CA1271900A - Toxin conjugates - Google Patents

Abstract

TOXIN CONJUGATES

Abstract A novel class of polypeptides of the general formula (F-(Pro)n)mF, wherein F represents a flexible amino acid sequence wherein each amino acid is individually selected from the group consisting of serine, glycine, and threonine, and n is an integer from 4-8 inclusive and m is an integer from 1-4 inclusive, is disclosed. These polypeptides are useful in the construction of conjugates between antibodies and peptide toxins. The preparation of such conjugate toxins by linking antibodies to toxin/spacer composites prepared by recombinant techniques is also disclosed.

Description

~7~900 TOXI~ CONJUG~T~S

Description Technical Field The present invention concerns the fields of cytotoxins, biochemistry, genetic engineering, and medicine. More ~articularly it concerns novel toxin conjugates, and components and uses thereof.

Background ~rt Bacterial and plant toxins, such as diphtheria toxin ~DT), Pseudomonas a. toxin A, abrin, --ricin, mistletoe, modeccin, and Shigella toxin, are potent cytocides due to their ability to disrupt a critical cellular function. For instance, DT and ricin inhibit cellular protein synthesis by inactiva-tion of elongation factor-2 and inactivation of ribo-somal~60s subunits, respectively. Bacterial Toxins_nd;Cell Membranes, Eds. Jelajaszewicz, J. and t~adstrom,~ T. (1978) Academic Pressr p. 291. These toxins are extremely potent because they are en~ymes - and~a~t catalytically~rather than~;stoichiometrically.
The molecules of these toxins are composed of an enzymatically active polypeptide chain or ~fragment, ~ : .
commonly called an "A" chain or fragment, linked to :

~ ~, ~. :

, 1 %71900 one or more ~olypeptide cnains or fragmen's, commonly called "8" chains or fragments, that bind the molecule to the cell surface and enable the A chain to reach its site of action, eg, the cytosol, and carry out its disruptive function. The act of gaining access to the cytosol is called variously "internalization", "intoxication", or "translocation". It is believed that the A chain must be timely liberated from the 8 :. :

. .

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

chain--~requently by reduction of a dislllfide bond--in order to make the A chain functional. These natural toxins are generally not selective for a given cell or tissue type because their B chains recognize and bind 5 to receptors that are present on a variety of cells.
~ erivatives of these bacterial and plant toxins have been prepared as therapeutic agents, pri-marily as antineoplastic agents, that are made speci-fic for tumor cells or other target cells by replacing 10 the native B chain(s) of the toxin molecule with a surrogate B chain that is specific for the tumor cell or adding a B chain having such specificity to the toxin molecule. Pharmacoloqy of Bacterial Toxlns, Eds. Drews, J. and Dorner, F. Pergamon Press. Syn-15 thetic cytotoxins containing the active fragment of abacterial or plant toxin or a cytotoxic drug are called variously "chimeric toxins", "toxin conju-gates", "cytotoxic conjugates", "hybrid toxins" or, when the surrogate ~ moiety is an antibody, "immuno-20 toxins". Antibodies tpolyclonal, monoclonal, andantigen binding fragments), hormones, lectins and various other compounds that are recognized by recep-tors on tumor cell surfaces have been used as surro-gate ~ moieties. See European patent application 25 publication no 0044167, and US Pats Nos 4,340,535, 4,350,626, 4,357,273, 4,359~457 4,363,758, 4,368,149, and 4,379,145.
Surrogate B moieties have been chemically linked to toxin A chains by a variety of coupling 30 agents. Heterobifunctional agents that include a disulfide group have been used extensively. The most popular of these agents is ~-succidimidyl-3-(2-pyri-dyldithio)propionate (SPDP). Immun_Rev, 62:185-216 (1982) reports using a longer disulfide-containing ,'.
I ,:

coupling aq~nt, 7-a~a-8-oxo-'0-~2-p5~ri~yl-11t'lio~d~-an-oic acid, to investigate whether greater separation between the A chain and antibody would increase acti-vity. Increased activity was observed in an acellular 5 system but not on intact cells, and the report conclu-ded that the longer disulfide containing linker was not advantageous. Replacement of the disulfide bridge by a stable thioether bridge using a derivative o~
6-maleimidocaproic acid as a bifunctional coupling 10 agent ~or an A-antibody conjugate caused a 99~ loss of activity relative to the disulfide conjugate in an intact cellular system.
The toxicity of diphtheria toxin for human lymphohlastoid cells was increased hy covalent linkage 15 to anti-lymphoblastoid (anti-CLA 4 and anti-Daudi) globulin. (Ross, W.C.J., et al, Eur J Biodiem (1980) 104:381). Thyrotropin releasing hormone derivatives have also been conjugated to DT fragments such as CRM 45 to study the translocation function (Bacha, D., 20 et al, J Biol Chem (1983) 238:1565). DT-A conju~ated using SPDP to a monoclonal antibody against guinea pig hepatocarcinorna cells showed specific cytotoxicity ln vitro and in vivo (Bernhard, M.I. et al, Cancer Res (1983) 43:4420).
Dextran, polyglutamic acid, and oligopep-tides containing up to four amino acid residues have been used as spacer arm bridges between cytotoxic drugs an-l antibodies. Nature (1975) 255:487-488, FEBS
letters (1980) 119:181-186, PNAS (1982) 79:626-629, .
30 and Immun Rev, 62:1-27 (1982). These conjugates were reported to have higher toxicity in vitro than drug coupled directly to antibody.
Moolten, F.L., et al, Immun Rev, (19823 62:47-72 conclude that A chain-antibody conjugates may . .
., , . -- ' : ' :' .

be more specific than native toxins, but lack much of the potency of the native toxin. They speculate that the loss of efficacy is because the internalization function is present in native B chains and surrogate 5 B chains are inefficient substitutes as regards inter-nalization. Use of toxins that lack a binding func-tion but are otherwise intact, such as the (cross reacting mutant) CRM45 of DT or toxins whose binding function is chemically or enzymatically abrogated, are 10 suggested, but no evi.lence of increased efficacy is given.
In summation, the efficacy of prior toxin-antibody con-iugates has been highly variable due, inter alia, to variations in immunogenicity, target 15 cell specificity, nonspecific toxicity, serum stabil-ity, effective concentration at the target cell, bind-ing efficiency, and internalization efficiency. In this re~ard the main objects of the present invention are to provide (l) means for improving the efficacy of 20 toxin-antibody conjugates and (2) novel toxin conju-gates that inclu~e those means. The present invention provides an A - surrogate B geometry which permits more facile translocation of the A portion into the target cell, and a more stable mode by which antibody 25 can be linked to the A portion. This latter property prevents premature decomposition prior to transloca-tion of the A portion into the target cell.

Summary of the Invention The present invention provides novel toxin 30 conjugates which are peculiarly effective in reco~ni-zing target cells and in effecting their demise. It also inclu~es certain components of these toxin conjugates. These conjugates comprise a cytotoxic .. , " ~ ;
, ' ', ~L~

component which is an enzymatically active portion of the molecule, capable of killing cells in which it is internalized, a specific binding moiety, typically an antibody or fragment of an antibody, which is capable 5 of recognizing a specific antigenic determinant or target cell, and a spacer which provides the proper geometry between the cytotoxic component an~ the binding fragment. These conjuqates represent an improvecl delivery system for naturally occuring or 10 modified cytotoxins ("A chains") which in their natural environment are bound to a relatively non-s~ecific binding component "B chain" (which is generally no~~an antibody). The naturally occuring toxins also contain, within the A-B chain fusion, 15 se~uences which are capable of cleavage intracellu-larly, ie which effect cleavage once the cytotoxic component has migrated to within the target cell, but are stable prior to this entry, and a translocation region which permits the ~esired cytotoxic A chain to 20 enter the target cell. This intracellular cleavage exposes and labilizes the link (typically disulfide) between the A and B chain.
In the conjugates of the present invention, the non-binding portion of the molecule is constructed so as to retain the foregoing ~esired cytotoxic, intracellularly cleavable/e~tracellularly stable, and translocation properties of the natural molecule in a geometry suitable for connecting to a "bin~in~
fragment" and permitting activity. Thus, typically, the conjugates of the invention include: antibody or fragment of an antibody which is covalently linked, preferably through a bifunctional linker to a non-binding entity. The non-binding entity is an amino acid se~uence which containsr to serve as the cyto-. ...~ :
, , : ~ . ~.
':,,, ~
'' ' toxic component, an enzymatically active site, anintracellularly cleavable/extracellularly stable site and translocation sequence and, a.s an extension of this amino acid sequence, a further sequence which 5 serves as a spacer between the cytotoxic component and the binding fragment to be linked.
The spacer confers the additional advantage of enhanced solubility in some instances where intermediates for desired immunotoxin may be sufficiently insoluble to 10 interfere with their purification. By supplying a relatively more soluble portion, or by altering the conformation of the remainder of the molecule, the spacer thus permits more options in the selection of components and conjugation methods.
The invention, therefore, relates to these conju~ates and to the novel components used for their construction. Both the spacer segment and the non binding portion of the conjugate forme~ by fusion of the spacer with a cytotoxic component, ie an extended spacer containing the cytotoxic component are aspects of the invention.
Thus, in one aspect, the invention relates to novel polypeptides (the spacer) consisting essen-tially of at least one rigid amino acid sequence bracketed by two flexible amino sequences - ie com-ponents of the formula ~flex-rigid)nflex, where "flex"
represents the flexible sequence and "rigid" the rigi~
- sequence. In a preferred emhodiment l'flex" is about 4 to 8 amino acids each individually selected from the group consisting of threonine, serine and glycine and "rigid" is 4-8 proline residues. Such novel poly-peptides are useful as spacers for separating the .
:, : ~ `'`
.

cytotoxic component and the target cell binding compon~nt of a toxin conjugate.
In the alternative, the spacer may be described in functional terms and comprises an amino acid sequence that:
~ a) is substantially stable in serum;
(b) is substantially nonhydrophobic so as not to affect adersely the water solubility of the toxin conjugate;
(c) is at least about lS A long;
(d) has a substantially extended structure; and (e) is sufficiently flexible to permit thcee dimensional movement of the cytotoxic component and the target cell binding component.
The spacer may also be a solubilization conferring sequence of amino acids, as set forth hereinbelow.
The foregoing spacer descriptions are not mutually exclusive, but are alternative characterizations of successful peptide sequences. The spacer may further contain a reactive amino acid residue proximate one of its termini so as to provide a conjugation site for the additional binding fragment.
Other aspects o the invention are fused polypeptides for use in making toxin conjugates comprising: ;
(a) a cytotoxic~portion; and (b) one of the above-described polypeptide spacers.
The invention also concerns the DNA sequences encoding the spacer or fused polypeptide (cytotoxic/

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

spacer); expression vectors containing these DNA
sequences, and cells transformed with these vectors.
Still another aspect of the invention is a toxin conjugate which comprises:
(a) a cytotoxic portion:
(b) a polypeptide spacer bound to the cytotoxic component; and (c) a target cell binding moiety conjugated to the cytotoxic component via the polypeptide spacer.
(The "cytotoxic portion" includes a site for intracellular cleavage within a serum stable domain and an internalization facilitating domain.) In summary, the invention is designed to provide a toxin conjugate which has the appropriate geometry for translocating the cytotoxic fragment into the target cell, the capacity to retain its binding fragment prior to such translocation, and/or the ability to solubilize the cytotoxic portion. In one aspect of the invention, the spacer is designed so as to permit the cytotoxic portion of the molecule ready access to the cell membrane. As the size of a typical antibody (binding) fragment is very much greater than that of most cytotoxic fragments, there is considerable steric hinderance of the access to the cell membrane by the cytotoxic portion imposed by the sheer bulk of the antibody or antibody fragments. Accordlngly, the conjugate toxins of the invention have a geometry schematically represented in (a) rather than that given without the spacer (b).

.; ~.- .. .

- . .
' . ~:
:.:
.. ..

~27~9~) In order to effect this translocation, the spacer needs to he sufficiently flexihle to allow the ~
portion to reach the cell membrane, and suficiently extende~ to permit it to have sufficient reach.
The performance of the conjugated toxin can also be improved by securing the binding portion tightly to the remainder of the molecule with respect to a serum environment. This is done in one preferred embodiment of the invention, by utilizing a linker between the antibody and spacer which employs bonds not readily cleaved by reducing agents or by hydroly-sis under extracellular conditions. The cleavage of the A-chain analo~ ~ie the enzymatically active site) from the other end of the spacer arm can be achieved by permitting the linkage at the A end of the spacer to be more readily cleavable. This can best be done by retaining a portion of the original B chain of the native toxin in linkin~ the A portion to the spacer.
In this manner, the normal extracellular-resis-tant/intracellular-cleavable configuration of the A-B

. - . ; :

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

chain pair is retained but with the loss of the bind-ing capacity of the original ~ portion. Thus, the specific binding capability conferred by the antibody on the conjugate is not lost prior to intracellular incorporation of the cy,totoxic fragment.
Since the binding portion confers specific recognition o certain target cells, the conjugated toxins are useful in killing specified undesirable cells within a subject. Thus, in two other aspects, the invention relates to pharmaceutical compositions containing effective amounts of these toxins and to methods of treatment e~ploying them.
.. _ rieE Descri~tion of the Drawings Figure 1 shows the base sequence for the DT
gene along with the corresponding de~uce~ amino aci sequence.
Figure 2 shows the construction of an Msp-Spacer arm clone, pMspSA2.
Figure 3 shows the construction of two Msp-Spacer fragment expression vectors wherein the coding sequence is under the control of the PL pro-moter: pPLMspSA and pPLOPMspSA2.
Figure 4 shows the construction of pTrpSmlMbo.

Modes for Carrying Out the Invention A. Definitions As used herein the terms "fragment", "domain'i, and "region" and '1portion" are interchange-able and refer to functionally but not necessarily physically distinct portions of the conjugated toxin moIecule.

.

~ '; :

~271~0 The ter~ "specificity" as used to describe the target cell binding portion of the conjugated toxin means that the moiety has the ability to distin-guish a target cell from other cells, typically due to 5 the presence of a cell surface receptor that is unique to the target cells.
The term "selective" means that the cyto-toxin has the ability to kill target cells preferen-tially, typically due to the specificity of the 10 binding moiety or a differential in the respective ~uantities of receptors on target cells and other cells.
The term "target cells" meAns those cells which thè cytotoxin is intended to ~ill. ~lthough 15 target cells will usually be tumor cells, they may be nontumorous cells whose selective destruction is desired for therapeutic or ~iagnostic purposes tfor instance in certain assays o~ peripheral blood cells it is desired to selectively kill one or the other of 20 B cells or T cells). The target cells may be present in living organisms or they may be preserved or main-tained in vitro. The cells may be individual or asso-ciated to form an organ.
As used herein the term "polypeptide" or "protein" refers to an a~ino acid polymer. It is understood that such polymers exist in a variety of ionization states dependant on ambient pH, and that they may, further, be associated with accessory moie-ties - eg glycosylated, phosphorylated or conjugated 30 ~o lipids. The peptides or proteins of the invention include all such forms, including unassociated forms.
The term "intracellular" is intended to include intracytoplasmic sites and sites within vesi-cular compartments such as lysosomes.

~ .
.

~7~0~

As used herein with respect to the enzymat-ically active polypeptide fragment the term "epitope"
means a domain (amino acid sequence) of the fragment that is a causative factor in an immune response thereto.
"Target cell binding portion" refers to that fragment of the toxin conjugates o~ the invention ~hich binds specifically to the target cells. In this invention the "binding" portion is an antihody or fragment thereof. The "non-binding" portion or frag-ment of the toxin conjugate comprises the remainder of the molecule. It thus includes the cytotoxin portion and the spacer.
The term "cytotoxic portion" includes the "enzymatically active polypeptide fragment" ie an A-fragment analogous sequence, the intracellularly cleavable/extracellularly stable do~ain and the translocation domain.
As used herein with respect to the construc-tion of the spacer domain, "reactive amino acid resi-~ue" refers to an amino acid (or residue) which pro-vides a site for lin~ing or conju~ation, as further set forth in B.4, below~
"Solubilizing conferring" sequence of amino acids in the context of the invention refers to a form of spacer sequence which continues at the (preferably C) terminus of the cytotoxic portion and which results in the fused protein being soluble in aqueous media, even when the cytotoxic portion is itself insoluble or when the cytotoxic portion would otherwise be rendered insoluble by the addition of a cysteine residue.

.

.
.:' .. ' , ~;~7~

. Structure of ~he Com ounds of the Invention p Toxin conjugates have classically been conceptualized as combinations of an A fragment and a surrogate B moiety, attached through linkinq group that binds these fragments via a labile bridge such as a disulfide hridge. The rationale hehind this concep-tualization was that the function of the B moiety was primarily to bind the conjugate to the cell surface via interaction with a cell surface receptor and that the disulfide hridge provided means for joining the A

, .

. .~.:
, ~2~

fragment and B moiety that could be bro~en in vivo to liberate the A fragment. The present invention is based on an expanded conceptualization o~ the mode of operation of cytotoxic conjugates and takes into 5 account factors including:
the spatial relationship between the cyto-toxic moiety and the binding moiety, the role of the non-A portion of the conju-gate in internalization, and the extracellular lability of the bond between the cytotoxic moiety and the binding moiety.
In the most preEerred embodiment of the present invention with respect to toxin conju~ates each of these factors is taken into account in syn-15 thesizing a novel toxin conjugate having five func-tional ele~ents (which may overlap in the structure of the conjugate):
(1) an enzymatically active domain capable of the toxic activity of the A-~ragment;

(2) an intracellular cleavage site within a extracellularly stable (or serum stable) domain that provides a site for liberating the enzymatically active domain from the remainder of the toxin conju-gate after the toxin conjugate has been internalized:

(3) a translocation or internalization facilitating domain that acts as an adhesive to anchor the toxin conjugate molecule to the target cell mem-brane and/or to facilitate internalization;

(4) the novel polypeptide spacer described 30 above; and

(5) a target cell hinding moiety that recog-nizes and binds to a receptor on the target cell sur-face which by virtue o~ the receptor's quantity or nature causes the toxin conjugate to localize selec-tively on target cells relative to other cells.

, - ,, ~ . ;

~2~)~

In the most preferred embodiment the binding moiety is bound to the non-binding portion of the mol-ecule (ie the remainin~ elements 1-4) by a chemical bond that is substantially resistant to cleavage ln 5 vivo (either extracellularly or intracellularly) thereby preserving the specificity of the molecule once it is administered, while the intracellular clea-vage site within a serum stahle domain is provided at the enzy~aticall~ active or "A" end of the spacer.
10 In order to effect this linkage using one preferred embodiment the cytotoxic porticn of the toxin needs to be provided wi~h a free cysteine residue, capable of forming a thio-ether linkage through a suitable bifunctional linker. The presence of this cysteine may interfere with 15 the solubility of the toxic portion, and the spacer then serves the additional function of providing solubilization. To accomplish this purpose, the detailed "spatial" requirements relating to rigidity and flexibility of the sequence are n~t required; all that is needed are the hydrophilic or neutral solubility properties of the chain.

n.l. The_~nzymatically Active Domaln The enzymatically active fragment of the 25 conjugate may be the A chain of a bacterial or plant toxin or be a natural protein that has enzymatic activity similar to the A chain of a bacterial or plant toxin. As used herein the terms "enzymatically active fra~ment" and "A chain" are intended to include 30 such similar acting natural proteins. Examples of such A chains are diphtheria A chain, exotoxin A chain (from Pseudmonas aeruginosa), ricin A chain, abrin A
chain, modeccin A chain, alpha-sarcin, Aleurites . , .

~27~

fordii proteins, dianthin proteins, Phytolacc_ americana proteins (PAP I, P~P II, and PAP-S), momor-__ din, curcin, crotin, gelonin, mitogellin, restric-tocin, phenomycin, and enomycin.
The derivation of an A chain from a whole natural toxin molecule involves breaking the bond(s) between the A and B chain(s) (eg, reducing the disul-fide honds(s) between the A and B chain(s) with an 10 appropriate reducing agent, such as 2-mercaptoethancl or dithiothreitol) and isolating the A chain from the B chain(s). The isolation may be carried out chroma-tographically or by other conventional protein frac-tionation techniques. The natural proteins that have enzymatic activity similar to the A chains of the natural protein toxins may be isolated from their sources, typically plant parts, by conventional pro-tein extraction and isolation techniques. Following 5 the isolation it may be possible depending upon the location and nature of the A chain epitopes an~l adja-cent residues to remove one or more of the epitopes by partial proteolytic digestion or by chemical modifi-cation without affecting the enzymatic activity of the 10 A chain. The enzymatically active fragment-is perhaps most conveniently produced by cloning and expressing the gene encoding its amino acid sequence using the techniques of recombinant technology. When it is thus ~enerated by genetic engineering, the epitopes may be 15 removed at the DNA level by recombinant DN~
techniques.

B.2. The Intracellular Cl_avage Site The intracellular cleavage site domain of the cytotoxin is preferably one that functions in a 20 way that ~imics the manner in which a natural toxin liberates its A chain. Three cleavage mechanisms are postulate~ currently: (l) proteolysis either enzy-matic or chemical (eg, p~ change), (2) disulfide reduction, and most commonly (3) a combination of (1) 25 and (2). The cleavage site(s) of such domains is substantially stable extracellularly and is labile intracellularly. In the cleavage mechanisms involving proteolysls and disulfide reduction, extracellular stability is probably due to the position of the 30 disulfide cleavage site in the extracellular tertiary structure of the conjugate. That is, the site is not exposed to cleavage agents in the extracellular envi-ronment, but is exposed in the intracellular environ-~27~

ment due to a change in the tertiary structure of themolecule. Cleavage sites whose lability depends on pH
are stable in extracellular environments, eg, bloo~, having a substantially neutral pH. The lower pH of 5 certain intracellular compartments (eg, within a lyso-some or receptosome) makes the site lahile. The clea-vage site domain comprises a sequence of amino acids that includes residues that are susceptible to pro-teolysis such as by lysosomal proteases. When disul-fide reduction is involved the sequence will obviouslycontain a disulfide bridge formed by spaced cysteine residues. When both proteolysis and reduction are involved th-e A chain is liberated from the remainder of the toxin conjugate by proteolysis at the sensitive 15 residues that causes a nick or break in the polypep-tide backbone of the molecule and reduction of the disulficle bond. Examples o residues that are lyso-somal protease sensitive are arginine, lysine, phenyl-alanine, tyrosine, and tryptophan.
In a preEerred embodimerlt, this cleavage site in a serum stable domail1 is proximate the enzymatically active domain and forms an extension of the C-terminus thereof. Such a configuration can be provided most conveniently by providing an extended "~" fragment from a naturally occurring toxin into a portion of the natural ~ chain. Recombinant tech-niques are best suited to this embodiment, since convenient peptide cleavage technigues sepcific for the desired extension may not exist for a given toxin. However, the coding sequence can be cleaved and modified so as to encode for just the desired fragment.

~27~

B.3. The Translocation Domain The internalization facilitating or "trans-location" domain of the conjugate participates in interacting with the cell or vesicle wall whereby the wall is penetrated, opene~, or disrupted to enable the conjugate to reach the intracellular co~partment. The internalization facilitating domain may be identical to, or substantially similar in, amino acid content an~ se~uence to an internalization facilitating domain of a bacterial or plant toxin or a polypeptide that is known to interact similarly with cell membranes. In the case of DT, the domain has been identified as being "hydrophobic" and located within the B frag-ment. The sequence of that segment possibly includes the followin~:
.
Val-Ala-Gln-Ala-Ile-Pro-Leu-Val-Gly-Glu-Leu Val-Asp-Ile-Gly-Phe-Ala-Ala-Tyr-Asn-Phe-Val Glu-Ser-Ile-Ile-Asn-Leu-Phe-Gln-Val-Val Examples o~ polypeptides that are known to have a similar interaction with cell membranes are melittin:

Gly-Ile-Gly-Ala-Val-Leu-Lys-Val-Leu-Thr-Thr-Gly-Leu-Pro-Ala-Leu-Ile-Ser-Trp-Ile-Lys-Arg-Lys-Arg-Gln-Gln and delta lysine:

Met-Ala-Gln-Asp-Ile-Ile-Ser-Thr-Ile-Gly-Asp-Leu-Val-Lys-Trp-Ile-Ile-Asp-Thr-Val-Asn-Lys-Phe-Thr-Lys-Lys.

The position of the domain in the toxin conjugate molecu]e may vary. It will usually be located adja-cent to the carboxy terminus of the A chain but may .

, ' ' " ' also be located adjacent to the amino terminus of the A chain. More than one internalization facilitating domain may be included in the conjugate if desired.
Since this domain is positioned adjacent the 5 A chain, its inclusion in the toxin conjugate is most conveniently accomplished by recombinant DN~ tech-niques. In DT, for example, the extension of the polypeptide sequence at the C terminus approximately 193 amino acids into the B chain provides such an 10 internali~ation domain. Therefore, cloning and expression of the coding sequence ~or this portion o~
the DT toxin would provide the desired configuration.
~lternatively, the oligonucleotides encoding known internalizing domains such as those exemplified above 15 can be ligate~ to the nucleotides encoding the desired fragment for cloning and expression.

B.4. The Spacer _ _ _ _ .
The spacer comprises a sequence of amino acids that can be described in one oE three non-mutually 20 exclusive ways. In any case, the end of the spacer that attaches to the binding moiety may have a "reactive amino acid" residue at or near the terminus that provides a site for conjugating the binding moiety. For instance, if a reactive amino group at or near the end of the 25 spacer is desired one or more lysine residues may be located near one terminus of the spacer fragment. If a reactive sulfhydryl group is desired, a cysteine residue may be situated similarly.

~~

. ~ .

-~2-In one aspect, the spacer is described as having one or more extended structure portions (segments in which the peptide bond angles are enlarged) linked by segments that are flexible. ~ach end of the spacer terminates with a flexible segment. The spacer thus links the target cell binding moiety to the remainder of the molecule with the extended structure portion(s) serving to separate the binding moiety and the remainder of the molecule and the flexible segments permitting three dimensional movement of the binding moiety and the remainder of the molecule. Thus the spacer can be described as being formed from a series of extended structure portions in tandem with intermediate flexible regions (the flexible regions lie on either side of the extended structure regions), ie, has the general formula (flex-rigid)m flex. The extended (rigid) portion(s) of the spacer is preferably formed of a series of 4 to 8 prolines while the flexible portions are preferably composed of 4-8 amino acid residues each selected individually from the group consisting of serine, glycine, or threonine. The series of prolines form a left-handed proline II helix. These preferred spacers (less terminal reactive eesidue) may be represented by the formula (F-(Pro)n)mF
wherein F represents a flexible sequence composed of amino acids each selected independently from the group consisting of serine, glycine, or threonine, n is an integer form 4 to 8 inclusive, and m is an integer from l to 4 inclusive. The fle~ible sequences may be the same or different. A particularly preferred spacer domain (less terminal reactive residue) is defined by ~he sequence Gly-Thr-Gly-se{-Gly-(pro)n-ser-Gly-ser-Gly-Thr where ~27~

n is an integer from 4 to 8, inclusive, most preferably

6. A particularly preferred terminal reactive residue i5 CyS .
In a second aspect, the spacer is substantially nonhydrophobic so that it has a neutral or positive effect on the water solubility of the conjugate. The spacer's hydrophobicity may be determined by summing the hydrophobicities of the individual amino acids (measured by partition coefficient tests) of which it is compose2.
A substantially nonhydrophobic sequence will measure neutral or hydrophilic. The hydrophilic nature of this segment wlll also place it on the surface of the configured molecule, thereby permit~ing accessibility for conjugation.
If the function of a particular spacer is merely to provide solubility during processing to a cytotoxic portion amino acid sequence, this property is, indeed, the only property reguired of it. Generally, the spacer is a relatively hydrophilic sequence, typically containing a cysteine residue.
In a third aspect, the spacer can also be described in functional terms as substantially stable in human serum, having a length selected such that it provides an extended structure link at least about 15 A
long, preferably about 30 to about lO0 A long, between the binding moiety and the remainder of the conjugate molecule, as being substantially non-hydrophobic so as not to adversely affect solubility, and having sufficient flexibility to permit ~hree dimensional movement of the cytotoxic component with respect to the binding component.

.5. The Target Cell Binding MoietY
The binding moiety may be any ligand that has the required target cell specificity. Antibodies, particularly monoclonal antibodies or their antigen 5 binding fragments, are preferred binding moieties.
Monoclonal antibodies against surface receptors of target cells may be made by the somatic cell hybri-dization procedure first described by Kohler, G. and Milstein, C., Nature, (1975) 256:495-497. The cell 10 lines, reagents, and conditions used in this procedure are well known and have been reviewed extensively in the literature (Somatic Cell Genetics, (1979) 5:957-972). ~riefly the procedure involves immunizing a host with the immuno~en of interest, collecting 15 antibody-producing cells from the immunized host, ; fusing the antibody-producing cells from the immunized host with an appropriate tumor cell line using a fuso-gen such as polyethylene glycol, growing the cells in . , , a selective medium to eliminate unhybridized partners, identifying hybridomas that produce antibody against the immunogen, growing such hybridomas, and collecting monoclonal antibodies from the resulting culture 5 medium (or body fluid when grown 1n vlvo). Antigen binding fragments ~Fab, Fab', F(ab')2, Fv) of the monoclonal antibodies may be made by digesting the whole Ig with an appropriate protease, for instance papain in the case of Fab and pepsin in the case of ~(ab')2. Antigen binding fragments will be particu-larly useful in instances where it is desired that the binding moiety lack its natural effector function.
Antibo~ies ~ current interest will typically be of human, rat or murine origin since rat, mouse and human tumor cell lines are available for fusion~ Also, a variety of antitumor monoclonal antibody reagents are rapidly becoming available. ~luman monoclonal anti-bodies are preferred for use in making conjugates for use in human therapy because of the reduced likelihoo-l of immunogenicity.
In Sections C-E, all temperatures are in degrees Celsius.
C. r1ethods of ~eparation C.l _The Non-binding Portion (cytotoxic-spacer Portions) Depending on the sizes of the four polypep-tide domains that make up the nonbin~ing portion ofthe conjugate toxins, these domains may be synthesized individually or as subunits by conventional polypep-tide synthesis techniques (Margolin, A. and Merrifield, R.B., nn Rev Biochem, (1970) 39:841) and combined in sequence by ~nown procedures.
Recombinant DNA methodology provides an alternative and preferre~ way of synthesizing the nonhinding portion of the conjugate, either as indi-~27~

vidual subunits, or as an entire fused polypeptide.
This process involves obtaining a DNA sequence that encodes the nonbinding portion (or a particular domain or comt)ination of domains) inserting the DNA sequence S into a suitable expression vector, transforming micro-organisms or cells with the vector, growing transfor-mants that produce the desired fra~ment and harvesting the d`esired fragment from the transformants or their growth medium. The coding sequence may he made by synthesis of DNA subunits that encode portions of the enzymatically active fragment, translocation domain, cleavage site domain, and spacer domain and assemblin~
them by lig~tion techniques known in the art, or by cloning those portions of naturally occuring genes which encode the desired protein. The D~l~ sequence that encodes the enzymatically acti~e fra~ment will normally be isolated from naturally occurring DNA; as may the sequence encoding the cleavage and transloca-tion domains. These and the spacer encodin~ sequence may also he made by conventional DNA synthesis tech-niques, and may be reproduced by conventional DN~
cloning methods. Partial structural genes of bacter-ial or plant toxins that lack a binding function but retain their enzymatic, cleavage and internalization functions (eg, the CR~ 45 mutant of DT) may be cloned and ligated to a DN~ sequence that encodes the spacer.
The ligation product is inserted into suitable expres-sion hosts and expressed to make the nonbinding portion of the conjugate.

C.2. The ~inding Portions The binding portions of the conjugates of the invention are antibodies or fragments thereof.
Use of monoclonal antibodies is preferred. Antibodies ,:

.

are prepared by means well known in the art, and can be isolated from serum, from spleen, or most prefer-ably prepared and isolated from antibody secreting hybridomas. Many are commercially available (see B.5 5 herein).

C.3. Conlugation of Bin~inq and Non-bindin~
Portions Since the antibodies or fragments thereof are prepared by reactions of immunoglobulins isolated 10 from serum or from hydridomas, they must be linked in vitro to the non-bin~ing domains. In a preferred configuration of the non-binding fragment, the cyto-toxic portion of the molecule is extended fro~ its C
terminus hy a sequence providing a spacer which has 15 been provided with a reactive amino acid residue, cysteine, located at or near its C terminus. While bifunctional reagents, such as SPDP, that result in a lahile disulfide ~C-S-S-C) bridge between the spacer antl binding moiety may be used as coupling agents, 20 (and have often been so used where a cleavable link was desired) bifunctional reaqents that form a stable ~nonre~ucible) bond, such as a thioether (C-S-C) link are preferred. Such bonds are not susceptible to cleavage in the environment of use, that is, they are 25 not cleaved extracellularly in vivo. Several protein couplin~ agents that form thioether bonds are avail-able. These agents usually contain on one end of the molecule an activated carboxyl group that will react with free amino groups, eg, an ~-amino of a lysine of 30 one of the conjugation partners (in this case the bindin~ portion) and a maleimido or haloacetyl group that will react with a sulfhydryl of the other partner (in this case the non-binding portion) on the other enA. Examples of such thioether-forming agents are active esters of the following aeids:
6-maleimidocaproic acid, 2-bromoacetic acid, 2-iodoacetic acid, o o ¢ N-CH2 ~ COOH , O

¢N{~--C112-C112-C112-Coo~

Activation of the carhoxyl groups is required to permit reaction with amino groups of the ehosen protein. The desired activated derivatives of the foregoing acids inelude most preferably the suceinimidyl ester, ie R

R-C -O-N
\C--CH2 o although others, espeeially esters, have been used, such as the water soluble ester formed from l-hydroxy-2-nitro-4-sulfonic acid sodium salt, r~
R-C(O)-O- ~ ~S3Na Other coupling agents that may be used are various hifunctional derivatives of imidoesters such as ~imethyl adipimidate HCl, active esters such as 5 disuccinirnidyl suberate, aldehydes such as glutaralde-hyde, bis-azido compounds such as bis(~-azi~o-benzoyl)hexanediamine, bis-diazonium derivatives such as bis-(~-diaoziumbenzoyi)-ethylene diamine, diisocy-anates such as tolylene 2,6-diisocyanate, and bis-10 active fluorine compoun~s such as 1,5-difluoro-2,4-dinitrobenzene. Heterologous permutations of such bifunctional derivatives may also be used as well as peptide bond-generating reagents such as carbodi-imides.
In a typical, preferred approach, a thio-ether linkage is formed between a sulfhydryl on the non-binding portion of the cOnjUgAte and the coupling agent and an amide linkage is formed between an ~-N~2 of a lysine contained in the binding portion an~ the 20 carboxyl of the coupling agent.

D. Mode of Administration When used to kill target cells in vitr_ the conjugates will typically be added to the target cell culture medium in amounts ranging from about 1000 to about 100,000 conjugate molecules per target cell.
The formulation and mode of administration for in itro use are not critical. Aqueous formulations that - are compatible with the culture or perfusion medium will normally be used.
~Jhen used ln vivo for prophylaxis or therapy of humans or animals (e~, farm~ laboratory, sport or , :.
. :

:. "., ~2~

pet animals~ the cytotoxins are administered to the patient in a manner and dose that induce the desireA
target cell reduction response. They will normally be administered parenterally, preferahly intravenously.
The dose and dosage re~imen will depend upon the nature of the target cell and its population, the characteristics of the particular cytotoxin, eg its therapeutic index, the patient, and the patient's history. The amount of cytotoxin administered will typically be in the range of about 0.01 to about l m~/kg of patient weiyht.
For parenteral administration the cytotoxins will be formulated in a unit dosage injectable form (solution, suspension, emulsion) in association with a pharmaceutically acceptable parenteral vehicle. Such vehicles are inherently nontoxic and nontherapeutic.
Examples of such vehicles are water, saline, ~inger's solution, dextrose solution, and ~an~'s solution.
Nonaqueous vehicles such as fixed oils and ethyl oleate ~ay also be used. Liposomes may be used as carriers. The vehicle may contain minor amounts of additives such as substances that enhance isotonicity and chemical stability, eg buffers and preservatives.
The cytotoxin will typically be formulated in such vehicles at concentrations of about l mg/ml to 10 mg/ml.

~. Detailed Decri tion of a Preferred Embodiment P _ An illustrative and preferred embodiment of the invention comprises the components and inter-mediates in a specific conjugated toxin. In thistoxin, the bin~ing fragment consists of anti~Daudi antibodies which are linked by reaction with the succinimidyl ester of m-maleimidoben7.0ic acid to a spacer portion of the formula ~, ~, 9~0 Gly-Thr Gly-Ser-Gly(Pro)6 Ser-Gly-Ser-Gly-Thr-Cys spacer reactive amino acid which is in turn a C terminal extension of a portion of the naturally occurring diphtheria toxin. Thus, the conjugate toxin here exemplified can be repre-5 sented by the formula:

DT-B ' ,, SPACEr~ I N~ICHCII -S- c C-7 wherein DTA represents the enzymatically active portion, or "A" chain of diphtheria toxin, DT-~' represents the first approximately 190 amino acids of the diphtheria ~ chain, and AB represents the anti-Daudi antibody.
In this embodiment, three of the elements of the non-binding portion o~ the conjugate toxin are derived from diphtheria toxin; the enzymatically active domain, the cleavage site domain and the trans-location domain. The mature diphtheria toxin molecule including both A and B chains contains 535 amino acid residues of which the A chain (the amino terminal fragment) contains 193 residues and the B chain (the carboxy terminal fragment) contains 342 residues. It appears that the intracellular cleavage site between the ~ and B chains in the native toxin is after one of ,'~. , ~.

~2~

the three Arg residues. Such cleavage readily takes place in vitro catalyzed by trypsin-like enzymes. It is thought that a similar cleavage takes place in vivo, thus exposing the disulfide link between the cysteine at position 186 and that at 201 for reductive intracell~lar cleavage. It is known that the amino terminal 193/342 of the B fragment contains sequences that cause the toxin to insert into artlficial lipid bilayers under appropriate conditions forming ion conductive channels (Kagan, B.L., et al, Proc Natl Acad Sci (USA), (1981) 78:4950; Donovan, J.J., et al Proc Natl Acad Sci, (1972) 78:172 (1972); Kaiser, G., et al Bioche~ Biophys Res Commun, (1981) 99-358 FEBS
Letters (1983) 160:82). Thus the portions of the 15 diphtheria toxin which are embodied in the illustra-tive conjugate toxin contain the intracellular clea-vage/extracellularly stable connection between the A
an~ B chain and at least a portion o~ the hydrophohic ~omain responsible for translocation of the cytotoxic portion into the cytosol.
Fig. 1 shows the sequence of the ~liphtheria toxin gene and the flanking regions, along with the deduced amino acid sequence. The deduced se~uence is in reasonable agreement Wit}l the previously reporte~
primary amino acid sequence data tDelange, R.J., et al, Proc Natl Acad Sci (USA) (1976) 73:69; Delange, ~.J., et al, J Bio Chem (1979) 254:5827, Drazin, R.E., et al, (ibid) 5832 (1979); ~elange, R~J., et al, (ibid) 5838 (1979); Falmagne, P , et al, Biochim Biophys Acta (1978) 535:S4; Falmagne, P., et al, Toxicon (1979) 17:supp 1 46; Lambotte, P., et al, J
Cell Biol (19801 87:837; Capiau, C., et al, Arch Phys (1982) 90:B-96; Falmagne, P., et al, Toxicon (19823 20:243). The deduced sequence assumes a leader .

sequence as shown, consistent with the fact that DT~
is secreted from the natural source, C. Diphtheriae and with the fact that the sequence in this region strongly resembles known signal peptides (r1ichaeli 5 S., et al, Ann_Rev Microbiol ~l982) 36:435). It is presently believed that the GTG codon at position -25 serves as a start codon and encodes methionine rather than the valine there shown.
The entire toxin gene sequence is carried by lO bacteriophage~ and can be isolated from the phage by restriction with Xba l and EcoR l. A shorter MspI
fragment within this sequence (see Fig l) comprises most of the~sequence used in the illustrative construct herein, this ~ragment results from Mspl 15 restriction about 300 bp preceding the first amino acid codol1, and at the site shown at the codon encoding amino acid 3~2, approximately in the middle of the B fragment.
The construction of the illustrated conju-20 gate toxin may be summarized as follows:
the Msp portion of the gene containing thecodons for the A chain and approximately half of the B
chain are ligated to synthetic DNA encoding the desired spacer with its reactive amino acid (cysteine) 25 terminus. The resulting oligonucleotide is then modi-fied to delete the prornoter and ribosome binding regions as well as the codons for the leader sequence.
It is then ligated into an operably linked position with respect to the PL promoter an~ N-gene riboso~e 30 binding site in suitable expression vector, along with an ATG start codon immediately preceding the glycine residue at position l. ~acteria transformed with this expression vector produced the entire non-binding region of the molecule. The transforlned cells are sonicated and the non-binding portion recovered from the sonicate. The non-binding portion is then linked to the anti-Dau~i antibo~y in vitro usin~ an activated m-maleimidobenzoic ester as linker.

E.l Methods and Procedures . . _ . . .
Construction of suitable vectors containing the ~esired coding and control sequences employs standard li~ation and restriction techniques which are well understood in ~he art. Isolated plasmids, DNA
10 sequences, or synthesized oli~nucleotides are cleaved, tailored, and religated in the form desired.
Cieavage is performed by treating with a suitable restriction enzyme (or enzymes) under condi-tions which are generally understood in the art, and 15 the particulars of which are specified by the manufac-turer of these commercially available restriction enzymes. In general, about 50 ~g of plasmid or DNA
sequence is cleaved by 50 units o~ enzyme in about 100 ~1 of buffer solution; in the examples herein, 20 typically an excess of restriction enzyme is used to insure cleava~e. Incubation times of about one hour to two hours at about 37C are workable, although variations can be tolerated. After each incubation, protein is removed by extraction with phenol/chloro-25 form an~ the nucleic acid recovered from aqueousfractions by precipitation with ethanol followed by ~r~ desalting over a Sephadex~G-50 spin column. If `-'3 desired, size separation of the cleaved fragments may be performed by polyacrylamide gel electrophoresis 30 using standard techniques. A general description of size separations is found in Methods in Enzymolo~y (1980) 65:499-560.

~ ~rc~ i~a r1~

,,, ,, . ., ~ . , . .. .... ., .. ... ~ .. .. . .. ... .. . . . . . . .. . . .. . ... .

. .

Restriction cleaved fragments may be blunt ended by treating with E.coli DNA polymerase I
(Klenow) in the presence of 0.01 mM of the four nucleotide triphosphates (dNTPs) using incubation 5 times of about 15 to 25 min at 20 to 25C in 50 mM Tris pH 7.6, 50 mM NaCl, 6 mM MgC12 and 6 mM
DTT. The Klenow fragment fills in at 5' sticky ends but chews back single strands even though the four dNTPs are present at 3' sticky ends. After treatment 10 with Klenow, the mixture is extracted with phenol/chloroform and ethanol precipitated and ~'~ desalted by running over a Sephadex~G-50 spin column. Treatment under appropriate conditions with Sl nuclease results in hydrolysis of any single 15 strande-1 portion.
Synthetic oligonucleotides are prepared by the triester method of Matteucci, M., et al, J Am Chem Soc (1981) 103:3185. Kinasing of single strands prior to annealing or for labeling is achieved using approx-20 imately 10 units of kinase to 1-10 nmoles substrate in the presence of suitable buffers, ATP, Mg~2, and EDTA
(50 mM Tris, pH 7.6, 10 mM MgC12, 5 mM DDT 1-2 mM ATP, 0.1 m~1 spermidine, 0.1 mM EDTA, and 1.7 pmoles ~2P-ATP 12.9 mCi).
Ligations are formed using approximately equimolar a~ounts of the desired components, suitably end tailored to provide correct matching, by treabment with about 0.4-l Weiss units T4 DNA ligase per ~g vector DNA. Ligation mixtures are buffered at approx-30 imately pH 7.6 using 66 mM Tris along with 5 mM magne-sium ion, 5 mM dithiothreitol, 1 mM ATP, 0.1 mg/ml , BSA. For blunt ended ligation, 4-10 units of RNA
; ligase are added. Incubations are carried out at approximately 14C overnight. The foregoing describes ~ ~rc~le rn~rk conditions suitable for ligation of blunt ends; sticky end conditions are conducted as above, but can be somewhat milder, employing a lower concentration of ligase and ATP, as is understood in the art.
In vector construction, the vector fragment is commonly treated with bacterial alkaline phospha-tase (BAP) in order to remove the 5' phosphate and prevent self ligation of the vector. BAP digestions are conducted at pH 8 in approximately 150 mM Tris, in 10 the presence of Na+ and Mg-~2 using about 1 unit of BAP
per 1l9 of vector at 60 for about one hr. In order to recover the nucleic acid fragments, the preparation is extracted with phenol/chloroform and ethanol precipi-~r ~ tated and desalted by application to a Sephadex~G-50 -~ 15 spin column. Alternatively, the religation can be prevented by additional restriction of one of the fragments.
In the constructions set forth below, correct ligations for plasmicl construction are con-20 firmed by transforming E.coli strain MM294 (obtainedfrom the E.coli Genetic Stock C~nter, CGSC#6135) with the ligation mixture, unles~s the ~ phage PL promoter is used; in this case E.coli strain MC1000 Lambda SN7N53CI857SusP8o is used (ATCC 39531 deposited 25 December 2, 1983.) This strain is hereinafter referred to as MC1000-39531. Successful transformants are selected by ampicillin, tetracycline or other antibiotic resistance or using other markers depending on the mode of plasmid construction as is understood 30 in the art. Plasmids from the transformants are then prepared accordin~ to the method of Clewell, D.B., et al Proc Natl Acad Sci (1969) 62:1159, following chlor-amphenicol amplification (Clewell, D.F~., J Bacteriol (1972) 110:667) and analyzed by restriction and/or rc~ nc~

., .,, . _ , . , . , :

sequenced by the method of Messing, et al, Nucleic Acids Res, (1981) _:309 or by the method of Maxam, et al, Methods in Enz molo~ (1980) 65:499.
Transfonnations were performed using the 5 calcium chloride method described hy Cohen, S.N., et al, Proc Natl Acad Sc_(USA) (1972) 6g:2110.
Two host strains are used in cloning an expression of the plasmids set forth below:
For most constructions, E.coli strain MM294 10 (CGSC#61353, Talmadge, K., et al, Gene (1980) 12:235;
Meselson, M., et al, Nature (1968) 217:1110 is used as the host. However, when expression is under control of the PL pr~moter the E.coli strain MC1000 Lambcla SN7Ns3cIg57susp8o is used (ATCC 39531). This strain 15 contains a lambda prophage which codes for a tempera-ture sensitive CI repressor, which at the permissive temperature (30-34C) is active. However, at the non-permissive temperature (38-48C), the repressor is inactive and transcription from the PL promoter can 20 proceed. The N7 and N53 mutations prevent excision of the prophage from the chromosome and phage production is thus inhibited in this strain.

E.2 Isolation and Clonin~_of the MsE~l Fra~ment from the DT Gene DNA was isolated from corynephage 13ToX+
grown on Corynebacterum diphtheriae C7(-)tX-. (The host and phage are obtainable from J. Collier, University of California, Los Angeles; see Tweten, P~.K., et al, J Bacteriol (1983) 156:680.
3() To prepare DNA, high-titered B phage stocks were prepared in "TY~ medium" (15 g/l bactotryptone, 10 g/l yeast extract, 5 g/l NaCl supplemented with 1 mM CaC12J, by the method of Holmes, R.K., et al J

,,'~ '' ,, : ' ~7~0~

Virology (1969~ 38:586. Upon completion of lysis, debris was removed by centrifugation at 13,000 x g for 5 min, and MaCl added to 0.5 M, followed by PEG to 100 g/l, and the mixture was stirred overnight at 5 4C. The phage were concentrated by centrifugation at 13,000 x q for 15 min and resuspended in 100 mM
Tris HCl pH 7.5, lO0 mM NaCl, 20 mM EDTA. Pronase was added to 1 mg/ml and the mixture was incubated at 37C
for 2 hr. After removal of PEG by addition of potas-10 sium phosphate (dibasic:monobasic/2:1~ to 23% and cen-trifugation at 6,000 x g for 5 min, the lower phase was extracted with phenol, ethanol precipitated and the DNA purified by banding in a CsCl-EtBr gradient.
~pproximately 500 ~g of the phage DNA
15 (MW = 22 x 106 daltons) was treated with EcoRI and XbaI and the resulting mixture run on 1.7 liters 1%
agarose gel at 90 volts for 35 hr. The XbaI/EcoRI
fragment (1.5 x 106 daltons) containing the toxin gene was cut out, run through a syringe, and electroeluted 20 in l/lO TBE for 4 hrs at 500 volts onto a spectropore dialysis membrane. The DNA was retreived from the membrane using 0.25% SDS in l/lO T~E, phenol extrac-ted, ether extracted, ancl ethanol precipitated.
The resulting DNA was further restricted 25 with MspI, the DNA resolved on 5~ PAGE, and the two MspI fra~ments isolated by the crush and soak method.
The large Msp fraction (see Fig 1) which contained control sequences, leader, A, and partial B sequences from the toxin was cloned hy ligating approximately 30 5 ng of the fragment with 2 ~9 of ClaI-restricted, ~APed, pBR322. The ligation mixture was transformed into E.coli MM294, and the desired clones determined . . _ by isolation of plasmids, restriction analysis and sequencing. The desired cloning vector was designated i ~Z7~900 pMsp. Although this cloning was accomplished as above, constructions to provide the non-binding por-tion were obtained using phage directly as the source of Msp fraqment.

E 3. Synthesis a_d Cloning of Spacer Coding Sequence A ~NA fragment encoding the amino acid sequence Gly-Thr-Gly-Ser-Gly-(Pro)6-Ser-Gly-Ser-Gly-Thr-Cys 10 and flanked ~y sequences defining convenient restriction sites and a stop codon was designed and synthesized by conventional DNA synthesis procedures.

The sequence, AG CTT CCA GGC ACT GGA TCT GGC-Gly Thr Gly Ser Gly-CCG CCG CCA CCG CCG CCT TCT GGA TCC GGT ACC TGC TGA G
Pro Pro Pro Pro Pro Pro Ser Gly Ser Gly Thr Cys Stop and its complement were prepared using the triester method of Matteucci (supra) and annealed and kinased to give the double stranded sequence:

P-AGCTTCCAGGC-----------ACCTGCTGAG
AGGTCCG-----------TGGACGACTCAGCT-P
HindIII SalI
The annealed sequence was cloned as fol-lows: pBR322 (25.72 ~g) was restricted with SalI and HindIII, BAPed, phenol extracted ~nd desalte~ over a ' ""' ' ' :

' one cc Sepha~e ~G-50 column. The kinased annealed, double stranded spacer encoding sequence (0.2 pmoles) was ligated with l ~g of the plasmid vector fragment, and the ligation mixture was used to transform E.coli 5 strain MM294. AmpRTetS colonies were screened for plasmid size. The desired plasmid, pS~l was confirmed by restriction analysis and sequencing.

E.4. Cloning of Spacer onto Mspl Fraqment The plasmid, prlspSA2 which contains the MspI
10 fragment coding sequence ligated to the spacer coding sequence was contructed as outline~ in Fig 2.
pS~l plasmid DNA (77 ~g~ was restricted with SalI and ClaI, run on a 12% polyacrylamide gel and the fra~ment ccntaining the spacer arm sequence isolated 15 by the crush and soak metllod. One half the sample was further restricted with AluI to give "fragment A". As shown in ~ig 2, the Alu cleavage results in a blunt end 6 bp upstream from the glycine codon.
The large MspI fra~ment (10 ng) isolated 20 from phage as in E.2 was blunt ended by filling in with Klenow fragment and dNTPs. The mixture was run over a Sephadex~G-50 column~ treated with HindIII, and re run over a 1 cc Sephadex G-50 column to give "fragment B". As seen from Fig l, ilindIII restriction 25 deletes a portion of the DT control sequences, but probably leaves at least a portion of the promoter and ribosome binding site, and the entire leader sequence.
For the vector, 77 ~g pSAl was restricted with HindIII and SalI, treated with BAP, and the Q~
30 vector DNA fragment purifie~ with a 1 cc Sephadex G-50 column to qive "fragment C".
The ligation mixture consisted of 4 ~g of fragment C, 3 ng of fragment B, and 20 ng of fragment ~Tf~ r1lC

A under standard ligation conditions. Following liga-tion overnight at 12C, the mixture was transformed into E.coli strain MM294, and AmpRTetS colonies screened for plasmid size. The desired construct was identified hy restriction analysis and confirmed by Maxam-Gilbert sequencing. This plasmid, pMspSA2 contains, between the ~indIII and SalI sites of pBR322, a portion of the DT control sequence, leader sequence, A fragment, B fragment through the codon for lO amino acid 382, and the spacer arm codons.

E.5. ~eletion of the DT Promoter and Leader Se~uences ~
The preparation of pATGMspSA is outlined in Fig 3-~ pTrpSmlMbo ~55 ~g) was double digested with AccI and ClaI and the short fragment spanning the ATG
start codon and a portion of the A fragment isolated.
(See Figure 4 Eor relevant sequences in pTrpSml~Sbo anA
paragraph E.lO for its construction.) A vector fragment was prepared by digesting 25 IJg of pMspSA with ClaI, and treating with BAP. The missing portions of the Msp toxin fragment were sup-plied by a digest of 50 ~g of pMspSA with AccI and ClaI and isolating the 764 bp fragment between these sites in the coding sequence.
A ligation mixture containing 250 ng of the ATG-partial A fragment from pTrpSmlMbo, 700 ng of the partial A- partial ~ fragment from pMspSA and 2 ~g of the spacer-vector fragment from pMspSA was transformed into E.coli MM294 and AmpRTetS colonies selected. The correct construction was confirmed by restriction analysis.

~Z7~900 E.6. Preparation of Expression Vectors ppLMspsA
and ppLOPMspSA and Expression of the Toxin-Spacer .
Construct One expression vector, pPLMspSA was con-5 structed by inserting the appropriate portions of pATGMspSA behind the PL promoter. The construction is shown in Fig 3. pATGMspSA was digested with ~indIII, PstI, (EcoRI to prevent religation) anc3 the large vector fr~glrlent c~ ntaining the ATt~ start codon, the A
10 and B' DT to~in and spacer coding sequences used in subsequent ligation. This fragment was then ligate(1 with a PstI, HindIII, BAPed preparation of pFC5 (see parag. E.9, ~elow) which fragment corresponcls to the portion of pFC5 containing the PL promoter and N-gene 15 ribosome binding site. The ligation mixture was then used to transform MC1000-39351 and transformants selected by AmpR. The correct construction of the desired plasmid pPLMspSA was confirmed by restriction analysis.
A second vector, pPLOPMspSA was constructed by a two-way ligation of a fragment obtained by res-tricting pCS3 (See para. E.ll. below) with EcoRI, SalI
followed by BAP treatment and a fragment obtained from pPLMspSA by restriction with EcoRI, SalI and PstI.
(see Fig. 3). The ligation mixture was then used to transform ~1C1000-39351 and transformants selected by AmpR. The correct construction o~ the desired plasmid pPLOPMspSA was confirmed by restriction analysis an(l Maxam-Gilbert sequencing.
The colonies transformed with each of the foregoing plasmids were grown at 30C in TYE medium containing 100 ~g/ml ampicillin, and at the end of log phase the temperature raised to 42C. After 1.5 hr the cells were centrifuged and sonicated and the soni-cate assayed by the assay of Chung, D.W., et al, Infect Immun (1977) 16:832 for enzymatic activity. Activity corre~.ponding to DTA-B'-spacer at levels of 1-10 ~g/ml medium was found when pPLMspSA was u~ed and 20-150 ~g/ml when pPLOPMspSA was used.
Production of the desired DT A-B'-spacer was confirmed by Western Blot.

In a similar manner, expression vectors were constructed for DT-A-B'-Cys, an analogous cytotoxic portion containing a C-terminal cysteine. Details of -this construction are disclosed in copending Canadian serial no. 472,560, filed 22 January 198~, assigned to the same assignee. This vector, denoted there pPLOPMspCys, was used to transform E. coli MC1000, and production of the DT-A~B'-Cys protein confirmed by Western Blot.

E.7. Isolation_of DT-A-B'-spacer To produce sufficient protein for isolation, cultures of E. coli MC1000 transformed with, respectively, pPLOPMspRT (which encodes the DT-A-B' fragment alone; see Canadian serial no. 472,560, supra) PPLOPMSPCYS, and pPLOPMsPSA were grown in fermenters on minimal media supplemented with casamino acids and ampicillin using glucose as carbon source. Each fermenter was inoculated from a seed grown at 30C to an initial dry weight of 0.5 mg/l. When the OD6go reached 2-4, the temperature was rapidly shifted from 30 to 43C, and growth continued for an additional 5 hrs. The cells were harvested and stored at -70C. Analysis by DSD=PAGE showed estimates for yields of the desired proteins in the range of 500-1000 mg/l culture.

~1 ,, ~ ~ .
""' .'' -44_ E.7.a. DT-A-B ' Peptide Purlfication 10-20 g cells were resuspended in 20 ml 50 mM
Tris, 1 mM EDTA, pH 8.2 and sonicated. Following centrifugation, the supernatant was diluted approximately 5 5 fold in 5 mM Na phosphate, pH 6.8, 0.5 M NaCl and applied to a phenylsepharose column. The desired peptide eluted in approximately 5 mM Na phosphate, pH
6.8, giving a fraction of greater than 80% purity.

E.7.b. DT-A-B'-Cys Pe~tide A similar procedure attempted on the cells harboring PPLOPMSPCYS did not result in any soluble peptide. The desired peptide remained with the pelleted cells during centrifugation, and was not recoverable in the supernatant.

E.7.c._ DT-A-B'-spacer Sixteen grams of cells transformed with PPLOPMspSA were resuspended in 20 ml 50 m~ Tris, pH 8.2, 1 mM EDTA, 10 mM DDT and sonicated. The supernatant from centrifugation WAS diluted 5-10 fold in 5 mM Na 20 phosphate~ pH 6.8, 10 mM DTT and applied to a DEAE
~, ; Sephacel column. The protein was eluted using a 0-300 mM
NaCl gradient and 5 mM Na phosphate, p~ 6.8, 10 mM DTT.
The fractions containing the desired peptide were pooled, dialyzed against 5 mM Na phosphate, pH 6.8, 10 mM DTT and 25 loaded onto an NAD-Agarose~(P.L. Biochemical TYPEl) column.
Following elu~ion using a 0-1 M NaCl gradient in the same buffer, desired fractions were pooled, concentrated, and run over a Sephacryl~S-200 sizing column and the resulting fractions estimated to be 80% pure.

~'~QG~e ho~k !

~7~

E.8. Assay for Cytotoxicity The DT-A-B'-spacer was conjugated with antibreast monoclonal antibody 260F9, (hybridoma deposited at the ATCC on 27 January 1984 under accession number HB8488 and the conjugat~s were assayed for immunotoxicity. Controls utilized reduced ricin toxin A
chain (RTA) or diph~heria toxin A chain (DTA) which, therefore, contain free sulfhydryl groups for analogous conjugation with the antib~dy.

E.8 a. Conjugation of Cytotoxic Portion to A~tibody To orm the conjugate, breast monoclonal anti-body 260F9 or other antibodies as specified below were first derivatized with SPDP or Il - (CH2~ 5 - C - O ~ 503~a ~.~

(mal-sac-HNSA). The antibodies derivatized to SPDP were used to form disulfide links to the free cysteine sulfhydryls of DT-A-B'-spacer, DTA or RTA. Those derivatized with mal-sac-HNSA were used to form thioether linkages with DT-A-B'-spacer.
For SPDP, a 10-20 fold molar excess of SPDP was added to a solution containing 20 mg/ml of antibody in PBS and incubated at room temperature for 1 hr, and then dialyzed against PaS to remove unreacted SPDP. It was calculated that approximat21y 2-5 pyridyl-disulfide moieties were introduced into each antibody using this procedure.

~;27~

To complete the conjugation with SPDP to give a disulfide linkage to the cytotoxlc portions, DT-A-B'-spacer solution or solutlon of RTA or DTA containing 1-2 mg/ml which had been stored in reducing agent in 4C was !" 5 passed over a Sephadex G-25 column equilibrated in PBS to remove the reducing agent, and the DT-A-B'-spacer or other cytotoxic portion was mixed with derivatized antibody in 2-4 molar excess cytotoxic portion~ Conjugation was confirmed by spectrophotometric determination of released pyridine-2-thiol and by SDS-PAGE.
For mal-sac-HNSA, approximately 0.2 ml of mal-sac-HNSA solution containing 1 mg/ml was added to l ml antibody solution containing 3-8 mg/ml in PBS. The mixture was kept at room temperature and monitored until 5 mal-sac-HNSA moieties were incorporated per antibody.
The reaction was then stopped by desalting the mixture on a G-25 column equilibrated in 0.1 M Na phosphate, pH 6.
The DT-A-B'-spacer, stored in reducing agent at 4C, was passed over PBS-equilbiated Sephadex G-25 to remove reducing agent, and the protein (1-2 mg/ml) mixed in 2-4 molar excess with the derivatized antibody.
Conjugation was confirmed by SDS-PAGE.

E.8.b Assay In a typical protocol, breast tumor cells (MCF-7) were seeded in 8 ml glass vials and dilutions of the immunoconjugates were added. Following incubation for 22 hr at 37C, the medium was removed and replaced with medium containlng 35S methionine. Following a 2-hr pulse, the medium was aspirated, the monolayer was washed 3n twice with lO~ trichloroacetic acid containing l mg/ml ~2 .1 methionine and the vials were dried. ~ollowing~the add~tion of 3 m ~ of 4ad scintillation fluid containing 20% (v/v~ Triton X-100, the vials were counted.
Toxicity was expressed as the concentration of protein required to inhibit protein synthesis by 50% (TCID50)~
The results of these assays are shown in Table 1 both for 260F9, and other antibody partners.
:

Table 1 Monoclonal TCIDso% (nM) Antib~ Ab-DTA Ab-DT-A-B'-spacer Ab-RTA
10ATRl 10 2 0.1 260F9 30 0.3 0.1 106A10 ~ 100 7 208D112 ~ 100 40 4 245E73 ~ 100 50-100 10 lSMOPC214 ~ 100 ~ 100 ~ 100 1. positive control, anti-transferrin receptor antibody 2. hybridoma deposited 1/27/84 at ATCC, number 3. hybridoma deposited 1/27/84 at ATCC, number 4. negative control, purchased from Zymed Labs The foregoing assay was run as set forth above, but substituting alternate cell lines for MCF-7. The 25 results are shown in Table 2.
~ aC~ l~ark .. .-. .
.

~Z7~9~

Table 2 TCID50% (nM) RTA DT-A-B'-spacer Cell LineDisulfideDisulfide Thioether MCF-7 0.1 0O3 ND
CAMA-l 0.4 1.0 0.5 5BT-20 9 1.6 1.4 SKBR3 0.06 0.6 0.2 CC95 >100 ~100 ND

The cell lines shown, CAMA-1, BT-20, and SKBR-3 are other breast tumor cell lines: a normal fibroblast cell line CC-95 was also used. The DT-A-B'-spacer was comparably active with respect to the alternative breast tumor lines, but relatively inactive against the normal cells.

~.9 Construction oE Plasmids with a Portable PLNRBS EcoR l-Hind III Cassette .. . .
Three plasmids were constructed which can serve as sources for the EcoRI (or PstI) - ~indIII
Pl,NR~S cassette: pFC5, pPL322, and pPLKan.
For each of these plasmids, the DNA se~uence containing PL ~ phage promoter and the ribosome bind-ing site for the N-gene (NRBS) is obtained from a derivative of pKC30 described by Shimatake and Rosenberg, Nature (1981) 292:128. pKC30 contains a 2.34 kb fragment from ~ phage cloned into the HindIII-BamHI vector fragment from pBR322. The PL promoter and NRBS occupy a segment in pKC30 between a BglII and HpaI siteO

7~

The BglII site imme~iately preceding the PL
promoter was converted into an EcoRI site as fol-lows: pKC30 was digested with BglII, repaired with ~lenow and dNTPs, and li~ated with T4 ligase to an EcoRI linker (available from New England Biolabs) and transformed into E.coli MM294. Plasmids were isolated from AmpRTetS transformants and the desired sequence confirmed by restriction analysis and sequencing. The resulting plasmid, pFC3, was double-digested with PvuI
and HpaI to obtain an approximately 540 bp fragment framing the desired sequence. This fragment was par-tially digested with HinfI and the 424 bp fragment isolated and treated with Klenow and dATP, followed by Sl nuclease, to generate a blunt-ended fragment with 3' terminal sequence -AGG~GAA where the -AGGAGA
portion is the NRBS. This fragment was restricted ~L2~0~

with EcoR 1 to give a 347 base pair DNA fragment with 5'-EcoRI/Hinf(partial repair Sl blunt)-31 termini.
To obtain plasmids containing desire-l EcoRI/HindIII cassette containing PLNRBS~ the resul-5 ting fragment was ligated into an EcoRI/ilindIII
(repaired) cleaved pla.smid vector fragment obtaine-l from one of three such host plasmids: p~l-Z15, pBR322, and pDG144.
pBl-Z15, deposited January 13, 1984 10 1983 ATCC No. 39578, was prepared by fusing a sequence containing ATG plus 140 bp of t3l-IFN fused to lac Z into pBR322. In pBl-Z15 the EcoRI site of p~3R322 is re~-ained, and the insert contains a HindIII
site immediately preceding the ATG start codon.
15 pnl-Z15 was restricted with HindIII, repaired with Klenow and dNTP, and then digested with EcoRI. The resulting EcoRI/HindIII (repaired) vector fra~ment was ligated with the EcoRI/llinfI (repaired) fragment above. The li~ation mixture was used to transform 20 MC1000-39351 and transformants containing the success-ful construction were identified by ability to grow on lactose minimal plates at 34 but not at 30~. (Trans-formants were plated on X-gal Amp plates at 30 and 34 and minimal-lactose plates at 30 and 34. Trans-25 formants with the proper construction are blue onX-gal-Amp plates at both temperatures, but grow on minimal lactose plates only at 34.) The successful construct was designated pFC5.
In the alternative, pBR322 may also be used 30 as the cloning vector to carry the desired III PL NRBS cassette. pBR322 was digestecl with HindIII, repaired with Klenow and dNTPs, and then further digested with Ecor~I. The vector fragment was t}len ligated to the EcoRI/I~infI ~repaired) fragment ,: ,' . . . .
. ~ .

~a.z~

prepared above. The ligation mixture was then trans-formed into MM1000-39351, and successful transformants identified by AmpRTetS. Plasmids were isolated from successful transformants and a successful liqation confirmed by sequencing, and designated pPL322.

The thlrd hGst plasmid vector to obtain and provide the cassette was pDG144, deposited January 13, 1984 ! ATCC No. 39579. pDG144 is extensively described in another appllcation and does not constitute a part of the herein 10 invention. It is an altered pB~322 containing an intact AmpR gene, and a coding sequence for a protein con-ferring resistance to kanamycin (Kan ) preceding a synthetic polylinker. The polylinker seq~lence immediately preceding the ATG start codon for the 15 kanamycin gene can be removed by digesting with EcoRI
and HindIII and PLNRBS inserted.
Accordingly, pDG144 was digested wl~n HindIII, blunt-ended with Klenow and dNTPs, and then digested with EcoRI. The vector fragment was ligated 20 with the above-prepared EcoRI/~infI (repaired) frag-ment and transformed into MC1000-39351 AmpR Kan R
colonies were selected and plasmids isolated and the correct construction verified by restriction analysis and sequencing. One plasmid containing the correct 25 sequence was designated pPLKanO
Each of the above resulting vectors, pFC5, pPL322, and pPLKan, may be used to clone and provide the EcoRI/l~indIII PLNRBS cassette. The cassette can then conveniently be placed behind an ATG start codon 30 having a HindIII site immediately preceding it.

~L;27~0~

E.10 Construction of pTrpSmlMbo pTrpSmlMbo contains the DT-A fragment coding sequence followed ~y the Mbo terminator sequence (supra) under the control of the trp promoter. The 5 construction is ~rom pTS12 (a plasmid containing the DT-A and Mbo terminator) and pDG141, which contains the trp promoter (see Fig 4) pTS12 (53.5 ~g) was restricted with HhaI
blunt-ended with Klenow, 18 ~g of the resulting frag-10 ments ligated to 3.15 nmoles of the oligomeric linkerGCCCGGGG, and then treated with SmaI. The resultin~
sequence at~the 5' terminus was thus modified to give the sequence GGGGCTGA which encodes the peptide sequence beginning with amino acid 1 of the DT-A
fragment, (See Fi~ 5). The 3' end of the ligation product terminates in the first HhaI site of pBR322 following the SalI site, and the fragment contains the entire co~lin~ sequence along with in reading frame with terminator for the small Mbo fragment. The 20 desired 654 bp fragment was isolated using 6% PAGE, and elution by crush and soak.
One picomole of this modified prepared fragment was ligated with 0.7 ~g of pDG141 which had been restricted with SacI, blunt-ended with Klenow, and BAPed (the preparation of pDG141 is described ; below). The pDG141 derived fragment has an ATG start codon operably linked to the trp promoter. The resul-ting ligation mixture was transformed into E~coli MM294, an~ resistant colonies grown in 10 ml TYEI
medium containing 100 ~g/ml ampicillin and screened for plasmid size. Those colonies which cont~ined plasmids larger than pDG141 were screened ~or expres-sion o the DT-A fragment.

J

/

The cells were grown to log phase in 10 ml of the TYE-Amp 100 medium at 37 for 4 hr. To demon-strate expression, one ml of 'he culture was centri-fuged and the pellet resuspended in 20 ~1 buffer con-5 taining 625 mM Tris, pH 6.8, 3~ SDS. After heating at95C for 5 min, samples were run in 12.5% SDS-PAGE
with a 3~ stacking (Laemmli, et al, Nature (1970) 22_:680). Two clones which showed an additional pro-tein band at the expected molecule weight were con-firmed hy the EF-2 ADP-ribosylation assay according to the procedure of Chung, D.W., et al, Infect Immun (1977) 16:832. These colonies, designated ptrpSmlMbo, produced 20 ~r~ of DT-A per ml of culture. The anti-genicity and molecular weight of the product were confirmed by Western Blot.

E.lO.a Preparation of pDG141 pDG141, deposited 24 January, 1984, ATCC No-39588, contalnS
the trp control s~quences immediately upstream from an ATG start codon. The sequence downstream of the ~TG provides a SacI clea-vage site which cuts between thc G and the succeedingbp. Thus this plasmid con~ains a trp control sequence-ATP cassette excisable by digestion with PstI
or EcoRI and SacI. pBR322-Trp3 is used to provide a trp (PstI/HindIII repaired) cassette containing the promoter and R~S. p~W20 is used to provide an ATG
start codon followed by a SacI site.
pBR322-Trp3 (12 ng) restricted with PstI, and ~lindIII was ligated with 1.34 ng of similarly restricted pBW20. The ligation mixture was subse-quently ~igested with BamHI to linearize any ligationproducts which contained the HindIII/PstI unwanted vector fragment from p~R322-Trp3. The ligation mixture was used to transform E.coli MM294, and the ,:

~2~

desired colonies selected using L broth medium containing 50 ~g/ml ampicillin on plates pre-spread with 500 mg tryptophan. Correct construction was confirmed by sequencing.

E~l a.l. Construction of pBR322-trp pBR322-trp has the trp promoter/opera-tor/ribosome bindin~ site se~uence absent the attenu-ator region, and was obtained from pVH53 a plasmid obtained from C. Yanofsky, Stanford University~ A
number of other plasmids containing these control sequences are available as is known in the art.
p~153 was treated with HhaI (which cuts leaving an exposed 3' sticky end just 5' of the trp promoter) blunt ended with Klenow, and partially digested with TacI. The 99 bp fra~ment corresponding to restriction at the TacI site immediately preceding the ATG start codon of trp leader was isolated, and then ligated into EcoRI/ClaI digested, BAPed pBR322 to provide pBR322-Trp3~ The HindIII site immediately downstream from the pBR322 ClaI site permits excission of the deslred trp ragment as an EcoRI/ElindIII
cassette.

E.lO.a.2. The Construction of pBW20 pBW20 is a HindIII (repaired)/PvuII digest of pBR322 containing an insert of the double-stranded dodecamer: TATGAGCTCATAo This insert was made by ligating the blunt-ended fragments, transforming competent E.coli MM294, and selecting AmpRTetS colo-nies of appropriate construction as confirmed by sequencing. The sequence resulting in the region of the insert is as follows:

pBR322 dodecamer pBR322 --3 1 I <--TCGAT~AGCTTATGAGC~CATACTG
HindIII Sacl E.lO.b pTS12 The oligonucleotide GA TCT GTT GGC TCG AGT TGA
Arg Ser Val Gly Ser Ser Term which encodes the amino acid sequence subsequent to the Mbo cleavage site for six additional amino acids prior to a termination codon was synthesized using the 10 triester met~lod of Matteucci, et al (supra): kinased and hybridized to the complementary synthetic fragment to give 5' HO GATCTGTTGGCTCGAGTTGA
ACAACCGAGCTCAACTAGCT P
~glII SalI

One pmole of the double-stranded oligonucleotide was then placed in a three way ligation mixture with 1.4 pmoles (0.8 ~g) o Mbo fragment 1 and the vector fragment formed from 1 ~g pBR322 which had been treated with BamHI, SalI and BAP. The mixture was ligated overnight before transforming into E.col_ MM294. AmpRTetS colonies were selected and the desired construction confirmed by ~NA isolation, restriction analysis and sequencing. The correct plasmid was designated pTS12.

~-, .

~2~

E.ll Construction of PCs3 pCS3 i5 constructed from pOP9, a high copynumber derivative of pOP6 (Gelfand, D., et al, Proc Natl Acad Sci (USA) (1978) 75:5869), with a tempera-5 ture sensitive fragment from pEW27 which is describedby E.M. Wong, Proc Natl Acad Sci (USA) (1982) 79:3570.
The EcoRI/PvuII shorter fragment from pE~727 contains mutations in the region surrounding the origin of replication such that replication is inhibited at 10 lower temperatures but increased at high ones.
To construct to pCS3, pOP6 was first modi-fied through several steps: 50 llg of pOP6 was diges-ted to compl~tion with 20 units each of BamHI and SstI. In order to eliminate the SstI 3' protruding 15 ends and fill in the BamHI 5' ends, the digested pOP6 DNA was treated with E.coli DNA polymerase I (Klenow) in a two-stage reaction, first, at 20C for elimina-tion of the SstI protruling end and then at 0C for repair at the 5' end. This blunt-ended fragment was 20 ligated and 0.02 picomoles used to transform competent DG75 (O'Farrell, P., et al, J Bacteriol (1978) 134:645). Transformants were selected on L plates containing a 50 ~g/ml ampicillin and screened for a 3.3 kb deletion, loss of an Sst restriction endonu-25 ~lease site, and presence of a newly formed BamHIsite.
One candidate, designated pOP7 was chosen and BamHI site deleted by digesting 25 llg of pOP7 with 20 units BamHI an-l religating with T4 DI~A ligase.
30 Competent DG75 was treated with O.l ~g of the ligation mixture DNA, and transformants selected on plates con-taining 50 ~Jg/ml ampicillin. Candidates were screened for loss of the BamHI restriction site.

,, .

pOP8 was selected and modified to result in pOP9. The AvaI (repaired)/ EcoRI TetR fragment from pBR322 was isolated and ligated to the isolated PvuII
~partial)/EcoRI 3560 bp fragment from pOP8. Ligation 5 of the 1.42 kb EcoRI/AvaI (repaired) TetR (fragment A) and the 3.56 kb EcoRI/PvuII AmpR (fragment B) used 0.5 ~g of fragment B and 4.5 ~g fragment A in a two-stage reaction in order to favor intermolecular liga-tion of the EcoRI end~s. Competent DG75 was trans-10 formed with 5 ~1 of the ligation mixture, an~ trans-formants selected on ampicillin 50 ~g/ml or ampicillin or tetracycline (15 ~g/ml). pOP9 isolated from AmpRTetR transformants showed a high copy number, colicin resistance single restriction sites for EcoRI, 15 BamHI, PvuII and HindIII; two restriction sites for l-lincII, and the appropriate size and HaeIII digestion pattern.
50 ~Ig pEW27 DNA was digested to completion with PvuII and EcoRI. Similarly, 50 ~g of pOP9 was 20 digested to completion with PvuII and EcoRI and the 3.3 kb fragment isolated.
0.36 ~g (.327 picomoles) p~l27 fragment and 0.35 ~g (0.16 picomoles) pOP9 fragment were ligated and used to transform E.coll MM294. Ampicillin-tetra-25 cycline resistant transformants were selected. Suc-cessful colonies were screened at 30C and 40C on beta lactamase assay plates and then for plasmid DN~
levels following growth at 30C and 41C. Plasmids isolated from a colony showing improved ampR and increased plasmid DNA levels at the higher temperature were confirmed by restriction analysis and designated pCS3 . ~

Claims (3)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE PROPERTY OR
PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. A recombinant polypeptide, useful in making toxin conjugates, comprising:
a diptheria toxin, or an enzymatic fragment of diptheria toxin;
and a spacer covalently linked to said diphtheria toxin comprising an amino acid sequence which has the formula:
Gly-Thr-Gly-Ser-Gly-(Pro)6-Ser-Gly-Ser-Gly-Thr-Cys.

2. A conjugated toxin comprising the recombinant polypeptide of claim 1 covalently linked to a binding moiety wherein said binding moiety is covalently linked to said spacer amino acid sequence and comprises an antibody or fragment thereof which is capable of binding to an antigenic determinant.

3. A pharmaceutical composition effective in killing undesirable cells in mammals comprising a cytotoxically effective amount of a conjugated toxin comprising: an antibody or fragment thereof which binds to a target cell which comprises the corresponding antigenic determinant; and a recombinant protein comprising:
a spacer covalently bound to said antibody or fragment thereof comprising an amino acid sequence which has the formula:
Gly-Thr-Gly-Ser-Gly-(Pro)6-Ser-Gly-Ser-Gly-Thr-Cys;
and a cytotoxic protein fragment covalently bound to said spacer selected from the group consisting of diptheria toxin and enzymatically active fragments of diptheria toxin, in admixture with one or more pharmaceutically acceptable excipients.