Note: Descriptions are shown in the official language in which they were submitted.
~7~900
--1--
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
:. :
. .
:~ ., , ...... ; . ~.. ..
--3--
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
. .
., , . -
- ' : ' :' .
--5--
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-
. ...~ :
, , : ~ . ~.
':,,, ~
'' '
--7--
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~)
--10--
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
-12-
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~
-13-
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~
-14-
. 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~)~
-16-
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~
-17-
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
-18-
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~
--19--
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~
-20-
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
.
,
' ' " '
-21-
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.
-24-
.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~
-26-
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
,:
.
-27-
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
-28-
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,
-29-
r~
R-C(O)-O- ~ ~S3Na
N02
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~
-30-
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
-31-
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~
-32-
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
-34-
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
--36--
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~
., .,, . _ , . , . , :
--37--
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~
-38-
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
-39-
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
' ""' ' ' :
'
-40-
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
--42--
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-
-43-
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~
-45-
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~
-46-
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
-47-
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
HB8487
3. hybridoma deposited 1/27/84 at ATCC, number
HB8489
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~
-48-
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~
-49-
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~
--50--
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~
-51-
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~
-52-
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~
-54-
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:
-55-
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~
--56--
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
. ~
