeled as molecular weight markers.
'~

., ~' ' .

WO94/11507 PCT/US93/1,1295~
~3y,009 32 ~ Results:
T15-sPecific primers. In slot blot and Northern analyses of the 3SR amplification products, heavy chain variable region products were detected in amplifications of R2-09 total RNA, and light chain variable region products were detected in amplifications from R2-26 poly(A)+ RNA.
Northern analyses of 3SR amplifications and agarose gel analysis of PCR re-amplifications indicated that heavy and light chain variable region products generated from T15-specific primers and R2-09 total RNA and/or R2-26 poly(A+) RNA were of the expected sizes.
~;~ Deqenerate ~rimers: The degenerate primer pools generic for mouse IgG families yielded similar results in 3SR
amplification as the the T15-specific primers. Thus these degenerate primer pools are suitable for the isolation of heavy and light chain variable region IgG
genes of unknown gene types from secondary mouse spleen fragments producing monoclonal antibodies of desired affinity and specificity for a target antigen.
Svnthesis of anti-PC sina,le chain antibody: The 3SR
amplification products of anti-PC light and heavy chains ~'!',''~"~ ` from hybridoma R2-09 were cloned into pBR~22 as described below for anti-E2 antibodies. Plasmids pSHC-5 and pSLC-2 contained heavy and light chain anti-PC clones, respectively (Figure 4). pSHC-5 and pSLC-2 were' linearized'with PstI, and the heavy and light chain inserts were joined using a two-step PCR reaction. The ~, .
first PCR reaction employed primer pairs 92-194 and 92-188, and 92-195 and 92-201 (Table 1) to amplify the heavy , and light chain inserts! respectively (Figure 4).
Primers 92-188 and 92-201 contain overlapping linker ~, sequences based on the Ganex 212 linker (Bird, R.E., et al., 1988 Science 242:423-426). ~The PCR fragments ,- 35 produced by the first PCR step were then joined by a second PCR reaction with the primer pair 92-194 and 92-,~' ' , ~:
, ~ v v u u ~J `
~-: WO94/11507 PCT/U~93/11295 l95 (Figure 4, Table l). The resulting PCR fragment contained SalI and XmaI clonLng sites at the termini (figure 4).

, f ~ W O 94/11507 PCT/US~3/11295 r~, : .

~: Table 1 ~3aoo~
; ...

~' .

Oligonucleotides used for the Synthesis of Anti-PC Single Chain Antibody ~ .
Tar~et Restliction ~2 ~3 9~188 AGAGcI~TA~cAcIAccGGAAGTAGATGAGGAGAcG~GAccGlG~l`-cc~ AS VH
92194 ACIA~GAGC~GAAGCIGGI~GAAI~GGAGGA S VH Sall 92-19; ACI'ACCCGI:;G~AC;CTCCAGCll~CC~GCA AS VL X~nal 92-201 GGI'AGI~AAGAGCI~A~GGI`AAAG~;TATI~;TGAI'C;ACI'CAGTCICCAACIT S YL nor~e S ~sense) refers to sequences that are identical to the target RNA, whereas AS (anti-sense~ refers to the sequences that are complementary to the target;RNA.
;~ZHeavy:chain variable region (VH3, light chain variable region (VL).
3SalI and SmaI sites are underlined in sequence.

, 1 ~ ~
~, ~ -.'~

, :' ,, ~ .

. ~ :RECrIFIED SHEET (RUEE 91) '~. :: ~-~ : ' ~ :~

,", WO94/11507 2 1 3 8 0 0 9 PCT/US93/11~95 Construction of pT15-l1o expression vector: Plasmid ~ pT15-110 was constructed for the expression of anti-PC sc ;'~, antibody in E. coli. In vector pT15-lO0, the gene for PC
sc antibody is under the regulation of the alkaline phosphatase (~_A) promoter region and the Bacillus thurinqiensis crY transcription terminator (Wong, H.C., et al., 198~ PNAS 83:3233-3237). The ~hoA leader sequence is used to direct secretion of the antibody.
Additionally, in pT15-100, a nucleotide sequence encoding five histidine residues is fused to the 3' end of the PC
sc antibody gene. These histidine residues function as a ~,, metal binding site (MBS) and can be used in a rapid partial purification of the antibody. A nucleotide ~, 15 sequence encoding the 13 amino acids from the carboxy,,'~ terminus of the human c-myc protein (Ramsey, G., et al., ~, 1985 Mol. Cell_Biol. 5:3610-3616) is also fused to the 5' -~ ` end~of the antibody gene. These amino acids provide an -~ immunological tag us;eful for monitoring the expression and purification in Western blots and functionality in EL/SA assays.

,~ Expression vector P15-llO was constructed as follows (Fig~re 4). The ~_A promoter,and leader sequence and ~- ~ 25 the crv terminator were obtained in plasmid pSYC 1087.
,~ Oligonucleotide 92-183 (5'GGCGCCGTCGCCCCGGGCATCACCATCATCACTAGGGATCC 3') was '~ ' inserted into the narI/BamHI sites (italics) of pSYC 1087 ,i ~ to form vector pMBS4. This resulted in the insertion of ' 30 the codons for five histldine residues as well as '~ restriction enzyme sites NarISalI, and SmaI for ~,~ , subsequent cloning steps.

i-~ The original polylinker positioned 5' to the E~_A
promoter in pSYC 1087 was removed by digestion of pMBS-4 ; with ~l~I and XbaI. The ends were repaired by Klenow DNA

~1 ' ' ~
~, : :
.
,i ~` . . .

WO94/l1507 PCT/US93/11295 ~3~ polymerase and religated to form plasmid pMBS-101.

To form pPHo 101, oligonucleotide 93-011 (5'GAACAAAAACTCATCTCAGAAGAGG~TCTGGGTGCAGTCGAC 3') was inserted into pMBS 101 which had been digested with NarI, treated with Klenow DNApolymerase and digested with SalI.
This resulted in the addition of the sequence encoding the c-myc tag.

The anti-PC sc antibody gene, constructed by a series of PCR reactions described above (Figure 4) was then inserted into the SalI/XmaI sites of pPHO 101 to form expression vector pT15-110.
' ExPression of anti-PC sc antibodY: Cell culture conditions.
E. coli strain MM294 was transformed with pT15 110 and used as a host for the expression of anti-PC 5C antibody.
Cells transformed with plasmid pPHO-101 served as a negativè control. Individual E. coli transformants were ~ ~ -grown at 30C to approximately 4-5 OD600 in phosphate medium (lx MOPS, 0.4~ glucose, 0.15% vitamin-free cas~amino acids, 10 mg/mI Bl, 100 mg/ml ampicillin) (Ne~idhart, F.C., et al., 1974 Enterobacteria. J.
Bacteriol. 119:736-747) containing 10 mM KH2PO4. This - media represses ~__A expression. The cells were washed with phosphate medium and then suspended in phosphate medium containing 0.1 mM KH2PO4 at an OD600 of ; approx~imately 0.08. E. coli cultures were grown under ~ 30 low phosphate conditions for 7 hours to achieve maximal - anti-PC sc antibody expression.

Cell lysis and antibody purification: Cell lysis was accomplished by resuspending transformed cells (20 OD600) in 1.0 ml of sonication buffer (50 mM Na phosphate, pH
~ 8.0, 300 mM NaCl, 0.25% Tween-20, 0.1 mM EGTA, 1 mM
-- ~,.
, ", ~ ' .
'''' ~ ' "~

: ~ wo g4~llso7 2 :i 3 8 0 0 9 PCT/US93/11295 phenylmethylsulfonylchloride). The cells were frozen on dry ice/ethanol, thawed and sonicated on ice (l0 cycles of l0 second bursts with l min cooling at 20 watts with a Bronson sonifier, [Danbury, CT3, Model 450). The lysed cells were centrifuged at ll,00~ rpm for 20 min at 4C.
The supernatant represented the total soluble protein.
The total protein concentration in the supernatant was measured by the Lowry method with a DC Protein Asssay Kit (BioRad, Richmond, CA).
Partial purification of anti-PC sc antibody: Anti-PC sc antibody was partially purified from cell lysates using a Nickel (Ni)-NTA resin (Qiagen, Chatsworth, CA) according to the manufacturer's recommendations. Briefly, a 50%
slurry of Ni-NTA resin (previously epuilib~ated in sonication buffer) was added to an aliquot of supernatant cell lysate and agitated for l hour at 4~C. The resin was centrifuged at 14,000 rpm and the unbound fraction -~ ;was collected. The~resin was then washed with l.0 ml aliquots of l0 to l00 mM imidazole dissolved in sonication buffer was applied to the resin in l.0 ml aliguots. Aliquots were analyzed by Western blot using monoclonal antisera specific to c-mvc (Oncogene Sciences, Uniondale, NY). ELISA assays were performed by coating each well of a 96 well microtiter plate with 50 ~l of 50 g/ml of phosphorylcholine conjugated to bovine serum albumin (PC-BSAj in 0.05 M carbonate buffer. Aliquots of E.~coli lysates or fractions eluted from the Ni-NTA resin were~added and~reacted at room temperature for 4 hours.
The sècondary antibody reaction, in which the mouse anti-c-myc binds to the anti-PC sc antibody, was allowed to incubate 4 hours at room tèmperatùre. The bound anti-c-myc antibody was visualized by incubation with alkaline phosphatase-labeled goat anti-mouse Ig antibody followed 35~ by reaction with p-nitrophenyl-phosphate (NPP). After color ~evelopment, the plates were read at 405 nm on a ~3~ S7 PCT/US93/11295' ' s Dynatech (Chantilly, VA) MR5000 microtiter plate reader.

Results: Processed (approx.32 kDA) and unprocessed (approx. 33 kDa) monomeric as well as dimeric (approx. 60 kDa) and trimeric (approx. 90 kDa) forms of sc an~i-PC
antibody were observed on Western blot analyses of whole cells lysates. However, Western blot analysis of a periplasmic space fraction extracted with Tween 20 and EGTA revealed that approximately 75% of the total sc antibody expressed by E. coli is processed and transported into the periplasmic space where it apparently is associated with the inner membranes requiring release by non-ionic detergent.
.
Partial purification of the anti-PC sc antibody was achieved by adsorption through the metal binding polyhistidine tract inserted at the 3' end of the sc antibody. Denaturation conditions of 6M urea was required for the sc antibody to bind quantitatively to the resin, otherwise the majority of the sc antibody was observed in the unbound fraction. The peak of the elution profile of sc antibody occurred at 40 mM
~ . .
imidazole.

In the quantitative ELISA assay, the fraction eluted with 40 mM imidazole from the Ni-NTA resin was the only material that demontrated quantitative binding to the immobilized phosphorylcholine. Because the undenatured anti-PC sc antibody that was present in the unbound fraction from the Ni-NTA resin did not react in the ELISA
assay, it may be condluded that the majority of the anti-- PC sc antibody made in E. coli is folded such taht it does not recognize the immobilized phosphorylcholine andjor the c-myc tag is also unavailable for the second antibody. Such improperly folded sc antibody can be denatured and refolded into a functional form as ~' ,~ ~
''~:
, ,~ ~

btt ~ U ~
WO94~11507 PCT/US93/11295 previously demonstrated for an anti-PC sc antibody (Glockshuber, R., et al., 1992 Biochem. 31:1270-1279).
The amount of the sc anti-PC antibody produced by pT15-110 in E. coli was estimated using silver-stained SDS-PAGE to be approximately 0.1% of the total cell protein.

ISOLATED FROM SECONDARY MOUSE SPLEEN FRAGME~TS
- Donor BALBlc mice were each injected twice with 100 ~g estradiol b coupled to Limulus PolvD-hemu-s hemocyanin (Hy) (gift from F; Boches) at two month intervals, as described in Linton, et al(su~ra). One to two months after the second lnjection, approximately 1-2 x lo8 whole ~- spleen cells were collected from the donor mouse. MHC
syngeneic~, Hy (carrier) primed recipient mice were ethally irradiated with 1300R,`and approximately 4 x 106 ~20 ~ donor spleen~cells~were tranferred intravenously. Within 24 hQurs~of cell transfer, the recipient's spleen was removed and dissected into 1-mm cubic fragments. Each sp~leen frangment was cultured in individual wells of a m}crotiter dish in the presence of E2-Hy for 2-3 days.
25~ Five to seven days l,ater, culture fluids from the wells were~screened~by an ELISA assay for antibodies specific for~E2~(anti-E2). When limiting numbers of B cells are ~- ~ tranferred, subsequent responses are expected to be monoc~lonal. In~this experiment, most of the microtiter ~ ~ 30 wells were positive, suggesting that responders were - ~ j likely to be polyclonal. Antibodies obtained from positive fragments were further analyzed for rèlative - ~ affinity by an inhibition ELISA assay wherein 10 5 to 10 8 molar dilutions of competing E2 were added to the ~ culture-produced antibodies during the ELISA.

~;
, ~
~ ~, WO94/11507 PCT/US93/11295i~
~3~09 40 Culture fragments which were positive for anti-E2 antibody were subjected to extraction of total nucleic acids (Stallcup, M.R., et al., (1983) J Biol Chem 258:2802-2807). This total nucleic acid was used for 3SR
amplification and subsequent cloning of Ig light and heavy chain cDNAs.

An aliquot of total RNA from each E2-specific secondary spleen fragment was amplified by 3SR as described in Example 3 with the following exceptions: the concentration of each primer in the reaction was 0.25 ~M, and the reaction was allowed to proceed for 90 minutes to 1.5 h. Two pools of generic heavy chain primers were used to amplify the anti-E2 mRNAs: 92-(107,108) (leader) or 92-109j92/110 (variable) (Figure 2). The 3'-end primer used was 91-267, which hybridizes to the CH1 region of the heavy chain mRNA. Generic light chain primers 92-,- 99~to 92-102 (Figure 2) hybridize to the 5'-end of the j lig~ht chain variable region and`are composed of four pool of degenerate oligonucleotides. Primer 92-002 hybridizes to the~C~ region.

To convert the 3SR RNA amplification products to DNA, 2 l o~ each 50 ~l 3SR amplification reaction was amplified 25~ ~ by the reverse transcript;ase (RT)/polymerase chain reaction (PCR) protocol~recommended by the manufacturer ^- ~ [Perkin Elmer Cetus, (GENE~mp~RNA kit)]. Heavy chain -~ ; olig;onucleotide primers 92-107/92/108 (leader region) and -~ 92-~109/92-llO (variable region) (Figure 2) were used with ",-. . , ~ .
~ AMV RT to synthesize the first strand of cDNA. Following the RT reaction, primer 92-140 (5'ACTAGAATTCAGTGGATAGACAGATGGGGGTG 3') which contains an inserted EcoR1 site was used to complete the PCR primer pairs. For the light chain PCR reaction, a primer pool consisting of 92-99 to 92-102 (Figùre 2) was used to create the first strand cDNA. Primer 92-012 (Fig~re 2) ,.", :
-, ,, .',~
"'`' , ~ V ~

4l !
was then used to complete the PCR primer pairs. During the RT reaction, the primer concentration was set at 1.25 mM, whereas, during the PCR reactian, the final concentration of each 5'- and 3l-end primer was 0.25 and 1.0 mM, respectively. The conditions used for the thermal cycling were: 30 cycles, each cycle consisting of 1 minute at 94 (denaturation), 1 min at 55 (annealing), 2 minutes at 72 (primer extension), followed by 10 minutes at 72 and a 4 termination step.
Northern blot analysis was performed on the 3SR
amplification products with a NuPage 8% RNA Gel Xit (Novex, San Diego, CA) according to the manufacturer's instructions. Aliquots (10~1) of 3SR amplification products were denatured by diluting each sample with an equal volume of 2x NuPage Urea Sample Buffer followed by heating for 2 minutes at 85C before e1ectrophoresis.
Following electrophoresis, transfer of the nucleic acids ,~ .
~ onto a Zeta Probe~ nylon filter (BioRad, Richmond, CA) -~ 20 was carried out for 45 minutes at 0.4 amps in lx NuPage Running Buffer on a TE22 transfer apparatus (Hoefer Scientific, San Francisco, CA). The DNA was cross-linked to the filter by W irradiation at 0.15 joules/cm2. The filters were probed with 2P-labeled oligonucleotides 92-" ~ ~ .
009 or 92-Q13 (Figure 2). An MsPI digest of pBR322 DNA
(N~ew England Biolabs, Beverly, MA) which had been pre-labeled with 32p was included on each blot as molecular weight markers.
",, . ~ - .
Southern blot analyses were performed on aliquots (1-5~1) of PCR amplification products separated on pre-cast 6%
~- TBE polyacrylamide geis (Novex, San Diego, CA).~- Following electrophoresis, the gel was soaked for 10 minutes in 0.05 M NaOH and 5 minutes in lx TBE. Transfer of the nucleic acids onto Zeta Probe~ nylon filters was carried out for 45 minutes at 0.4 amps in lx TBE buffer ", i' ,"~
:~

WO94~11507 PCT/US93/11295' ~3~o9 on a TE22 transfer apparatus. After the transfer, the nylon membrane was washed for 5 minutes each in 0.l M
NaOH and H2O. The DNA was cross-linked to the filter by W irradiation at 0.15 joules/cm . The filters were probed with 32P-labeled oligonucleotides 92-009 o~ 92-013 (Figure 2).

Cloninq. nucleic acid sequencinq, and computer analysis of anti-E2 heavy and light chain clones.
Because generic primers were used to amplify heavy and light chain immunoglobulin mRNAs, the amplification products were cloned and the clones were individually analyzed by sequencing. A l0 ~l aliquot of the anti-E2 leader heavy chain amplification reaction was treated with Klenow DNA polymerase and then digested with EcoRI.
The blunt end/EcoRI-cleaved fragments were gel purified - and ligated into pUCl8 which has been previously digested with SmaI and EcoRI and treated with calf alkaline phosphatase (CAP). The anti-E2 heavy chain variable ~ region amplification reaction (l0 ~l) was digested with - SalI and~EcoRI because both restriction endonuclease sites were engineered into primers 92-(l09-ll0) and 92-140, respectively, for cloning purposes. The digested PCR fragments were gel purified and ligated into pUCl8 which had been previously digested with SalI and EcoRI
; and treated with CAP. The anti-E2 light chain amplification reaction (l0 ~l) was treated with Klenow ~ .
-- DNA polymerase and digested with SalI. The PCR fragments , were gel purified and ligated into SmaI and SalI digested 30~ pUCl8 or pBR322 which had been treated with CAP.
Dilutions of each ligation were used to transform competent Escherichia coli MCl061 (recA dr HBl0l (recA
-~ cells).
.~ .~, :
: .
Nucleotide sequence analyses of pUCl8 and p8R322 heavy ~i ~ and light chain clones were performed by the modified "~ ~

. .

~15~VU~ ~
WO94/11507 PCTtUS93/11295 dideoxy nucleotide chain termination method described by Hsiao, et al (Nucleic Acid Res. l9:2787). Computer analysis of the derived sequence was performed with MacVector (IBI, New Haven, CT) and Line-Up from UWGCG.
.~
Cloning and Sequence Analysis: During the initial screening step, cells in a splenic fragment, D1A6, were identified as producing antibody specific for estradiol with an affinity constant of approximately 107 liters per mole. Total nucleic acid from cells of this fragment was amplified by 3SR with the light and heavy chain generic primers shown in Figure 2. The 3SR RNA products of the heavy and light chain amplifications were converted to DNA, digested with the appropriate restriction endonucleases and cloned into either pUC 18 or l9.
.

A total of 39 heavy and 36 light chain p~oductive clones : derived from fragment D1A6 were sequenced. Comparative sequence analysis of the clones using the UWGCG LineUP
program revealed the presenoe of four alleles each for the heavy and light chain groups (Table 2). The nucleotide sequence representative of clones from each allelic group is presented in Figure 5. The heavy chain clones from groups 2, 3 and 4 exhibited 81-86% similarity with the J558 VH heavy chain family (Kabat, et al., sura;
Brodeur, P.H., et al., J ExP-Med 168:2261-2278), whereas :, clones from group l were marginally similar to the Sl07 ~ (79.7%) VH family~ However, comparison of the sequences `: ~ of clones of group 1 to sequences in GENBANK revealed a 9~% similarity to a mouse anti-DNA rearranged heavy chain variable region (Kofler, R., et ?1-, J Clin Invest 82:852-860). Only in clones from groups!2 and 4 coùld the family of the DN minigenes be unambiguously determined ~: as being ~rom DFLl6.1 and DST4 minigenes, respectively . 35 (Rabat, et al., suPra). The JH minigenes utilized were JH
and ~. The light chain clones from fragment D1A6 also I

WO94/11507 PCT/~S93/11295 ~3 a~ consisted of four V~ alleles (Figure SB). Clones from groups 2 and 4 were highly similar (96-97%) to the lgf28 V~ family (Strohol, R., et al., Immunoqenetics 30:475-493), whereas clones from groups l and 3 were similar to 5 the V~4/5 (92%) and V~ lO (88%) families, respectively.
Light chain clones used the Jx5 minigene while clones from group l used the J~4 minigene.

Unlike the variable region degenerate primers, the leader ~ lO region primers allowed for unambiguous determination of the sequences present at the amino terminus of the VH
^ minigene. In group l clones, derived from the leader region primers, the nucleotide sequences predicted amino acids Val-5, Glu-6 and Thr-i in place of Gln-5, Gln-6, . 15 and Ser-7 as predicted by using the generic variable -j region primers. Likewise, in ~ ' , ~

,~ ~
.
:~ :
~ , ~:

' .

? r.~ "" ~ """~ ", ,~ " " ,~ " ~

. ; WO94/11507 2138009 PCI/US93/11295 Heavy and Ligh~ Chain Clones Obtained from Spleen Fragment Culture DlA6 `' ~ ~ ~ Minigene ~amilies # of Clones V D J
~- Allelec _ _Analyzrd~ hom~lo~
.: : VH Clones _ _ . ~ 1 11 S107 DSP 2.3, 4, 6JH2 (79) 2 6 J558 DFL 16.1 ~H3 (81) 9 3 15 J558 DSP 2.3, 7 JH43 (~6) 4 ? J558 ~ DST 4 JH3 , ~

Vl ~lones _ _ ~
16 V~ 4/5 - J~
~ . ~ (92) ! ,~, ~ 2~ 10 V~ /28 - JK5 3 6 VK 10 1*

3 VICI9/28 ~JICS

:-, '.',' ~: ' :
;~ '~

i~

;~ . ':,. ..
,,', ~
'~"

W094/11507 PCT/US93/11295 - ', ~3a~o9 group 4, clones derived from the leader primers contained , the codon for the amino acid Leu at position 3 in place of Lys as predicted by the variable region primers.

The clones sequenced in each of the four groups of light and heavy chain variable regions were productive, ` indicating that more than one memory B cell colonized the . splenic fragment D1A6. To avoid obtaining multiple sequences, limiting dilution of transferred lymphocytes ' 10 can be performed to obtain only one colonizing antigen ' responsive B cell. However, even in this case with '~ multiple heavy and light chain alleles represented, it is ~ relatively straightfo~ward to join together the four,-~ light and heavy chain clones into all 16 possible anti-E2 ,,:; 15 antibody combinations. One of the combinations will ~- reflect the original VH and V~ pairings. Even in this ~- case, the need to create and screen larger phage display libraries is avoided.

Produ~tion of recompinant monoclonal antibodY aqainst estradiol.
RNA~encoding heavy and light chains of anti-estradiol, as ~ produced in Example 4, is converted to cDNA. The cDNA, -,~ 25~ when conve~ted to double stranded DNA, is incorporated into expression vectors, which are then transfected into E. coli host cells. The host cells are induced to express the heavy chain and the light chain recombinant proteins for the anti-estradiol antibody. The heavy and -~ 30 light chains associate and form the appropriate di-- I sulfide,bridgesltolf~r,m,the compl,ete anti-estradiol ;~ antibody. The antibody is isolated and purified from the !~ ;'~ , host cell debris. The resultin~ recombinant monoclonal - antibody possesses the desired specificity and affinity for estradiol as identified in the asssays of the ~; positive spleen fragment culture in Example 4.

,`:: ~ . ..
'~
.~.
'~' ' :~

Claims (44)

We claim:

1. A method for obtaining RNA encoding a predetermined chain of a desired antibody against a specific antigen, comprising:
a) isolating a B cell population from a first mammal, said B cell population containing responsive B
cells, b) injecting an amount of said B cell population into a second mammal, said second mammal having no viable endogenous B cells, c) maintaining said second mammal for sufficient time to allow said injected B cells to colonize the spleen of said second mammal, d) removing the spleen of said second mammal, e) cutting said spleen into fragments of a predetermined size, said size being sufficiently small so that essentially each fragment contains no more than one responsive B cell, f) maintaining each fragment in a separate culture, each culture containing a culture medium, g) contacting each fragment with said specific antigen, h) maintaining each fragment in culture for sufficient time to allow proliferation of responsive B
cells within each fragment and antibody production by at least some of the responsive B cells, i) assaying the culture medium of each well for content of desired antibody, thereby identifying a positive fragment containing antibody-producing B cells, j) extracting heterogeneous RNA from said positive fragment, said heterogeneous RNA containing a first target RNA and a second target RNA, said first target RNA
encoding the light chain of said desired antibody, and said second target RNA encoding the heavy chain of said desired antibody, k) contacting said heterogeneous RNA with a first degenerate primer pool or with a second degenerate primer pool, said first degenerate primer pool containing primer nucleotide sequences complementary to at least a portion of said first target RNA, said second degenerate primer pool containing primer nucleotide sequences to at least a portion of said second target RNA, said first or second primer nucleotide sequences hybridizing to said first or second target RNA respectively, 1) amplifying said first target RNA or said second target RNA using isothermal self-sustained sequence replication, thereby obtaining an amplified RNA
consisting essentially of said first target RNA or said second target RNA.

2. The method of claim 1 wherein step (k) comprises contacting said heterogeneous RNA with said first degenerate primer pool, said first primer nucleotide sequences hybridizing with said first target RNA, and step (1) comprises amplifying said first target RNA using isothermal self-sustained sequence replication, thereby obtaining an amplified RNA consisting essentially of said first target RNA encoding the light chain of said desired antibody.

3. The method of claim 1 wherein step (k) comprises contacting said heterogeneous RNA with said second degenerate primer pool, said second primer nucleotide sequences hybridizing with said second target RNA, and step (1) comprises amplifying said second target RNA
using isothermal self-sustained sequence replication, thereby obtaining an amplified RNA consisting essentially of said second target RNA encoding the heavy chain of said desired antibody.

4. The method of claim 1 wherein, prior to step (a), said first mammal is injected with said antigen.

5. The method of claim 1 further comprising injecting said second mammal with at least a portion of said antigen, thereby stimulating T helper cell function in said second mammal.

6. The method of claim 1 wherein said first and said second mammals are congeneic mice.

7. The method of claim 1 wherein said first and said second mammals are members of different species.

8. The method of claim 6 wherein said amount of injected B cells is sufficiently low such that the second mammal's spleen will be on the average colonized by not more than one responsive B cell in each 1mm3 segment of spleen tissue.

9. The method of claim 8 wherein said fragment size is approximately 1mm3.

10. The method of claim 8 wherein steps (g) through (i) are repeated at least once.

11. The method of claim 2 wherein said first mammal is a mouse, said first target RNA encodes a mouse immunoglobulin light chain, and said generic degenerate primer pool comprises a 5' primer pool having an equal proportion of I, II, III, and IV, wherein I is:

II is:

III is:

IV is:

wherein X1=T or C, X2=T or C, X3=G or C, X4=G, C, or T, X5=
A, C, or G, and X6=A or G.

12. The method of claim 11 wherein said immunoglobulin is an IgG, and said generic degenerate primer pool further comprises a 3' primer specific for mouse IgG
light chain constant region, said 3' primer having the sequence designated 92-002* in Figure 2, said degenerate primer pool comprising one parts of said 3' primer and four parts of said 5' primer pool.

13. The method of claim 11 wherein said immunoglobulin is an IgG, and said generic degenerate primer pool further comprises a 3' primer specific for mouse IgG
light chain constant region, said 3' primer having the sequence designated 92-006 in Figure 2, said degenerate primer pool comprising one part of said 3' primer and four parts of said 5' primer pool.

14. The method of claim 3 wherein said first mammal is a mouse, said second target RNA encodes a mouse immunoglobulin heavy chain and said generic degenerate primer pool comprises a 5' primer pool having an equal proportion of I and II, wherein I is:

II is:

wherein X1 = G or A, X2 = G or C, X3 = G or C, X4 = T or G, X5 = T or C, X6 = A or C, X7 - A or G, X8 = G or T, X9 = G
or A, X10 = G or A, X11 = G or C, X12 = G or C, X13 = C or T, X14 = T or A, X15 = T or C.

15. The method of claim 14 wherein said immunoglobulin is an IgG, and said generic degenerate primer pool further comprises a 3' primer specific for mouse IgG
heavy chain constant region, said 3' primer having the sequence designated 91-267* in Figure 2, said degenerate primer pool comprising one part of said 3' primer and two parts of said 5' primer pool.

16. The method of claim 14 wherein said immunoglobulin is an IgG, and said generic degenerate primer pool further comprises a 3' primer specific for mouse IgG
heavy chain constant region, said 3' primer having the sequence designated 91-268* in Figure 2, said degenerate primer pool comprising one part of said 3' primer and two parts of said 5' primer pool.

17. The method of claim 14 wherein said immunoglobulin is an IgG, and said generic degenerate primer pool further comprises a 3' primer specific for mouse IgG
heavy chain constant region, said 3' primer having the sequence designated 92-004 in Figure 2, said degenerate primer pool comprising one part of said 3' primer and two parts of said 5' primer pool.

18. The method of claim 3 wherein said first mammal is a mouse, said second target RNA encodes a mouse heavy chain and said generic degenerate primer pool comprises a S' primer pool having an equal proportion of I, II, III, and IV, wherein I is:

II is:
CAG GTXt CAA CTX2 CAG CAX3 TCT GG
III is:
CAG GTX1.CAA CTX2 CAG CAX3 TCC GG
IV is:

wherein X1 = G, C, or A, X2 = G, C, or A, X3 = A or G.

19. The method of claim 18 wherein said immunoglobulin is an IgG, and said generic degenerate primer pool further comprises a 3' primer specific for mouse IgG
heavy chain constant region, said 3' primer having the sequence designated 91-267* in Figure 2, said degenerate primer pool comprising one part of said 3' primer and four parts of said 5' primer pool.

20. The method of claim 18 wherein said immunoglobulin is an IgG, and said generic degenerate primer pool further comprises a 3' primer specific for mouse IgG
heavy chain constant region, said 3' primer having the sequence designated 91-268* in Figure 2, said degenerate primer pool comprising one part of said 3' primer and four parts of said 5' primer pool.

21. The method of claim 18 wherein said immunoglobulin is an IgG, and said generic degenerate primer pool further comprises a 3' primer specific for mouse IgG
heavy chain constant region, said 3' primer having the sequence designated 92-004 in Figure 2, said degenerate primer pool comprising one part of said 3' primer and four parts of said 5' primer pool.

22. A method for producing a recombinant protein having a desired binding affinity and specificity for a specific antigen, comprising:
a) isolating a B cell population from a first mammal, said B cell population containing responsive B
cells, b) injecting an amount of said B cell population into a second mammal, said second mammal having essentially all endogenous B cells destroyed, c) maintaining said second mammal for sufficient time to allow said injected B cells to colonize the spleen of said second mammal, d) removing the spleen of said second mammal, e) cutting said spleen into fragments of a predetermined size, said size being sufficiently small so that essentially each fragment contains no more than one responsive B cell, f) maintaining each fragment in a separate culture, each culture containing a culture medium, g) contacting each fragment with said specific antigen, h) maintaining each fragment in culture for sufficient time to allow proliferation of responsive B
cells within each fragment and antibody production by at least some of the responsive B cells, i) assaying the culture medium of each well for content of desired antibody, thereby identifying a positive fragment containing antibody-producing B cells, j) extracting heterogeneous RNA from said positive fragment, said heterogeneous RNA containing a first target RNA and a second target RNA, said first target RNA
encoding the light chain of said desired antibody, and said second target RNA encoding the heavy chain of said desired antibody, k) contacting said heterogeneous RNA with a first degenerate primer pool, said first primer pool containing nucleotide sequences complementary to at least a portion of said first target RNA, l) amplifying said first target RNA using isothermal self-sustained sequence replication, thereby obtaining an amplified RNA consisting essentially of said first target RNA, m) synthesizing a first cDNA corresponding to said first target RNA, n) contacting said heterogeneous RNA from step (j) with a second degenerate primer pool, said second primer pool containing nucleotide sequences complementary to at least a portion of said second target RNA, o) synthesizing a second cDNA corresponding to said second target RNA, p) constructing at least one expression vector containing at least a portion of said first cDNA and at least a portion of said second cDNA, q) delivering said vector into host cells, r) causing said host cells to express at least one recombinant protein encoded by said vector, and s) isolating said recombinant protein.

23. The method of claim 22 wherein, prior to step (a), said first mammal is injected with said antigen.

24. The method of claim 22 further comprising injecting said second mammal with at least a portion of said antigen, thereby stimulating T helper cell function in said second mammal.

25. The method of claim 22 wherein said first and said second mammals are congeneic mice.

26. The method of claim 22 wherein said first and said second mammals are members of different species.

27. The method of claim 25 wherein said amount of injected B cells is sufficiently low such that the second mammal's spleen will be on the average colonized by not more than one responsive B cell in each 1mm3 segment of spleen tissue.

28. The method of claim 27 wherein said fragment size is approximately 1mm3.

29. The method of claim 22 wherein steps (g) through (i) are repeated at least once.

30. The method of claim 22 wherein said first mammal is a mouse, said first target RNA encodes a mouse immunoglobulin light chain, and said first degenerate primer pool comprises a 5' primer pool having an equal proportion of I, II, III, and IV, wherein I is:

II is:

III is:

IV is:

wherein X1=T or C, X2=T or C, X3=G or C, X4=G, C, or T, X5=
A, C, or G, and X6=A or G.

31. The method of claim 30 wherein said immunoglobulin is an IgG, and said first degenerate primer pool further comprises a 3' primer specific for mouse IgG light chain constant region, said 3' primer having the sequence designated 92-002* in Figure 2, said degenerate primer pool comprising one parts of said 3' primer and four parts of said 5' primer pool.

32. The method of claim 30 wherein said immunoglobulin is an IgG, and said first degenerate primer pool further comprises a 3' primer specific for mouse IgG light chain constant region, said 3' primer having the sequence designated 92-006 in Figure 2, said degenerate primer pool comprising one part of said 3' primer and four parts of said 5' primer pool.

33. The method of claim 22 wherein said first mammal is a mouse, said second target RNA encodes a mouse immunoglobulin heavy chain and said second degenerate primer pool comprises a 5' primer pool having an equal proportion of I and II, wherein I is:

II is:

wherein X1 = G or A, X2 = G or C, X3 = G or C, X4 = T or G, X5 = T or C, X6 = A or C, X7 = A or G, X8 = G or T, X9 = G
or A, X10 = G or A, X11 = G or C, X12 = G or C, X13 = C or T, X14 = T or A, X15 = T or C.

34. The method of claim 33 wherein said immunoglobulin is an IgG, and said second degenerate primer pool further comprises a 3' primer specific for mouse IgG heavy chain constant region, said 3' primer having the sequence designated 91-267* in Figure 2, said degenerate primer pool comprising one part of said 3' primer and two parts of said 5' primer pool.

35. The method of claim 33 wherein said immunoglobulin is an IgG, and said second degenerate primer pool further comprises a 3' primer specific for mouse IgG heavy chain constant region, said 3' primer having the sequence designated 91-268* in Figure 2, said degenerate primer pool comprising one part of said 3' primer and two parts of said 5' primer pool.

36. The method of claim 33 wherein said immunoglobulin is an IgG, and said second degenerate primer pool further comprises a 3' primer specific for mouse IgG heavy chain constant region, said 3' primer having the sequence designated 92-004 in Figure 2, said degenerate primer pool comprising one part of said 3' primer and two parts of said 5' primer pool.

37. The method of claim 22 wherein said first mammal is a mouse, said second target RNA encodes a mouse heavy chain and said second degenerate primer pool comprises a 5' primer pool having an equal proportion of I, II, III, and IV, wherein I is:

II is:

III is:

IV is:

wherein X1 = G, C, or A, X2 = G, C, or A, X3 = A or G.

38. The method of claim 37 wherein said immunoglobulin is an IgG, and said second degenerate primer pool further comprises a 3' primer specific for mouse IgG heavy chain constant region, said 3' primer having the sequence designated 91-267* in Figure 2, said second degenerate primer pool comprising one part of said 3' primer and four parts of said 5' primer pool.

39. The method of claim 37 wherein said immunoglobulin is an IgG, and said second degenerate primer pool further comprises a 3' primer specific for mouse IgG heavy chain constant region, said 3' primer having the sequence designated 91-268* in Figure 2, said second degenerate primer pool comprising one part of said 3' primer and four parts of said 5' primer pool.

40. The method of claim 37 wherein said immunoglobulin is an IgG, and said second degenerate primer pool further comprises a 3' primer specific for mouse IgG heavy chain constant region, said 3' primer having the sequence designated 92-004 in Figure 2, said second degenerate primer pool comprising one part of said 3' primer and four parts of said 5' primer pool.

41. A method for making a total pool of generic degenerate 5' primers for the efficient amplification of RNA encoding a predetermined immunoglobulin chain from a population of heterogeneous RNA, comprising a) identifying areas of sequence heterogeneity among all isotypes of said chain, b) cataloging each nucleic acid substitution at each heterogeneous site, c) selecting a 5' region comprising 10 to 30 nucleotides, said region encoding a leader sequence or a 5' coding region for said chain, d) selecting from within said 5' region a first annealing sequence, said first annealing sequence comprising 6 to 9 nucleotides at the 3' and of said 5' region, e) identifying each nucleotide substitution catalogued for each substituted site within said first annealing sequence, f) synthesizing a separate pool of degenerate primers for each substitution at each substituted site within said first annealing sequence, such that each pool contains an equal representation of nucleic acid substitutions for that site, and g) combining said separate pools into a total pool such that said total pool contains an equal representation of nucleotide substitutions at each substituted site within said first annealing sequence.

42. A pool of degenerate 5' primers, to be used in conjunction with a 3' primer, for the efficient amplification of RNA encoding a mouse immunoglobulin light chain from a population of heterogeneous RNA, said pool comprising an equal proportion of I, II, III, and IV, wherein I is:

II is:

III is:

IV is:

wherein X1=T or C, X2=T or C, X3=G or C, X4=G, C, or T, X5=
A, C, or G, and X6=A or G.

43. A pool of degenerate 5' primers, to be used in conjunction with a 3' primer, for the efficient amplification of RNA encoding a mouse immunoglobulin heavy chain from a population of heterogeneous RNA, said pool having an equal proportion of I and II, wherein I is:

II is:

wherein X1 = G or A, X2 = G or C, X3 = G or C, X4 = T or G, X5 = T or C, X6 = A or C, X7 = A or G, X8 = G or T, X9 = G
or A, X10 = G or A, X11 = G or C, X12 = G or C, X13 = C or T, X14 = T or A, X15 = T or C.

44. A pool of degenerate 5' primers, to be used in conjunction with a 3' primer, for the efficient amplification of RNA encoding a mouse immunoglobulin heavy chain from a population of heterogeneous RNA, said pool having an equal proportion of I, II, III, and IV, wherein I is:

II is:

III is:

IV is:

wherein X1 = G, C, or A, X2 = G, C, or A, X3 = A or G.