CA2032914A1 - Use of bacterial lipoprotein amino terminus in fusion plasmids for in vivo expression of lipid modified polypeptides - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE
The present invention relates to a fusion plasmid for in vivo synthesis of a lipid modified polypeptide. The plasmid comprises a first DNA sequence encoding a lipoprotein signal peptide, preferably a bacterial lipoprotein signal peptide, and at least the first amino acid of a mature lipoprotein, preferably cysteine. In the DNA sequence, the preferred cysteine codon may be followed by codons for a few amino acids forming a .beta.-turn structure. These amino acids may also include a specific exogenous protease recognition cleavage site. A second DNA sequence encoding the desired polypeptide can be inserted into the fusion plasmid to produce the desired lipid modified polypeptide.

Description

~ o~

TITLE OF THE INYENTION
Use of bacterlal llpoproteln amlno termlnus in fusion plasmids for in v~vo expression of lipid modlfied polypeptldes.
FIELD OF THE INVENTIO~
The present inventlon relates to a fusion plasmid suitable for the expression of polypeptides of any compositlon or size modifled in vivo by the additlon of a llpld moiety.

LACKGROUND OF THE INVENTION
The chemical synthesis of peptides ls known to be ~ubstantially sequence dependent. The repeated presence of certaln amino acld residues such as tryptophan, histldine, methionlne or cysteine makes synthesis of even small peptides (i.e. less than twenty amino acids) difficult. In general, the chemical synthe~is of peptldes wlth more than forty amlno acid residues i5 difficult.
The bacterial expression of peptldes rom DNA
20 sequcnces carried on multlcopy plasmids doe~ not suffer from the ~ame constraints on amino acid compositlon and size of th~ peptide products and results ln a lower co~t of synthesis. However, peptldes synthesized ln vlvo are easlly degraded by tlle proteases contalned ln the tran~formed host cells. In an attempt to alleviate thls problem, the working concept of tandem gene repetition described by Shen in (1984) Proc. Natl. Acad. Sci. 81, 4627-4631, allowed the stabllization of a specific gene product through the introduction of multiple copies of the forelgn gene to be expressed. Unfortunately, apart from the fact that a repeated DNA sequence may be a target for deletion, this method requires extensive manipulation~ to obtain the proper gene configuration.
The synthesis of peptide fragment~ of proteins as well as subsequent immunlzatlon uslng these fragments allow~ productlon of ~peclfic antl-peptlde antibodies useful for the lmmunologlcal ldentlflcatlon of particular clones from a recombinant DNA expression llbrary. The capacity of many anti-peptide antibodies to react with sequences of the native protein ls valuable ln studies of protein structure and functlon. Thls property can al80 lead to the development of synthetlc peptide vacclnes.
However, short peptides cannot generally elicit production of antibodles by them¢elves. Thus, synthetic peptide~ are commonly rendered lmmunogenlc by being coupled to a hi~h-molecular wclght carrler proteln, such as keyhole limpet haemocyanin ~KLH), whlch has beèn shown to induce a helper T-cell re¢pon~e.
Typically, these lmmunogenic peptide derlvatlves have been subsequently in~ected in the presence of ad~uvant6 which are nonspecific stimulators o~ the immune response. Commonly used ad~uvants are the complete ~FCA) 2~3~' or the incomplete (FIA) Freund's ad~uvants. FCA 1~ a water-in-oil emulsion containlng killed mycobacterla, whereas FIA is devoid of bacteria. The actlve ingredient in FIA that can substltute for the mycobacteria ln FCA ls muramyl dlpeptide (MDP; N-acetyl-muramyl-L-alanyl-D-l~oglutamine). Although MDP has been shown to be an effective adjuvant in various systems, it ls not without some undesirable effects. Hence, the~e immunlzatlon procedures require long time periods, multlple ln~ectlons of antigens and often do not lmprove antlbody titres signlficantly above the low levels obtalned by ln~ectlng the free peptides.
Effort# have been made to develop ad~uvants that can slmultaneously perform the function of carriers. It has been demonstrated by ~essler et al. ln (1985) Immunoblol, 170,239-244 and by Jung et al in ~1985) Angen Chem. Int. Ed. Engl. 24,872-873 that a speclflc peptide of the epidermal growth factor receptor ~EGF-R amino acid number 516-529) coupled covalently to a synthetic N-termlnal moiety of the lipoprotein rom the outer membraneo~ E.coli was able to elicit an increased production of specific antlbodles.
The chemlcal nature of the synthetic analog of the bacterial lipoprotein is N-palmitoyl-S-~2RS)-2,3-bis-(palmitoyloxy)propyl))-cysteinyl-serine. Its abbreviation i~ Pam3 Cys-Ser. Previously, this analog as well as the natlve lipoproteln of the outer membrane of E. coll, were found to be polyclonal activators for antibody production by B-lymphocytes as shown by Bessler et al. in (1985) The J. of Immunol. 135,1900-1905. In another study, Hopp demonstrated ln (1984) Mol. Immunol. 21,13-16, that attachment to a dlpalmityl-lyslne moiety was able to convert small peptides of hepatitis B surface antigen into good immunogens.
In both of these studies immune response was slgnificantly enhanced in comparison to the corresponding peptide-KLH con~ugate. These results suggest that con~ugation of a peptlde to a fatty acid carrier has an enhanciny effect on the lmmunogenlcity of the peptide concerned. Unfortunately, ln both cases, the peptide and the lipid moiety are chemically synthesized and coupled in vltro whlch leads to long and costly procedures.
SUMMARY OF THE INVENTION
In accordance wlth the pre~ent lnventlon there is provlded a fuslon pla3mid for in vlvo synthesis of lipid modified polypeptides. ~his fu~lon plasmid compri~es a DNA sequence encoding a llpoproteln ~ignal peptide, and at lea~t the flr~t amlno acid of a mature lipoprotein. Also, lt may optionally comprlse a short amlno acid segment with a ~-turn structure or an exogenous protease recognitlon sequence. A DNA sequence encoding the desired polypeptide can then be inserted into the -5~

fuslon plasmid downstream from the above-mentloned DNA
sequence to produce a lipid modified polypeptide.
The fusion plasmid of the present inventlon is preferably derived from a prokaryotlc expresslon vector with signals for the strong transcription of the fu~ion genes and effective translation. Also, the lipoprotein signal peptlde contalned ln the fuslon plasmld will preferably be a bacterial lipoprotein slgnal peptide.
Furthermore, the flrst amlno acld of the mature llpoproteln wlll more preferably be cystelne. Immedlately following the preferred cysteine codon, there may preferably be a DNA sequence codlng for a short amlno acld sequence with a ~-turn structure or an exogenous protea~e recognitlon cleavage sequence.
This fusion plasmid therefore allows for 6ynthe~is of polypeptldes as part of recomblnant llpopeptides secreted across the cytoplasmlc membrane and recovered from the cell~. Furthermore, the llpld moiety of thc lipopeptides fiynthesized through the fuslon plasmld of the presellt invcrltloll can fll~o ac~ a~ an intrln.~lc carrler/ad~uvant for anti-peptlde antibody production, making the lipopeptlde~ u3eful in the developmet~t of vacaines. The presence of a hydrophoblc region in the lipopeptlde derlvatives may facilltate their passage across biologlcal membranes. In addltlon, the llpid portlon can render a peptide more fat-soluble and can ~3 therefore be useful ~n enhanciny the delivery of polypept~de drugs. Finally, the æosslble incorporatlon of specific enzymatic cleavage sites in the fusion system allows the release of free polypeptides from the lipopeptides.
Also within the scope of the present invention ls a recomblnant fuslon protein comprising a polypeptlde and at least the first amlno acid of a mature lipoproteln, said amino acid having attached thereto at least one fatty acld. Preferably, the first amino acld is a cysteine residue bearing fatty acid~ at its sulfhydryl and a-NH2 groups. More preferably, the fusion protein wlll further comprlse a short amlno acld sequence comprlsing a ~-turn or the recognitlon sequence of an exogenous protease.
IN THE DRAWINGS
Figures la and lb represent a schematic representation of the con~truction of the fuslon plasmid pKLY3 from the commerclally avallable prokaryotic expre~ion vector pKK233~2.
E'igure lc represents the insertlon of the codiny sequellce ~or the P2 peptide of the epldermal growth factor receptor (EGF-R) lnto plasmid pKLY3 to pro~uce plasmid pKLY4.
Figure 2 represent~ the nucleotlde sequence of the NcoI-HindIII DNA fragment inserted into plasmid pKLY2.

Figure 3 represents the nucleotlde sequence of the SphI-BamHI DNA fragment in6erted into pla~mid pKLY3.
DETAILED DESCRIPTION OF THE INVENTION
The present invention relates to fusion plasmids for the expression and in vivo lipld modification of polypeptides. The DNA sequences coding for the polypeptides are inserted ln frame with a DNA reglon of the plasmid correspondlng to the 5' end of a lipoprotein gene. Preferably, the llpoproteln amino terminus gene includes codons for the llpoprotein signal peptlde, and at least the first amlno acld of the mature llpoprotein, preferably a cystelne resldue. Immedlately following the preferred cystelne codon, there may be added codons for a short amino acid segment capable of formlng a ~-turn.
Thls amino acld segment may also comprlse a protease recognltlon cleavage slte. The fuslon plasmld ls under the control of transcrlptlon and translatlon elements allowlng for induclble expresslon ln transformed bacterla.
The nature of the expression vector, llpoprotein amino termlnus, ~-turn pep~ide and polypeptide 1~ a~ follows.
Ex~re~ion vector rhe fu310ll ula~mlds can be derlved from prokaryotlc expresslon vectors replicated at a moderate or higll number of coples per cell. Suitable levels of expresslon of the fusion genes can be obtained through the use of a strong inducible promoter giving no or very low transcription under non-induced conditlons in order to prevent inhibition of bacterial growth by toxic fusion products. The presence of strong terminators of transcription downstream of the fusion gene prevents readthrough transcription that could make the plasmid unstable. The translational signals of the fusion gene consist of a ribosome binding site and an AUG initiation codon, possibly provided by the plasmid vector, whieh codes for the first amino acid of the lipoprotein signal peptide.
The presence in the fuslon plasmld of the orlgin of replication of a slngle-stranded DNA phage allows, upon phage infection, the synthesis of high yields of slngle stranded DNA duriny replication. The single stranded DNA
ean be used for site directed mutagenesis on the plasmid.
LiPoProtein amino terminus Typically, the lipoprotein ~lgnal peptide to be used in the context of the present lnvention ls a baeterlal lipoprotein ~lgnal peptide havlng between 16 and 29 amino acids in length such as that described ln Klein et al., 1988, Protein Eng., 2l15-20.
This slgnal peptide will contaln a posltlvely ellarged amino terminal region, a central hydrophobie region and a carboxyl-termlnal reglon which has a 2S recognition sequence for signal peptldase II (SPaseII).
The SPaseII reeognition sequence which is sometimes , d ~, _ 9_ referred to as a "lipoproteln box" consists of Leu-X-Y-Cys (the positlon of cys is numbered +1 and leu -3). In thls sequence, the cysteine is an invariant residue, leu at the -3 position can be replaced by a chemically 6imllar residue, and X and Y are small amino acids.
Cleavage of a llpoproteln slgnal peptlde in vivo through the actlon of the approprlate cell protease requlres prlor lipid modificatlon of the flrst amino acld of the mature lipoproteln. In the case of cystelne, this llpid modlficatlon malnly consists in palmltoylatlon at the sulfhydryl group of the cystelne residue at posltlon ~1 .
In the recognitlon sequence of protease SPaseII, the site of cleavage is on the amlno termlnal side of the modlfied cysteine of the mature protein (viz. Leu-X-Y-Cys). Thus, once the signal sequence has been cleaved, the attachment of another fatty acld to the a-N1~2 group of the cysteine residue of the mature protein, through further action of the cell' 8 enzymes, generate~ a fully modified and proces~e~ llpoprotein.
Following the cysteine residue, mature lipoprotein3 contain a tetrapeptide ~uch a3 ~ln~Ala-A~rl-Tyr (QANY) forming a ~-turn structure whlch may favor the cleavage action of SPaseII. These natural tetrapeptldes, which may be part of the plasmid of the present inventlon, can be replaced by Ile-Glu-Gly-Arg tIEGR) which is a 2 ~ ~`3 2 ~

specific recognition cleavage site for factor Xa protease as described by Nagar h Thoger~en in 1984, Nature, 309:810-812. I'his protease cleaves on the carboxy terminal side of the arginine residue.
It is to be understood by those skilled in the art that other protease recognition sequences such as Asp-Asp-Asp-Asp-Lys~ for enterokinase and Pro-X~ Gly-Pro-Y~
for collagenase (the arrows indicate the cleavage sites~
also fall within the scope of the present invention. In fact, a wide variety of protease recognition sequences pos~e8sing the appropriate characteristics may be used in the context of the present invention.
PolvpePti(le According to the maturatlon process of lipoproteins, any sequence coding for a desired polypeptide, once fused to the ~ene portlon coding for the deslred lipoproteln amino terminus and preferably to those amino aclds constitutin~ a ~-turn is expccted to produce, ater expression in a plasmld alld in vivo processing, a 2.0 polypeptide with an N-termlnal llpid mo~ifled cystelne.
I'l~e DN~ sequence codlng for the de3ired poly~eptide wlll be eitller syntlletic oligonucleotldes or an approprlate restriction endonuclease fragment. The nature of the polyueptide that can be expressed in vivo is basically independent of its amino acid composition or size contrary to experience in chemical synthesis of peptides.

3~9~

The llpopeptides thus syntheslzed can be used ~or many applications. ~ence, they can serve to raise specific antl-peptlde antibodies useful to ldentify particular clones in a recombinant DNA expression library.
Antibodies recognlzlng peptide fragments in a native protein will help to study the structure and function of the protein and to develop synthetic peptide vaccines.
Also, because the lipid portion of these derivatives could facilitate the passage across ~iological membranes, there 18 potential for the creation of efflcient delivery systems for certaln peptlde drug~.
It may, ln some lnstances, be desirable to obtaln a peptide free of a llpld-modlfied N-termlnus. In thls case, the presence of a speclfic cleavage site for an exogenous protease at the amino termlnus of the lipopeptlde wlll allow for the re}ease of free polypeptides through extracellular dlge~tlon of the lipoprotein wlth the appropriate enzyme.
Process or Producin~ the Recombinant DNA of the Present IQyQ~5~e~
Processes through whlch ~uslon genes are constructed are well known to those skllled in the art.
In the case of the present lnventlon, a DNA sequence codlng for the deslred llpoproteln amlno terminus will be generally fused to the 5'-phosphate end of ollgonucleotldes or restriction endonuclease fragments 2~32~

codlng for the desired polypeptlde sequence. The following descriptlon of a preferred embodiment of the present lnventlon can be used as a basis for the construction of other fusion genes and ls not to be interpreted as limlting the scope of the present invention.
Description_of a preferred embodiment As a model system, the gene coding for both the lipoprotein signal sequence of the ColE2-plasmld coded lysls protein, also known as bacteriocin release protein and the N-terminal 5 amlno acld re~idues of the mature lysi~ protein, i8 fused to the synthetic oligonucleotides coding for the P2 peptide correspondlng to the C-terminal phosphorylation site of EGF-R described by Hayden et al.
ln tl986) Proc. West. Pharmacol. Soc. 29, 459-461 uslng the followlng procedure.
The usion vector WflS constructed uslng plasmid pKK233-2 ~hown in Figure la. Tllls plasmld derivatlve of pBR322, ls expres~ed in a moderate number o~ coples per cell. It provide~ a strong regulated trp-lac fu~lon promoter (pTac) induclble by IPTG, an analog of lactose.
Downstream o~ this promoter are a ribo~ome-blndlng slte followed by an ATG tran~latlon initlation codon, a multiple cloning slte and the rrn B transcrlptlon terminators.

-13- ~32~

Th~ orlgin of repllcation of phage Fl ( obtained as a 514 bp RsaI fragment from plasmid pEMBL8) was inserted into the PvuII site of pKK233-2 to obtain pKLYl shown in Figure la.
Upon infection wlth M13 phage, the vlral origin present on the plasmid permits phage-directed plasmid replication in a clockwise direction. Thls process results in the productlon of high ylelds of slngle stranded DNA that can be used for slte directed mutagenesls. The method of Kunkel in (1985~, Proc. Natl.
Acad. Sci., 82, 488-492 was used on slngle stranded DNA
from pKLYl to elimlnate the restriction sltes BamHI and SphI in between SalI and EcoRI to produce pKLY2 shown in Figure lb.
Plasmld pKLY2 was dl~ested at the unlque restriction sites NcoI and HlndIII for insertlon of the syntlletic oligonucleotide of Figure 2 w~th the N-termlnal portion of the colicin E2 1YB15 gene. Digestlon at NcoI
expose~ tlle ATG start codoll. After gene a3~embly, the ATG
ls fused in frame wlth the signal sequence of the lys.ts gene followed by flve codons for the amino acids Cy~-Gln-Ala-A~n-Tyr (CQANY) of the mature proteln, a GTA codon for valine creating a unlque Sna BI restrlctlon slte, a TAA
stop codon and a BamHI restrlctlon slte close to the 2S HindIII ligation site. Tetrapeptide QANY can contribute to the formation of a ~-turn. In the synthetic DNA, the ~2~ ;2~
-lg-cysteine codon TGT of the colicin E2 lysis gene wa~
replaced by TGC. This silent change introduced a unlque SphI restriction site ln the recombinant plasmid pKLY3 shown in Figure lb.
Restrlction dlgests of pKLY3 at the restriction sltes SphI and ~amHI allow for directional insertion of annealed oligonucleotides or restriction fragments with sequences codlng for the desired polypeptide followed by a t.erminatlon codon (Method I). Alternatively, the fusion can be created by loop-in mutagenesls on pKLY3 slngle stranded DNA using ollgonucleotides with complementary sequences for annealing (Method II). In both these methods, the nature of the coding sequences between the cysteine codon and the coding sequences of the polypeptides can be varled. The presence in the oligonucleotides of a restriction slte absent on pKLY3 facilitated the selection of the mutagenized plasmid6.
The unique restrlctlon slte Sna~I ln pKLY3 allows for blunt-end ligation of restrlction fra~ments or annealed oligonucleotides with the desired polypeptlde codlng capacity immedlately downstream of tlle tyroslne resldue o the mature colicin E2 lysis proteln ~Metho~ III).
Tlle fuslon of tlle N-termlnal portlon of the colicin E2 lysis protein to the peptlde P2 whlc~l con~lsts of 12 amino acids from the C-terminal pllosphorylation slte of the epidermal growth factor receptor (EGF-R amino acids 1137-1148) was made according to method I. The ~ynthetic oligonucleotides contained the sequence for the tetrapeptide QANY and the peptide P2 followed by the TAA
stop codon and the diagnostlc restrictlon site XbaI as seen in Figure 3.
Ligation of the assembled oligonucleotides to plasmid pKLY3, digested with the restriction enzymes SphI
and BamHI, created plasmid pKLY4 shown in Flgure lc.
The expresslon of the llpld-modlfled peptlde from this pla~mld was studled in the E. coll rec A lac Iq straln DH5aF'IQ which overproduces the lac repres~or active on the Tac promoter. Cells transformed with the vector (pKLY2), the fusion plasmld (pKLY3), or the recombinant plasmld (pKLY4) were grown in a deflned medium containing lX M63 salts, lmM MgS0~, 2~ (w/v) casamino aclds, 0.5~ ~v/v) glycerol, 40 ~g/ml kanamycln (for malntenance o~ the F plasmld with the lac I~ gene) and 50 ~g/ml ampicillin Ifor ~electlon of the pKLY plasmlds with the bla gene). Tlle cell3 were incubated at 37C with agitation ~250 rpm).
The proce6sin~ of the fusion pro~eins into mature lipopaptides wa3 determined by in vivo labelling experlments with tri.tiated palmltate ([3~ palmitate).
Tllis compound was added at a concentratlon of 50 ~Cl/ml to early exponential phase cultures (A600 ~ 0-3 - 0.5) of the transformants. Incubatlon for 10 min was followed by ~ 5, ~ 3J~

lnductlon of pT~c ~lth the addltion of 1 mM IPTG and further incubation for 30 Inin.
Total cell proteins were then preclpitated with 10~ (w~v) trlchloroacetic acid (TCA) for 30 min. at 4 C.
The proteln pellets obtained by centrifugation at 18,000 g for 15 minutes were washed twice in methanol to remove lipids. The dried pellets were resuspended in sample buffer and analyzed by electrophoresis on dlscontlnuous Tricine-Sodlum Dodecyl Sulfate (SDS) polyacrylamlde gels for the separatlon of low molecular weight proteins following the method described by Schagger and Von Jagow in (1987) ~nal. Biochem 166, 368-379. The labelled proteln bands were detected by fluorography.
Compared to unlnduced cells, transformants containing plasmid pKLY4 induced by IPTG synthesized an additlonal substance, with an apparent molecular weight of approximately 3000 daltons, constitutlng the most abundant [3H]-palmltate-contalnillg compourl(l. Stmllarly, transf-ormants wlth the fuslon plasmld pKLY3 produced, upon lnduction, a palml~ate-contalnll-g compound of le~s than 2,300 daltons. These two compound3 were not found in cells containing the vector pKLY2 independently of lnduction. The lnduclble products of pKLY3 and pKLY4 would be translocated across the cytoplasmic membrane as indicated by the loss of mature products in cells treated Witll globomycin ~an inhibitor of signal peptidase II) for -17- 2 ~ f~

indicated by the loss of mature products in cells treated with globomycin (an inhibitor of signal peptidase II) for 5 min before adding [3H]-palmitate in in vivo labelling experiments. The results showed the presence of a precursor form of the induced product from pKLY3.
The nature of the lipid-modified products of pKLY3 and pKLY~ was analyzed by ln vivo labelling with [3H]-tyrosine of transformed cells gro~n in the above medium without casamino acids. Tyrosine is at C-terminal end of botl~ peptide derivatives expected. Upon induction with IPTG tyrosine containing compounds of size correspor1dil1cJ to those of the [311]-palmState products of pKLY3 and pKLY~ were found. The unique bands found for the lipopeptides ~abelled with either l311]-palmitate or [3}1]-tyrosine indicated their chemical ;tability.
The codiny recJion o~ the mature aLkaline phosp11atase gene of E. coli was inserted in fr~me immedia~ely dowl1c;trean1 of the tyrosil1e codon o ~l1e snature colicln E2 lysLs gel1e on pli1.Y3. This fuslol-1 crcated plas~nic1 p1~l.Y5. ~; s11owl1 hy tl1noil ant1 ~eckwi.~1) in (13)3G) ~ciel1ce ~;1s l~ Q~ k~l`L~ 31~ c:e l~rovl~
.stml~le way to 1nollitor the level of exSpresC;iorl and secretiol1 of fus1on products since it is active in cells only w1~el1 prese1-t in t1~e periplcl;lll. Upon intluction with IPTG DH5~E IQ transformants containing plasmid p~LYS
produced an enzymatically active palmitate-containing derivative of CQANY-alkaline phosphatase. This product had an apparent molecular weight of approximately 50 000 daltons on a discontinuous Tricine-SDS polyacrylamide gel and was absent from transformants treated ~ith globomycin.
The fractionation of IPTG induced transformants containing PKLY4 or P~LY5 into periplasmic and intracellular outer membrane and cytoplasmic membrane comuonents was performed by standard procedures. Tlle radiolabell.ed cells were harvested and suspended in 10 mM
sodlum phc)sphate (pH 7.0). The cells were disrupted by two passages through a French pressure cell at 20 000 lb/in2. Tlle cell lysate was centrifuged at 4 500 X y for 10 min at 4C and the cell envelope fraction was obtained by centrifugation at 100 000 X ~ for ~0 min at 4C. The cytoplasmic membrane was solul)ilized dl~erentially with tl-e deter~ent sodiulll lauryl sarcosirlclt.e. Ihe outer membrane was sedimented by centrif:llclclt:ioll at 100 000 X g fOL` 40 Illill c.lt 4`~ ~rld solubili.ztl~l ill W~ .t~ L' . ~llcl Lyc;is of tht TCA precipLtatetl alld methc,rlol. wa;hed colllporlellts of the variou~ Lr~ct.l.ons by e~lec~roplloL-esLc: On ~l:iscorl~inuous lLicLIle-Sl~S po~.y~cLyl.~ cJ~ ow~ tll.lt t;llo :I.ipid derivatives o CQANY-P2 and of CQANY-alkaline pllosl)hcltclc;e were most.Ly associ.att!d wlth the outer l~*mhrarle. rhus the lipicl modified polypeptides synthesizecl in E. coli were secreted across the cytoplasmic membrane. In the periplasm these products would become associated probably unspecifically, via anchoring of their fatty acids in the lipid layer of the outer membrane. An E.
coli rec A lac I- strain DH5aF'IQ transformed with the fusion plasmid pKLY3 was deposited at the American Type S Culture Collection, 12301 Parklawn Drive, Rockville, MD.
20852, U.S.A. under accession number A recomhinant plasmid similar to pKLY4 in which the coding se~uence for the tetrapeptide QANY was replaced by that of IEGR, a specific recognition cleavage site for factor Xa protease, was constructed. The chancJe in the nature of the ~-turn forming tetrapeptide had no effect on t;he in vivo production of a stable lipi(l modified derivative of peptide P2. Cleavage of the IEGR-containing lipoueptide by factor Xa protease should allow tlle release of the free P2 peptide or any other peptide of interest.

Claims (10)

1. A fusion plasmid for in vivo synthesis of a lipid modified polypeptide, said plasmid comprising a first DNA sequence encoding a lipoprotein signal peptide, and at least the first amino acid of a mature lipoprotein, whereby a second DNA sequence encoding said polypeptide can be inserted into said plasmid to produce the desired lipid modified polypeptide.

2. A fusion plasmid according to claim 1, wherein said lipoprotein signal peptide is a bacterial lipoprotein signal peptide.

3. A fusion plasmid according to claim 1, wherein said first DNA sequence encodes a lipoprotein signal peptide, a cysteine and an amino acid sequence with a a-turn or an exogenous protease recognition sequence.

4. A fusion plasmid according to claim 1, wherein said lipoprotein signal peptide contains a positively charged amino terminal region, a central hydrophobic region and a carboxyl-terminal region comprising a recognition site for signal peptidase II.

5. A fusion plasmid according to claim 1, wherein said lipoprotein signal peptide is the lipoprotein signal sequence of the ColE2-plasmid coded lysis protein.

6. A fusion plasmid according to claim 3, wherein the portion of said first DNA sequence encoding the amino acid sequence contributing to the formation of a .beta.-turn is tetrapeptide QANY.

7. A fusion plasmid according to claim 6, wherein the .beta.-turn is conferred at the SPaseII cleavage site of said lipoprotein signal peptide.

8. A recombinant fusion protein comprising a polypeptide and at least the first amino acid of a mature lipoprotein, said amino acid having attached thereto at least one fatty acid.

9. A recombinant fusion protein according to claim 8, wherein said first amino acid is a cysteine residue bearing fatty acids at its sulfhydryl and .alpha.-NH2 groups.

10. A recombinant fusion protein according to calim 8, further comprising a short amino acid sequence comprising a .beta.-turn and the recognition sequence of an exogenous protease.