CA2083487A1 - Chimeric toxins with improved inter-domain geometry - Google Patents

Abstract

A chimeric toxin including protein fragments joined together by peptide bonds, the chimeric toxin comprising, in sequential order, beginning at the amino terminal end of the chimeric toxin: (a) the enzymatically active Fragment A of diphtheria toxin; (b) a first fragment including the cleavage domain 11 adjacent Fragment A of diphtheria toxin; (c) a second fragment comprising at least a portion of the hydrophobic transmembrane region of Fragment B
of diphtheria toxin, the second fragment having a deletion of at least 50 diphtheria toxin amino acid residues, the deletion being C-terminal to the portion of the transmembrane region, and the second fragment not including domain 12; (d) a spacer; (e) a portion of a cell-specific polypeptide ligand, the cell-specific polypeptide ligand being a cell growth factor, the portion including at least a portion of the binding domain of the polypeptide ligand, the portion of the binding domain being effective to cause the chimeric toxin to bind selectively to the target cell.

Description

3~$7 W091/19745- 1 - . PCT/US91~04187 CHIMERIC TOXINS WITH IMPROVED INTER-DOMAIN GEOMETRY
Backqround of the Invention This invention relates to the use of recombinant DNA techniques to construct chimeric toxin molecules.
The highly selective effects asserted by many hormones, toxins, and oth~ biologically active proteins are in part possible because such proteins possess more than one functionally distinct polypeptide domain. Some plant and bacterial toxins, e.g., have evolved with separate domains responsible for cell binding, membrane translocation, and intoxication. The combination of properties conferred by the various domains results in extremely potent bioactive molecules.
The diphtheria toxin (DT) is an example of a lS naturally occurring multi-domain protein. DT consists of a number of domain~, each of which confers a particular function, and all of which, in combination, result in an extraordinarily active toxin molecule. DT
can be characterized, starting at the amino terminal end of the molocule, as follows: a hydrophobic leader ~ignal equence ~ ~amino acids Val_2s - Ala_l);
n~ymat~cally-activ- Fragment A tamino acids Glyl -A~gl93) whlch tnclud-~ a domain which catalyzes the th- nicotinamide adenine dinuclootide-depondont ad no-ln dipho~phate (ADP) ribosylation of tho u~iryotic protein synthe~is factor termed "Elongation Factor 2"; the ~rotease-sen~itive disulf~de loop 1 ~amino acld8 Cysl86 - Cys201), wh~ch contains a cl-a~age domain; and Fragment B ~amino acids 8erl9~ -~0 8-r535), which includes a translocation domain and a g n rali~ed binding domain flanking a ~econd disulfide . .

~ ~3~7 - 2 - PcT/us91/o4~8?
loop (12, amino acids Cys461 - Cys471).
process by which DT intoxicates sensitive eukaryotic cells involves at least the following steps: (1) the binding domain of diphtheria toxin binds to specific receptors on the surface of a sensitive cell; (ii) while bound to its receptor, the toxin molecule is internalized into an endocytotic vesicle; (iii) either prior to internalization, or within the endocytotic vesicle, the toxin molecule undergoes a proteolytic cleavage in the 11 cleavage domain between Fragments A
and B; (iv) as the pH of the endocytotic vesicle decreases to below 6, the toxin spontaneously inserts into the endosomal membrane; (v) once embedded in the membrane, the translocation domain of the toxin facilitates the delivery of Fragment A into the cytosol;
- (vi) the catalytic activity, which resides in a domain in Fragment A, causes the death of the intoxicated cell. The mechanism of cell killing by Pseudomonas:
exotoxin A, and possibly by certain other naturally-occurring toxins, is very similar.
me therapeutic potential of man-made toxin molecules containing various functional domains (whether polypeptide or nonpeptide) has been appreciated for many years. Paul Ehrlich, Ehrlich (1906) in Collected 2S 8tudl-s on Immunity 2:~2-~7, was the first to suggest th- oon~truotion of bifunctional molecules that combined a mol-cul- with an af~inlty for a speoiflc target ~a targeting or cell-binding domaln) with a molecule that ct~ a~ a cytotoxic agQnt (a cytotoxic domain).
~ince that time numerous biological and chemioal moieties have been coupled in a variety of way~
ln attempts to create molecules that exhibit a degree of ~-lecti~e, i.e., targeted, cytotoxicity. Early efforts w-r- typified by the con~ugation of non-peptide toxins , . , , .......... ., . , .- . - . .

.. .
' ' . . '. . ~ "' ~ . ''; ' ,` ' " ' , ;'`

. ' ' .: , . . .... ~ . -Z ~ 37 W O 91/19745 _ 3 _ PC~r/US91/04187 to solubilizing molecules or to antibodies. More recently, the advent of molecular genetics has allowed the engineering of chimeric genes that encode novel multi-domain chimeric toxin proteins. See, e.g., Murphy S U.S. Patent No. 4,675,382, hereby incorporated by refeeence, which teaches the construction of hybrid proteins that include the enzymatically active (toxin) domain of DT, the cleavage domain of DT, the tran~location domain of DT, and a cell-specific binding domain derived from a second protein. The hybrid proteins described in Murphy suPra combine DT toxin domain~, which confer toxicity, with a domain from a different protein, e.g., interleukin 2 (IL-2), which confers extremely selective cell binding properties.
lS Improvements in the intrinsic properties of the constituent components, i.e., the u~e of more highly-spec~fic cell-binding agents, e.g., monoclonal antibodies, and the use of toxins of increased potency, e.g., plant or bacterial toxins, have been the primary route~ to improved toxin conjugates. The way in which the cell binding and cell-killing entities of blfunc:io~al molecule are coupled has also received attention in attempts to improve the performance of the-e molecules. These efforts ha~e been directed to 2~ pr-~-rving the act~vitie~ of the primary component~, incr-a~ing ~olubllity, lncre~ing the ratio of cytotoxic ~g nt bound to o~ eclflc bindlng agent, or l~cr~ g the ea~e with which a reguired step in intoxloation, e.g., the cleavage of a toxin domain from ~0 th- r-~t of the molecule, is carried out.
Various spacer or l~n~er mQmbers have been plaoed ~etween the domains ~or nonp~ptide functional e~tlti-r) of con~ugate-toxins in efforts to realize the advantag ~ discu~sed above. Rowland et al. (1975) . ~ . . . ~ .
, ............. , - . ;, . .,..... ., : .

,, , . , . . : ~ ~.. . .
.. : , . . . . . .

- ` ' - ~:, :. ' ' ' - , Wo91/19745 2 ~ S~ '7 PCT/US91/W187 Nature 25s:487 reports that the use of polyglutamic acid (molecular weight=35,000) as a linker between chemical toxins and antibodies was preferable to direct linkage of the toxin to the antibody.
"By minimizing interference with the chemical structure of the Ig in this linkage step, a conjugate is produced with both a high concentration of drug and little loss of antibody activity . . . The concept of linking cytotoxic drugs to antibody through an inert intermediate carrier offers a wide scope for improved cancer chemotherapy in the future, with the possibility of using a variety of drugs, different carriers and antibody preparations of greater purity."
Monsigny et al. (1980) FE~S Letters ll9:l8l reports miaximizing the activity of a drug, e.g., daunorubicin con~ugated to a carrier, e.g., an antibody, by the insertion of a peptide containing spacer between the drug and the carrier. The spacer arm, 2-(l-thio-B-D-glycopyranosyl)-ethanoyl-L-arg-L-leu, ~a~
be cleaved by lysosomal but not by serum proteases.
~For technical reasons, and because the activity of a drug is partially or totallY lost when it is substituted or chem~cally modified, we devised a spacer arm such that the drug carrier con~ugate is stable in serum and can be specifically split by 2S lysosomal proteases leading to the free drug in~ide the target cells."
~rnon t al. (~9~2) ~mmunol Rev. 6Z:S re~orts the use of d-xtr n to llnk daunomycln to an antibody.
"Th- rea~on for employing this procedure was two-fold. Fir~t, it i8 exp4cted to re~ult in higher extent of _ drug bint1ng per antibody molecule, wh~ch ~hould lead to higher cytotoxic actlvity; econd, the use of a macromolecule as a spacer arm between th- drug and the carrier could prove advantageous in permitting higher ...
',, , . , . . - . ' ' , .: . ' -. ; .; . ~ :

, ~.. ..

:,. . . .. , ~ .: : .

. ,~ , ~ - .. , ...
~r . . . - .

~ 7 Wogl/19745 _ 5 _ PCT/US91/04187 exposure of the drug moiety on the conjugate surface with less steric hindrance and hence higher efficacy "
Truet et al (1982) Proc Natl Acad Sci USA, 79 626-629 reports coupling daunorubicin to succinylated bov~ne serum albumin (BSA) by a spacer arm one to four am~no acid residues in length The purpose of the spacer arm was to allow lysosomal cleavage between the drug and the ~SA molecule Neville et al (1989) JBC 264 14653-14661 reports that a cleavable cross-linker enhances potency of a DT-antigen conjugate three to ten fold The crosslinkers are cleavable at acid pH, and are thus cleaved in an acidic compartment The increased potency i8 believed to be due to an enhanced intracellular toxln-toxin receptor interaction which leads to increased translocation the con~ugate is thought to be ~terically hindered prior to, but not after, cleavage A nonpeptide crosslinker, bis-maleimidoethyoxy propane, was used Greenfield et al PCT/US85/00197 describes a toxin-antibody con~ugate wherein the tox~n and antibody are coupled by means of a ~pacer peptide "the ~nvent~on i~ designed to provide a toxin con~ate which has the 2~ a~pro~ri-te geometry ~or tran810cating th- cytotox~c fr~gment into the t~rget c-ll, th- c-~aclty to ret-in its blndlng ~r-gm nt ~r~or to ~uch trunrloc~tlon, and/or the ability to olublli~- the cytotoxic portion In-on a~poct of the invention, the spacer 1~ de~lgned ~o a~ to permit the cytotoxlc ~ortion of the molecule ready acc-~ to the c-ll membrane A~ the ~ of a typical antibody ~binding) fragm nt 1~ v ry much greater than that of mo~t cytotoxlc fragments, there i8 con~id rable teric hinterance of the . . . . . .
. . -, . , . . , , :
: . ~ . . - . . -.

:. . . . - . , - . -.
:..... ; - , . . . . .
,. . , ` ::: :. , `............ . -~ . . . ,::, .
- - .. . .
. . .

: . . ~ . .. . :- ... . .
.. . . . - - . ~ , . . ..... -.~ . . .. . . . . . . . .

WO9l/19745 - 6 - PCT/US91/WI87 2~ 7 '~~
access to the cell membrane by the cytotoxic portion imposed by the sheer -bulk of the antibody or antibody fragments In order to effect this translocation, the spacer needs to be sufficiently flexible to allow the A
S portion to reach the cell membrane, and sufficiently extended to permit it to have sufficient reach "
Summarv of the Invention In general, the invention features a chimeric toxin including protein fragments joined together by peptide bonds, including, in sequential order, beginning at the amino termi~al end of the chimeric toxin (a) the enzymatically active Fragment A of diphtheria toxin; (b) a fragment including the cleavage domain 11 adjacent Fragment A of diphtheria toxin; (c) a fragment includinq (i) at least a portion of the hydrophobic transmembrane region of Fragment B of diphtheria toxin, the fragment having a deletion of at least 50, preferably of at least 80, diphtheria toxin amino acid residues, the deletion being C-terminal to the portion of the transmembrane region, and the fragment not including domain 12~ or, (ii) a fragment including at least a portion of the hydrophobic transmembrane region of Fragment B of diphtheria toxin 2~ wher-in said Fragment B o diphtheria toxin does not lnclud- any diphther~ toxln 6eguences C-terminal to amlno acld r-rldue 386 of native diphtheria toxin; (d) a ~paa-r ~d-flned lnfra); and ~e) a portion of a c-ll-rp~cl~ic polypeptide ligand, the cell-specific poly~ ptlde ligand being a cell growth factor preferably a lym~ho~ine, g , interleukin 2 (IL-2), or interleukin ) The portlon of the cell-specific polypeptide ligand lnclud-~ at lea~t a portion of the binding domain of th polypeptide l~g~nd, the portion of the binding ...., .,, ~; ..
-: , : . ~ . . .

2~3~
WO91/19745 _ 7 _ PCT/US91/04187 domain being effective to cause the chimeric toxin to bind selectively to the target cell e.g., lymphocytes, e.g., T-cells or ~-cells bearing receptors for the ligand. Cell growth factor, as used herein, means a S protein that binds to a cell surface receptor found on a mammalian cell and causes proliferation of the cell.
Preferred cell growth factors are lymphokines, i.e., cell growth factors that bind to and stimulate the proliferation of lymphocytes.
In preferred embodiments DAB389 is the cytotoxic portion, i.e., (a), (b), and (c) above.
(DA~389 consists of me~hionine followed by residues 1-386 of native DT followed by residues 484 and 485 of native DT.) The construction of DAB389 is discussed IS in US8N 488,603, filed March 2, l990, hereby ~ncorporated by reference.
Preferred embodiments include those in which the spacer: is at least 5 amino acids in length, preferably 10-30 amino acids in length; when placed between the seguence of DT fragment DAB~85 (DAB485 consists of methionine followed by the first 484 amino acld residues of native DT) and amino acid residues 2-133 of IL-2, ha~ a Bnorm value of l.000 or greater, more preferably of 1.12~ or greater, and most preferably 2~ o~ 1.135 or greater; is ¢ompo-ed of at least 60~, and ~r-f-r~bly of ~t l-a~t 80~ of amino acids from the group o~ ly~lne, ~-rlne, glyclne, proline, aspartic acid, glutumlo acld, glutamin-, threonine, aspara~ine, or ~rglnlne; i~ at least 60%, and preferably at least 80~, ~0 ho~ologous to any of Ala-Pro-Thr-8er-8er-8er-Thr-~ys-Lys-Thr, hereinafter r-f-rred to a~ ~l-lO), Pro-Ly8-8er-Gly-mr-Gln-Gly, h reinafter referr-d to a~ ~l-7 Gly), P~ro-Thr-8er-8er-8er-Thr-Lys, (hereinafter referred to as , - . ~

. - , . . . ............................... .. .

.. .. . . . . .. .. . ....... . .

WO 91/19745 ~ 7 - 8 - PCI`/US91/04187_ ( 1-7 Lys), or multiples thereof (the number of multiples hereinafter referred to with a superscript, e g , ~ o)2, which indicates 2 tandem copies of the (1-10) subunit; results in an affinity of the chimeric toxin for the tarqet cells that is greater than the affinity of a second chimeric toxin for the target cells, the second chimeric toxin being identical to the chimeric toxin except that the second chimeric toxin lacks the spacer; or, results in the chimeric toxin exhibiting cytotoxicity for the target cells that is a least 2 times greater than the cytotoxicity exhibited by a -second chimeric toxin for the target cells, the second chimeric toxin being identical to the chimeric toxin except that the second chimeric toxin lacks the spacer Diphtheria toxin, or native diptheria toxin, as used herein, means the 535 amino acid residue mature form of diphtheria toxin protein secreted by CorYnebacterium diDhtheriae The seguence of an allele of the gene wh~ch encodes native diphtheria toxin can be found in Greenfield et al (1983) Proc Natl Acad Sci U8A 80 6853-6857, hereby incorporated by reference Enzymatlcally active Fragment A, as used herein, means umlno acid residues Gly 1 through ~rg 193 of native D~, or an enzymatically active derivative or analog of the 2S natural seguence Cleavage domain 11, as used herein, m an~ th- protea-e en~ltive domain within the region ~annlng Cy~ 186 and Cy- 201 of nitive DT Fragment B, ~ ur~d h-r-ln, m n~ th- region from 8er 19~ through 8-r 535 of natlve DT The hYdrophobic transmembrane ~0 r-glon, or hydrophobic domain, of Fragment B, as u~ed h r-ln, means the amino acid ~eguence bearing a ~tructural l~ilarity to the bilayer-sp~nning hellces of lnt-gral membrane proteins and located approximately at or d rlv d from amlno acid residue 3~6 through amino . ;.: .. . . ,,: . .. -.. . . : :

. : . , ,, : ,.. , . . .::
- - , . ~
, , ., -- . .. ~ . .- .:
.. . .

~S~ 7 WO91J19745 _ ~ _ PCT/~S91/04187 acid residue 371 of native diphtheria toxin Domain 12, as used herein, means the region spanning Cys 461 and Cys 471 of native DT The generalized eukaryotic binding site of Fragment B, as used herein, means a region with~n the C-terminal 50 amino acid residues of native DT responsible for binding DT to its native receptor on the surface of eukaryotic cells The generalized eukaryotic binding site of Fragment B is not included in the chimeric toxins of the invention A spacer, as used herein, is a polypeptide which possesses one or more of the following characteristics (1) when placed between the sequence of the diphtheria toxin fragment DAB485 and amino acid re~idue~ 2-133 of IL-2, it possesses an amino acid lS residue with a normalized B value (Bnorm) (as defined in Karplu~ et al (1985) Naturwissenschaften 72 212, hereby incorporated by reference) of 1 000 or ~reater, preferably of 1 125 or greater, and most preferably of 1 135 or greater; (2) at least 60% and preferably at least 80% of its ~mino acid residues are from the group of ly~ine, serine, glycine, proline, aspartic acid, glutamic acid, glutamine, threonine, a~paragine, or arglnlne; (3) it possesses the seguence, or multiples of th- 8eguence Ala-Pro-Thr-8er-~er-8er-Thr-Lys-Lys-Thr;
25 ~) lt 1~ ~t lea~t 60% and preferably at least 80%
homologou~ to the ~-guence, or to multiples of the ~ gu-nc~ t-d ln ~3) (S) lt possesse~ the sequence, or multlple8 of the sequence Pro-Ly~-8er-~ly-Thr-41n-Gly; ~6) it is at least 60% and ~0 ~r-f-rably at least 80% homologous to the seguence, or multl~ of the seguence, li8ted in ~S) ~7) lt ~0-~ 8 the seguence, or multiples of the seguence of ~ro-Thr-8-r-8er-8er-Thr-Lys; or (8) it 18 at least 60%
~d ~r-f-rably at least 80% homologou~ to the seguence, .. ...
.. .,.. .. . . . ~.
,;, ~ . ~ . , .. . . . . . -. . ~ .... . . . . . .. . .
: - : .' : .. ,. .. :

3~7 Wog1/19745 - 1o - PCT/US91/0418 or multiples of the sequence, listed in (7).
The Bnorm value of residues in a polypeptide sequence can be determined with computer programs, e.g., with FLEXPRO (Intelligenetics, Mountain View (CA)) which predicts the flex~bility between alpha carbon atoms at each point of a selected protein sequence, using the method (the Bnorm method) of Karplus et al., suPra.
FLEXPRO calculates the chain flexibitiy at a selected amino acid residue from the average values of the atomic temperature factors (also called B values or Debye-Waller factors) of the alpha carbon atoms in ad~acent amino acids. The B value is the mean sguare displacement of the atom from its average position in the protein. Flexible locations have high ~ values because their displacement can be large.
The Bnorm value for an amino acid in FLEXPRO
is affected by the Bnorm values of its neighboring amino acids. The predicted flexibility at an amino acid calculated by FLEXPRO is the weighted sum of the Bnorm value~ (taking account of neighbors) of the seven amino acids closest to that point in the sequence. The weight for the two outermo~t amino ac~ds is 0.25; for the two n-xt to them, 0.5; for the two ad~acent to the central amino acid, 0.75; and for central amino acit it~elf, l.
For aoh amino wld, the welght is multiplied by the Bnorm v-lU- ~t~klng account of neighbor~). When an uml~o acid -qu-nce i~ input to FLEXPRO, the program calculat-~ the ~norm values and displays 7-amino ac~d r-~ldue ~ections and the pea~ Bnorm value for each of ~0 th- 7-amino acid residue sections.
A normalizod B value of less than l indicates a rigid umino acid; ~ value greater than l indicate~ a fl-~ible amino acid. Observations in a number of prot-in~ ~uggest that the amino acids alanine, valine, .. . . . . .. .
' ' ' - ': '' . ' , - , , . ~. . .. .
.

.
. , .
, , ., :.
: - - . .. .. . . . .. .

3~ 3'7 Wo91/19745 - 11 - pcT/uss leucine, isoleucine, tyrosine, phenylalanine, tryptophan, cysteine, methionine, and histidine tend to be rigid and that the amino acids lysine, serine, glycine, proline, aspartic acid, glutamic acid glutamine, threonine, asparagine, and arginine tend to be flexible (Karplus et al suPra) The sequence used as a spacer may be derived from any source e g , from one of the polypeptides used to construct the chimeric toxin, from other naturally occuring sequences, or from synthetic sequences regardless of whether they are naturally occuring The invention also features a chimeric toxin encoded by a fused gene including regions coding for the protein ragments, a DNA sequence encoding the chimeric toxin, an expression vector containing the DNA sequence encoding the chimeric toxin, a cell transformed with the expre~ion vector, and a method of producing the chimeric toxin including culturing the transformed cell ~nd isolating said chlmeric toxin from the cultured cell or supernatant The invention also features spacer peptides having the seguence Ala-Pro-Thr-8er-Ser-Ser-~hr-Lys-Lys-Thr, or tandem repeats of that 10 residue subunit ~with or without one or two additional re-~due- at each end (or between the indi~idual ubunit~ of the tand~m repQat) that ari-e ~ro~ th- inolu~ion o~ function-l or disfunctional 1/2 r-~trlction n~yme r-cognltion site linker sequences in th- DNA that encodes the spacer), DNA segments encoding th- ~pac-r ~eptides ~with or without functional or dl-~unctional 1/2 re-triction enzyme recognition site lln~-r ~-quence~), vector~ containing these DNA
~-gm-nt~, c-lls trunsformed with the~e vectors, and m thod~ of ma~ing the ~pacor peptide including culturing :. -. . ......... . : . .
. : ,. ............... . - ... ..

.. . . , ., . - :: : : - .. . -- - .

W091/19745 2 C~ '7 PCTlUS91/0418 the transformed cell, and isolating the spacer from the cultured cell or supernatant Preferred embodiments include a DNA segment encoding the spacer wherein the spacer is a tandem repeat of the subunit Ala-Pro-Thr-Ser-Ser-Ser-Thr-Lys-Lys-Thr, and the DNA
~eguence encoding each occurence of the subunit in the spacer is nonhomologous with the DNA sequence encoding every other occurence of the subunit in the spacer or the chimeric toxin, the nonhomology being sufficient to prevent recombination between seguences encoding tandemly repeated subunits The invention also features a method of preventing recombination between the tandemly repeated lS subunits of spacer-peptide-encoding DNA by choosing the codons of each subunit-encoding sequence such that the DNA encoding each subunit is nonhomologous with the DNA
encoding e~ery other ~ubunit, the nonhomology being sufficient to prevent recombination between tandemly repeated subunits ~ he invention also features DNA encoding a spacer (with or without functional or disfunctional 1/2 rertriction enzyme recognition site linker sequences), an expression vector containing that DNA, a cell tran~formed with that expre~sion vector, and a method of ~roducing the ~ac-r includlnj culturing the transformed c-ll nd lrol~tinq ~id ~ac-r from the cu~tured cell or cU~ rn~t~nt Molecules of the invention exhibit improved ~0 blnding affinity and improv d cytotoxicity for cells b-arlAg th rece~tor to which the ligand portion of the ohim rlc toxin bind- In the management of autoimmune dlr-ar-, allograft re~ection, and other lym~hocyt~-dependent conditions, a chimeric toxin that . . .
. . . .' ,., ' . ,: ' . ' ' " .
: - ., - , .: :; , - - , . .: .
: ' ' . :. : , , - . .. : . - :
, . : . .
, . ., - .. ~ .
. - . . :-. ' :. ' - . , . - ' . :

. . ., : . . - ~ - . - -,, ~:
. ~. . . - - . . .

4~ - 13 - ~ ~ S3~ ~7. - PCT/US91/~187 employs a cell bindinq portion that recognizes an interleukin (or other growth factor~ receptor must compete with indigenous interleukin (or other growth factor) for sites on the target cell. Thus, optimizat~on of the early, cell-binding, step is partlcularly cr~tical in chimeric toxins in which the cell binding portion recognizes a ligand such as an interleukin receptor.
Other features and advantages of the invention will be apparent from the following description of the preferred embodiments and from the claims.
Description of the Preferred Embodiments The drawings will first be briefly described.
Drawinq~
lS Fig. 1 i5 a diagram of the DT molecule and various fu~ion proteins:
Fig. 2 i~ a depiction of the construction of the plasmids of a preferred embodiment;
Structure and 8Ynthe~is of chimeric toxins with imProved inter-domain qeometrY
DA~86-(1-10)-IL-2 i8 a chimeric toxin ~olypeptide con~ifiting of, in the following order: Met;
umino acid residues 1 through His~84 of mature native DT: ~ly; the amino acid residues Al~-Pro-Thr-8-r-8er-8er-Thr-Lys-~y~-Thr (indicated a8 ~1-10)); ~i~; Ala~85 of mature natlve DT; and ~mino aold r~ldu-~ 2 through 133 of S~-2. The DT portion of th ohim ric toxin DAB~8C-(1-10)- IL-2 includes all of DT fragm nt ~ and the portion of DT fragmont B extending - ~0 to r-~ldu- ~85 of mature nat~ve DT. See Fig. la for the ~t~uotur- of DT. Fig. lb ~hows the structure of DAB~6-1~--2. ln Fig. lb a wide bar indicate8 the fu~lon ~rotein, narrow connecting lines repre~ent d~l-tion~, numbers above the bars ~re amino acid residue . .

. ; : - : - , . -W091/19745 ~ 7 - 14 - PCT/USs1/04182 num~ers in the DAB nomenclature (described below), numbers below the bars correspond to the amino acid residu numbering of native DT, cross hatching indicated amphiF -hic regions, darkened areas correspond to the transr ~brane region, IL-2-2-133 indicates amino acid re81d 3 2-133 of IL-2, Ala ~ alanine, Asn = asparagine, A~p ~ ;partic acid, Cy8 - cysteine, Gly = glycine, His ~ hist;dine, Ile - isoleucine, Met = methionine, Thr =
threonine, Tyr = tyrosine, and Val = valine. (The nomenclature adopted for DT-IL-2 toxins is illustrated by DAB486-(1-10)-IL-2, where D indicates diphtheria toxin, A and B indicate wild type sequences for these DT
fragments, the number in the parenthesis represents a spacer polypeptide, and IL-2 indicates mature human lS interleukin-2 seguences. The numerical subscript indicates the number of DT-related amino acids in the fusion protein, the last of which is at the C-terminal end of the spacer where a spacer is inserted. Note that the last two codons of DT also function as a l/2 S~hI
site. 8ince the deletion of the tox signal sequence and expres~ion from the trc promoter re6ults in the addition of a methionine residue to the N-terminus, the numbering o DAB-IL-2 fu~ion toxins is ~l out of phase with that of native diphtheria toxin.) DAB~86-~1-lO~-rL-2 was ¢onstructed from DAJ~J6-IL-2. pDW2~, which c~rries DAB~86-IL-2 wa~
¢on~truct-d a~ follow~. pUCl8 (New England 8ioLabs) was dige-ted with P~tI and B9lI and the PstI-~glI fragment ¢arrying th~ E.coli origin of replication, the ~olylln~-r region, and the 3' portion of the B-lacatama~e gene ~ampr) was recovered. Plasmid pR~-233-2 ~Pharmacia) was digested w~th PstI and L~lI
and th- P8tI-B~lI fragment carrying, two transcr~ption t~r~nator~ and the 5' portion of the ~-lactamase gene ' , - ~ : :
- . ,., - .

W091/19745 - 15 - 2~ S~ 7 PCT/US91/~187 was recovered. pDW22 was constructed by ligating these two recovered fragments together.
pDW23 was constructed by isolating a BamHI-SaII
fragment encoding human IL-2 from plasmid pDW15 ~Williams et al. (1988) Nucleic Acids Res. 16:10453-10467) and ligating it to BamHI/SalI digested pDW22 ~described above).
pDW24 was constructed as follows. A BamHI-NcoI
fragment carrying the trc promoter and translational initiation codon (ATG) was isolated from plasmid pKX233-2 (Pharmacia). The DNA seguence encoding amino acid residues 1 through 485 of DT was obtained by digesting pABC508 (Williams et al. (1987) Protein Eng~neering 1:493_498) with SDhI and HaeII and recovering the HaeII-SPhI fragment containing the ~equence encoding amino acid residues 1 through 485 of DT. A NcoI/HaeII linker (5'CCATGGBCGC 3') was ligated to the HaeII-8~hI fragment and that construction wa6 then ligated to the previously isolated 8amHI-NcoI
fragment carrying the trc promoter. This results in a BamHI-80hI fragment bearing, in the following order, the trc promoter, the NcoI site (which supplies the ATG
lnltiator codon for Met), and the seguence encoding re~idue~ 1 through ~85 of native DT. This fragment was ln--rt d lnto DDW23 that had been dlge~ted wlth ~amHI
and a~l. Th- r-rulting pla-mid wa~ de-ignated pDW24.
~D~2~ 1J ~OWn ln Fig. 2. The inr-rt correspondlng to DAB~86-IL-2 18 ~hown a8 a heavy llne. In Fig. 2 flll d clrcl-~ indi¢ate NcoI sites, open circlos i~dicat- N~lI slt-~, open dlamond~ ~ndicate ClaI sltes, ~lll d gu~res lndioate HDalS sltes, open ~guare-lndi¢at- 8~hI slt-s, and fllled trlangles indicate 8alI
it--. The fusion protein ~DAB~86~ 2) encoded by ~DN2~ i~ expres~ed from the trc promoter and consists of , .
' ' '~ ~ ',, ' '. ' ' '~. , ' , ' ' .' , ' ' .
' ,' . ~,'. ,., ' ,`"'.,' .',' "'~' ', ,,"'.' '" " '' '' ' .

. ' : . ' ' . .: : . ' ' . . : ' . ' : ' ' . . . ' .

W091/19~45 ~ ~ 3 ~ S 7 - 16 - PCT/US91/W187,_ Met followed by amino acids l through 485 of mature DT
fused to amino acids 2 through 133 of human IL-2.
DNA encoding the polypeptide spacer, (l-lO), was synthesized and inserted into pDW24. pDW24 was cut S at the SphI site at the 3' end of the DT sequence and a synthetlc seqyence encoding a spacer (or multiples thereof) inserted. The sequence of the synthetic seguences encoding 1, 2, and 3 copies of the (l-lO) spacer are shown in Table l.

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

~ 91/1~45 - 17 - 2~ pcT/usgl/o4l~ ~

TABLE 1: SPACER SEQUENCES

1.Ala Pro Thr Ser Ser Ser Thr Lys Lys Thr His CT CCG ACC AGC TCT AGC ACT AAA AAG ACT CAT G
GTA CGA GGC TGG TCG AGA TCG TGA TTT. TTC TGA
S Functional SphI site Disfunctional SphI site 2. Gly Ala Pro Thr Ser Ser Ser Thr Lys Lys Thr-GT GCA CCG ACT AGC AGC TCT ACT AAG AAA ACA-GTA CCA CGT GGC TGA TCG TCG AGA TGA TTC TTT TGT-Di~functional SphI ~ite 11 12 13 1~ 15 16 17 18 19 20 . .
Ala Pro Thr 8er Ser Ser Thr Lys Lys Thr His lS GCT CCT ACC TCT TCT AGC ACG AAG AAG ACG CAT G
CGA GGA TGG AGA AGA TCG TGC TTC TTC TGC
Functional SphI site 20 3. Gly Ala Pro Thr Ser 8er Ser Thr Lys Ly~ Thr-~T GCA CCG ACT AGC AGC TCT ACT AAG AAA ACA-OTA CCA CGT GGC TGA TCG TCG AGA TGA TTC TTT TGT-Dl~u~ctlo~l 8~h1 lte .

2~ 11 12 13 1~ lS 16 17 18 19 20 21 22 Al~ Pro Thr 8er 8er 8er Thr Lys rys Thr H1~ Ala-oCT CCT ACC TCT TCT AGC ACG AAG AAG ACG CAT GCT-CCA G4A TG4 AGA AaA TCG TGC TTC TTC TGC GTA CGA-Functional 8phI site ~O
23 2~ 25 26 27 28 29 30 31 32 Pro Thr 8er 8~r 8er Thr ~ys Lys Thr ~i~
CCC ACC AGC TCT AGC ACT AAA AAG ACT CAT G
~GC TG~ TCG AGA TCG TGA TTT TTC TGA .
Disfunctional SphI site ;~f ~ 7 5 - 18 - PCT/US91/04187~
A 4-base extension capable of annealing to a l/2 SPhI site is present at either end of the DNA fragment encoding each spacer sequence In each spacer, codon substitutions destroy one of these SPhI recognition sites such that only one SPhI site is regenerated Refering to Table l, sequence l, the use of T in place of G in the third position of the 10th codon prevents the creation of an SPhI
~lte ~t the 3' end of the spacer In sequence 2, the use of G in place of C in the 2nd position of the Gly codon prevents the creation of an S~hI site at the spacer's S' end Sequence 3 is made by maintaining the functional SPhI site at the 5' end of sequence 1 and ligating the spacer of sequence 2 at that site, leaving a regenerated SPhI site 3' of the spacer of sequence 2, and disfunctional SPhI sites at either end of the spacer of sequence 3 Thus, insertion of sequence l from Table l into the SphI
15 site of DAB486-IL-2 or DAB389-IL-2 results in a fusion protein with splice/~unctions as follows the His48 residue of native DT; Ala-Pro-Thr-8er-8er-Ser-Thr-Lys-Lys-Thr; a His residue encoded by the 3' extension; Ala485 of native DT; and the second amino acid of the IL-2 sequence The insertion of sequence 2 from Table l into the SphI site of DAB~86-IL-2 or DA~389-IL-2 results in a fusion protein with rplice/~unctions as follows the His48~ residue of native DT; a Gly residue derived from the 3' extension at the 5' end of the oligonucleotide; Ala-Pro-Thr-8er-8er-8er-Thr-Lys-Lys-Thr-2~ Ala-Pro-Thr-8er-8er-8er-Thr-Ly8-Lys-Thr; a His re~idue encoded by th- 3' xt-n~ion; A1~85 of n~tive DT; and the second amino acit o~ th- IL-2 ~-~u-nc-~ h in~-rtion of seqUence 3 from Table l into the 8PhI site o~ D~8~86-IL-2 or DAB389-IL-2 results ln a fusion protein with ~0 rplic-/~unctlon~ as follow8 the His48~ re~idue of nttive DT; a Oly r--~due der~ved from the 3' extension at the 5' end of the oligonucl-otid-;
Al~-Pro-Thr-8-r-8er-8er-Thr-Lys-Lys-Thr-Ala-Pro-Thr-8er-8er-8er-Thr-Ly~-~y~-Thr-His-Ala-Pro-Thr-8er-8er-8er-Thr-Lys-Lys-Thr; a His .~ , . . . . -... , - .: . . . .
~ . . . .
.. . . . - :. .. . .
.. . . . . . . . .
.

... ... . . ..

.. .. . . . ..

WO 91/19745 -- 19 _ 2~?~37 PCI`/US9l/04187 residue encoded by the 3' extensioni Ala 85 of native DT; and the second amino acid of the IL-2 sequence. Three subunits may be added by ligating sequence 1 and 2 in vitro to achieve sequence 3, which i8 then inserted into the chimeric toxin, or, seguence 1 may be S inserted into the SphI site of a chimeric toxin that harbors 6equence 2.
Insertion of sequences 2 or 3 from Table 1, which encode 2 ~nd 3 copies o (1-10) respectively, result in analagous structures. When repeats of the (1-10) se~uence are present codons are chosen such that DNA homology between sequential (1-10) encoding 8eqUenCe8 i8 minimized. This prevents rearrangements that would arise from homolo~ous recombination between sequential copies of (1-10) encoding DNA .
The presence of the Gly and His residues (encoded by the lS linkers at each end of the spacer seguence) do not affect the Bnorm value of the linker (data not shown).
The (1-10) spacer geguence was identified by applying the FLEXPRO program (described above) to the seguence of DAB486-IL-2.
Table 2 shows the 10 most flexible 7-amino acid segments in DAB486-IL-2 as determined by FLEXPRO.
TABLE 2: FLEXPRO Analysis of DA84g6-IL-2 Rank From To Bnorm Sequence 1 ~87 ~93 1.13S Pro-Thr-Ser-8er-Ser-Thr-Lys 2 39 ~S 1.12S Pro-Lys-8er-Gly-Thr-Gln-Gly 3 1~2 1~8 1.107 Ala-Glu-Gly-8er-8er-Ser-Val S~l 587 1.107 Leu-~ys-Gly-8er-Glu-Thr-Thr S 172 178 1.102 Gly-Lys-Arg-Gly-Gln-Asp-Ala 6 7 13 1.098 Val-A~p-8er-8er-Lys-8er-Phe 7 232 238 1.097 8er-Glu-8er-Fro-Agn-Lys-Thr 8 63 7~ 1.036 Val-Asp-Asn-Glu-Asn-Pro-Leu ~ 9 ~11 417 1.08S Phe-Gln-Gly-Glu-Ser-Gly-His 266 272 1.08 Thr-Val-Thr-Gly-Thr-Asn-Pro The ~o~t flexible seven amlno acid stretch of the fusion ~rot-in wa~ found to be am~no acids, residues ~87 .- - . . . ~ . ~ . . . . . .
"'; ' ,. ' ' ,` ",` ' "', ., ,' ' '', . '' ' ' ' . "~ ' . ' " ~ . . --, , ' " ~ ' ,. .
, , . . ~ - :

,- - - : : - , .. ,-WO91/19745 ~ f S~ 7 - 20 - PCT/USsl/04187 through 493 of DA~486-IL-2, with a Bnorm 1.135. These seven amino acids correspond to amino acid residues 2-8 of IL-2. These seven amino acids were also the most flexible found in DAB389-IL-2 and DAB295-IL-2 (a molecule entirely devoid of the toxic ch~racteri~tics of the other DT chimeric toxins te~ted). 9ecause the first 10 amino acid residues of IL-2 are known not to orm an orderly array in crystals ~Brandhuber (1987) Science 238:1707-1709), and are thus flexible, the entire first 10 amino acid residues of IL-2 were used as the (1-10) spacer. When inserted into DA~486-IL-2, the (1-10) spacer is the most flexible seqyence of the new construction, as shown in Table 3.

2~

~O

WO 91/19745 - 21 - ~ ~ 3~ ~r ~7 PCr/US91/04187 TABLE 3:
A: FLEXPRO Analysis of DAB486-(1-1o)-IL-2 Rank From To B~Norm] Sequence S 1 487 493 1.135 Pro-Thr-Ser-Ser-Ser-Thr-Lys 2 498 504 1.135 Pro-Thr-Ser-Ser-Ser-Thr-Lys 3 39 4S 1.125 Pro-Lys-Ser-Gly-Thr-Gln-Gly ~ 142 148 1.107 Ala-Glu-Gly-Ser-Ser-Ser-Val 592 598 1.107 Leu-Lys-Gly-Ser-Glu-Thr-~hr 6 172 178 1.102 Gly-Lys-Arg-Gly-Gln-Asp-Ala 7 7 13 1.098 Val-Asp-Ser-Ser-Lys-Ser-Phe 1O 8 232 238 1.097 Ser-Glu-Ser-Pro-Asn-Lys-Thr 9 68 74 1.086 Val-Asp-Asn-Glu-Asn-Pro-Leu 411 417 1.085 Phe-Gln-Gly-Glu-Ser-Gly-His B: FLEXPRO Analysis of DAB486-(1-10)2-IL-2 RANK FROM TO BtNorm] SEQUEUOE
15 1 488 4941.135 Pro-Thr-Ser-Ser-Ser-Thr-Lys 2 ~98 50~1.135 Pro-Thr-Ser-Ser-Ser-Thr-Lys 3 509 51S1.135 Pro-Thr-Ser-Ser-Ser-Thr-Lys 39 ~5 1.125 Pro-Lys-Ser-Gly-Thr-Gln-Gly S 1~2 1481.107 Ala-Glu-Gly-Ser-Glu-Thr-Thr 6 603 6091.107 Leu-Lys-Gly-Ser-Glu-Thr-Thr 7 172 1781.102 Gly-Lys-Arg-Gly-Gln-Asp-Ala 20 8 7 13 1.098 Val-Asp-Ser-Ser-Lys-8er-Phe 9 232 2381.097 Ser-Glu-Ser-Pro-Asn-Lys-Thr 68 74 1.086 Val-Asp-A~n-Glu-Asn-Pro-Leu ~ =.

2~

` ' ,`',, ` "' . ~, ." '. ".'' '. ',' ..'. .',' . ;' `' ,'' ., .. - .: ':

wO9l/l9745 ~ 7 - 22 - PCT/US91~W187 DA~389-(1-10)-IL2 was constructed from DAB389-IL-2 DAB389-IL-2 was constructed by removing a 309 bp HpaII - SPhI restriction fragment from pDW24 and replacing it with oligonucleotide linker 261/274 (Table 4) to generate plasmid pDW27 (Fig 1) TABLE 4 Oligonucleotide Linkers con~truct oligonucleotide linker DAB389-IL-2 5~-CG-GGT-CAC-AAA-ACG-CAT-S' 1/2 HpaII-1/2SphI
CCA-GTG-TTT-TGC

S5 Thi~ linker restores fragment B sequerces from Pro383 to Thr387, and allows for in-frame fusion to IL-2 seguences at this position Thus, in DA9389-IL-2 the 97 amino acid~ between Thr387 and His~85 have been deleted DA9389-~1-10)-IL-2 was constructed by inserting DNA encoding the spacer (see Table 1) into pDW27 at the 8phI site, as de~cribed above DA~389-~1-10)-IL-2 may al~o be generated directly from ~DW2~ ~DA9~86-IL-2) by remo~al of the 309 bp H~aII-8phI fragment and replacing it with a linker that 2~ r-~tor-- fragm nt ~ ~eguences from Pro383 to Thr387, ncod - th ~oly~ ~tid- linker, and allows for the ln-~rum ~u~on to th IL-2 ~equence~
The -gue~ae of DT i8 given in Greenfield et ul ~1983) Proc Natl Acad 8ci ~8A 80 6853-68S7 The ~0 ~ qu-~ce enaoding IL-2 wa~ synthesized on an A~plied ~o-y~t-m- DNA-8ynthe-izer, a8 de8aribed in Williams et ~1 ~1988) Nucleic Acids Res 16 10~S3-10~67, hereby incorporat d by r-feronce The sequence of I~-2 i8 ~ound in Williams et al ~1988) Nucleic Acid~ Res - . . . ~ :

. . , , . ,, - :

~, . .

WO91/19745 - 23 ~ 35~7 PCT/US9~/04187 16:10453-10467. Fusion of the seguence encoding mature DT to ATG using an oligonucleotide linker is described in Bishai et al ( 1987) J. Bact 169 5140-5151 hereby incorporated by reference Chimeric toxins in which DT fragments were fused to murine interleukin 4 (IL-4) with or without the presence of a spacer were also constructed 486 2 3~9 DAB389-(1-10) -IL-4 were constructed by methods known to those skilled in the art Briefly DAB389-(1-10)2-IL4 was constructed by first digesting a plasmid containing the DAB389-(1-10)2-IL-2 fusion (derived from pDW27) with SphI and HindIII to remove the IL-2 coding sequence and lS then inserting a ~egment of DNA that encodes IL-4 The IL-4 encoding fragment includes linkers to allow ln~ertion and in-frame fusion to the 3 end of the spacer encoding DNA DAB486-IL-4 and DA~389-IL-4 were made by analogous treatment of pDW24 and pDW27 respectively The æequence of murine IL-4 may be found in Lee et al (1986) Proc Natl Acad Sci U8A
83 2061-2065 hereby incorporated by reference Thie -guenae u~ed in these construction~ was obtained from DN~X ~Californi~) Ollgonuoleotite8 and nuoleic acids were ~ynth-~lr-d and mani~ulat-d ~ follow~
Ollgonucl-otid-~ w r- ynthe-i~ed u~ing cyanoethyl ~iho~phor~midlt- chemi-try on an Applied Biosystems 380A
D~A ~ynth~ r (Ap~lied Biosystems Inc Foster City CA) Following ynthe~is oligonucleotides were purlfl-d by chromatography on Oligonucleotide Purification Cartridg~s ~Applied Bio~ystem8 ~nc Fo8ter City CA) a- dir-cted by the munufacturer Purified ~ligonualeotide~ were re~u~pended in IE buffer (10 mM

r ' ,.
. ' '` ,'', - . :" ," '............... ' ' ' - ' ~ ' ':
.~ . . . , ~: . , , , , ,:

: ., '; . . ' WO9l/19745 2 ~ ~ 7 - 24 - PCT/US91/0418 Tris base, 1 mM EDTA, pH 8.0). To anneal complementary strands, equimolar concentrations of each strand were mixed in the presence of 100 mM NaCl, heated to 90C for 10 min, and allowed to cool slowly to room temperature.
S Plasmid DNA was purified by the alkaline lysi6/cesium chloride gradient method of Ausubel et al.
~1989) Current Protocols in Molecular Biology, John Wiley & Sons, N.Y. D~A was digested with restriction endonucleases as recommended by the manufacturer (New England Biolabs, Beverly, MA and ~ethesda Research Laboratories, Gaithersburg, MD). Restriction fragments for plasmid construction were extracted from agarose-TBE
gels, ligated together (with or without oligonucleotide linkers) and used to transform E. coli using standard 1S methods. 8ee Ausubel et al (1989), su~ra, and Maniatis et al. (1982), Molecular Cloning Laboratory Manual, Cold 8pring Harbor Laboratory, Cold Spring Harbor, N.Y.
Plasmid DNA seguencing was performed according to the dideoxy chain termination method of Sanger et al. (1987) Proc. Natl. Acad. 8ci. U&A 74:5463-5467, as modified by Kraft et al. ~1988) Bio Technigues 6:544-5~7, using 8eguenase (United 8tates Biochemicals, Cleveland, OH).
ExDression ~nd Purification of Chimeric Toxins Expression and purification of chimeric toxins 2S wa- a~ follow~. All DT-related I8-2 fusion proteins u~-d h r-ln w r- expre--ed in the cytoplasm of E. coli ~tr~ln JM~Ol ~rom th- trc promoter, Amann et al. (l9~S), ~-n- ~0:183-190, hereby incorporated by reference.
R-comblnant E. coli were grown in M9 minlmal medium ~MaAiatis et al. (1982) supra) supplemented with 10 m~ml ca~amino acids ~Difco, Detroit, MI), 50 ~g/ml ampiclllin, and O.S ng/ml thym~ne in 10 liter volumes in a Microgen Fermentor tNew 8runswick 8clenctific, Edl~on, N.J.). Bacterial cultures were grown at 30-C, and ,, , . . , . . - , .:: . ,' ' ~ . .. .

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

. .j .

WO91/1~45 - 2s - 2~$~7 PCT/US91/~187 sparged with air at 5 L/min. When the absorbance (A590nm) of the culture reached o.3, expression of chimeric tox gene was induced by the addition of isopropyl-B-D- thiogalactopyranoside. Two hours after induction, bacteria were harvested by centrifugation, re~uspended in buffer ~lOl (50 mM ~H2P04, lO mM
EDTA, 750 mM NaCl, 0.1% Tween 20, pH 8.0), and lysed by sonication (~ranson Sonifier). Whole cells and debris were removed by centrifugation at 27,000 x g, and the clarified extract was then filter sterilized and applied to an anti-diphtheria toxin immunoaffinity column.
~ound proteins were eluted with 4M guanidine hydrochloride, reduced by the addition of ~-mercaptoethanol to 1% and then sized by high pressure lS liguid chromatography on a 7.5 x 600 mm G4000PW column ~To~oHass). Prior to uce, fusion toxins were exhauetively dialysed against HEPES buffered Hank's balancet salt solution (Gibco), pH 7.4. Purified diphtheria toxin was purchased from List Biological Laboratories (Campbell, CA). The concentration of all purified proteins was determined by using Pierce Protein As~ay reagent ~Pierce Chemical Co., Rockford, IL).
CYtotoxicltv The do~e re~ponse capacity of various chimeric 2~ toxln- to block ~l~C~-leuclne lncorporation by HUT
~0~/6I~ c-llJ ~whlch bQar the high afinity IL-2 r-c-ptor) and YT2C2 cells (whi¢h bear only the interm-diate affinity ~p7S) receptor) was determined.
Table ~ shows the concentration in mole~ of toxln r-gulred to lnhibit ~l~C~-leucine incorporation by ~0% ~IC50).

. ~ ~ ... . . .. . . . .
.` ' . ' -. ' ' .

' . '; I , - . ' '. , ' .

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

2f 3~ 7 wosl/1s74s - 26 - PCT/US91/~18 TABLE s CYTOTOXICITY

TOXIN ICso ~M) IL-2 -_ __ DAB486-IL-2 120 x 1o-12 70 x lo-9 DA~389-IL-2 40 x lo-12 11 x lo-9 DAB389-~1-1o)-IL-2 8 x 1o-12 2 x lo-9 DAB389-(1-10)2-IL-2 8 x 1o-12 2 x lo-9 DAB389-(1-10)3-IL-2 11 x 1o-12 2.2 x lo-9 As seen in Table S, the toxicity of the DAB389 - IL-2 chi~orie toxin is increased approximately 5-fold by the addition of ~ sp~eer peptide. The effect of two copies of the (1-10) spacer (~AB389-(1-10)2-IL-2) or three eopies of the (1-10) spaeer (DAB389-(1-10)3-IL-2) ha~e essentially the same effeet as does one copy of the 8pacer (DA8389-(1-10)-IL-2), Chimeric toxins in which DA~6 or DAB389 wa~ fu-ed to I~-~ (with or without the same spacer used on th- SL-2 con-truetion-, se- Table 1) were also an~ly~-d. ~h-n t-~t-d for eytotoxicity to rL-~-r-a~tor-b-aring o~ DAJ389-(1-10)2-IL-4 was 8 - -A to b- ~-10 tim 8 more cytotoxic than DAB389-IL-~whieh wa~ n to be about 10 timos more cytoto,xic thun, DA~ 86-S~
The addition of a spaeer to fusions o~ DT
fra~ nt~ to melanoeyte ~timulating hormone had no ff-ct on aytotoxicity.
Cytotoxicity assays were performed a~ follows.
F~or lS-2 eh~m~ric toxins cultured HVT 102/6TG (Tsudo et WO91/19745 - 27 - ; r' , , ~ PCT~USsl/04 2~ S~
al. (1986) Proc. Natl. Acad. Sci. USA 83:9694) or YT2C2 (Teshigawari et al. (1987) J. Exp. Med 165:223) cells were maintained in RPMI 1640 medium (Gibco, Grand Island, N.Y.) supplemented with 10% fetal bovine serum (Cellect, GIBCO), 2 mM glutamine, and penicillin and stceptomycin to 50 IU and 50 ~g/ml, respectively.
Cells were seeded in 96-well V-bottomed plates (Linbro-Flow Laboratories, McLean, VA) at a eoncentration of S x 104 per well in complete medium.
Toxin~, or toxin-related materials, were added to varying concentrations (10 12M to 10 6M) and the eultures were incubated for 18 hrs at 37C in a 5% CO2 atmosphere. Following incubation, the plates were eentrifuged for 5 min. at 170 x g and the medium removed and replaced with 200 ~1 leucine-free medium (MEM, Gibeo) eontaining 1.0 ~Ci/ml ~14C~-leucine (New England Nuclear, ~o~ton, MA). After an additional 90 min. at 37-C, the plates were centrifuged for 5 min. at 170 x g, the medium was removed and the cells were lysed by the addition of 4 M KOH. Protein was precipitated by the addition of 10% trichloroacetie acid and the insoluble material was then collected on glass fiber filters uslng a cell harvester (Skatron, Sterling, VA).
Filters were washed, dried, and counted aecording to standard method~. Cells cultured with medium alone r-rv-d a~ th- eontrol. IL-~ ehimerie toxin~ were tested ln a rlmll~r mann~r exc-~t that CT~R eells ~W~lliam E.
~ul, NSH), ~15 e~ ATCC), or CTL~2 ~ATCC) were a--d d at 1 x 10~ cells per well and incubated for 40 ~0 hourr.
Como-tit~ve disDlacement ex~eriments T~ble 6 ~how8 the competit~ve displacem-nt of tl25Il-1 bel d IL-2 from the high affinity IL~2 ~-e-~tor ~HnT102 eells) and the intermediate affinity IL-2 ree-ptor ~Y~2C2 eells) by various chimeric toxins.

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

': ~ : 1 , . .

WO91/19745 2 ~ 3 3A-~ 7 - 28 - PCT/US91/041~L~

TABLE 6 Competitive Displacement Toxin 50% displacement or HUTl02 YT2C2 Control .

IL-2 l x lO lO 2 x lO 9 DA~486-IL-2 35 x lO lO 120 x lO 9 DA~389-IL-2 29 x lO lO llO x lO 9 DA~389-(l-lO)-IL_2 13 x lO lO 30 x lO 9 DA~389-(1-10)2-IL-2 12 x lO lO 28 x lO 9 DAB389-(1-10)3-IL-2 13 x lO lO 32 x lO 9 A- r-en in Table 6, the bind~ng affinity is increased by the lnsertion of the (l-lo) spacer peptide into DAB389-IL-2 Insertion of two copies of the spacer (DAB389-~1-10)2-IL-2) or 3 copies of the spacer ~DAB389-~1-10)3-IL-2) has eissentially the sa~e ~ ct a~ does insortion of a single copy of the spacer ~DA1~389--~1-10)-IL-2) .
Compotitive displacement of ~125I]-rIL-2 by r-co~blnaAt IL-2 ~rlL-2) a~d ch~merlc toxins wa~
2~ d~t-r~in d ar tollow~ The ridiolabeled IL-2 blnding ~aaay waa ~ rtorm d ~a-ntlally ar dercribed by Wang et ~1 ~19~7 J ~x~ M d 166 1055-1069 Cells ~ere har~ ~t d, warh d with cell culture medium, and r-aua~nd d to 5 x 106 per ml and ~ncubated with ~ tl2~ rlL~ ~0 7 ~Ci/~mol) in the ~rosence or ~br-~- o~ i~ar---ing ~on~entrationr of unlabel d rIL-2 or chlm rlc toxin for 30 mi~ at 37 C under 5~ CO2 r-action wais then overlayed on a mixture o~ 80~ 550 - : ' ' . .`. ` . - ' .`..... ~ `-. ,. :. ' ',.. . . . , ~ ' :

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

3~37.
WO91/19745 - 29 ~ PCT/US91/~187 fluid (Accumetric Inc., Elizabethtown, KN) : 20~ parafin oil (d = 1.03 g/ml) and microcentrifuged. The aqueous phase and the pellet of each sample, representing free and bound lig,~nd, respectively, was then counted in a S Nuclear Chicago gamma counter. Apparent dissociation con6tants, Kd, were determined from the concentrations of unlabeled ligand required to displace 50% o radiolabeled rIL-2 ~inding to receptors.
U
The improved chimeric toxins of the invention are administered to a mammal, e.g., a human, suffering from a medical disorder, e.g., cancer, or other conditions characterized by the presence of a class of unwanted cells to which a polypeptide ligand can lS selectively bind. The amount of protein administered will vary with the type of disease, extensiveness of the disease, and size of species of the mammal suffering from the disease. Generally, amounts will be in the range of those used for other cytotoxic agents used in the treatment of cancer, although in certain instances lower amounts will be needed because of the specificity and increased toxicity of the improved chimeric toxins.
The improved chimeric toxins can be admnistered u~ing any conventional method: e.g., via in~ection, or ~ia ~ tlm d-relea~e implant. The improved chimeric toxin~ c n b- oomb~n d with any non-toxic, ph~r~ac-utically-acc-ptable carrier substance.
Other embodiments are within the following claimJ.
What i8 claimed is:

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

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

Claims (48)

WO 91/19745 - 30 - PCT/US91/04187

1. A chimeric toxin including protein fragments joined together by peptide bonds, said chimeric toxin comprising, in sequential order, beginning at the amino terminal end of the chimeric toxin, (a) the enzymatically active Fragment A of diphtheria toxin, (b) a first fragment including the cleavage domain l1 adjacent said Fragment A of diphtheria toxin, (c) a second fragment comprising at least a portion of the hydrophobic transmembrane region of Fragment B of diphtheria toxin, said second fragment having a deletion of at least 50 diphtheria toxin amino acid residues, said deletion being C-terminal to said portion of the transmembrane region, and said second fragment not including domain l2, (d) a spacer, (e) a portion of a cell-specific polypeptide ligand, said cell-specific polypeptide ligand being a cell growth factor, said portion including at least a portion of the binding domain of said polypeptide ligand, said portion of said binding domain being effective to cause said chimeric toxin to bind selectively to said target cell.

2. The chimeric toxin of claim 1, wherein said cell growth factor is a lymphokine.

3. The chimeric toxin of claim 1, wherein said deletion is at least 80 diphtheria toxin amino acid residues in length.

4. The chimeric toxin of claim 1, wherein said Fragment B of diphtheria toxin does not include any diphtheria toxin sequences C-terminal to amino acid residue 386 of native diphtheria toxin.

5. The chimeric toxin of claim 1, wherein (a), (b), and (c) comprise DAB389.

6. The chimeric toxin of claim 1, wherein said spacer is at least 5 amino acids long.

7. The chimeric toxin of claim 1 wherein said spacer is 10-30 amino acids in length.

8. The chimeric toxin of claim 1, wherein said spacer, when placed between the sequence of DT fragment DAB485 and amino acid residues 2-133 of IL-2, has a Bnorm value of 1.000 or greater.

9. The chimeric toxin of claim 1, wherein said spacer, when placed between the sequence of DT fragment DAB485 and amino acid residues 2-133 of IL-2, has a Bnorm value of 1.125 or greater.

10. The chimeric toxin of claim 1, wherein said spacer, when placed between the sequence of DT
fragment DAB485 and amino acid residues 2-133 of IL-2, has a Bnorm value of 1.135 or greater.

11. The chimeric toxin of claim 1, wherein at least 60% of the amino acids in said spacer are from the group lysine, serine, glycine, proline, aspartic acid, glutamic acid, glutamine, threonine, asparagine, or arginine.

12. The chimeric toxin of claim 1, wherein at least 80% of the amino acids in said spacer are from the group lysine, serine, glycine, proline, aspartic acid, glutamic acid, glutamine, threonine, asparagine, or arginine.

13. The chimeric toxin of claim 1, wherein said spacer is Ala-Pro-Thr-Ser-Ser-Ser-Thr-Lys-Lys-Thr.

14. The chimeric toxin of claim 1, wherein said spacer is at least 60% homologous to the sequence Ala-Pro-Thr-Ser-Ser-Ser-Thr-Lys-Lys-Thr.

15. The chimeric toxin of claim 1, wherein said spacer is at least 80% homologous to the sequence Ala-Pro-Thr-Ser-Ser-Ser-Thr-Lys-Lys-Thr.

16. The chimeric toxin of claim 1, wherein said spacer is a multiple of Ala-Pro-Thr-Ser-Ser-Ser-Thr-Lys-Lys-Thr.

17. The chimeric toxin of claim 1, wherein said spacer is at least 80% homologous to a multiple of Ala-Pro-Thr-Ser-Ser-Ser-Thr-Lys-Lys-Thr.

18. The chimeric toxin of claim 1, wherein said spacer is at least 60% homologous to a multiple of Ala-Pro-Thr-Ser-Ser-Ser-Thr-Lys-Lys-Thr.

19. The chimeric toxin of claim 1, wherein said spacer is Pro-Lys-Ser-Gly-Thr-Gln-Gly.

20. The chimeric toxin of claim 1, wherein said spacer is at least 60% homologous to the sequence Pro-Lys-Ser-Gly-Thr-Gln-Gly.

21. The chimeric toxin of claim 1, wherein said spacer is at least 80% homologous to the sequence Pro-Lys-Ser-Gly-Thr-Gln-Gly.

22. The chimeric toxin of claim 1, wherein said spacer is a multiple of Pro-Lys-Ser-Gly-Thr-Gln-Gly.

23. The chimeric toxin of claim 1, wherein said spacer is at least 80% homologous to a multiple of Pro-Lys-Ser-Gly-Thr-Gln-Gly.

24. The chimeric toxin of claim 1, wherein said spacer is at least 60% homologous to a multiple of Pro-Lys-Ser-Gly-Thr-Gln-Gly.

25. The chimeric toxin of claim 1, wherein said spacer is Pro-Thr-Ser-Ser-Ser-Thr-Lys.

26. The chimeric toxin of claim 1, wherein said spacer is at least 60% homologous to the sequence Pro-Thr-Ser-Ser-Ser-Thr-Lys.

27. The chimeric toxin of claim 1, wherein said spacer is at least 80% homologous to the sequence Pro-Thr-Ser-Ser-Ser-Thr-Lys.

28. The chimeric toxin of claim 1, wherein said spacer is a multiple of Pro-Thr-Ser-Ser-Ser-Thr-Lys.

29. The chimeric toxin of claim 1, wherein said spacer is at least 80% homologous to a multiple of Pro-Thr-Ser-Ser-Ser-Thr-Lys.

30. The chimeric toxin of claim 1, wherein said spacer is at least 60% homologous to a multiple of Pro-Thr-Ser-Ser-Ser-Thr-Lys.

31. The chimeric toxin of claim 1, wherein said spacer results in an affinity of said chimeric toxin for said target cells that is greater than the affinity of a second chimeric toxin for said target cells, said second chimeric toxin being identical to said chimeric toxin except that said second chimeric toxin lacks said spacer.

32. The chimeric toxin of claim 1 wherein said spacer results in said chimeric toxin exhibiting cytotoxicity for said target cells that is a least 2 times greater than the cytotoxicity exhibited by a second chimeric toxin for said target cells, said second chimeric toxin being identical to said chimeric toxin except that said second chimeric toxin lacks said spacer.

33. The chimeric toxin of claim 1, wherein said portion of said polypeptide ligand is a portion of interleukin-2 effective to cause said chimeric toxin to bind to T cells.

34. The chimeric toxin of claim 1, wherein said portion of said polypeptide ligand is a portion of interleukin-4 effective to cause e said chimeric toxin to bind to B cells.

35. The chimeric toxin of claim 1, wherein said chimeric toxin is encoded by a fused gene comprising regions coding for said protein fragments.

36. A DNA sequence encoding the chimeric toxin of claim 1.

37. An expression vector containing the DNA
sequence of claim 35.

38. A cell transformed with the vector of claim 36.

39. A method of producing the chimeric toxin of claim 1 comprising culturing the cell of claim 37, and isolating said chimeric toxin from the cultured cell or supernatant.

40. A spacer peptide said spacer peptide having the sequence Ala-Pro-Thr-Ser-Ser-Ser-Thr-Lys-Lys-Thr, or tandem repeats thereof.

41. A DNA segment encoding the spacer peptide of claim 40.

42. The DNA segment of claim 41, further characterised in that it includes a linker at each end.

43. The peptide encoded by the DNA segment of claim 42.

44. A vector containing the DNA segment of claim 41.

45. A cell transformed with the vector of claim 44.

46. A method of making the spacer peptide of claim 40 comprising culturing the cell of claim 45, and isolating said spacer from the cultured cell or supernatant.

47. The DNA segment of claim 41, wherein said spacer is a tandem repeat of the subunit Ala-Pro-Thr-Ser-Ser-Ser-Thr-Lys-Lys-Thr, and the sequence encoding each occurence of said subunit in said spacer is nonhomologous with the sequence encoding every other occurence of said subunit in the molecule, said nonhomology being sufficient to prevent recombination between sequences encoding tandemly repeated subunits.

48. A method of preventing recombination between the tandemly repeated subunits of spacer-peptide-encoding DNA comprising choosing the codons of each subunit-encoding sequence such that the DNA encoding each subunit is nonhomologous with the DNA
encoding every other subunit in the molecule, said nonhomology being sufficient to prevent recombination between tandemly repeated subunits.