CA2300804A1 - Monoclonal antibodies that bind testosterone - Google Patents

Abstract

The present invention provides monoclonal antibodies and derivatives thereof suitable for use in immunodiagnostic assay of testosterone. In particular, such reagents are disclosed having affinity for testosterone of at least about 109 M-1 and exhibiting less than 0.05 % cross-reactivity with dehydroepiandrosterone sulphate (DHEAS), less than 10 % cross-reactivity with 5.alpha.-dihydrotestosterone (5.alpha.-DHT) and less than 5 % cross-reactivity with androstenedione determined on the basis of ED50 values.

Description

MONOCLONAL ANTIBODIES THAT BIND TESTOSTERONE
This invention relates to monoclonal antibodies and antibody engineering technology. More particularly, the present invention relates to monoclonal antibodies and derivatives thereof that bind testosterone with high affinity and specificity. The present invention also relates to processes for making and engineering such testosterone-binding monoclonal antibodies and to methods for using these antibodies and derivatives thereof in the field of inlmunodiagnostics enabling qualitative and quantitative determination of testosterone in biological samples.
1o Backtzround of the invention Determination of serum testosterone (TES) is important in the evaluation of hirsutism, virilism, infertility and other conditions associated with hyperandrogenism in women such as polycystic ovarian syndrome (Wheeler 1994, Wilke and Utley 1987, Osborn and Yannone, 1973). In men is TES levels are used to diagnose reasons for delayed or premature puberty, hypogonadism, impotence and problems in spermatogenesis (Wang and Swerloff 1992, Ismail et al., 1986, Wu et al., 1981, Odel and Swerdloff, 1978). It has also been used to monitor hormone-replacement therapy (Ismail et al., 1986) and the treatment of patients with congenital adrenal hyperplasia (Korth-Schutz et al., 1978). Measurement of serum TES levels is also important to control the 2o forbidden use of steroid hormones, e.g. TES, to increse the performance of athletes, both female and male, in sports.
The great number of closely related steroid structures and significant variation in their relative concentrations even between normal healthy individuals necessitates great qualitative 2s performance of any measurement technique. This, together with high conservation of steroid structures in the animal kingdom, has made it extremely difficult to produce monoclonal anti-steroid antibodies with sufficiently high specificity and affinity for clinically useful routine immunoassays. The majority of commercially available diagnostic test kits for steroid hormones utilise polyclonal antibodies, not because of optimal performance and product quality, but 3o because of apparent lack of monoclonal antibodies good enough to fulfil the clinical requirements. Even in a good antiserum, over 90% of the Ig molecules have little or no affinity for the antigen, and the "specific antibodies" themselves represent a whole spectrum of molecules with different affinities directed against different determinants on the antigen. The large amount of non-specific relative to antigen-specific Ig in an antiserum means that unspecific binding in any given immunological test may be uncomfortably high.
The continuous supply of good polyclonal antibody reagents of uniform quality is a severe problem for the immunodiagnostic industry and requires continuous immunisation of large numbers of laboratory animals. These drawbacks of the currently used polyclonal anti-testosterone antibodies have motivated the search for more specific reagents for quantitative TES
determination.
In contrast to polyclonal antiserum, all antibody molecules produced by a given hybridoma are ~o identical having identical structure and binding properties. In vitro cell culture production of these monoclonal antibodies provides an unending supply of a single standard material. Today, monoclonal antibodies and binding fragments can also be produced by recombinant methods.
Microbial expression offers the means for efficient production and rapid engineering of antibody fragments for different specific applications (Better et al., 1988, Skerra and Pliickthun is 1988). The possibility of displaying active antibody fragments on the surface of a fllamentous bacteriophage provides a powerful tool for selection of specific or improved binders from a heterogeneous mixture of antibody fragments (McCafferty et al., 1990, Barbas et al., 1991, Marks et al., 1991 ). In the absence of detailed structurai data of the interactions involved in antigen or hapten binding, the in vitro maturation strategy, where an initially selected antibody 2o fragment is subjected to mutagenesis and improved variants are selected from large mutant libraries, has proven to be a very ei~cient method. Random mutagenesis of the whole variable regions is possible by a number of different approaches such as chemical mutagenesis (Myers et al., 1985), polymerise-induced mutagenesis (Leung et al., 1989) and in vivo mutagenesis using mutator strains ofE.coli (Schaaper, 1988). Focused mutagenesis, in which several residues are z s targeted, allows all possible mutations in a defined region to be explored. For antibodies, this strategy involves targeting all or some of the three complementarity deterrnirung regions (CDRs) present both in the light and heavy chain and residues immediately flanking these regions to mutagenesis since these most likely have effect on the binding properties (Berek and Milstein, 1987). Numerous recent publications describe successful use of random mutagenesis 30 of CDRs or of whole variable domain encoding gene regions of the heavy and light chain combined to phage display library selection to isolate antibodies having greatly improved affinity towards different protein antigens (Barbas et al., 1994, Yang et al., 1995, Schier et al., WO 99/12974 ~ PCT/FI98/00689 1996a,b,c), haptens or carbohydrates (Hawkins et al., 1992, Deng et al., 1995, Short et al., 1995, Yelton et al., 1995).
Interactions of the CDR3 hypervariable loops of the light and heavy chains with small antigens s are known to dominate the overall binding interactions (Wilson and Sta~eld, 1994). The molecular basis of steroid antibody interactions has been studied in detail by determining the 3D
structures of a progesterone binding antibody (DB3) in its free and complex form and with different high- affinity cross-reactive progesterone derivatives (Arevalo et al., 1993a,b, Arevalo et al., 1994). The progesterone molecule binding site is a hydrophobic pocket in which binding ~o specificity is determined by the van der Waals interactions and the few hydrogen bonds formed with the Iigand (Arevalo et al., 1993a). The high-affinity binding mode of a number of progesterone derivatives is due to two alternative binding sites for the A
ring of the steroid skeleton (Arevalo et al., 1993b, 1994). The restricted number or complete lack of direct interactions between antibodies and small, hydrophobic and rigid ligand molecules, is further ~s demonstrated by the 3D structure of the anti-digoxin antibody (26-10) in the complex form (Jeffrey et al., 1993). The high-affinity binding mode to digoxin (1 x 101° M'1 ) is solely determined by hydrophobic shape complementarity, no direct interactions are formed between the antibody and the ligand (Jeffrey et al., 1993, Schildbach et al., 1994).
The limited number of the direct interactions in the combining sites of the steroid hormone binding antibodies 2 o makes the improvement of these binding interactions highly demanding.
Another possible approach to developing high-affinity, high-specificity steroid hormone binding antibodies in vitro has been published recently. Pope et al. (1996) isolated from a large, naive human antibody phage display library a single chain antibody (scFv) binding oestradiol with 2s relatively high specificity. The affinity of this antibody to oestradiol was 3.7 nM and it had over 1000-fold selectivity in binding to other steroids. However, the performance of this scFv antibody in a clinical oestradiol immunoassay was not evaluated in the publication.
The present invention makes use of the unique property of monoclonal antibodies to bind their 3o antigens with high affinity and specificity. Monoclonal antibodies binding TES with high affnity and fine specificity were discovered in a stepwise process of engineering anti-testosterone antibody originally derived from a hybridoma cell line by random mutagenesis of the CDRs and phage display selection. We have previously published preliminary data on the binding WO 99/12974 , PCT/FI98/00689 properties of some moderately improved anti-testosterone Fab fragments selected at the early stage of this work from the randomised CDR-libraries of the TES-binding 3-C4Fs Fab fragment.
(Abstract: Fine tuning of an antibody binding site by mutagenesis and phage display, Exploring and Exploiting Antibody and Ig Superfamily Combining Sites, Keystone Symposia on Molecular and Cellular Biology, Taos, New Mexico February 22-28, 1996 and Seventh Annual International Conference on Antibody Engineering, December 2-4, 1996, Coronado, CA; Abstract: Can man improve evolution ? In vitro amity and specificity maturation of an antibody binding site, Kemian P~ivat, November 12-14, 1996, Helsinki; also poster presentation at INPEC-97, International Network of Protein Engineering Centres Annual io Meeting, Lillenhammer, Norway, January 26-28, 1997). These Fab fragments, with too low specificity and affinity for diagnostic immunoassay, were selected using panning methods still under optimization and also the sequence diversity of the CDR mutant libraries, especially the CDR3 libraries, was improved upon later. These Fab fragments are described in greater detail in Examples 1,2 and 3. It is shown in specific examples of the present invention that the is monoclonal antibodies and derivatives thereof produced by genetic engineering techniques bind TES with high affinity and specificity and can be successfully used as reagents in clinical immunoassays designed for the quantitative measurement of TES. To our knowledge, monoclonal antibody fragments herein described are the first shown to fulfil the tight qualitative requirements of a diagnostic TES immunoassay for clinical use. Because of these favourable 2o effects, new, economical and high-sensitivity diagnostic assay methods ofuniform quality can be set up.
Obiects of the Invention 2s One object of the present invention is to provide monoclonal antibodies or fragments or other derivatives thereof, that bind testosterone with affinity and specificity high enough to allow qualitative and quantitative measurement of testosterone in clinical samples.
The antibodies of the present invention demonstrate an affinity of at least about 109 M-' for testosterone, less than 0.05% cross-reactivity with dehydroepiandrosterone sulfate (DHEAS), less than 10% cross-so reactivity with Sa-dihydrotestosterone (Sa-DHT) and less than 5% cross-reactivity with androstenedione. Peferably, such antibodies demonstrate less than 0.01% cross-reactivity with DHEAS.
.. __ It is another object of the present invention to provide cDNA clones encoding testosterone-specific antibody chains, as well as constructs and methods for expression of such clones to produce testosterone-binding monoclonal antibodies or fragments or other derivatives thereof.
It is a further object of this invention to provide methods of using such testosterone-binding monoclonal antibodies or fragments or other derivatives thereof or combinations of them for qualitative and quantitative measurement of testosterone in clinical samples.
Other objects, features and advantages of the present invention will become apparent from the to following figures and detailed description. It should be understood, however, that the detailed description and the specific examples, while indicating preferred embodiments of the invention, are given by way of illustration only, since various changes and modifications within the spirit and scope of the invention will become apparent to those skilled in the art from this detailed description.
Brief Description of the Figures The figures of the constructions are not in scale.
2 o Fig. 1. shows the schematic presentation of an intact murine IgGI subclass antibody, Fab fragment and single chain antibody (scFv). The antigen binding site is indicated by a triangle.
Fig. 2. shows the deduced amino acid sequence of the light chain variable region of the 3-C4F5 antibody. CDRs are underlined. Numbering is according to Kabat (Kabat et al., 1991 ).
Fig. 3. shows the deduced amino acid sequence of the heavy chain variable region of the 3-C4F5 antibody. CDRs are underlined. Numbering is according to Kabat {Kabat et al., 1991).
Fig. 4. shows the schematic presentation of the E.coli expression vector pKKtac used for the production of soluble TES-binding Fab and scFv fragments.
Fig. 5. shows schematically the competitive specificity panning procedure.
Fig. 6. shows the deduced amino acid sequences of the light and heavy chain CDR3 loops of 3o the wild-type and selected mutant clones (numbering is according to Kabat et al., 1991). Only amino acid changes are indicated for the mutants.
Fig. 7. shows schematically the affinity panning procedure.

Fig. 8. shows the deduced amino acid sequences of the LCDR1, LCDRZ and HCDR1 loops of the monclonal antibody 3-C4F5 and selected mutant clones (numbering is according to Kabat et al., 1991). Only amino acid changes are indicated for the mutants.
Fig. 9. shows the standard curve of a competitive one-step fluoroimmunoassay for TES
s utilising monoclonal TES-binding antibody Fab fragment. The standard curve of a commercial DELFIA~ Testosterone kit is shown for comparision.
Fig. 10. shows the correlation of the results obtained by using TES-binding Fab (A60/HCDR1/LCDR2) in a competitive one-step fluoroimmunoassay with the results obtained by gas chromatography - mass spectrometry analysis.
io Fig. 11. shows the deduced amino acid sequences of the re-optimised CDR3 loops (numbering is according to Kabayt et al., 1991). Only amino acid changes are indicated for the mutants.
Fig. 12 shows the amino acid sequences of the VL regions of Fab fragments S73, 577, S83 and S88 according to the invention as referred to in Example 4 aligned with the amino acid sequence of the VL region of monclonal antibody 3-C4Fs (WT-VL) [character to show that a ~s position in the alignment is prefectly conserved: '; character to show that a position is well-conserved: ~; CDRs underlined in the WT-VL sequence.]
Fig.13 shows the amino acid sequences of the VH regions of Fab fragments 573, 577, S83 and S88 aligned with the amino acid sequence of the VH region of monclonal antibody 3-C4Fs (WT-V H) [character to show that a position in the alignment is perfectly conserved: '; character to 2o show that a position is well-conserved: ~; CDRs underlined in the WR-VH
sequence.]
Fig. 14 shows the values obtained from 48 female patient samples using either Fab fragments in the competitive one-step immunoassay protocol or a commercial kit designed for in vitro measurement of the concentration of TES in human serum or plasma.
Fig. 15 shows the values obtained from 32 male patient samples using either Fab fragments in 2 s the competitive one-step immunoassay or a commercial TES kit.
Fig. 16 shows schematically the stepwise in vitro affinity and specificity maturation process of the 3-C4F5 Fab fragment.
Summary of the Invention Theme is now described the development and characterisation of monoclonal antibodies that bind TES with affnity and specificity high enough to be utilised as reagents in immunoassays designed for the quantitative measurement of TES in clinical samples.
Specifically, the present WO 99/12974 ~ PCT/FI98/00689 invention describes the development of hybridoma cell line 3-C4F5 , the cloning and random mutagenesis of the 3-C4F5 antibody Fab fragment encoding genes, the selection of highly improved variants of the 3-CaFs by the phage display technique and the characterisation of the binding properties of the engineered Fab fragments produced in E.coli.
The usefulness of the deveioped TES-binding monoclonal recombinant Fab fragments was shown by comparing the optimised Fab fragments in the context of competitive immunoassay to a currently used commercial diagnostic kit, in which rabbit polyclonal antibodies are used as TES-specific reagents. By taking account of simultaneous improvements in affinity and io specificity the overall functioning of the recombinant antibodies was improved considerably and achieved the utmost demands for cross-reactivities to fulfil the diagnostic requirements.
This invention thus provides new reagents to be utilised in different kinds of immunoassay protocols. The invention also permits guaranteed continuous supply of these highly specific reagents of uniform quality eliminating inherent batch-to-batch variation of polyclonal antisera.
~s These advantageous effects permit the manufacture of new, highly sensitive and economical immunodiagnostic assays of uniform quality.
Detailed Description 2o The following definitions are provided for some terms used throughout this specification.
The terms, "immunoglobulin", "heavy chain", "light chain" and "Fab" are used in the same way as in Eur. Pat. Appln. No. 0125023, of Cabilly et al., published November 14, 1984, corresponding to U. S. Serial No. 483,457, filed April 8 1983.
"Antibody" in its various grammatical forms is used herein as a collective noun that refers to a population of immunoglobulin molecules and/or immunologically active portions of immunoglobulin molecules, i.e., molecules that contain an "antigen binding site" or paratope.
ao An antigen binding site is the structural portion of an antibody molecule that specifically binds an antigen.

WO 99/12974 ,. PCT/FI98/00689 Exemplary antibodies are intact immunoglobulin molecules, substantially intact immunoglobulin molecules and those portions of an immunoglobulin molecule that contain the paratope, including those portions known as Fab, Fab', F(ab')Z and F".
s Fab, and F(ab')2 portions of antibodies can be prepared by the proteolytic reaction of papain and pepsin, respectively, on substantially intact antibodies by methods that are well known. See for example, U.S. Patent No. 4, 342,566 to Theofilopolous and Dixon. Fab' antibody portions are also well known and are produced from F(ab')2 portions by reduction of the disulphide bonds linking the two heavy chain portions, as with mercaptoethanol, followed by alkylation of ~o the resulting protein mercaptan with a reagent such as iodoacetamide.
Antibodies and binding fragments can also be produced by recombinant methods, which are well known to those skilled in the art. See, for example, U. S. Patent 4, 949, 778 to Lander et al.
"Domain" is used to describe an independently folding part of a protein.
General structural is definitions for domain borders in natural proteins are given in Argos, 1988 (Argos, Protein En-gineering, 2:101-113,1988).
A "variable domain" or "Fv" is used to describe those regions of the immunoglobulin molecule which are responsible for antigen or hapten binding. Usually these consist of approximately the 2o first 100 amino acids of the N-termini of the light and the heavy chain of the immunoglobulin molecule.
"Single chain antibody" {scFv) is used to define a molecule in which the variable domains of the heavy and light chain of an antibody are joined together via a linker peptide to form a con-zs tinuous amino acid chain synthesised from a single mRNA molecule (transcript).
"Linker" or "linker peptide" is used to describe an amino acid sequence that extends between adjacent domains in a natural or engineered protein.
so A "TES-binding antibody" is an antibody which specifically recognises and binds to TES due to interactions mediated by its variable domains.

WO 99/12974 ~ PCT/FI98/00689 As hereinbefore indicated, antibodies of the present invention demonstrate an affinity for testosterone of at least about 109M'~, less than 0.05% cross-reactivity with dehydroepiandrosterone sulfate (DHEAS), less than 10% cross-reactivity with Sa-dihydrotestosterone (Sa-DHT) and less than S% cross-reactivity with androstenedione Preferably, such antibodies demonstrate less than 0.01% cross-reactivity with DHEAS, for example, less than 0.01% cross-reactivity with DHEAS, less than 8% cross-reactivity with Sa-DHT and no more than 2.5% cross-reactivity with androstenedione.
As examples of fragments of such antibodies falling within the scope of the invention, there are io herein disclosed Fab fragments S73, S77, S83 and S88 derived from the 3-CaFs Fab fragment by substitution of the heavy chain constant domain and mutation of the VL and VH regions as shown in Figures 12 and 13. In one preferred embodiment, the present invention thus provides testosterone-binding monoclonal antibodies and derivatives thereof, e.g. Fab fragments or ScFv antibody derivatives, wherein the CDRs of the VL region are selected from the VL region CDRs ~s of the Fab fragments S73, 577, S83 and S88 and the CDRs of the VH region are selected from the VH region CDRs of the same antibody fragments. Such a monoclonal antibody or antibody derivative may have VL and VH regions corresponding to complete VL and VH
regions selected from those regions of the Fab fragments 573, S77 S83 and S88 and additionally the VH region ofFab fragment S77 in which the C-terminal Serine is substituted by Alanine.
It will be 2o appreciated that mutant versions of the same CDR sequences or complete V~
and VH sequences having one or more conservative substitutions which do no substantially affect binding capability may alternatively be employed. The invention also encompasses fragments and extensions of the Fab fragments S73, S77 S83 and S88 which retain high affinity and good specificity for testosterone as hereinbefore defined.
An antibody or derivative thereof according to the invention may be labelled with any detectable label known in the art. This may include for example a radiochemical label, an enzyme label, a fluorescent label, a chemical label or a lanthanide label. The enzyme label may be alkaline phosphatase; the flourescent label may be fluorescein or a rhodamine; the chemical label may be biotin (which may be detected by avidin or streptavidin conjugated to peroxidase);
the lanthanide label may be europium. These labels are given by way of example only.

LO
Furthermore, the antibodies and derivatives thereof according to the invention may be attached to or coated upon a solid support for example the walls of a micro-titre plate or filter paper.
Antibodies and derivatives thereof according to the invention have a variety of applications.
s For example, use of antibodies for in vivo diagnosis, imaging and therapy currently attracts considerable interest from the pharmaceutical industry. In in vivo therapeutic use low immunogenicity (small size) and high affinity for antigen are typically the main requirements for the development of antibody characteristics, thus an antibody or derivative thereof according to the invention is especially suitable for in vivo therapeutic applications.
Another potential io application is in immunoassays, both non-competitive and competitive, especially for the qualitative or quantitative determination of testosterone in clinical samples.
Immunoassay methods include radio immunoassay, immunoradiometric assay, enzyme immunoassay, immunoenzymometric assay, time-resolved fluoroimmunoassay, immunofluorometric assay, chemiIuminescence immunoassay, anodic or cathodic electrochemiluminescene immunoassays, 15 homogenic immunoassays such as proximity assays with two different labels and particle immunoassays.
The reagents needed to perform the above assays according to the invention can conveniently be packaged into a kit. Thus, the invention extends to a kit comprising an antibody or 2o derivative according to the invention in a suitable container for storage and transport. The kit preferably comprises the antibody or derivative thereof according to the invention attached to or coated upon a solid support.
In another aspect, the present invention also provides DNAs encoding an antibody or antibody 2s derivative of the invention and fragments of such DNAs which encode the CDRs of the VL
and/or VH region. Such a DNA may be in the form of a vector, more particularly, for example, an expression vector which is capable of directing expression of an antibody or antibody derivative of the invention or at least one component antibody chain.
3o In a further aspect of the invention, there are provided host cells, which may be any of bacterial cells, yeast cells, fungal cells, insect cells, plant cells, and mammalian cells containing a DNA of the invention, including host cells capable of expressing an antibody or antibody derivative of the invention. Thus, an antibody or antibody derivative of the invention may be prepared by culturing a host cell or host cells of the invention expressing the required antibody chains) and either directly recovering the desired protein or, if necessary, initially recovering and combining component chains.
s As hereinbefore indicated, Fab fragments 573, S77 S83 and S88 were obtained by an in vitro maturation procedure starting from the anti-testosterone monoclonal antibody 3-C4F5. The hybridoma cell line producing monoclonal antibody 3-CaFs was developed by hyperimmunizing mice with a TES-3-carboxymethyloxime(CMO)-thyroglobulin conjugate. 3-C,oFs has a relatively high affinity (0.3x109 M'') for TES and has a reasonable low cross-reactions 011.6% and io ~1%) with Sa-dihydrotestosterone (Sa-DHT) and dehydroepiandrosterone sulfate (DHEAS).
The murine monoclonal anti-testosterone antibody, clone 3-C4F5, IgG26, x, was purified from the hybridoma cell culture supernatant on a Protein A-Sepharose (Pharmacia, Sweden). The proteolytic production of the Fab fragment was carried out using a minor modification of the is method described by Parham et al. (1982). The TES affinity and DHEAS cross-reactivities of the proteolytic Fab fragment were comparable to those of the original monoclonal 3-C4F5 antibody.
The N-terminal amino acid sequences of clone 3-C4F5 were determined and complementary 2 o DNAs of the light and heavy chains were synthesised from mRNAs isolated from hybridoma cells. The light and heavy chain cDNAs were amplified by PCR, cloned using the restriction sites introduced into the PCR primers and sequenced (Figs. 2 and 3). The cloned antibody genes were modified by PCR and the assembled Fab fragment expression unit was cloned under the tac promoter in the pKKtac vector (Takkinen et al., 1991, Alfthan et al., 1993). The first 2 s constant domain of the heavy chain (IgG26) was replaced by a gene fragment encoding the first constant domain of the IgG, subclass heavy chain to promote secretion and folding of the Fab fragment in E. coli (Alfthan et al., 1993; MacKenzie et al., 1994).
The 3-C4F5 Fab fragment expression unit in the pKKtac vector was transformed into the E.coli so RV308 strain. The Fab fragment was secreted as an active and soluble protein into the periplasmic space and was released to the culture medium during the overnight inductions. The 3-CaFs Fab fragment was purified from the culture supernatant in a single step by a TES
affinity column. Binding characteristics of the parental mouse monoclonal anti-testosterone 4 12 PCTlFI98/00689 antibody 3-CaFS (IgGzb), its proteolytic Fab fragment and E.coli-derived recombinant Fab fragment for TES and DHEAS were characterised by a competitive two-step time-resolved fluoroimmunoassay as described in Example 1. The TES affinity and DHEAS cross-reactivity of the recombinant Fab fragment were comparable to those of the original monoclonal 3-C4F5 antibody and proteolytic Fab fragment. The heavy chain constant domain subclass replacement did not affect the binding properties of the recombinant Fab fragment.
A molecular model was constructed for the 3-CaFs Fab fragment as described in Example 2.
According to the model most of the contacts between TES and the antibody originate from the third CDR loops of the heavy and the light chains. Thus, these CDRs were targeted for mutagenesis by using spiked PCR primers (Hermes et al., 1989; 1990). The randomness of PCR primers was adjusted by using a defined, biased mixture of nucleotides during the oligonucleotide synthesis for CDRs (80% wild-type and 20 % equal mixture of three other nucleotides). The CDR3 mutant libraries were constructed by the overlapping PCR method and is cloned in the separate pComb3 phagemid vectors (Barbas et ad.,1991).
The mutant libraries were selected by phage display using a competitive panning procedure.
The concentration of DHEAS was increased stepwise from 0.1 pM to 0.3 pM during seven panning cycles to achieve at least 50% inhibition of binding onto immobilised TES on each 2o round of panning. By performing the last panning cycle without amplifying the sixth phage eluate in E.coli cells, and eluting the binders with 100 pM soluble TES, the number of different clones was limited and the elution was done in a more specific way. After 7 rounds of panning, soluble Fab fragments were produced from isolated phages. Relative afFlnities and cross-reactivities of the selected mutant anti-testosterone Fab fragments were determined by the 2 s competitive time-resolved fluoroimmunoassay and compared to the recombinant wild-type Fab fragments produced either by the pKKtac or the phagemid pComb3 vectors. In one light chain mutant the DHEAS cross-reactivity was decreased to 0.52% and in one heavy chain mutant to 0.09%. By combining the light and heavy chain CDR3 mutations, the TES
afI'lnity was preserved on the wild-type level but the DHEAS cross-reactivity was further decreased to 30 0.03%. A surprising finding was that by the competitive panning procedure the overall binding specificity of the 3-C4F5 antibody was refined, since the cross-reactivities to Sa-DHT and androstenedione were also significantly decreased in the combined mutant.

The combined CDR3 mutant clone was used as a template for a subsequent round of modification by targetting parallely the remaining CDRs. A higher frequency of mutations was allowed for these generally less important loops. The VL and VH CDR1 and CDR2 loops (as defined in Kabat et al., 1991) of the combined CDR3 mutant Fab fragment of the antibody were targeted for mutagenesis by using spiked PCR primers. To restrict the frequency of amino acid mutations in single clones, the randomness of PCR primers was adjusted by nucleotide doping during the oligonucleotide synthesis for CDRs (using 62.5%
wild-type and 12.5% each three other nucleotides in oligo synthesis).
to The mutant libraries were selected by the phage display technique. To improve the TES affinity the new mutant libraries were selected by using limiting, decreasing concentrations of biotinylated TES in solution and capturing the binders on streptavidin. To improve the specificity, mutant libraries were selected by preincubating the libraries in solution with high concentration of the cross-reacting steroid, DHEAS, and then isolating TES
binders using ~5 microtiter plate wells coated with a TES conjugate. In two different light chain CDR/ mutant clones isolated from the affinity pannings, the TES affinity was increased over tenfold while the cross-reactivities to related steroids were preserved on the same level as in the parental combined CDR3 mutant clone. New heavy chain CDR1 and light chain CDR2 sequences were isolated from specificity selections showing decreased cross-reactivity to related steroids. By 2o combining compatible mutant CDRs together a Fab fragment with 18-fold lower cross-reactivity to DHEAS and over 11-fold higher TES affinity compared to original monoclonal antibody {3-C4F5) was created. Evaluation of the two best mutant Fab fragments in a competitive one-step research immunoassay protocol with TES standards in serum showed good correlation with gas chromatography - mass spectrometry analysis.
Next the CDR3 loops of the combined mutant clone were retargetted in parallel by oligo-directed PCR mutagenesis. The bias of the oligonucleotides for the parental sequence was weaker than in the previous CDR3 mutant libraries (62.5% parentai type and 12.5% each three other nucleotides) and the mutant libraries were cloned by utilising unique restriction sites 3o created by site-directed mutagenesis nearby the CDRs thus improving the quality and size of the libraries. A combined selection procedure allowing simultaneous selection with respect to affinity and specificity was used in addition to the previously utilized panning technique. A high concentration of cross-reacting steroid as a protein conjugate (DHEA-3-SucH-BSA) was used in phage solutions to inhibit binding of cross-reacting phages to limiting concentrations of biotinylated TES. The contact time of phages and biotinylated TES was decreased to 10 min and the biotinylated TES was immobilised in the streptavidin wells beforehand to ensure complete immobilisation. The elution of phages was done using 50 mM NaOH, pH
12.6 for 15 s min and immediate neutralisation with 1 M Tris, pH 7.5. The mutant libraries were also affinity selected without any competing steroid by using limiting, gradually decreasing concentrations of biotinylated TES to catch the high-affinity binders. The concentration of TES used in this approach was clearly lower (1 nM to 10 pM) compared to approaches where cross-reacting steroid was used.
After panning steps, about 50 individual clones were grown up from each library to produce soluble Fab fragments, which were analysed first on a competitive ELISA test in order to select clones having either decreased DHEAS cross-reactivity and/or improved TES
affinity. The selected clones were cloned into the pICKtac expression vector and transformed into the E. coli ~s strain RV308 for small scale production and the Fab fragments were characterised by the competitive time-resolved fluoroimmunoassay. In the case of the best light chain CDR3 mutant clone isolated from the combined affinity/specificity panning approach, the DHEAS and Sa-DHT cross-reactivities were clearly decreased (3-fold and 2-fold, respectively) while the TES
affinity was preserved at the same level as in the parental mutant clone. The TES affinity of one 20 light chain CDR3 mutant clone selected by affinity panning was improved 3-fold but all cross-reactivities were increased.
To further improve the overall binding properties of the of the TES-binding antibody, new light chain CDR3 sequences were created by combining mutations from different light chain CDR3 2 s mutant clones. Finally; a number of new CDR combinations were created by combining different CDR loop sequences selected during the stepwise optimisation of the CDRs of the 3-C4F5 antibody. The optimisation of the light chain CDR3 sequence and the CDR
combination resulted in clones having an excellent overall binding profile. A TES-binding Fab fragment having mutations in five CDRs and showing 32-fold higher affinity to TES, 60-fold lower 3o cross-reactivity to DHEAS, 2-fold lower cross-reactivity to Sa-DHT and over 9-fold lower cross-reactivity to androstenedione compared to the original monoclonal antibody was created.
The binding properties of many mutant Fab fragments were now comparable to those of a rabbit polyclonal anti-testosterone antiserum good enough for diagnostic immunoassay.

As described herein, the stepwise optimisation of the CDRs of a reasonably good hybridoma cell-derived monoclonal TES-binding antibody using random mutagenesis and phage display techniques and in the final step analysis of the additivity of different mutations and different CDRs is an efficient and feasible approach to develop recombinant monoclonal anti-testosterone antibodies for sensitive diagnostic applications.
While one successful in vitro maturation strategy for obtaining antibody fragments of the invention has been described, numerous variations whereby antibodies or antibody fragments of io the invention may be obtained will be apparent to those skilled in the art.
Depending upon the starting TES-binding antibody either one or both of the VL and VH regions may be mutated. It may prove possible to select an antibody fragment or ScFv of the invention directly from a phage or microbial display library of antibody fragments and/or ScFv antibody derivatives. A
phage or microbial cell which presents an antibody fragment or ScFv or the invetnion as a is fission protein with a surface protein represents a still further aspect of the invention.
While microbial expression of antibodies and antibody derivatives of the invention oilers means for efficient and economical production of uniform quality highly specific reagents suitable for use in immunodiagnostic assays, alternatively it may prove possible to produce such a reagent, 20 or at least a portion thereof, synthetically. By applying conventional genetic engineering techniques, initially obtained antibodies or antibody fragments of the invention may be altered, e.g. new sequences linked, without substantially altering the binding characteristics. Such techniques may be employed to produce chimeric, humanised and CDR-grafted anti-testosterone antibodies as well as novel testosterone-binding hybrid proteins which retain both 2 s affinity and high specificity for testosterone as hereinbefore defined.
The development of the original TES-binding monoclonal antibody and the hybridoma cell line producing it, the development and characterisation of the highly improved TES-binding recombinant monoclonal antibodies, and their usefulness in immunoassays is now described in 3o more detail in the following examples.

AN ANTIBODY AND FAB FRAGMENTS SPECIFICALLY BINDING TO
TESTOSTERONE
The anti-testosterone monoclonal antibody 3-CaFs was developed by hyperimmuruzing mice with a TES-3-carboxymethyloxime(CMO)-thyroglobulin conjugate. The 3-C4F5 was purified from the hybridoma cell culture and proteolytic production of the Fab fragment was carried out. The genes of the 3-C4F'S were cloned and the recombinant Fab fragment was produced in io E.coli. Binding characteristics ofthe parental monoclonal antibody and Fab fragments were analysed by the competitive two-step time-resolved fluoroimmunoassay.
Immunisation and screening procedures, cloning and characterisation of the anti-testosterone monoclonal antibody 3-CaFs are described in this example.
i5 I. Immunisation of mice and cell fusion Male mice (8 weeks old) of BALB/c were used for immunisation. Cell fusion was according to the method of Galfre & Milstein (Galfre and Milstein, 1981).
2o The mice used in a B-cell fusion, leading to clone 3-C4F5 were immunised with TES-3-carboxymethyloxime(CMO)-thyroglobulin conjugate (70 fig) emulsified in the Complete Freund's Adjuvant (Difco, USA). The immunisation was administered intraperitoneally. The mice were then reimmunized six times with TES-3-CMO-thyroglobulin conjugate (70 ,ug) emulsified in the Incomplete Freund's Adjuvant (Difco) at four weeks' intervals. A few days 2s after the last immunisation, the antibody response was tested. The mouse with the highest antibody response was chosen for cell fusion and given a booster injection (70 ~cg in saline) intravenously into the tail.
Three days after the booster injection, the spleen cells of the mouse were fused with mouse 3o myeloma cells (X63-Ag8.653) by treating the cells with 50 % PEG 4000 in DMEM (Gibco, USA). Fusion cells were then grown in HAT (Gibco) medium with 20 % Fetal Calf Serum (Gibco). About three weeks after the fusion, the hybridoma cells had grown enough to be screened. A competitive RIA was used at the beginning as a fast screening method. The anti-body-producing cell lines were further cultured until they reached a cell density of I .5 x 106 cells/ml (estimated with hematocytometer and Trypan Blue stain). The cells were then cloned at least two times by the limiting dilution method in 96-well plates and between and after the clonings the cultures were tested with the competitive RIA. The antibody-producing cell lines were further cultured, frozen, and stored in a liquid nitrogen.
II. Purification of antibody bindin~t to testosterone and preparation of proteolytic Fab fragment The monoclonal 3-C4F'S antibody, IgG26, x, was purified from the hybridoma cell culture to supernatant on a Protein A-Sepharose (Pharmacia, Sweden) column using the protocol of the manufacturer. The proteolytic production of the Fab fragment was carried out using a minor modification of the method described by Parham et al. ( 1982). Papain (Pharmacia) was preactivated before use with 50 mM cysteine in 100 mM Tris-HCI, 50 mM NaCI, 2 mM EDTA, pH 7.5 for 30 minutes at 37°C and was added to the purified antibody equilibrated in the same ~s buffer without cysteine to yield an antibody:enzyme ratio (w/w) of 33:1.
The reaction was stopped after 2 hours incubation at 37°C by adding iodoacetamide to the final concentration of mM. The Fc fragment and undigested complete antibody were immobilised on a Protein A-Sepharose column and the unbound Fab fragment in the effluent was gel filtrated and equilibrated in the phosphate-buffered-saline, pH 7.4 (PBS). The purity ofthe Fab fragment 2o was analysed by SDS-polyacrylamide gel electrophoresis (Laemmli, I970). The N-terminal amino acid sequences of the purified 3-CøFS antibody were determined on an Applied Biosystems 477 A on-line I20 A modified pulsed liquid phase/gas phase sequencer as described by Baumann ( 1990).
2 s III. Clonin; and sequencing of the 3-C4F'S Fd and L chain cDNAs All basic recombinant DNA methods were done essentially as described (Sambrook et al., 1990). mRNA was isolated from ~5 x 10' hybridoma cells using the FastTrack~
mRNA
Isolation Kit (Invitrogen Co., San Diego, CA) according to manufacturer's protocol. The first 3o strand cDNA was synthesised from poly(A) RNA either with oligo-dT priming (for the light (L) chain) or with a primer complementary to the heavy (H) chain CH2 domain coding region.
Polymerase chain reaction (PCR) was performed after the first strand cDNA
synthesis (Orlandi et al., 1989). The V,.,-5' and V,,-5' PCR primers were synthesised according to the N-terminal WO 99/12974 1g PCT1FI98/00689 amino acid sequences of the 3-C4F5 H and L chains, respectively. The CH 1-3' and CL-3' PCR
primers were complementary to the conserved sequences in the 3' end of the CH1 and C~
regions of the heavy and Iight chains, respectively. The restriction sites, introduced to the 5' end of oligonucleotides Xba I (VL-5'), Hind III (C~-3'), Xba I (VH-5') and Xho I
(Cul-3') were used s for cloning of the light and heavy chain cDNA fragments into the pSP73 vector (Promega Co., Madison, Wisconsin). The nucleotide sequences of the cloned L and Fd chains were determined from both strands of the double-stranded plasmid DNA by dideoxy DNA sequencing (Sanger et al., 1977). The deduced amino acid sequences are shown in Figs. 2. and 3.
io The cDNAs of the L and Fd chains were modified with PCR (Saiki et al., 1988) to provide them with restriction sites allowing precise in frame fusions in a Fab fragment expression unit in the pKKtac vector as described earlier (Alfthan et al., 1993; Takkinen et al., 1991). The first constant domain of the heavy chain (IgG~,) was replaced by a gene fragment encoding the first constant domain of the murine IgG, subclass heavy chain to promote secretion and folding of i s the Fab fragment in E. coli (Alfthan et al., 1993; MacKenzie et al., 1994). This replacement was done by utilising the unique BsaA I restriction site near the VH-3' end and did not alter the amino acid sequence of the VH. The nucleotide sequence of the final expression construct (Fig.
4) was verified by dideoxy sequencing.
2o IV. Production and purification of the recombinant Fab fragment The expression vector was transformed into the E.coli strain RV308. Overnight culture was diluted 1:50 and grown to an A6oo"", of 0.7-1.5, after which isopropyl ~i-D-thiogalactopyranoside (IPTG) was added to a final concentration of 1 mM and the growth was 2s continued overnight at 30°C. The cells were harvested by centrifixgation (5000 g, IO min.) and the culture supernatant was analysed on SDS-PAGE (Laemmli, 1970) followed by electrophoretic transfer of proteins onto a nitrocellulose filter (Towbin et al., 1979). The produced Fab fragment was reacted with a commercial alkaline phosphatase-conjugated goat anti-mouse kappa IgG (Southern Biotechnology Associates, Inc., USA) and detected using so BCIP/NBT as substrate (Promega). For further characterisation, the E.coli-produced Fab fragment was purified from the culture supernatant in a single step by a TES
affinity column.
Briefly, TES-3-O-(6-aminohexyl)oxime was coupled to vinylsulfone agarose (Kem-En-Tec, Denmark) according to the manufacturer's protocol and packed into a column.
The gel was saturated with BSA and washed with elution buffer (0.1 M citrate, pH 2.5) and equilibrated with PBS. The overnight supernatant of an induced culture was concentrated using the Pellicon Cassette System (Millipore, USA), applied to the column (1.2 cm x 3.0 cm) and the affinity chromatography was performed as described (Harlow and Lane, 1988): The eluted Fab was immediately neutralised with 1 M Tris-HCl buffer, pH 8.5. The pooled fractions were concentrated and then gel filtrated on an Econo-Pac l ODG column (Bio-Rad Laboratories, Richmond, CA) to change the buffer to PBS. The purified Fab fragment was analysed by SDS-PAGE. Protein concentrations were determined from the absorbance at 280 nm based on the Trp and Tyr content of the Fab protein sequence (Wetlaufer, 1962).
io V. Characterisation of testosterone-bindins~ antibodies The relative affinities and cross-reactivities of the purified parental monoclonal antibody and its Fab fragments were determined by a competitive two-step time-resolved fluoroimmunoassay.
is Briefly, goat anti-mouse IgG (Fab-specific, Southern Biotechnology Associates, Inc., USA) was coated on microtiter plate wells. The wells were blocked and samples were incubated in the wells for 2 hours. After a washing step, dilution series of different steroids (in PBS) were added together with TES-3-CMO-polylysine labelled with an Europium-chelate (Wallac, Finland), and incubated for 1 hour. After fluorescence development with the DELFIA~
2o enhancement solution (Wallac), the amount of bound label was measured with a 1230 Arcus fluorometer (Wallac). Dilution series of TES and DHEAS (in PBS) were used to analyse the relative affinities (ED-50 % concentrations) for the corresponding steroids.
DHEAS cross-reactivity was calculated from these values. The TES affinity and the DHEAS
cross-reactivity of the proteolytic and recombinant Fab fragments were comparable to those of the original 2s monoclonal 3-C4F5 antibody (Table I). The heavy chain constant domain subclass replacement did not affect the binding properties of the recombinant Fab fragment.
Table I. Relative affinities and cross-reactivities of the anti-testosterone 3-C4F5 antibody and its proteolytic and recombinant Fab fragments. EDso is the concentration of TES
required to 3o reduce binding by 50 percent.

EDso (nM) Cross-reaction Antibody # TES DHEAS
s IgG26 (3-C4F,) 13 1.6 Fab prot. 14 2.0 Fab rec. 14 1.2 The heavy chain constant domain subclass replacement did not affect the binding properties of ~o the recombinant Fab fragment.
Before the competitive immunoassay described above, the samples were titrated without soluble steroid in the same assay system in order to choose suitable dilutions for the antibody samples.
is The constructed molecular model within the ligand TES was docked revealed that several amino acid residues in the third CDRs of the heavy and light chains were in close proximity to the ligand. The CDR3 loops were randomised using spiked synthetic oligonucleotides and mutant Fab libraries were selected by a competitive panning procedure. After panning Fab 2s fragments with lowered DHEAS cross-reaction were isolated. The cross-reactivity was further decreased by combining the light and heavy chain CDR3 mutations.
I. Molecular modelling so The crystal structures of the immunoglobulin Fab fragments were extracted from the Brookhaven Databank (Bernstein et al., 1977). For the identification of the best conserved sequences, the light and heavy chain sequences of 3-C4F5 were aligned with the databank sequences using the computer program MALIGN (Johnson et al., 1993; Johnson and Overington, 1993). The two 3-D structures with the highest sequence similarity to either the H
or the L chain of 3-C,aFs were selected for model building. The anti-oxazolone antibody (Alzari et al., 1990) has 72 % sequence identity with the heavy chain of 3-C4F5 and the immunoglobuIin 4-4-20 Fab fragment (Herron et al., 1989) has 90 % sequence identity with the light chain of the 3-C,dFs. Optimal sequence alignment of the complete Fab fragments of the anti-oxazolone antibody, the 4-4-20 and the 3-C4F5 was obtained by MALIGN.
The computer program COMPOSER (Sali et al., 1990; Blundell et al., 1990) was used in building the structural model for the Fab fragment of 3-C4F3. Five out of the six CDRs were io built in as canonical conformations. The CDR3 ofthe heavy chain does not have any canonical conformation available and therefore fragment that fitted best to the structural model and had highest sequence similarity with the sequence of the CDR3 of the VH was selected from the database for that CDR3. First, the selected 3-D structures of the anti-oxazolone antibody and the 4-4-20 antibodies were simultaneously aligned and a rigid body superposition of their C, is backbones was performed in order to define the structurally conserved regions (SCRs) for the model. Next, the backbone of the 3-C4F5 SCRs was built and the side chain atoms were added using rules derived from the known 3-D structures and from the library of the most probable conformations of amino acid side chains (Sutcliffe et al., 1987). Finally, the structurally variable regions (SVRs) were built by selecting fragments from a database containing C, - C, distances 2o generated from known protein structures of resolution better than 3.0 ~.
The final model was refined by energy minimising using CFIARMm (Brooks et al., 1988).
The crystal structure of the ligand, TES, was extracted from the Cambridge Structural Database (Allen et al., 1991) and it was energy minimised. TES was then docked into the same 2s position as the oxazolone hapten in the anti-oxazolone antibody. The initial position was selected from the basis that the size of the oxazolone hapten is very similar to that of TES and the anti-oxazolone antibody was one of the crystal structures used as a template for the model.
The A-ring of TES must be solvent-exposed, because conjugated derivative of TES
(conjugated through the A-ring C3 carbonyl oxygen atom) was used for immunisation.
so However, the TES molecule can be accommodated within the binding site in a number of different orientations of the ring system. Four different starting conformations of TES were used. Beside the initial conformation, where the methyl groups point towards the antibody light chain, three other starting conformations were used. In these cases, the ring system was rotated by 90, -90 and 180 degrees with respect to the initial conformation. Each complex was energy minimised and simulated 100 picoseconds at the room temperature by using a stochastic boundary molecular dynamics method, SBMD, (Brooks and Karplus, 1989). This method allows detailed study of interactions between TES and the antibody including water molecules s and explicit hydrogen atoms into the simulation. The model of the complex in which TES had lowest interaction energy with the antibody was selected for fi~rther studies.
Our earlier experience has shown that longer simulation time improves the quality of the model (Hofl'ren et al., 1992). Thus the low energy model was further simulated 400ps by SBNiD in order to optimise the structure of the complex.
io The molecular model was inspected by molecular graphics to identify the regions of the variable domains of the heavy and light chains which are in close proximity of the antigen binding site. The identified contact residues (<3.5 ~ distance from TES), most of them locating in the third hypervariable loops of the heavy and light chains, were in the light chain G1y91, ~s His93, Va194, Pro96 and Phe98, and in the heavy chain Ser35, Trp47 Ser50, Tyr58, G1u95, Tyr99, Va1100, Glyl00J and Leul00K (numbering is according to Kabat et al., 1991).
II. Construction and selection of the mutant libraries of the 3-C4Fs an,~
tibody 2o The CDR3 loops of the light and heavy chains were targeted for oligonucleotide-directed random mutagenesis by using spiked nucleotides in PCR primer synthesis (Hermes et al., 1989;
1990). To restrict the number of amino acid mutations in single clones and thus to limit the number of the clones not retaining original structural features of the CDRs biased mixtures of nucleotides in the oligonucleotide synthesis for CDR3 loops (80% wild-type and 20 % equal 2s mixture of three other nucleotides) were used. The CDR3 region mutant libraries of the light and heavy chains were cloned separately by the overlapping PCR method (Ho et al., 1989) into the 3-C4F5 wild-type Fab fragment expression unit cloned in the phagemid vector pComb3, kindly provided by Dr. C. Barbas (Barbas et al., 1991).
so The light and heavy chain CDR3 mutant libraries were separately selected by the phage display technique (McCafI'erty et al., 1990, Barbas et al., 1991) using a competitive specificity panning procedure (Fig. 5). Briefly, the mutant Fab fragment libraries displayed on the surface of the bacteriophage were first incubated with soluble DHEAS for at least 30 minutes to inactivate those Fab phages having high cross-reactivity to DHEAS. Thereafter, the phage pools were transferred to microtiter plate wells coated with the probe, TES-4-mercaptopropionic acid-BSA conjugate (1 ~cg/well), to catch the TES binders from the libraries. After 1 hour incubation, the wells were washed 10 times during 1 hour period with washing solution (TBS-1% BSA-0.5% Tween 20) and the binders were eluted with acidic buffer {0.1 M
HCl - 1 mg/ml BSA, pH adjusted to 2.2. with glycine), and immediately neutralised with 2 M
Tris solution.
For the next panning rounds the eluted phage pools were amplified by infecting E.coli XL1-Blue cells. The inhibition effect of the soluble DHEAS was analysed during each round of selection by panning the same phage pools with and without DHEAS and titering eluted io binders on selective plates. The concentration of DHEAS was adjusted to achieve at least 50%
inhibition on each round of panning. Seven rounds of panning were performed, and during the cycles the concentration of DHEAS was increased stepwise from 0.1 pM to 0.3 pM. The seventh panning cycle was performed without amplifying the sixth phage eluate in E.coli XL1-Blue cells and on this cycle the binders were eluted with 100 pM TES in TBS-1%
BSA for 1 is hour.
N. Characterisation of the mutant Fab fragments After 7 rounds of panning soluble Fab fragments were produced from isolated clones as 2o described (Barbas et al., 1991). A number of individual clones were picked and the binding properties of the recombinant Fab fragments were preliminarily characterised by a competitive ELISA. Briefly, microtiter plate wells were coated overnight with a TES-3-CMO-BSA-conjugate (2 pg/ml) and blocked by incubating for 1 hour at room temperature with 0.5%
BSA-PBS. Samples were added with dilution series of TES and DHEAS (in PBS), and after 2 2s hours incubation and a washing step (three times with PBS), goat anti-mouse IgG (x-specific) labelled with alkaline phosphatase (Southern Biotechnology Associates, Inc., USA) was added and incubated for 1 hour. After washing, p-nitrophenylphosphate in diethanolamine-MgClz -buffer (Onion, Finland) was added and after an incubation period the absorbance in each well was read at 405nm.
Clones having decreased DHEAS cross-reactivity but still retaining the TES
binding activity on the wild-type level were selected for further characterisation. The most interesting ones were purified using the HiTrap protein G column (Pharmacia Biotech) using the protocol recommended by the manufacturer for purification of IgG molecules. Purified Fab fragments were characterised using microtiter plates coated with goat anti-mouse IgG
(Fab-specific) and Europium-labelled TES-3-CMO-polylysine as fluorescent label (as described in Example 1).
Dilution series of TES, DHEAS, Sa-DHT and androstenedione {all in PBS) were used to s analyse the relative affinities (ED-50 % concentrations) for the corresponding steroids. Cross-reactivities were calculated from these values (Table I). The amino acid sequences of the variable regions of the interesting clones were deduced by dideoxy sequencing.
Table I. Affinities and cross-reactivities of the mutant anti-testosterone Fab fragments io compared to the recombinant wild-type Fab fragments produced either by pKKtac or pComb3 vectors.
EDSO (nM) Cross-reaction rFab TES DHEAS Sa-DHT Androstenedione wt Fab (pKKtac) 6.7 1.214 1 I.9 8.3 wt Fab (pComb3) 13.8 1.463 15.0 6.7 2o mutant 44 9.7 0.085 5.4 1.8 mutant 28 18.3 0.524 2.5 12.9 combined {44+2g) 9.7 0.033 2.4 1.4 Mutant clone 44 had two mutations in the heavy chain CDR3 (Fig. 6): E95A and AlOIV
2s (Kabat et al., 1991). In this mutant clone the DHEAS cross-reactivity was decreased to 0.085 while the TES affinity was retained almost at the same level as in the wild-type {wt) Fab fragment. Mutant clone 28 had one mutation in the light chain CDR3 (Fig. 6):
H93T (Kabat et al., 1991). The DHEAS cross-reactivity was decreased to 0.524 % but this mutation also reduced the TES binding activity three-fold compared to the wild-type produced from the 3o pKKtac vector. The EDso (TES) value of the wild-type Fab fragment produced from the pKKtac vector was almost two times lower than that of the wild-type Fab fragment produced from the pComb3 vector (Table I). In the pComb3 vector the cloning sites of the cDNAs have been created within the very N-terminal regions of the heavy and light chain genes, resulting in WO 99/12974 2g PCT/FI98/00689 changes of these sequences (two amino acid changes in each chain). This is the probable reason for the decreased affinity. The same phenomenon has also been noticed earlier with an anti-digoxin Fab fragment (Short et al., 1995). When the individual CDR3 mutations identified in the heavy chain (clone 44) and light chain (clone 28) were combined into the same Fab expression unit in the pComb3 vector, the specificity of the anti-testosterone Fab fragment was still improved by a factor of 2.6 (combined (44+28)). In this mutant the cross-reactivity of DHEAS was 0.033 %. Surprisingly, also the cross-reactivities to Sa-DHT and androstenedione were much lower than in the wild-type although these steroids were not used in the panning to inhibit cross-reactive phages. The TES EDS° value of this combined mutant to was only a bit higher than in the wild-type. When expressed from the pKKtac vector with authentic N-terminal amino acids in the heavy and light chains, the affinity of the (44+28) mutant Fab was restored to the wild-type level.

In this example the mutagenesis strategy of the TES-binding Fab fragment was extended by mutating the other CDRs and selecting the mutant libraries using two different approaches in order to further improve the specificity and affinity. To improve the TES
affinity, the new mutant libraries were selected by using limiting, decreasing concentrations of a biotinylated TES in solution and capturing the binders on streptavidin. To improve the specificity, the mutant libraries were selected by preincubating the libraries in solution with a high concentration of the cross-reacting steroid, DHEAS, and then isolating the TES
binders using 2 s microtiter plate wells coated with a TES conjugate. By combining compatible mutant CDRs together, a Fab fragment with lowered cross-reactivity to DHEAS and raised TES
affinity was created.
I. Construction of the mutant libraries The VL and VH CDR1 and CDR2 regions (as defined in Kabat et al., 1991) of the combined CDR3 mutant Fab fragment (mutant 44+2g) (Example 2) of the 3-C4F5 antibody were targeted to mutagenesis by using spiked PCR primers (Hermes et al., 1989; 1990). To restrict the frequency of amino acid mutations in single clones and thus to limit the number of the clones not retaining original structural features of the CDR regions, the randomness of PCR primers was adjusted by nucleotide doping during the oligonucleotide synthesis for CDR
regions (62.5% wild-type and 12.5% each three other nucleotides).
The heavy chain CDRZ mutant library was cloned by the overlapping PCR method into the pComb3 vector containing the mutant Fab fragment (44+28) (Ho et al., 1989, Barbas et al., 1992). The three other libraries were constructed into the pComb3 vector (Barbas et al.,1991) containing the mutant Fab fragment (44+28) by utilising unique restriction sites nearby the io CDRs. For this reason, the two amino terminal codons of the light chain CDR1 containing a Bgi II restriction site used for the cloning of the light chain CDR1~ region were not randomised.
II. Selection of the mutant libraries is The mutant libraries were selected by the phage display technique (McCafferty et al., 1990, Barbas et al., 1991). To improve specificity, the libraries were panned using the same method as described earlier for the CDR3 mutant libraries of this 3-CaFs antibody (Example 2). Briefly, phage libraries were first incubated with a high concentration of soluble DHEAS in order to inactivate Fab phages with cross-reactivity to DHEAS. The phage pools were then transferred 2o to microtiter plate wells coated with 1 pg of TES-3-CMO-BSA and the TES
binders were eluted with an acidic buffer after washes. For the next panning rounds the eluted phage pools were amplified by infecting E.coli XL,-1 Blue cells. The inhibition effect of the soluble DHEAS
was analysed during each round of selection by panning the same phage pools with and without DHEAS inhibition and titering the eluted binders on selective plates. The concentration of 2 s DHEAS was adjusted to achieve at least 50% inhibition on each round of panning. A total of six cycles were performed and during the cycles the concentration of DHEAS was increased stepwise from 0.15 mM to 10 mM.
To increase affinity, the libraries were selected by using limiting, gradually decreasing so concentrations of a biotinylated hapten to catch the high-affinity binders (Fig. 7; Hawkins et al., 1992; Schier et al., 1996b). The libraries were first incubated with a biotinylated TES
(TES-3-biotin) for at least 30 minutes, then the phage pools were transferred into streptavidin-coated microtiter plate wells (Boehringer Mannheim GmbH, Germany) and incubated for 1 S

minutes. After washing the plate 10 times during a 1 hour period with the washing solution (TBS-i% BSA-0.5% Tween 20) the TES binders were eluted by adding the XLI-Blue cells (200 gel) into the wells together with high concentration of soluble TES (100 pM final). After 15 minutes infection the cells were transferred from the plate and amplified as usually. The amount of unspecific binding was analysed on each panning cycle by panning the same phage pools without adding the biotinylated TES in the first step, the binders eluted from streptavidin plates were titered on selective plates. This control panning was done because we noticed that the amount of unspecific binders eluted from the streptavidin-coated wells was quite high. The difference between the specific and unspecific binding was used to monitor the efficiency of the io panning and to adjust the concentration of the biotinylated TES during the panning steps.
Totally four rounds of panning were performed, and during the cycles the concentration of biotinylated TES was decreased gradually from 176 nM to 100 pM.
III. Characterisation of the mutants After the last panning step phagemid DNA was isolated and the pIII gene region removed by Spe I and Nhe I digestion from the pComb 3 vector (Barbas et al., 1991). The re-ligated vector DNA was transformed and about 100 individual clones were grown in a small scale to produce soluble Fab fragments for preliminary characterisation. Clones were analysed on a competitive 2o ELISA test using TES-3-CMO-BSA-coated wells (0.1 pg) to catch the TES
binders and soluble TES and DHEAS to achieve inhibition of binding (as described earlier, Example 2). The most promising clones having either decreased DHEAS cross-reactivity and/or improved TES
binding affinity were sequenced (Sanger et al., 1977) and selected for further characterisation.
These seiected clones were cloned into the pKKtac expression vector (Takkinen et al., 1991 ).
All possible different combinations of the isolated CDR regions were also collected into the pKKtac vector, and transformed into the E.coli strain RV308 for small scale production.
Clones were characterised using goat anti-mouse IgG (Fab-specific)-coated microtiter wells and TES-3-CMO-polylysine labelled with Europium-chelate (from Wallac) was used as the fluorescent label (Example 1). Dilution series of TES, DHEAS, Sa-DHT and androstenedione (all in PBS) were used to analyse the relative affinities (ED-50 %
concentrations) for the corresponding steroids. Cross-reactivities were calculated from these values (Tables I, II and III). The most interesting clones were produced in larger scale and purified using the HiTrap protein G columns (Pharmacia Biotech) following the same protocol as recommended by the manufacturer for purification of IgG samples.
TABLE I . Affinities and cross-reactivities of the anti-testosterone Fab fragments after specificity panning.
EDSO (nNl) Cross-reaction io rFab TES DHEAS Sa-DHT Androstenedione combined (44+28)5.9 0.044 2.6 1.4 HCDRI (2B5) 7.6 0.031 1.4 1.2 LCDR2 (1D7) 3.1 0.021 3.4 1.6 HCDRI/LCDR2 4.8 0.036 2.5 1.4 The specificity panning resulted in moderate improvements (Table I). In the light chain CDRZ
mutant clone 1D7, the DHEAS cross-reactivity was decreased from 0.044 % to 0.021 % while the TES affinity was improved almost 2-fold. In the LCDR2 1D7 cross-reactivities to Sa-DHT
and androstenedione were slightly increased. The heavy chain CDRI mutant 2B5 bound TES
2o with slightly lower affinity than the parental combined CDR3 mutant clone 44+28 but still had decreased cross-reactivities to all other three steroids. Interestingly, when LCDR2 and HCDRl mutations were combined into the same Fab expression unit in the pKKtac expression vector, the affinity and cross-reactivities of the Fab fragment were essentially the same as in the clone 44+28.
The light chain CDR1 mutant clones A58 and A60 selected by amity panning procedure showed about i0-fold improved affinity to TES while cross-reactivities to related steroids remained essentially unchanged (Table II).

TABLE II Affinities and cross-reactivities of the anti-testosterone Fab fragments after affinity panning.
EDso (nNl7 Cross-reaction rFab TES DHEAS Sa-DHT Androstenedione combined (44+28) 6.4 0.037 4.2 1.7 LCDR1 (A58) 0.7 0.054 4.3 2.6 io LCDR1 (A60) 0.5 0.040 3.4 2.1 By combining the affinity-selected CDRs (LCDR1: A58 and A60) with specificity-selected CDRs {HCDRI 2B5 and LCDR2 1D7) the specificity and affinity of the anti-testosterone Fab were further improved (Table III). In the best combined mutant, clone A60/HCDRI/LCDR2, is the cross-reactivity to DHEAS was 0.024 % and also the cross-reactivities to Sa-DHT and androstenedione retained on the level of the parental Fab clone (44+28). The TES EDso value of this combined mutant was 11.6-fold lower than in the wild-type.
TABLE III Affinities and cross-reactivities of the anti-testosterone Fab fragments after 2o combining the best mutations found in specificity and affinity pannings.
EDso (nlV1) Cross-reaction rFab TES DHEAS Sa-DHT Androstenedione wt Fab (pKKtac) 5. 8 0.43 5 I 1.0 8.1 A58 0.5 0.047 2.4 1.4 AS 8lHCDR 1 0. 9 0. 041 5 . 0 2.1 A58/HCDRl/LCDR2 1.3 0.043 6.1 2.0 ao A60 0.6 0.055 6.3 2.1 A60/HCDR1 0.8 0.026 4.7 2.3 A60/HCDR1/LCDR2 0.5 0.024 4.8 1.7 The deduced protein sequences of the LCDR1 mutant clones A58 and A60 showed significant differences compared to the wild-type sequence (Fig. 8). In the LCDR2 mutant (1D7) four out ofthe seven amino acid residues have been mutated (VS1A, S52R, N53Y and S56P) whereas the HCDRl mutant (2B5) had only one amino acid substitution (T31R).
IV. Kinetic measurements and affinity constant determinations by BIAcore~'~'' The kinetics of binding of the purified mutant Fab fragments to the TES-3-CMO-BSA
immobilised on a dextran-coated sensor chip were determined from the dependence of the surface plasmon resonance response upon the concentration of the purified Fab fragments injected into the biosensor (BIAcore~ , Pharmacia Biosensor AB, Sweden) (Karlsson et al., 1991; Jonsson et al., 1992). The antigen, TES-3-CMO-BSA, was coupled to the CMS sensor chip through its amine groups using the Amine Coupling Kit (Pharmacia Biosensor AB)(Johnsson et al., 1991). The antigen was diluted in 10 mM acetate buffer (pH 4.0) to the is concentration of 6.25 ~cg/ml and 398 resonance units (RL~ was immobilised using automated immobilisation cycle at a constant flow rate of 5 ~d/min as follows: 70 ~cl of 0.4 M N-ethyl-N'-(dimethylamino-propyl)carbodiimide (EDC) and 0.1 M N-hydroxysuccinimide (NHS) were transferred to the mixing vial, whereafler 35 ~1 of the mixed solution was injected over the surface. The antigen solution, 35 ~1, was then injected, followed by injection of 35 ul of 1 M
2o ethanolamine hydrochloride solution. Kinetic measurements were carried out at 25 °C in 10 mM HEPES complete buffer, pH 7.4 (HBS)(containing 3.4 mM Titriplex III, 0.15 M
sodium chloride, and 0.05% surfactant P20 (Pharmacia)) with a constant flow rate of 5 ~cUmin and injected sample volume of 30 /.d. In the measurements six different concentrations (between 12.5 - 400 nNi) of the purified Fab fragments were used, and the antigen was regenerated with 2s 10 ~1 of 50 mM NaOH, pH 12.6, between samples. The kinetic evaluation of data was performed using the BIAevaluation 2.1 software (Pharmacia Biosensor AB).
As shown in table IV, the most significant change in the rate constants was the about 20-fold slower dissociation rate of the best mutants compared to the wild-type.
Changes in the 3o association rate were smaller but together with dissociation rate improvements resulted in over 40-fold higher affinities of the best mutants to the TES conjugate compared to the wild-type Fab.

TABLE IV. Binding kinetics and affinities of the wild-type and three CDR
mutant library-derived, combined mutant Fab fragments. The ko" and kog rates for the Fab fragments were determined by BIAcore~'~"''. Affinity constants were calculated from these values.
s Fab ka" kon K,--k~,/l~,r M'ls 1/105 s 1/10'' Nfl/109 wt Fab 1.29 21.10 0.06 combined (44 + 28) 1.53 9.03 0.17 io A58/HCDR1 3.12 1.14 2.74 A60/HCDRl/LCDRZ 2.13 0.87 2.45 V. Evaluation of the A60/HCDRI/LCDRZ mutant Fab fragment in a competitive one-step fluoroimmunoassay is The improved sensitivity of the best mutant Fab fragment (A60/HCDR1/LCDR2) allowed use of serum based standards and a competitive one-step immunoassay protocol to test the performance of the mutant Fab in a clinically relevant TES concentration area.
The developed competitive one-step immunoassay protocol was compared to a currently used commercial kit, 2o in which rabbit polyclonal antibodies are used as TES specific reagents, designed for in vitro measurement of the concentration of TES in human serum or plasma (DELFIAc~
Testosterone kit, Time-resolved fluoroimmunoassay kit 1244-050 (Wallac) and to gas chromatography -mass spectrometry analysis.
2s First, suitable dilution for the A60/HCDRI/LCDRZ Fab (in 0.5% BSA/PBS) was titrated.
Briefly, 50 pl of TES-free serum, 100 pl of TES-Eu tracer (from the DELFIA~
Testosterone kit, diluted 1:50 to Testosterone Assay Buffer from the same kit) and 100 pl of different dilutions of the Fab (in 0.5% BSA/PBS) were pipetted to microtiter plate wells precoated with goat anti-mouse IgG. After 2 hours incubation at room temperature with shaking, the wells 3o were washed three times with DELFIA~ wash solution and then 200 pl of DELFIA~
Enhancement Solution were added to the wells. After fluorescence development at room temperature the fluorescence was measured by a 1230 Arcus fluorometer (Wallac). Suitable dilution of the Fab for subsequent TES measurements was chosen from a linear part of a curve representing the fluorescence as a function of a dilution of the Fab.
With the suitable dilution of the Fab and using different concentrations (0-50 nmol/1) of TES
s standards (in TES-free serum) unknown patient serum specimens were analysed using the same one-step assay protocol as described. After fluorescence measurement a standard curve, representing BB"""~ (%) as a function of TES concentration, was generated by the Multicalc program (Wallac) and TES concentrations of unknown samples were read from the curve using the same program. The same unknown patient serum specimens were analysed by using io gas chromatography - mass spectrometry analysis and by the currently used commercial kit, in which rabbit polyclonal antibodies are used as TES specific reagents ( DELFIA~
Testosterone kit, Time-resolved fluoroimmunoassay kit 1244-OSO,Wallac).
The standard curve for the modified research assay method utilizing the developed Fab fragment was essentially identical to that of the kit and covered the clinically relevant ~s concentration range (Fig. 9). The lowest standard (0.5 nmol/1) gave 85.4%
and 84.5% (BB,",x %) binding for the A60/HCDRl/LCDR2 Fab and the kit, respectively. The highest standard (50 nmoUl) gave 11.8% and 10.2% binding for the A60BCDR1/LCDR2 and the kit, respectively. The determined TES ED-50% values were 5.6 nmol/1 for the A60/HCDR1/LCDR2 and 4.3 nmol/1 for the kit. No proper binding inhibition of the parental 2o wild-type monoclonal was achieved with clinically relevant standard concentrations in the modified research assay system used.
Figure 10. shows the correlation of the results obtained for unknown serum samples using the A60/HCDR1/LCDR2 Fab-based immunoassay to the results determined by gas 2s chromatography - mass spectrometry analysis. Totally 28 samples were analysed in the range 0.501 to 31.851 nmol/1 with the mean value 9. A good correlation was found (r = 0.983, slope = 1.11). However, the obtained intercept value (= 5) revealed that although the overall binding specificity of the A60/HCDR1/LCDRZ Fab is significantly improved compared to the wild-type, it still has to be improved in order to allow determination of low serum TES values. The so correlation to the DELFIA~ Testosterone kit results confirmed the same (r =
0.989, slope =
1.17, intercept 3.95).
The use of the A60/HCDRIILCDR2 Fab with patient serum specimens revealed that the developed Fab fragments do work in complex assay medium containing human serum and, e.g., components intended to facilitate the release of TES from binding proteins in the sample.
However, particularly the low TES concentrations deviate significantly from the gas chromatography - mass spectrometry analysis. This positive bias is most probably as a consequence of cross-reactions with related steroids, notably DHEAS, normal concentration in female samples being 300 -10 000 times higher than TES cancentration.

In this example the CDR3 loops of the combined mutant clone were retargetted to oligo-directed PCR mutagenesis. The bias of the oligonucleotides for the parental sequence was weaker than in the previous CDR3 mutant libraries and the mutant libraries were cloned by utilising unique restriction sites created by site-directed mutagenesis nearby the CDRs thus is improving the quality and size of the libraries. Combined selection procedure allowing simultaneous selection with respect to affinity and specificity was used in addition to the previously utilized afFnity panning approach.
I Reoptimization~Mutagenesis) of the CDR3 loops of the A60/HCDR1/LCDR2 antibody The VL and VH CDR3 loops (as defined in Kabat et al., 1991) of the mutant Fab fragment (after CDR1, 2 and 3 metagenesis, A60/HCDRl/LCDR2, Example 3) of the 3-C4F5 antibody (Example 1) were retargetted to mutagenesis by using spiked PCR primers (Hermes et al., 1989; 1990). To restrict the frequency of new amino acid mutations in single clones, and thus 2 s to limit the number of the clones not retaining previously optimised structural features of the CDRs the randomness of PCR primers were adjusted by nucleotide doping during the oligonucleotide synthesis for CDR loops (62.5% parental-type (A60/HCDR1/LCDR2) and 12.5% each three other nucleotides). Both mutant libraries were constructed into the pComb3 vector (Barbas et al.,1991) containing the mutant Fab fragment (A60lHCDRI/LCDR2) by so utilising unique restriction sites created before cloning by site-directed mutagenesis sites nearby the CDR3s thus improving the quality and size of the libraries.

II. Selection of the mutant libraries The mutant libraries were selected by the antibody phage display technique (McCafferty et al., 1990, Barbas et al., 1991). Briefly, on every panning round mutant libraries were first s incubated with a high concentration of soluble DHEA-BSA conjugate (RHEA-3-SucH-BSA, 17-20 mol DHEA/mol BSA, 0.4 mg/ml) for at least 30 min at 37°C in order to inactivate Fab phages with cross-reactivity to DHEAS. On two first rounds the phage pools were then transferred into microtiter plate wells coated with 1 pg of TES-3-CMO-BSA and on next three rounds to streptavidin-coated microtiter plate wells (Boehringer Mannheim GmbH, Germany) ~o preincubated for 1 h with TES-3-biotin. After a 10 min incubation at 37°C, the wells were washed (TBS-1% BSA-0.05% Tween 20) 15 times during an 1 h period before elution of binders with 50 mM NaOH, pH 12.6 for 15 min and immediate neutralisation with 1 M Tris, pH 7.5. Between panning rounds the eluted phage pools were amplified by infecting E.coli XL,-1 Blue cells. A total of five panning cycles were performed and during the last three cycles is the concentration of TES-3-biotin used for preincubation of streptavidin-coated wells was decreased stepwise from 22 nM to 0.2 nM.
The mutant libraries were also affinity selected without any competing steroid by using limiting, gradually decreasing concentrations of biotinylated TES to catch the high-amity 2o binders (Hawkins et al., 1992; Schier et al., 1996b). The libraries were first incubated with a biotinylated TES (TES-3-biotin) for S min, then the phage pools were transferred to streptavidin-coated microtiter plate wells (Boehringer Mannheim GmbH, Germany) and incubated for 25 min at 37°C. After washing the plate 15 times during an 1 h period with the washing solution (TBS-1% BSA-0.05% Tween 20) the TES binders were eluted with 50 mM
2 s NaOH, pH 12.6 for I S min and then immediately neutralised with 1 M Tris, pH 7.5. A total of five rounds of panning were performed, and during the cycles the concentration of biotinylated TES was decreased gradually from 1 nM to 10 pM.
3o III. Characterisation of the mutants After the last panning step phagemid DNA was isolated and the pIII gene region was removed by Spe I and Nhe I digestion from the pComb 3 vector (Barbas et al., 1991).
The re-ligated vector DNA was transformed and about 50 individual clones were grown in a small scale to produce soluble Fab fragments for preliminary characterisation. Clones were analysed on a competitive ELISA test using TES-3-CMO-BSA-coated wells (0.1 pg) to catch the TES-binders and soluble TES and DHEAS to achieve inhibition of binding (as described earlier, Example 2). The most promising clones having either decreased DHEAS cross-reactivity and/or improved TES-binding affinity were sequenced (Sanger et al., 1977) and selected for further characterisation.
These selected clones were cloned into the pKKtac expression vector (Talckinen et al., 1991) io and transformed into the E.coli strain RV308 for small scale production.
Clones were characterised using goat anti-mouse IgG-coated microtiter wells, and TES-3-CMO-polylysine labelled with Europium-chelate (from Wallac) was used as the fluorescent label. A one incubation step assay protocol was used for characterisation of these clones i.e. the label, antibody sample and competing steroid were all added in the same time onto the wells i5 (Example 3/V). Dilution series of TES, DHEAS, Sa-DHT and androstenedione (all in serum) were used to analyse the relative affinities (ED-50 % concentrations) for the corresponding steroids. Cross-reactivities were calculated from these values (Table I ).
The most interesting clones were produced in a larger scale and purified using the HiTrap 2o protein G columns (Pharmacia Biotech) following the same protocol as recommended by the manufacturer for purification of IgG samples.
TABLE I. Affinities and cross-reactivities of the best anti-testosterone Fab fragments after combined affinity/specificity and affinity panning procedures.
EDso (n~ Cross-reaction rFab TES DHEAS Sa-DHT Androstenedione 3o A60/HCDR1/LCDR2 6.1 0.020 14.9 2.0 LCDR3 A4 8.7 0.007 6.3 2.4 LCDR3 A46 2.3 0.030 22.6 3.1 HCDR3 A52 / LCDR3 A4 8.4 0.009 7.8 3.4 In the best light chain CDR3 mutant clone isolated from the combined affinity/specificity panning approach, A4, the DHEAS and Sa-DHT cross-reactivities were clearly decreased (3-fold and 2-fold, respectively) while the TES affinity was preserved on the same level as in the s parental mutant clone. The TES affinity of the light chain CDR3 mutant clone A46, selected by affinity panning, was improved 3-fold but all cross-reactivities were increased. The binding properties of the affinity-selected heavy chain CDR3 mutant clone A52, combined to new light chain CDR3 mutant A4, were essentially the same as those of the clone A4 alone, although the same CDR combination showed positive effect when expressed from the pComb3 vector. The io deduced amino acid sequences of the re-optimised CDR3 loops are shown in Figure 11.
To fi~rther improve the overall binding properties of the TES antibody, new light chain CDR3 sequences were created by combining mutations from three different LCDR3 mutant clones (28, A4 and A46). Finally, a number of new CDR combinations were created by combining ~s different CDR loop sequences selected during the stepwise optimisation of the CDRs of the 3-C,eFs antibody.
TABLE II. Affinities and cross-reactivities of the best anti-testosterone Fab fragments after combining the LCDR3 mutations from different mutant clones and by optimisation of the 2o CDR combination.
EDso (nM) Cross-reaction rFab TES DHEAS Sa-DHT Androstenedione A60/HCDR1/LCDRZ 6.1 0.020 14.9 2.0 S58 2.7 0.009 11.3 2.3 S73 4.8 0.006 7.2 2.3 S77 4.8 0.007 7.9 2.2 3o S83 2.4 0.005 6.9 1.8 S88 4.6 0.007 5.4 2.5 As shown in Table II, the optimisation of LCDR3 sequence and CDR combination resulted in clones having excellent overall binding profile. No proper inhibition of the original monoclonal antibody was achieved in the modified research assay system with the same clinically relevant standard concentrations. The extrapolated TES EDso value for the original antibody was 76.5 nM and the cross-reactivities 0.3% (RHEAS), 13.7% (Sa-DHT) and 16.6%
(androstenedione). Based on the extrapolated values, we were able to produce a TES-binding Fab fragment having mutations in five CDRs and showing 32-fold higher affinity to TES, 60-fold lower cross-reactivity to RHEAS, 2-fold lower cross-reactivity to Sa-DHT
and over 9-fold lower cross-reactivity to androstenedione compared to the original monoclonal antibody.
to The amino acid sequences of the VL and VH regions of Fab fragments S73, S77, S83 and S88 aligned with the V~, and VH sequences of the 3-C4F5 antibody are shown in Figures 12 and 13 respectively.

PERFORMANCE OF SELECTED RECOMBINANT FAB FRAGMENTS IN A CLINICAL
IIVllvIUNOASSAY
In this example the performance of the best mutant Fab fragments were analysed in the context of competitive immunoassay protocol and compared to currently used commercial kit, in which rabbit polyclonal antibodies are used as TES specific reagents, designed for in vitro measurement of the concentration of TES in human serum or plasma (DELFIA~
Testosterone kit, Time-resolved fluoroimmunoassay kit 1244-050, Wallac).
I. The immunoassay protocol First, suitable dilutions for the Fab fragments (in 0.5% BSA/PBS) were titrated as described in Example 3/V. Next, different TES standards (0-50 nmoUl in human TES-free serum) and unknown patient serum specimens were analysed using the same assay protocol as described in 3o Example 3/V. The unknown patient serum specimens were also analysed using the DELFIA~
Testosterone kit as recommended by the manufacturer.

II Performance of the selected recombinant mutant Fab fragments in the clinical immunoass~
The standard curves for all the analysed Fab fragments were in the clinically relevant concentration range. The lowest standard (0.49 nmoUl) gave 86.5% - 93.1%
(BB,",x %) s binding for the Fab fragments and 85.9% for the kit. The highest standard (50.9 nmol/l) gave 8.6% - 15.6% and 10.4% binding for the Fab fragments and the kit, respectively. The determined TES ED-50% values were 3.4 - 8.8 nmol/1 for the Fab fragments arid 4.3 nmoUl for the kit.
~o Figures 14 and 15 show the results obtained for 48 female and for 32 male patient samples, respectively, using either Fab fragments in the one-step immunoassay or the commercial DELFIA~ testosterone kit designed for in vitro measurement of the concentration of TES in human serum or plasma.
Table I shows the comparision of correlation parameters between different Fab fragments in the one-step immunoassay and the commercial DELFIA~ Testosterone kit when 80 patient serum samples were measured.
TABLE I. The comparison between Fab fragments and commercial DELFIA~
Testosterone kit used for measuring 80 patient serum samples.

TES EO-60 8,8 8,8 5,1 5,6 3,4 4,8 4,3 Sample 0,98 1,09 0,47 0,1 0,69 0,34 0,13 min Sample 35,62 33,89 30,8 38,45 29,53 30,77 28,57 max Slope 1.161 1.187 1.094 1.105 1.059 1.043 Intercept 1.765 1.217 0,372 0,005 0,877 0,059 ~-value 0,971 0,971 0,964 0,943 0,982 0,968 A good correlation was found for ali the Fab fragments (r = 0.943 - 0.971).
However, the obtained intercept values of the A4 and A52/A4 Fab fragments (1.765 and 1.217, respectively) 2 s revealed that the analysis of the additivity of different mutations and different optimised CDRs was a necessary and efficient step to further fine-tune functioning of the Fab fragments. After the final optimisation the binding properties of the mutant Fab fragments, notably S77 and S88 Fabs, in this modified research assay system were comparable to those of a rabbit polyclonal anti-testosterone antiserum good enough for diagnostic immunoassay. By taking account of simultaneous improvements in affinity and specificity we were able to improve the overall functioning of the recombinant antibody in the context of competitive immunoassay considerably and have now achieved the utmost demands for cross-reactivities to fulfil the diagnostic requirements.

WO 99/12974 4p PCT/FI98/00689 References Alfthan, K., Takkinen, K., Sizmann, D., Seppala, L, Immonen, T., Vanne, L., Keranen, S., Kaartinen, M., Knowles, J.K.C. and Teeri, T.T. (1993) Gene, 128, 203-209.
s Allen, F.H., Davies, J.E., Galloy, J.J., Johnson, O., Kennard, O., Macrae, C.F., Mitchell, E.M., Mitchell, G.F., Smith, J.M. and Watson, D.G. (1991) J Chem. Inf. Comp. Sci., 31, 187-204.
Alzari, P.M., Spinelli, S., Mariuzza, R.A., Boulot, G., Poljak, R.J., Jarvis, J.M. and Milstein, C. ( 1990) EMBO Journal, 9, 3 807-3 814.
Arevalo, J.H., Stura, E.A., Taussig, M.J. and Wilson, LA. (1993a) J. Mol.
Biol., 231, 103-io 118.
Arevalo, J.H., Taussig, M.J. and Wilson, LA. (1993b) Nature, 365, 859-863.
Arevalo, J.H., Hassig, C.A., Stura, E.A., Sims, M.J., Taussig, M.J. and Wilson, LA. (1994) J
Mol. Biol., 241, 663-690.
Argos, P. {1988) Protein Ef~neerin~, 2, 101-113.
is Barbas III, C.F., Bain, J.D., Hoekstra, D.M., and Lerner, R.A. (1992 Proc.
Natl. Acad. Sci., USA., 89, 4457-4461.
Barbas III, C.F., Hu, D., Dunlop, N., Sawyer, L., Cababa, D., Hendry, R.M., Nara, F.L. and Burton, D.R. (1994) Proc. Natl Acaa~ Sci..USA, 91, 3809-3813.
Barbas III, C.F., Kang, A.S., Lerner, R.A. and Benkovic, S.J. {1991) Proc.
Natl Acad. Sci.
2o USA, 88, 7978-7982.
Baumann, M. (1990) Anal. Biochem., 190, 198-208.
Berek, C. and Mllstein, C. (1987) Immunol. Rev., 105, 5-23.
Bernstein, F.C., Koetzle, T.F., Willliams, G.J.B., Meyer, E.F., Jr, Brice, M.D., Rodgers, J.R.,Kennard, O., Shimanouchi, T. and Tasumi, Y. (1977) J. Mol. Biol., 112, 535-542.
2s Better, M., Chang, C.P., Robinson, R.R. and Horwitz, A.H. (1988) Scie»ce, 240, 1041-1043.
Blundell, T.L., Johnson, M.S., Overington, J.P., and Sali, A. (1990) In Hook, J.B. and.Poste, G. (eds.), Protein Design and the Development of New Therapeutics and Vaccines. Plenum Press, USA, pp. 209-227.
Brooks, B.R., Bruccoleri, R.E., Olafson, B.D., Sates, D.J., Swaminathan, S.
and Karplus, M.
30 (1983) J. Comp. Chem., 4, 187-217.
Brooks, C.L., III, Karplus, M. and Pettit, B.M. (1988) Proteins: A Theoretical Perspective of Dynamics Structure and Thermodynamics Advances in Chemical Physics, Vol. LX3~, Wiley, New York, pp. 33-58.

Deng, S.-J., MacKenzie, C.R., Himara, T., Brousseau, R., Lowary, T.L., Young, N.M., Bundle, D.R. and Narang, S.A. (1995) Proc. Natl Acad Sci. USA, 92, 4992-4996.
Galfre, G. and Milstein, C. (1981) Methods Enzymol., 73, 3-46.
Harlow, E. and Lane, D. (1988) Antibodies. A LaboratoryManual. Cold Spring Harbor s Laboratory Press, Cold Spring Harbor, NY.
Hawkins, R.E. and Winter, G. (1992) Eur. J. Immunol., 22, 867-870.
Hermes, J.D., Parekh, S.M., Blacklow, S.C., Koster, H. and Knowles, J.R.
(1989) Gene, 84, 143-151.
Hermes, J.D., Blacklow, S.C. and Knowles, J.R. (1990) ) Proc. Natl Acad. Sci.
USA, 87, 696-io 700.
Herron, J.N., He, X., Mason, M.L., Voss, Jr., W. and Edmundson, A.B. (1989) PROTEINS

5, 271.
Ho, S.N., Hunt, H.D., Horton, R.M., Pullen, J.K. and Pease, L.R. (1989) Gene, 77, 51-59.
Hoffren, A-M., Holm, L., Laaksonen, L., Teeri, T.T, and Teleman, O. (1992) 15 ImmunoMethods. 1, 80-90.
Ismail, A.A.A., Astley, P., Cawood, M., Short, F., Wakelin, K. and Wheeler, M.
(1986) Ann.
Clin. Biochem., 23, 135-145.
Jeffrey, P.D., Strong, R.K., Sieker, L.C., Chang, C.Y.Y., Cambell, R.L., Petsko, G.A., Haber, E., Margolies, M.N. and Sheriff, S. (1993) Proc. Natl Acad Sci. USA, 90, 10310-10314.
2o Johnson, M.S., Overington, J.P. and Blundell, T.L. (1993) J. Mol. Biol., 231, 735-752.
Johnson, M.S. and Overington, J.P. (1993) J. Mol. Biol., 233, 716-738.
Johnsson, B., Lofas, S., and Lindqvist, G. (1991) Anal. Biochem., 198, 268-277.
J~insson, U., Fagerstam, L., Ivarsson, B., Johnsson, B., Karlsson, R., Lundh, K., Lofas, S., Persson, B., Roos, H., Ronnberg, L, Sjtilander, S., Stenberg, E., St$hlblrrg, R., Urbaniczky, C., 2s bstlin, H., and Malmqvist, M. (1991) BioTechnigues, 11, 620-627.
Kabat, E.A., Wu, T.T., Reid-Miller, M., Perry, H.M. and Gottesman, K.S.
(1991Lc~uences f Proteins of Immulogical Interest, 4th Ed., U. S. Dept. of Health and Human Services, Bethesda, MD.
Karlsson, R., Michaelsson, A., and Mattson, A. (1991 J. Immunol. Methods, 145, 229-240.
3o Korth-Schutz, S., Virdis, R., Saenger, P., Chow, D.M., Levine, L.S. and New, M.I. (1978) J
Clin. Errdocrinol. Metab., 46, 452-457.
Laemmli, U.K. (1970) Nature, 227, 680-685.
Leung, D.W., Chen, E. and Goeddel, D.V., (1989) Tech. J. Methods Cell Mol.
Biol., 1, 11.

MacKenzie, C.R., Sharma, V., Brummell, D., Bilous, D., Dubuc, G., Sadowska, J., Young, N.M., Bundle, D.R. and Narang, S.A. (1994) BiolTechnolopv, 12, 390-395.
Marks, J.D., Hoogenboom, H.R., Bonnert, T.P., McCafferty, J., Griffiths, A.D.
and Winter, G.
(1991) J. Mol. Biol., 222, 581-597.
s McCafferty, J., Griffiths, A.D., Winter, G. and Chiswell, F.J. (1990) Nature, 348, 552-554.
Myers, R.M., Lerman, L.S. and Maniatis, T. (1985) Science, 229, 242-247.
Odel, W.D. and Swerdloff, R.S. (1978) Clin. Endocrinol., 8, 149-180.
Orlandi, R., Gussow, D.H., Jones, P.T. and Winter, G. (1989) Proc. Natl Acad.
Sci. USA, 86, 3833-3837.
1o Osborn, R.H. and Yannone, M.E. (1973) Gynecol. Surv., 26, 195-228.
Parham, P., Androlewicz, M.J., Brodsky, F.M., Holmes, N.J. and Ways, J.P.
(1982) J
Immunol. Methods. 53, 133-173.
Pope, A., Pritchard, K., Williams, A., Roberts, A., Hackett, J.R., Mandecki, W. and Johnson, K.S. (1996) Immunotechnolo~y, 2, 209-217.
is Saiki, R.K., Gelfand, D.H., Stoffel, S., Scharf, S.J., Higuchi, R., Horn, G.T., Mullis, K.B. and Erlich, H.A (1988) Science. 239, 487-491.
Sali, A., Overington, J.P., Johnson, M.S. and Blundell, T.L. (1990) TIBS. 15, 235-240.
Sambrook, J., Fritsch, E.F. and Maniatis, T. (1990) Molecular Cloning: A
Laboratory Manual, 2nd Ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
2o Sanger, F., Nicklen, S. and Coulson, A.R. (1977) Proc. Natl Acad. Sci. USA, 74, 5463-5467.
Schaaper, R.M. (1988) Proc. Natl Acad. Sci. USA, 85, 8126-8130.
Schier, R., Balint, R.F., McCall, A., Apell, G., Larrick, J.W. and Marks, J.D.
(1996b) Gene.
169, 147-155.
Schier, R., Bye, J.M., Apell, G., McCall, A., Adams, G.P., Malmqvist, M., Weiner, L.M. and 2s Marks, J.D. (1996a) J. Mol. Biol., 225, 28-43.
Schier, R., McCall, A., Adams, G.P., Marshall, K.W., Merritt, H., Yim, M., Crawford, R.S., Weiner, L.M., Marks, C. and Marks, J.D. (1996c) J. Mol. Biol., 263, 551-567.
Schildbach, J.F., Shaw, S.Y., Bruccoleri, R.E., Haber, E., Herzenberg, L.A., Jager, G.C., Jeffrey, P.D., Panka, D.J., Parks, D.R., Near, R.L, Novotny, J., Sheriff, S.
and Margolies, so M.N. (1994) Protein Sci., 3, 737-749.
Short, M.K., Jeffrey, P.D., Kwong, R.-F. and Margolies, M.N. (1995) J. Biol.
Chem., 270, 28541-28550.
Skerra, A. and Pliickthun, A., (1988) Science, 240, 1038-1041.

Sutcliffe, M.J., Hayes, F.R.F. and Blundell, T.L. (1987) Protein Engin., 1, 385-392.
Takkinen, K., Laukkanen, M-L., Sizmann, D., Alfthan, K., Immonen, T., Vanne, L., Kaartinen, M., _Knowles, J.K.C. and Teeri, T.T. (1991) Protein Engin., 4, 837-841.
Towbin, H., Staehlin, T. and Gordon, J. (1979) Proc. Natl Acad. Sci. USA, 76, 4350-4354.
s Wang, C. and Swerloff, R.S. (1992) Baillieres Clin. Endocrinol. Metab., 6, 405-435.
Wetlaufer, D. (1962) Adv. Protein Chem., 17, 303-390.
Wheeler, M.J. (1994) In: Wild, D. (ed.) The immunoassa~handbook. Macmillan Press, Hants, UK, pp.365-78.
Wilke,T.J. and Utley, D.J. (1987) Clin. Chem., 33, 1372-1375.
Wilson, LA. and Stanfield, R.L. (1994) Current Opinion in Structural Biology!, 4, 857-867.
Wu, F.C.W., Edmond, P., Raap, G. and Hunter, W.M. (1981) Clin. Endocrinol., 14, 493-507.
Yang, W.-P., Green, K., Pinz-Sweeney, S., Briones, A.T., Burton, D.R. and Barbas, C.F.
(1995) J. Mol. Biol., 254, 392-403.
Yelton, D.E., Rosok, M.J., Cruz, G., Cosand, W.L., Bajorath, J., Hellstrom, L, Hellstrom, 1s K.E., Huse, W.D. and Glaser, S.M. (1995) J. Immunol., 155, 1994-2004.

WO 99/12974 PCTlFI98/00689 SEQUENCE LISTING
(1) GENERAL INFORMATION:
{i) APPLICANT:
(A) NAME: Orion-yhtyma Oyj Orion Diagnostica (B) STREET: Orionintie 1 (C) CITY: Espoo (E) COUNTRY: Finland (F) POSTAL CODE (ZIP): FIN-02200 (A) NAME: Wallac Oy (B) STREET: P.O. Box 10 (C) CITY: Turku (E) COUNTRY: Finland (F) POSTAL CODE (ZIP): FIN-20101 (ii) TITLE OF INVENTION: Monoclonal Antibodies (iii) NUMBER OF SEQUENCES: 26 (iv) COMPUTER READABLE FORM:
{A) MEDIUM TYPE: Floppy disk (B) COMPUTER: IBM PC compatible (C) OPERATING SYSTEM: PC-DOS/MS-DOS
(D) SOFTWARE: PatentIn Release #1.0, Version #1.25 (EPO) (2) INFORMATION FOR SEQ ID NO: l:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 112 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single {D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (light chain variable region of the monoclonal 3-C, FS antibody) (ix) FEATURE:
(A) NAME/KEY: Region (B) LOCATION: 24..39 (D) OTHER INFORMATION: /label= CDR
/note= "CDR1"
(ix) FEATURE:
(A) NAME/KEY: Region (B) LOCATION: 55..61 (D) OTHER INFORMATION: /label= CDR
/note= "CDR2"
{ix) FEATURE:
(A) NAME/KEY: Region (B) LOCATION: 94..102 (D) OTHER INFORMATION: /label= CDR
/note= "CDR3"
(xi) SEQUENCE DESCRIPTION: SEQ ID NO: l:
Asp Val Val Val Thr Gln Thr Pro Leu Ser Leu Pro Val Ser Leu Gly Asp Gln Ala Ser Ile Ser Cys Arg Ser Ser Gln Sex Ile Val His Ser Asn Gly Asn Ser Tyr Leu Glu Trp Tyr Leu Gln Lys Pro Gly Gln Ser Pro Lys Leu Leu Ile Tyr Lys Val Ser Asn Arg Phe Ser Gly Val Pro Asp Arg Phe Ser Gly Ser Gly Ser Gly Thr Asp Phe Thr Leu Lys Ile Ser Arg Val Glu Ala Glu Asp Leu G1y Va1 Tyr Tyr Cys Phe Gln Gly Ser His Val Pro Pro Thr Phe Gly Gly Gly Thr Lys Leu Glu Ile Lys (2) INFORMATION FOR SEQ ID NO: 2:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 118 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (heavy chain variable region of the monoclonal 3-C,FS antibody) (ix) FEATURE:
(A) NAME/KEY: Region (B) LOCATION: 31..35 (D) OTHER INFORMATION: /label= CDR
/note= "CDRl"
(ix) FEATURE:
(A) NAME/KEY: Region (B) LOCATION: 50..65 (D) OTHER INFORMATION: /label= CDR
/note= "CDR2"
(ix) FEATURE:
(A) NAME/KEY: Region (B) LOCATION: 98..107 (D) OTHER INFORMATION: /label= CDR
/note= "CDR3"
(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 2:
Glu Val Lys Leu Val Glu Ser Gly Gly Gly Leu Val Lys Pro Gly Gly Ser Leu Lys Leu Ser Cys Ala Ala Ser Gly Phe Thr Phe Ser Thr Tyr Ala Leu Ser Trp Val Arg Gln Thr Ala Asp Lys Arg Leu Glu Trp Val Ala Ser Ile Val Ser Gly Gly Asn Thr Tyr Tyr Ser Gly Ser Val Lys Gly Arg Phe Thr Ile Ser Arg Asp Ile Ala Arg Asn Ile Leu Tyr Leu Gln Met Ser Ser Leu Arg Ser Glu Asp Thr Ala Met Tyr Tyr Cys Ala Arg Glu Tyr Tyr Gly Tyr Val Gly Leu Ala Tyr Trp Gly Gln Gly Thr Leu Val Thr Val Ser Ala (2) INFORMATION FOR SEQ ID NO: 3:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 9 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (light chain CDR3-loop of the wild-type clone) (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 3:
Phe Gln Gly Ser His Val Pro Pro.Thr (2) INFORMATION FOR SEQ ID NO: 4:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 9 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (light chain CDR3-loop of the mutant clone 28) (xi) SEQUENCE DESCRIPTION: SEQ ID N0: 4:
Phe Gln Gly Ser Thr Val Pro Pro Thr (2) INFORMATION FOR SEQ ID NO: 5:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 10 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (heavy chain CDR3-loop of the wild-type clone) (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 5:
Glu Tyr Tyr Gly Tyr Val Gly Leu Ala Tyr (2) INFORMATION FOR SEQ ID NO: 6:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 10 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide {vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (heavy chain CDR3-loop of the mutant clone 44) (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 6:
Ala Tyr Tyr Gly Tyr Val Gly Leu Val Tyr (2) INFORMATION FOR SEQ ID NO: 7:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 16 amino acids {B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear {ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (light chain CDR1-loop of the wild type clone) (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 7:
Arg Ser Ser Gln Ser Ile Val His Ser Asn Gly Asn Ser Tyr Leu Glu (2) INFORMATION FOR SEQ ID NO: 8:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 16 amino acids (8) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (light chain CDR1-loop of the mutant clone A60) (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 8:
Arg Ser Ser Glu Val Ile Val Thr Arg Asn Gly Tyr Thr Pro Ile Glu (2) INFORMATION FOR SEQ ID NO: 9:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 16 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (light chain CDR1-loop of the mutant clone A58) (xi) SEQUENCE DESCRIPTION: SEQ ID N0: 9:
Arg Ser Ser Gln Arg Met Val Gln Arg Asn Gly His Thr Pro Leu Glu (2) INFORMATION FOR SEQ ID NO: 10:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 7 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (light chain CDR2-loop of the wild type clone) (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 10:
Lys Val Ser Asn Arg Phe Ser (2) INFORMATION FOR SEQ ID NO: 11:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 7 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (light chain CDR2-loop of the mutant clone 1D7) (xi) SEQUENCE DESCRIPTION: SEQ ID N0: 11:
Lys Ala Arg Tyr Arg Phe Pro (2) INFORMATION FOR SEQ ID NO: 12:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 5 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: {heavy chain CDR1-loop of the wild-type clone) (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 12:
Thr Tyr Ala Leu Ser (2) INFORMATION FOR SEQ ID NO: 13:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 5 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (heavy chain CDR1-loop of the mutant clone 2B5) (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 13:
Arg Tyr Ala Leu Ser (2) INFORMATION FOR SEQ ID NO: 14:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 9 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear {ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (light chain re-optimised CDR3-loop of the wild-type clone) (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 14:
Phe Gln Gly Ser His Val Pro Pro Thr (2) INFORMATION FOR SEQ ID NO: 15:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 9 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (light chain re-optimised CDR3-loop of the mutant clone A4) (xi) SEQUENCE DESCRIPTION: SEQ ID N0: 15:
Phe Asp Gly Ser His Val Pro Pro Lys (2) INFORMATION FOR SEQ ID N0: 16:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 9 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single {D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (light chain re-optimised CDR3-loop of the mutant clone A46) (xi) SEQUENCE DESCRIPTION: SEQ ID N0: 16:
Phe Glu Gly Ser Gln Val Pro Pro Thr (2) INFORMATION FOR SEQ ID NO: 17:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 10 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (heavy chain re-optimised CDR3-loop of the wild-type clone) (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 17:
Glu Tyr Tyr Gly Tyr Val Gly Leu Ala Tyr (2) INFORMATION FOR SEQ ID NO: 18:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 10 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (heavy chain re-optimised CDR3-loop of the mutant clone A52) (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 18:
Ala Tyr Tyr Gly Tyr Val Gly Leu Val His (2) INFORMATION FOR SEQ ID NO: 19:

(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 112 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (VL region of the Fab fragment of S73) (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 19:
Asp Val Val Val Thr Gln Thr Pro Leu Ser Leu Pro Val Ser Leu Gly Asp Gln Ala Ser Ile Ser Cys Arg Ser Ser Glu Val Ile Val Thr Arg Asn Gly Tyr Thr Pro Ile Glu Trp Tyr Leu Gln Lys Pro Gly Gln Ser Pro Lys Leu Leu Ile Tyr Lys Ala Tyr Lys Arg Phe Pro Gly Val Pro Asp Arg Phe Ser Gly Ser Gly Ser Gly Thr Asp Phe Thr Leu Lys Ile Ser Arg Val Glu Ala Glu Asp Leu Gly Val Tyr Tyr Cys Phe Asp Gly Ser Thr Val Pro Pro Lys Phe Gly Gly Gly Thr Lys Leu Glu Ile Lys (2) INFORMATION FOR SEQ ID NO: 20:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 112 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (VL region of the Fab fragment of S77) (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 20:
Asp Val Val Val Thr Gln Thr Pro Leu Ser Leu Pro Val Ser Leu Gly Asp Gln Ala Ser Ile Ser Cys Arg Ser Ser Glu Val Ile Val Thr Arg Asn Gly Tyr Thr Pro Ile Glu Trp Tyr Leu Gln Lys Pro Gly Gln Ser Pro Lys Leu Leu Ile Tyr Lys Ala Tyr Lys Arg Phe Pro Gly Val Pro Asp Arg Phe Ser Gly Ser Gly Ser Gly Thr Asp Phe Thr Leu Lys Ile Ser Arg Val Glu Ala Glu Asp Leu Gly Val Tyr Tyr Cys Phe Asp Gly Ser Thr Val Pro Pro Lys Phe Gly Gly Gly Thr Lys Leu Glu Ile Lys (2) INFORMATION FOR SEQ ID NO: 21:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 112 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (V,, region of the Fab fragment of S83) {xi) SEQUENCE DESCRIPTION: SEQ ID NO: 21:
Asp Val Val Val Thr G1n Thr Pro Leu Ser Leu Pro Val Ser Leu Gly Asp Gln Ala Ser Ile Ser Cys Arg Ser Ser Glu Val Ile Val Thr Arg Asn Gly Tyr Thr Pro Ile Glu Trp Tyr Leu Gln Lys Pro Gly Gln Ser Pro Lys Leu Leu Ile Tyr Lys Ala Tyr Lys Arg Phe Pro Gly Val Pro Asp Arg Phe Ser Gly Ser Gly Ser Gly Thr Asp Phe Thr Leu Lys Ile Ser Arg Val Glu Ala Glu Asp Leu Gly Val Tyr Tyr Cys Phe Asp Gly Ser Gln Val Pro Pro Lys Phe Gly Gly Gly Thr Lys Leu Glu Ile Lys (2) INFORMATION FOR SEQ ID NO: 22:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 112 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (V,, region of the Fab fragment of S88) (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 22:
Asp Val Val Val Thr Gln Thr Pro Leu Ser Leu Pro Val Ser Leu Gly Asp Gln Ala Ser Ile Ser Cys Arg Ser Ser Glu Val Ile Val Thr Arg Asn Gly Tyr Thr Pro Ile Glu Trp Tyr Leu Gln Lys Pro Gly Gln Ser Pro Lys Leu Leu Ile Tyr Lys Ala Tyr Lys Arg Phe Pro Gly Val Pro Asp Arg Phe Ser Gly Ser Gly Ser Gly Thr Asp Phe Thr Leu Lys Ile Ser Arg Val Glu Ala Glu Asp Leu Gly Val Tyr Tyr Cys Phe Asp Gly Ser His Val Pro Pro Lys Phe Gly Gly Gly Thr Lys Leu Glu Ile Lys (2) INFORMATION FOR SEQ ID NO: 23:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 118 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (V" region of the Fab fragment of S73) (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 23:
Glu Val Lys Leu Val Glu Ser Gly Gly Gly Leu Val Lys Pro Gly Gly Ser Leu Lys Leu Ser Cys Ala Ala Ser Gly Phe Thr Phe Ser Arg Tyr Ala Leu Ser Trp Val Arg Gln Thr Ala Asp Lys Arg Leu Glu Trp Val Ala Ser Ile Val Ser Gly Gly Asn Thr Tyr Tyr Ser Gly Ser Val Lys Gly Arg Phe Thr Ile Ser Arg Asp Ile Ala Arg Asn Ile Leu Tyr Leu Gln Met Ser Ser Leu Arg Ser Glu Asp Thr Ala Met Tyr Tyr Cys Ala Arg Ala Tyr Tyr Gly Tyr Val Gly Leu Val Tyr Trp Gly Gln Gly Thr Leu Val Thr Val Ser Ala (2) INFORMATION FOR SEQ ID NO: 24:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 118 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:

(C) INDIVIDUAL ISOLATE: (V" region of the Fab fragment of S83) (xi) SEQUENCE DESCRIPTION: SEQ ID N0: 24:
Glu Val Lys Leu Val Glu Ser Gly Gly Gly Leu Val Lys Pro Gly Gly Ser Leu Lys Leu Ser Cys Ala Ala Ser Gly Phe Thr Phe Ser Arg Tyr Ala Leu Ser Trp Val Arg Gln Thr Ala Asp Lys Arg Leu Glu Trp Val Ala Ser Ile Val Ser Gly Gly Asn Thr Tyr Tyr Ser Gly Ser Val Lys Gly Arg Phe Thr Ile Ser Arg Asp Ile Ala Arg Asn Ile Leu Tyr Leu Gln Met Ser Ser Leu Arg Ser Glu Asp Thr Ala Met Tyr Tyr Cys Ala Arg Ala Tyr Tyr Gly Tyr Val Gly Leu Val Tyr Trp Gly Gln Gly Thr Leu Val Thr Val Ser Ala (2) INFORMATION FOR SEQ ID NO: 25:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 118 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (VH region of the Fab fragment of S88) (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 25:
Glu Val Lys Leu Val Glu Ser Gly Gly Gly Leu Val Lys Pro Gly Gly Ser Leu Lys Leu Ser Cys Ala Ala Ser Gly Phe Thr Phe Ser Arg Tyr Ala Leu Ser Trp Val Arg Gln Thr Ala Asp Lys Arg Leu Glu Trp Val Ala Ser Ile Val Ser Gly Gly Asn Thr Tyr Tyr Ser Gly Ser Val Lys Gly Arg Phe Thr Ile Ser Arg Asp Ile Ala Arg Asn Ile Leu Tyr Leu Gln Met Ser Ser Leu Arg Ser Glu Aap Thr Ala Met Tyr Tyr Cys Ala Arg Ala Tyr Tyr Gly Tyr Val Gly Leu Val Tyr Trp Gly Gln Gly Thr Leu Val Thr Val Ser Ala (2) INFORMATION FOR SEQ ID NO: 26:
(i) SEQUENCE CHARACTERISTICS:
(A) LENGTH: 118 amino acids (B) TYPE: amino acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: peptide (vi) ORIGINAL SOURCE:
(C) INDIVIDUAL ISOLATE: (VH region of the Fab fragment of S77) (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 26:
Glu Val Lys Leu Val Glu Ser Gly Gly Gly Leu Val Lys Pro Gly Gly Ser Leu Lys Leu Ser Cys Ala Ala Ser Gly Phe Thr Phe Ser Arg Tyr Ala Leu Ser Trp Val Arg Gln Thr Ala Asp Lys Arg Leu Glu Trp Val Ala Ser Ile Val Ser Gly Gly Asn Thr Tyr Tyr Ser Gly Ser Val Lys Gly Arg Phe Thr Ile Ser Arg Asp Ile Ala Arg Asn Ile Leu Tyr Leu Gln Met Ser Ser Leu Arg Ser Glu Asp Thr Ala Met Tyr Tyr Cys Ala Arg Ala Tyr Tyr Gly Tyr Val Gly Leu Val His Trp Gly Gln Gly Thr Leu Val Thr Val Ser Ser

Claims (26)

1. A monoclonal antibody or derivative thereof which binds testosterone with affinity of at least about 109 M-1 and exhibits less than 0.05% cross-reactivity with dehydroepiandrosterone sulphate (DHEAS), determined on the basis of ED50 values.

2. A monoclonal antibody or derivative thereof as claimed in claim 1 which exhibits less than 0.01% cross-reactivity with DHEAS.

3. A genetically modified derivative of the monoclonal antibody as claimed in claim 1 or 2, which further exhibits less than 10% cross-reactivity with 5.alpha.-dihydrotestosterone (5.alpha.-DHT) and less than 5% cross-reactivity with androstenedione.

4. The genetically modified derivative as claimed in claim 3, which exhibits less than 8% cross-reactivity with 5.alpha.-DHT and no more than 2.5%

cross-reactivity with androstenedione.

5. A monoclonal antibody or derivative thereof as claimed in any one of claims 1 to 4 wherein the complementarity determining regions (CDRs) of the V
L
region are selected from the V L region CDRs of the Fab fragments S73, S77, and S88 as shown in Figure 12 and mutants thereof having one or more conservative substitutions which do not substantially affect binding capability, and the CDRs of the V H region are selected from the V H region CDRs of the Fab fragments S73, S77, S83 and S88 as shown in Figure 13 and mutants thereof having one or more conservative substitutions which do not substantially affect binding capability.

6. A monoclonal antibody or derivative thereof as claimed in claim 5 wherein the V L region is selected from the V L regions of the Fab fragments S73, S77, S83 and S88 as shown in Figure 12 and mutants thereof having one or more conservative substitutions which do not substantially affect binding capability, and the V H region is selected from the V H regions of the Fab fragments S73, S77, and S88 as shown in Figure 13, the V H region of Fab fragment S77 in which the C-terminal Serine is substituted by Alanine, and mutants of said V H regions having one or more conservative substitutions which do not substantially affect binding capability.

7. A testosterone-binding antibody fragment according to any one of claims 1 to 6.

8. An antibody fragment as claimed in claim 7 selected from the Fab fragments S73, S77, S83 and S88, fragments thereof and extensions thereof which retain binding specificity and affinity for testosterone as defined in claim 1.

9. A derivative according to any one of claims 1 to 6 which is a ScFv.

10. A monoclonal antibody or derivative thereof according to any one of claims 1 to 9 which is labelled.

11. A DNA encoding a monoclonal antibody or derivative thereof according to any one of claims 1 to 9 and fragments thereof which encode the CDRs of the V L and/or V H region.

12. A DNA as claimed in claim 11 in the form of a vector.

13. A DNA as claimed in claim 12 which is an expression vector capable of expressing a monoclonal antibody or derivative thereof as claimed in any one of claims 1 to 9 or at least one antibody chain of said antibody or antibody derivative.

14. A host cell containing a DNA according to any one of claims 11 to 13.

15. A host cell as claimed in claim 14 capable of expressing a monoclonal antibody or derivative thereof as claimed in any one of claims 1 to 9 or at least one antibody chain of said antibody or antibody derivative.

16. A method of preparing a monoclonal antibody or derivative thereof as claimed in any one of claims 1 to 9 which comprises culturing a host cell or host cells according to claim 15 capable of expressing the required antibody chain(s) and recovering said antibody or antibody derivative, if necessary after initially recovering and combining component chains.

17. A method of preparing a monoclonal antibody or derivative thereof as claimed in any one of claims 1 to 9 wherein at least a portion of said antibody or antibody derivative is produced synthetically.

18. A method as claimed in claim 16 or claim 17 which further comprises the step of labelling said antibody or antibody derivative and/or attachment of said antibody or antibody derivative to a solid support.

19. A phage or microbial cell which presents an antibody fragment as claimed in claim 7 or a ScFv as claimed in claim 9 as a fusion protein with a surface protein.

20. A method of selecting an antibody fragment as claimed in claim 7 or a ScFv as claimed in claim 9 wherein said antibody fragment or ScFv is selected from a display library of antibody fragments and/or ScFv antibody derivatives containing a phage or cell according to claim 19.

21. A method of obtaining a monoclonal antibody or derivative thereof according to any one of claims 1 to 9 which comprises selecting a testosterone-binding antibody or antibody fragment, mutating the V L and/or V H regions of said antibody or fragment and selecting from the mutants thus obtained an antibody or antibody fragment having the required affinity and specificity for testosterone.

22. A method of assaying testosterone wherein a monoclonal antibody or derivative thereof according to any one of claims 1 to 10 is employed.

23. A kit comprising an antibody or derivative thereof according to any one of claims 1 to 10 in a suitable container for transport and storage.

24. A monoclonal antibody or derivative thereof according to any one of claims 1 to 10 which is attached to or coated upon a solid support.

25. A monoclonal antibody or derivative thereof according to any one of claims 1 to 10 for use in in vivo diagnosis.

26. A monoclonal antibody or derivative thereof according to any one of claims 1 to 10 for use in in vivo treatment.