Note: Descriptions are shown in the official language in which they were submitted.
CA 01341615 2011-05-24
1341615
TARGETED MULTIFUNCTIONAL PROTEINS
1 The United States Government has rights in
this application pursuant to small business
innovation research grant numbers SSS-4 R43
CA39870-01 and SSS-4 2 R44 CA39870-02.
Background of the Invention
This invention relates to novel compositions
of matter, hereinafter called targeted
multifunctional proteins, useful, for example, in
specific binding assays, affinity purification,
biocatalysis, drug targeting, imaging, immunological
treatment of various oncogenic and infectious
diseases, and in other contexts. More specifically,
this invention relates to biosynthetic proteins
expressed from recombinant DNA as a single
polypeptide chain comprising plural regions, one of
which has a structure similar to an antibody binding
site, and an affinity for a preselected antigenic
determinant, and another of which has a separate
function, and may be biologically active, designed to
30
CA 01341615 2011-05-24
2 -
1341615
bind to ions, or designed to facilitate
immobilization of the protein. This invention also
relates to the binding proteins per se, and methods
for their construction.
There are five classes of human antibodies.
Each has the sale basic structure (see Figure 1), or
multiple thereof, consisting of two identical
polypeptides called heavy (H) chains (molecularly
weight approximately 50,000 d) and two identical
light (L) chains (molecular weight approximately
25,000 d). Each of the five antibody classes has a
similar set of light chains and a distinct set of
heavy chains. A light chain is composed of one
variable and one constant domain, while a heavy chain
is composed of one variable and three or more
constant domains. The combined variable domains of a
paired light and heavy chain are known as the Fv
region, or simply "Fv". The Fv determines the
specificity of the immunoglobulin, the constant
regions have other functions.
Amino acid sequence data indicate that each
variable domain comprises three hypervariable regions
or loops, sometimes called complementarity
determining regions or "CDRs" flanked by four
relatively conserved framework regions or "FRs"
(Kabat et. al., Sequences of Proteins of
Immunological Interest [U.S. Department of Health and
Human Services, third edition, 1983, fourth edition,
1987]). The hypervariable regions have been assumed
to be responsible for the binding specificity of
individual antibodies and to account for the
diversity of binding of antibodies as a protein class.
CA 01341615 2011-05-24
3 -
1341615
Monoclonal antibodies have been used both as
diagnostic and therapeutic agents. They are
routinely produced according to established
procedures by hybridomas generated by fusion of mouse
lymphoid cells with an appropriate mouse myeloma cell
line.
The literature contains a host of references
to the concept of targeting bioactive substances such
as drugs, toxins, and enzymes to specific points in
the body to destroy or locate malignant cells or to
induce a localized drug or enzymatic effect. It has
been proposed to achieve this effect by conjugating
the bioactive substance to monoclonal antibodies
(see, e.g., Vogel, Immunoconiugates. Antibody
Conjugates in Radioimagina and Therapy of Cancer,
1987, N.Y., Oxford University Press; and Ghose et al.
(1978) J. Natl. Cancer Inst. 11:657-676, ). However,
non-human antibodies induce an immune response when
injected into humans. Human monoclonal antibodies
may alleviate this problem, but they are difficult to
produce by cell fusion techniques since, among other
problems, human hybridomas are notably unstable, and
removal of immunized spleen cells from humans is not
feasible.
Chimeric antibodies composed of human and
non-human amino acid sequences potentially have
improved therapeutic value as they presumably would
elicit less circulating human antibody against the
non-human immunoglobulin sequences. Accordingly,
hybrid antibody molecules have been proposed which
consist of amino acid sequences from different
mammalian sources. The chimeric antibodies designed
CA 01341615 2011-05-24
1341615
thus far comprise variable regions from one mammalian
source, and constant regions from human or another
mammalian source (Morrison et al. (1984) Proc. Natl. Acad.
Sci. U.S.A., 81:5851-6855; Neuberger et al. (1984) Nature
312:604-608; Sahagan et al. (1986) J. Immunol. 137:1066-
1074; EPO application Nos. EP 0 125 023, published on
November 14, 1984, Genetech; EP 0 171 496, published on
February 19, 1986, Research Development Corporation of
Japan; EP 0 173 494, published on March 5, 1986, Stanford;
Patent Publication WO 86/01533, published on March 13,
1986, Celltech Limited).
It has been reported that binding function is
localized to the variable domains of the antibodymolecule
located at the amino terminal end of both the heavy and
light chains. The variable regions remain noncovalently
associated (as VHVL dimers, termed Fv regions) even after
proteolytic clevage from the native antibody molecule, and
retain much of their antigen recognition and binding
capabilities (see, for example Inbar et al., Proc. Natl.
Acad. Sci. U.S.A. (1972) 69:2659-2662; Hochman et al.
(1973) Biochem. 12:1130-1135; and (1976) Biochem. 15:2706-
2710; Sharon and Givol (1976) Biochem. 15: 1591-1594;
Rosenblatt and Haber (1978) Biochem. 17:3877-3882; Ehrlich
et al. (1980) Biochem. 19:4091-40996). Methods of
manufacturing two-chain Fv substantially free of constant
region using recombinant DNA techniques are disclosed in
U.S. Patent 4,642,334, which issued on February 10, 1987.
CA 01341615 2011-05-24
-
1341615
Summary of the Invention
in one aspect the invention provides a
single chain multifunctional biosynthetic protein
expressed from a single gene derived by recombinant
.5 DNA techniques.. The protein comprises a biosynthetic
antibody binding site (GABS) comprising at least one
protein domain capable of binding to a preselected
antigenic determinant. The amino acid sequence of
,the domain is homologous to at least a portion of the
sequence of a variable region of an immunoglobulin
molecule capable of binding the preselected antigenic
determinant. Peptide bonded to the binding site is a
polypeptide consisting of an effector protein having
a conformation suitable for biological activity in a
mammal, an amino acid sequence capable of
sequestering ions, or an amino acid sequence capable
of selective binding to a solid support.
In another aspect, the invention provides
biosynthetic binding site protein comprising a single
polypeptide chain defining two polypeptide domains
connected by a polypeptide linker. The amino acid
sequence of each of the domains comprises a set of
complementarity determining regions (CDRs) interposed
between a set of framework regions (FRs), each of
which is respectively homologous with at least a
portion of the CDRs and FRS from an immunoglobulin
molecule. At least one of the domains comprises a
set of CDR amino acid sequences and a set of FR amino
acid sequences at least partly homologous to
different immunoglobulins. The two polypeptide
CA 01341615 2011-05-24
6 -
1341615
domains together define a hybrid synthetic binding
site having specificity for a preselected antigen,
determined by the selected CDRs.
In still another aspect, the invention
provides biosynthetic binding protein comprising a
single polypeptide chain defining two domains
connected by a polypeptide linker. The amino acid
sequence of each of the domains comprises a set of
CDRs interposed between a set of FRs, each of which
is respectively homologous with at least a portion of
the CDRs and FRs from an immunoglobulin molecule.
The linker comprises plural, peptide-bonded amino
acids defining a polypeptide of a length sufficient
to span the distance between the C terminal end of
one of the domains and N terminal end of the other
when the binding protein assumes a conformation
suitable for binding. The linker comprises
hydrophilic amino acids which together preferably
constitute a hydrophilic sequence. Linkers which
assume an unstructured polypeptide configuration in
aqueous solution work well. The binding protein is
capable of binding to a preselected antigenic site,
determined by the collective tertiary structure of
the sets of CDRs held in proper conformation by the
sets of FRs. Preferably, the binding protein has a
specificity at least substantially identical to the
binding specificity of the immunoglobulin molecule
used as a template for the design of the CDR
regions. Such structures can have a binding affinity
of at least 106, M-1, and preferably 108 M-1.
In preferred aspects, the FRs of the binding
protein are homologous to at least a portion of the
FRs from a human immunoglobulin, the linker spans at
CA 01341615 2011-05-24
7 -
1341615
least about 40 angstroms; a polypeptide spacer is
incorporated in the multifunctional protein between
the binding site and the second polypeptide; and the
binding protein has an affinity for the preselected
antigenic determinant no less than two orders of
magnitude less than the binding affinity of the
inununoglobulin molecule used as a template for the
CDR regions of the binding protein. The preferred
linkers and spacers are cysteine-free. The linker
preferably comprises amino acids having unreactive
side groups, e.g., alanine and glycine. Linkers and
spacers can be made by combining plural consecutive
copies of an amino acid sequence, e.g., (Gly4
Ser)3. The invention also provides DNAs encoding
these proteins and host cells harboring capable of
expressing these DNAs.
As used herein, the phrase biosynthetic
antibody binding site or BABS means synthetic
proteins expressed from DNA derived by recombinant
techniques. BABS comprise biosynthetically produced
sequences of amino acids defining polypeptides
designed to bind with a preselected antigenic
material. The structure of these synthetic
polypeptides is unlike that of naturally occurring
a5 antibodies, fragments thereof, e.g., Fv, or known
synthetic polypeptides or "chimeric antibodies" in
that the regions of the GABS responsible for
specificity and affinity of binding, (analogous to
native antibody variable regions) are linked by
peptide bonds, expressed from a single DNA, and may
themselves be chimeric, e.g., may comprise amino acid
sequences homologous to portions of at least two
CA 01341615 2011-05-24
1341615
different antibody molecules. The BABS embodying the
invention are biosynthetic in the sense that they are
synthesized in a cellular host made to express a synthetic
DNA, that is, a recombinant DNA made by ligation of plural,
chemically synthesized oligonucleotides, or by ligation of
fragments of DNA derived from the genome of a hybridoma,
mature B cell clone, or a cDNA library derived from such
natural sources. The proteins of the invention are
properly characterized as "binding sites" in that these
synthetic molecules are designed to have specific affinity
for a preselected antigenic determinant. The polypeptides
of the invention comprise structures patterned after
regions of native antibodies known to be responsible for,
antigen recognition.
As used herein, the phrase biofunctional domain
means a polypeptide domain, for example, an effector
protein having a confirmation suitable for biological
activity in vivo, such that the biological function of the
domain is maintained in embodiments of the present
invention.
Accordingly, it is an object of the invention to
provide novel multifunctional proteins comprising one or
more effector proteins and one or more biosynthetic
antibody binding sites, and to provide DNA sequences which
encode the proteins. Another object is to provide a
generalized method for producing biosynthetic antibody
binding site polypeptides of any desired specificity.
CA 01341615 2011-05-24
1341615
1 Brief Description of the Drawing
The foregoing and other objects of this
invention, the various features thereof, as well as
the invention itself, may be more fully understood
from the following description, when read together
with the accompanying drawings.
Figure 1A is a schematic representation of
an intact IgG antibody molecule containing two light
chains, each consisting of one variable and one
constant domain, and two heavy chains, each
consisting of one variable and three constant
domains. Figure 1B is a schematic drawing of the
structure of Fv proteins (and DNA encoding them)
illustrating VH and VL domains, each of which
comprises four framework (FR) regions and three
complementarity determining (CDR) regions.
Boundaries of CDRs are indicated, by way of example,
for monoclonal 26-10, a well known and characterized
murine monoclonal specific for digoxin.
Figure 2A-2E are schematic representations
of some of the classes of reagents constructed in
accordance with the invention, each of which
comprises a biosynthetic antibody binding site 3.
Figure 2A depicts a single chain protein construct
comprising a polypeptide domain 10 having an amino
acid sequence analogous to the variable region of an
immunoglobulin heavy chain, bound through its
carboxyl end to a polypeptide linker 12, which is in
turn bound to a polypeptide domain 14 having an amino
acid sequence analogous to the variable region of an
immunoglobulin light chain. The protein construct
also defines binding protein segment 2. In
Figure 2B, the protein construct also includes a
CA 01341615 2011-05-24
- 9A -
1341615
1 helically coiled polypeptide structure 16 linked to
the amino terminal end of domain 10 via spacer 18.
Figure 2C illustrates a bifunctional protein having
an effector polypeptide 20 linked via spacer 22 to
the carboxyl terminus.of polypeptide 14. Figure 2D
depicts a trifunctional protein having protein
domain 20 attached to the N-terminus of domain 10.
Figure 2E depicts a protein structure wherein
effector polypeptide 20 is attached through spacer 22
to polypeptide 14.
Figure 3 discloses five amino acid sequences
(heavy chains) in single letter code lined up
vertically to facilitate understanding of the
invention. Sequence 1 is the known native sequence
CA 01341615 2011-05-24
- 10 -
1341615
of VH from murine monoclonal glp-4
(anti-lysozyme). Sequence 2 is the known native
sequence of VH from marine monoclonal 26-10
(anti-digozin). Sequence 3 is a BABS comprising the
!Rs from 26-10 VH and the CDRs from glp-4 V0.
The CDRs are identified in lower case letters;
restriction sites in the DNA used to produce chimeric
sequence 3 are also identified. Sequence 4 is the
known native sequence of VH from human myeloma
antibody NEWM. Sequence 5 is a BABS comprising the
FRs from NEWM VH and the CDRs from glp-4 VH1
i.e., illustrates a "humanized" binding site having a
human framework but an affinity for lysozyme similar
to murine glp-4.
Figures 4A-4F are the synthetic nucleic acid
sequences and encoded amino acid sequences of (4A)
the heavy chain variable domain of murine
anti-digoxin monoclonal 26-10; (4B) the light chain
variable domain of murine anti-digoxin monoclonal
26-10; (4C) a heavy chain variable domain of a BABS
comprising CDRs of glp-4 and FRs of 26-10; (4D) a
light chain variable region of the same BABS; (4E) a
heavy chain variable region of a BABS comprising CDRs
of glp-4 and FRs of NEWM; and (4F) a light chain
variable region comprising CDRs of glp-4 and FRs of
NEWM. Delineated are FRs, CDRs, and restriction
sites for endonuclease digestion, most of which were
introduced during design of.the DNA.
CA 01341615 2011-05-24
-11- 1341615
Figure 5 is the nucleic acid and encoded
amino acid sequence of a.host DNA (VR) designed to
facilitate insertion of CDRs of choice. The DNA was
designed to have unique 6-base sites directly
flanking the CDRs so that relatively small
oligonucleotides defining portions of CDRs can be
readily inserted, and to have other sites to
facilitate manipulation of the DNA to optimize
binding properties in a given construct. The
framework regions of the molecule correspond to
murine FRs (Figure 4A).
Figures 6A and 6B are multifunctional
proteins. (and DNA encoding them) comprising a single
chain BABS with the specificity of murine monoclonal
26-10, linked through a spacer to the FB fragment of
protein A, here fused as a leader, and constituting a
binding site for Fc. The spacer comprises the 11
C-terminal amino acids of the FB followed by Asp-Pro
(a dilute acid cleavage site). The single chain BABS
comprises sequences mimicking the VH and VL (6A)
and the VL and VH (6B) of murine monoclonal
26-10. The VL in construct 6A is altered at
residue 4 where valine replaces methionine present in
the parent 26-10 sequence. These constructs contain
binding sites for both Fc and digoxin. Their
structure may be summarized as;
(6A) FB-Asp-Pro-VH (Gly4-Ser)3-VL,
and
(6B) FB-Asp-Pro-VL (Gly4-Ser)3-VH,
where (Gly4-Ser)3 is a polypeptide linker.
CA 01341615 2011-05-24
- 12 -
1341615
1 in Figures 4A-4E and 6A and 6B, the amino
acid sequence of the expression products start after
the GAATTC sequences, which codes for an EcoR1 splice
site. translated as Glu-Phe on the drawings.
Figure 7A is a graph of percent of undiluted
units bound versus concentration comparing the
binding of native 26-10 (curve 1) and the construct
of Figure 6A and Figure 2B renatured using two
different procedures (curves 2 and 3). Figure 7B is
a. graph demonstrating the bifunctionality of the FB
(26-10) BABS adhered to microtiter plates through the
specific binding of the binding site to the
digoxin-BSA coat on the plate. Figure 7B shows the.
percent inhibition of 125I-rabbit-IgG binding to
the FB domain of the FB BABS by the addition of IgG,
protein A, FB, murine IgG2a, and murine IgG1.
Figure 8 is a schematic representation of a
model assembled DNA sequences encoding a
multifunctional biosynthetic protein comprising a
leader peptide (used to aid expression and thereafter
cleaved), a binding site, a spacer, and an effector
molecule attached as a trailer sequence.
Figure 9A-9E are exemplary synthetic nucleic
acid sequences and corresponding encoded amino acid
sequences of binding sites of different
specificities: (A) FRs from NEWM and CDRs from 26-10
having the digoxin specificity of murine monoclonal
26-10; (B) FRs from 26-10, and CDRs from G-loop-4
(glp-4) having lysozyme specificity; (C) FRs and CDRs
from MOPC-315 having dinitrophenol (DNF) specificity;
CA 01341615 2011-05-24
- 13 -
1341615
(D) FRs and CDRs from an anti-CEA monoclonal
antibody; (8) FRs in both V. and VL and CDR1
and CDR3 in VR, and CDR1, CDR2, and CDR3 in
VL from an anti-CEA monoclonal antibody; CDR2 in
VR is a CDR2 consensus sequence found in most
immunoglobulin VB regions.
Figure 10A is a schematic representation of
the DNA and amino acid sequence of a leader peptide
(MLE) protein with corresponding DNA sequence and
some major restriction sites. Figure 10B shows the
design of an expression plasmid used to express
MLE-BABS (26-10). During construction of the gene,
fusion partners were joined at the EcoRl site that is
shown as part of the leader sequence. The pBR322
plasmid, opened at the unique SspI and PstI sites,
was combined in a 3-part ligation with an SspI to
EcoRI fragment bearing the tro promoter and MLE
leader and with an EcoRI to PstI fragment carrying
the BABS gene. The resulting expression vector
confers tetracycline resistance on positive
transformants.
Figure 11 is an SDS-polyacrylamide gel (15%)
of the (26-10) BABS at progressive stages of
purification. Lane 0 shows low molecular weight
standards; lane 1 is the MLE-BABS fusion protein;
lane 2 is an acid digest of this material; lane 3 is
the pooled DE-52 chromatographed protein; lanes 4 and
5 are the same oubain-Sepharose pool of single chain
BAGS except that lane 4 protein is reduced and lane 5
protein is unreduced.
*Trade Mark
CA 01341615 2011-05-24
- 14 -
1341615
Figure 12 shows inhibition curves for 26-10
BABE and 26-10 Fab species, and indicates the
relative affinities of the antibody fragment for the
indicated cardiac glycosides.
Figures 13A and 13B are plots of digozin
binding curves. (A) shows 26-10 BABE binding
isotherm and Sips plot (inset), and (B) shows 26-10
Fab binding isotherm and Sips plot (inset).
Figure 14 is a nucleic acid sequence and
corresponding amino acid sequence of a modified FB
dimer leader sequence and various restriction sites.
Figure 15A-15H are nucleic acid sequences
and corresponding amino acid sequences of
biosynthetic multifunctional proteins including a
single chain BABS and various biologically active
protein trailers linked via a spacer sequence. Also
indicated are various endonuclease digestion sites.
The trailing sequences are (A) epidermal growth
factor (EGF); (B) streptavidin; (C) tumor necrosis
factor (TNF); (D) calmodulin; (E) platelet derived
growth factor-beta (PDGF-beta); (F) ricin; and (G)
interleukin-2, and (H) an FB-FB dimer.
CA 01341615 2011-05-24
- 15 -
1341615
Description
The invention will first be described in its
broadest overall aspects with a more detailed
description following.
A class of novel biosynthetic, bi or
multifunctional proteins has now been designed and
engineered which comprise biosynthetic antibody
binding sites, that is, "BABS" or biosynthetic
polypeptides defining structure capable of selective
antigen recognition and preferential antigen binding,
and one-or more peptide-bonded additional protein or
polypeptide regions designed to have a preselected
property. Examples of the second region include
amino acid sequences designed to sequester ions,
which makes the protein suitable for use as an
imaging agent, and sequences designed to facilitate
immobilization of the protein for use in affinity
chromatography and solid phase immunoassay. Another
example of the second region is a bioactive effector
molecule, that is, a protein having a conformation
suitable for biological activity, such as an enzyme,
toxin, receptor, binding site, growth factor, cell
differentiation factor, lymphokine, cytokine,
hormone, or anti-metabolite. This invention features
synthetic, multifunctional proteins comprising these
regions peptide bonded to one or more biosynthetic
antibody binding sites, synthetic, single chain
proteins designed to bind preselected antigenic
determinants with high affinity and specificity,
constructs containing multiple binding sites linked
CA 01341615 2011-05-24
- 16 -
1341615
together to provide multipoint antigen binding and
high net affinity and specificity, DNA encoding these
proteins prepared by recombinant techniques, host
cells harboring these DNAs, and methods for the
production of these proteins and DNAa.
The invention requires recombinant
production of single chain binding sites having
affinity and specificity for a predetermined
antigenic determinant. This technology has been
developed and is disclosed herein. In view of this
disclosure, persons skilled in recombinant DNA
technology, protein design, and protein chemistry can
produce such sites which, when disposed in solution,
have high binding constants (at least 106,
preferably 108 M-1 ) and excellent specificity.
The design of the BABS is based on the
observation that three subregions of the variable
domain of each of the heavy and light chains of
native immunoglobulin molecules collectively are
responsible for antigen recognition and binding.
Each of these subregions, called herein
"complementarity determining regions" or CDRs,
consists of one of the hypervariable regions or loops
and of selected amino acids or amino acid sequences
disposed in the framework regions or FRs which flank
that particular hypervariable region. It has now
been discovered that FRs from diverse species are
effective to maintin CDRs from diverse other species
in proper conformation so as to.achieve true
immunochemical binding properties in a biosynthetic
protein. It has also been discovered that
CA 01341615 2011-05-24
- 17 -
1341615
1 biosynthetic domains mimicking the structure of the
two chains of an immunoglobulin binding site may be
connected by a polypeptide linker while closely
approaching, retaining, and often improving their
collective binding properties.
The binding site region of the
multifunctional proteins comprises at least one, and
preferably two domains, each of which has an amino
acid sequences homologous to portions of the CDRs of
the variable domain of an immunoglobulin light or
heavy chain, and other sequence homologous to the FRs
of the variable domain of the same, or a second,
different immunoglobulin light or heavy chain. The
two domain binding site construct also includes a
polypeptide linking the domains. Polypeptides so
constructed bind a specific preselected antigen
determined by the CDRs held in proper conformation by
the FRs and the linker. Preferred structures have
human FRs, i.e., mimic the amino acid sequence of at
least a portion of the framework regions of a human
immunoglobulin, and have linked domains which
together comprise structure mimicking a VH_VL or
VL-VH immunoglobulin two-chain binding site. CDR
regions of a mammalian immunoglobulin, such as those
of mouse, rat, or human origin are preferred. In one
preferred embodiment, the biosynthetic antibody
binding site comprises FRs homologous with a portion
of the FRs of a human immunoglobulin and CDRs
homologous with CDRs from a mouse or rat
immunoglobulin. This type of chimeric polypeptide
displays the antigen binding specificity of the mouse
or rat immunoglobulin, while its human framework
CA 01341615 2011-05-24
- 18 -.
1341615
minimizes human immune reactions. In addition, the
chimeric polypeptide may comprise other amino acid
sequences.. It may comprise, for example, a sequence
homologous to a portion of the constant domain of an
inmiunoglobulin, but preferably is free of constant
regions (other than !Rs).
The binding site region(s) of the chimeric
proteins are thus single chain composite polypeptides
comprising a structure which in solution behaves like
an antibody binding site. The two domain, single
chain composite polypeptide has a structure patterned
after tandem VH and VL domains, but with the
carboxyl terminal of one attached through a linking
amino acid sequence to the amino terminal of the
other. The linking amino acid sequence may or may
not itself be antigenic or biologically active. It
preferably spans a distance of at least about 40A,
i.e., comprises at least about 14 amino acids, and
comprises residues which together present a
hydrophilic, relatively unstructured region. Linking
amino acid sequences having little or no secondary
structure work well. Optionally, one or a pair of
unique amino acids or amino acid sequences
recognizable by a site specific cleavage agent may be
included in the linker. This permits the VH and
VL-like domains to be separated after expression,
or the linker to be excised after refolding of the
binding site.
Either the amino or carboxyl terminal ends
(or both ends) of these chimeric, single chain
binding sites are attached to an amino acid sequence
which itself is bioactive or has some other function
CA 01341615 2011-05-24
- 19 -
1341615
to produce a bifunctional or multifunctional
protein. For example, the synthetic binding site may
include a-leader and/or trailer sequence defining a
polypeptide having enzymatic activity, independent
affinity for an-antigen different from the antigen to
which the binding site is directed, or having other
functions such as to provide a convenient site of
attachment for a radioactive ion, or to provide a
residue designed to link chemically to a solid
support. This fused, independently functional
section of protein should be distinguished from fused
leaders used simply to enhance expression in
prokaryotic host cells or yeasts. The
multifunctional proteins also should be distinguished
from the "conjugates" disclosed in the prior art
comprising antibodies which, after expression, are
linked chemically to a second moiety.
Often, a series of amino acids designed as a
"spacer" is interposed between the active regions of
the multifunctional protein. Use of such a spacer
can promote independent refolding of the regions of
the protein. The spacer also may include a specific
sequence of amino acids recognized by an
endopeptidase, for example, endogenous to a target
cell (e-.g., one having a surface protein recognized
by the binding site) so that the bioactive effector
protein is cleaved and released at the target. The
second functional protein preferably is present as a
trailer sequence, as trailers exhibit less of a
tendency to interfere with the binding behavior of
the BABS.
CA 01341615 2011-05-24
-20-
1341615
The therapeutic use of such "self-targeted"
bioactive proteins offers a number of advantages over
conjugates of immunoglobulin fragments or complete
antibody molecules: they are stable, less
immunogenic and have a lower molecular weight; they
can penetrate body tissues more rapidly for purposes
of imaging or drug delivery because of their smaller
size; and they can facilitate accelerated clearance
of targeted isotopes or drugs. Furthermore, because
design of such structures at the DNA level as
disclosed herein permits ready selection of
bioproperties and specificities, an essentially
limitless combination of binding sites and bioactive
proteins is possible, each of which can be refined as
disclosed herein to optimize independent activity at
each region of the synthetic protein. The synthetic
proteins can be expressed in procaryotes such as
coli, and thus are less costly to produce than
immunoglobulins or fragments thereof which require
expression in cultured animal cell lines.
The invention thus provides a family of
recombinant proteins expressed from a single piece of
DNA, all of which have the capacity to bind
specifically with a predetermined antigenic
determinant. The preferred species of the proteins
comprise a second domain which functions
independently of the binding region. In this aspect
the invention provides an array of "self-targeted"
proteins which have a bioactive function and which
deliver that function to a locus determined by the
binding site's specificity. It also provides
biosynthetic binding proteins having attached
CA 01341615 2011-05-24
- 21 -
1341615
polypeptides suitable for attachment to
immobilization matrices which may be used in affinity
chromatography and solid phase immunoassay
applications, or suitable for attachment to ions,
e.g., radioactive ions, which may be used for in vivo
imaging.
The. successful design and manufacture of the
proteins of the invention depends on the ability to
produce biosynthetic binding sites, and most
preferably, sites comprising two domains mimicking
the variable domains of immunoglobulin connected by a
linker.
As is now well known, Fv, the minimum
antibody fragment which contains a complete antigen
recognition and binding site, consists of a dimer of
one heavy and one light chain variable domain in
noncovalent association (Figure 1A). It is in this
configuration that the three complementarity
determining regions of each variable domain interact
to define an antigen binding site on the surface of
the VH-VL dimer. Collectively, the six
complementarity determining regions (see Figure 1B)
confer antigen binding specificity to the antibody.
FRs flanking the CDRs have a tertiary structure which
is essentially conserved in native immunoglobulins of
species as diverse as human and mouse. These FRs
serve to hold the CDRs in their appropriate
orientation. The constant domains are not required
for binding function, but may aid in stabilizing
VH-VL interaction. Even a single variable domain
(or half of an Fv comprising only three CDRs specific
CA 01341615 2011-05-24
-22-
1341615
1 for an antigen) has the ability to recognize and bind
antigen, although at a lower affinity than an entire
binding site (Painter et al. (1972) Biochem.
21:1327-1337).
-5 This knowledge of the structure of
immunoglobulin proteins has now been exploited to
develop multifunctional fusion proteins comprising
biosynthetic antibody binding sites and one or more
other domains.
The structure of these biosynthetic proteins
in the region which impart the binding properties to
the protein is analogous to the Fv region of a
natural-antibody. It comprises at least one, and
preferably two domains consisting of amino acids
defining VH and VL-like polypeptide segments
connected by a linker which together form the
tertiary molecular structure responsible for affinity
and specificity. Each domain comprises a set of
amino acid sequences analogous to immunoglobulin CDRs
held in appropriate conformation by a set of
sequences analogous to the framework regions (FRs) of
an Fv fragment of a natural antibody.
The term CDR, as used herein, refers to
amino acid sequences which together define the
binding affinity and specificity of the natural Fv
region of a native immunoglobulin binding site, or a
synthetic polypeptide which mimics this function.
CDRs typically are not wholly homologous to
hypervariable regions of natural Fvs, but rather also
may include specific amino acids or amino acid
sequences which flank the hypervariable region and
have heretofore been considered framework not
CA 01341615 2011-05-24
- 23 -
1341615
1 directly determinitive of complementarity. The term
FR,.as used herein, refers to amino acid sequences
flanking or interposed between CDRs.
The CDR and FR polypeptide segments are
designed empirically based on sequence analysis of
the Tv.region of preexisting antibodies or of the DNA
encoding them. In one embodiment, the amino acid
sequences constituting the FR regions of the BARS are
analogous to the FR sequences of a first preexisting
antibody, for example, a human IgG. The amino acid
sequences constituting the CDR regions are analogous
to the sequences from a second, different preexisting
antibody, for example, the CDRs of a murine IgG.
Alternatively, the CDRs and FRs from a single
preexisting antibody from, e.g., an unstable or hard
to culture hybridoma, may be copied in their entirety.
Practice of the invention enables the design
and biosynthesis of various reagents, all of which
are characterized by a region having affinity for a
preselected antigenic determinant. The binding site
and other regions of the biosynthetic protein are
designed with the particular planned utility of the
protein in mind. Thus, if the reagent is designed
for intravascular use in mammals, the FR regions may
comprise amino acids similar or identical to at least
a portion of the framework region amino acids of
antibodies native to that mammalian species. On the
other hand, the amino acids comprising the CDRs may
be analogous to a portion of the amino acids from the
hypervariable region (and certain flanking amino
acids) of an antibody having a known affinity and
specificity, e.g., a murine or rat monoclonal
antibody.
CA 01341615 2011-05-24
- 24 -
1341615
Other sections of native immunoglobulin
protein structure, e.g., CH and CL, need not be
present and normally are intentionally omitted from
the biosynthetic proteins. However, the proteins of
the invention normally comprise additional
polypeptide or protein regions defining a bioactive
region, e.g., a toxin or enzyme, or a site onto which
a toxin or a remotely detectable substance can be
attached.
The invention thus can provide intact
biosynthetic antibody binding sites analogous to
VH-VL dimers, either non-covalently associated,
disulfide bonded, or preferably linked by a
polypeptide sequence to form a composite VH VL or
VL-VH polypeptide which may be essentially free
of antibody constant region. The invention also
provides proteins analogous to an independent VH or
VL domain, or dimers thereof. Any of these
proteins may be provided in a form linked to, for
example, amino acids analogous or homologous to a
bioactive molecule such as a hormone or toxin.
Connecting the independently functional
regions of the protein is a spacer comprising a short
amino acid sequence whose function is to separate the
functional regions so that they can independently
assume their active tertiary conformation. The
spacer can consist of an amino acid sequence present
on the end of a functional protein which sequence is
not itself required for its function, and/or specific
sequences engineered into the protein at the DNA
level.
CA 01341615 2011-05-24
-25-
1341615
The spacer generally may comprise between 5
and 25 residues. Its optimal length may be
determined using constructs of different spacer
lengths varying, for example, by units of 5 amino
acids.. The specific amino acids in the spacer can
vary. Cysteines should be avoided. Hydrophilic
amino acids are preferred. The spacer sequence may
mimic the sequence of a hinge region of an
immunoglobulin. It may also be designed to assume a
structure, such as a helical structure. Proteolytic
cleavage sites may be designed into the spacer
separating the variable region-like sequences from
other pendant sequences so as to facilitate cleavage
of intact BABS, free of other protein, or so as to
release the bioactive protein in vivo.
Figures 2A-2E illustrate five examples of
protein structures embodying the invention that can
be produced by following the teaching disclosed
herein. All are characterized by a biosynthetic
polypeptide defining a binding site 3, comprising
amino acid sequences comprising CDRs and FRs, often
derived from different immunoglobulins, or sequences
homologous to a portion of CDRs and FRs from
different immunoglobulins. Figure 2A depicts a
single chain construct comprising a polypeptide
domain 10 having an amino acid sequence analogous to
the variable region of an immunoglobulin heavy chain,
bound through its carboxyl end to a polypeptide
linker 12, which in turn is bound to a polypeptide
domain 14 having an amino acid sequence analogous to
CA 01341615 2011-05-24
-26-
1341615
1 the variable region of an immunoglobulin light
chain. Of course, the light and heavy chain domains
may be in reverse order. Alternatively, the binding
site may comprise two substantially homologous amino
.acid sequences which are.both analogous to the
variable region.of an immunoglobulin heavy or light
chain.
The linker 12 should be long enough (e.g.,
about 15 amino acids or about 40 A to permit the
chains 10 and 14 to assume their proper
conformation. The linker 12 may comprise an amino
acid sequence homologous to a sequence identified as
"self" by the species into which it will be
introduced, if drug use is intended. For example,
the linker may comprise an amino acid sequence
patterned after a hinge region of an immunoglobulin.
The linker preferably comprises hydrophilic amino
acid sequences. It may also comprise a bioactive
polypeptide such as a cell toxin which is to be
targeted by the binding site, or a segment easily
labelled by a radioactive reagent which is to be
delivered, e.g., to the site of a tumor comprising an
epitope recognized by the binding site. The linker
may also include one or two built-in cleavage sites,
i.e., an amino acid or amino acid sequence
susceptible to attack by a site specific cleavage
agent as described below. This strategy permits the
VH and VL-like domains to be separated after
expression, or the linker to be excised after folding
while retaining the binding site structure in
non-covalent association. The amino acids of the
CA 01341615 2011-05-24
- 27 -
1341615
linker preferably are selected from among those
having relatively small, unreactive side chains.
Alanine and glycine are preferred.
Generally, the design of the linker involves
considerations similar to the design of the spacer,
excepting that binding properties of the linked
domains are seriously degraded if the linker sequence
is shorter than about 20A in length, i.e., comprises
less than about 10 residues. Linkers longer than the
approximate 40A distance between the N terminal of a
native variable region and the C-terminal of its
sister chain may be used, but also potentially can
diminish the BABS binding properties. Linkers
comprising between 12 and 18 residues are preferred.,
The preferred length in specific constructs may be
determined by varying linker length first by units of
5 residues, and second by units of 1-4 residues after
determining the best multiple of the pentameric
starting units.
Additional proteins or polypeptides may be
attached to either or both the amino or carboxyl
termini of the binding site to produce
multifunctional proteins of the type illustrated in
Figures 2B-2E. As an example, in Figure 2B, a
helically coiled polypeptide structure 16 comprises a
protein A fragment (FB) linked to the amino terminal
end of a VH like domain 10 via a spacer 18. Figure
2C illustrates a bifunctional protein having an
effector polypeptide 20 linked via spacer 22 to the
carboxyl terminus of polypeptide 14 of binding
protein segment 2. This effector polypeptide 20 may
CA 01341615 2011-05-24
- 28 -
1341615
consist of, for example, a toxin, therapeutic drug,
binding protein, enzyme or enzyme fragment, site of
attachment for an imaging agent (e.g., to chelate a
6 radioactive ion such as indium), or site of selective
attachment to an immobilization matrix so that the
DABS. can be used in affinity chromatography or solid
phase binding assay. This effector alternatively may
be linked to the amino terminus of polypeptide 10,
although trailers are preferred. Figure 2D depicts a
trifunctional protein comprising a linked pair of
BABS.2 having another distinct protein domain 20
attached to the N-terminus of the first binding
protein segment. Use of multiple BABS in a single
protein enables production of constructs having very
high selective affinity for multiepitopic sites such
as cell surface proteins.
The independently functional domains are
attached by a spacer 18 (Figs 2B and 2D) covalently
linking the C terminus of the protein 16 or 20 to the
N-terminus of the first domain 10 of the binding
protein segment 2, or by a spacer 22 linking the
C-terminus of the second binding domain 14 to the
N-terminus of another protein (Figs. 2C and 2D). The
spacer maybe an amino acid sequence analogous to
linker sequence 12, or it may take other forms. As
noted above, the spacer's primary function is to
separate the active protein regions to promote their
independent bioactivity and permit each region to
assume its bioactive conformation independent of
interference from its neighboring structure.
CA 01341615 2011-05-24
- 29 -
1341615
Figure 2E depicts another type of reagent,
comprising a GABS having only one set of three CDRs,
e.g., analogous to a heavy chain variable region,
which retains a measure of affinity for the antigen.
Attached to the carbonyl end of the polypeptide 10 or
14 comprising the FR and CDR sequences constituting
the binding site 3 through spacer 22 is effector
polypeptide 20 as described above.
As is evidenced from the foregoing, the
invention provides a large family of reagents
comprising proteins, at least a portion of which
defines a binding site patterned after the variable
region of an immunoglobulin. It will be apparent
that the nature of any protein fragments linked to
the BABS, and used for reagents embodying the
invention, are essentially unlimited, the essence of
the invention being the provision, either alone or
linked to other proteins, of binding sites having
specificities to any antigen desired.
The clinical administration of
multifunctional proteins comprising a BABS, or a BABS
alone, affords a number of advantages over the use of
intact natural or chimeric antibody molecules,
fragments thereof, and conjugates comprising such
antibodies linked chemically to a second bioactive
moiety. The multifunctional proteins described
herein offer fewer cleavage sites to circulating
proteolytic enzymes, their functional domains are
connected by peptide bonds to polypeptide linker or
spacer sequences, and thus the proteins have improved
stability. Because of their smaller size and
efficient design, the multifunctional proteins
CA 01341615 2011-05-24
-30-
1341615
described herein reach their target tissue more
rapidly, and are cleared more quickly from the body.
They also have reduced immunogenicity. In addition,
their design facilitates coupling to other moieties
in drug targeting and imaging'application. Such
coupling may be conducted chemically after expression
of the BABS to a site of attachment for the coupling
product engineered into the protein at the DNA
level. Active effector proteins having toxic,
enzymatic, binding, modulating, cell differentiating,
hormonal, or other bioactivity are expressed from a
single DNA as a leader and/or trailer sequence,
peptide bonded to the BASS.
Design and Manufacture
The proteins of the invention are designed
at the DNA level. The chimeric or synthetic DNAs are
then expressed in a suitable host system, and the
expressed proteins are collected and renatured if
necessary. A preferred general structure of the DNA
encoding the proteins is set forth in Figure 8. As
illustrated, it encodes an optimal leader sequence
used to promote expression in procaryotes having a
built-in cleavage site recognizable by a site
specific cleavage agent, for example, an
endopeptidase, used to remove the leader after
expression. This is followed by DNA encoding a
VH-like domain, comprising CDRs and FRs, a linker,
a VL-like domain, again comprising CDRs and FRs, a
spacer, and an effector protein. After expression,
CA 01341615 2011-05-24
- 31 -
1341615
folding, and cleavage of the leader, a bifunctional
protein is produced having a binding region whose
specificity is determined by the CDRs, and a
peptide-linked independently functional effector
region.
The ability to design the NABS of the
invention depends on the ability to determine the
sequence of the amino acids in the variable region of
monoclonal antibodies of interest, or the DNA
encoding them. Hybridoma technology enables
production of cell lines secreting antibody to
essentially any desired substance that produces an
immune response. RNA encoding the light and heavy
chains of the immunoglobulin can then be obtained
from the cytoplasm of the hybridoma. The 5' end
portion of the mRNA can be used to prepare cDNA for
subsequent sequencing, or the amino acid sequence of
the hypervariable and flanking framework regions can
be determined by amino acid sequencing of the V
region fragments of the H and L chains. Such
sequence analysis is now conducted routinely. This
knowledge, coupled with observations and deductions
of the generalized structure of immunoglobulin Fvs,
permits one to design synthetic genes encoding FR and
CDR sequences which likely will bind the antigen.
These synthetic genes are then prepared using known
techniques, or using the technique disclosed below,
inserted into a suitable host, and expressed, and the
expressed protein is purified. Depending on the host
cell, renaturation techniques may be required to
attain proper conformation. The various proteins are
then tested for binding ability, and one having
CA 01341615 2011-05-24
-32-
1341615
appropriate affinity is selected for incorporation
into a reagent of the type described above. If
necessary, point substitutions seeking to optimize
binding may be made in the DNA using conventional
casette mutagenesis or other protein engineering
methodology such as is disclosed below.
Preparation of the proteins of the invention
also is dependent on knowledge of the amino acid
sequence (or corresponding DNA or RNA sequence) of
bioactive proteins such as enzymes, toxins, growth
factors,.cell differentiation factors, receptors,
anti-metabolites, hormones or various cytokines or
lymphokines. Such sequences are reported in the
literature and available through computerized data
banks.
The DNA sequences of the binding site and
the second protein domain are fused using
conventional techniques, or assembled from
synthesized oligonucleotides, and then expressed
using equally conventional techniques.
The processes for manipulating, amplifying,
and recombining DNA which encode amino acid sequences
of interest are generally well known in the art, and
therefore, not described in detail herein. Methods
of identifying and isolating genes encoding
antibodies of interest are well understood, and
described in the patent and other literature. In
general, the methods involve selecting genetic
material coding for amino acids which define the
proteins of interest, including the CDRs and FRs of
interest, according to the genetic code.
CA 01341615 2011-05-24
- 33 -
1341615
Accordingly, the construction of DNAs
encoding proteins as disclosed herein can be done
using known techniques involving the use of various
restriction enzymes which make sequence specific cuts
in DNA to produce blunt ends or cohesive ends, DNA
ligases, techniques enabling enzymatic addition of
sticky ends to blunt-ended DNA, construction of
synthetic DNAs by assembly of short or medium length
oligonucleotides, cDNA synthesis techniques, and
synthetic probes for isolating immunoglobulin or
other bioactive protein genes. Various promoter
sequences and other regulatory DNA sequences used in
achieving expression, and various types of host cells
are. also known and available. Conventional
transfection techniques, and equally conventional
techniques for cloning and subcloning DNA are useful
in.the practice of this invention and known to those
skilled in the art. Various types of vectors may be
used such as plasmids and viruses including animal
viruses and bacteriophages. The vectors may exploit
various marker genes which impart to a successfully
transfected cell a detectable phenotypic property
that can be used to identify which of a family of
clones has successfully incorporated the recombinant
DNA of the vector.
One method for obtaining DNA encoding the
proteins disclosed herein is.by assembly of synthetic
oligonucleotides produced in a conventional,
automated, polynucleotide synthesizer followed by
ligation with appropriate ligases. For example,
overlapping, complementary DNA fragments comprising
15 bases may be synthesized semi manually using
CA 01341615 2011-05-24
-34-
134161,5
phosphoramidite chemistry, with end segments left
unphosphorylated to prevent polymerization during
ligation. One and of the synthetic DNA is left with
a 'sticky end" corresponding to the site of action of
a particular restriction endonuclease, and the other
end is left with an end corresponding to the site of
action of another restriction'endonuclease.
Alternatively, this approach can be fully automated.
The DNA encoding the protein may be created by
synthesizing longer single strand fragments (e.g.,
50-100 nucleotides long) in, for example, a Biosearch
oligonucleotide synthesizer, and then ligating the
fragments.
A method of producing the BABS of the
invention is to produce a synthetic DNA encoding a
polypeptide comprising, e.g., human FRs, and
intervening "dummy" CDRs, or amino acids having no
function except to define suitably situated unique
restriction sites. This synthetic DNA is then
altered by DNA replacement, in which restriction and
ligation is employed to insert synthetic
oligonucleotides encoding CDRs defining a desired
binding specificity in the proper location between
the FRs. This approach facilitates empirical
refinement of the binding properties of the BABS.
This technique is dependent upon the ability
to cleave a DNA corresponding in structure to a
variable domain gene at specific sites flanking
nucleotide sequences encoding CDRs. These
restriction sites in some cases may be found in the
native gene. Alternatively, non-native restriction
sites may be engineered into the nucleotide sequence
CA 01341615 2011-05-24
-35-
1 '1615
1 resulting in a synthetic gene with a different
sequence of nucleotides than the native gene, but
encoding the same variable region amino acids because
of the degeneracy of the genetic code. The fragments
resulting from endonuclease digestion, and comprising
.FR-encoding sequences, are then ligated to non-native
CDR-encoding sequences to produce a synthetic
variable domain gene with altered antigen binding
specificity. Additional nucleotide sequences
encoding, for example, constant region amino acids or
a bioactive molecule may then be linked to the gene
sequences to produce a bifunctional protein.
The expression of these synthetic DNA's can
be achieved in both prokaryotic and eucaryotic
systems via transfection with an appropriate vector.
In .. coli and other microbial hosts, the synthetic
genes can be expressed as fusion protein which is
subsequently cleaved. Expression in eucaryotes can
be accomplished by the transfection of DNA sequences
encoding CDR and FR region amino acids and the amino
acids defining a second function into a myeloma or
other type of cell line. By this strategy intact
hybrid antibody molecules having hybrid Fv regions
and various bioactive proteins including a
biosynthetic binding site may be produced. For
fusion protein expressed in bacteria, subsequent
proteolytic cleavage of the isolated fusions can be
performed.to yield free BABS, which can be renatured
to obtain an intact biosynthetic, hybrid antibody
binding site.
CA 01341615 2011-05-24
-36-
1341615
1 Heretofore, it has not been possible to
cleave the heavy and light chain region to separate
the variable and constant regions of an
inwaunoglobulin so as to produce intact Iv, except in
specific cases not of com ercial utility. However,
one method of producing NABS in accordance with this
invention is to redesign DNAs encoding the heavy and
light chains of an immunoglobulin, optionally
altering its specificity or humanizing its FRs, and
incorporating a cleavage site and "hinge region"
between the variable and constant regions of both the
heavy and light chains. Such chimeric antibodies can
be produced in transfectomas or the like and
subsequently cleaved using a preselected
endopeptidase.
The hinge region is a sequence of amino
acids which serve to promote efficient cleavage by a
preselected cleavage agent at a preselected, built-in
cleavage site. It is designed to promote cleavage
preferentially at the cleavage site when the
polypeptide is treated with the cleavage agent in an
appropriate environment.
The hinge region can take many different
forms. Its design involves selection of amino acid
residues (and a DNA fragment encoding them) which
impart to the region of the fused protein about the
cleavage site an appropriate polarity, charge
distribution, and stereochemistry which, in the
aqueous environment where the cleavage takes place,
efficiently exposes the cleavage site to the cleavage
agent in preference to other potential cleavage sites
that may be present in the polypeptide, and/or to
CA 01341615 2011-05-24
-37-
1341615
improve the kinetics of the cleavage reaction. In
.specific cases, the amino acids of the hinge are
selected and assembled in sequence based on their
known properties, and then the fused polypeptide
sequence is expressed. tested, and altered for
refinement.
The hinge region is free of cysteine. This
enables the cleavage reaction to be conducted under
conditions in which the protein assumes its tertiary
conformation, and may be held in this conformation by
intramolecular disulfide bonds. It has been
discovered that in these conditions access of the
protease to potential cleavage sites which may be
present within the target protein is hindered. The
hinge region may comprise an amino acid sequence
which includes one or more proline residues. This
allows formation of a substantially unfolded
molecular segment. Aspartic acid, glutamic acid,
arginine, lysine, serine, and threonine residues
maximize ionic interactions and may be present in
amounts and/or in sequence which renders the moiety
comprising the hinge water soluble.
The cleavage site preferably is immediately
adjacent the Fv polypeptide chains and comprises one
amino acid or a sequence of amino acids exclusive of
any one or sequence found in the amino acid structure
of the chains in the Fv. The cleavage site
preferably is designed for cleavage by a specific
selected agent. Endopeptidases are preferred,
although non-enzymatic (chemical) cleavage agents may
be used. Many useful cleavage agents, for instance,
cyanogen bromide, dilute acid, trypsin,
CA 01341615 2011-05-24
-38-
1341616
Staphylococcus aureus V-8 protease, post proline
cleaving enzyme, blood coagulation Factor Xa,
enterokinase, and resin, recognize and preferentially
or exclusively cleave particular cleavage sites. One
currently preferred cleavage agent is V-8 protease.
The currently preferred cleavage site is a Glu
residue. Other useful enzymes recognize multiple
residues as a cleavage site, e.g., factor Xa
(Ile-Glu-Gly-Arg) or enterokinase
(Asp-Asp-Asp-Asp-Lys). The principles of this
selective cleavage approach may also be used in the
design of the linker and spacer sequences of the
multifunctional constructs of the invention where an
exciseable linker or selectively cleavable linker or
spacer is desired.
Design of Synthetic V.. and V, Mimics
FRs from the heavy and light chain murine
anti-digoxin monoclonal 26-10 (Figures 4A and 4B)
were encoded on the same DNAs with CDRs from the
murine anti-lysozyme monoclonal glp-4 heavy chain
(Figure 3 sequence 1) and light chain to produce VH
(Figure 4C) and VL (Figure 4D) regions together
defining a biosynthetic antibody binding site which
is specific for lysozyme. Murine CDRs from both the
heavy and light chains of monoclonal glp-4 were
encoded on the same DNAs with FRs from the heavy and
light chains of human myeloma antibody NEWM (Figures
4E and 4F). The resulting interspecies chimeric
antibody binding domain has reduced immunogenicity in
humans because of its human FRs, and specificity for
lysozyme because of its murine CDRs.
CA 01341615 2011-05-24
-39-
1341615
A synthetic DNA was designed to facilitate
CDR insertions into a human heavy chain FR and to
facilitate empirical refinement of the resulting
chimeric amino acid sequence. This DNA is depicted.
in Figure S.
A synthetic, bifunctional FB-binding site
protein was also designed at the DNA level,
expressed, purified, renatured, and shown to bind
specifically with a preselected antigen (digoxin) and
Fc. The-detailed primary structure of this construct
is shown in Figure 6; its tertiary structure is
illustrated schematically in Figure 2B.
Details of these and other experiments, and
additional design principles on which the invention
is based, are set forth below.
GENE DESIGN AND EXPRESSION
Given known variable region DNA sequences,
synthetic VL and VH genes may be designed which
encode native or near native FR and CDR amino acid
sequences from an antibody molecule, each separated
by unique restriction sites located as close to
FR-CDR and CDR-FR borders as possible.
Alternatively, genes may be designed which encode
native FR sequences which are similar or identical to
the FRs of an antibody molecule from a selected
species, each separated by "dummy" CDR sequences
containing strategically located restriction sites.
These DNAs serve as starting materials for producing
BABS, as the native or "dummy" CDR sequences may be
excised and replaced with sequences encoding the CDR
CA 01341615 2011-05-24
- 40 -
1341615
amino acids defining a selected binding site.
Alternatively, one may design and directly synthesize
native or near-native FR sequences from a first
antibody molecule, and CDR sequences from a second
antibody molecule. Any one of the VR and VL
sequences described above may be linked together
directly, via an amino acids chain or linker
connecting the C-terminus of one chain with the
N-terminus of the other.
These genes, once synthesized, may be cloned
with or without additional DNA sequences coding for,
e.g., an antibody constant region, enzyme, or toxin,
or a leader peptide which facilitates secretion or
intracellular stability of a fusion polypeptide. The
genes then can be expressed directly in an
appropriate host cell, or can be further engineered
before expression by the exchange of FR, CDR, or
"dummy" CDR sequences with new sequences. This
manipulation is facilitated by the presence of the
restriction sites which have been engineered into the
gene at the FR-CDR and CDR-FR borders.
Figure 3 illustrates the general approach to
designing a chimeric VH; further details of
exemplary designs at the DNA level are shown in
Figures_4A-4F. Figure 3, lines 1 and 2, show the
amino acid sequences of the heavy chain variable
region of the murine monoclonals glp-4
(anti-lysozyme) and 26-10 (anti-digoxin), including
the four FR and three CDR sequences of each. Line 3
shows the sequence of a chimeric VH which comprises
26-10 FRs and glp-4 CDRs. As illustrated, the hybrid
protein of line 3 is identical to the native protein
CA 01341615 2011-05-24
- 41 -
1341615
of line 2, except that 1) the sequence TFTNYYIHWLK
has replaced the sequence IFTDFYMNWVR, 2)
EWIGWIYPGNGNTKYNENFKG has replaced
DYIGYISPYSGVTGYNQKFKG, 3) RYTHYYF has replaced
GSSGNKWAM, and 4) A has replaced V as the sixth amino
acid beyond CDR-2. These changes have the effect of
changing the specificity of the 26-10 VH to mimic
the specificity of glp-4. The Ala to Val single
amino acid replacement within the relatively
conserved framework region of 26-10 is an example of
the replacement of an amino acid outside the
hypervariable region made for the purpose of altering
specificity by CDR replacement. Beneath sequence 3
of Figure 3, the restriction sites in the DNA
encoding the chimeric VH (see Figures 4A-4F) are
shown which are disposed about the CDR-FR borders.
Lines 4 and 5 of Figure 3 represent another
construct. Line 4 is the full length VH of the
human antibody NEWM. That human antibody may be made
specific for lysozyme by CDR replacement as shown in
line 5. Thus, for example, the segment TFTNYYIHWLK
from glp-4 replaces TFSNDYYTWVR of NEWM, and its
other CDRs are replaced as shown. This results in a
VH comprising a human framework with murine
sequences determining specificity.
By sequencing any antibody, or obtaining the
sequence from the literature, in view of this
disclosure one skilled in the art can produce a BABS
of any desired specificity comprising any desired
framework region. Diagrams such as Figure 3
comparing the amino acid sequence are valuable in
suggesting which particular amino acids should be
CA 01341615 2011-05-24
-42-
1341615
replaced to determine the desired complementarity.
Expressed sequences may be tested for binding and
refined by exchanging selected amino acids in
relatively conserved regions, based on observation of
trends in amino acid sequence data and/or computer
modeling techniques.
Significant flexibility in VH and VL
design is possible because the amino acid sequences
are determined at the DNA level, and the manipulation
of DNA can be accomplished easily.
For example, the DNA sequence for murine VH
and VL 26-10 containing specific restriction sites
flanking each of the three CDRs was designed with the
aid of a commercially available computer program
which performs combined reverse translation and
restriction site searches ("RV.exe" by Compugene,
Inc.). The known amino acid sequences for VH and
VL 26-10 polypeptides were entered, and all
potential DNA sequences which encode those peptides
and all potential restriction sites were analyzed by
the program. The program can, in addition, select
DNA sequences encoding the peptide using only codons
preferred by L._. Coll if this bacterium is to be host
expression organism of choice. Figures 4A and 4B
show an example of program output. The nucelic acid
sequences of the synthetic gene and the corresponding
amino acids are shown. Sites of restriction
endonuclease cleavage are also indicated. The CDRs
of these synthetic genes are underlined.
CA 01341615 2011-05-24
-43-
1341615
1
The DNA sequences for the synthetic 26-10
VN,and VL are designed so that one or both of the
restriction sites flanking each of the three CDRs are
unique. A six base site (such as that recognized by
Barn I or BspM I) is preferred, but where six base
sites are not possible, four or five base sites are
used. These sites, if not already unique, are
rendered unique within the gene by eliminating other
occurrences within the gene without altering
necessary amino acid sequences. Preferred cleavage
sites are those that, once cleaved, yield fragments
with sticky ends just outside of the boundary of the
CDR within the framework. However, such ideal sites
are only occasionally possible because the FR-CDR
boundary is not an absolute one, and because the
amino acid sequence of the FR may not permit a
restriction site. In these cases, flanking sites in
the FR which are more distant from the predicted
boundary are selected.
Figure 5 discloses the nucleotide and
corresponding amino acid sequence (shown in standard
single letter code) of a synthetic DNA comprising a
master framework gene having the generic structure:
Rl-FR1-X1-FR2-X2-FR3-X3-FR4-R2
where R1 and R2 are restricted ends which are to
be ligated into a vector, and X1, X2, and X3
are DNA sequences whose function is to provide
convenient restriction sites for CDR insertion. This
particular DNA has murine FR sequences and unique,
6-base restriction sites adjacent the FR borders so
CA 01341615 2011-05-24
- 4'1 -
1341615
1 that nucleotide sequences encoding CDRs from a
desired monoclonal can be inserted easily.
Restriction endonuclease digestion sites are
indicated with their abbreviations; enzymes of__choice
for CDR replacement are underscored. Digestion of
the gene with the following restriction endonucleases
results in 3' and 5' ends which can easily be matched
up with and ligated to native or synthetic CDRs of
desired specificity; KpnI and BstXI are used for
ligation of CDR1; XbaI and Dral for CDR2; and
BssHII and C1aI for CDR3.
OLIGONUCLEOTIDE SYNTHESIS
The synthetic genes and DNA fragments
designed as described above preferably are produced
by assembly of chemically synthesized
oligonucleotides. 15-100mer oligonucleotides may be
synthesized on a Biosearch*DNA Model 8600
Synthesizer, and purified by polyacrylamide gel
electrophoresis (PAGE) in Tris-Borate-EDTA buffer
(TBE). The DNA is then electroeluted from the gel.
Overlapping oligomers may be phosphorylated by T4
polynucleotide kinase and ligated into larger blocks
which may also be purified by PAGE.
CLONING OF SYNTHETIC OLIGONUCLEOTIDES
The blocks or the pairs of longer
oligonucleotides may be cloned into E. coli using a
suitable, e.g., pUC, cloning vector. Initially, this
vector may be altered by single strand mutagenesis to
*Trade Mark
CA 01341615 2011-05-24
-45-
1341615
eliminate residual six base altered sites. For
example, VH may be synthesized and cloned into pUC
as.five primary blocks spanning the following
restriction sites: 1. gcoRI to first Nari site; 2.
first Narl to XbaI; 3. Xbal to Sall; 4. Sall to Ecol;
5. NcoI to BamBI. These cloned fragments may then be
isolated and assembled in several three-fragment
ligations and cloning steps into the pUC8 plasmid.
Desired ligations selected by PAGE are then
transformed into, for example, coil strain JM83,
and plated onto LB Ampicillin + Xgal plates according
to.standard procedures. The gene sequence may be
confirmed by supercoil sequencing after cloning, or
after subcloning into M13 via the dideoxy method of
Sanger.
PRINCIPLE OF CDR EXCHANGE
Three CDRs (or alternatively, four FRs) can
be replaced per VH or VL. In simple cases, this
can be accomplished by cutting the shuttle pUC
plasmid containing the respective genes at the two
unique restriction sites flanking each CDR or FR,
removing the excised sequence, and ligating the
vector with a native nucleic acid sequence or a
synthetic oligonucleotide encoding the desired CDR or
FR. This three part procedure would have to be
repeated three times for total CDR replacement and
four times for total FR replacement. Alternatively,
a synthetic nucleotide encoding two consecutive CDRs
separated by the appropriate FR can be ligated to a
pUC or other plasmid containing a gene whose
CA 01341615 2011-05-24
-46-
1341615
1 corresponding CDRs and FR have been cleaved out.
This procedure reduces the number of steps required
to perform CDR and/or FR exchange.
EXPRESSION OF PROTEINS
The engineered genes can be expressed in
appropriate prokaryotic hosts such as various strains
of coli, and in eucaryotic hosts such as Chinese
hamster ovary cell, murine myeloma, and human
myeloma/transfectoma cells.
For example, if the gene is to be expressed
in coli, it may first be cloned into an expression
vector. This is accomplished by positioning the
engineered gene downstream from a promoter sequence
such as trp or tac, and a gene coding for a leader
peptide. The resulting expressed fusion protein
accumulates in refractile bodies in the cytoplasm of
the cells, and may be harvested after disruption of
the cells by French press or sonication. The
refractile bodies are solubilized, and the expressed
proteins refolded and cleaved by the methods already
established for many other recombinant proteins.
If the engineered gene is to be expressed in
myeloma cells, the conventional expression system for
immunoglobulins, it is first inserted into an
expression vector containing, for example, the Ig
promoter, a secretion signal, immunoglobulin
enhancers, and various introns. This plasmid may
also contain sequences encoding all or part of a
constant region, enabling an entire part of a heavy
or light, chain to be expressed. The gene is
CA 01341615 2011-05-24
1341615
1 transfected into myeloma cells via established
electroporation or protoplast fusion methods. Cells
so transfected can express VL or VH fragments,
VI-12 or VH2 homodimers, VL-VH heterodimers,
VH-VL,or VL-VHsingle chain polypeptides,
complete heavy or light immunoglobulin chains, or
portions thereof, each of which may be attached in
the various ways discussed above to a protein region
having another function (e.g., cytotoxicity).
Vectors containing a heavy chain V region
(or V and C regions) can be cotransfected with
analogous vectors carrying a light chain V region (or
V and C regions), allowing for the expression of
noncovalently associated binding sites (or complete
antibody molecules).
In the examples which follow, a specific
example of how to make a single chain binding site is
disclosed, together with methods employed to assess
its binding properties. Thereafter, a protein
construct having two functional domains is
disclosed. Lastly, there is disclosed a series of
additional targeted proteins which exemplify the
invention.
I EXAMPLE OF CDR EXCHANGE AND EXPRESSION
The synthetic gene coding for murine VH
and VL 26-10 shown in Figures 4A and 4B were
designed from the known amino acid sequence of the
protein with the aid of Compugene* a software
program. These genes, although coding for the native
amino acid sequences, also contain non-native and
*Trade Mark
CA 01341615 2011-05-24
- 48 -
1341615
1 often unique restriction sites flanking nucleic acid
sequences encoding CDR's to facilitate CDR
replacement as noted above.
Both the 3' and 5' ends of the large
synthetic oligomers were designed to include 6-base
restriction sites, present in the genes and the pUC
vector. Furthermore, those restriction sites in the
synthetic genes which were only suited for assembly
but not for cloning the pUC were extended by "helper"
cloning sites with matching sites in pUC.
Cloning of the synthetic DNA and later
assembly of the gene is facilitated by the spacing of
unique restriction sites along the gene. This allows
corrections and modifications by cassette mutagenesis
at any location. Among them are alterations near the
5' or 3' ends of the gene as needed for the
adaptation to different expression vectors. For
example, a PstI site is positioned near the 5' end of
the VH gene. Synthetic linkers can be attached
easily between this site and a restriction site in
the expression plasmid. These genes were synthesized
by assembling oligonucleotides as described above
using.a Biosearch Model 8600 DNA Synthesizer. They
were ligated to vector pUC8 for transformation of
coli.
Specific CDRs may be cleaved from the
synthetic VH gene by digestion with the following
pairs of restriction endonucleases: HpHI and BstXI
for CDR1; XbaI and Dral for CDR2; and Banli and
BanI for-CDR 3. After removal on one CDR, another
CDR of desired specificity may be ligated directly
CA 01341615 2011-05-24
- 49 -
341fi15
into the restricted gene, in its place if the 3' and
5' ends of the restricted gene and the new CDR
contain complementary single stranded DNA sequences.
In the present example, the three CDRs of
each. of murine VR 26-10 and VL 26-10 were
replaced with the corresponding CDRs of glp-4. The
nucleic acid sequences and corresponding amino acid
sequences of the chimeric VH and VL genes
encoding the FRs of 26-10 and CDRs of glp-4 are shown
in Figures 4C and 4D. The positions of the
restriction endonuclease cleavage sites are noted
with their standard abbreviations. CDR sequences are
underlined as are the restriction endonucleases of
choice useful for further CDR replacement.
These genes were cloned into pUC8, a shuttle
plasmid. To retain unique restriction sites after
cloning, the VH like gene was spliced into the
EcoRl and Hindlil or BamHI sites of the plasmid.
Direct expression of the genes may be
achieved in coli. Alternatively, the gene may be
preceded by a leader sequence and expressed in L
soli as a fusion product by splicing the fusion gene
into the host gene whose expression is regulated by
interaction of a repressor with the respective
operator. The protein can be induced by starvation
in minimal medium and by chemical inducers. The
VH-VL biosynthetic 26-10 gene has been expressed
as such a fusion protein behind the trp and tac
promoters. The gene translation product of interest
may then be cleaved from the leader in the fusion
protein by e.g., cyanogen bromide degradation,
tryptic digestion, mild acid cleavage, and/or
CA 01341615 2011-05-24
-50-
1341615
1 digestion with factor Xa protease. Therefore, a
shuttle plasmid containing a synthetic gene encoding
a leader peptide having a site for mild acid
cleavage, and into which has been spliced the
synthetic NABS gene was used for this purpose. In
addition, synthetic DNA sequences encoding a signal
peptide for secretion of the processed target protein
into the periplasm of the host cell can also be
incorporated into the plasmid.
After harvesting the gene product and
optionally releasing it from a fusion peptide, its
activity as an antibody binding site and its
specificity for glp-4 (lysozyme) epitope are assayed
by established immunological techniques, e.g.,
affinity chromatography and radioimmunoassay.
Correct folding of the protein to yield the proper
three-dimensional conformation of the antibody
binding site is prerequisite for its activity. This
occurs spontaneously in a host such as a myeloma cell
which naturally expresses immunoglobulin proteins.
Alternatively, for bacterial expression, the protein
forms inclusion bodies which, after harvesting, must
be subjected to a specific sequence of solvent
conditions (e.g., diluted 20 X from 8 M urea 0.1 M
Tris-HC1 pH 9 into 0.15 M NaCl, 0.01 M sodium
phosphate, pH 7.4 (Hochman et al. (1976) Biochem.
J :2706-2710) to assume its correct conformation and
hence its active form.
Figures 4E and 4F show the DNA and amino
acid sequence of chimeric VH and VL comprising
human FRs from NEWM and murine CDRs from glp-4. The
CA 01341615 2011-05-24
1341615
1 CDRs are underlined, as are restriction sites of
choice for further CDR replacement or empirically
determined refinement.
These constructs also constitute master
framework genes, this time constructed of human
framework sequences. They may be used to construct
BABS of any desired specificity by appropriate CDR
replacement.
Binding sites with other specificities have
also been designed using the methodologies disclosed
herein. Examples include those having FRs from the
human NEWM antibody and CDRs from murine 26-10
(Figure 9A), murine 26-10 FRs and G-loop CDRs
(Figure 9B), FRs and CDRs from murine MOPC-315
(Figure 9C), FRs and CDRs from an anti-human
carcinoembryonic antigen monoclonal antibody
(Figure 9D), and FRs and CDRs 1, 2, and 3 from VL and
FRs and CDR 1 and 3 from the VH of the anti-CEA
antibody, with CDR 2 from a consensus immunoglobulin
gene (Figure 9E).
Ii. Model Binding Site:
The digoxin binding site of the IgG2a,k
monoclonal antibody 26-10 has been analyzed. The
26-10 V region sequences were determined from both
amino acid sequencing and DNA sequencing of 26-10 H
and L chain mRNA transcripts. The 26-10 antibody
exhibits a high digoxin binding affinity [Ko = 5.4 X
109 M-1] and has a well-defined specificity profile,
providing a baseline for comparison with the
biosynthetic binding sites mimicking its structure.
CA 01341615 2011-05-24
-52-
1341615
1 Protein Design:
Crystallographically determined atomic
coordinates for Fab fragments of 26-10 were obtained
from the Brookhaven Data Bank. Inspection of the
available three-dimensional structures of Fv regions
within their parent Fab fragments indicated that the
Euclidean distance between the C-terminus of the VH
domain and the N-terminus of the VL domain is about
35.A. Considering that the peptide unit length is
approximately 3.8 A, a 15 residue linker was selected
to bridge this gap. The linker was designed so as to
exhibit little propensity for secondary structure and
not to interfere with domain folding. Thus, the 15
residue. sequence (Gly-Gly-Gly-Gly-Ser) 3 was
selected to connect the VH carboxyl- and VL
amino-termini.
Binding studies with single chain binding
sites having less than or greater than 15 residues
demonstrate the importance of the prerequisite
distance which must separate VH from VL; for
example, a (Gly4-Ser)1 linker does not
demonstrate binding activity, and those with
(Gly4-Ser)5 linkers exhibit very low activity
compared to those with (Gly4-Ser)3 linkers.
Gene Synthesis:
Design of the 744 base sequence for the
synthetic binding site gene was derived from the Fv
protein sequence of 26-10 by choosing codons
CA 01341615 2011-05-24
- 53 -
1341615
frequently used in Z. coli. The model of this
representative synthetic gene is shown in Figure 8,
discussed previously. Synthetic genes coding for the
jp.promoter-operator, the modified t= LE leader
peptide (MLE), the sequence of which is shown in
Figure 1OA, and-VH were prepared largely as
described previously. The gene coding for VH was
assembled from 46 chemically synthesized
oligonucleotides, afl 15 bases long, except for
terminal fragments (13 to 19 bases) that included
cohesive cloning ends. Between 8 and 15 overlapping
oligonucleotides were enzymatically ligated into
double stranded DNA, cut at restriction sites
suitable for cloning (Narl, XbaI, Sall, Sacli, Saci),
purified by PAGE on 8% gels, and cloned in pUC which
was modified to contain additional cloning sites in
the polylinker. The cloned segments were assembled
stepwise into the complete gene mimicking VH by
ligations in the pUC cloning vector.
The gene mimicking 26-10 VL was assembled
from 12 long synthetic polynucleotides ranging in
size from 33 to 88 base pairs, prepared in automated
DNA synthesizers (Model 6500, Biosearch, San Rafael,
CA; Model 380A, Applied Biosystems, Foster City,
CA). Five individual double stranded segments were
made out of pairs of long synthetic oligonucleotides
spanning six-base restriction sites in the gene
(AatII, BstEII, PpnI, Hindill, BglII, and PstI). In
one case, four long overlapping strands were combined
and cloned. Gene fragments bounded by restriction
sites for assembly that were absent from the pUC
polylinker, such as AatII and BstEII, were flanked by
EcoRI and BamHI ends to facilitate cloning.
CA 01341615 2011-05-24
-54-
1341615
The linker between VH and VL, encoding
(Gly-Gly-Gly-Gly-Ser)3, was cloned from two long
synthetic oligonucleotides, 54 and 62 bases long,
spanning Sacl and AatII sites, the latter followed by
an EcoRI cloning end. The complete single chain
binding site gene was assembled from the VH, VL,
and linker genes to produce a construct,
corresponding to aspartyl-prolyl-V H-clinker)-VL,
flanked by EcoRI and PstI restriction sites.
The txp'promoter-operator, starting from its
SspI,site, was assembled from 12 overlapping 15 base
oligomers, and the MLE leader gene was assembled from
24 overlapping 15 base oligomers. These were cloned
and assembled in pUC using the strategy of assembly
sites flanked by cloning sites. The final expression
plasmid was constructed in the pBR322 vector by a
3-part ligation using the sites SspI, EcoRI, and PstI
(see Figure 10B). Intermediate DNA fragments and
assembled genes were sequenced by the dideoxy method.
Fusion Protein Expression:
Single-chain protein was expressed as a
fusion protein. The MLE leader gene (Fig. 10A) was
derived from X. coli Lt LE sequence and expressed
under the control of a synthetic tõro promoter and
operator. E. coil strain JM83 was transformed with
the expression plasmid and protein expression was
induced in M9 minimal medium by addition of
indoleacrylic acid (10 Ng/ml) at a cell density
with A600 : 1. The high expression levels of the
CA 01341615 2011-05-24
-55-
1341615
1 fusion protein resulted in its accumulation as
insoluble protein granules, which were harvested from
cell paste (Figure 11, Lane 1).
Fusion Protein Cleavage:
The WA leader was removed from the binding
site protein by acid cleavage of the Asp-Pro peptide
bond engineered at the junction of the MLE and
binding site sequences. The washed protein granules
containing the fusion protein were cleaved in 6 M
guanidine-HC1 + 10% acetic acid, pH 2.5, incubated at
37 C for 96 hrs. The reaction was stopped through
precipitation by addition of a 10-fold excess of
ethanol with overnight incubation at -20 C, followed
by centrifugation and storage at -20 C until further
purification (Figure 11, Lane 2).
Protein Purification:
The acid cleaved binding site was separated
from remaining intact fused protein species by
chromatography on DEAE cellulose. The precipitate
obtained from the cleavage mixture was redissolved in
6 M guanidine-HC1 + 0.2 M Tris-HC1, pH 8.2, + 0.1 M
2-mercaptoethanol and dialyzed exhaustively against 6
M urea + 2.5 ON Tris-HC1, pH 7.5, + 1 mM EDTA.
2-Mercaptoethanol was added to a final concentration
of 0.1 M, the solution was incubated for 2 hrs at
room temperature and loaded onto a 2.5 X 45 cm column
of DEAE cellulose (Whatmari DE 52), equilibrated with
*Trade Mark
CA 01341615 2011-05-24
- 56 -
1341615
6 N urea + 2.5 mM Tris-HC1 + 1 mM EDTA, pH 7.5. The
intact fusion protein bound weakly to the DE 52
column such that its elution was retarded relative to
that.of.the binding protein. The first protein
fractions which eluted from the column after loading
and washing with urea buffer contained BAGS protein
devoid of intact fusion protein. Later fractions
contaminated with some fused protein were pooled,
rechromatographed on DE 52, and recovered single
chain binding protein combined with other purified
protein into a single pool (Figure 11, Lane 3).
Eefoldina:
The 26-10 binding site mimic was refolded as
follows: the DE 52 pool, disposed in 6 M urea + 2.5
mM Tris-HC1 + 1 mM EDTA, was adjusted to pH 8 and
reduced with 0.1 M 2-mercaptoethanol at 37 C for 90
min. This was diluted at least 100-fold with 0.01 M
sodium acetate, pH 5.5, to a concentration below 10
Ng/ml and dialyzed at 4 C for 2 days against
acetate buffer.
Affinity Chromatography:
Purification of active binding protein by
affinity chromatography at 4 C on a
ouabain-amine-Sepharose*column was performed. The
dilute solution of refolded protein was loaded
directly onto a pair of tandem columns, each
containing 3 ml of resin equilibrated with the 0.01 N
acetate buffer, pH 5.5. The columns were washed
Trademark
CA 01341615 2011-05-24
-57-
1341615
individually with an excess of the acetate buffer,
and then by sequential additions of 5 ml each of 1 N
NaCl. 20 AN ouabain, and 3 K potassium thiocyanate
dissolved in the acetate buffer, interspersed with
acetate buffer washes.- Since digozin binding
activity was still present in the eluate, the eluate
was pooled and concentrated 20-fold by
*
ultrafiltration (PM 10 membrane, 200 ml concentrator;
Amicon), reapplied to the affinity columns, and
eluted. as' described. Fractions with significant
absorbance at 280 nm were pooled and dialyzed against
PBSA-or`the above acetate buffer. The amounts of
protein in the DE 52 and ouabain-Sepharose pools were
quantitated by amino acid analysis following dialysis
against 0.01 M acetate buffer. The results are shown
below in Table 1.
*Trade Mark
CA 01341615 2011-05-24
-58-
1341615
TABLE 1
Estimated Yields of BABS Protein During Purification
Yield
Cleavage relative
Net wt. Mg yield (t) to fusion
At Per I_. protein prior step protein
Cell 12.0 g 1440.0 mga
paste
Fusion 2.3 g 480.0 mga,b 100.0% 100.0%
protein
Granules
Acid 144.0 mg 38.0e 38.0e
Cleavage/
DE 52
pool
Ouabain- 18.1 mg 12.6d 4.7e
Sepharose
.pool
aDetermined by Lowry protein analysis
bDetermined by absorbance measurements
CDetermined by amino acid analysis
dCalculated from the amount of BABS protein
specifically eluted from ouabain-Sepharose relative
to that applied to the resin; values were determined
by amino acid analysis
ePercentage yield calculated on a molar basis
CA 01341615 2011-05-24
- 59 -
1341615
Sequence Analysis of Gene and Protein:
The complete gene was sequenced in both
directions using the dideozy method of Sanger which
confirmed, the gene was correctly assembled. The
protein sequence was also verified by protein
sequencing. Automated Edman degradation was conducted
on intact protein (residues 1-40), as well as on two
major.CNBr fragments (residues 108-129 and 140-159)
with a Model 470A gas phase sequencer equipped with a
Model 120A on-line phenylthiohydantoin-amino acid
analyzer (Applied Biosystems, Foster City, CA).
Homogeneous binding protein fractionated by SDS-PAGE
and eluted from gel strips with water, was treated
with a 20,000-fold excess of CNBr, in 1%
trifluoroacetic acid-acetonitrile (1:1), for 12 hrs at
250 (in the dark). The resulting fragments were
separated by SDS-PAGE and transferred
electrophoretically onto an Immobilort membrane
(Millipore, Bedford, MA), from which stained bands
were cut out and sequenced.
Specificity Determination:
Specificities of anti-digoxin 26-10 Fab and
the BABS were assessed by radioimmunoassay. Wells of
microtiter plates were coated with affinity-purified
goat anti-murine Fab fragment (ICN ImmunoBiologicals,
Lisle, IL) at 10 pg/ml in PBSA overnight at 40C.
After the plates were washed and blocked with 1% horse
serum in PBSA, solutions (50 pl) containing 26-10
Fab or the BABS in either PBSA or 0.01 M sodium
acetate at pH 5.5 were added to the wells and
*Trademark
CA 01341615 2011-05-24
-60-
1341615
1 incubated 2-3 hrs at room temperature. After unbound
antibody fragment was washed from the wells, 25 K1
of a series of concentrations of cardiac glycosides
(10-4 to 10-11 M in PBSA)'were added. The cardiac
glycosides tested included digoxin, digitoxin,
digoxigenin, digitoxigenin, gitoxin, ouabain, and
acetyl strophanthidin. After the addition of
125I-digoain (25 pl, 50,000 cpm; Cambridge
Diagnostics, Billerica, MA) to each well, the plates
were incubated overnight at 4 C, washed and counted.
The inhibition curves are plotted in Figure 12. The
relative affinities for each digoxin analogue were
calculated by dividing the concentration of each
analogue at 50% inhibition by the concentration of
digoxin (or digoxigenin) that gave 50% inhibition.
There is a displacement of inhibition curves for the
BABS to lower glycoside concentrations than observed
for 26-10 Fab, because less active BABS than 26-10 Fab
was bound to the plate. When 0.25 M urea was added to
the BABS in 0.01 M sodium acetate, pH 5.5, more active
sFv.was bound to the goat anti-murine Fab coating on
the plate. This caused the BABS inhibition curves to
shift toward higher glycoside concentrations, closer
to the position of those for 26-10 Fab, although
maintaining the relative positions of curves for sFv
obtained in acetate buffer alone. The results,
expressed-as normalized concentration of inhibitor
giving 50% inhibition of 1251-digoxin binding, are
shown in Table 2.
CA 01341615 2011-05-24
- 61 -
1341615
1
TABLE 2
26-10
Antibody Normalizing
species Glycoside Q pQ pQ pQQ A-S Q Q
Fab Digoxin 1.0 1.2 0.9 1.0 1.3 9.6 15
Digoxigenin 0.9 1.0 0.8 0.9 1.1 8.1 13
GABS Digoxin 1.0 7.3 2.0 2.6 5.9 62 150
Digoxigenin 0.1 1.0 0.3 0.4 0.8 8.5 21
D = Digoxin
DG = Digoxigenin
DO = Digitoxin
DOG Digitoxigenin
A-S = Acetyl Strophanthidin
G = Gitoxin
0 = Ouabain
Affinity Determination:
Association constants were measured by
equilibrium binding studies. In immunoprecipitation
experiments, 100 p1 of 3H-digoxin (New England
Nuclear, Billerica, MA) at a series of concentrations
(10-7 M to 10-11 M) were added to 100 K1 of
26-10 Fab or the BABS at a fixed concentration.
After 2-3 hrs of incubation at room temperature, the
protein was precipitated by the addition of 100 pl
goat antiserum to murine Fab fragment (ICN Immuno-
CA 01341615 2011-05-24
-62- 1341615
Biologicals), 50 p1 of the IgG fraction of rabbit
anti-goat IgG (ICN ImmunoBiologicals), and 50 pl of
a 10% suspension of protein A-Sepharose (Sigma).
Following 2 hrs at 40C, bound and free antigen were
separated by vacuum filtration on glass fiber filters
(Vacuum Filtration Manifold, Millipore, Bedford,
MAY. Filter disks were then counted in 5 ml of
scintillation fluid with a Model 1500 Tri-Carb* Liquid
Scintillation Analyzer (Packard, Sterling, VA). The
association constants, K0, were calculated from
Scatchard analyses of the untransformed radioligand
binding data using LIGAND, a non-linear curve fitting
program based on mass action. Kos were also
calculated by Sips plots and binding isotherms shown
in Figure 13A for the BABS and 13B for the Fab. For
binding isotherms, data are plotted as the
concentration of digoxin bound versus the log of the
unbound digoxin concentration, and the dissociation
constant is estimated from the ligand concentration
at 50% saturation. These binding data are also
plotted in linear form as Sips plots (inset), having
the same abscissa as the binding isotherm but with
the ordinate representing log r/(n-r), defined
below. The average intrinsic association constant
(Ko) was calculated from the modified Sips equation
(39), log (r/n-r) = a log C - a log Ko, where r
equals moles of digoxin bound per mole of antibody at
an unbound digoxin concentration equal to C; n is the
number.of moles of digoxin bound at saturation of the
antibody binding site, and a is an index of
heterogeneity which describes the distribution of
association constants about the average intrinsic
*Trade Mark
CA 01341615 2011-05-24
- 63 -
1341615
1 association constant Ko. Least squares linear
regression analysis of the data indicated correlation
coefficients for the lines obtained were 0.96 for the
BABS and 0.99 for 26-10 Fab. A summary of the
calculated association constants are shown below in
Table 3.
TABLE 3
Association Constant, K 0
Method of Data K0'(BABS), M 1 K0 (Fab), M 1
Analysis
Scatchard plot (3.2 0.9) X 107 (1.9 0.2) X 108
Sips plot 2.6 X 107 1.8 X 108
Binding
isotherm 5.2 X 107 3.3 X 108
III. Synthesis of a Multifunctional Protein
A nucleic acid sequence encoding the single
chain binding site described above was fused with a
sequence encoding the FB fragment of protein A as a
leader to function as a second active region. As a
spacer, the native amino acids comprising the last 11
amino acids of the FB fragment bonded to an Asp-Pro
dilute acid cleavage site was employed. The FB
binding domain of the FB consists of the immediately
preceding 43 amino acids which assume a helical
configuration (see Fig. 2B).
CA 01341615 2011-05-24
- 64 -
1341615
1 The gene fragments are synthesized using a
Biosearch DNA Model 8600 Synthesizer as described
above. Synthetic oligonucleotides are cloned
according to established protocol described above
5- using the pUC8 vector transfected into Z., coll. The
completed fused gene set forth in Figure 6A is then
expressed in K, coll.
After sonication, inclusion bodies were
collected by centrifugation, and dissolved in 6 M
guanidine hydrochloride (GuHC1), 0.2 M Tris, and 0.1 M
2-mercaptoethanol (BME), pH 8.2. The protein was
denatured and reduced in the solvent overnight at room
temperature. Size exclusion chromatography was used
to purify fusion protein from the inclusion bodies. A
Sepharose 4B column (1.5 X 80 cm) was run in a solvent
of 6 M GuHCl and 0.01 M NaOAc, pH 4.75. The protein
solution was applied to the column at room temperature
in 0.5-1.0 ml amounts. Fractions were collected and
precipitated with cold ethanol. These were run on SDS
gels, and fractions rich in the recombinant protein
(approximately 34,000 D) were pooled. This offers a
simple first step for cleaning up inclusion body
preparations without suffering significant proteolytic
degradation.
For refolding, the protein was dialyzed
against 100 ml of the same GuHC1-Tris-BME solution,
and dialysate was diluted 11-fold over two days to
0.55 M GuHC1, 0.01 M Tris, and 0.01 M BME. The
dialysis sacks were then transferred to 0.01 M NaCl,
and the protein was dialyzed exhaustively before being
assayed by RIA's for binding of 125I-labelled
digoxin. The refolding procedure can be simplified by
CA 01341615 2011-05-24
-65- 1341615
making a rapid dilution with water to reduce the GuHC1
concentration to 1.1 M, and then dialyzing against
phosphate buffered saline (0.15 M NaCl, 0.05 M
potassium phosphate, pH 7, containing 0.03% NaN3),
so that it is free of any GuHC1 within 12 hours.
Product of both types of preparation showed binding
activity, as indicated in Figure 7A.
Demonstration of Bifunctionality:
This protein with an FB leader and a fused
BABS is bifunctional; the BABS can bind the antigen
and the FB can bind the Fc regions of
immunoglobulins. To demonstrate this dual and
simulataneous activity several radioimmunoassays were
performed.
Properties of the binding site were probed by
a modification of an assay developed by Mudgett-Hunter
et al. (J. Immunol. (1982) 129:1165-1172; Molec.
Immunol. (1985) 22:477-488), so that it could be run
on microtiter plates as a solid phase sandwich assay.
Binding data were collected using goat anti-murine Fab
antisera (gAmFab) as the primary antibody that
initially coats the wells of the plate. These are
polyclonal antisera which recognize epitopes that
appear to reside mostly on framework regions. The
samples of interest are next added to the coated wells
and incubated with the gAmFab, which binds species
that exhibit appropriate antigenic sites. After
washing away unbound protein, the wells are exposed to
1251-labelled (radioiodinated) digoxin conjugates,
either as 125I-dig-BSA or 125I-dig-lysine.
CA 01341615 2011-05-24
-66- 1341615
1 The data are plotted in Figure 7A, which
shows the results of a dilution curve experiment in
which the parent 26-10 antibody was included as a
control. The sites were probed with 125I-dig-BSA as
described above, with a series of dilutions prepared
from initial stock solutions, including both the
slowly refolded (1) and fast diluted/quickly refolded
(2) single chain proteins. The parallelism between
all three dilution curves indicates that gAmFab
binding regions on the BABS molecule are essentially
the same as on the Fv of authentic 26-10 antibody,
i.e., the surface epitopes appear to be the same for
both proteins.
The sensitivity of these assays is such that
binding affinity of the Fv for digoxin must be at
least 106. Experimental data on digoxin binding
yielded binding constants in the range of 108 to
109 M 1. The parent 26-10 antibody has an
affinity of 7 X 109 M 1
_ Inhibition assays also
indicate the binding of 1251-dig-lysine, and can be
inhibited by unlabelled digoxin, digoxigenin,
digitoxin, digitoxigenin, gitoxin, acetyl
strophanthidin, and ouabain in a way largely parallel
to the parent 26-10 Fab. This indicates that the
specificity of the biosynthetic protein is
substantially identical-to the original monoclonal.
In a second type of assay, Digoxin-BSA is
used to coat microtiter plates. Renatured BABS
(FB-BABS) is added to the coated plates so that only
molecules that have a competent binding site can stick
to the plate. 125I-labelled rabbit IgG
(radioligand) is mixed with bound FB-BABS on the
CA 01341615 2011-05-24
- 67 -
1341615
1 plates. Bound radioactivity reflects the interation
of IgG with the FB domain of the BABS, and the
specificity of this binding is demonstrated by its
inhibition with increasing amounts of FB, Protein A,
rabbit IgG. IgG2a, and IgGl, as shown in Figure 7B.
The following species were tested in order to
demonstrate authentic binding: unlabelled rabbit IgG
and IgG2a monoclonal antibody (which binds
competiviely to the FB domain of the BABS); and
protein A and FB (which bind competively to the
radioligand). As shown in Figure 7B, these species
are found to completely inhibit radioligand binding,
as expected. A monoclonal antibody of the IgGi
subclass binds poorly to the FB, as expected,
inhibiting only about 34% of the radioligand from
binding.. These data indicate that the GABS domain and
the FB domain have independent activity.
IV. OTHER CONSTRUCTS
Other BABS-containing protein constructed
according to the invention expressible in R. coli and
other host cells as described above are set forth in
the drawing.. These proteins may be bifunctional or
multifunctional. Each construct includes a single
chain BABS linked via a spacer sequence to an effector
molecule comprising amino acids encoding a
biologically active effector protein such as an
enzyme, receptor, toxin, or growth factor. Some
examples of such constructs shown in the drawing
include proteins comprising epidermal growth factor
(EGF) (Figure 15A), streptavidin (Figure 15B), tumor
CA 01341615 2011-05-24
-68- 1341615
necrosis factor (TNF) (Figure 15C), calmodulin (Figure 15D) the beta
chain of platelet derived growth factor (B-PDGF) (15E) ricin A (15F),
Interleukin 2 (15G) and FB dimer (15H). Each is used as a trailer and
is connected to a preselected BABS via a spacer (Gly-Ser-Gly) encoded
by DNA defining a BamHI restriction site. Additional amino acids may
be added to the spacer for empirical refinement of the construct if
necessary by opening up the BAM HI site and inserting an
oligonucleotide of a desired length having BamHI sticky ends. Each
gene also terminates with a PstI site to facilitate insertion into a
suitable expression vector.
The BABS of the EGF and PDGF constructs may be, for example,
specific for fibrin so that the EGF or PDGF is delivered to the site
of a wound. The BABS for TNF and ricin A may be specific to a tumor
antigen, e.g., CEA, to produce a construct useful in cancer therapy.
The calmodulin construct binds radioactive ions and other metal ions.
Its BABS may be specific, for example, to fibrin or a tumor antigen,
so that it can be used as an imaging agent to locate a thrombus or
tumor. Alternatively, other ion sequestering polypeptides useful in
the practice of the invention may include metallothionein or an amino
acid sequence rich in at least one of glutamic acid, aspartic acid,
lysine, and arginine. The streptavidin construct binds with biotin
with very high affinity. The biotin may be labeled with a remotely
detectable ion for imaging purposes. Alternatively, the biotin may be
immobilized on an affinity matrix or solid support. The BABS-
streptavidin protein could then be bound to the matrix or support for
affinity chromatography or solid phase immunoassay. In addition, the
effector protein capable of selective binding to a solid support may
include, for example, a positively or negatively charged amino acid
sequence, a cysteine-containing amino acid sequence, or a fragment of
CA 01341615 2011-05-24
-69- 1341615
Staphylococcus protein A. The interleukin-2 construct could be
linked, for example, to a BABS specific for a T-cell surface antigen.
The FB-FB dimer binds to Fc, and could be used with a BABS in an
immunoassay or affinity purification procedure linked to a solid phase
through immobilized immunoglobulin.
Figure 14 exemplifies a multifunctional protein having an
effector segment as a leader. It comprises an FB-FB dimer linked
through its C-terminal via an Asp-Pro dipeptide to a BABS of choice.
It functions in a way very similar to the construct of Fig. 15H. The
dimer binds avidly to the Fc portion of immunoglobulin. This type of
construct can accordingly also be used in affinity chromatography,
solid phase immunoassay, and in therapeutic contexts where coupling of
immunoglobulins to another epitope is desired.
In view of the foregoing, it should be apparent that the
invention is unlimited with respect to the specific types of BABS and
effector proteins to be linked. Accordingly, other embodiments are
within the following claims.
