CA1334944C - Signal peptide, dna sequences coding for the latter, expression vectors carrying one of these sequences, gram-negative bacteria transformed by these vectors, and process for the periplasmic production of a polypeptide - Google Patents

Abstract

The invention relates to a signal peptide of the formula :

MXKSTLLLLFLLLCLPSWNAGA

in which:

A = Alanine M = Methionine C = Cysteine N = Asparagine F = Phenylalanine P = Proline G = Glycine S = Serine K = Lysine T = Threonine L = Leucine W = Tryptophan and X represents a direct bond between M and K, an amino acid selected from the group comprising the 20 amino acids of the genetic code, or a peptide containing 2, 3 or 4 amino acids selected, each independently of the other, from the group comprising the 20 amino acids of the genetic code.
Application : Use for the periplasmic production of polypeptides.

Description

r S 1 1 3 3 4 9 ~ ~

~1~n~1 pept;~e, nN~ ~egl~ence~ co~ln~ for th~ lAtter, ex~re~;on vector~ cArry~ nE o~ of the~e ~e~ n~ rA~-n~eAt;ve hActeriA trAn~for~ hy th~.~e vector~, An~
proce~ for the ~er;~l~ic pro~uction of A ~01 y~e~t~ ~e 05 The present invention relates to a novel signal peptide, to the DNA sequence~ coding for the latter, to the sxpression vectors carrying one of these se~uences, to the Gram-negative bacteria transformed by the~e vec-tor~ and to a process for the periplasmic production of a polypeptide with the aid of these bacteria.
It is known that Gram-negative bacteria naturally produce polypeptides synthesized in the form of a pre-cursor in the cytoplasm and exported towards the peri-plasm - the space between the cytoplasmic membrane and the bacterial wall - where they accumulate in the form of a mature polypeptide, i e a polypeptide which iA capable of en~uring it~ specific biological action. These poly-peptide~ include enzymes in particular, such as alkaline phosphatase.
It i~ al~o known that Gram-negative bacteria can be made to produce, in their periplasm, a polypeptide which is foreign to them. This kind of peripla3mic pro-duction is of definite interest because it is easier to separate said polypeptide from the other constitusnts of the periplasm than to separate the polypeptide from the other components of the cytoplasm, as is required in the case of a production with accumulation in the cytoplasm.
It is al~o of intsrest because the polypept~de accumu-lates in its mature form without the addition of an N-terminal methionine, which would then have to be L-e~.oved, and without the adoption of an unfavorabls 3econdary conformation.
It i~ known that, for its production to be peri-plasmic, a polypeptide must be synthesizsd in the form of a precursor corresponding to the maturs polypsptids extended at it~ N-terminal end by a peptide, called a signal peptide, which generally consists of 15 to 30 amino acid3. This signal peptide, which has a decisi~e rôle in the ~ecretion of the polypeptide, is clea~ed 05 during the proce~s, thereby releasing the mature poly-peptide in the periplasm.
The first studies concerned with adaptin~ bac-teria to the periplasmic production of a polypeptide which was foreign to them, and e~pecially a polypeptide of eukaryotic origin, consisted in tran~forming the bacteria with the aid of an expression vcctor carrying a DNA ~equence coding for a natural precursor of said polypeptide. Thi~ ~trategy has repeatedly proved to be rather unsuitable for indu~trial production in term~ of 1~ quantity.
An attempt to provide a satisfactory solution con~i~ted in replacing the natural signal pe~tide of the polypeptide with that of a bacterial polypeptide synthe-sized in the form of a precur~or European patent appli-cation A-0177343 givel3 Examples of how to u~e ~3uch f3ignal peptides.
The applicant, observing that the choice of a bacterial signal peptide can determine the adoption, by the precursor of a heterologous polypeptide (especially 2~ of eukaryotic origin), of an inappropriate secondary conformation as ~aid precursor i~ synthesized in the bacterium, ha~ de~igned a novel ~ignal peptide which affords a good yield for the periplasmic production of biologically active heterologous polypeptides.
The invention therefore relates to a novel signal peptide of the formula MXKSTLLLLFLLLCLPSWNAGA

in which ~ 133~944 A = ~lAnine M = Methionine C ~ Cysteine N = A~paragine F = Phenylalanine P = Proline G = Glycine S - Serine 05 K = Lysine T = Threo~ n~
L = Leucine W = Tryptophan and X represent~ a direct bond between M and B, an amino acid ~elected from th~ group comprising the 20 amino acids of the genetic code, or a peptide contA~n~n~ 2, 3 or 4 amino acidA sel~cted, each independently of the other, from the group comprising the 20 amino acids of the genetic code.
A particularly ~aluable 3ignal peptide i8 the one in which X repre~ents a direct bond between M and R, or all or part of the peptide of the sequencs APSG.
According to one aspect, the invention relates to a DNA se~uence coding for the signal peptide described above, wherein X represents APSG or a direct bond between M and K. Any sequences permitted by the degenerescence of the genetic code can be used. The following two sequences are particularly valuable:

5' AT&-GCT-CCA-TCT-GGC-AAA-TCC-ACG-CTG
CTT-CTC-TTA-TTT-CTG-CTC-CTG-TGC-CTG
CCC-TCT-TGG-AAC-GCC-GGC-GCT 3' which code~ for the ~ignal peptide of formula (1):
MAPSGKSTLLLLFLLLCLPSWNAGA
and S' ATG - AAA - TCC - ACG - CTG - CTT - CTC - TTA - TTT - CTG
CTC - CTG - TGC - CTG - CCC - TCT - TGG - M C - GCC- GGC -GCT - 3' .~ .i ...

_ 4 _ 13349 ~4 which codes for the signal peptide of formula (2):
MKSTI,LLLFT.T.T CT~PSWNAGA
The invention further relates to an expression vector for Gram-negative bacteria, comprising a DNA
sequence coding for a precursor of a polypeptide capable of being secreted into the periplasmic space of the bacteria, the precursor being a mature polypeptide extended at its N-terminal end by a signal peptide, wherein the portion of the DNA sequence which codes for the signal peptide is a DNA sequence according to the invention and as defined above.
The signal peptide according to the invention, the DNA seguences coding for it and the expression vectors carrying these sequences can be applied to the periplasmic production of polypeptides by bacteria gi~ing a negative re~ponse to Gram~s stAinin~ test (so-called Gram-negative bacteria), transformed by these vectors.
The invention therefore further relates to the Gram-negative bacteria transformed by a vector as defined above. Among these bacteria, those belonging to the species ~cher~ch; A COl 1 are of value. Preferably, the latter carry one or more mutations (stable if possibls), for example mutations by deletion, affecting the cya gene and/or the crp gene.
According to another aspect, the in~ention relates to a process for the periplasmic production of a polypeptide, which con~ists in cultivating the cells of Gram-negative bacteria defined above, in subjecting the cells to an osmotic shock and in separating the recom-binant polypeptide from the osmotic shock supernatant.
The process according to the invention i8 suitable for a production according to an inducible mode, where the expression of the DNA seguence codin~ for the precursor is placed under the control of an inducible promoter, as well as a production according to a con-stitutive mode, where the production of the polypeptide is continuous as soon as culture o the transformed strain has started.
The process according to the invention i8 suitable for the production of all kinds of polypeptides which are heterologous relative to the ~train u~ed. Thus it is appropriate for the production of polypeptides of eukaryotic origin. These can be proteins in the strict Q5 sense, ~uch a~ human growth hormone (hGH) in particular, or peptides of smaller size, such as, in particular, a natural form or a variant of hirudin, for example the yariant (Lys~7~ HV?.
The present invention will now be described in greater detail with the aid of three Examples in which refsrence will be made to the five Figures attached.
Figure 1 shows a restriction map of pla~mid pl63,1. The different restriction segments are labeled arbitrarily according to the following le~end:

~~~~ ~ ~~~ = Location of the origin of replication (ORI).

= DNA seg~ent derived from plasmid _ _ _ _ _ _ _ _ _ pBR322.

= DNA segment contA 1 n ~ n~ the sequence coding for a natural precursor of hGH.

~ m~ = DNA segment of phage fd contA~nin~
a transcription terminator.

//////////, = DNA segment contA~ng a W 5 trypto-phan-lactose hybrid promoter-operator.

= DNA segment codin~ for ~-lactamase (ApR: ampicillin resistance).

Figure 2 shows the restriction map of plasmid pl60,1, whose PvuI-XhoI-BamHI(1) and PvuI-ORI-BamHI(2) fragment~ originate from plasmid~ pl63,1 and pBR327 re~pectively and whose small BamHI(2)-BamHI(1) fragment is fragment 3 described in Example 1 below.
05 Figure 3 Ahow~ a re~triction map common to pla~-mid~ p380,1 and p373,2. The different restriction seg-ments are labeled arbitrarily according to the following legend:
0 + ~ ~ = PvuI-BamHI(2~ seguence dsrived from plasmid pBR327.

____ _ _ _ __ = PvuI-XhoI sequence derived from plasmid pl63,1 ~////// /// = XhoI-HincII se~uence derived from pla~mid pl63,1.
ClaI
NdeI ~tI = Fragment 4 described in Example 1 20 (~incll) I I below.
O--XXX X XXXXX = Fragment 3 described in Example 1 below.

,, ~ ; . . 5' . - .~` '. . = DNA segment of phage fd contA;n;no a transcription terminator.

= PstI-HindIII fragment of the DNA seg-ment contA; n; nE the se~uence coding for the natural precursor of hGH.

Figure 4 ~how~ the restriction map of pla~mid p400,18. The different restriction fragments are defined arbitrarily according to the following legend:

13349~4 _ + _ ~ _ + = PvuI-BamHI(2) ~e~uence derived from pla~mid pBR327.

= PvuI-XhoI sequence derived from pla~mid _ _ _ _ _ _ _ _ 05 pl63,1.

//~ = XhoI-HincII sequence derived from plasmid pl63,1 XXXX X X X X = Fragment 3 described in Example 1 below.

= DNA segment of phage fd cont~ n ~ ng a transcription terminator.

15 ~ = DNA segment comprising in particular the promoter of the precur~or of hirudin (HincII-NdeI part of fragment 4).

f X X X X = Coding se~uence of the precursor of hirudin, comprising the sequence coding for the signal peptide of formula (1~, represented by the black part of this symbol (combination of fragments ~ and 7).

Figure 5 shows the restriction map of plasmid p460. The different fragment~ are defined arbitrarily according to the following legend:
_ + ~ = PvuI-BamHI(2) se~uence derived from plasmid pBR327.

- = PvuI-XhoI sequence derived from plasmid _ _ _ _ _ _ _ _ pl63,1.

133~944 //~ = XhoI-HincII sequence deri~ed from plasmid pl63,l.
XXXX X X X X = Fragment 3 described in Example l below.

` = DNA segment of phage fd contA~n;n~ a transcription terminator.

~ ~ = DNA segment comprising in particular the promoter of the prscursor of hirudin (HincII-NdeI part of fragment 4).

X X X X = Coding seguence of the precursor of hirudin, comprising the se~uence coding for the ~ignal peptide of formula (2), represented by the black part of this symbol .
0 ~x~le 1: PERIPLASMIC PRODUCTION OF HUMAN GROWTH HORMONE
WITH THE SIGNAL PEPTIDE OF THE FORMULA
MAPSGKSTLLLLFLLLCLPSWNAGA t1) The ~train used i~ a strain of the species R~ch~r1~h; A COl; which i3 directly related to the strain described in European patent application A-0245138, depo-sited in the Collection Nationale de Cultures de Micro-organisme~ (CNCM, Paris, France) on 17 February 1986 under the reference I-529. This ~train carries a cya mutation by deletion and a crp mutation by deletion.
A plasmid carrying a DNA 8e~uence coding for a S 9 133~944 precursor of hGH, whose signal peptide i8 the one accor-ding to the invention of formula (l) -MAPSGKSTLLLLFLLLCLPSWNAGA - was prepared. This plasmid was called p398.

l. Con~trl~ct;on of gl A~ p39R

la) Construction of plasmid p373,2 The ~trategy employed utilizes fragment~ obtalned from already existing plasmids available to the public and fragments prepared by synthesis according to the technique~ now in common use. The cloning techniques employed are tho~e described by T. MANIATIS, E.F. FRITSCH
and J. .~AMRRQoK in "Molecular cloning, a laboratory manual" (Cold Spring Harbor Laboratory, 1984). The oligonucleotides are synthesized using a Biosearch 4600 DNA synthesizer.
Plasmid pl63,l (Figure l), described in European pat~nt application A-0245138 (~hown in Figur~ ~ of thi~
document, which does not mark the BamHI(1) site shown in Figure l of the present patent application) and present in the strain depo~ited in the CNCM under the reference I-530 on 17 February 1986, was digested with the enzymes PvuI and BamHI. This plasmid contains the gene coding for hGH. The PvuI-Ba~HI(1) fragment - hereafter fragment l - contA~n;n~ the action site of the restriction enzyme XhoI, shown in Figure l, was purified.
Likewise, plasmid pBR327, well known to those skilled in the art (q.v. SOBERON, X. et al., Gene, 9 (1980) 287-305), was digested with the enzyme~ PvuI and BamHI. The PvuI-BamHI(2) fragment - hereafter fragment 2 - contA;n~n~ the origin of replication, was purified.
Fragment 3 was then prepared; this is a ~ynthetic BamHI(l)-BamHI(2) fragment cont~n~no the lac igene and .i~

- lo - 1334944 it~ promoter and having the following DNA se~uence, on which the two ends of the ~trand are identified by the numbers 1 and 2 in order to specify the orientation of the fragment in the plasmid~ de~cribed in Figur~ 2 and 0~ 3:

FRA~ T 3 8amHI(l) 5' GATCC GCGGAAGCAT MAGTGTAAA GCCTGGGGTG CCTAATGAGT
GAGCTAACTT ACATTAATTG CGTTGCGCTC ACTGCCCGCT TTCCAGTCGG
GAAACCTGTC GTGCCAGCTG CATTAATGAA TCGGCCAACG CGCGGGGAGA
GGCGGTTTGC GTATTGGGCG CCAGGGTGGT TTTTCTTTTC ACCAGTGAGA
CGGGCAACAG CTGATTGCCC TTCACCGCCT GGCCCTGAGA GAGTTGCAGC
AAGCGGTCCA CGCTGG-TTTG CCCCACCACC CGAA M TCCT GTT-TGATGGT.
GGTTAACGGC GGGATATAAC ATGAGCTGTC TTCGGTATCG TCGTATCCCA
CT-ACCGAGAT ATCCGCACCA ACGCGCAGCC. CGGACTCGGT AATGGCGCGC
ATTGCGCCCA GCGCCATCTG ATCGTTGGCA ACCAGCATCG CAGTGGGAAC
GATGCCCTCA TTCAGCATTT GCATGGTTTG TTGAAAACCG GACATGGCAC
TCCAGTCGCC TTCCCGTTCC GCTATCGGCT GAATTTGATT GCGAGTGAGA
TATTTATGCC AGCCAGCCAG ACGCAGACGC GCCGAGACAG AACTTAATGG
GCCCGCTAAC AGCGCGATTT GCTGGTGACC CAATGCGACC AGATGCTCCA
CGCCCAGTCG CGTACCGTCT TCATGGGAGA AAATAATACT GTTGATGGGT
GTCTGGTCAG AGACATCAAG AAATAACGCC GGAACATTAG TGCAGGCAGC
TTCCACAGCA ATGGCATCCT GGTCATCCAG CGGATAGTTA ATGATCAGCC
CACTGACGCG TTGCGCGAGA AGATTGTGCA CCGCCGCTTT ACAGGCTTCG
ACGCCGCTTC GTTCTACCAT CGACACCACC ACGCTGGCAC CCAGTTGATC
GGCGCGAGAT TTAATCGCCG CGACAATTTG CGACGGCGCG TGCAGGGCCA
GACTGGAGGT GGCAACGCCA ATCAGCAACG ACTGTTTGCC CGCCAGTTGT
TGTGCCACGC GGTTGGGAAT GTAATTCAGC TCCGCCATCG CCGCTTCCAC
TTTTTCCCGC GTTTTCGCAG AAACGTGGCT GGCCTGGTTC ACCACGCGGG
AAACGGTCTG ATAACAGACA CCGGCATACT CTGCGACATC GTATAACGTT
ACTGGTTTCA CATTCACCAC CCTGAATTGA CTCTCTTCCG GGCGCTATCA
TGCCATACCG CGAAAGGTTT TGCGCCATTC GATGGTGTCC G 3' BamHI(2) ~ 1334944 Fragments 1, 2 and 3 were then ligated to give plasmid pl60,1 shown in Figure 2.
This plasmid was subjected to partial digestion with the restriction enzymes HincII and PstI. The large HincII-PstI fragment, containing the origin of replication and shown in Figure 2 was then ligated to fragment 4 shown below, which is a synthetic DNA fragment carrying a sequence coding for the first 44 amino acids of a natural precursor of hGH and, upstream from this sequence, regulatory signals.

ClaI
S' TCGAGCTGACTGAC~.~l~GCTTATATTACATCGA
AGCTCGACTGACTGGACAACGAATATAATGTAGCT
NdeI
TAGCGTATAA~ GGAATTGTGAGCGr-ATAAt'AATTTr~rArAt:TTTAACTTTAAGAAGGAGATATACAT
ATCGCATATTACACACCTTAACACTCGCCTATTGTTAAAGTGTGTCAAATTGAAATTCTTCCTCTATATGTA
ATG GCT ACC GGA TCC CGG ACT AGT CTG CTC CTG GCT TTT GGC CTG CTC TGC CTG
TAC CGA TGG CCT AGG GCC TGA TCA GAC GAG GAC CGA AAA CCG GAC GAC ACG GAC
~M A T G S R T S L L L A F G L L C L
-26 .
XbaI
CCC TGG CTT CAA GAG GGC AGT GCC TTC CCA ACC ATT CCC TTA TCT AGA CTT TTT
GGG ACC GAA GTT CTC CCG TCA CGG AAG GGT TGG TAA GGG AAT AGA TCT GAA AAA
P W L Q E G S A F P T I P L S R L F

GAC AAC GCT ATG CTC CGC GCC CAT CGT CTG CAC CAG CTG GCC TTT GAC ACC TAC
CTG TTG CGA TAC GAG GCG CGG GTA GCA GAC GTG GTC GAC CGG AAA CTG TGG ATC
D N A M L R A H R L H Q L A F L T Y
PstI
CAG GAG TTT GAA GAA GCC TAT ATC CCA AAG GAA CAG AAG TAT TCA TTC CTG CA
GTC CTC AAA CTT CTT CGG ATA TAG GGT TTC CTT GTC TTC ATA AGT AAG G 5' In this fragment, the amino acid3 are designated by letter3 according to the following code:

A = Alanine M = Methionine 05 C = Cysteine N = Asparagine D = Aspartic acid P = Prolins E = Glutamic acid Q = Glut~m~e F = Phenylalanine R = Arginine G = Glycine S = Serine H = Hi~tidine T = Threonine I = Isoleucine V = Valine K = Lysine W = Tryptophan L = Leucine Y = Tyro~ine Sequences -35 tTTGCTT~ and -10 (TATA~T) of the promoter sequence, and the Shine-Dalgarno se~uence well known to those skilled in the art, are underlined in that order in this fragment.
Plasmid p380,1 was obtained in this way.
Pla~mid p380, 1 (Figure 3) wa~ then dige~ted with the restriction enzyme~ ClaI and NdeI 80 as to remove ths small ClaI-NdeI fragment of fragment 4 above and replace it with the ClaI-NdeI fragment below:

ClaI
5' CGATAGCGTATAATGTGTGGAATTGTGAGCGGATAACA
TATCGCATATTACACACCTTAACACTCGCCTATTGT

NruI NdeI
ATTTCACACAGTTTTTCGCGAAGAAGGAGATATACA
TAAAGTGTGTCAAAAAGCGCTTCTTCCTCTATATGTAT 5' 5_ ~
~ 13 - 1334914 The re~ulting plasmid is plasmid p373,2 (Figure 3).

lb3 Construction of plasmid p398 Finally, plasmid p373,2 was di~ested with the re-striction enzymes NdeI and XbaI 80 as to ~eL.~Ve the NdeI-XbaI fragment of fragment 4 above and replace it with the synthetic NdeI-XbaI fragment shown below:
NdeI
5 ' ~TG GCT CCA TCT GGC A~A TCC ACG CTG
AC CGA GGT AGA CCG TTT AGG TGC GAC
CTT CTC TTA TTT CTG CTC CTG TGC CTG
GAA GAG AAT AAA GAC GAG GAC ACG GAC

CCC TCT TGG AAC ~CC GGC GCT-~TTC CCA
GGG AGA ACC TTG CGG CCG CGA AAG GGT
XbaI
ACC ATT CCC TTA T
TGG TAA GGG AAT AGATC 5' Plasmid p398 obtained in this way contains a par-ticularly valuable DNA sequence coding for the signal peptide of formula (1). This sequence is delimited above by two arrows.
2. ~.enerA~ ~th~olo~y The experiments were performed on 6 clones of plasmid p398 (clone~ 2, 3, 5, B, 7 and 8), the results being assessed relative to plasmid p373,2, which contains a DNA sequence coding for the natural precursor of hOEH.
The experiments consisted in cultivating the host-vector systems in question, prepared beforehand (cf. ~2.1), - 14 - 133~944 under conditions ~uch as to give an adeguate biomass (cf.
~2.2) and such that the cells subjected to induction produce hGH (cf. ~2.3), in collecting the proteins con-tained in the periplasmic ~pace by osmotic shock (cf.
05 ~2.4), in subjecting the bacteria to total ly~i8 to give a total protein extract (cf. ~2.~), in determ;nine the periplasmic hGH collected in Z.4 (cf. ~2.6) and in analyzing the eupernatant~ obtained in 2.4 and in 2.5 by the Western Blot technique (cf. ~2.7).
2.1 Pre~r~t; on of the ho~t-vector ~y~te~
The ho~t-vector system~ were prepared according to the bacterial transformation techni~ues known to thoqe skilled in the art, which are described especially in the following book~:
- Molecular cloning - A Laboratory MAnl~Al - T. Maniatis, E.F. Fritsch and J. Sambrook - Cold Spring Harbor Laboratory - 1982.
- Experiments in Molecular Genetics - J.H. MILLER - Cold Spring Harbor Laboratory - 1972.
2.2 Clllt-~re a) Inocl~l~t; on An isolated colony obtained on a solid medium (LB
medium + agar-agar) was suspended in 5 ml of a medium (LB
medium).
The LB medium used has the following characteris-tics:
- its components introduced before autoclaving are:
Bactotryptone* 10 g yeast extract 5 g sodium chloride 5 g di3tilled water qs 1 1 - its pH i~ ad~usted to 7.3 before autoclaving;
- ampicillin is added after autoclaving at a rate of 100 l~g/ml.
* - Trademark ~f~
~,, 1334~44 b) In~llh~t;on The suspension prepared in a) wa3 incubated at 37C for 18 h in order to allow the culturs to reach the stationary growth phase. The dense suspension obtained 06 was diluted in LB medium to give an optical density value close to O.03 when read at 600 nm - OD at 600 nm - and 2 ml of this bacterial suspension were then incu~ated at 37C, with agitation, until the OD at 600 nm was of the order of 0.3.
10 2 . 3 In~l~ct; or Isopropyl-~-D-thiogalactose (or IPTG) was added to the bacterial su~pension obtained according to 2.2.b in an amount such that its final concentration was e~ual to 1 mM, IPTG was used here to initiate and maintain the synthesis of the precursor of hCH by neutralizing the action of the repressor which normally binds to the lactose operator.
The suspension, with IPTG added, was agitated at 37C for 2 h 30 min.
2 . 4 O~n -)t; c ~e;hnck R~ference was made to the protocol described by N.G. NOSSAL and L.A. HEPPEL in "The Journal of Biological Chemistry, 241 (1966) 3055-3063".
a) WA~; n~ w~ th Tri~ An~ ~nTA
A sample of the suspension as obtained in 2.3 after induction was taken and centrifuged for 5 minutes at 6000 g.
The residue was taken up in a volume of buffer at pH 7 (solution A) (cf. above) such that the suspension obtained had an OD at 600 nm of the order of 10.
The buffer used was prepared by ~; n~ the fol-lowing to distilled water:
tri(hydroxymethyl)Am;no~ethane-HCl, or Tris-HCl, added 80 as to give a final concentration of 30 mM.
3~ ethylenediaminetetraacetic acid, or EDTA, added 80 as to give a final concentration of 1 mM.
b) ~ct;o~ of ~l~cro.~e The suspension obtained in 2.4.a was centrifuged for 5 minutes at 6000 g.
05 The residue was taken up very carefully, at con-stant volume, in a solution B prepared for immediate use and corresponding to solution A to which sucrose has been added at a rate of 15 ~ per 100 ml.
The suspension was left for 10 minute~ at 20C.
It wa8 then centrifuged for 6 minutes at 6000 g. The centrifuge tubes were placed in meltin~ ice.
The supernatant was carefully removed and re-placed (at constant volume) with deionized water which had been cooled beforehand to the temperature of melting iC8.
The suspension prepared in this way (having an OD
at 600 nm of the order of 10) was left for 5 minutes at OC.
c) Collect;on of the ~rote~n~ locAte~ ~n th~
~er;D~A~
The suspen~ion obtained in 2.4.b was centrifuged for 10 minutes at 18,000 g.
The supernatant, which contained the proteins located in the periplasm, wa~ collected.
2~ 2.5 TotAl ly.~;~
A sample of the suspension as obtA~ne~ in 2.3 after induction was taken and centrifuged in an Eppendorf tube for 5 minutes at 6000 g.
The residue was resuspended in a volume of buffer such that 1 ml of suspension had an OD at 600 nm of 0.2.
The buffer was prepared from a twice concentrated buffer comprising a solution of the following in distil-led water:
- Tris-HCl 0.1~5 M, pH 6.8;
- sodium dodecylsulfate (4% (w/v));

- glycerol (20% (w/v));
- ~-mercaptoethanol (10% (w/v));
- bromophenol blue (0.02% (v/v)).
The tube was placed for 10 minutes in a water 05 bath set at 100C, the suspension was then centrifuged for 5 minutes at 6000 g and the supernatant was col-lected.
2.6 neter~nAt; on of the ~er1~1 A~i C h~7~
The supernatant obtained in 2.4.c was sub~ected to high pressure li~uid chromatography using an apparatus equipped with a calibrated inJection system and a detec-tor set at 220 nm.
The following were used:
a C8 - 300 Angstrom reversed-phase column made of steel, with a length of 10 cm and an internal diameter of 4.6 mm (SYNCHROM reference C8 R103-10), a mobile phase consisting of a linear gradient passing from 70 volumes of solution S1 and 30 volumes of solution S2 to 40 volumes of solution S1 and 60 volumes of ~olution S2 in 20 minut~.
Solutions Sl and S2 had the following charac-teristics:
S1 = purified water contA;nino 0.1% (v/v~ of trifluoro-acetic acid, S2 = acetonitrile for HPLC, contAinina 0.08X (v/v) of trifluoroacetic acid.
The flow rate was 1 ml per minute.
The optical density of the fractions was measured and the amount of periplasmic hG~, expressed in micro-grams per ml of supernatant, was determi ne~ by comparisonwith a previously established standard scale.
2.7 ~nAly~i~ hy the We~tern Rlot techn~
The following operations wcre carried out in succession:
- separation by gel electrophoresis (according to the - 18 - 1~34~44 protocol de~cribed by TA~MMrT, U.K., Nature, 22I (1970) 680-685) of the different proteins contained in each of the supernatants obtained according to 2.4.c and 2.5;
the gel used was a polyacrylamide gel (15% w~v) con-05 taining 0.5% of sodium dodecylsulfate;
- transfer of said proteins contained in the gel on to a nitrocellulo~e filter (according to the techni~ue of H. TOWBIN et al., Proc. Natl. Acad. Sci. USA, 76 (1979 4350-4354);
- ;~tlnodetection performed according to the technigue of BURNETTE (W.W. BURNETTE, Anal. Biochem., 112 (1981) 195-203), this entails the following successive operations:
rinsing the nitrocellulose filter for 10 minutes with a buffer A (Tri~-HCl 10 mM, NaCl 170 mM, KI 1 mM);
bringing the nitrocellulose filter into contact with a buffer B (buffer A with bovine serum albumin added at a rate of 3 g per 100 ml) for 30 minutes at 37C, bringing the nitrocellulose filter into contact with an i m~l~n~ serum (a polyclonal antibody r~cognizing mature hGH and its precursor) for 18 h at 20C;
rinsing the nitrocellulose filter with buffer B;
bringing the nitrocellulose filter into contact with a solution of protein A labeled with iodine 125 at a rate of 0.1 microcurie per ml, for 6 h at 20C;
rinsing the filter with buffer A;
drying the filter between two absorbent sheets;
bringing the filter into contact with an X-ray film;
developing the film.
3. ~ur~T~
3 1 neter~1nAt;on of the ~eri~l A~1 C h~
The results are reported in the Table below:

- 19 - I3~494~

pr~A~Tn TESTED
Control PLASMID 398 373,2 398,2 398,3 398,5 398,6 398,7 3g8,8 Periplasmic hGH
expre~sed in micrograms per 1.5 2.5 2.9 2.8 3.1 3.1 3.4 ml of super-natant collected after osmotic shock and brought to a turbidity such that OD at 600 nm = 1 It i~ clearly apparent that plasmid 398 affords a periplasmic production which is about twice that afforded by plasmid 373,2.
3.2 AnAly~;.q by the We~tern Rlot techn;al~
Analysis of the autoradiographic films reveals that the precursor has not been detected in the extracts obtained after total lysis of the bacteria transformed with plasmid p398, wherea3 it is detected in the extracts obtained after total lysis of the bacteria transformed with clone p373,2. This show~ that the signal peptide according to the invention is capable of permitting, with a high efficacy, the passage of the precur~or through the cytoplasmic,membrane and its concomitant maturation.
These results emphasize the great advantage of using the signal peptide according to the invèntion of formula (1) for the periplasmic production of a protein such as human growth hormone.

RxA~ple ~: PERIPLASMIC PRODUCTION OF A HIRUDIN VARIANT
WITH THE SIGNAL PEPTIDE OF THE FORMULA
MAPSGKSTLLLLFLLLCLPSWNAGA (1) 05 1. ~trA;n An~ plA~
The strain described in Example 1 was used.
A plasmid called p400 was constructed from plasmid p373,2. It carries a DNA sequence coding for the variant (Lys47) HV2 described in European patent appli-cation A-0273800, the formula of which is reproduced below:

Ile Thr Tyr Thr Asp Cys Thr Glu Ser Gly Gln Asn Leu Cys Leu Cys Glu Gly Ser Asn Val Cys Gly Lys Gly Asn Lys Cys Ile Leu Gly Ser Asn G1y Lys Gly Asn Gln Cys Val Thr Gly Glu Gly Thr Pro Lys Pro Glu Ser His Asn Asn Gly Asp Phe Glu Glu Ile Pro Glu Glu ~yr ~eu Gln said DNA sequence being preceded by a sequence coding for the signal peptide according to the invention of formula (1) .

~o~trl~ction of pl A~i~ ~400 A synthetic AccI-HindIII fragment (fragment 5) contA; n; ne a sequence coding for this variant (except the first 3 amino acids) wa~ prepared. It i8 shown below, the codon corresponding to the last amino acid being designated by an arrow:

FRA~ T 5 AccI
5' AT ACA GAC TGC ACA GAA TCG GGT

CAA ATT TTG TGC CTC TGC GAG GGA AGC AAT
GTT TTA AAC ACG GAG ACG CTC CCT TCG TTA

GTT TGC GGT AAA GGC AAT AAG TGC ATA TTG
CAA ACG CCA TTT CCG TTA TTC ACG TAT AAC

GGT TCT AAT GGA AAG GGC AAC CAA TGT GTC
CCA AGA TTA CCT TTC CCG TTG GTT ACA CAG

ACT GGC GM GGT ACA CCG A M CCT GAA AGC
TGA CCG CTT CCA TGT GGC TTT GGA CTT TCG

CAT AAT AAC GGC GAT TTC GAA GAA ATT CCA
GTA TTA TTG CCG CTA AAG CTT CTT TAA GGT

GAA GAA TAT TTA C M-~TGA M A ATG A M GAA
CTT CTT ATA AAT GTT ACT TTT TAC TTT CTT
Eco RI
TAT CAA TCA TAG AGA ATT TTG ATT TGG G M
ATA GTT AGT ATC TCT TAA AAC TAA ACC CTT

TTC GAA AAG TTC GCT GGT ACT AAG GCT TTG
AAG CTT TTC AAG CGA CCA TGA TTC CGA AAC

TTA GAC GAA GTT GTC AAG AGC TCT GCT GCT
AAT CTG CTT CAA CAG TTC TCG AGA CGA CGA

GGT AAC ACC GTC ATC ATT GGT GGT GGT GAC
CCA TTG TGG CAG TAG TAA CCA CCA CCA CTG

TGA CGG TGA CAG CGA TTC TTC ATG CCA CAG
HindIII
ACT GAC AAG ATC CCA
TGA CTG TTC TAG GGT TCG A - 5' Plasmid p373,2 was digested with the re~triction enzyme~ NdeI and HindIII and the NdeI-HindIII fragment (fragment 6), contA;n;n~ the origin of replication, a~
shown in Figure 3, was purified.
A ~ynthetic NdeI-AccI fragment (fragment 7) was prepared; its sequence is given below:

FR~ NT 7 NdeI
5' T~TGGCTCCATCTGGCAAATCCACGCTGCTTCTCTTATTTCTGCTCCTGTGCCTGCCCTCT-ACCGAGGTAGACCGTTTAGGTGCGACGAAGAGAATAAAGACGAGGACACGGACGGGAGA-AccI
TGGAACGCCGGCGCT~ATTACGT
ACCTTGCGGCCGCGATAATGCATA 5' This fragment contains a particularly valuable DNA sequence coding for the signal peptide of formula (1), which sequence has been delimited by two arrow~ and is followed by nucleotides corresponding to the fir~t 3 codonc2 of the hirudin variant (Lys~7) HV2.

Fragments 5, 6 and 7 were ligated; the plasmid obtained i3 plasmid p400, which is shown in Figure 4.
2. ~enerAl ~tho~olo~y Plasmid p400 was introduced by tran~formation 05 into the bacterial strain described in Example 1.
The experiments were performed in parallel on two different clones (clones p400,18 and p400,24) in accor-dance with the procedure described in sections 2.1 and 2.2 of Example 1. The cultures were induced by the method indicated in section 2.3 of ~xample 1, with two modifications: induction was initiated by ~ n~ IPTG
when the culture had reached an OD at 600 nm of about 0.~, and it wa~ maintained for 3 h 30 min in a first experiment and for 17 h in a second experiment.
After induction, the cells were subjected to an o~motic shock (cf. ~2.4, Example 1~ and the antithrombin activity of the hirudin in the supernatant collected was measured.
This activity was deter~;ne~ using the techni~ue de~cribed by Markwardt, F . et al . (Thromb Haemostas., 52 (19) 160-163) and discus~ed in detail by Harvey, R.P. et al. (Proc. Natl. Acad. Sci. USA, 83 (1986) 1084-1088).
The hirudin variant obtained from one of the clones was purified by high pre~sure liguid chromato-graphy and its NHz-terminal sequence was determined.
3. R~UTT~
The results are reported in the Table below.
They are expressed in antithrombin unit~ per ml of supernatant collected after osmotic ~hock and brought to a turbidity such that OD at 600 nm = 10.

13~34~
- ~4 -Plasmid p400 clone clone p400,18 p400,24 05 Induction 3 h 30 min 241 369 time 17 h 1595 983 As the known values of the specific antithrombin activity of hirudin are between 13,000 and 17,700 anti-thrombin unit~ per mg of hirudin (Loison, G. et al., Bio/Technology, 6:72-77 (1988)), it can be deduced that, after 17 h of induction, from 5 to 10 mg of hirudin have been extracted per liter of supernatant of OD at 600 nm =

Such an amount is much greater than that des-cribed by Dodt, J. et al., FEBS, 202 (1986) 373-377, with the hirudin variant HV1 produced in the periplasm in a strain of ~ col; transformed with a plasmid carrying a se~uence coding for a hybrid precursor of hirudin, the signal peptide of which is that of alkaline phosphata3e.
It has furthermore been ob~erved that the hirudin produced by clone p400,18 doe~ indeed have the NH2-terminal end characteri~tic of the variant (Lys~7) HV2.
These results show that the signal peptide accor-ding to the invention of formula (1) i8 appropriate for the efficient periplasmic production of a peptide such as the hirudin variant (Lys~7) HV2.

~xAm~le 3: PERIPLASMIC PRODUCTION OF A HIRUDIN VARIANT
WITH THE SIGNAL PEPTIDE OF THE FORMULA
MKSTLLLLFLLLCLPSWNAGA

1. ~trA;n ~ ~l~mi~
The ~train described in Example 1 was used.

1334~44 The ~trategy employed to construct plasmid p460, comprising a sequence coding for the hirudin variant (Ly~7) HV2 described in European patent application A-0273800, which 8e~uence i8 preceded by a sequence coding 05 for the signal peptide according to the invention of formula (2):

MKST~LLLFLLLCLPSWNAGA

utilizes DNA fragments obtained from plasmid p400,18 des-cribed in Example 2 and a fragment obtained after directed mutagenesis in phage Ml3mpl9 marketed by Amersham.

Co~trl~t; on of ~1~ ~ p460 la~ Fragments obtained from plasmid p400,18 a) Plasmid p400,18 was digested with the enzymes Pst and EcoRI. The Pst-EcoRI fragment of 3868 bp contA 1 n; n~
the origin of replication, hereafter called fragment 8 (~hown in Figure 5 as F8), wa8 purified.
~) Plasmid p400,18 was digested with the enzymes NruI
and PstI. The ~mall NruI-PstI fragment of 1062 bp contAin;n~ the promoter, hereafter called fragment 9 (shown in Figure 5 as F9), was purified.
lb~ NruI-EcoRI fragment obtA; neA from phage M13mp19 a) The XhoI-EcoRI fragment of 650 bp deri~ed from plas-mid p400,18 by digestion with the enzymes XhoI and ~coRI and purification, which contains the se~uenc~
coding for the signal peptide of formula (1), was inserted into the polylinker (cloning polysite) of phage M13mpl9 (Amersham) at the restriction ~ite~
SalI/EcoRI. Ligation of the XhoI and SalI sites cau~ed these two sites to disappear.
35 ~3) An oligonucleotide of 63 nucleotides having the following - 26 - 13 349 ~

sequence:

S' - ATG - AAA - TCC - ACG - CTG - CTT - CTC - TTA - TTT -CTG - CTC - CTG - TGC - CTG - CCC - TCT - TGG - M C -05 GCC - GGC - GCT - 3' -was synthesized. Thi~ se~uence codes for the signal peptide of formula (2).
The technique of directed mutagenesis in vitro, per-formed with the aid of the Amersh_m 1523 kit, wa~
used to con~truct a fragment which carried a muta-tion, as regards the sequence coding for the signal peptide, relative to the fragment obtained in a).
This technique, which is described in detail in the booklet accompanying this kit, consi~t~ of the intro-duction of an XhoI-EcoRI fragment of 650 bp (cf.
section a) above) into the double-stranded form of phage M13mpl9, the purification of the single~stranded form of thi~ recombinant phage, the hybridization of th~ above-mentioned oligonucleotide of 63 nucleotides and the action of the Klenow fragment of ~NA polymerase and then T4 ligase to give a double-stranded circular form of the recombinant phage, one of the strand~ of which carries the desired mutation.
~) The phage contA; n i ne the mutated ~NA fragment was digested with the enzymes NruI and EcoRI. The NruI-EcoRI fragment contA; n ~ nE the sequence coding for the ~ignal peptide of formula (2) and the sequence coding for the hirudin variant (Lys47) HV2, hereafter called fragment 10 (shown in Figure 5 a~ F10), was purified.
Fragments 8, 9 and 10 were ligated; the plasmid obtained is plasmid p460, which i8 shown in Figure 5.

2. t~.en~rA1 n ~t.ho~n 1 o~y Plasmid p460 was introduced by transformation into the bacterial strain described in Example 1.
The experiment3 were performed in parallel on two different clones (clones p460,2 and p460,4), in which the presence of the above-mentioned sequence of 63 nucleotides was 05 checked, and on control clone p400,18, in accordance with the procedure described in sections 2.1 and 2.2 of Example 1. The cultures were induced by the method indi-cated in section 2.3 of Example 1, with two modifica-tions: induction was initiated by A~ IPTG when the culture had reached an OD at 600 nm of about 0.6, and it was maintained for 2 hours.
After induction, the cells were subjected to an osmotic shock (cf. section 2.4~ Example 1) and the hirudin variant (Lys~7) HV2 thereby released into the culture supernatant was determined by HPLC.

neterm;nAt;on of the h;rl~;n v~rt~nt fr~A47) HV~
The supernatant obtA; ne~ after osmotic shock was subjected to high pressure liguid chromatography, HPLC, using an apparatus equipped with a calibrated in~ection system and a detector ~et at 220 nm.
The following were u~ed:
a C8 - 300 A reversed-phase column made of st~el, with a length of 7.~ cm and an internal diameter of 4.6 mm (BECKMAN Ultrapore* reference 238 771).
a mobile phase consisting of a linear gradient pas~ing from 85 volumes of solution Sl and 15 volumes of solution S2 to 50 volumes of solution S1 and 50 volumes of solution S2 in 10 minutes.
Solutions Sl and S2 had the following charac-teristics:
S1 = purified water contA~n;n~ 0.1~ (v/v) of trifluoro-acetic acid.
S2 = acetonitrile for HPLC, contA~n;n~ 0.08% (v/v) of 3s trifluoroacetic acid.
* - Trademark ~,j,.:

~334944 The flow rate was 2 ml per minute.
The optical density of the fractions was measured and the amount of periplasmic variant (Ly~47) HV2, expre~sed in milligrams per liter of 3upernatant, was 05 determined by comparison with a st~n~A~d solution of variant (Lys~7) HV2.

III - ~UT.T~
The results are reported in the Tables below.
They are expressed in mg/l of supernatant collected after osmotic shock and brought to an optical density at 600 nm of 1.

PLASMID ~ L~
1.~t exper~m~nt Control 400,18 460,2 Hirudin variant (Lys~7) HV2 in mg/l of supernatant collected after osmotic 1.3 3.1 shock and brought to OD at 600 nm = 1 PLASMID
~n~ exper;m~t Control 400,18 460,2460,4 Hirudin variant (Lys~7) HV2 in mg/l of supernatant collected after osmotic 1.3 5.2 4.
shock and brought to OD at 600 nm = 1 It is clear from the above Tables that the peri-pla~mic production of the hirudin variant (Lys~7) HV2 afforded by plasmids 460,2 and 460,4 i3 considerably greater than that affordcd by plasmid 400,18.

These re~ults show that the signal peptide accor-ding to the invention of formula (2) i~ particularly appropriate for the efficient periplasmic production of a peptide such as the hirudin variant (Lys~7) HV2.

Claims (10)

1. A DNA sequence coding for a signal peptide of the formula MXKSTLLLLFLLLCLPSWNAGA
in which:
A = Alanine M = Methionine C = Cysteine N = Asparagine F = Phenylalanine P = Proline G = Glycine S = Serine K = Lysine T = Threonine L = Leucine W = Tryptophan and X represents APSG or a direct bond between M and K.

2. The sequence according to claim 1 which has the formula

3. The sequence according to claim 1 which has the formula

4. An expression vector for Gram-negative bacteria, comprising a DNA sequence coding for a precursor of a polypeptide capable of being secreted into the periplasmic space of said bacteria, said precursor being a mature polypeptide extended at its N-terminal end by a signal peptide, wherein the portion of said DNA sequence which codes for said signal peptide is a sequence according to claim 1.

5. A Gram-negative bacterium which is transformed by a vector according to claim 4.

6. The bacterium according to claim 5 which belongs to the species Escherichia coli.

7. The bacterium according to claim 6 whose chromosomal DNA comprises a cya mutation by deletion and a crp mutation by deletion.

8. The bacterium according to claim 5 or 6, wherein the polypeptide is a natural form or a variant in amino acid sequence of hirudin.

9. The bacterium according to claim 8 wherein the hirudin variant is (Lys47) HV2.

10. The bacterium according to claim 5, wherein the polypeptide is human growth hormone.