Ba£k _jj r£und

The human CD2 (Til) molecule is a 50KD surface glycoprotein expressed on >95% of thy ocytes and virtually all peripheral T lymphocytes which mediates both adhesion between these cells and their cognate partners as well as subsequent activation events. Specific combinations of antibodies against the surf ce-bound molecule can activate IL-2 dependent T cell proliferation, helper T cell function and cytotoxicity by natural killer cells and cytolytic T lymphocytes (Meuer, S.C. , ______ ___!___ ,

Cell 3^:897-906 (1984); Brottier, P. , et al^, J^ Immunol. , 135:1624-1631 (1985); Siliciano, R.F. , et λ__. _» Nature 317 : 428-430 (1985)) in the absence of cellular adhesion. In addition, thymocyte activation can be mediated via CD2 (Fox, D.A. , e t H± - _> J. Immunol. 134:330-335 (1985); Denning, S.M. , H ≤li. » ii_l2S2HS°I_i 139:2573-2578 (1987)). The role of CD2 in approximation of T cells to various cell types including human thymic epithelial cells, B cells, target cells and sheep erythrocytes has been demonstrated to depend on direct interaction between CD2 and the broadly distributed human lymphocyte function-associated, antigen 3 (LFA-3) surface glycoprotein or its sheep homologue, TUTS (Denning, S.M. , et l^. , ^..Immunol^ 133: 2573-2578 (1987); Hunig, T. , J^._E £^._Med^ 162 : 890-901 (1985); Hunig, T. , et al _Λ , Nature 325: 298 - 301 (1987); Shaw, S- et a , Nature 3_23: 262-264 (1986); Selvaraj , P. , et aL., Nature 3_2.£: ⁴⁰ °- ⁴⁰³ (1987) ; Vollger, et a 1___ , J. Immunol. 138:358-363 (1987)).

Biochemical analyses using specific monoclonal antibodies show that CD2 is T lineage-specific and exists on the cell surface in several differentially glycosylated forms (Howard, F.D., e_t al^, J. Immunol. 126:2117-2122 (1981); Brown, M.H. , et al^, In Leucocyte Typing III ed McMichael, A.J. , Oxford University Press pp. 110-112 (1987); Sayre, P.H. , et al. , Proc. Natl. Acad. Sci. USA 4:2941-2945 (1987)). CD2 cDNA clones predict a cleaved signal peptide of 24 amino acid residues, an extracellular segment of 185 residues, a transmembrane domain of 25 residues and a cytoplasmic region of 117 residues (S yre, P.H. , et al_. , Proc. Natl. Acad. Sci^ USA .84:2941-2945 (1987); Sewell, W.A. , et al. , Proc^ Natl. Acad. Sci. USA 83:8718-8722 (1986); Seed, B. and A. Aruffo , Proc.. _Natl _Jι Acad. Sci. USA 4:3365-3369 (1987); Clayton, L.K. , et a , Eur. J. Immunol. 17:1367-1370 (1987)). The corresponding genomic organization reveals a single exon encoding the signal peptide (less four residues), two exons encoding the extracellular segment, one exon encoding the transmembrane domain and charged membrane anchor segment, and one exon encoding the cytoplasmic region (Diamond, D.J. , et ___1___ , Proc. Natl. Acad. Sci. USA 85:1615-1619 (1988)).

ummarγ_£f_the_I _; nvent^£n

This invention pertains to a soluble peptide having a lymphocyte function-associated antigen 3 (LFA-3) binding domain and antigenic epitopes recognized by antibodies raised against native

surface CD2 on resting T lymphocytes. The soluble peptide is capable of forming at least two intra¬ molecular disulfide bonds and binds the surface- bound CD2 legand, LFA-3. The soluble peptides of this invention exist as monomers in an aqueous medium. Preferably, the soluble peptides have an amino acid sequence encoded by the two extracellular segment exons of the gene encoding human CD2 glyco¬ protein. Most preferably, the soluble peptides have about 182 amino acid residues encoded by the two extracellular segment exons of the gene encoding human CD2 glycoprotein. Soluble peptides of this invention can be made by enzymatic fragmentation, peptide synthesis, or recombinant DNA technology. They can be used to block T cell function which is dependent upon antigen activation.

Bri f_D £ri£ti£n_£f__ h _Drawin£

Figure 1 shows the amino acid sequence and exon organization of human CD2.

Figure la shows the DNA sequence of human CD2 glycoprotein.

Figure 2 shows the structure of recombinant soluble Til „ . native CD2 and its genomic ex2 ° organization.

Figure 3 shows the construction of expression plasmid for production of the CD2 external segment molecule Til „ . ex2

Figure 4 shows SDS-PAGE analysis of purified, radioiodinated and endoglyosidase digested Til _.

° ex2

Figure 5 shows equilibrium sedimentation data.

Figure 6 shows the circular dichroism spectra of Til _. ex2

Figure 7 shows the competitive inhibition of radioiodinated Til „ binding of JY cells. ex2 °

Figure 8 shows the saturation binding of Til „ to JY cells and Scatchard analysis.

Detailed Description of the Invention

This invention pertains to a soluble peptide having a lymphocyte function-associated antigen 3 (LFA-3) binding domain and antigenic epitopes recognized by antibodies raised against native surface CD2 on resting T lymphocytes. The soluble peptide is capable of forming at least two intra¬ molecular disulfide bonds and binds the surface- bound CD2 ligand, LFA-3. Specifically, the anti¬ genic epitopes which are localized on the peptide are Til. , TIL and llg . The soluble peptides are further capable of inhibiting CD2-mediated T cell activation and are soluble in aqueous medium and exist as monomers in the same.

In one embodiment, the soluble peptides have an amino acid sequence encoded by the two extracellular segment exons of the gene encoding human CD2 glycoprotein. The soluble peptides are capable of binding the surface-bound CD2 ligand, LFA-3, and can react with antibodies raised against native CD2 on the surface of human T cells. . The soluble peptides are further capable of inhibiting CD2-mediated T cell activation. The peptides exist as monomers in an aqueous medium.

In the preferred embodiment, the soluble peptide comprises about 182 amino acid residues

encoded by the two extracellular segment exons of the gene encoding human CD2 glycoprotein. The soluble peptide is monomeric in aqueous medium and comprises four cysteine residues located in the carboxy-terminal region of the amino acid sequence. The peptide is capable of forming at least two sets of intramolecular disulfide bonds between the amino- terminal cysteines and the carboxy-terminal cysteines. Preferably, the peptide has the following amino acid sequence:

The soluble peptide of about 182 amino acid residues in length has an LFA-3 binding domain and comprises the antigenic epitopes TIL , TIL and Til- which are localized on the soluble peptide. The epitopes, which are localized on this peptide are antigenic determinants which are recognized by antibodies against the native surface CD2 structure on resting T lymphocytes. Thus, the peptides of this invention can be used to inhibit CD2-mediated T cell activation.

The above amino acid sequence corresponds with a portion of the naturally occurring extracellular domain of CD2 that is capable of binding LFA-3, inhibiting CD2-mediated T cell activation and reacting with antibodies raised against native CD2 on the surface of human T cells. Amino acid sequences embraced by this invention include analogous or homologous sequences which encode proteins capable of binding LFA-3, inhibiting CD2-mediated T cell activation and reacting with antibodies raised against native CD2. In addition, the peptide structure can be modified by deletions, additions, inversions, insertions or substitutions of one or more amino acid residues in the sequence to yield peptides having the characteristics of the peptides of this invention. All such modifications of the above amino acid sequence are embraced by this invention without essentially detracting from the properties of the peptide, i.e. , the capacity of the peptide to bind LFA-3, inhibit CD2-mediated T cell activation and react with antibodies raised against native CD2 on the surface of human T cells.

Naturally occuring allelic variations and modi ications are included within the scope of the invention so long as the variation does not substantially reduce the ability of the peptide to bind LFA-3 and inhibit CD2-mediated T cell activation.

The soluble peptides of this invention carry epitopes recognized by antibodies against native surface CD2 structure on resting T lymphocytes and interacts specifically with the surface-bound CD2 ligand, LFA-3. Additionally, the soluble peptide exists as a monomer in aqueous medium and includes a proteolytically-resistant amino-terminal fragment encoded by the first extracellular segment exon of the gene encoding human CD2 glycoprotein. The proteolytically resistant fragment comprises about 100 amino acid residues which correspond with the amino- terminal portion of the extracellular domain of human CD2 glycoprotein and is capable of inhibiting CD2-mediated T cell activation. The 100 amino acid fragment has been described in U.S. Patent Application Serial No. 07/293,330 filed January 4, 1989, by Reinherz et a ______ , the teachings of which are incorporated herein by reference. The entire amino acid sequence and exon organization of human CD2 glycoprotein is shown in Figure 1. This amino acid sequence of human CD2 was deduced from cDNA (Sayre, P.H. , et a ______, Proc^ atl^Acad^Sci^

US_A 4:2941-2945 (1987) ; Sewell, W.A. , et aL., Proc^ Na11^_Acad^__S£i ___J_SA £3 : 8718- 8722 (1986) ; Seed, B. and A. Aruffo , Pro£__Natl __A£ad^_££i_ _: __U£A £4:3365-3369 (1987)) and genomic clones (Diamond, t _________, Proc^_Natl^_Acad _i _Sci _Λ _ϋSA £.5:1615-1619

(1988)) .

This invention also pertains to an isolated DNA sequence that encodes a soluble monomeric human CD2 peptide having an LFA-3 binding domain and antigenic epitopes recognized by antibodies raised against native CD2 on the surface of human T cells and is capable of forming at least two intramolecular disulfide bonds. Specifically, the antigenic epitopes are TIL , Til. and TIL . The peptide that is encoded by the isolated DNA of this invention and is capable of binding LFA-3, inhibiting CD2-mediated T cell activation and reacting with antibodies raised against native CD2 on the surface of human T cells. In a preferred embodiment, the isolated DNA sequence encodes a soluble peptide having an amino acid sequence encoded by the two extracellular segment exons of the gene encoding the human CD2 glycoprotein. The encoded peptide Is capable of binding LFA-3, inhibiting CD2-mediated T cell activation and reacting with antibodies raised against native CD2. Preferably, the isolated DNA of the invention encodes the 182 amino acid sequence shown above or substantial coding equivalent thereof. The DNA sequence can be modified by deletion, insertion or substitution of nucleotides to yield peptides which exhibit substantially the same properties of the above peptide of about 182 amino acid residues. All such modifications of the DNA sequence are within the scope of this invention so long as the DNA sequence encodes a soluble peptide that is capable of binding LFA-3 and inhibiting CD2-mediated T cell activation. The isolated DNA sequences of the invention can be made

using recombinant DNA technology or chemically synthesized. The DNA sequence for CD2 glycoprotein is shown in Figure la.

This invention further pertains to a recombinant expression vector comprising the DNA sequence encoding a soluble, monomeric human CD2 peptide having an LFA-3 binding domain and the antigenic epitopes Til. , TIL and Til- . The antigenic epitopes are recognized by antibodies raised against native CD2 on resting T cells and interacts specifically with LFA-3. The encoded soluble protein is capable of forming at least two intramolecular disulfide bonds. Alternatively, the expression vector comprises a DNA sequence encoding a soluble CD2 protein encoded by the two extra¬ cellular segment exons of the gene encoding human CD2 which is capable of inhibiting CD2-mediated T cell activation and LFA-3 binding. Preferably, the expression vector is a baculovirus transfer vector and comprises a DNA sequence which encodes a peptide of about 182 amino acid residues as shown above. Other vectors, however, may be used, including prokaryotic and eukaryotic expression systems. Modifications of the peptide in which amino acid residues have been deleted, inserted or substituted without essentially detracting from the properties of the peptide are embraced by the invention. The invention further pertains to cells transformed with the above expression vector.

Soluble peptides of this invention can be made by enzymatic fragmentation of human CD2 glycoprotein

or a portion thereof, by peptide synthesis or recombinant DNA technology. Preferably, the soluble CD2 peptides will be produced by inserting DNA encoding a peptide sequence which is capable of binding LFA-3 and inhibiting CD2-mediated T cell activation (e.g. , CD2 DNA which represents the desired amino acid sequence of the extracellular domain of CD2) into an expression vector. The transformed cells then express the soluble human CD2 peptide encoded by the two extracellular segment exons of the gene encoding the human CD2 glycoprotein. The peptide expressed by the transformed cells is capable of binding LFA-3 and inhibiting CD2-mediated T cell activation. In addition to genetic engineering techniques for synthesizing soluble peptides of the invention, the soluble peptides can be synthesized directly by pro¬ cedures of chemical protein synthesis. For example, the above 182 amino acid sequence or modified equivalent thereof can be synthesized by the solid phase procedure of Merrifield.

This invention further pertains to a method of inhibiting T cell activation, comprising the step of administering to a patient, a soluble peptide having an LFA-3 binding domain, antigenic epitopes recog¬ nized by antibodies raised against native CD2 on the surface of human T cells and is capable of forming at least two intramolecular disulfide bonds. In another embodiment, a patient is administered a solution containing a soluble peptide having an amino acid sequence encoded by the two extracellular

segment exons of the gene encoding human CD2 glyco¬ protein. The peptide is capable of binding LFA-3, inhibiting T cell activation and reacting with antibodies raised against the native CD2 protein. Preferably, the soluble peptide comprises a sequence of about 182 amino acid residues which corresponds with the portion of the extracellular domain of human CD2 glycoprotein or fragment thereof which is capable of binding LFA-3 and inhibiting CD2-mediated T cell activation. In one embodiment of the method, the soluble peptide can be administered intravenously.

The soluble CD2 peptides of this invention generally will bind to human lymphocytes and human red blood cells which express a homolgous set of surface structures. The soluble peptides are also capable of competing with the naturally-present CD2 on the surface of a human lymphocyte, thus interfering with the ability of the lymphocyte to make contact with its target cells (if the lymphocyte is a cytolytic cell), or with macrophages having CD2 binding structures which permit the cell-to-cell contact necessary for lymphocyte proliferation. To test a soluble peptide for the ability to inhibit lymphocyte proliferation, or the cytotoxic effector. function, the soluble peptide is contacted with the lymphocytes prior to stimulation with mitogen, and degree of proliferation is measured, using standard techniques, and the result compared to a control in which the soluble peptide was not used.

The soluble peptides of this invention can be used in a variety of diagnostic and therapeutic applications in which the CD2 surface glycoprotein Is expressed on the surface of many human T cell malignancies, e.g. , T cell leukemias and lymphomas . In addition, autoimmune diseases, e.g. , rheumatoid arthritis and Systemic Lupus Erhthmatosis (SLE) , are characterized by the presence in the blood and lymph of large numbers of CD2-bearing T cells. Rapid cell turnover in these disease states can cause the shedding of the CD2 molecule into the bloodstream.

The CD2 soluble peptides of this invention can be used as an immunogen to produce polyclonal or monoclonal anti-CD2 antibodies, using conventional techniques. These antibodies can be labeled with any conventional label, e.g., radioisotopes , and used in conventional immunoassay methods to measure serum CD2 levels and thus monitor patients having T cell associated diseases. Particularly sensitive ELISA-type assays will employ two anti-CD2 anti¬ bodies, each to a different antigenic determinant on the surface of CD2 , in a sandwich format.

The disease states which can be treated using the soluble peptides of this Invention Include medical conditions characterized by unwanted activity of the immune system which results in excess T cell activation, which plays a key role in the amplification of the immune response. These conditions include SLE; juvenile onset diabetes; multiple sclerosis; allergic conditions; inflammatory conditions such as exze a, ulcerative

colitis, inflammatory bowel disease, and Crohn's disease; and allograft rejection (e.g. , rejection of a transplanted heart or kidney) . The soluble CD2 peptides compete with the surface-bound CD2 for its ligand on target cells thus dampening immune response amplification. The soluble CD2 peptide admixed with a pharmaceutically acceptable carrier substance such as saline, is administered intraveneously to a human patient in an effective amount, e.g. , 20 μg to 500 μg per kg body weight. Preferably, the soluble peptide admixed with a pharmaceutically acceptable carrier comprises about 182 amino acid residues which correspond to a portion of the extracellular domain of CD2 that is capable of binding LFA-3 and inhibiting CD2-mediated CD2 activation. For some conditions, a soluble CD2 peptide can be administered directly to the site where needed most; for example, a soluble CD2 peptide can be injected directly into the inflammed joint of a human patient suffering from rhematoid arthritis .

This invention is further illustrated by the following Exemplification.

EXEMPLIFICATION

The term Tilex„ , as referred to in the

Exemplification, is defined herein to be a soluble protein having an amino acid sequence encoding the two extracellular segment exons and a codon (182-Lys) derived from the transmembrane domain exon of the CD2 glycoprotein. The entire amino acid

sequence for CD2 has been described. See U.S. Patent Application Serial No. 932,871, filed November 18, 1986, by Reinherz et al . , the teachings of which are incorporated herein by reference.

Baculovirus Transfer Vector Plgsmid_Construct _ or

Tilex2-

The plasmid pAc373/Tll „ was constructed by digestion of pGEM-4-Sl, a pGEM derivative containing a 950 bp fragment of the CD2 cDNA PB2 (Sayre et al. , Proc. N tl..._Acad _ι . _τ Sci _. USA £4:2941-2945 (1987)) with PvuII, which digests the cDNA at nucleotide position 628 near the start of the transmembrane region. A double-stranded synthetic oligonucleotide linker:

CTGTCCAGAGAAATAAGGATCCT GACAGGTCTCTTTATTCCTAGGA

containing the last base for the Ser-178 codon, codons for Cys-Pro-Glu-Lys , a stop codon and a BamHI restriction site was synthesized and ligated to the blunt PvuII ends. After BamHI digestion, the insert was cloned into the BamHI site of the pAc373 baculovirus transfer vector (Smith ____£__________ , Annu.

Re __Immun£^ 2:319-333 (1985)). All restriction enzymes were produced from New England Biolabs (Beverly, MA) .

Re bina _£r ei£_£r du ti iι

Transfer of the Til _ sequences from the plasmid vector to the AcNPV genome to generate recombinant baculovirus Tilexz--AcNPV was ac-

complished essentially as described (Smith e t aL_, Proc. Natl. Acad. Sci. USA, £2:8404-8408 (1985) ;

Hussey e t aL_, Nature, ££1:78-81 (1988). Metabolic

35 labeling with S-cysteine of Til - -AcNPV- infected

SF9 cells was carried out as described (Hussey t:

≤I . Na ure ££1:78-81 (1988). Metabolically labeled culture supernatants were harvested, microfuged for

10 min. and subjected to immunoprecipitation for 16 h at 4°C with a monoclonal anti-CD2 antibody

(anti-TH ₁ , 3T4-8B5) linked to Affigel-10 (Biorad) beads (10 mg monoclonal antibody/ml gel) . After immunoabsorption the beads were washed twice with lysis buffer and bound material eluted with 0.1 M glycine, HC1 buffer, pH2. Eluates were analyzed by

SDS-PAGE in 12.5% gels. Large protein production was performed as described (Hussey £t aL_, Nature

££L78-81 (1988)) except that proteins were purified over an anti-HL (3T4-8B5) immunoaffinity column.

Z∑£__£ iS__5ϊi£_ϊ£≤. 3H S£iΞ£

Proteins were prepared for microsequencing by electrophoresis on 12.5% polyacrylamide gels, followed by electroblotting onto polyvinylidene difluoride membranes according to the method of Matsudaira, J^_Bi£ __Chem^ 262 : 10035-10038 (1987) . After visualization with Coomassie blue, stained bands were excised and loaded onto an Applied Biosystems (Foster City, CA) 470A sequencer and sequenced using the 03RPTH program.

Endoglycosidase F digestion of Til „

1 μg samples of purified Til „ dialyzed against PBS were incubated in Endo-F buffer (0.1 M sodium phosphate, 1% NP-40, 1% 2-mercaptoethanol, 50 mM EDTA, 1 mM phenylmethyl-sulfonyl fluoride, pH 6.1) in the presence of 0.7 units Endo-F (NEN) in 7 μl reaction volumes. Digestion was arrested at the indicated times by addition of 15 μl SDS sample buffer, boiling for 5 min. and freezing at -20°C prior to analysis by 12.5% SDS-PAGE.

Equilibriu sedimentation

Sedimentation studies were performed using the short column, high speed meniscus-depletion method of Yphantis Btochem^ £:297-317 (1964); and Richardson et al. , Biochem. J. 1£5: 87-92 (1973)). Standard double-sector cells, equipped with sapphire windows, 4 mm solution column lengths and a tempera¬ ture of 21°C were used.

Papain digestion

Papain (32 ng) was added to 8 μg samples of Til ₂ at 0.5 mg/ l in PBS containing 10 mM DTT for an enzyme :protein ratio of 1:250. Samples were incubated at 37°C for 15, 30, 45 or 60 min. Digestions were stopped by the addition of SDS buffer and boiling for 5 min. Samples were electrophoresed on a 12.5% polyacrylamide gel and stained with Coo assie blue.

Ci _: r ular_di£hr ₁ m

Far ultraviolet CD spectra were obtained on an Instruments SA Jobin Yvon circular dichrograph calibrated with (+) 10-camphorsulfonic acid and epiandosterone . Measurements were taken at 25, 50 and 80°C +/-0.1°C in 10 mM sodium phosphate pH 7.2 in a 1 mm cell. All spectra represent an average of 3 to 5 individual spectra with data taken at 0.5 mm intervals using a 10 second response time for each point. Protein concentrations were determined by quantitative amino acid analysis of aliquots taken from the sample ceils.

R t£ _inhibi£i£n_a jsay

Jurkat cells were washed with SMEM/2% FCS

(wash) and resuspended at 10 /ml. Sheep erythro- cytes were washed twice in HBSS and resuspended in wash to 5% v/v. 10 μl of sheep erythrocytes were aliquoted into 12-75 mm plastic tubes and 100 μl wash, Til - protein or control soluble CD4 T4 - ex2 ^r exl protein (Hussey et aL_, Nature ££1:78-81 (1988) was added, followed by incubation at 4 ^β C for 30 min. Subsequently, 20 μl of Jurkat cells were added after 5 min. , centrifuged at 800 rpm in a Sorvall RT6000, followed by incubation at 4°C for 1 h. The cell mixture was gently resuspended and mounted on glass slides with cover slips and rosette formation assessed on a Zeiss photomicroscope .

Radiolabeled Til - binding assays

The purified recombinant soluble CD2 extracellular domain molecule Til - was ex2 radioiodinated as follows: 50 μl Til „ (1 mg/ml) dialyzed against PBS was labeled with 10 μl immobilized lactoperoxidase/ glucose oxidase (Enzymobeads ; Biorad Laboratories, Richmond, CA) in

40 mM sodium phosphate, pH 7.2, 0.4% glucose and 1

125 mCI I for 5 min. After quenching the reaction for 20 min. with 20 mM sodium iodide and 0.02% sodium azide, 20 μl FCS was added and the free iodine separated on a 1.ml Bio-Gel P-6 column

(Biorad Laboratories) conditioned with 0.2M sodium phosphate pH 7.2, 10% FCS and run in the same buffer.

Cold competition studies: 5 x 10 cpm radio- o labelled Til - (2.8 x 10 cpm/nmole) was added at

0.1 μM to 1.8 x 10 JY cells overlayed onto 0.2 ml of a 1.5:1 mixture of dibutyl phthalate:dioctyl phthalate (Aldrich Chemical Co. , Milwaukee, WI) in

0.5 ml plastic tubes as described in Teshigawara e£ al. , J. Exp. Medicine 165 :223-238 (1987).

Increasing concentrations of unlabelled Til „ or ex2

T4 , were added in a final volume of 200 μl in RPMI exl ^

1640/10% FCS. After 1 h incubation at 4°C, the tubes were centrifuged (8,500 g for 1 min.), the tips of the tubes containing the cell pellets were cut, and the cell-bound and free radioactivity were determined in a gamma counter. To some tubes, anti-LFA-3 antibody TS2/9 (generously provided by Dr. Timothy Springer, Dana Farber Cancer Institute)

was added at 50 μg/ml as a separate determination of nonspecific binding. This concentration of anti- LFA-3 was independently shown to give maximal inhibition of Til - binding. Specific activity was calculated using a MW for Til „ of 30,000. Protein

° ex2 concentrations were determined by quantitative amino acid analysis of standard Til - samples. ^J ex2 ^r

Subsequent samples were compared to standards by

Coomassie blue staining of two-fold dilutions of standard and test samples run on the same gel and analyzed by densitometry.

Saturation binding studies: Increasing concentrations of radiolabelled Til - (1.31 x 10

, ex2 cpm/nmole) were added to 2.6 x 10 JY cells in the presence or absence of 50 μg/ml anti-LFA-3 antibody to determine nonspecific binding. Binding was carried out as above and the dissociation constant determined by Scatchard analysis after subtraction of nonspecific binding determined in the presence of anti-LFA-3.

RESULTS

Production and purification of Til _

A construct for expression of a soluble fragment of CD2 that included all the residues encoded by the leader and two extracellular segment exons was designed (Figure 2, exons 1-3) .

Figure 2 shows a comparison of the 182 extra¬ cellular CD2 amino acids comprising Til - (top) to CD2 protein structure (middle) . The positions of cysteine residues (C) , carbohydrate addition sites

(CHO) , the CD2 leader segment (L) and the CD2 transmembrane domain (TM) are indicated. In the CD2 gene (bottom) , exon 1 corresponds to CD2 amino acid residues -24 to -4, exon 2 to residues -4 to 104, exon 3 to residues 104 to 181, exon 4 to residues 181 to 222, and exon 5 to residues 222 to 327 (Diamond et al. , PNAS USA £5:1615-1619 (1988).

The plasmid pAc373/TH „ was constructed and encodes 182 amino acids of the predicted CD2 ex¬ ternal segment including all of the residues derived from the two extracellular exons (Figure 2) and part of one codon (for Glu-181) and all of a second codon (for Lys-182) derived from the transmembrane domain exon. This construction, thus, includes all four extracellular cysteine residues located in domain II of CD2 and thereby avoids problems associated with intermolecular disulfide exchange observed with a previous construction (Richardson, ______ aL_, Proc.

Natl. Acad. Sci USA ££: 5176-5180 (1988)).

Plasmid pAc373/TH _ was used to co-transfect SF9 with AcNPV baculoviral DNA. Recombinant baculo¬ virus, termed Til „-AcNPV, were selected, purified and used to infect small-scale cultures for metabolic labeling. Immunoprecipitation of radio- labelled supernatants with anti-TIL (3T4-8B5), an anti-CD2 specific monoclonal antibody (Meuer et aL_,

Cell 36:897-906 (1984), verified that Til -AcNPV — ex directed the production of a recombinant CD2 molecule in SF9 cells (data now shown) .

Til _ -AcNPV was therefore used to infect liter ex2 cultures for the production of large amounts of protein. Til - protein was purified from infected

cell supernatants by affinity chromtography on an anti-Til column.

The construction of expression plasmid for production of the CD2 external segment molecule Til - is shown in Figure 3. The plasmid pGEM-4-Sl carries a 950 bp fragment of the CD2 cDNA. After digestion of pGEM-4-Sl with PvuII, ligation of the double- stranded linker and further digestion with BamHI, the cDNA insert was isolated and ligated into the BamHI-digested baculoviral transfer vector pAc373. The resulting plasmid pAc373/TH _ encodes 182 amino acids of the mature CD2 extracellular segment. The promoter for the polyhedrin gene in the pAc373 transfer vector is shown by the black box and the polyhedrin gene is indicated by the open box. The 950 bp CD2 coding fragment in pGEM-4-Sl is shown in a stippled box. The position of the T7 polymerase promoter in the pGEM vector is shown.

Biochemical characterization of Tilex2

Figure 4 shows the purification, radioiodination and endoglycosodase digestion of

Til -. Lanes a-d contain 1 λg Til - purified from ex2 ^& ex2 ^r large scale cultures of SF9 cells infected with Til --AcNPV was analyzed by Coomassie staining on a 12.5% polyacrylamide gel in the presence of 50 mM DTT (lane a) or in nonreducing conditions (lane b). An aliquot of Til - radioiodinated with solid-phase lactoperoxidase/glucose oxidase was analyzed on the same gel in the presence (lane c) or absence (lane d) of 50 Mm DTT by autoradiography . Lanes e-k

contain 1 μg purified Til _ which was digested with

9.7 units Endo-F for varying amounts of time and the reaction stopped by the addition of SDS sample buffer prior to SDS-PAGE analysis and Coomassle staining. Lane e contains no enzyme; f, simultaneous addition of enzyme and sample buffer; g-j , 1 min, 5 min, 1 h, 18 h digestion; k, 0.7 units

Endo-F alone.

Til - migrates as a well-demarcated doublet in ex2 ° both reducing and non-reducing conditions in SDS- PAGE (Figure 4, lanes a and b). Two well-separated bands at 30-31KD are seen in the presence of 50 M DTT (lane a), which migrate at 27-28KD in the absence of reducing agent (lane b) . The clear-cut decrease in electrophoretic mobility after reduction with DTT strongly indicates that Til - contains intrachain disulfide bridges; it does not form interchain bridges. Although not shown, micro-

35 sequencing analysis of S-cysteine labeled peptides verifies _^ that there are two sets of intrachain disulfide bonds in Til „ between the amino-terminal ex2 cysteines and carboxy-terminal cysteines.

To investigate the difference between the two bands representing Til . , 160 pmole of purified protein was separated by SDS-PAGE and blotted onto a PVDF membrane (Matsudiara, J. BioL Chem. 262:10035-10038 (1987)). The upper and lower bands were cut separately from the membrane for amino- terminal sequencing. Each band yielded the CD2 amino-terminal sequence, suggesting that they differ from one another by post-translational modification.

As shown in Figure 4, endoglycosidase digestion generates at least five distinct bands. After short digestion times, two new lower molecular weight species are generated (lanes f-h) . Some glycans on

Til - are apparently . quite susceptible to digestion since even after simultaneous addition of enzyme and

SDS sample buffer, these new species are generated

(lane f) . After 1 h digestion, most of the Til _

° ex2 protein is digested to a 25KD species (line i); overnight digestion results in complete digestion to a single band at 25KD (lane j). Note that the band of approximately 45KD size represents the Endo-F enzyme since it appears in lane k, where an equiva¬ lent amount of enzyme alone has been analyzed. The origin of the faint band at 48KD in lane j is unclear.

To determine whether Til - exists as a non- ex2 covalently linked multimer in aqueous solution, it was subjected to equilibrium sedimentation by the high-speed meniscus depletion method (Yphantis , -_ __. .£._1_1E__. : 297-317 (1964) in both aqueous and dissociating conditions. As shown in Figure 5, the calculated molecular weights for both conditions are very similar (25.3KD in aqueous solution vs. 24.7KD in dissociating conditions). This result demonstrates that Til „ exists as a monomer in ex2 solution.

Figure 5 shows the equilibrium sedimentation analysis as a plot of log (fringe displacement) against square of distance from center of rotation,

2 r . Til - (0.05%) was analyzed by sedimentation

equilibrium on a Beckman model E analytical ultra- centrifuge in aqueous solution (PBS) at 30,000 rpm

3 (ω - 3.142 X 10 rad/sec) or in dissociating conditions (6 M guanidine hydrochloride) at 44,000

3 rpm (ω - 4.608 x 10 rad/sec). Data were obtained at 22°C (PBS) or 23 ^β C (guanidine-hydrochloride) .

Assuming a partial specific volume of 0.725, the calculated molecular weight from the displayed slope and using the equation: MW — 4.606 RT x slope/[ω(l - vbar ω) ] is 25,315 daltons In aqueous solution. b ____.___ ^* Assuming a partial specific volume (denoted as v ) of 0.725 - 0.1 in 6 M guanidine hydrochloride (Richardson, N.E. , et: al. , Biqchem J. l£5_:87-92 (1973)), the calculated molecular weight is 24,736 daltons in dissociating conditions.

The expression of CD2 epitopes was investigated by immunoprecipitation analysis. The Til _ mole¬ cule can be immunoprecipitated by both anti-Til. and a second monoclonal antibody to a different epitope termed anti-Til- . However, Til „ is not immuno- . e Z precipitated by the anti-CD2 antibody, anti-Til-, which defines an activation specific epitope on CD2.

Nevertheless, Til - was able to Inhibit the binding ex2 ^& of anti-TlL-FITC to the TIL + Jurkat cell line at a concentration of 10 μm, implying its presence on Til _ (data not shown). These results also suggest that the affinity of anti-Tll_ for its epitope is low.

Secondary structure predictions suggest the presence of both α-helical and J-sheet structure in the CD2 external domain (Clayton e£ aL_, Eur . J_^

ImmunoL 17:1367-1370 (1987)). To more directly

predict secondary structural characteristics, the

Til - molecule was evaluated by circular dichroism. ex2 ^J

Figure 6 shows the circular dichroism spectras of

Til _. Far ultraviolet circular dichroism spectra ex2 ^r represent the average of 3-5 individual spectra with data taken at 0.5 nm wavelenght intervals in lOmM sodium phosphate, pH 7.2 A: spectrum at 25°C of untreated Til ^• B: spectrum of Til _ reduced with ex2 ^r ex2

10 μM DTT and alkylated with 20μM iodacetamide ; C: thermal denaturation of the sample in A; D: thermal denaturation of the sample in B.

As shown in Figure 6A, the far ultraviolet CD spectrum of Til - in 10 mM sodium phosphate shows a positive absorption maximum at about 200 nm (Δe =

0.459) , a negative minimum at 215 nm (Δe = -1.94) and shoulder at 225 nm (Δe - -1.0) . When the Til ex molecule is reduced by 50 mM DTT and subsequently alkylated with iodoacetamide , the CD spectrum is substantially altered, pointing to a role for disulfide bridges in stabilizing secondary and tertiary structure (Figure 6B) . The fact that the spectrum of the non-reduced molecule reflects significant thermal denaturation at 80°C (Figure 6C) confirms that substantial secondary structure is present in soluble Tilex2-. As exp ^r ected, the p ^r attern after thermal denaturation is the same for the reduced as for the non-reduced molecule (compare Figures 6C and 6D) .

In its overall pattern, the CD spectrum resembles that for Thy-1 (Campell e t aL_, Nature 2££:341-342 (1979)) which is a well-recognized

member of the immunoglobulin superfamily (Williams and Barclay, Annu. Rev. ImmmunoL : 381-405 (1988)) and Is therefore predicted to consist entirely of 0-sheet. However, the shoulder at 225 n is absent from the Thy-1 profile. To obtain a more objective prediction of secondary structure from the CD spectrum, the digitalized absorption data (Table I) were deconvoluted according to the inverse matrix method of Compton and Johnson, Anal. Biochem. 155_:155-167 (1986). The resulting predictions for proportions of secondary elements are: α-helix, 20%; anti-parallel 0-sheet, 13%; parallel 0-sheet, 9%, turn, 20%, other, 46%. Since none of the protein In the data sets used to determine the matrix values are homologous to CD2, the predicted fractions of secondary structure are only approximate. When the same CD data are deconvoluted by the method of Yang et ____________ Meth. EnzymoL 130: 208-209 (1986a), 15% α and 40% β structures are predicted.

TABLE I

DIGITIZED PROTEIN CIRCULAR DICHROISM SPECTRUM

FOR Til -FROM 184-260 nm at 2 nm INTERVALS ex2

Spectrum was taken in 10 mM sodium phosphate (pH 7.2) in a 1 mm cell at 22°C. Secondary structure predictions were calculated by taking dot products with inverse CD spectra for the five secondary

structure categories as described In Table 6 of

Compton and Johnson, Aj2£L__Bi £hem_ ₁ . 1 5: 155-167

(1986). The calculated results are: α-helix, 20%; anti-parallel 0-sheet, 13%; parallel 0-sheet, 9%; turn, 20%; other, 46%. Units for Δe are M cm

To investigate the presence of protease- resistant domains in the Til „ molecule, limited ex2 papain digestions were performed. Fifteen min. digestion at an enzyme :protein ratio of 1:250 at 37°C yields a major band at 15KD when analyzed by SDS-PAGE (data not shown). Longer digestion times cause partial disappearance of the 15KD band. Approximately 400 pmol of this 15KD material was blotted onto PVDF and sequenced, yielding 19 un¬ ambiguous residues corresponding to the CD2 amino- terminus . These results demonstrate that the carboxy region of Til „ is much less resistant to papain digestion that its amino-terminal counter¬ part.

Til _ inhibits sheep erythrocyte rosetting

The ability of Til - to interact with the CD2 ligand expressed on the surface of various cell types was investigated. The capacity of Til - to inhibit sheep erythrocyte rosetting with T lympho¬ cytes was tested. Table II shows that rosetting is completely inhibited at concentrations of Til ex2 greater than 5 μM; half-maximal inhibition occurs between 0.63 and 1.25 μM Til „. Note that the exz antl-TIL (3T4-8B5) antibody abrogates rosetting at a concentration as low as 0.007 μM (Table II). This result suggests that any direct interaction between the soluble Tilex2 molecule and the CD2 lig ^& and is of relatively low affinity.

TABLE_II

C0NCENTRATI0N_DEPENDENCE_0F_T11 _INHIBITI0N_0F SHEEP ERYTHROCYTE-HUMAN T SELL ROSETTES

100

0

0

0

4.8

9.9

122

84

116

110

66 91 96

105 90 72

110 99

116

0 0 0

For rosette formation, sheep erythrocytes were preincubated with soluble protein or antibody at the indicated concentrations for 1 h at 4°C followed by the addition of Jurkat cells, co-centrifugation and a further 1 h incubation at 4°C. Rosettes were counted in a hemacytometer . Control value was the fraction of rosette formation in the absence of added protein. Anti-CD2 was the anti-TIL antibody of 3T4-8B5.

TH - blocks the binding of anti-LFA-3 monoclonal antibody

To investigate whether Til _ could interact

° ex2 with a previously defined ligand for CD2 on human cells (Selveraj et aL_, Nature 326:400-403 (1987), Its ability to block the binding of monoclonal anti-LFA-3 antibody TS9/2 (Sanchez-Madrid et al. , Proc. Natl. Acad. Sci. USA 79_:7489-7493 (1982) to LFA-3 bearing cells was tested. Specifically, anti-LFA-3 reactivity was measured by FACS analysis on the human B lymphoblastoid line JY, which expresses high levels of LFA-3. As shown in Table III, preincubation of JY cells with soluble CD4 does not affect this staining, whereas preincubation with 10 μM Til - causes a substantial decrease In observed fluorescence, reducing linear immuno- luoroescence from channel 120 to 30.6. This repre¬ sents a reduction of 85% of specific anti-LFA-3 re¬ activity (calculated after subtraction of the back¬ ground fluorescence of 13.2 linear units). Significant blocking is also seen at 1 μM but not at

0.1 μM. Addition of the control soluble CD4 T4 exl protein has no effect on anti-LFA-3 binding. Thus, by this measure, as well as by inhibition of sheep erythrocyte rosetting, the affinity of Til _ for its ligand is apparently in the micromolar range.

TABLE I I I

CONCENTRATION DEPENDENCE OF Til ^INHIBITION OF MONOCLONAL ANTI-LFA-3 ANTIBODY BINDING

Anti-LFA-3 Mean Fluorescense

Protein Added H__ti £^Z_Add d Intens_it:y_

Tilex2-:

300 μg/ml (10 μM) + 30.6

30 μg/ml (1 μM) + 66.1

3 μg/ml (0.1 μM) + 109.1

0.3 μg/ml (0.01 μM) + 119.8

T4 . : exl

300 μg/ml + 125.1 30 μg/ml + 120.2 3 μg/ml + 117.6 0.3 μg/ml + 116.9

None 123.7 13.2

Mean fluorescence intensities were obtained on a linear scale from 0-250.

11 - b inds to LFA- 3 qn_human_0 lymphoblas toid

££l.ls_

Since Til „ clearly inhibits sheep erythrocyte rosetting and blocks the binding of anti-LFA-3 antibody, we determined whether specific, saturable binding of Til _ to human cells bearing LFA-3 could be detected. Two types of binding assays were employed. In the first, increasing amounts of unlabeled Til _ was added to a mixture of the JY B ex2 lymphoblastoid cell line plus a constant amount of

¹²⁵ I labeled Til ₂ (ligand).

Figure 7 shows the competitive inhibition of radioiodinated Til _ binding to JY cells. 5 x 10 cp radiolabelled Til „ (2.8 x 10 cpm/nmole) was added at 0.1 μM to 1.8 x 10 JY cells overlayed onto

0.2 ml of a 1.5:1 mixture of dibutyl phthalate: dioctyl phthalate: dioctyl phthalate in 0.5 ml plastic tubes. Increasing concentrations of unlabeled Til ₀ (closed circles) or T4 .. as a ex exJL control (open circles) were added in a final volume of 200 μl in RPMI 1640/10% FCS. After 1 h incubation at 4°C, tubes were spun, cell pellets severed and bound radioactivity was determined. To one tube anti-LFA-3 antibody TS2/9 was added at 50 μg/ml as a determination of non-specific binding. Specific cpm are the total cpm minus cpm bound in the presence of anti-LFA-3 (9070 cpm). Binding in the presence on control anti-HLA monoclonal antibody W6/32 is also shown (triangle) .

As shown in Figure 7, binding of radiolabeled Til „ is progessively inhibited by the addition of

increasing amounts of unlabeled molecules.

Half-maximal inhibition occurs at about 0.5 μM

Til „ . In contrast and as expected, addition of ex2 soluble CD4 T4 . protein has no effect on ex1

¹²⁵ I-labeled Til - binding to JY. ex2 °

As a second measure of binding, increasing amounts of labeled Til _ were added to a constant ex/ number of JY cells. Figure 8 shows the saturation binding of Til - to JY cells and Scatchard ^& ex2 analysis: Increasing concentrations of radiolabeled

Til „ (1.31 x 10 cpm/nmole) were added to 2.6 x ,ex2

10 JY cells in the presence or absence of 50 μg/ml anti-LFA-3 antibody to determine nonspecific binding. Binding was carried out as above and the dissociation constant determined by Scatchard analysis after subtraction of nonspecific binding determined in the presence of anti-LFA-3. Top: total binding (open circles) and specific binding

(closed circles) . Botton: Scatchard analysis: correlation coefficient - 0.98; -1/slope - K, - 0.4 μM.

As shown in Figure 8 (top), specific binding is saturable. These saturation binding data were transformed by Scatchard analysis (Figure 8, bottom) . Specific Til . binding is saturated at about 300 x 10 molecules per cell. The Scarchard plot yields an estimated dissociation constant of

0.4 μM for the TH _eχ2 -LFA-3 interaction. Note that although a single incubation period (1 h at 4°C) of cells with ligand was used in these experiments, kinetic analysis demonstrates that maximal

Til _Λ -LFA-3 binding occurs within 5 min. exz

Cqnclusion

We have shown that a soluble, monomeric extra¬ cellular segment CD2 molecule, termed Til -, carries epitopes recognized by antibodies against the native surface CD2 structure on resting T lymphocytes and interacts specifically with the sur ace-bound CD2 ligand, LFA-3. The measured dissociation constant for this interaction is 0.4 μM, implying a low affinity relative to hormone receptor-ligand interactions (i.e. , for IL-2 K, *=

-11 10 M) (Smith, 1984) but equivalent to that of

-5 -6 primary antibody responses (K, = 10 to 10 ^" )

(Elsen and Siskind, 1964). The Til . molecule ex2 gives rise to a proteolytically-resistant 15KD amino-terminal papain fragment, suggesting that the amino-terminal -100 amino acid residues comprise a stable, well-folded domain corresponding to a polypeptide encoded by the first extracellular exon, In contrast, despite disulfide linkages, the carboxy- erminal encoded extracellular domain is labile to papain.

Deposit

The virus, AcNP/TH _ has been deposited at the American Type Culture Collection in Rockville, Maryland on January 19, 1989, and assigned the ATCC Accession Number

E££iχ l nts_

Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, many equivalents to the specific embodiment of the invention described herein. Such equivalents are intended to be encompassed by the following claims .