Note: Descriptions are shown in the official language in which they were submitted.
~)4~
SYNT}IETIC INTERLEUKIN-6
This application is a oontinuat.ion-in-part of copending
U.S. Application Serial No. 07/278,690, filed December 1,
198~.
_____ _
1. INTRODUCTION
The present invention relates to genes and their
encoded proteins which are recombinant mature interleukin-6
(hereinafter IL-6) and a synthetic cysteine-free protein that
retains IL-6 activity. These proteins are expressed in
unicellular hosts as the amino or carboxy terminal peptide
portion o a tri-hybrid~fusion protein comprising either
interleukin-6 or its modified synthetic form, together with a
portion of a cleavage site and carrier DNA. The recombinant
fusion protein is purified and digested i~ vitro with a
protease that ls specific for the cleavage site to liberate
the peptide with IL-6 activity, which is easily separated
from the large ~-galactosidase protein expressed by the
carrier DNA. The purified peptide can be used to stimulate
the production of proteins, including immunoglobulins and
hepatic proteins and may be used to prevent viral infections.
The protein can also be added to vaccine preparations as an
adjuvant.
2. BACK5ROUND OF THE INVENTION
2.1. PEPTIDE REGULATORS
Physiological agents that regulate metabolic activity
of distant cells were given the name hormone by English
scientists Bayliss and Starlinger in 190~. These agents
consist of amino acid derivatives, s~eroids and peptides.
More recently, a variety of peptides that activate and/or
inhibit cell proliferation have been identified and termed
35 stimulatory factors or growth factors. An alternative
general term for such cellular factors is cytokine although
more specific ter~inology indicates the cell of origin, i.e.
lymphokines which are produced by lymphocytes. Lymphokines
also belong to the interleukin class of molecules that
modulate the proliferation of cells in the immune system.
Other peptides produced by the immune system result in
specific antiviral activities; such peptides are termed
interferons. To qualify as an interferon a factor must be a
protein which exerts antiviral activiky through cellular
metabolic processes involving the synthesis of both RNA and
protein (Committee on Interferon Nonmenclature, 1980, Nature
86(2~:110).
2.~ INTERLEUKIN-6
Interleukin-6 (IL-6) is the term given to a peptide
described alternatively as IFN~2A, BSF-2, H5F, G-CSF, CS-309,
HPGF, or 26 kDa protein by a multiplicity of investigators.
These original factors now known a~ IL-6 include: I)
interferon-~2 (2ilberstein et al., 1986, ~MBO J. 5:2529) or
26 kDa protein (Haegeman et al., lg86, Eur. J. ~iochem.
159:625) which was first detected in poly(rI) poly(rC)
stimulated fibroblasts; 2) a potent T-cell derived lymphokine
termed B-cell stimulation factor-2 (BSF-2) ~Hirano et al.,
1986, Nature 324:73): 3) a fibroblast product called 8-cell
hybridoma/plasmacytoma growth factor (HPGF or HGF) (Van Snick
et al., 1986, Proc. Natl. Acad. Sci. 83:9679 Billiau, 1987,
Immunol. Today ~:84 Van Damme et al., 1987, Eur. J. Biochem.
168:543; Tosato et al., 1988, Science 239:502): and 4) a
peripheral blood monocyte protein called hepatocyte
stimulating factor (HSF) (Gauldie et al., 19~7, Proc. Natl.
Acad. Sci. 84:7251).
The functions which have been ascribed to the IL-6
peptide are basic to both the inflammatory and immune
response in human pathology. Those functions are diverse and
35 depend on the type of cells under examination~ IL-6 is
20~fi~
~3--
expressed in leukocytes, epithelial cells, IL-I treated
fibroblasts, hepatocytes, vascular endothelial cells, cardiac
myxoma tissue, certain bladder carcinom~s, cert~in cervical
cancer cells and glial cells. IL-6 is one of the peptides
involved in the interaction of T cells with B cells to result
in the prolifer~tion and diferentlation o antibody
producing cells. IL-6 significantly enhances secretion of
immunoglobulins in activated B-cells (Mura~uchi et al., 1988,
J. EXp. Med. ~ 332; Tosato et al., 1988, Science
239:502; Hirano et al., 1985, Proc. Nat:l. Acad. Sci.
ln
82:5490).
With regard to the antiviral activity responsible for
the initial identification of IL-6 as an interferon, the
transformation of Chinese hamster ovary cells with a
recombinant IL-6 plasmid which allowed constitutive
expression of IL-6 resulted not only in the detection of the
peptide in media, but also resulted in the detection of
antiviral activity (Zi]berstein et al., 1986, EM~O J.
5:2529).
Whether natural IL-6 has antiviral activity or not has
been subject to debate in the scientific literature.
Antiviral specific activities were first reported by
Weissenbach et al., (1980, Rroc. Natl. Acad. Sci. 77:7152);
and members of this scientific group have continued to report
antiviral activitie~ in their preparatlons (see Content et
al., 1985, Eur. J~ Biochem. 152:253: May et al., 1988, J.
Biol. Chem. 263:7760). The values reported for antiviral
activity range ~rom 5-10 x 1o6 U/mg protein to as low as 1-3
x 102 U/mg. In contrast, a large number of other
30 investigators have repoxted no significant antiviral e~fects
associated with their purified recombinant IL-6 preparations.
(Poupart et al., 1987, EMBO J. 6:1219 Van Damme et al.,
1988, J. Immunol. 140:1534 Reis et al., 1988, J. Immunol.
140Q5):1566; Hirano et al., I988, ImmunolO Lett. 17(1~:41).
IL-6 appears to suppress the action of TNF (Kohase et
2 ~ rZ ~
al., 1987, Mol. Cell Biol. 7:273). However, it stimulates
the growth of human B-lymphoblastoid cells infected with EBV
(Tosato et Al., 19~, Science 239:502), and of human
thymocytes and T-lymphocytes tLotz et al., 1988, J. Exp. Med.
1_7~ 1253). It has been identified as a growth factor for
murine hybridomas (Poupart et al., 1987, EMBO J. 6:1219) and
hybridoma plasmacytoma cell lines (Van Damme et al., 1987, J.
Exp. Med. 165:914). In combination with IL-3, IL-6 supports
the proliferation of hematopoietic progenitor cells (Van
Damme et al., 1987, J. Exp. Med. 165:914; Ikebuchi et al.,
1987, Proc. Natl. Acad. Sci. 84:9035), and modulates the
synthesis of a subset of hepatocyte proteins in response to
injury and infection (Gauldie et al,, 1987, Proc. Natl. Acad.
5ci. 84:7251: Andus et al., 1987, FEBS Lett. 221:18). The
multiplicity of its own actions, and its interactions with
other peptide regulators like IL-l, TNF, IFN~l and PDGF, have
led to the suggestion that IL-6 plays a pivotal role in a
complex cytokine network needed for homeostatic control of
cellular functions (Kohase et al., 19~7, Mol. Cell Biol.
7:273; Sehgal et al., 1987, Science 235:731; Billiau, 1987,
Immunol. Today 8:84; Sporn and Roberts, 1988, Nature
332:217).
2.3. DNA SEQUENCE OF THE INTERLEUKIN-6 PROTEIN
Comparison of the primary structure of factors that
were originally characterized by multiple investigators on
the basis of differing biological activitie~ revealed the
identity of their amino acid sequences and resulted in the
renaminy of the peptides as interleukin-6 (IL-6).
3a Fibroblasts treated with either poly(rI) (rC) or
cycloheximide and actinomycin D produced a 14S mRNA molecule
which coded for a protein capable of inducing antiviral
activity called IFN~2 (Weissenbach et al., 1g80, Proc. Natl.
Acad. Sci. 77:7152; British Patent No. 2,063,882~. Clones
35 produced from cDNA copies of this induced mRNA fraction
2~ fil.
provided a partial sequence of the IFN~2 promoter that was
clearly different from IFN~l, IFN7~A, IFN7~, and IFN~D
(Chernajovsky et al., 1984, DNA 3:297; Revel et al., 1983,
Interferon 5:205 1. Gresser ed., Academic Press, N.Y.). The
complete lFNa2 sequence derived ~rom multiple cDNA clones
defined a 212 a~ino acid long protein in addition to a
probable ATG start se~uence, polyadenylation sites~ TATA
boxes and the mRNA start site which was revealed by Sl
nuclease mapping (Zilberstein et al., 1986, EMB0 J. 5:2529;
European Patent Application 0220574A1, Publication Date 06-
05-1987). In a different analysis of the IFN~2 sequence, May
et al., 1986, Proc. Natl~ Acad. Sci. 83:8957, also defined
the 212 amino a~id protein sequence using a full length cDNA
clone rather than the partial cDNA clones of Zilberstein et
al., 1986, EMB0 J. 5:2529. These investigators also noted
that IFN~2 was induced by TNF.
Even though their 26 kDa protein had no detectable
antiviral activity, Haegemann et al., 1986, Eur. J. Biochem.
159:625 recogn.ized that the sequence of their 26 kDa protein,
induced in fibroblasts by treatment with cycloheximide or
interleukin-l, was identical to the IFN~2 of Zilberstein et
al., 1986, EMB0 J. 5:2529. Because the 5' terminus of the
protein was missing .in the cDNA clone collection~ a screen of
a human gene library, testing for complementarity with an
internal cDNA sequence of the 26 kDa peptide, yielded genomic
clones that provided the complete 212 amino acid sequence, as
well as the DNA sequence of a 162 bp intron in the 5'
terminus region of the human gene. A search o~ a protein
data base containing 3309 individual peptide sequences failed
to reveal any significant similarities with proteins in the
database.
In a separate study ~Hirano et al., 1986, Nature
324:73), BSF-2 protein was purified from a human T-cell line
that constitutively produced the factor and was established
using the HTLV-1 virus. The amino acid sequence data from
--6--
nine peptide fragments provided the information necessary to
produce synthetic oligomers which were used to probe cDNA
li~raries. The cDNA clones were se~lenced as was the amino
terminus o the purified IL-6 protein. The end of the mature
protein sequence was pro-val pro-pro indicating that the 212
amino acid prepeptid~ that was predicted from the nucleic
acid sequence contained a 28 amino acid long signal peptide
which is cleaved to produce the natural mature BSF-2 (IL-6
protein.
Continued sequence analysis of the entire BSF-2 (IL-6)
genomic DNA demonstrated that the chromosomal segment
contained five exons and four introns (Yasukawa et al., 1987,
EMBO J. 6:2939). The organization of the BSF-2 (IL-6) gene
was strikingly ~imilar ~o that of G-CSF when the two genes
were compared.
After completion of the sequence of H~F (Brakenhoff et
al., 1987, J. Immunol. 139:4116), the identity of HGF and
lL-6 was confirmed. In addition, a study of the sequence
differences in all reported analyses of the IL-6 gene (as
HGF, IFN~2, the 26 kDa protein or BSF-2) revealed f~w single
base changes. Neither of the two changes that occur within
the peptide reading frame produce an amino acid change. Van
Damme et al., 1988, J. Immunol. 140:1534, also noted that the
amino terminal sequence of HGF was identical to IFN~2, the 26
kDa protein and BSF 2.
Clark et al. (International Application Number
PCT/US87/01611, Publication Number W088/00206, published
January 14, 19~8) also reports the cDNA sequence of IL-6.
2.4. CONSERVATION OF CYSTEINE XESIDUE5 IN IL-6
All cysteine residue positions are conserved between
the two proteins, BSF-2 (IL-6) and G-CSF. Upon noting this
similarity, Hirano et al., 1986, Nature 324:73, suggested
that intramolecular disulphide bonds would be important in
35 the structure of BSF-2 (IL~6~ and G-C5F. Their suggestion
~O~fl2~
--7--
was based on a comparison of the ~SF-2 (IL-~) sequence with
other se~lences in a personal database of multiple growth
factors, interleukins and interferons which revealed that
B~F-2 (IL-6) was distantly relat~d to G-CSF. However,
comparison with the National Biomedical Research Foundation
database or the Genetic Sequence Data ~ank revealed no
significant similarity with other proteins in those
databanks.
Conservation of the cysteines was also suggested by Van
Snick et al., 1986, Proc. Natl. Acad. Sci. 83:9679. They
sequenced murine HP-1 and found conserved homology with human
IL-6. They also examined the positions of the cysteines of
HP-1, IL-6 and G-CSF and noted that the cysteine positions
were conserved within t~e three peptides.
As described by Yasukawa et al., ~ , the
organization of exons and introns within the BSF-2 (IL-6)
gene was strikingly similar to that o~ G-CSF when the two
genes were compared. This finding provides further support
for the shared evolutionary history-of the two proteins (see
also Kishimoto, 1987, J. Clin. Immunol. 7(5):343~.
The presence of cysteines in a protein can cause
problems in processing when the protein is being produced
recombinantly in ~ bacterial host. Microblally produced
cysteine-containing proteins may tend to form multimers which
greatly complicate puri~ication of the protein product.
Several additional purification steps, such as reduction and
reoxidation of the recombinant protein, may be required to
obtain the protein in the proper conformation. Removal of
one or more of the cysteine residues, with concurrent
30 replacement by a chemically equivalent neutral amino acid,
would be desirable, in oxder to simplify the isolation and
purification of the IL-6 molecule. However, the successful
removal of cysteines ~rom biologically active molecules is
unpredictable, in that the ~ertiary structure, in the absence
35 of the normally formed disulfide bridges, can be
2a~0~fi~.
substantially altered. It is often the case that one or more
of these residues may he essential to retaining the desired
activity.
Natural cysteine residues are expected to be
particularly critical for cyto~ines such as IL-6, since the
cysteine~ of IL-6 have been conserved (See above). Moreover,
the biological activity of other cytokines has been reported
to be affected if cysteines are removed. For example, only
one cysteine residue out of three can be removed without
af~ecting the biological activity of the IFN-~ (see U.S. Pat.
No. 4,737,462). Similarly, only one of four cysteines of
IFN~l can be removed without complete loss of activity (Mark
et al., 1984, Proc. Natl. Acad. Sci. 81: 5662).
2.5. PRODUCTION OF NATURAL INTERLEUXIN-6
AS A RECOMBINANT PROTÆIN
-
2.5.1. RECOMBINANT DNA TECHNOLOGY AND GENE EXPRESSION
Recombinant DNA technology involve~ insertion of
specific DNA sequences into a DNA vehicle (vector) to Eorm a
recombinant DNA molecule which is capable of replication in a
20 host cell. Generally, the inserted DNA sequence is foreign
to the recipient DNA vehicle, i.e , the inserted DNA sequence
and the DNA vector are derived from organisms which do not
exchange genetic information in nature, or the inserted DNA
sequence may be wholly or partially synthetically mad~.
25 Several general methods have been developed which enable
construction of recombinant DNA molecules.
Regardless of the method used for construction, the
recombinant DNA molecule must be compatible with the host
cell, i.e., capable of autonomous replication in the host
3~ cell or stably integra~ed into one or mor~ of the host cells
chromosomes. The recombinant DNA molecule should prefexably
also have a marker function which allows the selection o~ the
desired recombinant DNA molecule(s;. In addition, if all o~
the proper replication, transcription, and translation
signals are correctly arranged on the recombinant vector, the
foreign gene will be properly expressed in, e.g., the
transformed bacterial cells, in the case of bacterial
expression plasmids, or in permissive cell lines or hosts
infected with a recombis~ant virus or carrying a recombinant
plasmid having the appropriate orig~n of replication.
Different genetic signals and processing events control
levels of gene expression such as DNA transcription and
messenger RNA (mRNA) translation. Transcription of DNA is
dependent upon the presence of a promoter, which is a DNA
sequence that directs the binding of ~A polymerase and
thereby promotes mRNA synthesis. The DNA sequences of
eucaryotic promoters differ from those of procaryotic
promoters. Furthermore,~ eucaryotic promoters and
accompanying genetic signals may not be recognized in or may
not function in a procaryotic system and furthermore,
procaryotic promoters are not recogni2ed and do not function
in eucaryotic cells.
5imilarly, translation of mRNA in procaryotes depends
upon the presence of the proper procaryotic signals, which
differ from those of eucaryotes. Efficient translation of
mRNA in procaryotes requires a ribosome binding site called
the Shine-Dalgarno (S/D) sequenc~e on the mRNA (Shine, J. and
Dalgarno, L., 1975, Nature 254:34). This sequence is a short
nucleotide sequence of mRNA that is located before the start
codon, usually AUG, which encodes the amino-terminal
methionine of the protein. The S/D sequences are
complementary to the 3' end of the 16S rRNA (ribosomal RNA),
and probably promote ~inding of mRNA to ribosomes by
duplexing with the rRNA to allow correct positioning of the
ribosome.
Although the Shine/Dalgarno sequence, consisting of the
few nucleotides of complementarity between the 16S ribosomal
RNA and mRNA, has been identified as an important feature o~
the ribosome binding site (Shine and Dalgarno, 1975, Nature
Z63~
--10--
254: 34; Steitz, 1980, in Ribosomes: Structure, Function and
Geneti-cs ed. Chambliss et al. Baltimore, Md., University Park
Press pp. 479-495), computer analysis has indicated that
approximately one hundred nucleotides surrounding the AUG
initiating codon are involved in ribosome/mRNA interaction as
indicated by proper predlction of translation start signals
(Stormer et al., 1982, Nucl. Acids Res. 10:2971; Gold et al.,
1984, Proc. Natl. Acad. Sci. 81:7061). No prediction of
what actually proYides the best and complete ribosome binding
site for maximum translation of a specific protein can be
made (see Joyce et al~, 1983, Rroc. Natl. Acad. Sci.
80:1830).
Schoner and Schoner recognized the significance of the
entire ribosome/mRNA interaction region in the development of
recombinant expression vectors in their charactarization of a
72 bp sequence termed the "minicistron" sequence (See Figure
1 of Schoner et al., 1986, Proc. Natl. Acad. Sci. USA 83:
8506). A one base deletion in the first cistron of the
"minicistron" sequence was sufficient to increase the
production of the downstream recombinant protein Met-[Ala]bGH
from 0.4~ to 24% of total cell protein (See Figure 4, pCZ143
compared to pCZ145, Schoner et al., id.).
Alternati~ely a two base insertion also resulted in
significant expression of the coding peptide encoded by the
second cistron. Control experiments indicated that the
differences in expression were due to translational
differences because mRNA levels in these constructs were
essentially equi~alent (no more than 3 fold different) as
compared to the expressed protein differences ~which were
approximately 50 fold). The conclusion was that the position
of the stop codon that termina~es translation of the first
cistron of the minicistron sequence affected tha efficiency
o~ translation of the second cistron, con~aining the coding
sequence of the recombinant protein. Most importantly their
35 work indicated that one or two base changes in the sequence
--ll--
immediately preceding the coding sequence of a recombinant
protein can have tremendous effects on downstream expression.
Successful expression of a cloned gene requires
suficient transcripkion of DNA, translation o~ the mRNA and
in some .instan~es, post-translational modification of the
proteln. Expression vectors have been used to express genes
under the control of an active promoter in a suitable host,
and to increase protein production.
2.5.2. NONBACTERIAL PRODUCTION OF RECOMBINANT INTERLEUKIN-6
Although Weissenbach et al., 19~0, Proc~ Natl. Acad.
Sci. 77:7152 translated cDNA in oocytes to make recombinant
IFN~2 in vitro, Zilberstein et al., 1986, EMBO J. 5:2529
reported the first exam~le of IFN~2A (IL-6~ cDNA expression
in vivo. The restriction fragment containing the fused cDNAs
o~ different primary clones was positioned downstream from
the SV40 early promoter to produce the pSVCIFB2 plasmid. In
this plasmid, the entire IFN~2A cDNA sequence and some
adjacent nucleotides o~ the original cloning vector followed
the SV40 early promoter and the first ~0 nucleotides of the T
antigen RNA. The IFN~2A sequence was followed by the T
antigen splicing region and polyadenylation site. After the
steps of trans~ection into Chinese hamster ovary (CHO) tlssue
culture cells and methotrexate amplification selection, CHO
cell clones labeled with [35S] were screened for radioactive
IFN~2 protein produ¢tion by immunoprecipitation using
nonsaturating amounts of antibodie~ In the same study,
plasmid constructs, wherein the entire IFN~2A cDNA sequence
was fused to the T7 RNA polymerase promoter, also were
30 transcri~ed in vitro to produce mRNA that was subseqllently
translated in rabbit reticulocyte lysates. Following
immunoprecipitation of the lysate, the proper sized IFN~2
(IL-6) protein could be detected on SDS-polyacrylamide gels.
BSF-2 (IL-6) was functionally expressed in COS7 cells
35 following transfection of a plasmld containing the entire
.2~;~
-12-
coding region of BSF-2 adjacent to the SV40 early promoter
(Hirano et al., 1986, Nature 324:73). BSF-2 activity could
be detected after purification and concentration of the media
from the transfected cells using immunoaffinity gel methods.
Recombinant IL~6 produced usiny this method was used to
analyze the regulation of ibrinogen and albumin mRN~ in FAO
cells (Andus et al., 1987, FEBS Letts. 221: 18).
Clark et al., (International Publication Number
W088/00206, published January 14, 1988) described the
construction o~ pCSF30~ wherein the IL-6 CDNA sequence was
ligated into the p91023B plasmid which contains the SV40
enhancer, major adenovirus late promoter, DHFR coding
sequence, SV40 late message poly-A addition site and the VA I
gene. When COS cells are transfected with this plasmid, IL-6
hematopoietic stimulating activity could be recovered at a
10-4 dilution of conditioned tissue culture cell medium.
This type of recombinant preparation was used in an analysis
o~ the effect af IL-6 on T-cell proliferation in the presence
of either ConA or agarose beads coupled with F23.1 IgG2a
anti-[mouse T cell receptor B chain variable region seqments
8.1, 8.2, and 8.3]-mouse antibodies (Garman et al., 1987,
Proc. Natl. Acad. Sci. ~4:76~9). The preparation of
recombinant IL-6 was also used in a study of the ef~ect of
IL-6 on the differentiation of Ly-2+ cytolytic lymphocytes
from murine thymocytes in the present of interleukin-2 (Takai
et al~, 1988, J. Immunol. 140:508). Alternative methods o~
higher cell production of recombinant IL 6 include in vitro
synthesis following injection of IL-6 mRN~ into Xenopus
oocytes (Weissenbach et al., 1980, Proc. Natl. ~cad. Sci.
77:7152; Coulie et al., 1987t Eur. JO Immunol. 17:1435;
poupart et al., 1~87, EMB0 J. 6:1219), by translation of the
IL-6 mRNA in xabbit reticulocytes (Poupart et al., 1987, EM80
J. 6:1219), or by cDNA expression in yeast (La Pierre et al.,
1988, J. Exp. Med. 167:794~.
-13-
2.5.3. C _RIAL PRODUCTION OF RECOMBINANT_INTERLEUKIN-6
-Clark et al., (International Publication Number
Woa8/oo2o6~ published January 14, l9a~) reported the
expression and production o recombinant II,~6 in E. coli
using pCSF~09 (deposited July 11, 1986 as accession number
ATCC 67153). tlowever, the preferred emhodiment reported
therein selec~ively modified the sequenca of pCSF309 a) to
delete its signal peptide leader sequence and its 3'
noncoding sequence: and b) to co~ple the protein coding
sequence to a temperature inducible PL promoter in
association with a temperature sensitive CI repressor. This
strain, containing plasmid pAL309C-781, was not deposited
with the ATCC. In this strain, the bacterially expressed
protein was produced in an insoluble form which first had to
be solubilized and then refolded ~see example V of their
patent application). Alternatively, a plasmid termed pAL-
Sec-IL6-181 was constructed to produce IL-6 by coupling the
cDNA sequence of ~L-6 to a synthetic signal peptide leader
sequence under PL promoter control. Following high
temperature induction o~ the PL promoter in a temperature
sensitive CI repressor strain, IL-6 was isolated from the
periplasm of the transformed cells a~ a homogeneous protein
from which the signal peptide had been removed. No yield o~
purified protein was reported.
Brakenhoff et al., 1987, J. Immunol. 139:~116, reported
the expression of HGF, hybridoma growth factor (IL- 6), in
different E. coli strains containing any one o~ seven
specific IL-6 cDNA clones. Media was screened for IL-6
activity to iden~ify the clones producing IL-6. Th~ activity
varied at least 5,000 fold between the clone~. The most
active construct (clone 7) was missing the first 43 amino
acids of the peptide, but that construct produced more than
20X as much activity as the next active clone. The
constructs containing the complete amino acid coding region
of HGF (IL-6) had the least activity ~for example, clone I5
fl~
starting at position -62 from the ATG start site had 1.7
kU/ml of activity compared to clone 7 which had 10,000
kU/ml). In order to detect activity, recombinant IL-6 was
first separated from cytosolic proteins of clones 7 and 15.
Activity was assayed only after elutlon of indivldual SDS-
PAGE slices. Even though activity was detected~ no IL-6
protein was apparent ih the gel protein profile.
Tosato et al., 1983, Science 239:502, reported the
production of an antiserum after immunization with
recombinant IL-6 that was produced in E. coli but the details
of both the antiserum production methods and the recombinant
IL-6 protein synthesis and purification were not published
(see their reference 10, reported as manuscript in
preparation).
Recombinant BSF-2 (IL-6) was produced in E. coli using
pTBCDF-12 (Hirano et al., 1988, Immunol. Letters 17:413.
Induction of this plasmid resulted in the production of a
fusion protein in which the recombinant IL-6 peptide was
fused to the IL-2 peptide. Protease digestions using
Kallikrein and amino peptidase-P were required to obtain
mature IL-6 protein. However, the details of the procedure
other than this nonenabling description were to be published
elsewhere (see their manuscript in preparation citation).
May et al., 1988, J. Biol. Chem. 263:7760 reported the
production of an insoluble form of IL-6 in bacterla following
the fusion of the IL-6 cDNA in a REV expression vector
(Repligen Corp., Cambridge, MA) to produce a product
containing 34 amino acids of a prokaryotic leader peptide
fused to an IL-6 peptide portion of 182 amino acids. The
30 expressed pro~ein wa~ recovered in an insoluble pellet which
was solubilized in ~M urea, 5 mM DTT and 10 mM ~-
mercaptoethanol in a Tris based bu~fer. The fusion product
was partially purified from ~he numerous other E. coli pellet
proteins by D~AE-sepharose column chromatography using a salt
35 gradient with the solubilizing buffer, followed either by
~0[)4~
immunoaffinity chromatography with a monoclonal antibody to
the prokaryotic leader peptide or FPLC with a Mono Q column
(Pharmacia). No yields of purified protein were reported.
The DEAE-Sepharose peak fraction exhibited antiviral activity
of 0.5-1 IU/ml using a cytophatic effect reduction assay in
FS-4 cell cultures and vesicular stomatitis virus. The
recombinant fused IL 6 protein was injected into rabbits to
produce a polyclonal antlserum. The polyclonal antiserum was
used to characterize the natural IL-6 protein in fibroblastic
FS4 cells following TNF and cycloheximide induction of IL-6
or in human monocytes. Incubation of the antiserum with the
F54 cell medium prevented the induction of immunoglobulins in
CESS cells.
Multiple investig~tors have reported a recombinant I~-6
protein made in E. coli, but no disclosure of the methodology
for reco~binant IL-6 production was provided (Taga et al.,
1987, J. Exp. Med. ~ 4L:967; Gauldie et al., 1987, Proc.
Natl. Acad. Sci. 84:7251: Lotz et al., 1988, J. Exp. Med.
167(3):1253; Muraguchi et al., 198~, J. Exp. Med. 167(2):332;
Reis et al., 1988, J. Immunol. ~ L:1566).
2.5.4. PRODUCTION OF HIGH YIELDS OF RECOMBINANT
INTERLEUKIN-6 IS DIFFICULT
Although nonenabling reports of production of natural
IL-6 in bacteria exist (see Section 2.5.~), production of
25 high yields of the natural protein in bacteria has not been
successful even using recombinant DNA bioengineering methods
(Asagoe et al., 1988, Biotechnology 6:806) (see also Section
5.1 of this applicat~on). The Asagoe study reportC the
inability to produce natural IL~6 as a detectable protein in
30 induced cell extracts using a plasmid construct very
reminiscent to those reported in 5ection 2.5.3. supra. The
unsuccessful Asagoe expression plasmid contained the mature
processed BSF-2 (IL-6) protein coding sequence adjacent to
and in frame with the PTAC promoter and ATG start sequence~
)4~
-16-
Although the plasmid was analyzed in a variety of strain
backgrounds r IL-6 protein was not detected following
induction. 5ignificant amounts of insoluble recombinant IL-6
activity were produced by Asagoe et al. only after induction
of a plasmid containiny a multi-hybrid-fusion protein in
which 1) the PTRyp promoter and ~TG start site were fused to
human growth hormone coding sequence; 2) the human growth
hormone se~uence was fused to an oligonucleotide sequence
producing a Xa factor peptide recognition site (ile-glu-gly-
arg); and 3) the Xa factor recognition site was then fused to
either a glu-phe-met-BSF-2 sequence or an ala-BSF-2 coding
sequence.
This recombinant IL-6 (or BSF-2) peptide was thus the
carboxy terminal portio~ of the human growth hormone/Xa
factor sequence/BSF-2 hybrid fusion product in this
construct. The recombinant IL-6 activity, existing either as
glu-phe-met-BSF-2 peptide in one plasmid construct or ala-
BSF-2 peptide in the other, could be purified only after
solubilization of the fusion protein in 8M urea. Cleavage of
the fusion pro~ein with Xa factor required refolding the
purified fusion protein hy extensive dialysis. The refolding
process wa~ incomplete since the fusion protein preparation
was only incompletely digested by Xa factor.
The Xa factor cleavage produced a heterogeneous mixture
containing intact fusion protein, recombinant IL-6 peptides,
and human growth hormona peptide, in addition to other
partial cleavage products. In order to recover IL-6 activity
in the purified product, the Xa factor/cleavage mixture first
had to be denatured with 6M guanidinium hydrochloride. After
30 chromatographical purification of the recombinant IL-6
peptide, it was subjected to an additional round of extensive
dialysis in order to recover a refolde~ active peptide. The
rPsultant recombinant peptides, either glu-phe-met-BSF-2 or
ala-BSF-2, were examined only for their activity in a B-cell
35 stimulatory factor assay. The yield reported for the
2~
-17-
production process was approximately 5% (3 mg of purified
actiYity was recovered from an initial 100 mg of fusion
protein of which 58 m~J was IL-6 peptide; 3/58 = 5.17%).
A need thus still exists for a convenient, relatively
inexpensive method of prod~lcing high yields of IL-6.
Recombinant methods using mammalian cel1s tend to be rather
costly. Although bacterial production is a lower cost
al~ernative, previous attempts to produce IL-6 in transformed
bacteria have not resulted in commercially feasible yields.
The present invention provides DNA constructs and methods for
producing IL-6 in bacteria which permit the production of the
protein in high yields, typically about 20% of total cytosol
protein. The invention also provides a cysteine-free form of
IL-6 which retains the biological activity of the native IL-6
molecule.
1~
3. SUMMARY OF T~IE INVENTION
The present invention is directed to the economical
production of recombinant synthetic cysteine-free IL-6
proteins produced by bacteria as well as active peptide
fragments of each of the proteins. The proteins produced are
either cysteine-containing or cysteine-free, and retain IL-6
biological activity. Bacterial cultures express either
recombinant protein as a high percentage, generally at least
20%, of total soluble cytosol protein. The recombinant IL 6
protein does not require treatment with harsh denaturing
agents like 8M urea, or ~-mercaptoethanol to solubilize the
protein from a pellet. As used throughout the present
specification and claims, the phrase ~Isubstantially soluble
30 in wat~r" re~ers to this property.
In the present invention, the IL-6 is produced
recombinantly in a unicellular host by expression of a
tripartite fusion protein. The fusion protein comprises
first, second and third peptide portions. The first peptide
portion has IL-6 activity; the second peptide portion is a
X0~1~2Gl
-18-
chemically or enzymatically cleavable sequence which links
the first peptide portion to the third peptide portion; the
third peptide portion is a protein or portion thereof which
is capable of being expressed by the unicellular host, and
which pre~erably has a detectable function. The three
peptide portions ara referred to as "first", "second" or
"third" only for convenience, and should not ~e read as
requiring a specific order, relative to the promoter
controlling expression of the protein: the positions of the
first and third peptide portions are interchangeable.
The third peptide portion provides a protein or portion
thereof which the unicellular host cell can make. The
presence of the carrier DNA that expresses the third peptide
facilitates production Qf the aucaryotic IL-6 protein. In
its detectable function, (for example, enzymatic activity),
the third peptide portion also provides a means by which
transformed clones producing the fusion protein can be
readily identified and isolated. The first peptide portion,
which possesses the desired biological activity, is separated
from the remainder of the fusion protein by cleavage of the
second peptide portion. The recombinant IL-6 peptide is then
separated and purified from the protein mixture by methods
known in the art such as by higX performance liquid
chromatography ~HPLC) to yield a pure protein that is active
in IL-~ assays. The present invention also contemplates
nucleotide sequ~nces encoding the fusion protein, recombinant
vectors comprising these nucleotide sequences, as well as
unicellular hosts transformed with these vectorsO
In a particular embodiment, a synthet~c peptide having
IL-6 activity is produced with all four cysteines of the
native IL-6 sequence replaced by serine residues. Unless
otherwise specified, the term synthetic cysteine-free IL-6 or
recombinant cysteine-free IL-6 is used in the specification
and claims to identify a synthetic peptide having IL-6
activity with all four cysteines of the native IL-~ sequence
19--
replaced by serine residues. The peptide so produced
surpr~singly retains its biological activity. For example,
the cysteine-free form of IL-6 has heen shown to exhibit
hepatocyte activation and B-cell stimulation. The proteins
o~ the present invention may have antiviral activity and may
prevent viral ~nfection of cells. The cysteine-free
synthetic protein may be nonpyrogenic or at least
significantly less pyrogenic than native IL-6 protein.
The proteins of the present invention also include
different cysteine-free and cysteine-containing IL-6 peptides
which have deletad up to 27 amino acid residues from the IL-6
amino terminus, and/or which add up to 50 amino acid residues
on the amino terminus and/or up to 350 amino acids on the
carboxy terminus of the~IL-6 sequence, and retain the
relevant biological activity. Thus, surprisingly, and unlike
many other biologically active peptides, the basic IL-6
sequence can be modified substantially without any
significant loss of activity. The invention therefore
encompasses the gene sequences encoding the cysteine-free
peptides, as well as truncated and extended cysteine-free or
cysteine-containing peptides which retain IL-6 activity.
The present method represents an improvement over known
recombinant methods for producing IL 6 in that it provides
means for producing IL-6 in commercial quantities in a
recombinant/vector system. Previous IL-6 fusion protein
constructs, such as the one described by Asagoe, supra, have
not been successful in obtaining expression o~ the protein in
large quantities, and also require extensive harsh
purification and refolding procedures in order to obtain a
30 functional protein. This treatment with harsh denaturing
agents is not required in the present method, however, nor is
refolding of either the fusion protein or the cleavage
protein necessary.
The peptides producad by culturing these recombinant
35 hosts retain characteristic IL-6 activity in their
-20-
stimulatory effects on the immune system and on therapeutic
cell àctivity. The present invention also contemplates
method.s of treatment of viral disease, immunodeficiencies and
hepatic disorders, as well a~ overall modulation o~ immune
response, by administration of the claimed synthetic IL-G
peptide.
3.1. DEFINITIONS
ATCC : American Type Culture Collection
bp : base pair
~SF-2 : B-cell Stimulation Factor
cDNA : Complementary DNA
CHO : Chinese Hamster Ovary
CSF : Colony Stimulating Factor
DHFR : Dihydrofolate Reductase
DNase : Deoxyribonucleic acid nuclease
ELISA : Enzyme Linked Immunosorbent Assay
G-CSF : Granulocyta Colony Stimulating
Factor
HGF : Hepatocyte Growth Factor
HPGF : Hybridoma Plamacytoma Growth Factor
HPLC : High Pexformance Liquid
Chromatogxaphy
HSF : Hepatocyte Stimulating Factsr
IGF : Insulin-Like Growth Factor
Ig~, IgM : Immunoglo~ulin G, M
IL-1, IL-3,
or IL-6, : Interleukin 1, 3, or 6
IPTG : Isopropylthiogalactoside
INF : Interferon
kDa : kilo Dalton
Kb : Kilobases
LB : Luria Broth
mRNA : messenger RNA
PAGE : Polyacrylamide Gel Electrophoresis
fil
PBS : Phosphate buffered saline
PDGF : Platelet Derived Growth Factor
Poly ~rI)'~rC): RNA with a strand of ribo-Inosine
duplexed with a ribo-Cytosine strand
PL : Promoter l.eft of Lambda
PR : Promoter Right of Lambda
PTAC' PTRC Synthetic Promoter of Tryptophan-Lac
Combination
RNase : Ribonucleic acid nuclease
SDS : Sodium Dodecylsulfate
S~D sequence: Shine-Dalgarno sequence
SH : Sulfhydryl
TNF : Tumor Necrosis factor
X-gal ~Bromo-chloro-indolyl-~-D-
galactopyranoside
YT : Yeast-tryptone growth media
4. RIEF DESCRIPTION OF THE FIGURES
FIG. l. Diagram of the construction of p360. The
plasmid p360 is a plasmid expression vector constructed to
express the cysteine free synthetic IL-6 like peptide as a
protein of approximately 22-23 KDa in E. coli.
Oligonucleotide 737 [AGC TGA TT~ AAT AAG GA~ GAA TAA CCA TGG
CTG CA 3'] and 73 8 [ GCC ATG GTT ATT CCT CCT TAT TT~ ATC 3'~
were fused and replaced the Hind III-Pst I fra~ment of
pB~(+), a phagemid purchased from Stratagene Systems, to form
p350. The sequence for the synthetic cysteine free IL-6 like
peptide fro~ plasmid p337-l, which contains a peptide
terminating codon, replaced the Nco I-Eco RI fragment of p350
30 to produce plasmid p360. In p360, expression ~rom the
induc~ble lac promoter through the minicistron sequence and
through the syn~hetic cysteine free I~-6 like sequence fails
to produce a peptide of the size of IL-6 in extracts of E.
coli (see Figure 3 rox gel of expressed proteins).
FIG. 2. Diagram of the construction of p369. The
-22-
fragment containing the sequence for the synthetic cysteine
free IL-6 like peptide of plasmid p365, which contains no
peptide terminating codon is this construction~ replaced the
Nco I-Bam HI fragment of p350 tdescribed in Figure 1) to form
plasmid p369 wherein the synthetic cysteine free IL-6 like
peptide sequence was fused in frame to the alpha-
complementing fragment of ~-galactosida~e. Induction of the
lac promoter oP p369 fails to produce a detecta~le peptide
which would be the fusion product between the synthetic IL-6
like peptide and the alpha-complementing portion of ~-
galactosidase (see Figure 3 for gel of the expressed
proteins).
FIG. 3. Expression and nonexpresslon of plasmid
constructs containiny sequences for the synthetic cysteine
free IL-6 peptlde. Cultures of E. coli cells carrying
different plasmids were induced with IPTG. Samples of cell
cultures were lysed and electrophoresed in SDS-polyacrylamide
gels. The gels were stained with coomasie blue to visualize
proteins. Lanes a and b are duplicate aliquots from
different tubes of the same culture. Lanes la and b are from
cells bearing plasmid p369. The ëxpected induction protein
should be a ~usion protein o~ the alpha-complementing portion
of ~-galactosidas~ and the synthetic IL-6 like peptide. No
large molecular weight protein representing the fusion is
present. Lanes 2a and b are from cells bearing plasmid p367,
the successful expression vector ~or the fusion product of
the synthetic IL-6 like peptide/collagen/~-galactosidase.
Note the large amount o~ high molecular weight fusion protein
in the approximately 130 kDA position of the gel. Lanes: 3a
30 and b are from cells bearing plasmid p360 which were expected
to produce a peptide of 22-23 ~Da, the size of deglycosylated
IL-6. The arrowhead marXs the position of the expected small
peptide.
FIG. 4. Diagram of the construction of p340-1. The
35 steps in the construction of p~40-1 are described in Section
5.2. Oligonucleotide 237 was [AAT TCT AAT ACG ACT CAC TAT
~GG GTA AGG AGG TTT AAC CAT GGA GAT CTG 3']. Oligonucleotide
238 was ~GAT CCA GAT CTC CAT GGT TAA ACC TCC TTA CCC TAT AGT
GAG TCG TAT TAG 3'] Oli~onucleotide 703 was [CAT GTA TCG ATT
F AAA TAA GGA GGA ATA ACC 3']. Oligonucleotide 704 was [CAT
GGG TTA TTC CTC CTT ATT TAA TCG AT~ 3']. PTRC is ~he hybrid
promoter tryptophan/lac. "Term" represents the terminatlng
codon th~t is in the same reading frame as the preceeding
lnitiating methionine codon. APR represents the rightward
promoter from bacterlophage lambda. Ampr represents the
ampicillin resistance marker from pBR322. ori represents the
plasmid origin of DNA replication from pBR322. ~-gal
represents the E. coli ~-galactosidase gene region.
FIG. 5. Complete ,nucleotlde sequence (5'-3') of the
synthetic recombinant IL-6 gene and its predicted amino acid
sequence. The coding sequence o~ the fused gene begins at
nucleotide 3 (Met). The amino acid sequence of the
recombinant IL-6 protein extends to nucleotide 557 (Met)
where it is fused to the collagen linker of the fusion
protein via a ~erine residue. The asterisk desiqnate~ the
location of the TAG stop codon normally found in the native
cDNA sequence. Modified amino acid residues are shown in
script below the sequence.
FIG. 6. Assembly of the gene ~or the synthetic
cysteine-free IL-6 like peptide. Synthetic oligonucleotides
were annealed and ligated to form the four synthetic double
stranded sub-fragments of the synthetic gene. The insert
contained in p333 was formed by ligating two pairs of
complementary oligonucleotides~ each bearing internal ~ive-
30 base 5' complementary overhangs (nucleotides 59-63 in Figure
5). Construct p334 was composed o~ three pairs of annealed
oligonucleotides joined at residues 162-167 and 220-225.
Complementary overhangs for the assembly of p331 were located
at nucleotides 312-315 and 358-361. Construct p332 was
35 ligated at positions 457-46i and 512-517.
2~2fi~
-~4-
The synthetic fragments were ~loned into a modified pBS
M13+ Stratogene vector by insertion into the multiple cloning
site of the circular vector via synthetic 4-6 base overhangs
(represented by the hatched boxes). The coding sequence for
the synthetic cysteine-free IL-6 like gene is shown in black.
Open boxes represent the modified pBS M13~ sequences.
Fragments were assembled by digesting with appropriate
enzymes, puri~ying the desired bands by electroelution from
agarose gels and religating. Removal of the termination
codon in p337 was accomplished by substituting a second
synthetic fragment containing a serine resldue and a
compatible Eco RI overhang (RI). Ligation of this fragment,
to yield construct p365, resulted in the addition of another
Bgl II site and loss of.the Eco RI site. Nucleotide 564
identifies the position of the last base contained in the Bgl
II site required to fuse the insert in p365 with the collagen
linker of the expression vectort Restriction sites are
identified as follows: B, Bam HI; 8g, Bgl II; RI, Eco RI; RV,
Eco RV; H, HinD III; N, Nco I; S, Stu I; X, Xba I.
FIG. 7. Diagram showing the construction o~ plasmid
p367, the success~ul expression vector fox producing the
synthetic cysteine-free IL-6 peptide. The steps in the
construction of p367 are described in sections 5.2-5.5 in the
taxt.
FIG. 8a. Plasmid map of recombinant plasmid p367
showing the location of the IL-6 sequence (designated HSF).
The assembled synthetic cysteine free recombinant IL-6 gene
contained i~ an (Nco I - Bgl II) insert was introduced
between the Nco I and Bam ~ I restrîction sites of the
vector. The open box (C) indicates the location of the 60
amino acid collagen linker through which recombinant ~L-6 is
fused to ~- galactosidase. Expression of the complete fusion
gene produces a 130 kDa hybrid protein.
FIG. 8b. Plasmid map of recombinant plasmid p367
showing the location of the sequence for the synthetic
cysteine-free IL-6 like peptide (designated HSF). The vector
shown is a significantly modified version of pJG200 as
describin~ in the application in section 5. The assembled
synth~tic IL-6 likQ gene contained in an (Nco I-BGl II)
insert was introduced between the Nco I and ~am HI
r~striction sites of tha vector. The open box (C) indicates
the location of the 60 amino acid collagen linker through
which the synthetic IL-6 like peptide is fused to ~-
galactosidase. Expression of the complete fusion gene
produces a hyhrid protein of approximately 130 kDa (see
Figure 3 ~or expression of the plasmid in induced E. coli).
FIG. 9. NaDodS04-PAGE analysis of synthetic cysteine-
free IL-6 protein purification. Samples collected at various
stages o purification ~ere electrophoresed on 10%
polyacrylamide- NaDodS04 gels under reducing conditions as
shown. Protein molecular weight markers are shown in lanes
M. E. coli cells grown in 10 L batch culture were lysed by
addition of lysozyme followed by brief sonication (Lane 1).
The unusually hydrophobic fusion protein was purified by
repeated ammonium sulfate precipitation after solubilization
in PBS- Sarkosyl (Lane ~). Partially purified ~usion protein
was then digested with collagenase to release the 23 kDa
recombinant IL-6 moiety (Lane 3j. After removing the
majority of the cleaved, soluble ~-galactosidase by
precipitation in 40% ammonium sulfate, synthetic I~-6 was
recovered in near homogeneous form as a pellicle upon
increasing the ammonium sulfate concentration to 70%
saturation (Lane 4). Reverse phas~ ~PLC was used in the
final step of the purification~ The protein was recovered in
30 a single peak fraction eluting at 54% acetonitrile (Lane 5~.
FlG. 10. Stimulation of fibrinogen sythesis by
synthetic cysteine-free recombinant IL-6 as prepared
according to the methods described in ~5.2 to 5.8.
Confluent FAZ~ cell monolayers were treated with varying
35 concentrations of purified recombinant cysteine-free IL-6 in
'12~
- ? 6 -
the presence of 10 7 M dexamethasone. After five hours of
stimulation, cell medium was recovered and filtered through a
nitrocellulose membrane using a dot-blot apparatus. Secreted
fibrinogen level~ were cletermined by solid phase immunoassay
using a polyclonal fi~rinogen antiserum as primary antibody
and alkaline phosphatase conjugated anti-IqG antiserum as
secondary antlbody. Bound antibody was visualized by adding
chromog0nic substrate~ ~BCIP ar.d NBT) to the blots.
Concentrations of secreted protein were determined by
scanning laser densitmetry and comparison of signal
intensities against dilutions of a purified rat fibrinogen
standard.
FIG. 11. B-cell differentiation assay. CH12.LX cells
(2 x 10 cells/well) were co-cultured in the presence or
absence o~ antiyen (SRBC) prior to treatment with various
concentrations of recombinant synthetic cysteine-free I~-6,
as prepared according to the methods described in 5.2 to
5.8. After 48 hrs, treated CH12.LX aliquots were mixed with
a freshly washed suspension of SRBC containing guinea pig
complement, and transferred to Cunningham chambers for
incubation at 37 de~rees Centigrade. After 30 minutes, cells
were returned to room temperature allowing hemolytic plaques
to develop. Results of the assay are expressed as plaque
forming colonies per million viable cells (shown a~ Pfc on
the y axis). Bacterial lipopolysaccharide, included in the
assay as a positive stimulatory control, gave a response of
10,726 pfc/million cells. The results shown represent mean
values o~ duplicate assays.
FIG. 12a and b. Diagrammatic illustration of ~he
30 construction o~ pTrpE/EK/cfIL-6. The details of the
construction are found in the text in Section 5~10.1.
FIG. 13. Graphic depiction of ~ime 0 colony assay for
stimulation of progenitor cells by various growth factors and
combinations thereof.
3~ FIG. 14. Number of nonadherent cPlls after 7 days u~
-27-
liquid culture. This assay is described in Example ll.
FIG. 15. Imclone vs. Endogen IL-6 on 7tdrl~
FIG. 16. Imclone vs. bm IL-6 on 7td.1.
FIG. 17. Imclone v. Genzyme IL-6 using 7td.1.
FIG. 18. The sequence in pTrpE/EK/cfIL-6 from the
enterokinase site through the amino terminal sequence o~
cysteine-free IL-6 to the natural carboxy terminus of IL-6
followed by three stop codons.
5. DETAILRD DESCRIPTION OF THE INVENTION
5.1. INITIAL EXPRESSION CONSTRUCTS
Initial attempts to express high levels of
interleukin-6 in bacterial cells resulted in the production
of no detectable or commercially useful amounts of the
15 protein. One attempt to produce commercially useful amounts
of protein resulted in the synthesis of plasmid p360. In
this expression vector, the synthetic cysteine-free IL-6
sequences from p337-1 were inserted immediately downstream
from the minicistron sequence produced by the fusion of oligo
737 ~AGCTGATTA~ATAAGGAGGAATAACCATGGCTGCA-3'] and oligo 738
~GCCATGGTTATTCCTCCTTATTTAATC-3']. In this construct the
synthetic IL-6 peptide had a termination codon and was not
fused to the ~-galactosidase protein. Induction was expected
to yield a non-glycosylated synthetic IL-6 peptide slightly
larger than 22 KDa. The construction of plasmid p360 is
graphically diagrammed, but not to scale, in Figure 1. The
inductiQn of expresslon of the synthetic IL-6 peptide by the
addition of IPTG to cells carrying p360 resulted in no
detectable peptide (see Figure 3, lanes 3a and 3b for absence
30 of protei~ at the arrow).
Similarly induct~on of a strain carrying plasmid p369,
constructed as diagrammed in Figure 2, wherein the synthetic
IL-6 like p~ptide was fused in frame with the alpha-
complementing fragment of ~-galaotosidase resulted in no
35 detectable protein (see lanes la and b, Figure 3).
-28-
These unsuccessful attempts to produce synthetic IL-6
c~n ~e compared in Figure 3 with the successful production,
as described i _ a, of the synthetic cysteine-free
IL-6/collagen/~-galactosidase fusion protein (see the large
amount o~ the 130 KDa protein in lanes 2a and b o~ Figure 3).
5.2. CONSTRUCTION OF AN IL-6 EXPRESSION VECTOR
5.2.1. THE INITIAL VECTOR pJG200
Plasmid pJG200 was the starting material that was
modified to produce a successful IL-6 expression vector. The
initial plasmid, pJG200, contained target cistrons that were
fused in ~he correct reading frame to a marker peptide with a
detectable activity via a piece of DNA that codes for a
protease sensitive linker pepkide (Germino and Bastia, 1984,
Proc. Natl. Acad. Sci. USA 81:~692; Germino et al~, 1983,
Proc. Natl. Acad. Sci. USA 80:6848). The promoter in the
ori~inal vector pJG200 was the PR promoter of phage lambda.
Adjacent to the promoter is the CI857 thermolabile repressor,
followed by the ribosome-binding site and the AUG initiator
triplet of the Cro gene of phage lambda. Germino and Bastia
inserted a fragment containing the triple helical region of
the chicken pro-2 collagen gene into the Bam HI restriction
site next to the ATG initiator, to produce a vector in which
the collagen sequence was ~used to ths lacZ ~-galactosidase
gene sequence in the correct translational phase. A single
Bam HI restriction sites was regenerated and used to insert
the plasmid R6K replication initiator protein coding
sequence.
The plasmid pJG200 expressed the R6K replicator
initiator protein as a hybrid fusion product following a
temperature shift which inactivated the CI~57 repressor and
allowed transcription initiation ~rom the PR promoter. Both
the parent vector construct with the ~TG initiator adjacent
to and in frame with the collagen/~-galactosidase fusion
(non~nsert vector), and p~G200 containing tbe R6K r~plicator
-29-
initiator prot~in joined in frame to the ATG initiator codon
(5') and the collagen/~-galactosidase fusion t3') (insert
vector), produced ~-yalactosidase activity in bacterial cells
transformed witll the plasmids. As a result, strains
containing plasmids with inserts are not distinguishable from
strains containing the parent vector with no insert.
5~2.2. REMOVAL OF THE P CI857 REPRESSOR
AND AMINO TE~MINBS OF CRO
The first alteration to pJG200 in this invention was
10 the removal and replacement of the Eco RI~Bam HI fragment
that contained the PR promoter, CI857 repressor and amino
terminus of the cro protein which provided the ATG start site
for the fusion proteins. An oligonucleotide linker was
inserted to produce the p258 plasmid, which maintained the
15 Eco RI site and also encoded the additional DNA sequences
recognized by Nco I, Bgl II and Bam HI restriction
endonucleases. This modificatlon provided a new ATG staxt
codon that was out of frame with the collagen/~-galactosidase
fusion. As a result, there is no ~-galactosidase activity in
20 cells transformed with the p258 plasmid. In addition this
modification removed the cro protein amino terminus so that
any resultant recombinant fusion products inserted adjacent
to the ATG start codon will not have cro encoded amino acids
at their amino terminus. In contrast, recombinant proteins
25 expressed from the original pJG200 vector all have cro
encoded amino acids at their amino terminus.
5.2.3. AD~ITION OF THE P PROMOTER, SHINE
DA~GARNO SEQUENCET~D ATG CODON
In the second step oP construction of the IL-6
expression VeGtOr~ a restriction fragment, the Eco RI-Nco I
fragment of pKK233-2 (Pharmacia Biochemicals, MilwauXee, WI~,
was inserted into the Eco RI-Nco I restriction sltes o~
plasmid p25~ to produce plasmid p277. As a result, the p277
ZO~
-30-
plasmid contained the PTAC ~also known as PTRC) promoter of
pXK233-2, the lacZ ribosom~ binding site and an ATG
initiation codon. In the p277 plasmid, the insertion of a
target protein sequence allows its transcription from an IPTG
inducible promoter in an appropriate strain background. The
appropriate strain background provides sufficient lac
repressor protein to inhibit transcription from the uninduced
Pq,AC promoter. B~ca~se cells can be induced by the simple
addition of small amounts of the chemical IPTG, the p277
plasmid provides a significant commercial advantage over
promoters that require temperature shifts for induction such
as the PR promoter of pJG200. Induction of commercial
quantities of cell cultures containing temperature inducible
promoters would otherwise require heating large volumes of
cells and medium to produce the temperature shift necessary
for induction. For example, induction by the PR promoter
requires a temperature shift to inactivate the CI857
repressor inhibiting pJG200's promoter. One additional
benefit of the promoter change is that cells are not
subjected to high temperatures or temperatuxe shifts. High
temperatures and temperature shi~ts result in a heat shock
response and the induction of heat shock response proteases
capable of degrading recombinant proteins as well as host
proteins (See Grossman et al., 1984, Cell 38:383; Baker et
al., 1984, Proc. Natl. Acad. Sci. 81:6779).
5.2.4. IMPROVEMENT OF T~E RIBOSOME BINDING_SITE
The p277 expression vector was further modified by
insertion o~ twenty-nine base pairs, namely
5'CATGTATCGATTAAATAAGGAG~AAT~AC3' into the Nco I site of p277
to produce plasmid p340. This sequence i5 related to, but
different than, one portion of the Schoner "minicistron"
sequence (described in section 2.5.1). The inclusion of
these 29 base pairs provides an optimum Shine/Dalgarno site
for ribosomal/mRNA interaction. The final p340 vector
31-
significantly differs from pJG200 because it contains a
highly inducible promoter suitable ~or the high yields needed
for commercial preparations, and improved synthetic ribosome
binding site region to improve translation, and means to
provide a visual indicator of fragment insertion. The steps
in the construction of vector p340-1 are diagrammed in Figure
.
5.3. M FICATIONS OF THE INTERLEUKIN-6 SEQUENCE
The coding sequence used for the exprassion of
synthetic cysteine-free recombinant IL-6 was based on the
cDNA sequence of human IL-6. The coding sequence was
cons~ructed using 11 complementary synthetic oligonucleotide
pairs. Several mutatio~s were introduced into the
recombinant IL-6 coding sequence during oligonucleotide
synthesis to enable proper assembly of the gene sub-fragments
on the on2 hand, and to ensure efficient expression of the
assembled gene on the other. Nucleotide sequence
modifications, designed to introduce novel restriction sites
for use in joining the gene sub-fragments, were incorporated
within the coding sequence in such a manner as to avoid
altering the amino acid composition of tha synthetic gene
with respect to the native IL-6 protein sequence. The
modifications included: 1) replacement of the 5' terminal
118 nucleotides, which encode th2 28 amino acid signal
sequence normally found in the native IL-6 gene, with a
methionine codon (nucleotides 3-5); ii) replacement o~ the
[Pro] residue in the native IL-6 protein sequence with tGly]
(nucleotides 6-8) in the syn hetic IL-6 sequence; iii)
replacement of the normal TAG stop codon with a serine codon
(nucleotides 558-560) to effect fusion of the synthetic lL-6
protein with the collagen linker. and iv) replacement of the
four internal cysteine residues with serines (nucleotides
135-137, 153-155, 222-224, 252-25~) to produce a synthetic
35 lL-6 protein that is unable to form disul~ide bonds. ~he
~0~26~
eequence of the modified synthetic cysteine-free protein is
included in Figure 5. Those skilled in the art will
recognize that other modifications in the sequence of the
synthetic peptide are useful in the present invention. The
invention as contemplated includes the modi~ication of other
amino acids of the peptide sequence or other nucleotides of
the DNA sequence.
5.4. OLIGONUCLEOTIDE SYNTHESIS AND AS5EMBLY
Assembly o~ the synthetic oligomers was carried out in
three ~teps. Initially, oligonucleotides bearing
complementary overhangs were annealed and ligated to produce
four separate double stranded fragments, one composed of four
oligonucleotides (2 per strand) and three composed of six
oligonucleotides each (3 per strand). Before assembly,
synthetic oligomers were kinased with 10 units of T4
polynucleotide kinase. To prevent concatenation during
ligation, the 5' terminal oligomers on either strand were not
phosphorylated. Subsets of these double stranded
oliqonucleotides were assembled in separate annealing and
2 ligation reactions to produce four sub ~ra~ments, each
repre~entin~ approximately one fourth of the recombinant
synthetic cysteine-free IL-6 coding sequence.
A modified plasmid was constructed to allow for DNA
amplification and ease in sequencing each oligomer. A pBS
M13+ cloning vector (Stratagene) was modified by insertion of
a 28 base oligonucleotide adapter
(5'AGC~TCCATGGTCGCGACTCGAGCTGCA-3') between the Hind III and
Pst I site~ of its multiple cloning region. As a result, the
modified plasmid, designated p2~7, no longer contains its
original Sph I restriction site bu~ encodes additional sites
for Nco I, Nru I and Xho I.
The synthatic oligomers were separately cloned into the
modified pBS M13+ vector p287 to allow DNA amplification and
35 sequence verification by dideoxy-nucleotide sequencing.
~o~
Insertion of the assembled fragments into th~ modified vector
produced recombinant plasmids p333, p334, p331, and p332,
each containing a portion of the synthetic cysteine-free IL-6
protein coding region proceeding ~rom amino to carboxy
terminus respectively. Following ligation, each plasmid DNA
was transformed separately into competent E. coli JMlol.
The second step of the assembly involvQd the
construction of two vectors that encoded the amino and
carboxy halve~ ~f synthetic cysteine-free IL-6. These were
obtained by ligating the inserts of p333 and p334 to produce
the N- te~minal coding vector p336 and by joining the inserts
of p332 and p331, to form the C-terminal coding vector p335.
In the third and final step of the construction,
inserts were subsequently combined to yield the entire coding
sequence of the recombinant synthetic cysteine-free IL-6
gene. The insert released from plasmid p335 was ligated into
p336. The resulting plasmid p365 contained the complete
coding sequence o~ synthetic IL 6 inserted between the Nco I
and Eco RI sites o the modified pBS M13~ vector p2~7. Th~
constructions are shown diagrammatically in Figure 6.
2n
5.5. CONSTRUCTION OF THE p367 VECTO~ FOR
EXPRES5ING THE SYNTHETIC IL-6/COLLAGEN/~
GALACTOSIDASE FUSION PRODUCT
The assembled cysteine-free recombinant IL-6 gene was
25 excised from p365, and inserted into plas~id p340 between the
Nco I site, which encompasses the initiating methionine, and
the BamH I site adjacent to the collagen linker as depicted
in Figurs 7. The resulting vec~or p367, diagrammed in Figure
8 but not to scale, was used to transform E. coli JM101.
30 Recombinant colonies were selected on the basis of antibiotic
resistance and by appearance o~ blue coloration in the
presence o~ X-Gal. The size of ~he insert DNA was con~irmed
by mini-lysate extraçtion ~ollowed by polyacrylamide gel
electrophoresis.
3S
~4~fi~1.
-34-
A tripartite fusion protein composed of synthetic
cysteine-free IL-6, a sixty amino acid collagen linker and
fl-galactosidase was produced in tran~formed bacteria ~See
Figure 3 and Figure 9). As predicted from the gene sequence,
ampicillin resistant transformants carrying the modified IL-6
expression plas~id produced blue colonies upon addition of
the inducing agent IPTG and the chromogenic ~-galactosidase
substrate X-Gal. Synthesis of the fusion protein by IPTG-
induced transformants was independently confirmed by western
blot analysis of a total E.coli lysate using a monoclonal
anti-~-galactosidase antibody obtained from Promega BiotecO
The monoclonal anti-~-galactosidase antibody used in the
Western blot recognized a band with an apparent molecular
weight of 130,000 kDa.
5.6. INDUCTION OF LARGE AMOUNTS OF THE MODIFIED
INTERLEUKIN-6 FUSION PROTEIN
Transformed JM101 containing plasmid p367 were grown in
10 L batch cultures using a Magnaferm fermentor tNew
Brunswick Scienkific). Cells were grown in 2x YT containing
20 100 ug/ml ampicillin and induced at A550 = 1.5 by addition of
5 mM isopropylthio-~D-galactopyranoside (IPTG, Sigma). At
to 12 hours post-inoculation, when tlle c 11 numbers and ~-
galactosidase activity had reached maximal levels, cells were
pelleted by centrifugation at 5000 x g and stored frozen at
25 -20~C until needed.
5.7. PURIFICATION OF THE FUSION PROTEIN
Frozen E. coli cell pellets were processed in aliquots
of 100 g (wet wei~ht) by washing with TNS buffer ~30 mM
30 Tris.Cl, pH 7.4; 30 mM NaCl, 0.05% sodium lauroyl sarcosine~.
Washed cells were lysed in 450 ml TNS containing 1.5 mM EDTA
and 0O5 mg/ml lysozyme. After incubating the suspension on
ice for 30 minutes, complete lysis was ensured by subjecting
cells to three cycles of freeze-~hawing and brief sonication.
2~,~
Soluble prot2ins, devoid of ~-galactosidase activity, were
removed by three repeated washings in TNS followed by
centrifugation at 10,000 x g for 20 minutes. The ~inal
pellet, weighing approximately 40 g, was resuspended in 60 ml
of 10% sarkosyl and diluted to 2~4 liters with phosphate
buff~red saline ~PBS). Insoluble material was removed by
centrifugation and the supernatant made 40% with respect to a
saturated solution of ammonium sulfate. After incubating the
extract on ice for 30 min. precipitated protein was again
recovered by centrifugation and sub~ected to two additional
rounds of ammonium sulfate precipitation. The ~inal extract
was resuspended in 250 mls of 20 mM Tris.Cl, pH 7.4 and 150
mM NaCl, divided into 4 ml aliquots and stored at -20-C until
ready for further proce~sing.
5.8. CLEAVAGE OF THE FUSION PROTEIN AND PURIFICATION
OF THE SYNTHETIC CYSTEINE-FREE INTERLEUKIN-6
The recombinant synthetic cysteine-free IL-6 protein
was purified to homogeneity using r~verse phase HPLC. Thawed
extract (4 mls) was sonicated briefly to disperse aggregates,
20 added to pre- treated collagenase, and incubated for 45
minutes at room temperature. The majority of the cleaved ~-
qalactosidase was removed by adding 0.5 volumes of saturated
ammonium sulfate, incubating on ice for 30 minute~ and
pelleting the insoluble material. The cleaved recombinant
25 cysteine-frea IL-6 was concentrated by bringing the total
~olum~ of the supernatant to 13 mls with saturated ammonium
sulfate, incubating on ice for 30 minutes, and centrifuging
to compact the insoluble protein into a floating pellicle.
Liquid was drained by puncturing the tube, and the remaining
30 pellicle was resuspended in 0.5 mls of Tris buffered saline.
The resuspended recombinant cysteine-free IL-6 was
prepared for reverse phase HP~C by adding an equal volume 9f
60% acetonitrile, 0.1% trifluroacetic acid. Insoluble
material was pelleted and the clarified supernatant was
~ 3~
-36-
loaded onto a 250 mm, 4.6 mm ID reverse phase column (Vydac,
218Tp, Cl8, 10 ~Im- Alltech Associates) in an injection volume
of 0.5 to 1 ml. The mobile phase consisted of varying
concentrations of solvent B (60% acetonitrile and 0~1%
triflurooacetic acid) relative to solvent ~ (0.1
trifluoracetic acid). The flow rate was 0.5 ml 1 min , and
the system programmed to deliver two consecutive linear
gradients, from 25~ to 80% B in five minutes, and 80% to 100%
B over 54 minutes. Protein eluted from the column at 54%
acetonitrile and collected in a single peak fraction was the
purified synthetic cysteine-free interleukin-6 peptide which
migrated at 2~ kDa following SDS-PAGE. The steps of the
purification are indicated in Figure g and the yields at each
step are provided in Table 1.
x~
-37-
TABLE 1
PURIFICATION OF SYNTMETIC CYSTEINE FREE IL-6
Yield
Purification step _ (m~lOOg wet cell weiqh~)
Total Fusion
protein protPin Galactosidase rIL~6
Whole cells12000 1080 --- 156a
10 ~ashed Lysateb1420 570 ~~~ 81a
Collagenased lysate 1220 --- 491 75
Digest supernatantd 80 --- 22 27
HPLC fractione16 16
~5 Relative Purificatio
Purification step _(~raction of total_protein~
Fusion Protein Galactosidase rIL-6
Whole cells 0.09 --- 0.013a
Washed Lysateb0.40 --- 0.057a
20 Collagenased lysateC --- 0.40 0.061
Diges~ supernatantd --- 0.27 0.34
HPLC fractione ~ - 1.0
a Hypothetical value based on a calculated molecular weigh~
ratio of about 1:7 for rIL-6:fusion protein.
25 b Insoluble cellular material after washing with TNS buîfer,
detergent solubilization, and repeated ammonium sulfate
c precipitation.
Washed lysate fraction after digesting 1 hour with
d collagenase.
Supernatan~ fraction after precipi~a~ion of collagenase~
lysate with 3~% satura~ed ammonium sulfate.
30 e Fraction containing rIL-6 after running the digest
supernatant on reverse phase HPLC column.
,,
To confirm the identity of the 23 kDa protein, HPLC-
35 purified material was subjected to direct N-terminal
--38--
automated protein sequencing. The amino acid residues
tMGVPPGED) identified after seven cycles of sequential
degradation coincided with the predicted N-terminal amino
acid composition of the recombinant protein as deduced from
the synthetic recombinant cysteine free IL-6 gene.
Confirmation o~ the protein sequence revealed that the
preparation contained a small fraction of recombinant
cysteine-free IL-~ protein with a terminal methionine
residue.
In a particular embodiment the isolated active fraction
consisted of a mixture of amino terminal methionine-
containing and amino terminal methionine-free synthetic
cysteine-free IL-6 proteins. ~mino acid sequence analysis
indicates that in one preparation of the mixture 90% of the
mixture was methionine-free at the amino terminus and 10%
d contained an amino terminal methionine. In this embodiment,
the active preparation of the cysteine-free protein varies
from natural IL-6 in that 1) no intramolecular disulfide
bonds occur; 2) a methionine amino acid at position one in
the bioengineered protein replaces the signal sequence amino
20 acids 1-28 of the natural unprocessed protein; 3) amino acid
two in the bioengineered protein, glycine, replaces the
proline amino acid present ~n the natural IL-6 protein: 4)
the carboxy terminus contains additional amino acids.
Collagenasa generally cleaves after Y in the sequence
P-Y-G-P wherein Y represents a neutral amino acid. See Keil
et al. FEBS Letters 56: 29~-296 (1975). In the present
example of the fusion protein with II,-6 sequences the neutral
amino represented by Y i5 valine. Accordingly, the carboxy
terminus of the protein produced by induction of the fusion
protein coded by the p367 vector after digestion of the
fusion protein with collagenase is expected to be mainly . .
. P-G-P-V-G-P-V and/or . . . P-G-P-Vo If suf~icient
collagenase is present under sufficiently rigorous
~6 conditions, the carboxy terminus is exclusively . . .
-39-
P-G-P-V. I the amino acid sequence starting with the third
amino acid (i.e. V) in Figure 5 and ending with the
fourteenth amino acid from the end (i.e~ M) is called pep, a
mixture of the following peptides may be prepared by the
present example:
G-pep-SDPGPVGPV ~Protein I)
G-pep-SDPGPV (Protein II)
MG-pep-SDPGPVGPV (Protein III)
MG-pep-SDPGPV (Protein IV)
Such a mixture was used in the asYays described in
Sections 5.11l 7, R, 9, 10 and 11. The termini of the
peptides of this mixture differ from that of native IL-6 by
the presence of G or MG instead of P at the amino terminus
and by the pre~ence of SDPGPV and SDPGPVGPV at the carboxy
terminus. ~hese diffarences in the termini of cysteine-free
IL-6 do not affect the activity of the protein.
5.9. ALTERNATE METHODS OF PREPARATION
OF SYNTHETIC CYSTEINE-FREE IL-6
The foregoing description is but one specific example
of a useful method by which the Rynthetic cysteine-free IL-6
peptide of the present invention may be prepared. However,
~5 those ~killed in the art will readily recognize that other
vector constructs, as well as other unicellular host, are
also useful in the method of the present invention. In very
general term~, for example, the skilled artisan will
recognize that to eventually achieve transcription and
tran~lation of the inserted gene, the gene must be placed
under the control of a promoter compatible with the chosen
host cell. A Promoter is a region of DNA at which RNA
polymerace attaches and initiates transcription. The
promoter selected may be any one which ha~ been i olated
6~
-39a-
from the host cell organism. For example, E coli, a
commonly used host system, has numerous promoters such as
the lac or rec~ promoter associated with it, its
__
bacteriophages or its plasm.ids. Also, synthetic or
recombinantly produced
-40-
promoters, such as the ~ phage PL and PR promoters may be
used to direct hiqh leval production of the segments of DNA
adjacent to it. Similar promoters have also been identified
for other bacteria, and eukaryotic cells.
Signals are also necessary in order to attain efficient
transcription and translation of the gene. For example, in
E. coli m*NA, a ribosome binding site includes the
translational start codon (AUG or ~UG) and other sequence
complementary to the bases of the 3' end of 16S ribosomal
RNA. Several of these latter sequences (Shine-Dalgarno or
S-D) have been identified in E. coli and other suitable host
cell types. Any SD-ATG sequence which is compatible with the
host cell system, can be employed. These SD-ATG sequences
include, but are not limited to, the SD-ATG sequences o~ the
cro gena or N g~ne of coliphage lambda, or the E. coli
tryptophane E, D, C, B or A genes.
A number of methods exist for the insertion of DNA
fra~ments into cloning vectors ln vitro. DNA ligase is an
enzyme which seals single-stranded nicks between adjacent
nucleotides in a duplex DNA chain; this enzyme may therefore
be used to covantly join the annealed cohesive ends produced
by certain restriction enzymes. Alternatively, DNA ligase
can be used to catalyze the formation of phosphodiester bonds
between blunt-ended fragments. Finally, the enzyme terminal
deoxynucleotidyl transferase may be employed to form
homopolymeric 3'-single-stranded tails at the ends of
fragments; by addition of oligo ~dA) sequences to the 3' end
of one popula~ion, and oligo (dT) blocks to 3' ends of a
second population, the two types of molecules can anneal to
form dimeric circles. Any of these methods may be used to
ligate the control ~lements into specific sites in the
vector. Thus, the sequence coding for the cysteine-free or
cysteine-containing IL 6 fusion protein is ligated in~o the
chosen vector in a specific relationship to the vector
promoter and control elements, so that th~ sequence is in the
2~
-41-
correct reading frame with respect to the vector ATG
sequence. The vector employed will typically have a marker
function, such as ampicillin resistance or tetracycline
resistance, so that transformed cells can be identified. The
method employed may be any of the known expression vectors or
their derivatives; among the most frequently used are plasmid
vectors such as pB~ 3~2, pAC 105, pVA 5, pACYC 177, pKH 47,
pACYC 184, pUB 110, pmB9, pBR325, col El, pSClOl, pBR313,
pML21, RSF212~, pCRl or Rp4; bacteriophage vectors such as
lambda gtll, lambda gt-WES-lambda B, chain 28, chain 4,
lambda gt-I-lambda BC, lambda-gt-l lambda ~, M13mp7, M13mp8,
M13mp9; SV40 and adenovirus vectors; and yeast vectors. ThP
vector is selected for its compatibility with the chosen host
cell system. Although bacteria, particularly E. coli, have
proven very useful in high yield production of the synthetic
IL~6 peptide, and are the preferred host, the invention is
not so limited. The present method contemplates the use of
any cultuxable unicellular organism as host; for example,
eukaryotic hosts such as yeast, insect, and mammalian cells,
are also potential hosts for IL-6 production. The selection
of an appropriate expression system, based on the choice o~
host cell, is well within the ability of the s}cilled artisan.
One skilled in the art will readily recognize that
variations on the described fusion protein are also possible.
For example, the order of the first and third peptide
portions can be reversed, so that the third peptide segment
is positioned at the amino terminus and the sequence coding
fox peptides with IL-6 activity is a~ the carboxy terminu~,
with thQ second, cleavable peptide portion remaining as a
30 link between the two segments.
The identity of each of these segments may also be
varied. For example, substantial variation is possible
within and around the basic IL-6 peptide sequence. A
particularly interesting observation is that a substantial
35 portion of the amino terminus can be deleted, not only
-42-
without loss of activity, but with a resultant 2-3 fold
increase in activity in both cysteine-containing and cysteine
free IL 6 sequences. However, removal o~ the last 20
residues in the sequence results in a complete loss of
activity, in both cysteine-containing as well as cysteine-
free forms. Also, as noted above, the presence o~ additional
amino acid residues on the carboxy terminus of the IL-6
peptide doe~ not affect the biological activity of the
molecule.
It will also be understood by those ski:Lled in the art
that any amino acid in the known sequence of IL-6 may be
substituted with a chemically equivalent amino acid. In
other words, "silent changes" may be made in the amino acid
sequence without affecting the activity of the molecule as a
whole. For example, as has be~n shown, substitution of all
cysteine residues with serine residues allows the modified
IL-6 molecule to retain its biological activity. Alternative
choices as substitutes for cysteine are other ne-ltral amino
acids such as valine, proline, isoleucine and glycine,
serine, threonine or tyrosine. Negatively charged residues,
such as aspartic acid and glutamic acid may be interchanged,
as may be positively charged residues such as lysine or
arginine. Hydrophobic r~sidues including tryptophan,
phenylalanine, leucine, isoleucine, valine and alanine may
al50 be exchanged. Alteration of the sequence by amino acid
substitution, deletion, or addition and subsequent testing of
the resultant molacule to determins if biological activity is
retained is well within the ability of one skilled in the
art, without neces~ity for undue expPrimentation.
The identity of the cleavable linker peptide sequence
is also a matter of choice and may be accomplished using
chemical or enzymatic means. The sequence employed may be
any one which can be chemically cleaved, so tha~ t~e peptide
with the biological activity of IL-6 can be released from the
35 remainder of the fusion protein. In a preferred embodiment,
-
-43-
the cleavable sequence is one wllich is enzymatically
degradable. A collagenase-susceptible sequence is but one
example. Other useful sites include enteroklnase- or Factor
Xa-cleavable site. For example, enterokinase cleaves after
the lysine in the sequence Asp-Asp Asp-~,ys. Factor Xa i~
specific to a site having the sequence Ile~Glu-Gly-Arg, and
cleaves after th~ arginine. Another useful cleavage site is
that of thrombin which recognizes the sequence Leu-Val-Pro-
Ary-Gly-Ser-Pro. Thrombin cleaves between the Arg and Gly
residues. Other enz~me-cleavable sites will also be
recognized by those skilled in the art. Alternately, the
sequence may be selected so as to contain a site cleavable by
cyanogen bromide; cyanogen bromide attac~s methionine
residues in a peptide sequence.
It is preferabla, although not essential, to select a
linker sequence which, when cleaved, leaves a minimal number
of residues attached to the IL-6 sequence, so that the
terminus of the released IL-6 active peptide is as near to
the native sequence as possible. In a particular embodiment
the IL-6 active portion is at the carboxy terminal end of the
2 fusion protein, and the cleavage site i5 specific for a
protease that is capable of leaving the natural pro-val-pro
amino terminal peptide sequence. Examples of such cleavage
sites axe those that are cleaved by enterokinase or Factor
Xa.
The identity of the third peptide seguence may also be
varied. This portion of the tripartite structure potentially
serves two purposes: (1) the use of a correctly selected
protein, capable of being expressed in the chosen host, can
30 place the produotion of peptides having activity Ih-6 under
the control of a qtrong promoter, and thus ~acilitate the
production of those peptides; and (2) it can provide a
convenient means for identifying transformed clones producing
the fusion protein. For example, in the discussion provided
35 above, the full sequence encoding ~-galactosidase was used;
-44-
this protein provides a visual means of detection by the
addition of the proper substrate.
Alternatively, the third peptide portion can ~e a
fraction of such a protein, provided that the portion
remaininq i~ still readily expressed by the host cell. This
portion can also be a peptide which is not necessarily
visually detectable, but the presence of which may be
detectable by other means, such as by calculation of the
expected molecular weight of the fusion protein or insertion
into a vector with a detectable marker. Another useful
alternative sequence for use in a prokaryotic host is the
~E gene product, or a portion thereof (Kleid et al.,
Science 214:1125-1129, 1981). Additional choices include
sequences coding for the cro gene of ~ phage or othex
portions of the lac genes than the lac~ sequence coding for
~-galactosidase. Those skilled in the art will recognize
additional choices which may provide the basis for the third
peptide portion of the claimed fusion protein.
5.10. ALTERNATE DNA CONSTRUCTS
The followins Examples illustrate the preparation of
DNA constructs in which the position of the first and third
peptide portions of the tripartite ~usion proteins is
reversed. Also illustrated is the use of alternative linkers
and third peptide portions. The quantity of production of
IL-6 protein using this construct is also high, ranging ~rom
about 1-~0% of total soluble cellular protein.
5.10.1. TrpE - ENTEROKINASE CLEAVAGE SITE -
IL-6 MUTEIN FUSION PROTEIN _ _
~a) A fusinn protein that encodes beta alacto~idase,
followed by an enzymatic cleavage site, followed immediately
by a synthetic IL-6 peptide sequence having C-terminus and
N-terminus ends of the natural IL-6 peptide and all four
cysteines replaced by serinss, is expressed by a new
-45-
recombinant plasmid p-beta-Gal/cfIL-6. To prepare p-beta-
GalJcfII.-6, th~ plasmid p365 - which has been described in
Section 5.5 and Figure ~ of this specification - is digested
with the enzymes EcoRII (to cut the EcoRII site that is
located 14 bases from the 5' NcoI site) and ~glII (to cut the
~glII site ~hat is shown in Figure 13 aæ BglII'). The
plasmid is digested with 5 units of each enzyme per 10 ~g
plasmid at 37C :~or 2 hours. The 0.492 Kb EcoRII/BglII
fragmen-t is isolated by standard procedures such as electro-
el~tion. This 0.492 Kb fragment is called sequence A. This
sequence A and the subsequent sequences noted in the
following text of this section refer to the illustrations in
Figure 12.
(b) An ~dditional aliquot of plasmid p365
15 (approximately 10 ~g) is digested with HindIII and ~glII as
described above at pH 7.5 in a bu~fer of 25 mM Tris HCl, 100
mM MgC12, 10 mg~ml BSA and 2 mM BME. The large (3.0 Kb)
fragment that results from cutting the unique HindIII site
and the BglII site referred to in Figure 13 as BglII' is
called sequence B
~ .
(c) A synthetic, double-stranded oligonucleotide
(sequencP C) is prepared and ligated to sequence A and
sequence B. The oligonucleotide starts with overlapping
HindIII and BclI sites, encodes a sequence of amino acids
con~aining an enterokinase cleavage site followed immediately
by the first three amino acids of natural IL-6, Pro-Val-Pro
(PVP), and ends with an EcoRII site. The sequence o~ the
oligonucleotide is:
BclI EK sit~
5'AG CTT GAT CAG GCG GAT CCG GAA GGT GGT AGC GAC GAC GAC GAC AAA
3' A CTA GTC CGC CTA GGC CTT CCA CCA TCG CTG CTG CTG CTG TTT
P V P 3'
CCG GTT CCG
GCC CAA GGC GGT CC 5"
61.
-46--
EcoRL I
Each strand of the oligonucleotide ls prepared
separately, treated with polynucleotide kinase in the
pre~ence of 1 mM rATP in a ~uita~le reaction buffer at 37-C
for 30 minutes, ~nd annealed by heating to 85-C for 5
minutes, followed by slow cooling to 25-C.
To ligate sequence C to sequences ~ and B,
approximately 1 ~g of synthetic cysteine-free IL-6
EcoRII/BglII fragment tsequence A) is coprecipitated with 200
ng of the synthetic oligonucleotlde (sequence C) and ligated
to the HindIII/BglII vector component ~sequence B) of p36S.
Li~ation is accomplished in a 20 ~1 reaction volume
containing 20 mM Tris HCl, pH 7.6, 0.5 mM rATP, 10 mM MgC12,
5 mM DTT at 16- overnight. The new plasmid is pABC. pABC is
cloned by adding a 5 ~1 aliquot of the reaction mixture to
competent HB101 bacteria. ~mpicillin-resistant colonies are
selected after overnight incubation at 37C.
(d) The 3' end of the recombinant synthetic cysteine-
free IL-6 gene expressed in p365 is reconstructed to encode
the natural I~-6 carboxy terminus, which ends with
methionine.
To accomplish this, the following oligonucleotide is
synthesized as above~
S' GA TCT TTC AAA GAA TTC CTG CAG TCC TCC CTG CGT GCT CTG CGT
3' A AAG TTT CTT AAG GAC GTC AGG AGG GAC GTA CGA GAC GCA
CAG ATG TAA TGA TAG GTA C 3'
GTC TAC ATT ACT ATC 5'
This oligonucleotide, reading from left to right,
starts with a BglII site, encodes the natural amino acid
sequence of IL-6 that follows the BglII' site of p365, and
35 concludes with a methionine residue that is followed
Z~ 2~.
-47-
immediately by three stop codons an~ a KpnI site (sequence
D).
(e) p~BC (step C) is di~ested with HindIII and BglII
(sequence E).
(f) PATH 23 (available from A. Tzajaloff, Columbia
University, New YorJc City) is an ampicillin-resistance
plasmid containing a gene that encodes the amino-terminal 3~7
a~ino acids of TrpE (anthranilate synthetase component I)
adjacent, and in reading frame at its 3' end with, a
polylinker containing a HindIII site. A general description
of the TrpE operon may be found in ~iller and Reznikaff,
eds., The Operon, Cold Spring Harbor Laboratory, pp. 263-302
(1978). Other sources of DN~ that encode all or part of trpE
and lacZ are readily available. Such other sources may be
found, for example, in Pouwels et al., Cloning Vectors, A
Laboratory Manual, Elsevi~r, 1985. For example, trpE
sequences may be isolated from plasmids haviny the following
identifying codes in the Pouwels et al. manual:
I-A-ii-3 (pDF41 and 42), I-A-iv-23 (pRK353), I-B-ii-4
(pMBL24), I-B~ii-1 (ptrpED5~ D-i-3 (pEP70-pEP75~, and
I-D-i-4 (pEP165 and pEP168).
10 ~g PATH 23 is digested with 5 units each of HindIII
and XpnI at 37 C for 2 hours. The large fragment is isolated
by gel chromatography, followed by electro-elution and
ethanol precipitation (sequence F).
(g) Approximately ~00 ng of sequenc~ D are mixed with
approximately 1 ~g of sequence E. The resulting mixture is
coprecipitated with ethanol in tha presence o~ sequence F and
ligated as described above. The resulting fragment i5 called
PTrpE/EK/cfIL-6.
(h) Competent E. coli host cells are trans~ormed with
pTrpE/EK/cfIL-6. For example, the E. coli HB101 strain is
used as host cell for transformation in one embodiment.
Ampicillin-resistant colonies are gathered. These colonies
express a fusion protein comprising a TrpE segment, an amino
6~.
-48-
acid segment recogni2ed and cleaved by enterokinase, and the
synthetic cysteine-free IL-6 amino acid sequence that has
the termini at both carboxy and amino ends of the natural
IL-6 peptide, starting with PVP and ending with M.
The structure of the protein following cleavage is
shown in Figure 18.
5.10.2. ~ETA-~ALACTOSIDASE-ENTEROKINASE
CLEAVAGE SITE - SYNTHETIC CYSTEINE-FREE
IL-6 MUTEIN FUSION PRO?EIN
The protocol described in Section 5.10.1 for producing
the TrpE enterokinase - cleavage site synthetic cysteine-
free IL~6 fusion protein is ollowed, except the pEx-l
vetor is substituted for PAT~ 23 in step f. pEx~
digested with BamHI and XpnI. ~amHI and BclI have
compatible restriction sites. rrhe resulting construct
contains a thermoinducible ~-galactosidase gene followed by
a cloning polylinker. The truncated gene produces a peptide
that i3 approximately 48 Rd of the beta-galactosidase
protein (Stanley, K.K. and Luzio, J.P., 1984, EMBO J., Vol.
3, pp. 1429-1434). The resulting construct is called ~ gal/
EK/cfIL-6. Another source of beta-galactosidase DNA
includes pHg~000 described in 5.2.1. The fu~ion protein
is produced after transformation of E.coli N4830 cells with
p~gal/EK/cfIL-6 and thermoinduction. The plasmid is
replicated in E.coli strain N99. N99 and N4830 are
available from Pharmacia. The fusion protein is cleaved by
enterokinase using methods known and used in the art.
It is routine to cleave proteins having an enterokinase
recognition site with enterokinase. See, for example, Hopp
et al., ~iotechnology 6: 1204-1210 (1988~.
-48a-
5.10.3. FUSION PROTEINS WITH FACTOR Xa CLEAVAGE SITE
A factor Xa site is substituted for an enterokinase
site by modifying step C of Section 5.10.1 and 2. The
synthetic oligonucleotide (sequence C) shown in step (C) of
Section 5.10.1 comprises a DNA sequence that encodes
Asp.Asp.Asp.Asp.Lys. This DN~ sequence is modified so as.to
X~
-49-
encode the factor Xa cleavage recognition site,
Ile.Glu.Gly.Arg. The resulting construct is called
p~gal/Xa/cfIL-6. The plasmids pTrpE/EK/cfIL-6 and
p~gal/Xa/cfIL-6 are expressed as described in ~5.10.1 and
5.10.2. The resulting fusion proteins are cleaved with
factor Xa, which cleaves after the arg in its recognition
site by methods known in the art. See, for example, Nagal &
Thogerson, Nature 309: 810-812 (1984).
5.11. ASSAYS OF ACTIVITIES OF THE
SYNTHETIC CYSTEINE-FREE PROTEIN
5.11.1. HEPATOCYTE STIMULATION ASSAY
FAZA 967 rat hepatoma cells were grown in DMEM/F12
supplemented with 10% NuSerum (Collaborative Research),
15 penicillin and streptomycin. Assays were performed on one
day old confluent monolayers s~eded in 48 well plates
(Costar). Prior to treatment with cysteine-free recombinant
IL-6, as prepared according to the methods described in 5.2
to 5.8, and other conditioned media, cells were washed with
20 serum-free medium con~aining 107 M dexamethasone. Treated
cells were subsequently maintained in serum-
free/dexamethasone medium for the duration of the assay.
Cells were incubated in the pressnce o~ cysteine free
recombinant IL-6 and other conditioned media for five hours
25 at 37-C in a tissue culture incubator. After treatment, cell
supernatants were removed and ~tored at -20C or assayed
directly for fibrinogen as follows. Two-fold serial
dilutions of cell supernatants were prepared in a ~eparate
96-well microtiter plate and spotted onto a 0.45 7m
30 nitrooellulose filter using dot~blot apparatus (Bio-Rad).
~ibrinogen levels were datermined by solid phase enzyme-
linked immunoassayO Fibrinogen was detected using a 1:1000
dilution of rabbit anti-rat fibrinogen polyclonal anti~erum
(obtained from Dr. Gerald R. Crabtree, Stanford University).
2~
-50-
Secondary antibody was affinity purified alkaline
phosphatase-colljugated goat anti-rabbit antisera (Promega ,
Biotec). Bound antibody was visualized by addition of
substrates nitro-blue tetrazolium (NBT) and 5-bro~o-4-
chloro-3-indovl-phosphate, p-toluidine salt (BCIP; Sigma).
Positive controls consisted of cells treated with supernatant
obtained from PMA stimulated MRC-5 fibroblasts. Quantitation
of the assay was carried out by scanning laser densitometry.
5.11.2. _-CELL DIFFERENTIATION ASSAY
Murine B-cell clone CH12.LX (N , d+ LY-l+) was grown
and maintained in ~PMI 1640 containing 5~ heat-inactivated
fetal bovine serum, 300 Ng/ml glutamine, 0.04 mM 2-
mercaptoethanol and anti~iotics CH12.LX cells bear surf~ce
IgM spècific for the phosphatidyl choline moiety of sheep
erythrocytes (SRBC).
The differentiation assay was performed by culturing 2
x ~05 B-cells in the presence or absence of various
concentrations of synthetic cysteine-free IL-6, as prepared
according to the methods described in ~5.2 to 5.B, in 2 ml
B-cell medium in Costar 24-well plates. SRBC (ASA Biological
Products, Einston-Salem, NC) were washed three times in RPMI
1640 prior to use; 1 x 10~ erythrocytes were included in each
test culture. Positive controls consisted of mitogen-
stimulated B-cells using 50 Nq/ml lipopolysaccharide (Difco).
Cultures were incubated at 37C in an atmosphere of 5% CO2.
Direct hemoly~ic plaque forming colonies (pfc) in CH12.LX
cultures were determined.
5.11~3. IN VITRO ~ONE MA~ROW ASSAYS
Synthetic cysteine-free IL-S prepared in accordance
with the methods described in Section 5.2 - 5.8, has also
been shown to have therapeutic utility in art-recognized ln
vitro testing. Delta (~) assays of the e~fect u~ IL-6, on
~luorouracil treated bone marrow cells, indicates that
2~
synthetic cysteine--free IL-6 has good stimulatory activity on
progenitor cells. The protocol for these assays is found in
Figure 14. Particularly effective stimulation is observed
when synthetic cysteine-free IL-6 is combined with IL-l, and
also when these two cytokines are comhined with either M-CSF
or IL-3. Tabular presentation of ~ values are found in Table
2, graphic depiction of Time 0 colony assays are shown in
Figura 13.
5.11.4. 7TDl ASSAY OF DELETION MUTATIONS
... .
In a 7TDl assay (Van Snick et al~, Proc. Nat'l~ Acad.
Sci. USA 83:9679-9683, lg86), which utilizes the
proliferation of an IL-6 dependent murine hybridoma cell line
to quantify biological activity, various deletion~ of amino
acid residues from IL-6 sequence were tested. Table 3 shows
these results, expressed as percent activity compared with
equimolar amounts of recombinant cysteine-containing IL-6
(Amgen). The IL-6 sequences tested are all cysteine-
containing. Plasmid p478 i~ a plasmid construct identical to
plasmid p367 except that in p478, the cysteines in th2 IL-6
sequence have not been replaced. The IL-6 active protein
fraction is the IL~6 pep~ide cut from the fusion protein
("p478 cut" in Table 2) with collagenase. This peptide
fraction for each deletion mutation tested is, therefore~ a
mixture of cysteine-containing IL-6 peptide~ having discrete
amino acids added to the carboxy terminal end of the IL-6
peptide sequence. In addition, the proteins tested are the
fusion proteins themselves ('7p478 uncut'l). The wild typet
cysteine-containing, natural IL-~ sequence fusion protein
produced by p478 ("478 uncut") retains the biological
activity of IL-6 (3% activity) as compared to the cleaned
cysteine-containing IL-6 peptide ("478 cu~ hat i-~ produced
from it . The ~ 478 cut ep~ide is the s~andard against which
the nine deletion mutations are tested and as such, is 10û%
active in the test. Thus in Table 3 478 cut refers to the
~O~
-52-
wild type plasmid expressing cysteine-containing IL-6 fusion
protein that has had tlle IL-6 cleaved from the fusion protein
with collagenase and 478 uncut in Table 3 refers to the
uncut fusion protein. Mutant lAE3 is a deletion mutation
produced by deletion of amino acids 4 through 23 in the
cysteine-containing IL-6 peptide where met is amino acid
number one of the peptide sequence for the analogous IL-6
cysteine-free peptide found in Figure 8b. Mutant 2AB, 3AB
etc. all refer to the corresponding twenty amino acid
deletions at the appropriate position in the sequence as
identified in Table 3 in the second column labeled amino acid
residues deleted~
--53--
TABLE 2
Colony Synergizing ~ctivity in Liquid Culture
Stimulating
Actlvity in
Liquid
5 Culture
Mediu~ IL-l IL-6 IL-l ~ IL-6
Readout Value Readout Value Readout Value Readout Value
CSA CSA CSA CSA
G 0 G 5 G 7 G 149
10 Medium M 0 M 7 M 2 M 27
GM 0 GM 15 GM 4 GM 37
IL-3 0 IL-3 5 IL-3 1 IL-3 15
G 0 G 19 G 17 G 344
G-CSF M 0 M 7 M 3 M 65
(G) GM 0 ' GM 25 GM 7 GM 58
IL-3 0 IL-3 11 IL-3 4 IL-3 22
1~i
G 0 G 104 G 32 G 518
CSF-l M 0 M 29 M 2 M 97
(M) GM 0 GM 43 GM 9 GM 99
IL-3 0 IL-3 28 IL-3 3 IL-3 37
G 0 G 122 G 362 G 704
20GM-CSF M 0 M 55 M 19 M 236
(GM) GM 0 GM 54 GM 53 GM 157
IL-3 0 IL-3 34 IL-3 37 IL-3 85
G 14 G 509 G 866 G 1,781
IL-3 M 15 M 208 M 136 M 675
GM 19 GM 129 GM 132 GM 415
IL-3 24 IL 3 102 IL-3 62 IL~3 313
Z5 _ _ _ . _
-54~
TABLE 3
Amino
Acids
~utantResldues AcClvity in
Na~e Deleted 7TDl Assay
~i _
lAB 4-23 280%
2AB 24-43 0.001
3AB 44-63 0.0003
4AB 64-83 0.0003
5AB 84-103 0.0003
6AB 104-123 0.006
7AB 124-143 (0.0000)
8AB 144-163 (0.0000)
9AB 164-183 ~(0.0000)
478 Cut 100
478 Unc~t 3
6. _XAMPLE: MATERIALS AND METHODS
6.1. CONDITIONS FOR RESTRICTION ENZYME DIGESTION
Enzymes were obtained from commercial sources (New
England Biolabs) and digestion were carried out as
recommended ~y the manufacturer.
6.2. BACTERIAL STRAINS AND PLASMIDS
E. coli JM101 (P-L Pharmacia~ was transformed as
de~cribed in Hanahan, 1~83, J. Mol. Biol. 166:557. Plasmid
pKK233-2 was obtained from P-L Pharmacia; plasmid pBSt was
from Stratogene. Other plasmid constructs are as described
in this application.
6. 3 . OLIGONUCLEOTIDE ASSEMBLY
Oligonucleotides were synthesized from CED
phosphoramidites and tetrazole from American Bionetics.
Oligonucleotides were kinased with T4 polynucleotide kinase
according to manufacturers suggestions (New England Biolabs).
The kinase was inactivated by heating at 65C.
Oligonucleotide mixtures were annealed by heating at 85C for
15 minutes and cooled slowly to room temperature. The
annealed oligonucleotides were ligated with 10 U T4 ligase,
ligated products were separated on a 6% polyacrylamide gel,
and the fragments were recovered by elertroelution.
6.4. DNA SEQUENCING
The DNA sequence of inserted fragments and
oligonucleotides were determined by the chain termination
method of Sanger et al., 1977, Proc. Natl. Acad. Sci.
3~ 74:5463, in~orporating the modifications o~ Biggen et al.,
1983, Proc. Natl. Acad. Sci. 80 3963/ ~attori and Sakakai,
1986, Anal. Biochem. 152:232, and Bankier et al., 1988,
Methods Enz~mology, in press.
35~.5. PROTEIN BLOT ANALYSIS
2~
-56-
Samples equivalent to 50 ~L cell culture were run on 8
NaDodSO4-polyacrylamide gels under reducing conditions
according to L~emlli, 1970, Nature 227:680~ Gels were either
stained with Coomasie blue or electroblotted onto two layers
of nitrocellulose in order to have duplicate blots of the
same gel. Prestained molecular weight markers (BRL) were
used to monitor transfer. After being blocked with 0.25%
gelatin, the blots were incubated with a commercial antibody
to ~-galactosidase (Promega siotech). The cross reacting
bands were visualized with a phosphatase-linked, affinity-
purified, goa~ anti-mouse IgG antisera (1:7,500 dilution,
Promega Biotech) using bromo-chloro-iodoyl phosphate and
nitro-blue tetrazolium as recommended by Promega Biotech.
6.6. CELL LYSIS AND TRIHYBRID ASSAY
Cell lysis was performed according to Germino et al.,
Proc. Natl. Acad. Sci. 1983, 80:6848. E. coli were harvested
by centrifugation, and the cell pellets were suspended in
on~-fifth volume of 0.05 mol/L Tris-HCl p~8, 0.05 mol/L EDTA,
15~ sucrose with freshly dissolved lysozyme at l mg/ml.
After 15 minutes at room temperature, the lysates were ~rozen
at -70C, thawed rapidly at 37C, and sonicated briefly to
shear DNA. ~rihybrid fusion protein was quantitated by
colorimetric assay for ~-galactosidase activity using 0-
nitrophenyl-~-D galactopyranoside as substrate.
6.7. COLLAGENASE DIGESTION OF FUSION PROTEIN
_.
Prior to digestion of the fusion protein, non-speci~ic
proteolytic activities in the collagenas~ preparation were
reduced by treatment with ~-hydroxy-mercuribenzoate according
to Lecroisey et al., 1975, FEBS Le~. 59:167. One unit of
collagenase o~ Achromobacter iophagus ~EC 3~4-24.8;
Boehringer Mannheim~ was dissolved in 100 ~l of buffer
containing 100 mM Tris-HCl, pH 7.4; 250 mM NaCl; 1 mM CaC12:
and 40 ~g/ml ~-hydroxy-mercuribenzoate. The dissolved
collagenase was transferred to a 15 ml siliconized
polyp~opylene tube and incubated at room temperature for 30
minutes. Thawed cell extract (4 mls) was sonicated briefly
to disperse aggregates, added to the pretreated callagenase,
and incubated for ~5 minutes at room temperature. The
majority of the cleaved ~-galactosidase was removed by adding
o.~ volumes of saturated ammonium sulfate, incubating on ice
for 30 minutes and pelleting the insoluble material. The
cleaved recombinant synthetic cysteine-free IL-6 was
concentrated by bringing the total volume of the supernatant
to 13 mls with saturated ammonium sulfate, incubating on ice
for 30 minutes and centrifuging to compact the insoluble
protein into a floating pellicle. Liquid was drained by
puncturing the tube and the remaining pellicle was
resuspended in 0.5 ml of Tris buffered saline.
7. EXAMPLE: HEPATOCYTE STIMULATION
Synthetic cysteine-free IL-6, as prepared according to
the methods described in 5.2 to 5.8, stimulates hepatocytes
as shown in Figure 11. The hepatocyte stimulating activity
2 was detected according to the assay described in 5O9.2
supra. The hepatocyte stimulation exhibited ~y synthetic
cysteine-free IL-6 occurs at conc~ntrations of 10 8M,
indicating a specific activity o~ 104 U/mg protein. This
activity is 100 fold different than observed by May et al.,
1988, J. Biol. Chem. 263:7760~ The level of fibrinogen
synthesis observed in our FAZ~ cells is similar to the
response obtained using crude conditioned medium obtained
from PMA induced fibxoblasts. At concentrations greater than
30 1 ~M, the peptide causes a decrease in the number of
hepatocytes resulting in a decrease in the detectable
fibrinogen.
8. EXAMPLE: B-CELL DIFFERENTIATION
_
~5 The most sensitive method ~o evaluat~ the functionality
-58-
of the synthetic cysteine-free IL-6 protein, as prepared
according to the methods described in 5.2 to 5.8 was the
B-cell differentiation assay. A hemolytic plaque assay
assessed the ability of the synthetic IL-6 to induce Ig
secretion in resting ~-cells. The hemolytic assay is
superior for studies of differentiation and proliferation at
the cellular level (Gronowicz et al., 1976, Eur. J. Immunol.
6:5~8). Our protein had a maximal stimulatory concentration
of about 0.1 ng/ml (43 p) (see Figure 12). This value is
similar to values reported elsewhere for natural IL-6 (Van
Snick et al., 1986, Proc. Matl. Acad. Sci. 83:9679;
Brakenhoff et al., 1987, J. Immunol. 139:4116; Poupart et
al., 1987, EMBO 3. 6:1219; Kishimoto, 1985, Ann. Rev.
Immunol. 3:133). At concentrations above 43 pM, synthetic
cysteine-free IL 6 can cause a decreas~ in the number of
differentiated B-cells.
9. EXAMPLE: CYSTEINE-FREE IL-6 RETAINS BIOLOGICAL ACTIVITY
Threa assays o different activities of the natural
IL-6 protein have shown that cysteine-free IL-6 retains
biological activity. Our invention has shown that it i5 the
primary sequence of the peptide that is necessary to fold the
peptide chain into an active conformation.
Vaccines are often formulated and inoculated in
comb1nation with various adjuvants. The adjuvants aid in
attaining a more durable and higher level of immunity using
smaller amounts of antigen or fewer doses than if the
immunogen were administered alone. The mechanism of adjuvant
action is romplex. It may involve the stimulation of
30 production of cellular cytokines (such as the cytokine IL-6),
phagocytosis and other activities of the reticuloendothelial
system as well as a delayed release and degradation of the
antigen. Examples o~ adjuvants include Freund's adjuvant
(complete or incomplete), Adjuvant 65 (containing peanut oil,
35 mannide monooleate and aluminum monostearate), surface active
0
-59-
substances such as lysolecithin, pluronic polyols,
polyanions, peptides, oil emulsions, and mineral gels such as
aluminum hydroxide or aluminum phosphate. Freund's adjuvant
is no longer used in vaccine formulation for hum~ns because
it contains nonmetabolizable mineral oil and is a potential
carcinogen.
Purified synthetic IL~6 like peptides of the present
invention can be added to vaccine preparations to modulate an
immune response, i.e., to act as an adjuvant; this includes
vaccine preparations used to immunize animals such as mice,
guinea pigs, rabbits, chickens, horses, goats, sheep, cows,
chimpanzees as well as other primates, and humans. Methods
of introduction of the vaccine with peptides like IL-5 as
adjuvant include oral, intradermal, intramuscular,
intraperitoneal, intravenous, subcutaneous, intranasal, or
any other route of immunization. The purified IL-6 active
peptides can be used as an antigen for host immunization and
ultimate production of IL-6 specific monoclonal antibodies.
Such antibodies in turn may be used as Ih-6 inhibitors and as
such may be useful in the treatment of certain conditions in
which dysfunction in immunoglobulin production has been
implicated. This includes treatment of multiple myeloma, and
autoimmune disease. For example, intraarticular injection of
these monoclonal antibodies could be employed in treating
rheumatoid arthritis
The purified synthetic IL~6 like peptide of either the
cystelne-frae form or the form with cysteines i~ adjusted to
an appropriate concentration, ~ormulated with any suitable
additions such a~ o~her cytokine peptides or vaccines and
packaged for use. The peptide with IL--6 activity can also be
incorporated into liposomes for use as an adjuvant in vaccine
formulations or as a pharmaceutical product by itself. The
synthetic peptide with IL-6 activity can also be added to
preformed antibodies that are provided for passive
immunotherapy.
-60-
In an alternative embodiment, the purified peptides
with IL-6 activity can be used as an immunostimulant
pharmaceutical product; for example, IL-6 peptldes are useful
generally for stimulation of hemopoietic stem cells, and
specifically for the stimulation of antibody production in
disease-caused and drug or radiation-induced
immunodeficiencies. As such, the peptides are useful in the
treatment of immunosuppressed AIDS patients. They are also
useful for the stimulation of production of hepatic proteins
in hepatic dysfunction. The peptide can also be used as a
reagent for inducing antibodies in vitro, or to modify the
expression of other growth factors in culture. In another
embodiment, the purified peptides with IL-6 activity,
prefarably in combination with other antiviral compounds, can
be used for the prevention and/or the treatment of viral
dis~ases, including ~IV and HBV. In an alternative
embodiment, the purified pepti~es with IL-6 activity can be
used, alone or in combination with other purified cytokines,
to invoke the terminal differentiation of B-cells in
leukemias and other disease states. In another embodiment,
20 the synthetia IL-6 like peptide of either the cysteine-free
form or the form containing cysteines can be used to induce
antibodi~s that recognize any portion of the peptide.
.
10. EXAMPLE: EFFECT OF CYSTEINE-FREE IL-6
AND OTHER GROWTH FACTORS ON PROLIFERATION
AND DIFFERENTIATION OF BONE M~RROW CELL5
IN VIVO (DELTA ASSAY3
Delta assay is designed to determine whether there is
renewal of a highly proliferatiYe population (HPP) in bone
marrow (BM) stem cells tha~ are stimulated with variou~
growth factors~ Mice are treated with 5-fluorouracil (5FU)
to remove cells that are of low proliferative potential
(LPP).
The HPP cells are incubated in a first semi solid
35 agarose culture ~or 12 days in the presence of growth
factors. The growth factors used include IL-l, IL-6, and a
1:1 mixture of IL-l and IL-6. BM stem cells are grown in the
presence of each of these growth factors in the absence of
other growth factors, and in the presence of granulocyte
colony stimulating factor (G-CSF), macrophage colony
stimulating factor (M-CSF), granulocyte macrophaye colony
stimulating factors (GM-CSF), and IL-3. The number of
colonies stimulated by the growth factors is determined by a
double-layer agarose clonal assay and exhibited in Figure 16
labelled "Time O Colony Assay". This figure shows the total
colony forming units per 2.3 x 105 mouse bone marrow cells 2
hours after mice are treated with 5-FU. The first four bars
show the effect of IL-l, IL-6 (cysteine free), and a 1:1
mixture of IL-l and IL 6 (cysteine-free) on colony
stimulating activity in the absence of other growth factors.
The next set of four bars repxesents the effect of G-CSF
alone and in combination with IL-l, IL-6 ~cysteine-free), and
a 1:1 mixture of IL-1 and IL-6 (cysteine-free) on colony
stimulating activity. The next three sets of four bars each
represent the colony stimulating activity of M-CSF, GM-CSF,
and IL-3 alone and with IL-l, IL-6 ~cyst~ine-freP), and a 1:1
mixture of IL-1 and IL-6 (cysteine-free). The results show
that IL-1, IL-S (cysteine-free), and a combination of IL-l
and IL-6 (cysteine-free) stimulate colony formation when used
in combination with M-CSF more than in combination with IL-3,
GM-CSF and G-CSF.
In a separate assay, the HPP cells are incubated with
the same growth factors and combination of growth factors in
liquid culture. The number of non-adherent cells are counted
30 after 7 days, and the results are illustrated in Figure 15
labelled "Number o~ Non-Adherent Cells after 7 Dayq Liquid
Culture." This figure shows the to~al number of non-adherent
cells per ml. Each of ~he five sets of four bars has the
same significance as the corresponding set in the figurP
35 labelled "Time O Colony Assay." The largest number of non-
62-
adherent cells is observed when a mixture of IL-1 and IL-6
(cysteine-free) are used in combination with IL-3, GM-CSF and
M-CSF.
The cells from the liquid culture ar~ washed and again
grown in the pr~sence of the growth factors in semi-solid
agarose, and subjected to a second double-layer agarose
clonal assay. This second agarose clonal assay is referred
to as the "readout" assay. A Delta value is calculated by
dividing the number o bone marrow colonies in the readout
assay by the number of bone marrow colonies in the Time o
Clonal Assay. The results are shown in ~able 2.
The data from the Delta assay describes which growth
factors have synergizing activity in liquid culture when
various factors are used in the readout assay. When medium
alone is used in the liquid culture, the ability of the added
cytokines to facilitate colony formation is IL-1 + IL-6
IL-l or IL-6. This pre~erence of IL-l ~ IL-~ over the
ability of IL-l or IL-6 to synergize with other growth
factors is evident when ~-CSF, GN-CSF, M-CSF and IL-3 are
used in the liquid culture. The synergy is most evident when
G-CSF is used in the readout assay. When GM-CSF and IL-3 are
used in the liquid culture in conjunction with IL-6 or IL-l,
IL-6 causes a greater Delta value than IL-l. However, when
CSF-1 is used in conjunction with IL-1 or IL-6, IL~l causes a
greater 5ynergistic value. The best results in the entire
assay wer~ seen when IL-l ~ IL~6 were used in conjunction
with I~-3 in the liquid assay and G-CSF wa used in the
readout.
These data suggest that cysteine-free IL-6, when used
30 in conjunction with other growth factors, may be an important
tool in bone marrow transplantation and cancer treatment.
high Delta value indicates that a certain combination of
growth factors yields both growth and renewal of stem cells.
This particular charac~eristic would be required ln a growth
35 factor that would be used to stimulate bone marrow growth and
differentiation in vivo.
Protocol for Delta Assay
1) aDFl mice treated with 150 mg/kg 5-fluorouracil ~5FU).
2) Kill mice, harvest bone marrow (BM) 24 house after 5FU
treatment.
3) Wash BM cells in IMDM medium with 20% fetal bovine
serum (FBS3 and antibiotic (gentamicin).
4) Grow cells in a double layer agarose clonal assay:
Growth factors in IMDM with 20% FBS are plated in 35mm petri
dishes in 0.5% agarose. BM cells are added to an overlayer
at a concentration of 2 5 x 104 up to 2 x 10 BM cells in .5
ml per plate. This is called the Time O Clonal Assay. Grow
at 37C under approximately 7% 2 for 12 days.
5) BM cells are grown in 1 ml liquid culture (IMDM 20% FBS
+ antibiotic) in 24-well cluster plates. Cells are incubated
for 7 days starting at 2.5 x 105 cells/ml.
6) After 7 days liquid culture, the non-adherent BM cells
are collected. The cell numbers are counted, cytospin
preparations are made, and the remaining cell6 are washed
over 5ml of cold FBS to remove any growth factors.
7) The cells are hen diluted 20-100-fold and plated into
the clonal agarose assay. These cultures are grown for 7-12
days at 37C! 7% 2 in a fully hu~idified atmosphere. This
is called the Readout Assay.
8) The Delta value is calculated by dividing the number of
BM colonies in th~ Readout Assay by the number of BM colonies
in the Time O Clonal Assay
11. EXAMPLE: COMP~RISON WITH OTHER COMMERCIAL
SOURCES OF IL-6 IN 7td 1 ASSAYS
.
The biological activity o~ the cysteine-free IL-6
peptide purified ~rom the f~sion product induced by cells
-64-
harboring p367 was compared with cysteine-containing IL-6
from commercial sources using a 7td.1 assay (Van Snick et
al., Proc. Nat'l. Acad. Sci. USA 83:9679-9683, 1986). Those
commercial sources were Boehringer Mannheim (bm) which
produced an E. coli derived IL-6; Endogen which was a non-
recombinant product; and Genzyme which produced a yeast-
derived IL-6. When the cysteine-free synthetic peptide in
phosphate buffer~d saline purified from induced p367
containing cells was compared to Boehringer Mannheim's IL-6,
Boehringer's showed higher activity at low concentrations and
lower activity at high concentrations as shown in Figure 17.
The highest maximal activity was reached using the cysteine-
free synthetic IL-6.
Comparison of the cysteine-free synthetic IL-6 peptide
with Endogen IL-~ revealed that the cysteine-free IL 6
peptide had superior activity over all concentrations
analyzed (0.1 ng to So ng1 as shown in Figure 16.
Cysteine-free synthetic IL-6 also gave higher
biological activity than Genzyme's IL-6 when concentrations
between 0~1 ng and 50 ng were tested as shown in Figure 18.
Description of_the 7td.1 Growkh Assay
This 3.5 day long assay is performed in 96 well culture
plates. Peptides with IL-6 activity are measured in HG~
units where 100 HGF units equal approximately 1 PCT-GF unit.
Typically 2000 cells of the 7td.1 strain are incubated in
highly humidified chambers of 5% C0, in 200 ~1 of medium
containing a dilution of the peptide exhibiting IL-6
activity. After 8~ hours, cells are pulsed for ~ hours with
the tetrazolium salt 3-(4,5 dimethylthiazol-2-y~-2,5-
diphenylformazan bromide ~M~T) and the supernatants are then
removed following centrifugation at lOOOx G Por 5 minutes.
The formazan crystals ar dissolved with 100 1~ of DMSO and
the plates are read at 570 nm wavelength. The optical
density is proportional to the number o~ live cells~
~ 0 0~
-65-
Theamount of IL-6 in HGF units is defined as the reciprocal
of the dilution required to give 50% of the maximum optical
density.
~eaqents of the 7td.1 Assay
Diluent medium is RPMI 1640 with 10% fetal calf serum.
Supplemented RPMI contains RPMI 1640 with 10~ fetal calf
serum, 0.1 mM 2-mercaptoethanol and 2x antibiotic solution.
The labeling reagent is 5 mg/mL MTT in PBS.
12. DEPOSIT OF MICROORGANISMS
The following plasmids have been deposited with the
American Type Culture Collection (ATCC), ~ockville, MD on
November 29, 1988, and have been assigned the indicated
accession numbers:
Plasmid A C
p340 ATCC 40516
p367 ATCC 40S17
pTrpE/EK/cfIL-6 ATCC _ _
p~gal/EK/cfIL-6 ATCC
The present invention is not to be limited in scope by
the plasmids deposited since the deposited embodiments are
intended as single illustrations o~ one aspect of the
invention and any which are functionally eguivalent are
within the scope of this invention. Indeed, various
3~ modifications o~ the invention in addition to those shown and
described herein will become apparent to those skilled in the
art from the foregoing descrip~ion and accompanying drawings~
Such modification are intended to fall within the scope of
the appended claims.
2~
-66-
It is also to be understood that all base pair and
amino acid residue numbers and sizes given for nusleotides
and peptides are approximate and are used for purposes of
description.
3a
