WO1993022450A1 - Novel diphtheria toxin-based molecules - Google Patents

Novel diphtheria toxin-based molecules Download PDF

Info

Publication number
WO1993022450A1
WO1993022450A1 PCT/US1993/004335 US9304335W WO9322450A1 WO 1993022450 A1 WO1993022450 A1 WO 1993022450A1 US 9304335 W US9304335 W US 9304335W WO 9322450 A1 WO9322450 A1 WO 9322450A1
Authority
WO
WIPO (PCT)
Prior art keywords
molecule
diphtheria toxin
domain
binding
cdr
Prior art date
Application number
PCT/US1993/004335
Other languages
French (fr)
Inventor
Seunghyon Choe
David Eisenberg
Francis S. Genbauffe, Jr.
Original Assignee
The Regents Of The University Of California
Seragen, Inc.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by The Regents Of The University Of California, Seragen, Inc. filed Critical The Regents Of The University Of California
Publication of WO1993022450A1 publication Critical patent/WO1993022450A1/en

Links

Classifications

    • CCHEMISTRY; METALLURGY
    • C07ORGANIC CHEMISTRY
    • C07KPEPTIDES
    • C07K14/00Peptides having more than 20 amino acids; Gastrins; Somatostatins; Melanotropins; Derivatives thereof
    • C07K14/195Peptides having more than 20 amino acids; Gastrins; Somatostatins; Melanotropins; Derivatives thereof from bacteria
    • C07K14/34Peptides having more than 20 amino acids; Gastrins; Somatostatins; Melanotropins; Derivatives thereof from bacteria from Corynebacterium (G)
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61KPREPARATIONS FOR MEDICAL, DENTAL OR TOILETRY PURPOSES
    • A61K38/00Medicinal preparations containing peptides
    • CCHEMISTRY; METALLURGY
    • C07ORGANIC CHEMISTRY
    • C07KPEPTIDES
    • C07K2319/00Fusion polypeptide
    • C07K2319/33Fusion polypeptide fusions for targeting to specific cell types, e.g. tissue specific targeting, targeting of a bacterial subspecies
    • CCHEMISTRY; METALLURGY
    • C07ORGANIC CHEMISTRY
    • C07KPEPTIDES
    • C07K2319/00Fusion polypeptide
    • C07K2319/55Fusion polypeptide containing a fusion with a toxin, e.g. diphteria toxin

Definitions

  • the field of the invention is chimeric molecules.
  • Hybrid molecules in which all or part of an antibody is fused to another molecule have been suggested as a means for targeting molecules to particular sites.
  • Diphtheria toxin (DT) is an extremely potent cytotoxin which is secreted by Corynebacterium diphtheriae that has been lysogenized by a bacteriophage carrying the
  • Diphtheria toxin gene Naturally occurring Diphtheria toxin is a single polypeptide chain of 535 residues . Mild trypsinization and reduction of Diphtheria toxin in vitro generates two fragments, Fragment A (amino-terminal, ⁇ 21K) and Fragment B (carboxy-terminal, ⁇ 37K), as a result of cleavage at residue 190, 192, or 193. A similar proteolytic cleavage ('nicking') occurs in vivo before or soon after the toxin binds to a sensitive cell. Fragment B of the toxin binds the protein to receptors on the cell surface and promotes transfer of the Fragment A to the cytoplasm. Fragment A in the cytoplasm catalyzes the transfer of the ADP-ribosyl group of NAD + to
  • EF-2 elongation factor 2
  • EF-2 elongation factor 2
  • EF-2 elongation factor 2
  • Introduction of a single molecule of Fragment A into the cytoplasm can kill a cell. While the exact mechanism by which Diphtheria toxin enters a cell is not completely understood, it is known that Diphtheria toxin binds a receptor and is endocytosed and delivered to endosomes where it encounters acidic conditions. At a threshold pH of ⁇ 5.0 the toxin undergoes a conformational change, which promotes insertion and formation of an ion-selective channel in the membrane, and Fragment A is translocated and released into the cytoplasm. Summary of the Invention
  • the invention features a chimeric diphtheria toxin molecule wherein all or part of a complementarity determining region (CDR) of an antibody is inserted into a loop region of the Diphtheria toxin receptor binding-domain.
  • loop region is meant any of the portions of the Diphtheria toxin receptor binding domain lying between ⁇ strands as delimited herein.
  • the term encompasses single amino acids.
  • Diphtheria toxin receptor binding domain is meant the portion of Diphtheria toxin lying from amino acids 386 to amino acid 535, inclusive.
  • CDR is meant a portion of a complementarity determining region
  • complementarity determining region of an antibody as defined by sequence heterogeneity, e.g., according to Kabat et al (in Sequences of Proteins of Immunological Interest, U.S. Dept. of Health and Human Services, U.S. Government Printing Office, 1987).
  • sequence heterogeneity e.g., according to Kabat et al (in Sequences of Proteins of Immunological Interest, U.S. Dept. of Health and Human Services, U.S. Government Printing Office, 1987).
  • the antibody is capable of specifically binding a cell surface antigen expressed on a cell
  • the chimeric diphtheria toxin molecule is capable of
  • cell surface antigen is meant any cell surface marker, e.g., a protein or a carbohydrate.
  • specifically binding is meant does not substantially bind to other molecules.
  • diphtheria toxin receptor is meant the receptor for naturally-occurring Diphtheria toxin.
  • the molecule is capable of decreasing the viability of the cell. In an even more preferred embodiment the molecule kills the cell.
  • all or part of a first CDR is inserted into a first loop region
  • all or part of a second CDR is inserted into a second loop region
  • the first and the second CDR are of a single antibody chain.
  • of a single antibody chain is meant CDR sequences found within a single heavy or light chain.
  • the molecule lacks diphtheria toxin catalytic activity.
  • diphtheria toxin catalytic activity is meant the ability to inhibit translation.
  • the molecule lacks all or part of the catalytic domain of diphtheria toxin.
  • the loop region is RL3 and the CDR is a CDR1; and the loop region is RL9 and the CDR is a CDR3.
  • the invention features a hybrid molecule which includes a first and a second portion joined together covalently, the first portion includes a chimeric diphtheria toxin molecule wherein all or part of a CDR of an antibody is inserted into a loop region of the receptor binding-domain of diphtheria toxin, the antibody being capable of specifically binding a cell surface antigen expressed on a cell, the chimeric diphtheria toxin molecule being capable of specifically binding the cell surface antigen, being substantially incapable of binding to the diphtheria toxin receptor, and lacking Diphtheria toxin catalytic activity; and the second portion includes a molecule to be delivered to the cell.
  • the molecule to be delivered to the cell is a protein; is an enzyme; is a protein which modulates transcription; is a nucleic acid binding protein; is a nucleic acid-binding protein capable of binding a single-stranded nucleic acid; and is a nucleic acid.
  • the invention features a hybrid molecule which includes a first and a second portion joined together covalently, the first portion includes a chimeric diphtheria toxin molecule wherein all or part of a CDR of an antibody is inserted into a loop region of a first diphtheria toxin receptor binding-domain, the antibody being capable of specifically binding a cell surface antigen expressed on a cell, the chimeric diphtheria toxin molecule being capable of specifically binding the same cell surface antigen and being substantially incapable of binding to the
  • diphtheria toxin receptor wherein the amino-terminus of a second diphtheria toxin receptor-binding domain is connected to the carboxy-terminus of the first diphtheria toxin receptor-binding domain, the second diphtheria toxin receptor-binding domain being substantially
  • the carboxy terminus is
  • all or part of a first the CDR is inserted into a first loop region of the first diphtheria toxin receptor-binding domain and all or part of a second the CDR is inserted into a second the loop region of the first diphtheria toxin receptor-binding domain.
  • all or part of a third CDR is inserted into a first loop region of the second diphtheria toxin
  • the first and the second CDR are of a first antibody chain and the third and the fourth the CDR are of a second antibody chain.
  • the first and the second antibody chains are from antibodies recognizing the same antigen; the first and second antibody chains are from the same antibody molecule; and the first antibody chain is the light chain of an antibody and the second antibody chain is the heavy chain of the same antibody.
  • the invention features a chimeric diphtheria toxin molecule wherein all or part of a CDR-like sequence of a Iigand binding protein having an antibody variable domain-like Iigand binding-domain is inserted into a loop region of the receptor binding domain of diphtheria toxin.
  • CDR-like sequence is meant a sequence which is responsible for Iigand binding and which has the same relationship to overall structure of a protein as does the CDR of an antibody variable domain.
  • an antibody variable domain-like Iigand binding domain is meant a Iigand binding domain which has structural homology to an immunoglobulin variable domain.
  • tumor necrosis factor includes an antibody variable domain-like Iigand binding domain.
  • the chimeric molecules of the invention bind specifically to the same epitope (antigen) as the
  • molecules can enter cells to whihc they bind, they can be used to introduce any molecule into a specific class of cells.
  • Figure 1 is a schematic drawing of Diphtheria toxin in which each secondary structural segment is identified.
  • the first letter denotes the domain: C for catalytic, T for transmembrane, and R for receptor-binding domains.
  • the second letter denotes the secondary structure class: H for helix, B for ⁇ strand, L for loop.
  • the third symbol is the sequential number of each
  • Figure 2 is a representation of the C ⁇ skeleton of Diphtheria toxin from the same viewpoint as that of Fig. 1.
  • An ApUp molecule occupies the active site of
  • Figure 3 is a stereo pair representation of the electron density maps calculated at 2.5A from (2F ob -F c ) and the refined model phases. Maps are superimposed on the corresponding region of the refined model.
  • Figure 4 is a representation of the Diphtheria toxin dimer observed within the Form4 crystal. The two monomers are related by a crystallographic 2-fold
  • Figure 5 is a stereo pair representation of the C ⁇ skeleton of the C domain.
  • the entrance to the active site is at the lower right.
  • the four loops, CL1 to CL4 are highlighted. Notice that they form a hinge which may permit the C domain to form a more elongated structure.
  • FIG. 6 is a stereo pair representation of the C ⁇ skeleton of the T domain, with the direction of view from the right side of Diphtheria toxin in Fig. 1.
  • Helix TH1 lies in back, starting at residue 205.
  • Helix TH2 runs to the left at the bottom, followed by a turn and helix TH3 running to the right.
  • In front center is TH5 (running to the left) and above it are helices TH6 and TH7.
  • Behind these pairs of antiparallel helices is another pair of antiparallel helices, TH8 and TH9, with TH9 running upwards and ending at residue 378.
  • the Asp and Glu side chains are shown. Notice the tips of two helix layers, TL3 and TL5 contain a total of six acidic groups (on the left).
  • Figure 7 is a stereo pair representation of the T domain as in Fig. 6 except that the Lys, Arg and His side chains are shown. Notice the positive charge asymmetry, with all charges at the bottom and back of the domain, with an exception Lys 299 near the loop TL3 between TH5 and TH6.
  • FIG. 8 is a schematic representation of the R domain of Diphtheria toxin (panel A), an Ig variable domain (panel B) and tumor necrosis factor (panel C).
  • R domain is viewed in the direction from the back side of Diphtheria toxin in Fig. 1. Numbers from 2 to 10 of the R domain represent the strands RB2 through RB10 of
  • strands 3, 4, 5, 6, 7, 8, and 9 correspond well to strands C, D, E, F,
  • Figure 9 is a schematic drawing of a rearranged Diphtheria toxin receptor-binding domain.
  • panel A is a schematic representation of the receptor-binding domain of diphtheria toxin with each ⁇ strand labeled (RB1-RB10) and each loop region labeled (RL1-RL9).
  • the amino acid end points refer to the ⁇ sheets, e.g., RB3 consists of residues 412-424 inclusive and RL3 consists of residues 425-427 inclusive.
  • Panel B is a schematic representation of a Diphtheria toxin molecule which has undergone segment rearrangement. The notation for the regions is a in panel A except that 447, 483, 467, and 455 indicate the amino acid residues of immediately adjacent to residues 407, 455, 445, and 483 respectively.
  • diphtheria toxin as determined by x-ray crystallography.
  • the receptor-binding domain of diphtheria toxin has a structure similar to the variable domain of an antibody. Because of this
  • Diphtheria toxin can be modified so that certain portions of its receptor-binding domain are replaced by, or modified to include, antigen-binding portions (complementarity determining regions) of an antibody of choice. Such modification results in the creation of a chimeric diphtheria molecule which
  • a chimeric molecule recognizes and binds the same antigen as the selected antibody. If a chimeric molecule is modified so as to substantially eliminate binding to the diphtheria toxin receptor, it will selectively bind only to cells bearing the antigen recognized by the antibody from which the complementarity determining regions were derived.
  • Chimeric molecules of the type described above can be targeted to selected cell types. For example,
  • portions of an antibody directed against the interleukin-2 (IL-2) receptor can be used to make a chimeric
  • diphtheria toxin molecule which binds to cells bearing the IL-2 receptor. If the chimeric molecule is designed so as to retain the translocation and catalytic functions normally associated with diphtheria toxin, the chimeric molecule will enter and kill cells bearing the IL-2 receptor. If in the course of creating this chimeric diphtheria toxin molecules the receptor binding domain is altered so that the chimeric molecule does not bind to the diphtheria toxin receptor, this chimeric molecule will bind and kill cells bearing the IL-2 receptor while leaving all other cells unharmed.
  • chimeric diphtheria toxin molecules can be used to introduce any molecule into a selected group of cells. For example, if a chimeric diphtheria toxin molecule capable of binding to cells bearing the IL-2 receptor is modified so that the catalytic domain of diphtheria toxin is replaced by an enzyme, that enzyme can be selectively introduced into cells bearing the IL-2 receptor. Similar modifications would permit an
  • the catalytic domain may be substantially inactivated by mutation rather than
  • a chimeric diphtheria toxin molecule is created by: (1) generating (or selecting) an antibody which recognizes the antigen; (2) cloning and sequencing at least the variable domain of a heavy or light chain of the antibody; (3) identifying the complementarity
  • the molecules of the invention can be more completely understood by first detailing the structure of Diphtheria toxin itself. Accordingly, the overall structure of Diphtheria toxin is discussed below followed by a detailed discussion of the structure of its
  • diphtheria toxin receptor-binding domain The relationship between the diphtheria toxin receptor-binding domain and an antibody variable domain is then described. This is followed by a discussion of methods for generating and screening chimeric molecules. This discussion is followed by a description of the structure of other parts of Diphtheria toxin including the catalytic domain, the translocation domain and the domain junctions. Lastly, details of the structure determination are presented.
  • Diphtheria toxin consists of three abutting domains that are connected by interdomain linkers.
  • the amino-terminal domain (residues 1-193) is the catalytic (C) domain.
  • the middle domain (residues 205-378) is the transmembrane (T) domain, and the carboxy-terminal domain (residues 386-535) is the receptor binding (R) domain.
  • Diphtheria toxin is Y-shaped with the base formed by the T domain, one arm of the Y formed by the C domain, and the other arm formed by the R domain. The Y is about 9 ⁇ A high, 5 ⁇ A across the top of the Y, but only 30 ⁇ thick ( Figure l).
  • the C domain is a mixed structure of eight ⁇ strands (CB1-CB8) and seven ⁇ -helices (CH1-CH7).
  • the eight ⁇ strands form two ⁇ sheets of 3 and 5 strands each. These ⁇ sheets form a core that is surrounded by 7 short helices.
  • the overall folding of the C domain is similar to that, of Pseudomonas aeruginosa exotoxin A (ETA) especially near the active site (Allured et al., Proc. Natl . Acad.
  • ETA Pseudomonas aeruginosa exotoxin A
  • the R domain contains ten ⁇ strands (RB1-RB10), nine of which (RB2-RB10) build two ⁇ sheets. These two ⁇ sheets form a ⁇ sandwich with a topology similar to a jellyroll fold (Richardson, Adv. Protein Chem . 34:167, 1981).
  • the three-domain organization of Diphtheria toxin is shared by two other bacterial toxins, ETA and 5-endotoxin from Bacillus thuringiensis (Carroll, et al., Nature 353:815, 1991).
  • the catalytic domains of Diphtheria toxin and ETA are the closest among all these domains in their
  • the receptor-binding (R) domain is formed from two ⁇ sheets, ⁇ strands RB2 (residues 393-399), RB3 (residues 412-424), RB5 (residues 445-453), and RB8 (residues 483-495) form a four-stranded ⁇ sheet that faces a five-stranded ⁇ sheet containing ⁇ strands RB4 (residues 428-438), RB6 (residues 455-465), RB7 (residues 467-480), RB9 (residues 513-520), and RB10 (residues 525-534).
  • RB6 interacts with both ⁇ sheets through hydrogen bonds.
  • the connection of the strands is such that the R domain is similar to the jellyroll topology found in many proteins that are exclusively formed from antiparallel ⁇ strands (Richardson, J. Adv. Protein Chem . 34:167, 1981).
  • Jellyroll domains include viral coat proteins, tumor necrosis factor, and the receptor-binding domain of ETA.
  • the R domain differs somewhat from a strict jellyroll topology (Fig. 9) in having strand 2 in the "front” sheet, and having a strand 10 in the "back".
  • the R domain is also similar in structure to an immunoglobulin (Ig) variable domain (Fig. 9, panel B), but differs from the Ig fold in having an "insert" of strands 5 and 6 between 4 and 7, and also in lacking two short strands (C' and C" in Fig. 9, panel B) between 4 and 5.
  • the portion of the R domain that resembles a strict jellyroll in topology is the right side as viewed in Fig. 9; and the portion that resembles the Ig variable domain is the left side, the side that is away from the rest of the Diphtheria toxin monomer.
  • chimeric diphtheria toxin molecules all or part of one or more complementarity determining regions derived from an antibody are inserted into one or more loop regions of the Diphtheria toxin receptor-binding domain.
  • CDR sequence is inserted into each loop and the insertion may or may not be accompanied by deletion of all or a portion of the loop region.
  • An antibody consists of two identical light chains (L) and two identical heavy chains (H). Each light chain is attached to a heavy chain by one or more disulfide bonds.
  • a single antibody forms a "Y" shaped structure in which the carboxy-terminal portion of the heavy chains forms the base of the Y and the amino-terminal portion of a single heavy chain and the amino-terminal portion of a single light chain together form each arm.
  • Each chain, heavy or light is composed of structurally similar domains. The domains are referred to as constant or variable based on sequence heterogeneity. Proceeding from the carboxy-terminus, a heavy chain is composed of the CH 3 constant domain, the CH 2 constant domain, the CH 1 constant domain, and the V H variable domain. Proceeding from the carboxy-terminus, a light chain is composed of a C L constant domain followed by a V L variable domain.
  • variable domains are of particular interest since together they form the antigen binding site.
  • Each variable domain is approximately 110 amino acids long and is composed of three hypervariable or complementarity-determining regions (CDR1, CDR2, and CDR3) interspersed with four less-variable framework regions (FR1, FR2, FR3, and FR4).
  • CDR1 hypervariable or complementarity-determining regions
  • FR1 FR2, FR3, and FR4
  • CDR's are responsible for antigen recognition.
  • each variable domain consists of two ⁇ sheets which together form a structural motif often referred to as the immunoglobulin fold. (Constant regions, with a slight variation, also form an
  • One ⁇ sheet is composed of four ⁇ strands (A, B, D and E) the other sheet is composed of five ⁇ strands (C, C, C", F and G). There are loops between each ⁇ strand. Three of these loops, to a first approximation, correspond roughly to the three CDR's.
  • the identification of a CDR or a framework region is based primarily on sequence heterogeneity rather than secondary structure.
  • the identification of these regions within a given antibody molecule requires analysis of the amino acid sequence of the antibody.
  • the loop between B and C often includes all or part of CDR1; the loop between C and C" often includes all or part of CDR2; and the loop between F and G often includes all or part of CDR3.
  • a comparison between the structure of the receptor binding domain of Diphtheria toxin and immunoglobulin V domain illustrates the structural similarities (Fig. 9). There is a correspondence, not identity, between the ⁇ strands of the Diphtheria toxin receptor domain
  • Two of the Diphtheria toxin receptor binding domain loops, RL3 and RL9 thus correspond to CDR1, and CDR3, respectively.
  • the receptor domain can be engineered to more closely
  • CDR sequences can be inserted into RL3 and RL9 in a process referred to herein as 'loop grafting' to yield a chimeric molecule which includes a CDR1 sequence or a CDR3 sequence or both.
  • a process referred to herein as 'loop grafting' to yield a chimeric molecule which includes a CDR1 sequence or a CDR3 sequence or both.
  • all or part of either or both loop regions may be deleted.
  • the region of Diphtheria toxin from RB5 to RB8 can be rearranged so that it more closely resembles the region of a variable domain extending from strand C through strand E.
  • This process referred to herein as 'segment rearrangement', can provide a framework for the grafting of a CDR2 sequence into Diphtheria toxin.
  • these two approaches can be used to create a chimeric molecule into which three CDR's (CDR1, CDR3, CDR3)
  • Loop grafting is similar to CDR grafting in which the CDR of a first antibody are exchanged for those of a second antibody, and the techniques employed in CDR grafting will, in general, be useful for loop grafting.
  • Jones et al. (Nature 321:522, 1986), Riechmann et al. (Nature 322:323, 1988), Winter et al. (PCT/GB89/00113), Winter (EPA 0 239 400), and Clackson et al. (Nucl . Acids Res . 17:10163, 1989) describe CDR grafting techniques which can be applied to loop grafting.
  • RL3 and RL9 while precise in terms of structure, represent only approximate limits to the region which might be replaced by all or part of a CDR. Further insertion of a CDR can take place without the deletion of any loop region sequence. Thus, a CDR may replace a few residues of RB3 and a few residues of RB4 as well as all of RL3 (e.g., residues 422-429).
  • a CDR might replace only a part of RL3 (e.g., residues 426 and 427). Alternatively, no residues are deleted.
  • CDR's are identified by sequence hypervariability (Kabat et al., in Sequences of Proteins of Immunological Interest , U.S. Dept. of Health and Human Services, U.S. Government Printing
  • a CDR so-defined may include only a portion of the loop between two ⁇ -strands and likewise may include part or all of one or both j ⁇ -strands flanking the loop. Nevertheless, it should be understood that in many antibodies CDR's are found at similar positions.
  • CDRl is commonly located near amino acid 30
  • CDR2 is commonly located near amino acid 50
  • CDR3 is commonly located near amino acid 95 (Roitt et al.,
  • Useful CDR's may be derived from either
  • immunoglobulin H or L chains Further, antibodies derived from any species may be used as a source of
  • CDR's Because CDR's appear not to contain species specific motifs, CDR's from a first species can be used without substantially increasing the immunogenicity of the chimeric molecule in a second species.
  • Segment rearrangement essentially consists of reorganizing the portion of Diphtheria toxin from the beginning of RB5 to the end of RL7. This reorganization results in the formation of a rearranged R domain which more closely resembles an antibody variable domain than does the native R domain. This rearrangement in
  • combination with loop grafting of all or part of RL3 and/or RL9 can provide a molecule with improved antigen binding characteristics compared to a molecule which has undergone grafting of the same loops but has not
  • Domain rearrangement can also provide a location, between rearranged RB7 and rearranged RB6, for the addition of a CDR2.
  • a molecule in which the R domain has the following sequence of elements (beginning at its amino-terminus): RB1-RL1-RB2-RL2-RB3-RL3-RB4-RL4-RB7-RL7-RB6-RL6-RB5-RL5-RB8-RL8-RB9-RL9-RB10 (Fig. 10, panel B). All or part of a CDR2 sequence can be introduced by replacing all, part, or none of RL7 (in the rearranged molecule located between RB7 and RB6). Referring to Fig. 9, panel A, this rearranged receptor-binding domain more closely resembles an immunoglobulin variable domain (Fig. 9, panel B) than does the naturally occurring Diphtheria toxin receptor-binding domain (Fig. 8, panel A).
  • diphtheria toxin receptor binding ability of Diphtheria toxin receptor domain must be substantially reduced in chimeric molecules compared to native
  • Diphtheria toxin so that the chimeric molecules do not substantially bind to or enter non-targeted cells (i.e., cells not bearing the antigen recognized by the CDR's).
  • This can be accomplished by incorporating into the chimeric molecules certain mutations which reduce binding of Diphtheria toxin to its natural receptor.
  • CRM9 Human et al., Biochim . Biophys . Acta 902:24, 1987
  • CRM107 and CRM103 Greenfield et al., Science 238:536, 1987
  • the insertion of a CDR sequence into a loop of the R domain in the absence of any deletion may, in and of itself, substantially eliminate binding to the diphtheria toxin receptor.
  • diphtheria toxin receptor-binding ability of chimeric diphtheria toxin molecules can be assessed by standard techniques (Middlebrook et al., Can . J. Microbiol . 23:183, 1978; Middlebrook et al., J. Biol . Chem . 253:7325, 1978) using Vero cells or other cell lines.
  • the carboxy-terminus of RB10 of a diphtheria toxin molecule is fused to the amino-terminus of RB1 or RB2 of an R domain portion of diphtheria toxin.
  • CDR sequences can be introduced into these molecules in manner described above by modifying the loop regions of one or both of the R domains.
  • the result is a molecule which has a domain (R+R') resembling an sFv molecule fused to the diphtheria toxin translocation and catalytic domains.
  • it may be possible to generate antigen binding molecules by introducing CDR sequences into only one of the two R domains.
  • both receptor binding domains must be modified to essentially eliminate recognition of the diphtheria toxin receptor. If both R domains have been engineered to introduce CDR sequences, it is preferred all of the CDR sequences be derived from the same antibody (or at least antibodies recognizing the same epitope), and that the CDR sequences of one R domain be derived from a V H domain, and the that the CDR sequences of the other R domain be derived from a V L domain. This creates a molecule which more closely resembles an Fv fragment.
  • chimeric diphtheria toxin molecules having two R domains can be linked by disulfide bonds as described by Glockshuber et al. (Biochemistry 29:1362, 1990) for single-chain Fv molecules.
  • chimeric diphtheria toxin molecules having two R domains and CDR sequences can be modified to act as delivery vehicles rather than cytotoxins. This is accomplished by replacing the catalytic domain with a molecule to be introduced into cells thus creating a hybrid molecule. The R domains will then target the hybrid molecule to a selected class of cells and the translocation domain will mediate entry.
  • variable domains can be amplified for cloning using the polymerase chain reaction and oligonucleotide primers which recognize conserved sequences at each end of the heavy or light chain
  • variable region (Orlandi et al., Proc. Nat'l . Acad. Sci . USA 86:3833, 1989; Larrick et al., Biochem. Biophys. Res . Comm . 160:1250, 1989; Sastry et al., Proc. Nat 'l Acad. Sci . USA 86:5728, 1989).
  • This approach allows the cloning of the variable regions of human antibody genes from unstable human-mouse hybridomas as well as the cloning of variable regions from other unstable
  • hybridomas single hybridoma cells, and single B
  • lymphocytes These techniques permit the expression of antibody fragments in bacteria (Skerra et al., Science 240:1038, 1988; Better et al., Science 240:1041, 1988) or on the surface of phage (McCafferty et al., Nature
  • variable regions are sequenced CDR's are identified according to Kabat et al. (supra) .
  • the chimeric molecules are generated using the standard techniques of molecular biology (Current
  • the chimeric molecules themselves can be produces in bacterial cells, mammalian cells, or insect cells by standard techniques. In designing chimeric molecules the techniques of computer-based molecular modeling may be useful. The coordinates of the solved diphtheria toxin structure are included (appendix) to aid in this process. It should be understood that changes in the amino acid sequence may be introduced at any position to generate more stable molecules or molecules with higher binding specificity.
  • the chimeric molecules Once the chimeric molecules have been produced, they can be screened for antigen binding ability using any of the approaches described above for antibodies and antibody fragments.
  • the binding specificity of a chimeric molecules can be screened for antigen binding ability using any of the approaches described above for antibodies and antibody fragments.
  • cytotoxic chimeric molecules targeted to a particular antigen can be determined by comparing the toxicity of the molecule toward cells bearing the antigen to its toxicity towards cells not bearing the antigen.
  • a detectable label may be covalently linked to the chimeric molecule to facilitate comparison of binding to antigen-bearing cells and cells not bearing antigen.
  • a chimeric diphtheria toxin molecule capable of recognizing cells bearing the Campath-1 antigen can be constructed by replacing all of RL3 with the CDR1
  • a chimeric diphtheria toxin molecule capable of recognizing cells bearing the interleukin-1 receptor can be constructed by replacing all of RL3 with residues 26-33 of the anti-tac antibody light chain and replacing all of RL9 with residues 99-107 of the anti-tac antibody light chain, (supra) .
  • the dimer may represent a
  • Form3 there is one Diphtheria toxin chain per asymmetric unit and pairs of Diphtheria toxin chains are related by a 2-fold rotation axis.
  • the initial model was based on the structure determination of Form4 crystals at 3 . 0 ⁇ resolution, using the multiple isomorphous replacement (MIR) method
  • Some of the useful markers in the density maps were W 50 , W 153 , W 281 , W 398 , a 5-residue segment of M 178 , Y 179 , E 180 , Y 181 , M 182 , a 4-residue segment of F 355 , Y 358 , H 372 , Y 375 , a cluster of Y 514 , F 530 , F 531 , witn big side chains near the carboxy terminus (Fig. 3), and two disulfide bonds between C 186 and C 201 and C 461 and C 471 .
  • An initial improper fitting in the R domain was detected by profile window plots (L ⁇ thy et al., Nature 355:xxx, 1992) and then corrected.
  • the final model consists of 4137 non-hydrogen atoms with individual isotropic temperature factors.
  • the model also includes ApUp in the active site cleft of the catalytic (C) domain, but no solvent atoms.
  • Residues 190-195 are part of the protease-sensitive region of the first disulfide loop, where nicking occurs; this region may be intrinsically flexible. So may be the loop between the transmembrane (T) and R domains, which includes residues 389-390.
  • Crystal Forms, 1, 3 and 4 were used for the current study (Fuji, supra) .
  • Diffraction data were collected on a Rigaku AFC-6 diffractometer operating at 8.5 kW, equipped with a two-panel area detector of Xuong-Hamlin design (San Diego Multiwire Systems). Images were recorded as 0.1° oscillation frames, integrated and merged into batches of 50 frames (5°). Integrated intensities were scaled and merged for FOURIER scaling method (Weissman, Thesis , Univ. California, LA, 1979). Form4 native and derivative data were later collected to 2.5 °A with a RAXIS imaging plate system.
  • KOS K 2 OsO 4 , soaked for 3 days at the
  • Os derivatives of Form4 and Form3 have the same single site binding.
  • the C domain is formed from two ⁇ sheet subdomains, which subtend the active site cleft (Fig. 5). These ⁇ sheets are oriented roughly perpendicular to each other and form the core of the domain.
  • One subdomain consists of ⁇ strands CB2, CB4, and CB8, surrounded by ⁇ -helices, CH2, CH3, CH6, and CH7.
  • the other subdomain consists of ⁇ strands CB1, CB3, CB5, CB6, and CB7 surrounded by helices, CH1 CH4, and CH5.
  • the two subdomains are connected by extended loops, CL1 through CL4, which link the two subdomains. These four loops appear to suggest potential for flexibility or even extension to a longer and narrower shape. Conceivably the C domain can assume this partially unfolded structure during membrane translocation.
  • the active site cleft of the C domain identified by the binding of the dinucleotide ApUp, is formed primarily by ⁇ strands, CB2, CB3, CH3, CB7 and the loop, CL2 and is also bounded by ⁇ strand RB6 of the R domain.
  • Glu 148 which is believed to play a key role in catalysis (Carroll et al., Proc . Natl . Acad . Sci USA 81:3307, 1984), His 21 (Papini et al., J. Biol . Chem . 264:12385, 1989) and Tyr 65 (Papini et al., J. Biol . Chem . 266:2494, 1991), both of which have been implicated in NAD + binding, and various other residues suggested to be at or near the active site (Gly 52 (Carroll, supra and
  • the approximate position of the substance NAD + in the active site can be inferred, because the dinucleotide, ApUp, binds competitively with NAD + .
  • the high affinity of ApUp ( ⁇ 0.3nM as compared with ⁇ 8-16 ⁇ N for NAD + ; (Carroll et al., Biochem . 25:2425, 1986) may be a consequence of multiple contacts with the C domain and of salt bridges between the 3'-terminal phosphate of ApUp and the side chains of
  • the second disulfide bond makes a 9 residue loop between residues 461 and 471 within Fragment B. Residues near this loop (456, 458, 460, 472, 474) are also rich in positive charges and face the active site cleft, probably forming the so-called phosphate-binding or P-site (Lory et al., Proc. Natl . Acad. Sci . USA 77:267, 1980).
  • loop CL2 may allow a substantial movement of main chain atoms of the loop, permitting substrate entry to the active site.
  • the structure of the T domain exhibits two features that suggest how it might experience pH-triggered insertion into the membrane.
  • the first is that the T domain is entirely ⁇ -helical, similar to the known and proposed transmembrane proteins, and that some of the helices have hydrophobic
  • the nine helices are arranged more or less in three layers, each layer consisting of an antiparallel pair of helices.
  • the two long, carboxy terminal helices, TH8 and TH9 are unusually apolar and constitute the central core layer.
  • the other layer made up of helices TH1-TH3 is, in contrast, very hydrophilic even compared to globular proteins.
  • the second noteworthy feature of the T domain is the acidic composition of the loops that connect pairs of these helices.
  • loop TL3 between helices TH5 and TH6, and loop TL5 between hydrophobic helices TH8 and TH9 contain a total of six Asp and Glu residues (Fig. 6). At neutral pH, these loops are highly charged and water soluble. But at acidic pH, these residues would be at least partially
  • T domain has the capacity to insert into the membrane and can assist the translocation of the C domain.
  • the first is that the nearly parallel packing of the three helix layers would permit spreading on the membrane surface of the first helix layer (TH1-TH3) if other layers were inserted. This insertion would require local conformational changes in loops, but no alteration of the helices themselves. Also the pronounced hydrophobic asymmetry is compatible with the
  • the chimeric molecules of the invention can be used, for example, to kill particular classes of cells.
  • a chimeric diphtheria toxin molecule which binds specifically to cells bearing the interleukin-2 receptor can be used in treatment of various autoimmune diseases, e.g., arthritis.
  • Non-cytotoxic hybrid molecules in which a chimeric diphtheria toxin molecule is linked to a second molecule can be used to introduce the second
  • Tay-Sachs may be treated by introducing hexosaminidase A into appropriate cells.
  • diphtheria toxin catalytic domain of chimeric diphtheria toxin molecules can be replaced by the catalytic domain of other toxin molecules to generate other targeted cytotoxins.
  • Peptide toxins are preferred, but others are also useful. Many peptide toxins have a generalized eukaryotic
  • toxin must be modified to prevent intoxication of non-targeted cells. Any such modifications must be made in a manner which preserves the cytotoxic functions of the molecule.
  • Potentially useful toxins include, but are not limited to: cholera toxin, ricin, O-Shiga-like toxin (SLT-I, SLT-II, SLT II V ), LT toxin, C3 toxin, Shiga toxin, pertussis toxin, tetanus toxin, Pseudomonas exotoxin, alorin, saporin, modeccin, and gelanin.
  • the catalytic domain is to be removed for the purpose of creating a hybrid molecule which includes a molecule to be introduced into a selected class of cells, it is preferred that the molecule to be introduced be fused to the chimeric diphtheria toxin molecule just to the amino-terminal side of the junction of Fragment A and Fragment B, i.e., to the amino-terminal side of residue 190.

Abstract

The invention features a chimeric diphtheria toxin molecule wherein all or part of a complementarity determining region of an antibody is inserted into a loop region of the Diphtheria toxin receptor binding-domain.

Description

NOVEL DIPHTHERIA TOXIN-BASED MOLECULES
Background of the Invention
The field of the invention is chimeric molecules. Hybrid molecules in which all or part of an antibody is fused to another molecule have been suggested as a means for targeting molecules to particular sites. Diphtheria toxin (DT) is an extremely potent cytotoxin which is secreted by Corynebacterium diphtheriae that has been lysogenized by a bacteriophage carrying the
Diphtheria toxin gene. Naturally occurring Diphtheria toxin is a single polypeptide chain of 535 residues . Mild trypsinization and reduction of Diphtheria toxin in vitro generates two fragments, Fragment A (amino-terminal, ~21K) and Fragment B (carboxy-terminal, ~37K), as a result of cleavage at residue 190, 192, or 193. A similar proteolytic cleavage ('nicking') occurs in vivo before or soon after the toxin binds to a sensitive cell. Fragment B of the toxin binds the protein to receptors on the cell surface and promotes transfer of the Fragment A to the cytoplasm. Fragment A in the cytoplasm catalyzes the transfer of the ADP-ribosyl group of NAD+ to
elongation factor 2 (EF-2). This inactivates EF-2, stopping protein synthesis and killing the target cell. Introduction of a single molecule of Fragment A into the cytoplasm can kill a cell. While the exact mechanism by which Diphtheria toxin enters a cell is not completely understood, it is known that Diphtheria toxin binds a receptor and is endocytosed and delivered to endosomes where it encounters acidic conditions. At a threshold pH of ~5.0 the toxin undergoes a conformational change, which promotes insertion and formation of an ion-selective channel in the membrane, and Fragment A is translocated and released into the cytoplasm. Summary of the Invention
In general, the invention features a chimeric diphtheria toxin molecule wherein all or part of a complementarity determining region (CDR) of an antibody is inserted into a loop region of the Diphtheria toxin receptor binding-domain. By "loop region" is meant any of the portions of the Diphtheria toxin receptor binding domain lying between β strands as delimited herein. The term encompasses single amino acids. By "Diphtheria toxin receptor binding domain" is meant the portion of Diphtheria toxin lying from amino acids 386 to amino acid 535, inclusive. By "CDR" is meant a portion of a
complementarity determining region of an antibody as defined by sequence heterogeneity, e.g., according to Kabat et al (in Sequences of Proteins of Immunological Interest, U.S. Dept. of Health and Human Services, U.S. Government Printing Office, 1987). In a preferred
embodiment, the antibody is capable of specifically binding a cell surface antigen expressed on a cell, the chimeric diphtheria toxin molecule is capable of
specifically binding the same cell surface antigen and is substantially incapable of binding to the diphtheria toxin receptor. By "cell surface antigen" is meant any cell surface marker, e.g., a protein or a carbohydrate. By "specifically binding" is meant does not substantially bind to other molecules. By "diphtheria toxin receptor" is meant the receptor for naturally-occurring Diphtheria toxin. In a more preferred embodiment, the molecule is capable of decreasing the viability of the cell. In an even more preferred embodiment the molecule kills the cell.
In another preferred embodiment, all or part of a first CDR is inserted into a first loop region, all or part of a second CDR is inserted into a second loop region, and the first and the second CDR are of a single antibody chain. By "of a single antibody chain" is meant CDR sequences found within a single heavy or light chain. In a preferred embodiment, the molecule lacks diphtheria toxin catalytic activity. By "diphtheria toxin catalytic activity" is meant the ability to inhibit translation. In a yet more preferred embodiment, the molecule lacks all or part of the catalytic domain of diphtheria toxin.
In other preferred embodiments, the loop region is RL3 and the CDR is a CDR1; and the loop region is RL9 and the CDR is a CDR3.
In a related aspect, the invention features a hybrid molecule which includes a first and a second portion joined together covalently, the first portion includes a chimeric diphtheria toxin molecule wherein all or part of a CDR of an antibody is inserted into a loop region of the receptor binding-domain of diphtheria toxin, the antibody being capable of specifically binding a cell surface antigen expressed on a cell, the chimeric diphtheria toxin molecule being capable of specifically binding the cell surface antigen, being substantially incapable of binding to the diphtheria toxin receptor, and lacking Diphtheria toxin catalytic activity; and the second portion includes a molecule to be delivered to the cell. In various preferred embodiments, the molecule to be delivered to the cell is a protein; is an enzyme; is a protein which modulates transcription; is a nucleic acid binding protein; is a nucleic acid-binding protein capable of binding a single-stranded nucleic acid; and is a nucleic acid.
In a related aspect, the invention features a hybrid molecule which includes a first and a second portion joined together covalently, the first portion includes a chimeric diphtheria toxin molecule wherein all or part of a CDR of an antibody is inserted into a loop region of a first diphtheria toxin receptor binding-domain, the antibody being capable of specifically binding a cell surface antigen expressed on a cell, the chimeric diphtheria toxin molecule being capable of specifically binding the same cell surface antigen and being substantially incapable of binding to the
diphtheria toxin receptor, wherein the amino-terminus of a second diphtheria toxin receptor-binding domain is connected to the carboxy-terminus of the first diphtheria toxin receptor-binding domain, the second diphtheria toxin receptor-binding domain being substantially
incapable of binding to the diphtheria toxin receptor. In preferred embodiments, the carboxy terminus is
connected to the amino terminus through a polypeptide chain; and all or part of a CDR of an antibody is
inserted into a loop region of the second diphtheria toxin receptor-binding domain. By "connected" is meant linked via one or a series of covalent bonds, e.g., by a polypeptide chain. In a more preferred embodiment, all or part of a first the CDR is inserted into a first loop region of the first diphtheria toxin receptor-binding domain and all or part of a second the CDR is inserted into a second the loop region of the first diphtheria toxin receptor-binding domain. In an even more preferred embodiment, all or part of a third CDR is inserted into a first loop region of the second diphtheria toxin
receptor-binding domain and all or part of a fourth CDR is inserted into a second the loop region of the second diphtheria toxin receptor-binding domain. In a yet more preferred embodiment, the first and the second CDR are of a first antibody chain and the third and the fourth the CDR are of a second antibody chain. In still more preferred embodiments, the first and the second antibody chains are from antibodies recognizing the same antigen; the first and second antibody chains are from the same antibody molecule; and the first antibody chain is the light chain of an antibody and the second antibody chain is the heavy chain of the same antibody.
In a related aspect, the invention features a chimeric diphtheria toxin molecule wherein all or part of a CDR-like sequence of a Iigand binding protein having an antibody variable domain-like Iigand binding-domain is inserted into a loop region of the receptor binding domain of diphtheria toxin. By "CDR-like sequence" is meant a sequence which is responsible for Iigand binding and which has the same relationship to overall structure of a protein as does the CDR of an antibody variable domain. By "an antibody variable domain-like Iigand binding domain" is meant a Iigand binding domain which has structural homology to an immunoglobulin variable domain. For example tumor necrosis factor includes an antibody variable domain-like Iigand binding domain.
The chimeric molecules of the invention bind specifically to the same epitope (antigen) as the
antibody from which the inserted CDR sequences are derived. Thus it is possible to generate a molecule targeted to any antigen. Because these chimeric
molecules can enter cells to whihc they bind, they can be used to introduce any molecule into a specific class of cells.
Other features and advantages of the invention will be apparent from the following description of the preferred embodiments thereof, and from the claims.
Detailed Description
The drawings are first briefly described.
Figure 1 is a schematic drawing of Diphtheria toxin in which each secondary structural segment is identified. The first letter denotes the domain: C for catalytic, T for transmembrane, and R for receptor-binding domains. The second letter denotes the secondary structure class: H for helix, B for β strand, L for loop. The third symbol is the sequential number of each
secondary segment from the N-terminus of each domain.
The residue numbers in each segment are as follows:
CH1:2-7, CB1:11-14, CB2:16-24, CH2:28-34, CB3:52-57,
CH3:58-66, CB4:76-86, CB5:88-96, CH4:99-106, CH5:120-126, CB6:130-136, CB7: 147-152, CB8: 159-166, CH6:168-173,
CH7:176-186; TH1:205-221, TH2:225-231, TH3:238-257,
TH4:258-269, TH5:274-288, TH6:297-307, TH7:310-315,
TH8:326-346, TH9:356-378; RB1:386-390, RB2:393-399,
RB3:412-424, RB4:428-438, RB5:447-453, RB6:455-465,
RB7:467-480, RB8:483-495, RB9:513-520, and RB10:525-534.
Figure 2 is a representation of the Cα skeleton of Diphtheria toxin from the same viewpoint as that of Fig. 1. An ApUp molecule occupies the active site of
Diphtheria toxin.
Figure 3 is a stereo pair representation of the electron density maps calculated at 2.5A from (2Fob-Fc) and the refined model phases. Maps are superimposed on the corresponding region of the refined model.
Figure 4 is a representation of the Diphtheria toxin dimer observed within the Form4 crystal. The two monomers are related by a crystallographic 2-fold
rotation axis, which is vertical. The molecule at the left (in thick line) has the same orientation as that in Fig. 1.
Figure 5 is a stereo pair representation of the Cα skeleton of the C domain. The entrance to the active site is at the lower right. The four loops, CL1 to CL4, are highlighted. Notice that they form a hinge which may permit the C domain to form a more elongated structure.
Figure 6 is a stereo pair representation of the Cα skeleton of the T domain, with the direction of view from the right side of Diphtheria toxin in Fig. 1. Helix TH1 lies in back, starting at residue 205. Helix TH2 runs to the left at the bottom, followed by a turn and helix TH3 running to the right. In front center is TH5 (running to the left) and above it are helices TH6 and TH7. Behind these pairs of antiparallel helices is another pair of antiparallel helices, TH8 and TH9, with TH9 running upwards and ending at residue 378. The Asp and Glu side chains are shown. Notice the tips of two helix layers, TL3 and TL5 contain a total of six acidic groups (on the left).
Figure 7 is a stereo pair representation of the T domain as in Fig. 6 except that the Lys, Arg and His side chains are shown. Notice the positive charge asymmetry, with all charges at the bottom and back of the domain, with an exception Lys299 near the loop TL3 between TH5 and TH6.
Figure 8 is a schematic representation of the R domain of Diphtheria toxin (panel A), an Ig variable domain (panel B) and tumor necrosis factor (panel C). R domain is viewed in the direction from the back side of Diphtheria toxin in Fig. 1. Numbers from 2 to 10 of the R domain represent the strands RB2 through RB10 of
Diphtheria toxin. Notice that strands 2, 3, 4, 8, 9, and 10 of the R domain correspond well to strands A, B, C, E,
F, and G of the Ig variable domain. Also strands 3, 4, 5, 6, 7, 8, and 9 correspond well to strands C, D, E, F,
G, H, and I of tumor necrosis factor, a classical
jellyroll.
Figure 9 is a schematic drawing of a rearranged Diphtheria toxin receptor-binding domain. Each
structural segment is identified.
Figure 10, panel A is a schematic representation of the receptor-binding domain of diphtheria toxin with each β strand labeled (RB1-RB10) and each loop region labeled (RL1-RL9). The amino acid end points refer to the β sheets, e.g., RB3 consists of residues 412-424 inclusive and RL3 consists of residues 425-427 inclusive. Panel B is a schematic representation of a Diphtheria toxin molecule which has undergone segment rearrangement. The notation for the regions is a in panel A except that 447, 483, 467, and 455 indicate the amino acid residues of immediately adjacent to residues 407, 455, 445, and 483 respectively.
Chimeric Diphtheria Toxin Molecules
Described below is the structure of diphtheria toxin as determined by x-ray crystallography. As will be discussed more fully below, the receptor-binding domain of diphtheria toxin has a structure similar to the variable domain of an antibody. Because of this
similarity, Diphtheria toxin can be modified so that certain portions of its receptor-binding domain are replaced by, or modified to include, antigen-binding portions (complementarity determining regions) of an antibody of choice. Such modification results in the creation of a chimeric diphtheria molecule which
recognizes and binds the same antigen as the selected antibody. If a chimeric molecule is modified so as to substantially eliminate binding to the diphtheria toxin receptor, it will selectively bind only to cells bearing the antigen recognized by the antibody from which the complementarity determining regions were derived.
Chimeric molecules of the type described above can be targeted to selected cell types. For example,
portions of an antibody directed against the interleukin-2 (IL-2) receptor can be used to make a chimeric
diphtheria toxin molecule which binds to cells bearing the IL-2 receptor. If the chimeric molecule is designed so as to retain the translocation and catalytic functions normally associated with diphtheria toxin, the chimeric molecule will enter and kill cells bearing the IL-2 receptor. If in the course of creating this chimeric diphtheria toxin molecules the receptor binding domain is altered so that the chimeric molecule does not bind to the diphtheria toxin receptor, this chimeric molecule will bind and kill cells bearing the IL-2 receptor while leaving all other cells unharmed.
Alternatively, chimeric diphtheria toxin molecules can be used to introduce any molecule into a selected group of cells. For example, if a chimeric diphtheria toxin molecule capable of binding to cells bearing the IL-2 receptor is modified so that the catalytic domain of diphtheria toxin is replaced by an enzyme, that enzyme can be selectively introduced into cells bearing the IL-2 receptor. Similar modifications would permit an
antisense RNA molecule capable of blocking translation of selected RNA to be introduced into cells bearing the IL-2 receptor. Alternatively, the catalytic domain may be substantially inactivated by mutation rather than
deletion.
The approaches used to create chimeric diphtheria toxin molecules are completely general. Once the target antigen, e.g., a cell surface protein or carbohydrate, has been selected, a chimeric diphtheria toxin molecule is created by: (1) generating (or selecting) an antibody which recognizes the antigen; (2) cloning and sequencing at least the variable domain of a heavy or light chain of the antibody; (3) identifying the complementarity
determining regions within the antibody variable domain; (4) modifying diphtheria toxin to insert all or part of a complementarity determining region (s) into a loop
region(s) of the diphtheria toxin receptor binding domain; and (5) testing the ability of the chimeric molecule to bind to the selected antigen.
The molecules of the invention can be more completely understood by first detailing the structure of Diphtheria toxin itself. Accordingly, the overall structure of Diphtheria toxin is discussed below followed by a detailed discussion of the structure of its
receptor-binding domain. The relationship between the diphtheria toxin receptor-binding domain and an antibody variable domain is then described. This is followed by a discussion of methods for generating and screening chimeric molecules. This discussion is followed by a description of the structure of other parts of Diphtheria toxin including the catalytic domain, the translocation domain and the domain junctions. Lastly, details of the structure determination are presented.
Structure of Diphtheria toxin
Diphtheria toxin consists of three abutting domains that are connected by interdomain linkers. The amino-terminal domain (residues 1-193) is the catalytic (C) domain. The middle domain (residues 205-378) is the transmembrane (T) domain, and the carboxy-terminal domain (residues 386-535) is the receptor binding (R) domain. Schematically, Diphtheria toxin is Y-shaped with the base formed by the T domain, one arm of the Y formed by the C domain, and the other arm formed by the R domain. The Y is about 9θA high, 5θA across the top of the Y, but only 30Å thick (Figure l).
Each of the three domains has a distinctive fold. The C domain is a mixed structure of eight β strands (CB1-CB8) and seven α-helices (CH1-CH7). The eight β strands form two β sheets of 3 and 5 strands each. These β sheets form a core that is surrounded by 7 short helices. The overall folding of the C domain is similar to that, of Pseudomonas aeruginosa exotoxin A (ETA) especially near the active site (Allured et al., Proc. Natl . Acad. Sci USA 83:1320, 1986), a result that had been foreshadowed by a weak similarity in amino acid sequences (Carroll et al., Mol . Microbiol . 2:293, 1988; Brandhuber et al., Proteins 3:146, 1988). Sixma et al. (Nature 351:371, 1991) recently demon, rated that the folding of the active site region of E. coli heat labile enterotoxin also closely resembles that of ETA. The T domain contains nine helices (TH1-TH9) that are folded into three helix layers, each of which is formed by two or more antiparallel helices. A similar feature was observed in the structure of the channel-forming domain of colicin A (Parker et al., Nature 337:93, 1989). The R domain contains ten β strands (RB1-RB10), nine of which (RB2-RB10) build two β sheets. These two β sheets form a β sandwich with a topology similar to a jellyroll fold (Richardson, Adv. Protein Chem . 34:167, 1981). The three-domain organization of Diphtheria toxin is shared by two other bacterial toxins, ETA and 5-endotoxin from Bacillus thuringiensis (Carroll, et al., Nature 353:815, 1991). The catalytic domains of Diphtheria toxin and ETA are the closest among all these domains in their
structures and functions.
Receptor-binding domain
Referring to Fig. 8 (panel A) and Fig. 10 (panel A), the receptor-binding (R) domain is formed from two β sheets, β strands RB2 (residues 393-399), RB3 (residues 412-424), RB5 (residues 445-453), and RB8 (residues 483-495) form a four-stranded β sheet that faces a five-stranded β sheet containing β strands RB4 (residues 428-438), RB6 (residues 455-465), RB7 (residues 467-480), RB9 (residues 513-520), and RB10 (residues 525-534). RB6 interacts with both β sheets through hydrogen bonds. The connection of the strands is such that the R domain is similar to the jellyroll topology found in many proteins that are exclusively formed from antiparallel β strands (Richardson, J. Adv. Protein Chem . 34:167, 1981).
Jellyroll domains include viral coat proteins, tumor necrosis factor, and the receptor-binding domain of ETA. The R domain differs somewhat from a strict jellyroll topology (Fig. 9) in having strand 2 in the "front" sheet, and having a strand 10 in the "back".
The R domain is also similar in structure to an immunoglobulin (Ig) variable domain (Fig. 9, panel B), but differs from the Ig fold in having an "insert" of strands 5 and 6 between 4 and 7, and also in lacking two short strands (C' and C" in Fig. 9, panel B) between 4 and 5. The portion of the R domain that resembles a strict jellyroll in topology is the right side as viewed in Fig. 9; and the portion that resembles the Ig variable domain is the left side, the side that is away from the rest of the Diphtheria toxin monomer.
Chimeric Diphtheria Toxin Molecules
In chimeric diphtheria toxin molecules all or part of one or more complementarity determining regions derived from an antibody are inserted into one or more loop regions of the Diphtheria toxin receptor-binding domain. Generally, only one CDR sequence is inserted into each loop and the insertion may or may not be accompanied by deletion of all or a portion of the loop region.
The design of chimeric diphtheria toxin molecules can be more readily understood by first considering certain aspects of antibody structure. An antibody consists of two identical light chains (L) and two identical heavy chains (H). Each light chain is attached to a heavy chain by one or more disulfide bonds.
Likewise, the two heavy chains are attached to each other by one or more disulfide bonds. Overall, a single antibody forms a "Y" shaped structure in which the carboxy-terminal portion of the heavy chains forms the base of the Y and the amino-terminal portion of a single heavy chain and the amino-terminal portion of a single light chain together form each arm. Each chain, heavy or light, is composed of structurally similar domains. The domains are referred to as constant or variable based on sequence heterogeneity. Proceeding from the carboxy-terminus, a heavy chain is composed of the CH3 constant domain, the CH2 constant domain, the CH1 constant domain, and the VH variable domain. Proceeding from the carboxy-terminus, a light chain is composed of a CL constant domain followed by a VL variable domain.
The variable domains (VL and VH) are of particular interest since together they form the antigen binding site. Each variable domain is approximately 110 amino acids long and is composed of three hypervariable or complementarity-determining regions (CDR1, CDR2, and CDR3) interspersed with four less-variable framework regions (FR1, FR2, FR3, and FR4). The complementarity-determining regions (collectively, CDR's) are responsible for antigen recognition.
Structurally, each variable domain consists of two β sheets which together form a structural motif often referred to as the immunoglobulin fold. (Constant regions, with a slight variation, also form an
immunoglobulin fold.) One β sheet is composed of four β strands (A, B, D and E) the other sheet is composed of five β strands (C, C, C", F and G). There are loops between each β strand. Three of these loops, to a first approximation, correspond roughly to the three CDR's.
However, as will be discussed below, the identification of a CDR or a framework region is based primarily on sequence heterogeneity rather than secondary structure. Thus, the identification of these regions within a given antibody molecule requires analysis of the amino acid sequence of the antibody. This caveat notwithstanding, the loop between B and C often includes all or part of CDR1; the loop between C and C" often includes all or part of CDR2; and the loop between F and G often includes all or part of CDR3. A comparison between the structure of the receptor binding domain of Diphtheria toxin and immunoglobulin V domain illustrates the structural similarities (Fig. 9). There is a correspondence, not identity, between the β strands of the Diphtheria toxin receptor domain
(described above) and the β strands of an immunoglobulin variable domain as follows: RB2 ≡ A, RB3 ≡ B, RB4≡ C, RB5 ≡ D, RB6≡ C" , RB7 ≡ C, RB8 ≡ E, RB9 ≡ F, RB10≡ G. Two of the Diphtheria toxin receptor binding domain loops, RL3 and RL9 thus correspond to CDR1, and CDR3, respectively.
Given this understanding, it can be seen that the receptor domain can be engineered to more closely
resemble a variable domain. In particular, CDR sequences can be inserted into RL3 and RL9 in a process referred to herein as 'loop grafting' to yield a chimeric molecule which includes a CDR1 sequence or a CDR3 sequence or both. In the course of grafting all or part of either or both loop regions may be deleted. Further, the region of Diphtheria toxin from RB5 to RB8 can be rearranged so that it more closely resembles the region of a variable domain extending from strand C through strand E. This process, referred to herein as 'segment rearrangement', can provide a framework for the grafting of a CDR2 sequence into Diphtheria toxin. In combination, these two approaches can be used to create a chimeric molecule into which three CDR's (CDR1, CDR3, CDR3) have been introduced. It may also be possible to insert a CDR2 sequence into RL5 without segment rearrangement.
LOOP Grafting
Loop grafting is similar to CDR grafting in which the CDR of a first antibody are exchanged for those of a second antibody, and the techniques employed in CDR grafting will, in general, be useful for loop grafting. Jones et al. (Nature 321:522, 1986), Riechmann et al. (Nature 322:323, 1988), Winter et al. (PCT/GB89/00113), Winter (EPA 0 239 400), and Clackson et al. (Nucl . Acids Res . 17:10163, 1989) describe CDR grafting techniques which can be applied to loop grafting.
It should be understood that the precise limits of the regions to be grafted are a matter of experimental choice. All or part of RL3 (residues 425-427) could be replaced by all or part of the CDR1 of an antibody heavy or light chain of choice. All or part of RL9 (residues 521-524) could be replaced by all or part of the CDR2 of the same antibody heavy or light chain (or, less
preferably, an antibody heavy or light chain of an antibody recognizing the same epitope). Of course, the above boundaries of RL3 and RL9, while precise in terms of structure, represent only approximate limits to the region which might be replaced by all or part of a CDR. Further insertion of a CDR can take place without the deletion of any loop region sequence. Thus, a CDR may replace a few residues of RB3 and a few residues of RB4 as well as all of RL3 (e.g., residues 422-429).
ALternatively, a CDR might replace only a part of RL3 (e.g., residues 426 and 427). Alternatively, no residues are deleted.
In identifying a CDR to graft into Diphtheria toxin, it should be . understood that CDR's are identified by sequence hypervariability (Kabat et al., in Sequences of Proteins of Immunological Interest , U.S. Dept. of Health and Human Services, U.S. Government Printing
Office, 1987) and/or structural hypervariability (Chothia et al. J. Mol . Biol . 196:901, 1987) rather than by secondary structure (e.g., a loop). Thus, a CDR so-defined may include only a portion of the loop between two β-strands and likewise may include part or all of one or both jø-strands flanking the loop. Nevertheless, it should be understood that in many antibodies CDR's are found at similar positions. Thus, within a variable domain CDRl is commonly located near amino acid 30, CDR2 is commonly located near amino acid 50 and CDR3 is commonly located near amino acid 95 (Roitt et al.,
Immunology Gower Medical Publishing, London, 1985).
Useful CDR's may be derived from either
immunoglobulin H or L chains. Further, antibodies derived from any species may be used as a source of
CDR's. Because CDR's appear not to contain species specific motifs, CDR's from a first species can be used without substantially increasing the immunogenicity of the chimeric molecule in a second species.
Segment Rearrangement
Segment rearrangement essentially consists of reorganizing the portion of Diphtheria toxin from the beginning of RB5 to the end of RL7. This reorganization results in the formation of a rearranged R domain which more closely resembles an antibody variable domain than does the native R domain. This rearrangement in
combination with loop grafting of all or part of RL3 and/or RL9 can provide a molecule with improved antigen binding characteristics compared to a molecule which has undergone grafting of the same loops but has not
undergone segment rearrangement. Domain rearrangement can also provide a location, between rearranged RB7 and rearranged RB6, for the addition of a CDR2.
In detail domain rearrangement entails
constructing a molecule in which the R domain has the following sequence of elements (beginning at its amino-terminus): RB1-RL1-RB2-RL2-RB3-RL3-RB4-RL4-RB7-RL7-RB6-RL6-RB5-RL5-RB8-RL8-RB9-RL9-RB10 (Fig. 10, panel B). All or part of a CDR2 sequence can be introduced by replacing all, part, or none of RL7 (in the rearranged molecule located between RB7 and RB6). Referring to Fig. 9, panel A, this rearranged receptor-binding domain more closely resembles an immunoglobulin variable domain (Fig. 9, panel B) than does the naturally occurring Diphtheria toxin receptor-binding domain (Fig. 8, panel A).
Eliminating Binding to the Diphtheria Toxin Receptor
The diphtheria toxin receptor binding ability of Diphtheria toxin receptor domain must be substantially reduced in chimeric molecules compared to native
Diphtheria toxin so that the chimeric molecules do not substantially bind to or enter non-targeted cells (i.e., cells not bearing the antigen recognized by the CDR's). This can be accomplished by incorporating into the chimeric molecules certain mutations which reduce binding of Diphtheria toxin to its natural receptor. CRM9 (Hu et al., Biochim . Biophys . Acta 902:24, 1987) CRM107 and CRM103 (Greenfield et al., Science 238:536, 1987) are mutant Diphtheria toxin molecules with reduced receptor binding. The sequences changes in these mutants can be incorporated into chimeric diphtheria toxin molecules. It should also be recognized that replacing all or part of RL3 and/or all or part of RL9 with a CDR may
essentially eliminate binding of the chimeric molecule to the diphtheria toxin receptor. Further, the insertion of a CDR sequence into a loop of the R domain in the absence of any deletion may, in and of itself, substantially eliminate binding to the diphtheria toxin receptor.
While the diphtheria toxin receptor has not been
positively identified (Naglich et al., Proc . Nat 'l Acad . Sci . USA. 89:2170, 1992), the diphtheria toxin receptor-binding ability of chimeric diphtheria toxin molecules can be assessed by standard techniques (Middlebrook et al., Can . J. Microbiol . 23:183, 1978; Middlebrook et al., J. Biol . Chem . 253:7325, 1978) using Vero cells or other cell lines.
Chimeric Diphtheria Toxin Molecules Having Two R Domains In naturally occurring antibodies the antigen binding site is formed by a VH domain and a VL domain, and structural studies suggest that antigen binding is mediated by contacts with both domains. Fv fragments, which are non-covalently associated heterodimers of VH and VL domains, have been developed to provide small, engineered molecules with antigen binding activity similar to the intact antibody from which the domains were derived (Glockshuber et al., Biochemistry 29:1362, 1990). Because Fv molecules are prone to dissociation, single-chain Fv molecules (sFv) have been developed by linking the domains with a flexible hydrophilic
polypeptide (Bird et al., Science 423:423, 1988; Huston et al., Proc. Nat 'l . Acad . Sci. USA 85:5879, 1988). As an alternative, the domains can be linked by disulfide bonds (Glockshuber et al., Biochemistry 29:1362, 1990). In a similar fashion it is possible to generate chimeric diphtheria toxin molecules having two R domains. When properly constructed the R domains of such molecules resemble a single-chain Fv fragment. Loops within one or both R domains can be modified to include CDR sequences. These changes, in combination with modifications which prevent either receptor binding domain from recognizing the diphtheria toxin receptor, result in the creation of a molecule which will specifically recognize the same antigen as the antibody from which the CDR sequences were derived.
To generate chimeric diphtheria toxin molecules having two R domains, the carboxy-terminus of RB10 of a diphtheria toxin molecule is fused to the amino-terminus of RB1 or RB2 of an R domain portion of diphtheria toxin. CDR sequences can be introduced into these molecules in manner described above by modifying the loop regions of one or both of the R domains. The result is a molecule which has a domain (R+R') resembling an sFv molecule fused to the diphtheria toxin translocation and catalytic domains. In certain circumstances it may be possible to generate antigen binding molecules by introducing CDR sequences into only one of the two R domains. This is because the mere existence of an unmodified R domain may improve contacts between the CDR sequences in the other modified R domain and the antigen. In any case, if the chimeric molecule is to be specifically targeted, both receptor binding domains must be modified to essentially eliminate recognition of the diphtheria toxin receptor. If both R domains have been engineered to introduce CDR sequences, it is preferred all of the CDR sequences be derived from the same antibody (or at least antibodies recognizing the same epitope), and that the CDR sequences of one R domain be derived from a VH domain, and the that the CDR sequences of the other R domain be derived from a VL domain. This creates a molecule which more closely resembles an Fv fragment.
In general, methods used for the generation of sFv molecules (Bird et al., Science 423:423, 1988; Huston et al., Proc. Nat 'l . Acad . Sci . USA 85:5879, 1988) can be used to generate chimeric diphtheria toxin molecules having two R domains. In designing chimeric diphtheria toxin molecules in which the two R domains are linked by a polypeptide chain it is important that the linking polypeptide chain be selected so as to hold the two R domains in a configuration that resembles an Fv molecule. Lardner et al. (US Patent 4,946,778 and US Patent
4,704,692) describe techniques for selecting polypeptides to link VH and VL domains to form sFv. The same
techniques can be used to generate chimeric diphtheria toxin molecules having two R domains. As an alternative, the two domains can be linked by disulfide bonds as described by Glockshuber et al. (Biochemistry 29:1362, 1990) for single-chain Fv molecules. As discussed above for simple chimeric diphtheria toxin molecules, chimeric diphtheria toxin molecules having two R domains and CDR sequences can be modified to act as delivery vehicles rather than cytotoxins. This is accomplished by replacing the catalytic domain with a molecule to be introduced into cells thus creating a hybrid molecule. The R domains will then target the hybrid molecule to a selected class of cells and the translocation domain will mediate entry.
Generation of Antibodies and Identification of CDR's
In order to create a chimeric diphtheria toxin molecule directed against a selected antigen, it is first necessary to identify an antibody directed against that antigen. In many instances appropriate antibodies will already be available (see Kabat et al., supra; Catalogue of Cell Lines and Hybridomas, American Type Culture
Collection, Rockville, MD). Alternatively, antibodies (polyclonal or monoclonal) directed against the selected antigen can be generated and screened by standard methods (Current Protocols in Immunology, Wiley-Interscience, New York, 1991). Once a hybridoma secreting an antibody with the desired specificity has been isolated there are several approaches which can be used to sequence the variable domain for the purpose of identifying CDR sequences. The heavy and/or light chains can be cloned and sequenced. Alternatively, variable domains can be amplified for cloning using the polymerase chain reaction and oligonucleotide primers which recognize conserved sequences at each end of the heavy or light chain
variable region (Orlandi et al., Proc. Nat'l . Acad. Sci . USA 86:3833, 1989; Larrick et al., Biochem. Biophys. Res . Comm . 160:1250, 1989; Sastry et al., Proc. Nat 'l Acad. Sci . USA 86:5728, 1989). This approach allows the cloning of the variable regions of human antibody genes from unstable human-mouse hybridomas as well as the cloning of variable regions from other unstable
hybridomas, single hybridoma cells, and single B
lymphocytes. These techniques permit the expression of antibody fragments in bacteria (Skerra et al., Science 240:1038, 1988; Better et al., Science 240:1041, 1988) or on the surface of phage (McCafferty et al., Nature
348:552, 1990). Expression in one or another of these systems permits the use of a number of efficient
screening methods (Skerra et al., Analytical Biochem .
196:151, 1991; Huse et al.. Science 246:1275, 1989) which can be used to identify antibodies fragments that bind the selected antigen with the desired affinity.
In addition various non-immunization techniques (Marks et al., J. Mol . Biol . 222:581, 1991; Persson et al., Proc. Nat 'l Acad . Sci . USA 88:2432, 1991. Huse et al., supra) can be used to generate antibodies which can serve as a source of CDR sequences.
Once the variable regions are sequenced CDR's are identified according to Kabat et al. (supra) .
Generation and Screening of Chimeric Diphtheria Toxin Molecules
The chimeric molecules are generated using the standard techniques of molecular biology (Current
Protocols in Molecular Biology, Wiley-Interscience, New York, 1991). The primary approach involves the
generation of nucleic acids encoding the chimeric
molecules. The chimeric molecules themselves can be produces in bacterial cells, mammalian cells, or insect cells by standard techniques. In designing chimeric molecules the techniques of computer-based molecular modeling may be useful. The coordinates of the solved diphtheria toxin structure are included (appendix) to aid in this process. It should be understood that changes in the amino acid sequence may be introduced at any position to generate more stable molecules or molecules with higher binding specificity.
Once the chimeric molecules have been produced, they can be screened for antigen binding ability using any of the approaches described above for antibodies and antibody fragments. The binding specificity of a
cytotoxic chimeric molecules targeted to a particular antigen can be determined by comparing the toxicity of the molecule toward cells bearing the antigen to its toxicity towards cells not bearing the antigen. For non-toxic chimeric molecules, a detectable label may be covalently linked to the chimeric molecule to facilitate comparison of binding to antigen-bearing cells and cells not bearing antigen.
Examples
A chimeric diphtheria toxin molecule capable of recognizing cells bearing the Campath-1 antigen can be constructed by replacing all of RL3 with the CDR1
sequence identified by Waldman et al. (PCT/GB89/00113, hereby incorporated by reference) and replacing all of RL9 with the CDR3 sequence identified by Waldman et al. (supra) .
A chimeric diphtheria toxin molecule capable of recognizing cells bearing the interleukin-1 receptor can be constructed by replacing all of RL3 with residues 26-33 of the anti-tac antibody light chain and replacing all of RL9 with residues 99-107 of the anti-tac antibody light chain, (supra) .
Structure determination
The structure is based on analyses of Form1, Form3, and Form4 crystals. Form1 crystals of Diphtheria toxin complexed with adenylyl-3',5'-uridine monophosphate (ApUp) belong to triclinic space group P1 with unit cell dimensions of a=70.4Å, b=70.6Å, c=65.4Å, α=94.9°,
β=91.0°, and γ=99.6° with two chains per asymmetric unit. This dimeric asymmetric unit is consistent with the fact that a dimeric form of Diphtheria toxin is sometimes found in crude or urified preparations of the protein
(Collier et al., J. Biol . Chem . 257:5283, 1982). Dimeric Diphtheria toxin itself is not toxic, presumably because it does not bind to receptors, but it slowly dissociates to fully toxic monomers (Carroll et al., Biochem .
25:2425, 1986). The dimer may represent a
conformationally altered form of the biologically active monomeric toxin. Irreproducible crystallization
conditions for obtaining Form1 crystals hampered
crystallographic studies of structure determination until three new crystal forms were obtained (Fujii et al., J. Mol . Biol . 222:861, 1991). Form3 and Form4 belong to monoclinic space group C2 with unit cell dimensions for Form3 of a=107.3Å, b=91.7Å, c=66.3Å, and β=94.7°, and for Form4 of a=108.3Å, b=92.3Å, c=66.1Å, and β=90.4°. In both of these forms there is one Diphtheria toxin chain per asymmetric unit and pairs of Diphtheria toxin chains are related by a 2-fold rotation axis.
The initial model was based on the structure determination of Form4 crystals at 3 . 0Å resolution, using the multiple isomorphous replacement (MIR) method
followed by solvent flattening (Wang, Methods in Enzymol . 115:90, 1985). With the initial model, the structures of Form1 and Form3 were readily solved by molecular
replacement (Brunger, Acta Cryst. , A47:195, 1991; Rossman et al., Acta Cryst . , 15:24, 1962). Single isomorphous replacement (SIR) phases were also obtained for Form3. Native data were then collected to 2.5Å resolution, and the model was rebuilt into 2.5A maps with Form3 (SIR) and Form4 (MIR) after the phases had been extended and modified by the method of Zhang et al. Acta Cryst .
A46:377, 1991. This was followed by real-space density averaging between two forms. Sequence fitting was difficult in the ~120 C-terminal residues (part of receptor-binding or R domain) where the most ambiguous regions were near residues 408 and 510. Some of the useful markers in the density maps were W50, W153, W281, W398, a 5-residue segment of M178, Y179, E180, Y181, M182, a 4-residue segment of F355, Y358, H372, Y375, a cluster of Y514, F530, F531, witn big side chains near the carboxy terminus (Fig. 3), and two disulfide bonds between C186 and C201 and C461 and C471. An initial improper fitting in the R domain was detected by profile window plots (Lϋthy et al., Nature 355:xxx, 1992) and then corrected.
Iterative cycles of refinement were carried out
independently at 2.5A for each data set. The atomic model for each form is essentially identical except for crystal packing. Assessment of the accuracy of the model rests on the fit of the model to the MIR and density-modified maps, crystallographic R-factors, real-space
R-factors, (Jones et al., Acta Cryst . A47:110, 1991), the free R-value (Briinger Nature 355:472, 1992), which is only 4% higher than the crystallographic R factor, and profile window plots (Lϋthy, supra). At the present stage of refinement, the agreement of the atomic models to crystallographic data is characterized by R factors of 21.1, 21.6 and 21.9%, respectively, for Form1, Form3, and Form4 for all observed data having Fob greater than 1σ (Fob) between 6 and 2.5A resolution.
The final model consists of 4137 non-hydrogen atoms with individual isotropic temperature factors. The model also includes ApUp in the active site cleft of the catalytic (C) domain, but no solvent atoms. There are poorly-defined regions in the electron density maps where main chain densities for residues 170-172, 190-195, 389-390, 500-503, are not well defined. Residues 190-195 are part of the protease-sensitive region of the first disulfide loop, where nicking occurs; this region may be intrinsically flexible. So may be the loop between the transmembrane (T) and R domains, which includes residues 389-390. Aspects of data collection, phase determination and refinement are presented below and in Table 1.
Data Collection, Phase Determination and Refinement
Statistics for X-ray data collection, phase determination and refinement. Crystal Forms, 1, 3 and 4 were used for the current study (Fuji, supra) . Diffraction data were collected on a Rigaku AFC-6 diffractometer operating at 8.5 kW, equipped with a two-panel area detector of Xuong-Hamlin design (San Diego Multiwire Systems). Images were recorded as 0.1° oscillation frames, integrated and merged into batches of 50 frames (5°). Integrated intensities were scaled and merged for FOURIER scaling method (Weissman, Thesis , Univ. California, LA, 1979). Form4 native and derivative data were later collected to 2.5 °A with a RAXIS imaging plate system.
Heavy atom derivatives
KOS: K2OsO4, soaked for 3 days at the
concentration saturated in artificial mother liquor (12% PEG8000, 0.43M NaCl, 43mM Tris-HCl, pH 7.8); CNP: 4-chloro-2-nitro-mercury phenol, soaked for 5 days at the concentration saturated in artificial mother liquor; KNP: 1 to 1 mixture of KOS and CNP; CAP, trans-dichlorodiamine Platinum (II), soaked for 3 days at 2 mg/ml in artificial mother liquor; KAP: 1 to 1 mixture of KOS and CAP; GCL, HgCl2, soaked for 3 days at 2 mg/ml in artificial mother liquor. Heavy Atom Parameters
Heavy atom parameters were refined and MIR phases calculated using the program HEAVY (Terwilliger et al., Acta Cryst. A43:1, 1987). We initially obtained the Os derivative for Form3 crystals. From electron density maps based on the single isomorphous replacement (SIR) phases after solvent flattening at 3.5A resolution, the shape of the molecule was interpreted to have three domains. However, secondary structures were not easily interpretable and the course of the polypeptide chain was difficult to determine. A search for additional heavy atom derivatives was hampered by the lack of good quality crystals of Form3. We, therefore, shifted our efforts to Form4 crystals. MIR phases for Form4 were obtained from six heavy atom derivatives using isomorphous differences and anomalous differences. The Os and Pt derivatives were solved by isomorphous difference Patterson
functions, and the Hg derivative by a difference Fourier synthesis. Os derivatives of Form4 and Form3 have the same single site binding.
Solvent Flattening
Initial electron density maps of Form4 were calculated at 3.0°A resolution, with phases modified using an iterative solvent flattening procedure (Wang, Methods in Enzymol . 115:90, 1985) including phases extended to 3. oA from 3.2A by the Wang phase extension algorithm (Wang, supra) . A solvent volume of 45% was used to ensure that all protein density was included in the protein mask, somewhat smaller than the 57% estimated from the molecular weight. From these maps, all
secondary structures were identified and an initial model was built using a polyalanine chain. Model Building
Model building was expedited with the program FRODO (Jones Methods in Enzymol . 115:157, 1985) and the fragment-fitting routines of the program O (Jones, Acta Cryst . A47:110, 1991). Starting with α carbon
coordinates that were manually built, main chain atoms were added using the database of 34 well-refined protein structures. Then side chains were added using the rotamer database (Ponder, J. Mol . Biol . 193:775, 1987). Refinement
This initial model was adjusted by visual inspection of density maps before it was refined by the simulated annealing protocol of the program XPLOR
(Brϋnger et al. Acta Cryst . A46:585, 1990). The relative orientations of Diphtheria toxin in Forms 1, 3, and 4 were determined by a Patterson-space rotation and
translation search of the refined Form4 model against Form1 and Form3 data. Two top solutions (9σ) for Form1 data correspond to two Diphtheria toxin chains related by a noncrystallographic symmetry in asymmetric unit. The transformation from Form4 to Form1 is essentially a change of coordinate system from C2 to P1, where the crystallographic rotation axis of C2 becomes a
noncrystallographic rotation symmetry axis of P1 that is nearly parallel with (110) axis of P1. One top solution (7σ) for Form3 corresponds to a rotation of less than 0.5° in any direction. The transformation from Form4 to Form3 is essentially a 5Å translation along the a axis. This result is consistent with the observation that the average absolute difference of the amplitudes of
structure factors of Okl reflections between Form3 and Form4 is 15%, whereas those between hO1 or hkO
reflections between Form3 and Form4 are almost random (R=48%). Also, when the model was superimposed on the solvent-flattened electron density maps of Form3 based on the SIR phases, most of the secondary structures were recognized with the model as a guide. Real-space
averaging of densities between Form4 and Form3 with MIR and SIR phases at 3.0Å improved the density maps at this stage. Subsequently, experimental phases were extended to 2.5Å by the algorithm based on solvent flattening, histogram matching, and Sayre's equation (Brunger, supra) for Form3 and Form4. Form3 maps at 2.5Å were again skewed and averaged with Form4 maps. These were the most interpretable maps. Refinement of the atomic model was carried out independently for Form1, Form3, and Form4 with all observed data having Fob greater than 1σ(Fob) between 6 and 2.5A.
Figure imgf000031_0001
Figure imgf000032_0001
Catalytic domain
The C domain is formed from two β sheet subdomains, which subtend the active site cleft (Fig. 5). These β sheets are oriented roughly perpendicular to each other and form the core of the domain. One subdomain consists of β strands CB2, CB4, and CB8, surrounded by α-helices, CH2, CH3, CH6, and CH7. The other subdomain consists of β strands CB1, CB3, CB5, CB6, and CB7 surrounded by helices, CH1 CH4, and CH5. The two subdomains are connected by extended loops, CL1 through CL4, which link the two subdomains. These four loops appear to suggest potential for flexibility or even extension to a longer and narrower shape. Conceivably the C domain can assume this partially unfolded structure during membrane translocation.
The active site cleft of the C domain, identified by the binding of the dinucleotide ApUp, is formed primarily by β strands, CB2, CB3, CH3, CB7 and the loop, CL2 and is also bounded by β strand RB6 of the R domain. Located within the active site cleft are the following residues: Glu148 which is believed to play a key role in catalysis (Carroll et al., Proc . Natl . Acad . Sci USA 81:3307, 1984), His21 (Papini et al., J. Biol . Chem . 264:12385, 1989) and Tyr65 (Papini et al., J. Biol . Chem . 266:2494, 1991), both of which have been implicated in NAD+ binding, and various other residues suggested to be at or near the active site (Gly52 (Carroll, supra and
Giannini et al., Nuc . Acid Res . 12:4063, 1984), Trp50 (Collins et al., Biochim . Biophys . Acta 828:138, 1985), Lys474 (Proia, J. Biol . Chem . 255:12025,
1980). Least squares superposition of the α carbon coordinates of the C domains of Diphtheria toxin and ETA yields an r.m.s. difference of 1.44Å between 85 residues (16-33, 34-38, 49-66, 75-90, 91-96, 131-136, 147-164 of Diphtheria toxin and 437-452, 454-458, 465-482, 493-508, 511-516, 540-545, 552-569 of ETA).
The approximate position of the substance NAD+ in the active site can be inferred, because the dinucleotide, ApUp, binds competitively with NAD+. The high affinity of ApUp (~0.3nM as compared with ~8-16μN for NAD+; (Carroll et al., Biochem . 25:2425, 1986) may be a consequence of multiple contacts with the C domain and of salt bridges between the 3'-terminal phosphate of ApUp and the side chains of
Thr42 and Arg458, the latter of which is a residue of the R domain. Although the structure of bound ApUp resembles that of NAD+ and ApUp to make difficult the prediction of the conformation of NAD+ in the cleft. However, assuming that the adenine phosphate portion of NAD+ binds in the same conformation as that of ApUp, the nicotinamide ring will be
positioned close to the site of the uridine ring. This places the nicotinamide ring adjacent to side chains of His21, Tyr65, and Glu148.
Domain junctions
One of the two intramolecular disulfide bonds of Diphtheria toxin bridges a handle-like loop TL1 on the molecular surface (Fig. 1). This 14 residue loop (187-200) connects Fragment A to Fragment B; it is rich in Arg and known to be easily nicked by proteases (Moskaug et al., J. Biol . Chem. 264:15709, 1989; Collier, J. Biol . Chem . 246:1496, 1971). Once this loop is nicked, Fragment A and Fragment B are covalently linked only by the disulfide bond. There is evidence that nicking plays a role in the
cytotoxic action of Diphtheria toxin (4), and it is generally believed that nicked Diphtheria toxin separates into free Fragment A and Fragment B when this disulfide bond is exposed to the reducing environment of the cytoplasm during membrane
translocation of the toxin. The second disulfide bond makes a 9 residue loop between residues 461 and 471 within Fragment B. Residues near this loop (456, 458, 460, 472, 474) are also rich in positive charges and face the active site cleft, probably forming the so-called phosphate-binding or P-site (Lory et al., Proc. Natl . Acad. Sci . USA 77:267, 1980).
The structure suggests why whole Diphtheria toxin is inactive in catalyzing the ADP-ribosylation of EF-2 until the C domain dissociates, in the form of Fragment A, from Fragment B. As shown in Fig. 2, the active site is shielded by the 18-residue loop CL2 and the R domain. Thus, in whole Diphtheria toxin, the approach of EF-2 (Mr~100K) to the active site is blocked. The active site of whole Diphtheria toxin remains accessible to NAD+, however and catalyzes the NAD-glycohydrolysis (a slow side reaction that is probably physiologically
insignificant). The lack of secondary structural elements within loop CL2 may allow a substantial movement of main chain atoms of the loop, permitting substrate entry to the active site.
Transmembrane domain
The structure of the T domain exhibits two features that suggest how it might experience pH-triggered insertion into the membrane. The first is that the T domain is entirely α-helical, similar to the known and proposed transmembrane proteins, and that some of the helices have hydrophobic
characteristics more typical of transmembrane helices than of globular proteins (Rees, Science 245:510, 1989). The nine helices are arranged more or less in three layers, each layer consisting of an antiparallel pair of helices. The two long, carboxy terminal helices, TH8 and TH9, are unusually apolar and constitute the central core layer. One flanking layer, made up of helices TH5-TH7, also contains hydrophobic helices, TH6 and TH7. The other layer made up of helices TH1-TH3, is, in contrast, very hydrophilic even compared to globular proteins. The second noteworthy feature of the T domain is the acidic composition of the loops that connect pairs of these helices. Both loop TL3 between helices TH5 and TH6, and loop TL5 between hydrophobic helices TH8 and TH9 contain a total of six Asp and Glu residues (Fig. 6). At neutral pH, these loops are highly charged and water soluble. But at acidic pH, these residues would be at least partially
protonated, and hence more nearly neutral and membrane-soluble, especially near the surface of the membrane that has even higher concentration of protons due to the surface potential (McLaughlin, Curr. Topics Memb . Transport 9:71, 1977). Thus, the lower pH inside the endosome would tend to render these tip-shaped loops into membrane-soluble
"daggers" that would lead the two apolar helix pairs into the membrane.
Other structural characteristics of the T domain suggest that it has the capacity to insert into the membrane and can assist the translocation of the C domain. The first is that the nearly parallel packing of the three helix layers would permit spreading on the membrane surface of the first helix layer (TH1-TH3) if other layers were inserted. This insertion would require local conformational changes in loops, but no alteration of the helices themselves. Also the pronounced hydrophobic asymmetry is compatible with the
proposed rearrangement: 15 of 16 Lys and Arg
residues and all 6 His residues of the T domain are located on the opposite side from the "dagger" tips (Fig. 8), making the whole domain a hydrophobic dipole, once the Asp and Glu residues are
neutralized. It is possible that the hairpin loop TL5 and probably TL3 cross the membrane, where the Asp and Glu residues will once again be charged in the neutral pH of the cytoplasm.
The Diphtheria toxin dimer
Two monomers associate tightly to form a dimer with an interface between RB1/RB2 of one Diphtheria toxin molecule and RB2/RB1 of the other Diphtheria toxin molecule related by 2-fold rotation symmetry (Fig. 4). This interface is one of three major protein-protein contacts in crystal packing and involves 3 hydrogen bonds per monomer. These hydrogen bonds are well defined since they are formed between main chain N and C atoms of RB1 and RB2. The other interfaces are not common among three different crystal forms. The inability of the dimer to bind to the Diphtheria toxin receptor
(Carroll et al., Biochem . 25:2425, 1986) suggests that the dimer interaction sterically blocks the receptor binding domains of each monomer from the receptors on the surface of a target cell. The conformational differences between the monomer within the dimer and the native monomeric Diphtheria toxin remain uncertain, but biochemical evidence suggests they are not large. Binding data show that the affinity constant of the dimer for ApUp is the same as that of the monomer, and that the dimer binds 2 ApUp's (Carroll, supra) . In addition, comparable specific activities of NAD-glucohydrolase activity and affinities for NAD+ were found in the monomer and dimer. Further, the specific ADP-ribosyltransferase activity of Fragment A released from the dimer after reduction was the same as that from the monomer (Carroll, supra) . These findings show that the conformations of the C domain, and of the portion of the R domain interfacing the C domain, are relatively unperturbed in the dimer.
Use
The chimeric molecules of the invention can be used, for example, to kill particular classes of cells. As one example a chimeric diphtheria toxin molecule which binds specifically to cells bearing the interleukin-2 receptor can be used in treatment of various autoimmune diseases, e.g., arthritis. Non-cytotoxic hybrid molecules in which a chimeric diphtheria toxin molecule is linked to a second molecule can be used to introduce the second
molecule into selected cells, e.g., to correct an enzyme deficiency caused by a genetic disease. For example Tay-Sachs may be treated by introducing hexosaminidase A into appropriate cells.
Other Embodiments
The diphtheria toxin catalytic domain of chimeric diphtheria toxin molecules can be replaced by the catalytic domain of other toxin molecules to generate other targeted cytotoxins. Peptide toxins are preferred, but others are also useful. Many peptide toxins have a generalized eukaryotic
receptor binding domain; in these instances the toxin must be modified to prevent intoxication of non-targeted cells. Any such modifications must be made in a manner which preserves the cytotoxic functions of the molecule. Potentially useful toxins include, but are not limited to: cholera toxin, ricin, O-Shiga-like toxin (SLT-I, SLT-II, SLT IIV), LT toxin, C3 toxin, Shiga toxin, pertussis toxin, tetanus toxin, Pseudomonas exotoxin, alorin, saporin, modeccin, and gelanin.
If the catalytic domain is to be removed for the purpose of creating a hybrid molecule which includes a molecule to be introduced into a selected class of cells, it is preferred that the molecule to be introduced be fused to the chimeric diphtheria toxin molecule just to the amino-terminal side of the junction of Fragment A and Fragment B, i.e., to the amino-terminal side of residue 190.
What is claimed is:
Figure imgf000039_0001
Figure imgf000040_0001
Figure imgf000041_0001
Figure imgf000042_0001
Figure imgf000043_0001
Figure imgf000044_0001
Figure imgf000045_0001
Figure imgf000046_0001
Figure imgf000047_0001
Figure imgf000048_0001
Figure imgf000049_0001
Figure imgf000050_0001
Figure imgf000051_0001
Figure imgf000052_0001
Figure imgf000053_0001
Figure imgf000054_0001
Figure imgf000055_0001
Figure imgf000056_0001
Figure imgf000057_0001
Figure imgf000058_0001
Figure imgf000059_0001
Figure imgf000060_0001
Figure imgf000061_0001
Figure imgf000062_0001
Figure imgf000063_0001
Figure imgf000064_0001
Figure imgf000065_0001
Figure imgf000066_0001
Figure imgf000067_0001
Figure imgf000068_0001
Figure imgf000069_0001
Figure imgf000070_0001
Figure imgf000071_0001
Figure imgf000072_0001
Figure imgf000073_0001
Figure imgf000074_0001
Figure imgf000075_0001
Figure imgf000076_0001
Figure imgf000077_0001
Figure imgf000078_0001
Figure imgf000079_0001
Figure imgf000080_0001
Figure imgf000081_0001
Figure imgf000082_0001
Figure imgf000083_0001
Figure imgf000084_0001
Figure imgf000085_0001
Figure imgf000086_0001
Figure imgf000087_0001
Figure imgf000088_0001
Figure imgf000089_0001
Figure imgf000090_0001
Figure imgf000091_0001
Figure imgf000092_0001
Figure imgf000093_0001
Figure imgf000094_0001
Figure imgf000095_0001
Figure imgf000096_0001
Figure imgf000097_0001
Figure imgf000098_0001
Figure imgf000099_0001
Figure imgf000100_0001
Figure imgf000101_0001

Claims

Claims
1. A chimeric diphtheria toxin molecule wherein all or part of a CDR of an antibody is inserted into a loop region of the diphtheria toxin receptor binding-domain.
2. The molecule of claim 1, said antibody being capable of specifically binding a cell surface antigen expressed on a cell, said chimeric
diphtheria toxin molecule being capable of
specifically binding said cell surface antigen and being substantially incapable of binding to the diphtheria toxin receptor.
3. The molecule of claim 1 wherein all or part of a first said CDR is inserted into a first said loop region, all or part of a second said CDR is inserted into a second said loop region, and said first and second CDR are of a single antibody chain.
4. The molecule of claim 2 said molecule being capable of decreasing the viability of said cell.
5. The molecule of claim 4 wherein said molecule kills said cell.
6. The molecule of claim 1 wherein said loop region is RL3 and said CDR is a CDR1.
7. The molecule of claim 1 wherein said loop region is RL9 and said CDR is a CDR3.
8. The molecule of claim 2 wherein said molecule lacks diphtheria toxin catalytic activity.
9. The molecule of claim 8 wherein said molecule lacks all or part of the catalytic domain of diphtheria toxin.
10. A hybrid molecule comprising a first and a second portion joined together covalently, said first portion comprising the molecule of claim 8 and said second portion comprising a molecule to be delivered to said cell.
11. The hybrid molecule of claim 10 wherein said molecule to be delivered to said cell is a protein.
12. The hybrid molecule of claim 11 wherein said protein is an enzyme.
13. The hybrid molecule of claim 11 wherein said protein modulates transcription.
14. The hybrid molecule of claim 11 wherein said protein is a nucleic acid-binding protein.
15. The hybrid molecule cf claim 14 wherein said nucleic acid binding protein is capable of binding a single-stranded nucleic acid.
16. The hybrid molecule of claim 10 wherein said molecule to be delivered to said cell is a nucleic acid.
17. The molecule of claim 2 wherein the amino-terminus of a second diphtheria toxin receptor-binding domain is connected to the carboxy-terminus of said first diphtheria toxin receptor-binding domain, said diphtheria toxin second receptor-binding domain being substantially incapable of binding to the diphtheria toxin receptor.
18. The molecule of claim 17 wherein said carboxy terminus is connected to said amino terminus through a polypeptide chain.
19. The molecule of claim 17 wherein all or part of a CDR of an antibody is inserted into a loop region of said second diphtheria toxin receptor-binding domain.
20. The molecule of claim 19 wherein all or part of a first said CDR is inserted into a first said loop region of said first diphtheria toxin receptor-binding domain and all or part of a second said CDR is inserted into a second said loop region of said first diphtheria toxin receptor-binding domain.
21. The molecule of claim 20 wherein all or part of a third said CDR is inserted into a first loop region of said second diphtheria toxin
receptor-binding domain and all or part of a fourth said CDR is inserted into a second said loop region of said second receptor-binding domain.
22. The molecule of claim 21 wherein said first and said second CDR are of a first antibody chain and said third and said fourth said CDR are of a second antibody chain.
23. The molecule of claim 22 wherein said first and said second antibody chains are from antibodies recognizing the same antigen.
24. The molecule of claim 24 wherein said first and second antibody chains are from the same antibody molecule.
25. The molecule of claim 23 wherein said first antibody chain is the light chain of an antibody and said second antibody chain is the heavy chain of said same antibody.
26. A chimeric diphtheria toxin molecule wherein all or part of a CDR-like sequence of a Iigand binding protein having an antibody variable domain-like Iigand binding-domain is inserted into a loop region of the receptor binding domain of diphtheria toxin.
27. The chimeric diphtheria toxin molecule of claim 26 wherein said Iigand binding protein having an antibody variable domain-like Iigand binding-domain is tumor necrosis factor.
PCT/US1993/004335 1992-05-07 1993-05-07 Novel diphtheria toxin-based molecules WO1993022450A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US87963092A 1992-05-07 1992-05-07
US07/879,630 1992-05-07

Publications (1)

Publication Number Publication Date
WO1993022450A1 true WO1993022450A1 (en) 1993-11-11

Family

ID=25374540

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US1993/004335 WO1993022450A1 (en) 1992-05-07 1993-05-07 Novel diphtheria toxin-based molecules

Country Status (2)

Country Link
AU (1) AU4238893A (en)
WO (1) WO1993022450A1 (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1998021344A1 (en) * 1996-11-12 1998-05-22 Michigan State University Chimeric ltb vaccines
WO1998023731A3 (en) * 1996-11-27 1998-11-12 Univ Catholique Louvain Chimeric target molecules having a regulatable activity
US6027921A (en) * 1995-06-06 2000-02-22 Transkaryotic Therapies, Inc. Chimeric proteins for use in transport of a selected substance into cells and DNA encoding chimeric proteins

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4946778A (en) * 1987-09-21 1990-08-07 Genex Corporation Single polypeptide chain binding molecules

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4946778A (en) * 1987-09-21 1990-08-07 Genex Corporation Single polypeptide chain binding molecules

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
SCIENCE, Volume 238, issued 20 November 1987, VITETTA et al., "Redesigning Nature's Poisons to Create Anti-Tumor Reagents", pages 1098-1104. *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6027921A (en) * 1995-06-06 2000-02-22 Transkaryotic Therapies, Inc. Chimeric proteins for use in transport of a selected substance into cells and DNA encoding chimeric proteins
US6262026B1 (en) 1995-06-06 2001-07-17 Transkaryotic Therapies, Inc. Chimeric proteins for use in transport of a selected substance into cells
US6858578B2 (en) 1995-06-06 2005-02-22 Transkaryotic Therapies, Inc. Chimeric proteins for use in transport of a selected substance into cells
WO1998021344A1 (en) * 1996-11-12 1998-05-22 Michigan State University Chimeric ltb vaccines
US5993820A (en) * 1996-11-12 1999-11-30 Michigan State University Chimeric LTB vaccines
WO1998023731A3 (en) * 1996-11-27 1998-11-12 Univ Catholique Louvain Chimeric target molecules having a regulatable activity

Also Published As

Publication number Publication date
AU4238893A (en) 1993-11-29

Similar Documents

Publication Publication Date Title
US11291720B2 (en) Anti-CTLA4 monoclonal antibody or its antigen binding fragments, pharmaceutical compositions and uses
Choe et al. The crystal structure of diphtheria toxin
CN1054609C (en) Monoclonal antibody of human combining mosaic for recognizing epidermal growth factor receptor (EGF-R) and its diagnosis and therapeutical use
CA2103887C (en) Methods and materials for preparation of modified antibody variable domains and therapeutic uses thereof
AU696286B2 (en) Recombinant humanized anti-lewis Y antibodies
US5091513A (en) Biosynthetic antibody binding sites
EP0368684B1 (en) Cloning immunoglobulin variable domain sequences.
US5843711A (en) Diphtheria toxin receptor-binding region
US6207804B1 (en) Genetically engineered antibody analogues and fusion proteins thereof
EP1415002B1 (en) Focussed antibody technology
US5132405A (en) Biosynthetic antibody binding sites
JPH06508267A (en) immunoglobulin variants
CN112166121B (en) Monoclonal antibody of nerve growth factor, and coding gene and application thereof
HUT67796A (en) Specific binding antibodies
CN102395604B (en) There is the humanized antibody of anti-tumor activity
CN111718415A (en) anti-TIGIT nano antibody and application thereof
CN112480250B (en) Anti-human osteopontin antibody and application thereof
WO1993022450A1 (en) Novel diphtheria toxin-based molecules
EP0537489B1 (en) Monoclonal antibodies against tumor-associated antigens, method for their preparation and their use
CN114380917B (en) Bispecific single domain antibodies against IL-17A and TNF α and uses thereof
CN114920838B (en) anti-IL-17A single domain antibody and application thereof
CN115160434B (en) Humanized single domain antibody, application and medicine thereof
WO2022026775A1 (en) Compositions and methods for targeting coronavirus
CN116836268A (en) Antibodies against novel coronavirus SARS-CoV-2 and uses thereof
CN114716553A (en) Nano antibody targeting human LILRB4 and application thereof

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AT AU BB BG BR CA CH CZ DE DK ES FI GB HU JP KP KR LK LU MG MN MW NL NO NZ PL PT RO RU SD SE SK UA VN

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): AT BE CH DE DK ES FR GB GR IE IT LU MC NL PT SE BF BJ CF CG CI CM GA GN ML MR NE SN TD TG

121 Ep: the epo has been informed by wipo that ep was designated in this application
DFPE Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101)
REG Reference to national code

Ref country code: DE

Ref legal event code: 8642

122 Ep: pct application non-entry in european phase
NENP Non-entry into the national phase

Ref country code: CA