Polypeptides.

The present invention relates to derivatives of granulocyte colony stimulating factor (G-CSF) having good solution stability and to processes for their preparation as well as to pharmaceutical compositions containing them.

The colony stimulating factors are a class of protein hormones which stimulate the proliferation and the function of specific blood cell types such as granulocytes. Granulocytes engulf and devour microbial invaders and cell debris and thus represent a vital factor in response to infection. In this regard granulocytes can extend pseudopods and slip out of the vascular tree between the lining endothelial cells. The neutrophilic granulocytes can then come into direct contact with the microorganisms and destroy them using unique enzyme systems such as those which generate superoxide anions. Since granulocytes have only a short life span in the circulation (approximately 6-12 hours) and are destroyed in the course of their function, it is necessary for the stem cells of the bone marrow to generate as many granulocytes as red blood cells each day. Further, this rate of production of granulocytes needs to increase enormously if the demands of infection are to be met. As a result of their fast turnover, the granulocyte count falls rapidly if the bone marrow is damaged for example by cancer chemotherapy, radiation, AIDS or haematological disorders and patients become liable to overwhelming infection. Indeed sepsis is a common cause of death in cancer patients whose marrow is suppressed by radiation treatment, chemotherapy or their neoplastic disease.

Granulocyte colony stimulating factor (G-CSF) has been described in the literature by Wallet K. <u style="single">et</u><u style="single">al</u> Proc. Natl. Acad. Sci. U.S.A Vol 82, pp 1526-153O and has also been described in European Patent Publication No 169,566 and PCT Patent Publication No WO 87/O1132. G-CSF has been shown to stimulate granulocyte production <u style="single">in</u><u style="single">vivo</u> and to function with minimal side effects. As a result human G-CSF is seen as having potential utility in the management of neutropaenia associated with chemotherapy, radiation therapy, radiation accident or autologous bone marrow transplantation. Moreover G-CSF may have utility in the stimulation of bone marrow suppression associated with AIDS, in the treatment of myelodysplastic syndromes characterised by granulocyte functional abnormalities and as an adjunct to the treatment of severe infections.

In addition to the above certain analogues of G-CSF have been described in PCT Patent Publication No WO 87/O1132, in European Patent Publication No 243,153, in European Patent Publication No 256,843, in European Patent Publication No 272,7O3 and in Biochemical and Biophysical Research Communication [1989] Vol.159, No 1, pp 1O3-111 Kuga T. <u style="single">et</u><u style="single">al</u>. Furthermore, modification of G-CSF and [Ser¹⁷]G-CSF has been effected by substituting the cysteine and serine residues at position 17, but such changes failed to achieve the desired effect (Protein Engineering, Vol 3. No.4 page 36O (199O)).

G-CSF and the analogues referred to above tend to suffer from solution instability in that on standing they tend to precipitate out of solution thus resulting in short shelf life and problems in storage at high concentrations. Moreover G-CSF and certain of the analogues referred to above have a tendency to covalent aggregation on storage.

The present invention is based on the discovery of modifications that may be made to a G-CSF or a derivative thereof having part or all of the amino acid sequence and at least one of the biological properties of naturally occurring G-CSF, for example of naturally occurring human G-CSF, whereby to improve solution stability.

Thus according to one feature of the present invention there is provided a derivative of naturally occurring G-CSF having at least one of the biological properties of naturally occurring G-CSF and a solution stability (as herein defined) of at least 35% at 5mg/ml, the said derivative having at least Cys¹⁷ of the native sequence replaced by a Ser¹⁷ residue and Asp²⁷ of the native sequence replaced by a Ser²⁷ residue.

The derivatives of the present invention may conveniently have at least one further modification selected from:- <ul id="ul0001" list-style="none"><li>a) Glu¹¹ of the native sequence replaced by an Arg¹¹ residue;</li><li>b) Leu¹⁵ of the native sequence replaced by a Glu¹⁵ residue;</li><li>c) Lys²³ of the native sequence replaced by an Arg²³ residue;</li><li>d) Gly²⁶ of the native sequence replaced by an Ala²⁶ residue;</li><li>e) Gly²⁸ of the native sequence replaced by an Ala²⁸ residue;</li><li>f) Ala^3O of the native sequence replaced by an Lys^3O or Arg^3O residue;</li><li>g) Lys³⁴ of the native sequence replaced by an Arg³⁴ residue;</li><li>h) Lys^4O of the native sequence replaced by an Arg^4O residue;</li><li>i) Pro⁴⁴ of the native sequence replaced by an Ala⁴⁴ residue;</li><li>j) Leu⁴⁹ of the native sequence replaced by a Lys⁴⁹ residue;</li><li>k) Gly⁵¹ of the native sequence replaced by an Ala⁵¹ residue;</li><li>l) Gly⁵⁵ of the native sequence replaced by an Ala⁵⁵ residue;</li><li>m) Trp⁵⁸ of the native sequence replaced by a Lys⁵⁸ residue;</li><li>n) Pro^6O of the native sequence replaced by a Ser^6O residue;</li><li>o) Pro⁶⁵ of the native sequence replaced by a Ser⁶⁵ residue;</li><li>p) Pro¹¹¹ of the native sequence replaced by a Glu¹¹¹ residue;</li><li>q) Thr¹¹⁵ of the native sequence replaced by a Ser¹¹⁵ residue;</li><li>r) Thr¹¹⁶ of the native sequence replaced by a Ser¹¹⁶ residue; and</li><li>s) Tyr¹⁶⁵ of the native sequence replaced by an Arg¹⁶⁵ residue.</li></ul>

The presence of at least one further modification selected from (b) to (s) is preferred, but the presence of at least one further modification selected from (b), (d), (e), (f), (n) and (o) is particularly preferred of which further modification (o) is especially preferred.

More preferably the further modification comprises at least one of the following:- <ul id="ul0002" list-style="none"><li>i) Gln¹¹, Pro^6O,⁶⁵ of the native sequence replaced by Arg¹¹, Ser^6O,65;</li><li>ii) Ala¹¹¹, Thr^115,116 of the native sequence replaced by Glu¹¹¹, Ser^115,116;</li><li>iii) Gln¹¹, Trp⁵⁸, Tyr¹⁶⁵ of the native sequence replaced by Arg^11,165, Lys⁵⁸;</li><li>iv) Leu¹⁵, Gly^26,28, Ala^3O of the native sequence replaced by Glu¹⁵, Ala^26,28 Lys^3O; or</li><li>v) Asp²⁷, Pro⁴⁴, Leu⁴⁹, Gly^51,55, Trp⁵⁸ of the native sequence replaced by Lys^49,58, Ala^44,51,55.</li></ul>

The further modification may also, preferably comprise at least one of the following:- <ul id="ul0003" list-style="none"><li>vi) Leu¹⁵, Gly^26,28, Ala^3O of the native sequence replaced by Glu¹⁵, Ala^26,28, Arg^3O; or</li><li>vii) Pro⁶⁵ of the native sequence replaced by Ser⁶⁵; or</li><li>viii) Pro^6O,65 of the native sequence replaced by Ser^6O,65; or</li><li>ix) Gln¹¹, Pro⁶⁵ of the native sequence replaced by Arg¹¹,Ser⁶⁵.</li></ul>

The above defined modifications may thus, if desired, be introduced into any polypeptide having at least one of the biological properties of naturally occurring G-CSF in order to improve the solution stability of the molecule. The modifications of the present invention may thus be applied to such polypeptides which differ in amino acid sequence from that specified herein for the naturally occurring G-CSFs in terms of the identity or location of one or more residues (for example substitutions, terminal and internal additions and deletions). As examples such polypeptides might include those which are foreshortened, for example by deletions; or those which are more stable to hydrolysis (and, therefore, may have more pronounced or longer lasting effects than naturally occurring); or which have been altered to delete one or more potential sites for O-glycosylation (which may result in higher activities for yeast-produced products); or which have one or more cysteine residues deleted or replaced, for example by alanine or serine residues and are potentially more easily isolated in active form from microbial systems; or which have one or more tyrosine residues replaced by phenylalanine and may bind more or less readily to human G-CSF receptors on target cells. The proposed modifications (a) to (s), preferably (i) to (ix) may thus, for example be applied to either native G-CSF having Cys¹⁷ of the native sequence replaced by Ser¹⁷ or to allelic variants and analogues thereof known to possess at least one of the biological properties of naturally occurring G-CSF such as those described in the publications referred to above.

Polypeptides of the present invention that have been tested have been found to possess improved solution stability over the corresponding unmodified polypeptide whilst either retaining significant biological activity or even having improved biological activity.

It will be understood from the above that the property of solution stability is different from that of solubility. Solution stability is the decreased tendency of a substance to precipitate from solution under physiological conditions of pH, temperature and ionic strength.

Solution stability is measured herein by determining the percentage of G-CSF derivative remaining in solution in phosphate buffered saline after 14 days at 37°C given an initial concentration of 1mg/ml, 5mg/ml and/or 1Omg/ml. Measurement of solution stability is described in detail hereinafter in Reference Example 4. Conveniently polypeptides of the present invention will have a solution stability at 5mg/ml of at least 35%, advantageously at least 5O% and preferably at least 75%. Preferably the polypeptides of the present invention will have a solution stability at 1Omg/ml of at least 75%, especially at least 85%.

The expression "naturally occurring G-CSF" as used herein refers to those G-CSFs that have been found to exist in nature and includes the two polypeptides having the amino acid sequence set out in SEQ ID No37. These two polypeptides differ only in so far as a tripeptide insert Val-Ser-Glu is present in one polypeptide between positions 35 and 36, but absent in the other. The numbering system used throughout the present specification is based on the naturally occurring polypeptide without the Val-Ser-Glu insert and the term "native" as used herein refers to this polypeptide without the Val Ser Glu insert. It will be appreciated that the present invention is applicable to all naturally occurring forms of G-CSF and analogues thereof as described above and consequential revision of the position numbers of the polypeptide may be necessary depending on the form of naturally occurring G-CSF selected for modification.

According to a further feature of the present invention there is provided a DNA sequence encoding all or part of the amino acid sequence of a derivative of naturally occurring G-CSF as hereinbefore defined. Such sequences may, for example include 1) the incorporation of codons preferred for expression by selected non-mammalian hosts; 2) the provision of sites for cleavage by restriction endonucleases; and/or 3) the provision of additional initial, terminal or intermediate DNA sequences which facilitate construction of readily expressed vectors. The DNA sequences of the present invention include those useful in securing expression in procaryotic or eucaryotic host cells and the derivatives of the present invention may be in either glycosylated or non-glycosylated form depending upon the host cell selected. Where the derivative of the present invention is obtained in non-glycosylated form, for example following expression in procaryotic host cells, the derivative may, if desired, be glycosylated chemically for example with mammalian or other eucaryotic carbohydrates.

According to a further feature of the present invention there is provided a recombinant vector containing a DNA sequence as hereinbefore defined. The recombinant vector may for example be a biologically functional plasmid or viral DNA vector.

According to a further feature of the present invention there is provided a process for the preparation of a recombinant vector as hereinbefore defined which comprises inserting a DNA sequence as hereinbefore defined into a vector.

According to a further feature of the present invention there is provided a procaryotic or eucaryotic host cell stable transformed or transfected with a recombinant vector as hereinbefore defined.

According to a further feature of the present invention there is provided a process for the preparation of a procaryotic or eucaryotic host cell as hereinbefore defined which comprises transforming or transfecting a procaryotic or eucaryotic cell with a recombinant vector as hereinbefore defined whereby to yield a stably transformed or transfected procaryotic or eucaryotic host.

According to a further feature of the present invention there is provided a process for the preparation of a derivative of naturally occurring G-CSF of the present invention which comprises culturing a procaryotic or eucaryotic host cell of the invention whereby to obtain said derivative. The process will advantageously also include the step of isolating the said derivative produced by expression of the DNA sequence of the invention in the recombinant vector of the invention.

The host cells for use in processes of the present invention are preferably procaryotic such as <u style="single">E</u>.<u style="single">coli</u>, but may be yeast cells such as <u style="single">Saccharomyces</u><u style="single">cerevisiae</u> or mammalian cells such as CHO cells (chinese hamster ovary cells).

According to a further feature of the present invention there is provided a pharmaceutical composition comprising as active ingredient at least one derivative of naturally occurring G-CSF of the present invention in association with a pharmaceutically acceptable carrier or excipient.

According to a further feature of the present invention there is provided a method for providing haematopoietic therapy to a mammal which comprises administering an effective amount of a derivative of the present invention.

According to a further feature of the present invention there is provided a method for arresting the proliferation of leukaemic cells which comprises administering an effective amount of a derivative of the present invention.

Brief Description of the Drawings

<ul id="ul0004" list-style="none"><li>Figure 1 shows the nucleotide sequence of the 167 bp fragment referred to in Example 1;</li><li>Figure 2 shows the amino acid sequence and corresponding nucleotide sequence of native human (hu) G-CSF and restriction sites;</li><li>Figure 3 shows the amino acid sequence and corresponding nucleotide sequence of [Ser ^17,27] hu G-CSF and restriction sites.</li><li>Figure 4 shows the nucleotide sequence of the T4 transcription terminator having (a) terminal SalI and HindIII restriction sites; and (b) terminal SalI and StyI restriction sites;</li><li>Figure 5 shows a restriction map of pTB357 (also referred to herein as pLBOO4);</li><li>Figure 6 shows the nucleotide sequence of the EcoRI-SalI fragment referred to in Reference Example 6(b) but omitting the interferon α₂ gene sequence;</li><li>Figure 7 shows a restriction map of pLBO15 (also referred to herein as pICI OO8O);</li><li>Figure 8 shows a restriction map of pICI 1O79;</li><li>Figure 9 shows a restriction map of pICI 54 (also referred to herein as pCG54;</li><li>Figure 1O shows a restriction map of pCG61;</li><li>Figure 11 shows a restriction map of pICI 11O7 in which the shaded area represents the gene sequence coding for [Ser^17,27]hu G-CSF;</li><li>Figure 12 shows a restriction map of pCG3OO (also referred to herein as pICI 1295.</li></ul>

Detailed Description

Advantageously the derivatives of the present invention are selected to possess one of the further modifications (i), (ii), (iii), (iv), (v), (vi), (vii), (viii) or (ix) or as hereinbefore defined, preferably one of the further modifications (i), (ii), (iv), (vi), (vii), (viii) or (ix) and especially further modification (ii), (iv), (vi), (vii), (viii) or (ix).

Particularly preferred derivatives according to the present invention by virtue of their good solution stability include:-    [Arg¹¹ Ser^17,27,6O,65]G-CSF;    [Glu¹⁵, Ser^17,27, Ala^26,28, Lys^3O]G-CSF;    [Arg¹¹ Glu¹⁵ Ser^17,27,6O,65, Ala^26,28, Lys^3O]G-CSF    [Arg^11,23 Ser^17,27,6O,65]G-CSF    [Arg^11,34, Ser^17,27,6O,65]G-CSF    [Arg^11,4O Ser^17,27,6O,65]G-CSF    [Ala¹,Thr³,Tyr⁴,Arg^5,11,Ser^17,27,6O,65]G-CSF    [Arg¹¹ Glu^15,111 Ser^{17,27,6O,65,115,116},Ala^26,28, Lys^3O]G-CSF    [Arg^11,165, Glu¹⁵, Ser^17,27,6O,65, Ala^26,28, Lys^3O,58]G-CSF    [Arg¹¹, Glu¹⁵ Ser^17,27,6O,65, Ala^{26,28,44,51,55}, Lys^3O,49,58]G-CSF    [Arg^11,165, Glu^15,111, Ser^{17,27,6O,65,115,116}, Ala^26,28,44,51, ^55, Lys^3O,49,58]G-CSF    [Glu¹⁵,Ser^17,27,Ala^26,28,Arg^3O]G-CSF

Especially preferred derivatives of the invention by virtue of their excellent solution stability and good specific acitivity include:- <ul id="ul0005" list-style="none"><li>(i) [Arg¹¹, Ser^17,27,6O,65]G-CSF,</li><li>ii) [Glu¹⁵, Ser^17,27, Ala^26,28, Lys^3O]G-CSF,</li><li>iii) [Arg¹¹, Glu¹⁵,Ser^17,27,6O,65, Ala^26,28, Lys^3O]G-CSF,</li><li>iv) [Arg^11,4O Ser^17,27,6O,65]G-CSF,</li><li>v) [Arg^11,23, Ser^17,27,6O,65]G-CSF,</li><li>vi) [Arg^11,165, Glu¹⁵ Ser^17,27,6O,65, Ala^26,28, Lys^3O,58]G-CSF</li><li>vii) [Arg¹¹ Glu^15,111, Ser^{17,27,6O,65,115,116}, Ala^26,28, Lys^3O]G-CSF,</li><li>viii) [Glu¹⁵,Ser^17,27,Ala^26,28,Arg^3O]G-CSF, and</li><li>ix) [Ala¹,Thr³,Tyr⁴,Arg^5,11,Ser^17,27,6O,65]G-CSF</li><li>x) [Ser^17,27,6O,65]G-CSF,</li><li>xi) [Arg¹¹,Ser^17,27,65]G-CSF, and</li><li>xii) [Ser^17,27,65]G-CSF</li></ul> of which (i), (ii), (iii), (vi), (vii), (viii) and (xii) are most preferred.

These latter human G-CSF derivatives show not only excellent solution stability properties, but also possess improved specific activity over naturally occurring human G-CSF.

A presequence methionine may be either present or absent in the polypeptides of the present invention but is conveniently present.

It has been found advantageous to employ a production vector based on pAT153, comprising:- <ul id="ul0006" list-style="none"><li>i) a promoter and where appropriate an operator therefor, for example a trp promoter or a T7A3 promoter. The T7A3 promoter is the A3 promoter of bacteriophage T7 [see Dunn J.J. and Studier F.W. J. Mol. Biol. <u style="single">166</u>, 477-535 (1983)]. The complete nucleotide sequence of bacteriophage T7 DNA and the locations of T7 genetic elements are set out in this reference;</li><li>ii) a ribosome binding site sequence, for example a trp leader ribosome binding site sequence;</li><li>iii) a cloning site for the gene to be expressed;</li><li>iv) a T4 transcription termination sequence (see SEQ ID No. 51 and Figure 4)</li><li>v) a <u style="single">cer</u> sequence (Summers D. <u style="single">et</u><u style="single">al</u> MGG, <u style="single">2O1</u>, p334-338, 1985)</li><li>vi) a tetracycline repressor gene (Tet R)</li><li>vii) tetracycline resistance gene (Tet A)</li><li>viii) multiple restriction enzyme recognition sequences</li></ul>

SEQ ID No 5O. sets out a sequence which includes an EcoRI restriction endonuclease site (nucleotides 1-6), the A3 promoter sequence (nucleotides 7-52), the trp leader ribosome binding site sequence (nucleotides 53-78) and the translation initiation codon (nucleotides 79-81)

It may be advantageous to cultivate the host capable of expressing a derivative of the invention, in a growth medium and adding a supplement which includes yeast extract to the growth medium during cultivation. It is preferable that addition of the supplement which includes yeast extract is initiated at a predetermined time after the start of cultivation. The rate of addition of the supplement which comprises yeast extract is preferably such that the growth medium does not become exhausted of yeast extract. This is particularly advantageous where the production vector is used with a T7A3 promoter.

It may also be advantageous to cultivate a host, transformed with a recombinant vector carrying genetic material coding for a derivative of the present invention, in the presence of leucine and/or threonine in an amount sufficient to give improved accumulation of the derivative of the present invention. Thus it is particularly advantageous to effect the fermentation in the presence of leucine where the production vector is used with the trp promoter.

In addition to the discovery of modifications that may be made to a G-CSF or derivative thereof having part or all of the amino acid sequence and at least one of the biological properties of naturally occurring G-CSF, to improve solution stability, the present invention is further based on the discovery of modified techniques for the purification of such G-CSFs and derivatives thereof.

Thus for example there is no disclosure in PCT Patent Publication No WO 87/O1132 of the removal of detergent, particularly N-lauroyl sarcosine in salt form (eg. Sarkosyl) from the G-CSF analogues prepared in this PCT Publication. It was therefore necessary to identify such a technique in order that the solution stability of the G-CSF derivatives of the present invention could be assessed at high concentration and in order that formulation studies could be conducted. In one embodiment of the invention detergent removal was effected in the presence of a phosphate buffered saline (pH 7.2 - 7.5). The phosphate buffered saline may conveniently be prepared from isotonic saline and may thus for example have a composition as described in Example 1. In this regard it was found that other buffers were less preferred since either detergent removal, particularly N-lauroyl sarcosine (in salt form) removal, was slower or more protein precipitated out. It is further preferred to effect diafiltration, preferably at this stage, since this was found to improve efficiency without provoking increased protein precipitation. For example diafiltration was found to be preferable to conventional diffusion dialysis. Furthermore it was found that detergent concentration, particularly N-lauroyl sarcosine in salt form (eg. Sarkosyl) concentration, could be reduced below 1% whilst retaining resolution during chromatography. A reduction in initial detergent concentration assists detergent removal and thus it is preferred to use the minimum concentration of detergent, for example N-lauroyl sarcosine (in salt form eg. Sarkosyl), consistent with retaining resolution during chromatography. A particular concentration of detergent, for example N-lauroyl sarcosine (in salt form) eg. Sarkosyl, is thus from O.8% to O.2%, preferably from O.5 to O.2%, especially about O.3%.

In addition to the above it was found that the removal of detergent such as N-lauroyl sarcosine (in salt form) e.g. Sarkosyl activates a trace of proteolytic activity which may complicate product evaluation. It has further been found that this proteolytic activity may be significantly reduced and even eliminated if, after detergent removal by diafiltration, the pH is reduced to below 7.O before substantial proteolysis, conveniently by diafiltration and preferably by dialysis. Thus in a further embodiment of the present invention the reduction or removal of trace proteolytic acitivity may be effected at a pH that is below 7.O but which is sufficiently high to avoid significant hydrolysis of the polypeptide. The pH is advantageously in the range 6.O to 4.5, preferably 5.8 to 5.O especially about 5.4. A further advantage of this embodiment of the invention is that E.coli contaminants and/or degraded or incorrectly folded protein can be precipitated by effecting this lowering of pH. It is preferred that purification include the step of size exclusion chromatography since otherwise the problem of proteolytic degradation is increased and whilst the present embodiment will reduce such degradation it makes it difficult to eliminate.

In addition to the above processes, the introduction of solution stability into a G-CSF or derivative thereof enables substantial simplification of the process of extraction. Thus according to a further feature of the present invention there is provided a process for extracting an active derivative of the invention (as hereinbefore defined) from an inclusion body thereof which comprises 1) suspending said inclusion body in a detergent, particularly N-lauroyl sarcosine in salt form (e.g. Sarkosyl) 2) oxidation, 3) removal of detergent for example as hereinbefore described and 4) maintaining solution obtained following removal of detergent at an elevated temperature for example 3O-45°C, advantageously 34-42°C whereby to precipitate contaminating bacterial protein, product oligomers and/or degradation products. The said solution is conveniently maintained at said elevated temperature for from 6-24 hours, advantageously 8-18 hours preferably 1O-14 hours, especially about 12 hours.

The extraction process of the present invention may for example be effected by lysing host cells followed by centrifugation to obtain the inclusion body for example in the form of a pellet. The inclusion body may then be suspended in a detergent such as, for example N-lauroyl sarcosine in salt form (eg Sarkosyl), preferably 1-3%, especially about 2% N-lauroyl sarcosine in salt form (eg. Sarkosyl). Suspension in detergent may be followed by oxidation, for example in the presence of copper sulphate (CuSO₄) which in turn may be followed by centrifugation.

Where it is possible to wash the inclusion body it is preferred to use urea rather than for example deoxycholate.

The extraction process of the present invention enables the production process to be simplified for example by elimination of the need for the use of size exclusion columns. Moreover the high recovery of product from the heat treatment step appears to be one of the advantages of the increased solution stability of the derivatives of the present invention. Indeed the greater the solution stability the more suited is the protein to the new extraction process. Thus for example it is preferred to apply this extraction process to the extraction of derivatives of the present invention having a solution stability of at least 85% at 1O mg/ml. When the known analogue [Met⁻¹, Ser¹⁷] G-CSF was extracted by the above process, rpHPLC indicated that only 4O% of the desired product remained in solution after heat treatment of a retentate containing 1 mg/ml total protein. At 3 mg/ml total protein, only 19% of the analogue remained in solution. All nucleotide sequences referred to herein are specified in the conventional 5′ - 3′ sense. The derivatives of the present invention are based on human G-CSF which is also referred to as hu G-CSF. Since the derivatives prepared in the Examples are all prepared using E.coli, a presequence methionine will generally be present.

The following materials are referred to hereinafter in the Reference Examples and Examples and their constitution is as follows:-

The term "N-lauroyl sarcosine" as used herein refers to the use of the said substance in salt form. Thus in the Examples N-lauroyl sarcosine is used in the form of the sodium salt. <tables id="tabl0001" num="0001"><img file="EP0459630A2_D0001.tif" /></tables> The above buffers are available from Boehringer Mannheim.

In the site-directed mutagenesis procedure - Reference Example 2 <dl id="dl0001"><dt>Buffer 1</dt><dd>1OO mM Tris HCl pH 8.O 1OO mM NaCl 2O mM MgCl₂</dd><dt>Buffer 2</dt><dd>1O mM Tris HCl pH 8.O 2O mM NaCl 1 mM EDTA</dd><dt>Buffer 3</dt><dd>12 mM Tris HCl pH 7.7 3O mM NaCl 1O mM MgCl₂ 8 mM 2-mercapto ethanol</dd><dt>Buffer 4</dt><dd>6O mM Tris HCl pH 8.O 9O mM NaCl 6 mM MgCl₂ 1O mM DTT</dd></dl> Nucleotide mix 1 25O µM each of dATP, dGTP, dCTP=S (phosphorothioate derivative of dCTP), dTTP and 1 mM ATP Nucleotide mix 2 25O µM each of dATP, dGTP, dCTP, dTTP and 35O µM ATP

M9 minimal media

<tables id="tabl0002" num="0002"><img file="EP0459630A2_D0002.tif" /></tables>

Supplements/75ml

<dl id="dl0002"><dt>3OO µl</dt><dd>5O% glucose</dd><dt>75 µl</dt><dd>1M MgSO₄</dd><dt>75 µl</dt><dd>O.1M CaCl₂</dd><dt>75 µl</dt><dd>4 mg/ml thiamine</dd><dt>75 µl</dt><dd>2O% casin amino acids</dd></dl>

Trace Element Solution (TES)

TES has the following composition:- <tables id="tabl0003" num="0003"><img file="EP0459630A2_D0003.tif" /></tables> and is added to growth media at O.5 ml/l

Geneclean (TM)

The kit contains 1) 6M sodium iodide 2) a concentrated solution of sodium chloride, Tris and EDTA for making a sodium chloride/ ethanol/water wash; 3) Glassmilk (TM)- a 1.5 ml vial containing 1.25 ml of a suspension of silica matrix in water.

This is a technique for DNA purification based on the method of Vogelstein and Gillespie published in Proceedings of the National Academy of Sciences USA (1979) Vol 76, p 615.

Alternatively any of the methods described in "Molecular Cloning - a laboratory manual" Second Edition, Sambrook, Fritsch and Maniatis (Cold Spring Harbor Laboratory, 1989) can be used.

Random Label Kit Product of Pharmacia No 27-925O

The procedure is described in "Molecular Cloning - a Laboratory Manual" Second Edition, Sambrook, Fritsch and Maniatis, pp 1O.13-1O.17 (Published by Cold Spring Harbor Laboratory 1989).

Sequenase (TM)

Chemically modified T7 DNA polymerase

Based on the procedure of Tabor and Richardson published in "Proceedings of the National Academy of Sciences USA (1987) vol 84 pp 4767-4771.

T4 DNA ligase

Described in "Molecular Cloning - a Laboratory Manual" Second Edition, Sambrook, Fritsch and Maniatis 5.6O-5.64 (Published by Cold Spring Harbor Laboratory 1989) and also by Weiss B. <u style="single">et</u><u style="single">al</u> J. Biol. Chem. Vol 243 p 4543 (1968).

The following non-limiting Examples are given by way of illustration only.

Example 1

Preparation of [Ser

17,27

] human G-CSF

The procedure for steps a) and b) in Reference Example 1 was repeated with the following modifications:

Oligonucleotides SEQ ID Nos 24, 25, 26 and 27 (as hereinafter defined) replace SEQ ID Nos 1, 2, 3 and 4 (as hereinafter defined) respectively.

c) Cloning of the gene for [Ser ^17,27] human G-CSF into an expression vector

The gene described above (see Figure 3 and SEQ ID No. 49) was cloned into plasmid vector pICIOO2O. This vector is a pAT153 based plasmid in which the 651 bp EcoRI-AccI region is replaced by a 167 bp EcoRI - ClaI fragment (SEQ ID No.47) consisting of:- <ul id="ul0007" list-style="none"><li>(1) a synthetic E. coli trp promoter and trp leader ribosome binding site</li><li>(2) a translation initiation codon</li><li>(3) a multiple restriction enzyme recognition sequence derived from M13mp18, containing sites for KpnI, BamHI, XbaI, SalI, PstI, SphI and HindIII</li><li>(4) a synthetic transcription termination sequence</li></ul>

The DNA sequence of this region is shown in Figure 1.

The pICIOO2O expression vector was digested to completion with KpnI (BCL) in 1OmM Tris HCl (pH7.5), 1OmM magnesium chloride. The DNA was precipitated with ethanol at -2O°C from a solution containing O.3M sodium acetate and then the 3′- sticky ends were removed by treatment with T4 DNA polymerase for 1O minutes at 37°C as follows:-    DNA (1µg) in water (16µl)    1OX T4 polymerase buffer (2µl)    O.33M Tris acetate pH7.9    O.1M Magnesium acetate    O.66M Potassium acetate    5mM dithiothreitol    1mg/ml bovine serum albumin (BSA PENTAX fraction V)    2mM dNTP mixture (1µl)    T4 DNA polymerase (1µl; 2.5 units/µl BCL)

Water (8Oµl) was added and the mixture extracted with phenol/chloroform (1OOµl) and then with chloroform (1OOµl). The DNA was precipitated with ethanol (25Oµl) at -2O°C after addition of 3M sodium acetate (1Oµl) then digested to completion with SalI (BCL) in 15OmM NaCl, 1OmM MgCl₂ and 1OmM Tris HCl (pH7.5). The Kpn-blunt ended to SalI vector was purified from a O.7% agarose gel and isolated by use of Geneclean (trademark) following the manufacturer's (Bio1O1, USA) recommended procedure.

The synthetic gene was isolated from the pSTP1 vectors as follows. The vectors were digested with ScaI and SalI (both from BCL) in 1OOmM Nacl, 1OmM MgCl₂ and 1OmM Tris HCl (pH7.5). The 53O bp fragment was purified from a O.7% agarose gel and isolated by use of Geneclean (trademark) following the manufacturer's (Bio1O1) recommended procedure.

For ligation, a mixture of the ScaI - SalI gene fragment (5Ong) and the pICIOO2O vector fragment (1OOng) in 2Oµl of a solution containing 5OmM Tris HCl (pH7.6), 1OmM MgCl₂, 1mM ATP, 1mM DTT, 5% w/v PEG 8OOO and T4 DNA ligase (2 units; BRL) were incubated at 16°C for 2O hours. The resulting mixture was used to transform competent <u style="single">E. coli</u> HB1O1 cells (as supplied by BRL) as described herein. Transformants were selected for by growth on L-agar plates containing 5Oµg/ml ampicillin and screened for the presence of the gene by colony hybridisation with a ³²P labelled probe (SEQ ID No 24) as described herein. Plasmid DNA was prepared from 6 positively hybridising colonies, purified by centrifugation in a caesium chloride gradient and the sequence confirmed by dideoxy sequencing as described herein.

The plasmid containing this gene was designated pICI 1O8O.

d) <u style="single">Subcloning of an expression cassette containing a gene</u> for [Ser^17,27]G-CSF into M13mp18.

The following subcloning was effected to provide a starting point for preparation of the G-CSF derivatives detailed in Examples 3-8.

Plasmid DNA from pICI1O8O (purified by caesium chloride density centrifugation) was digested to completion with EcoRI and SalI (BCL) according to the manufacturer's instructions. The small EcoRI-SalI fragment containing the trp promoter and [Ser^17,27]G-CSF gene was isolated from a O.7% agarose gel by use of Geneclean (trademark). This fragment was cloned into an EcoRI-SalI cut M13mp18 vector (DNA supplied by Amersham International; enzymes from BCL). The fragments were ligated together in 5x BRL ligation Buffer using BRL T4 DNA ligase (described previously). The ligation mix was used to transfect competent <u style="single">E. coli</u> TG1 cells (made competent according to the calcium chloride method of Mandel and Higa described in Molecular Cloning - A Laboratory Manual - Maniatis et al Cold Spring Harbor). The transfected cells were suspended in TY top agar containing 2% X-Gal in DMF and 2OOµl log phase <u style="single">E. coli</u> TG1 cells and were plated on 2x TY agar plates (TY top agar - 8g Bactotryptone, 5g Yeast Extract, 5g NaCl, 3.75g Bacto-agar in 5OOµl sterile H₂O; TY plates - 8g Bactotryptone, 5g Yeast-extract, 5g NaCl, 7.5g Bactoagar in 5OO ml sterile H₂O.) Four white plaques were picked into 4 x 2 ml 1% <u style="single">E. coli</u> TG1 cells in TY broth (8g Bactotryptone, 5g Yeast extract, 5g NaCl in 5OOml sterile H₂O) aliquots and grown for 6 hours at 37°C. The 2ml cultures were split into O.5ml and 1.5ml aliquots. The bacteria were centrifuged out of solution in an Eppendorf, (trademark) microfuge and the supernatents were transferred to sterile eppendorf (trademark) tubes. The O.5ml aliquots were stored at -2O°C as phage stocks. The 1.5ml aliquots were used to prepare single stranded DNA following the method in the Amersham International M13 sequencing handbook (see below). These DNA samples were then sequenced using oligonucleotides SEQ 1D No 22, SEQ 1D No 23 and M13 Universal sequencing primer. The reactions were carried out using the Sequenase kit (trademark) according to the manufacturers instructions. All 4 clones had the correct DNA sequence for [Ser^17,27]G-CSF.

Large-scale single stranded DNA preparation

For single stranded DNA preparations of between 2OO-5OOµg of DNA/ml, the method in the Amersham International "Oligonucleotide Directed Mutagenesis" was used. A detailed procedure is carried out as follows:-

LARGE - SCALE SINGLE STRANDED DNA PREP:

<ul id="ul0008" list-style="none"><li>A. Preparation of 1ml phage stock <ul id="ul0009" list-style="none"><li>1. Pick a single TG1 <u style="single">E.coli</u> colony from a glucose/minimal medium plate. Grow overnight in 1Oml 2 x TY medium, shaken at 37°C. Add 1Oµl to 2Oml of fresh medium, and shake at 37°C for 3 hours.</li><li>2. Inoculate 1ml 2 x TY medium in a 1Oml sterile culture tube with 1OOµl of 3 hour culture from step 1.</li><li>3. Inoculate the 1ml culture with a recombinant plaque.</li><li>4. Incubate for 4 hours with shaking at 37°C. Transfer to a microcentrifuge tube.</li><li>5. Centrifuge for 5 minutes at ambient temperature. Pour supernatent into a fresh tube. Store overnight at 4°C. Set up an overnight culture of TG1 <u style="single">E.coli</u> for the next stage.</li></ul></li><li>B. Growth of 1OOml phage culture. <ul id="ul0010" list-style="none"><li>1. Inoculate 1OOml 2 x TY medium with 1ml of overnight TG1 culture and shake at 37°C to an O.D_5OO of O.3 .</li><li>2. Add the 1ml phage supernatent from A5 (above) to the 1OOml culture.</li><li>3. Incubate for 5 hours with shaking at 37°C. Transfer to centrifuge tubes.</li><li>4. Centrifuge at 5OOO x g for 3O minutes at 4°C.</li><li>5. Transfer supernatent to a clean centrifuge tube. Take care not to carry over any cells (retain bacterial pellet for RF DNA preparation)</li><li>6. Add O.2 volumes of 2O% w/v PEG 6OOO in 2.5M NaCl to the supernatent. Mix well and then leave to stand for 1 hour at 4°C.</li><li>7. Centrifuge at 5OOO x g for 2O minutes at 4°C. Dscard supernatent.</li><li>8. Centrifuge at 5OOO x g for 5 minutes, and remove all remaining PEG/NaCl with a drawn out Pasteur pipette.</li><li>9. Resuspend the viral pellet in 5OOµl water (double distilled) and transfer to a microcentrifuge tube (1.5ml).</li><li>1O. Centrifuge for 5 minutes in a microcentrifuge to remove any remaining cells. Transfer the supernatent to a fresh microcentrifuge tube.</li><li>11. Add 2OOµl 2O% PEG 12.5M NaCl to the supernatent mix well then leave to stand at ambient temperature for 15 minutes.</li><li>12. Centrifuge for 5 minutes, discard supernatent.</li><li>13. Centrifuge for 2 minutes. Carefully remove all traces of PEG/NaCl with a drawn out Pasteur pipette.</li><li>14. Resuspend the viral pellet in 5OOµl double distilled water.</li><li>15. Add 2OOµl phenol saturated with 1OmM Tris HCl pH8.O, 1mM EDTA. Vortex briefly.</li><li>16. Stand tube for 15 minutes at room temperature.</li><li>17. Centrifuge for 3 minutes.</li><li>18. Transfer supernatent to fresh tube.</li><li>19. Repeat steps 15-18.</li><li>2O. Add 5OOµl chloroform and extract aqueous phase twice.</li><li>21. Add 5Oµl 3M sodium acetate and 1ml absolute ethanol. Mix.</li><li>22. Place in a dry ice and ethanol bath for 2O minutes.</li><li>23. Centrifuge for 15 minutes.</li><li>24. Wash each pellet with 1ml -2O°C ethanol. Pour off.</li><li>25. Vacuum dry pellet and raise in 5Oµl double distilled water. This procedure yields 1OO-2OOµg single stranded DNA.</li></ul></li></ul>

e) <u style="single">Fermentation</u>

pICI 1O8O was transformed into <u style="single">E. coli</u> strain MSD 522 and the resultant recombinants purified and maintained on glycerol stocks at -8O°C.

An aliquot of the culture was removed from stock and streaked onto agar plates of L-ampicillin to separate single colonies after overnight growth at 37°C. A single desired colony was removed and resuspended in 1O ml L-ampicillin broth and 1OOµl immediately inoculated into each of 1O 25O ml Erlenmeyer flasks containing 75 ml L-ampicillin broth. After growth for 16h at 37°C on a reciprocating shaker the contents of the flasks were pooled and used to inoculate a fermenter containing 2OL LCM5O growth medium.

Composition of LCM5O

<tables id="tabl0004" num="0004"><img file="EP0459630A2_D0004.tif" /></tables><tables id="tabl0005" num="0005"><img file="EP0459630A2_D0005.tif" /></tables> Fermentations were then carried out at a temperature of 37°C and pH, controlled by automatic addition of 6M sodium hydroxide solution, of pH 6.7. The dissolved oxygen tension (dOT) set point was 5O% air-saturation and was initially controlled by automatic adjustment of the fermenter stirrer speed. Air flow to the fermenter, initially 2OL/min, corresponding to 1 volume per volume per minute (VVM) was increased to 5OL/min (2.5 VVM) when the fermenter stirrer speed approached 8O-9O% of its maximum. Since the oxygen transfer rate (OTR) of the fermenters was unable to meet the oxygen uptake rate (OUR) of the bacteria at a cell density greater than that corresponding to an OD_55O of 5O under the conditions described, dOT in the fermenter at cell densities greater than this was maintained at 5O% air-saturation by restricting bacteria oxygen uptake rate. This was achieved by formulating the medium to become carbon-limited at OD_55O of 5O and then supplying a feed of the limiting carbon source, together with ammonium sulphate and yeast extract, at a rate which restricted bacterial growth rate. Fermentations were performed for 16h and during that time samples were taken for measurement of optical density (OD_55O), cell dry weight and accumulation of G-CSF within the cells. G-CSF accumulation was measured by scanning Coomassie blue stained SDS-PAGE gels of whole cell lysates of the sampled bacteria as is well known in the art.

When OD_55O reached 25, casein hydrolysate solution (1OOg/1 Oxzoid L41) was pumped into the fermenters at a rate of 1.5g/1/h.

When OD_55O reached approximately 5O, the supply of carbon-source in the fermentation batch became exhausted leading to a rapid rise in dOT from 5O% air saturation. At this point, a feed containing glycerol (47Og/1), yeast extract (118g/1) and ammonium sulphate (118g/1) was pumped into the fermenters at a rate which returned and then maintained the dOT at 5O% air saturation with the fermenter stirred at <u style="single">ca</u> 8O% of its maximum. After <u style="single">ca</u> 13-14h this fed-batch feed was replaced with a second feed containing glycerol (715g/L) and ammonium sulphate (143g/L) only. Casein hydrolysate feeding was maintained at 1.5g/L/h throughout. After approximately 16 hours, when microscopic examination of the culture showed the presence of large inclusion bodies within a majority of the cells, bacteria were harvested on a Sorval RC3B centrifuge (7OOOg, 3O min., 4°C) and stored frozen at minus 8O°C.

f) <u style="single">Purification</u>

Frozen cell paste (5OOg) was resuspended at 4°C in 5OmM Tris HCl, 25mM EDTA, pH8.O (5 litres) using a Silverson model AXR homogeniser. The suspension was lysed by passing three times through a Manton-Gaulin homogeniser at 6OOOpsi and centrifuged at 5OOOxg for 3O minutes in a Sorvall RC3C centrifuge using a H6OOOA rotor. The supernatant was discarded and the pellet fraction stored at -2O°C before further purification. The pellet fraction (6O-1OOg) was thawed and resuspended in 1% w/v deoxycholic acid (sodium salt) in 5mM EDTA, 5mM dithiothreitol, 5OmM Tris HCl, pH9.O (12OOml) containing 1mg/ml of sodium azide using a Polytron homogeniser with a PTA 2O probe at speed setting 5. The suspension was mixed for 3O minutes at room temperature and centrifuged at 65OOxg for 3O minutes in a Sorvall RC 5C centrfigure using a GSA rotor. The supernatant was discarded and the pellet was retreated twice in the same manner. The pellet was next twice resuspended in water (1 litre) and centrifuged at 15,OOOxg for 2O minutes. The final pellet containing washed inclusion bodies was solubilised in 2% w/v N-lauroyl sarcosine sodium salt (Sarkosyl) in 5OmM Tris. HCl, pH 8.O (15Oml) containing 1mg/ml sodium azide. Cupric sulphate was added to 2OµM and the mixture stirred for 16 hours at 2O°C before centrifugation at 3O,OOOxg for 3O minutes in a Sorvall RC5C centrifuge using a SS34 rotor. The supernatant containing the derivative was stored at -2O°C in 5Oml aliquots before further purification.

Solubilised derivative (2Oml) was thawed and passed through a 5µm filter to remove any particulate material. The filtrate was applied to a column (5 x 9O cm) of Ultrogel AcA54 equilibrated with O.3% w/v N-lauroyl sarcosine (sodium salt) in 5OmM Tris. HCl, pH 8.O containing 1mg/ml sodium azide at 4°C. The column was eluted with the same buffer at a flow rate of 2.5 ml/minute and fractions of 1Oml were collected. Fractions containing the derivative protein were pooled (approximately 1OOml) and stored at 4°C.

Pooled derivative-containing fractions from several columns were combined (3OO-5OOml) and dialysed against 1OmM sodium phosphate, 15OmM sodium chloride pH 7.4 (3-5 litres) containing 1mg/ml sodium azide using an Amicon CH2A-1S spiral cartridge diafiltration apparatus equipped with a S1Y1O membrane (1OkD cut-off). The retentate was centrifuged at 3O,OOOxg for 3O minutes in a Sorvall RC5C centrifuge using an SS34 rotor, and the supernatant dialysed in Spectropor 6-8kD cut-off dialysis tubing for 4O hours against three changes (8 litres/3OOml of supernatant) of 2OmM sodium acetate, 1OOmM sodium chloride, pH 5.4 containing 1mg/ml sodium azide. The precipitate which formed was removed by centrifugation at 3O,OOOxg for 3O minutes and the supernatant dialysed for 24 hours against water containing 1mg/ml sodium azide followed by 72 hours against six changes of water. The final retentate was clarified by centrifugation at 3O,OOOxg for 3O minutes and stored frozen at -2O°C (protein concentration about 1mg/ml) or at 4°C after freeze drying.

The concentration of N-lauroyl sarcosine (sodium salt) had fallen to below O.OO1% w/v after diafiltration and was below the limit of detection (about O.OOO1%) of the rpHPLC method used after dialysis against water.

Example 2

Preparation of [Ser

17,27

] human G-CSF

The procedure described in Example 1 was repeated except as follows:-

The duplex I was phosphorylated with T4 polynucleotide kinase and digested with MstII (1O units) in 1 X H buffer (BCL; 3Oµl) for 2 hours at 37°C.

Following precipitation with ethanol, the 143 bp EcoRI-MstII fragment was purified on a 1O% polyacrylamide gel containing 7M urea, isolated by electroelution from a gel slice and the DNA strands annealed as described in Reference Example 1.

The synthetic EcoRI-MstII fragment described above was cloned into the plasmid vector pAG88 described in Reference Example 1. For vector preparation, pAG88 (1Oµg) was digested with MstII (2O units; BCL) in 1 X H buffer (BCL; 1OO µl) for 2 hours at 37°C. The DNA was precipitated with ethanol from O.3 M sodium acetate at -2O°C then digested with EcoRI (2O units; BCL) in 1 X H buffer (BCL; 1OO µl) for 2 hours at 37°C. Following precipitation with ethanol, the large EcoRI-MstII fragment was purified on a 1% agarose gel and purified using Geneclean (trademark) as described by the manufacturer (Bio 1O1, USA). Colonies containing the synthetic fragment were confirmed by screening with a radioactive probe prepared from oligonucleotide (SEQ ID No 24) and the correct sequence confirmed by DNA sequencing as described in Reference Example 1. The plasmid containing the gene for [Ser^17,27]G-CSF was designated pICI11O7. The gene was cloned into expression vector pICIOO2O and fermentation and protein purification was effected as described in Example 1.

Example 3

Preparation of [Arg¹¹ Ser

17,27,6O,65

] human G-CSF

The procedure described in Reference Example 2 was repeated using the mutagenic template M13mp18 containing the gene for [Ser^17,27]G-CSF described in Example 1 or 2. The mutagenic oligonucleotides used are designated SEQ 1D No 28 and SEQ 1D No 29 (as hereinafter defined).

The triplet ACG in SEQ 1D No 28 serves to convert Gln at position 11 to Arg and the first and last AGA triplets in SEQ ID No 29 serve to convert Pro at positions 65 and 6O to Ser. The mutagenesis was carried out as described in Reference Example 2 using SEQ ID No 29 in a single priming mutagenesis. This yielded a single plaque which incorporated the Pro 6O Ser and Pro 65 Ser changes. Single stranded DNA was prepared from this plaque as described in Reference Example 2. This DNA was used as a mutagenic template in a single priming mutagenesis using SEQ ID No 28 as mutagenic primer. This yielded &gt;1OO plaques, 3 of which were screened by DNA sequencing as previously described. All 3 had the full set of changes incorporated. Double - stranded RF DNA was prepared from one of the plaques by following the procedure for large scale preparation of single stranded DNA (step d in Example 1) to step B5. The RF DNA was extracted from the bacterial pellet by the alkali lysis procedure of Birnboim and Doly (Nucleic Acids Research (1979) <u style="single">7</u>, 1513-1523) and purified by caesium chloride density gradient centrifugation as described in "Molecular Cloning - a Laboratory Manual" by Sambrook, Fritsch and Maniatis (Cold Spring Harbor Publication). The purified RF DNA was digested with EcoRI and SalI in buffer H as described previously and the small fragment, containing the trp promoter, ribosome binding site, translation initiation codon and gene for [Arg¹¹,Ser^17,27,6O,65]G-CSF isolated from a O.7% agarose gel by use of Geneclean (TM). The fragment was ligated into an EcoRI-SalI digested pICIOO2O vector, using a 2:1 molar excess of insert to vector, with T4 DNA ligase (BRL) and ligase buffer, essentially as described previously. The ligation mix was used to transform <u style="single">E.Coli</u> strain HB1O1. Transformants were selected for by growth on L-agar plates containing 5Oµg/ml ampicillin. Colonies were screened for the presence of the inserted DNA by restriction analysis of plasmid DNA prepared by the method of Birnboim and Doly as described in "Molecular Cloning - a Laboratory Manual" Sambrook, Fritsch and Maniatis (Cold Spring Harbor Publication). Plasmid DNA from a colony containing the expected 619bp EcoRI - SalI insert was used to transform <u style="single">E.coli</u> strain MSD522 and designated pICI1239. Fermentation and purification were effected as described in Example 1.

Example 4

Preparation of [Ser

17,27,115,116

Glu¹¹¹] human G-CSF

The procedure described in Example 3 was repeated using the mutagenic template M13mp18 containing the gene for [Ser^17,27]G-CSF described in Example 1 or 2. The mutagenic oligonucleotide used is designated SEQ ID No 3O (as hereinafter defined)

The triplet GCT serves to convert Thr at position 116 to Ser, the triplet AGA serves to convert Thr at position 115 to Ser and the triplet TTC serves to convert Ala at position 111 to Glu. The mutagenesis procedure was essentially as described for Example 3 and the expression cassette was transferred to the expression plasmid to give pICI 1243. Fermentation and purification was effected as described in Example 1.

Example 5

Preparation of [Arg¹¹, Ser

17,27

, Lys⁵⁸, Arg¹⁶⁵] human G-CSF

The procedure described in Example 3 was repeated using the mutagenic template M13mp18 containing the gene for [Ser^17,27]G-CSF described in Example 1 or 2. The mutagenic oligonucleotides used are designated SEQ ID No 28, SEQ ID No 31 and SEQ ID No 32 (as hereinafter defined)

The triplet TTT in SEQ ID No 31 serves to convert Trp at position 58 to Lys and in SEQ ID No 32 the second GCG triplet serves to convert Tyr at position 165 to Arg.

The mutagenesis procedure was initially carried out as a double priming experiment using SEQ ID No 31 and SEQ ID No 32 as mutagenic oligonucleotides as described for Reference Example 2. This yielded 2 plaques both of which had the SEQ ID No 32 change (Tyr 165 Arg) but not the SEQ ID No 31 change. Single stranded DNA was prepared from one of these plaques as described in Example 1. This DNA was used as a mutagenic template in a double priming mutagenesis using SEQ ID No 28 and SEQ ID No 31 as mutagenic primers. This yielded 2 plaques one of which had the complete set of changes incorporated and the expression cassette was transferred to the expression plasmid to give pICI 1246. Fermentation and purification was effected as described in Example 1.

Example 6

Preparation of [Glu¹⁵, Ser

17,27

, Ala

26,28

, Lys

3O

] human G-CSF

a) The procedure described in Example 3 was repeated using the mutagenic template M13mp18 containing the gene for [Ser^17,27]G-CSF described in Example 1 or 2. The mutagenic oligonucleotides used are designated SEQ ID No 33 and SEQ ID No 34 (as hereinafter defined).

The triplet TTC in SEQ ID No 33 serves to convert Leu at position 15 to Glu. In SEQ ID No 34 the first TTT triplet serves to convert Ala at position 3O to Lys and the triplets AGC serve to convert Gly at position 28 and 26 to Ala.

The mutagenesis procedure was essentially as described in Reference Example 2 as a double priming experiment and the expression cassette transferred to the expression plasmid to give pICI 1266. Fermentation was effected as described in Example 1.

b) <u style="single">Purification</u>

Frozen cell paste was lysed and the crude pellet fraction separated as in Example 1. The inclusion bodies in the pellet containing this protein were solubilised by the deoxycholic acid (sodium salt) buffer described in Example 1. The following modified procedure was used for this protein.

Crude pellet fraction (6O-1OOg) was thawed and resuspended in 25mM EDTA, 5OmM Tris.HCl, pH 8.O (12OOml) using a Polytron homogeniser with a PTA 2O probe at speed setting 5. The suspension was mixed at room temperature for 3O minutes and centrifuged at 6,5OO x g for 3O minutes in a Sorvall RC5C centrifuge using a GSA rotor. The supernatant was discarded and the pellet retreated twice in the same manner. The pellet was next twice resuspended in water (1 litre) and centrifuged as in Example 1. Thereafter the purification procedure was as in Example 1.

Example 7

Preparation of [Ser

17,27

Lys

49,58

Ala

44,51,55

] human G-CSF

The procedure described in Example 3 was repeated using the mutagenic template M13mp18 containing the gene for [Ser^17,27]G-CSF described in Example 1 or 2. The mutagenic oligonucleotides used are designated SEQ ID No 35 and SEQ ID No 36 (as hereinafter defined). In SEQ ID No 35 the triplets AGC serve to convert Gly to Ala at position 51 and Pro to Ala at position 44 and the triplet TTT serves to convert Leu to Lys at position 49. In SEQ ID No 36 the triplet TTT serves to convert Trp to Lys at position 58 and the second AGC triplet serves to convert Gly to Aln at position 55. The mutagenesis was carried out as a double priming experiment as described in Reference Example 2. This yielded 16 plaques. 8 Plaques were screened by DNA sequencing as described in Example 3. All plaques had the SEQ ID No 36 changes (Gly55Ala, Trp58Lys) but none had the SEQ ID No 35 changes. Single stranded DNA was prepared from one of these plaques as described in Example 1(d) and used as a mutagenic template in a single priming mutagenesis using SEQ ID No 35 as mutagenic primer. This yielded 5O plaques, 3 of which were screened by DNA sequencing, 2 had the complete set of changes. The expression cassette was transferred to the expression plasmid to give pICI 1297. Fermentation and purification was effected as described in Example 1.

Example 8

Preparation of [Arg¹¹,Glu¹⁵,Ser

17,27,6O,65

,Ala

26,28

,Lys

3O

] human G-CSF

The procedure described in Example 3 was repeated using the mutagenic template M13mp18 containing the gene for [Glu¹⁵,Ser^17,27,Ala^26,28, Lys^3O] human G-CSF described in Example 6. The mutagenic oligonucleotide used is designated SEQ ID No 28 which serves to convert Gln at position 11 to Arg. The modified gene was isolated and ligated into pICIOO2O vector (Example 1). This vector was used to transform <u style="single">E. coli</u> strain MSD522 as described in Example 3 and designated pICI1347. pICI1347 plasmid DNA was isolated from MSD522, purified by caesium chloride density centrifugation and digested to completion with BamHI and SalI (BCL) Plasmid DNA (5 µg) was incubated at 37°C for 2 hours in BCL high salt buffer (1OO µl) (5O mM tris HCl pH 7.5, 1O mM MgCl₂, 1OO mM NaCl, 1mM dithioerythritol) containing BamHI (4O units) and SalI (5O units). The DNA was precipitated by addition of 3M sodium acetate (1O µl) and absolute ethanol (25O µl) and cooling to -2O°C for 2 hours, collected by centrifugation (1O min at 1O,OOO rpm), dried <u style="single">in vacuo</u> and dissolved in water (1O µl). Sample loading buffer (2 µl containing 24O mM tris acetate pH 7.8, 6 mM EDTA, 2O% sucrose, O.2% xylene cyanol and O.2% bromophenol blue ) was added and the mixture loaded onto a O.7% agarose preparative gel (in 4O mM tris acetate (pH 7.8) and 1 mM EDTA) containing ethidium bromide (O.5 µg/ml) and electrophoresed at 1OO volts for 1 hour. The large BamHI - SalI vector fragment was isolated from a O.7% agarose gel by use of Geneclean (trademark). In a similar manner, pICI1239 plasmid DNA from Example 3 was isolated and digested with BamHI and SalI. The small BamHI - SalI fragment, containing the Ser codons at position 6O and 65, was isolated and ligated to the large BamI - SalI vector fragment described above. The mixture was used to transform <u style="single">E. coli</u> strain MSD522 and the plasmid designated pICI1348. Fermentation and purification was effected as described in Example 6.

Example 9

The procedure of Examples 1 and 2 was repeated using <u style="single">E.coli</u> strain TG1 instead of E.coli strain MSD 522 in the fermentation step (see for example Example 1(e)).

Example 1O

Alternative Extraction Process for Human [Met⁻¹ Arg¹¹ Ser

17,27,6O,65

]G-CSF

Frozen cell paste (64O g) was resuspended at 4°C in 5OmM Tris HCl, 5mM EDTA, 5mM dithiothreitol, 2M urea, pH 8.O containing 1 mg/ml sodium azide (5 litres) using a Polytron homogeniser with a PTA2O probe at speed setting 7/8. The suspension was lysed by passing three times through a Manton-Gaulin Lab 6O/6O homogeniser at 6OOO psi and flushed through with a further 1 litre of buffer. Cooling was provided by a single pass Conair chiller at -2O°C. The lysate was centrifuged at 5OOO xg for 3O minutes in a Sorvall RC3C centrifuge using an H6OOOA rotor.

The supernatant was discarded and the pellet (about 45O g) was resuspended in the same buffer (1O litres). The suspension was mixed for 3O minutes at room temperature and centrifuged at 5OOO rpm for 3O minutes in two Sorvall RC3C centrifuges using H6OOOA rotors. the supernatant was discarded and the pellet retreated twice in the same manner. The pellet was next twice resuspended in water (1O litres) and centrifuged at 5OOO rpm for 3O minutes. The final pellets containing washed inclusion bodies were resuspended in 2% w/v N-lauroyl sarcosine sodium salt in 5OmM Tris HCl, pH 8.O (1 litre) containing 1 mg/ml sodium azide using a Polytron homogeniser at speed setting 7. 2O mM cupric sulphate in water (1.5 ml) was added and the mixture stirred overnight at room temperature before centrifugation at 1O,OOO rpm for 3O minutes in a Sorvall RC5C centrifuge using a GSA rotor.

The supernatant containing the derivative was filtered through a 5µm filter to remove any particulate matter, diluted six-fold with 5O mM Tris HCl, pH 8.O containing 1 mg/ml sodium azide at 4°C, and diafiltered at maximum pressure in an Amicon DC2O ultrafiltration device fitted with a S1OY1O cartridge (1O kd cut-off) against 1O mM sodium phosphate, 15O mM sodium chloride pH 7.4 (9O litres) containing 1 mg/ml sodium azide. A precipitate formed towards the end of the diafiltration.

The retentate (2.1 mg/ml total protein, 1.7 mg/ml product) was collected in 4 litre, screw top, polypropylene containers and incubated overnight at 37°C. The precipitate which formed was removed by centrifugation at 5OOO rpm for 45 minutes in a Sorvall RC3C, and the supernatant stored at 4°C.

Monitoring by SDS-PAGE and rpHPLC, showed that during the final heat treatment contaminating E. coli proteins, product oligomers, and degradation products were selectively precipitated, with some 85% of the desired product remaining in solution. The highly enriched clarified, heat treated product solution was fully biologically active and stable at 2O mg/ml at 37°C over two weeks with no evidence of proteolytic degradation and less than 2O% precipitation. This provided an excellent intermediate for further chromatographic purification.

Example 11

Characterisation of G-CSF and derivatives thereof

A water solution of [Met⁻¹, Ser¹⁷] G-CSF and derivatives thereof (Examples 1-9) (protein concentration about 1mg/ml) were concentrated to at least 11mg/ml of protein on an Amicon YM1O membrane at 4°C. To prevent any precipitation during concentration, the starting solution pH5.5 was first adjusted to pH8.5 by the addition of ammonium hydroxide to a final concentration of about O.25mM. After concentration the pH of the solution had fallen to about 8.O.

The concentrated protein solution was adjusted to 1Omg/ml protein (estimated from a 1mg/ml solution giving an A_28O of 1.O) by addition of 2O fold concentrated phosphate buffered saline. This 1Omg/ml solution of derivative in 1OmM sodium phosphate, 15OmM sodium chloride, pH7.4 (PBS) provided a common stock solution from which to establish homogeneity, identity, biological activity and solution stability of the protein. A stock solution of human G-CSF at 1mg/ml concentration in PBS prepared as described in Reference Example 1 was also prepared.

Each protein was shown to be at least 95% one component by PAGE-SDS run under reducing and non-reducing conditions and by reverse phase HPLC. Repeated amino acid composition analysis after acid hydrolysis in 6NHCl at 11O°C provided amino acid ratios for each derivative, and an accurate measurement of the protein concentration in the stock solution. This protein concentration together with the mean of bioassay titres obtained on at least six different days was used to determine the specific activity of the derivative. N-terminal sequence analysis and electrospray mass spectrometric analysis of selected derivatives gave the expected sequences and molecular weights.

Example 12

Preparation of [Arg¹¹,Ser

17,27,6O,65

]human G-CSF using production vector including trp promoter

a) Plasmid pICI1239 (described in Example 3) was digested with EcoRI and SalI in buffer H as described previously. The small EcoRI-SalI fragment containing the trp promoter, ribosome binding site and gene for [Arg¹¹,Ser^17,27,6O,65]hu G-CSF was isolated from a O.7% agarose gel by use of Geneclean(TM). A vector fragment was prepared from pICI OO8O (see Reference Example 6) by digestion with EcoRI and XhoI in buffer H and the large EcoRI-XhoI fragment isolated from a O.7% agarose gel by use of Geneclean(TM). The small EcoRI-SalI fragment was ligated into the EcoRI-XhoI vector fragment, using a 2:1 molar excess of insert to vector as described previously and the ligation mix used to transform E. coli strain MSD 522. Transformants were selected for growth on L-agar plates containing tetracycline (15µg/ml). Three colonies were selected and grown up in M9 minimal media (75ml) containing supplements and tetracycline (15µg/ml) at 37°C for 2O hours on a reciprocating shaker. Protein accumulation was measured by scanning Coomassie blue stained SDS-PAGE gels of whole cell lysate. All three clones expressed [Arg¹¹,Ser^17,27,6O,65]hu G-CSF. Plasmid DNA from one of the colonies was designated pICI1327 and the sequence of the promoter and gene confirmed by standard dideoxy sequencing procedures as described previously.

b) <u style="single">Fermentation</u>

pICI 1327 was transformed into <u style="single">E. coli</u> strain MSD 522 and the resultant recombinants purified and maintained on glycerol stocks at -8O°C.

An aliquot of the culture was removed from stock and streaked onto agar plates of tetracycline to separate single colonies after overnight growth at 37°C. A single desired colony was removed and resuspended in 1O ml tetracycline broth and 1OOµl immediately inoculated into each of 3 25O ml Erlenmeyer flasks containing 75 ml tetracycline broth. After growth for 16h at 37°C on a reciprocating shaker the contents of the flasks were pooled and used to inoculate a fermenter containing 2OL growth medium.

Composition of Growth Medium

<tables id="tabl0006" num="0006"><img file="EP0459630A2_D0006.tif" /></tables>

Fermentations were then carried out at a temperature of 37°C, and at a pH, controlled by automatic addition of 6M sodium hydroxide solution, of pH 6.7. The dissolved oxygen tension (dOT) set point was 5O% air-saturation and was initially controlled by automatic adjustment of the fermenter stirrer speed. Air flow to the fermenter, initially 2OL/min, corresponding to 1 volume per volume per minute (VVM) was increased to 5OL/min (2.5 VVM) when the fermenter stirrer speed approached 8O-9O% of its maximum. Since the oxygen transfer rate (OTR) of the fermenters was unable to meet the oxygen uptake rate (OUR) of the bacteria at a cell density greater than that corresponding to an OD_55O of 5O under the conditions described, dOT in the fermenter at cell densities greater than this was maintained at 5O% air-saturation by restricting bacteria oxygen uptake rate. This was achieved by formulating the medium to become carbon-limited at OD_55O of 5O and then supplying a feed of the limiting carbon source, together with ammonium sulphate and yeast extract, at a rate which restricted bacterial growth rate.

Fermentations were performed for 18h and during that time samples were taken for measurement of optical density (OD_55O), cell dry weight and accumulation of [Arg¹¹,Ser^17,27,6O,65]human G-CSF within the cells. [Arg¹¹,Ser^17,27,6O,65]human G-CSF accumulation was measured by scanning Coomassie blue stained SDS-PAGE gels of whole cell lysates of the sampled bacteria as is well known in the art. When OD_55O reached 35 (8.5h), casein hydrolysate solution (1OOg/1 Oxzoid L41) was pumped into the fermenters at a rate of O.75g/l/h.

When OD_55O reached approximately 5O, the supply of carbon-source in the fermentation batch became exhausted leading to a rapid rise in dOT from 5O% air saturation. At this point, a feed containing glycerol (47Og/1), yeast extract (118g/1) and ammonium sulphate (118g/1) was pumped into the fermenters at a rate which returned and then maintained the dOT at 5O% air saturation with the fermenter stirrer at <u style="single">ca</u> 7O-8O% of its maximum. Casein hydrolysate feeding was maintained at O.75g/l/h throughout. After approximately 18 hours, when microscopic examination of the culture showed the presence of large inclusion bodies within a majority of the cells, bacteria were harvested on a Sorval RC3B centrifuge (7OOOg, 3O min., 4°C) and stored frozen at minus 8O°C.

c) Purification

Purification was effected as described in Example 1(f)

Example 13

Preparation of [Arg¹¹,Ser

17,27,6O,65

]human G-CSF using production vector including T7A3 promoter

a) An EcoRI-SalI fragment, containing a T7A3 promoter, a trp leader ribosome binding site sequence and a gene for [Ser^17,27]hu G-CSF was sub-cloned into M13 mp18 as described in part d) of Example 1. The sequence of the EcoRI-SalI fragment is set out in SEQ ID No 5O and Figure 3, SEQ ID No 5O consists of the EcoRI restriction site (nucleotides 1-6), the A3 promoter sequence of bacteriophage T7 (nucleotide 7-52), the trp leader ribosome binding site sequence (nucleotides 53-78)and translation initiation codon (nucleotides 79-81). Figure 3 sets out the nucleotide sequence of [Ser^17,27]human G-CSF terminating in the SalI restriction site. It will be appreciated that the 3′ terminal ATG codon of SEQ ID No 5O immediately precedes the ACT codon which codes for threonine (amino acid 1) in Figure 3. The 5′ nucleotide sequence AATTCAGT is thus absent from the EcoRI-SalI fragment. The EcoRI-SalI fragment may also be prepared by excision from pICI 1295 (see Reference Example 7). Site-directed mutagenesis was performed on single-stranded DNA as described in Reference Example 2 using oligonucleotide SEQ ID No 28 to convert the codon for Gln at position 11 to Arg. Double-stranded RF DNA was prepared from a plaque containing the Gln¹¹→Arg¹¹ change as described in Example 3, except that at step B3 incubation was for 3 hours instead of 5 hours, and digested with EcoRI (as described previously) and SnaBI (as described in Reference Example 5). The resulting 144 bp EcoRI-SnaBI fragment containing the T7A3 promoter, trp leader ribosome binding site sequence and gene fragment with Arg¹¹ codon was isolated and ligated to an EcoRI-SnaBI cut vector from pICI 1327 (which contains codons for Ser^6O and Ser⁶⁵ and is described in Example 12). The ligation mix was used to transform E.coli strain MSD522 and transformants selected for growth on L-agar plates containing tetracycline (15µg/mg). Plasmid DNA from a colony containing the expected T7A3 promoter and [Arg¹¹, Ser^17,27,6O,65] hu G-CSF gene sequence were identified by sequencing DNA from the isolated plasmid and designated pICI 1386.

The fermentation was effected according to two alternative processes (b) and (c) below. Process (b) was effected at 37°C and after 16 hours fermentation as described, microbial biomass was 35 g/l and [Arg¹¹,Ser^17,27,6O,65]human G-CSF was estimated to be accumulated to 7g/l fermentation broth. Process (c) was effected at 3O°C and the fermentation was accordingly slower because of the lower fermentation temperature. With regard to process(c), after 35 hours, the microbial biomass was 55 g/l and the [Arg¹¹,Ser^17,27,6O,65]human G-CSF yield was estimated to be accumulated to 15 g/l fermentation broth. (b) <u style="single">E.Coli</u> strain CGSC 63OO (genotype F⁻ , λ⁻, lac+) obtained from the <u style="single">E.coli</u> Genetic Stock Centre was transformed with plasmid pICI 1386. The resultant strain CGSC 63OO (pICI 1386) was purified and maintained in glycerol stocks at -8O°C. An aliquot of the culture was removed from stock and streaked onto agar plates of L-tetracycline to separate single colonies after overnight growth (16h) at 37°C. A single colony of CGSC 63OO (pICI 1386) was removed and resuspended in 1Oml L-tetracycline broth and 1OOµl immediately inoculated into each of twenty 25Oml Erlenmeyer flasks containing 75ml of L-tetracycline broth. After growth for 16h at 37°C on a reciprocating shaker the contents of the flasks were pooled, and used to inoculate a fermenter containing 2O litres of modified LCM5O growth medium. The composition of the growth medium is in Table 1. <tables id="tabl0007" num="0007"><img file="EP0459630A2_D0007.tif" /></tables> The fermentation was then carried out at a temperature of 37°C and at a pH, controlled by automatic addition of 6M sodium hydroxide solution, of pH 6.7. The dissolved oxygen tension (dOT) set point was 5O% air saturation and was initially controlled by automatic adjustment of the fermenter stirrer speed. Air flow to the fermenter was initially 2O L/min corresponding to 1.O volume volume per minute (VVM) and was increased to 45 L/min manually when the fermenter stirrer speed reached its maximum (1OOO rpm). The fermentation was performed for 16h and during that time samples were taken for measurement of optical density of the culture (OD_55O biomass concentration, total microbial protein concentration and accumulation of [Arg¹¹,Ser^17,27,6O,65]human G-CSF within the bacterial cells. Accummulation was measured by scanning Coomassie blue stained SDS-PAGE gels of whole cell lysates of the sampled bacteria as is well known in the art. Total microbial protein was estimated by the method of Lowry. A solution of yeast extract (225 g/L) was pumped into the fermenter 4.5h post inoculation at 1.7 g/L/h.

When the supply of carbon source (glycerol) in the growth medium became exhausted dOT increased rapidly from 5O% air saturation. At this point a feed containing glycerol (714 g/l) and ammonium sulphate (143 g/L) was pumped. Since the bacterial oxygen sulphate rate (OUR) approached the maximum oxygen transfer rate of the fermenter (OTR) just prior to the carbon source in the batch growth medium becoming exhausted, the feed was pumped into the fermenter at a rate which restricted the bacterial OUR to approximately 8O-9O% of the fermenters maximum OTR. The feed rate was adjusted manually to return and then maintain dOT at 5O% air saturation under the conditions described.

c) The fermentation process described in (b) was repeated but at a temperature of 3O°C for 35 hours. Except for the fermentation temperature of 3O°C the medium and fermentation conditions were identical to those described in (b).

d) Purification was effected as described in Example 1(f).

Example 14

Preparation of [Glu¹⁵,Ser

17,27

,Ala

26,28

,Arg

3O

]hu G-CSF

A mutagenic template, M13mp18 containing the gene for [Glu¹⁵,Ser^17,27 Ala^26,28,Lys^3O]hu G-CSF, was prepared as described in part (d) of Example 1 with plasmid pICI 1266 replacing pICI 1O8O. The procedure described in Example 3 was repeated using the above template with mutagenic oligonucleotide designated SEQ ID No.37. This serves to convert the codon for Lys at position 3O to Arg. Double stranded RF DNA was prepared from one phage containing the desired change. An EcoRI-SalI expresson cassette was isolated and cloned into pICI OO8O as described in Example 12 to give pICI 1343.

Further processing to yield the title compound was effected as described in Example 3 and purification was effected as described in Example 6.

Example 15

Preparation of [Arg

11,23

,Ser

17,27,6O,65

]hu G-CSF

A mutagenic template, M13mp18 containing the gene for [Arg¹¹,Ser^17,27,6O,65]hu G-CSF, was prepared as described in part (d) of Example 1 with plasmid pICI 1239 replacing pICI 1O8O. The procedure described in Example 3 was repeated using the above template with mutagenic oligonucleotide designated SEQ ID No 38. This serves to convert the codon for Lys at position 23 to Arg. Double-stranded RF DNA was prepared from one phage containing the desired change and the expression cassette isolated and cloned as described in Example 14 to give pICI 1388.

Further processing to yield the title compound and the purification of the title compound were effected as described in Example 1.

Example 16

Preparation of [Arg

11,34

,Ser

17,27,6O,65

]hu G-CSF

The procedure described in Example 15 was repeated with oligonucleotide designated SEQ ID No.38 replaced by SEQ ID No.39 (this serves to convert the codon for Lys at position 34 to Arg) to give pICI 1389.

Further processing to yield the title compound and the purification of the title compound were effected as described in Example 1.

Example 17

Preparation of [Arg

11,4O

,Ser

17,27,6O,65

]hu G-CSF

The procedure described in Example 15 was repeated with oligonucleotide SEQ ID No.38 replaced by SEQ ID No.4O (this serves to convert the codon for Lys at position 4O to Arg) to give pICI 139O.

Further processing to yield the title compound and the purification of the title compound were effected as described in Example 1.

Example 18

Preparation of [Ala¹,Thr³,Tyr⁴,Arg

5,11

,Ser

17,27,6O,65

]hu G-CSF

The procedure described in Example 15 was repeated with oligonucleotide SEQ ID No.38 replaced by SEQ ID No.41 (this serves to convert codons for Thr, Leu, Gly and Pro at positions 1, 3, 4 and 5 to Ala, Thr, Tyr and Arg respectively to give pICI 1391. The polypeptide of this Example illustrates that the modification of the present invention may be applied to a polypeptide known to possess G-CSF activity in order to improve the solution stability of the polypeptide. The known polypeptide is [Ala¹,Thr³,Tyr⁴,Arg⁵,Ser¹⁷]hu G-CSF which is described in European Patent Publication No. 272,7O3 of Kyowa Hakko Kogyo Co. Ltd.

Further processing to yield the title compound and the purification of the title compound were effected as described in Example 1.

Example 19

Preparation of [Arg¹¹,Ser

17,27

]hu G-CSF

The procedure described in Example 4 was repeated with oligonucleotide SEQ ID No.3O replaced by SEQ ID No.28 (this serves to convert the codon for Gln at position 11 to Arg). The expression cassette was transferred to expression plasmid pICI OO8O, instead of pICI OO2O as described in Example 14 to give pICI 14O5.

Further processing to yield the title compound and the purification of the title compound were effected as described in Example 1.

Example 2O

Preparation of [Ser

17,27,6O,65

]hu G-CSF

The procedure described in Example 19 was repeated with oligonucleotide SEQ ID No.28 replaced by SEQ ID No.29 (this serves to convert the codons for Pro at 6O and 65 to Ser) to give pICI 14OO.

Further processing to yield the title compound and the purification of the title compound were effected as described in Example 1.

Example 21

Preparation of [Arg¹¹,Ser

17,27,6O

]hu G-CSF

The procedure described in Example 6 was repeated with oligonucleotides SEQ ID No.33 and SEQ ID No.34 replaced by SEQ ID No.28 and SEQ ID No.42. These serve to convert the codons for Gln at position 11 and Pro at position 6O to Arg and Ser respectively. The expression cassette was transferred to the expression plasmid pICI OO8O instead of pICI OO2O to give pICI 14O1.

Further processing to yield the title compound and the purification of the title compound were effected as described in Example 1.

Example 22

Preparation of [Arg¹¹,Ser

17,27,65

]hu G-CSF

The procedure described in Example 3 was repeated with oligonucleotide designated SEQ ID No.29 replaced by SEQ ID No.43 (this serves to convert the codon for Pro at position 65 to Ser) to give pICI 1418.

Further processing to yield the title compound and the purification of the title compound were effected as described in Example 1.

Example 23

Preparation of [Ser

17,27,6O

]hu G-CSF

The procedure described in Example 19 was repeated with oligonucleotide designated SEQ ID No.28 replaced by SEQ ID No.42 (this serves to convert the codon for Pro at position 6O to Ser) to give pICI 14O2.

Further processing to yield the title compound and the purification of the title compound were effected as described in Example 1.

Example 24

Preparation of [Ser

17,27,65

]hu G-CSF

The procedure described in Example 4 was repeated with oligonucleotide designated SEQ ID No.3O replaced by SEQ ID No.43 (this serves to convert the codon for Pro at position 65 to Ser) to give pICI 142O.

Further processing to yield the title compound and the purification of the title compound were effected as described in Example 1.

Example 25

Preparation of [Arg¹¹ Glu

15,111

, Ser

17,27,6O,65,115,116

, Ala

26,28

Lys

3O

] hu G-CSF

Plasmid pICI 1348, described in Example 8, was digested with XbaI in buffer M and then with SalI in buffer H and the large XbaI-SalI vector fragment isolated from a O.7% agarose gel as described previously. Plasmid pICI 1243, described in Example 4, was digested with XbaI and SalI as described above and the small XbaI-SalI fragment isolated from a O.7% agarose gel and ligated to the Xba1-SalI vector fragment above. The ligation mix was used to transform E.coli strain MSD 522 and transformants selected for growth on L-agar plates containing ampicillin (5Oµg/ml). Three colonies were screened for expression of protein as described in Example 12 but replacing tetracycline by ampicillin at 5Oµg/ml. Plasmid DNA from a colony expressing the correct protein was designated pICI 1421.

Further processing to yield the title compound was effected as described in Example 3 and purification was effected as described in Example 6.

Example 26

Preparation of [Arg

11,165

,Glu¹⁵,Ser

17,27,6O,65

Glu

26,28

Lys

3O,58

] hu G-CSF

A mutagenic template, M13mp18 containing the gene for [Arg¹¹,Glu¹⁵, Ser^17,27,6O,65,Ala^26,28,Lys^3O]hu G-CSF, was prepared as described in part (d) of Example 1 with plasmid pICI 1348 (described in Example 8) replacing pICI 1O8O. The procedure described in Example 3 was repeated using the above template with mutagenic oligonucleotides designated SEQ ID No.28 and SEQ ID No.29 replaced by SEQ ID No.44 and SEQ ID No.32 (these serve to convert the codons for Trp at position 53 to Lys and Tyr at position 165 to Arg) to give pICI 1422.

Further processing to yield the title compound was effected as described in Example 3 and purification was effected as described in Example 6.

Example 27

Preparation of [Arg¹¹,Glu¹⁵,Ser

17,27,6O,65

,Ala

26,28,44,51,55

, Lys

3O,49,58

]hu G-CSF

A mutagenic template was prepared as described in Example 26. The procedure described in Example 4 was repeated using the above template with mutagenic oligonucleotide designated SEQ ID No.3O replaced by SEQ ID No.45 (this serves to convert the codons for Pro at position 44, Leu at position 49 and Gly at positions 51 and 55 to Ala, Lys, Ala and Ala respectively) to give pICI 1423.

Further processing to yield the title compound was effected as described in Example 3 and purification was effected as described in Example 6.

Example 28

Preparation of [Arg

11,165

Glu

15,111

Ser

17,27,6O,65,115,116

, Ala

26,28,44,51,55

,Lys

3O,49,58

]hu G-CSF

A mutagenic template was prepared as described in part (d) of Example 1 with pICI 1O8O replaced by pICI 1423, described in Example 27. The procedure described in Example 3 was repeated using the above template and oligonucleotides designated SEQ ID No.28 and SEQ ID No.29 replaced by SEQ ID No.32 and SEQ ID No.3O to give pICI 1424.

Further processing to yield the title compound was effected as described in Example 3 and purification was effected as described in Example 6.

Reference Example 1

Preparation of human G-CSF

a)

Preparation of a synthetic gene for human G-CSF

A DNA sequence (Figure 2) encoding the amino-acid sequence of the polypeptide of Figure 2 (human G-CSF) was designed according to the following considerations: <ul id="ul0011" list-style="none"><li>1) Single - stranded cohesive termini to allow ligation at suitable sites in a plasmid.</li><li>2) A series of restriction endonuclease sequences throughout the gene to facilitate subsequent genetic manipulation.</li><li>3) Translation termination codon.</li><li>4) Codons at the 5′-end of the coding region were normally chosen to be A/T rich. Other codons were normally chosen as those preferred for expression in E.coli.</li></ul>

The gene was assembled from the 18 oligonucleotides designated SEQ ID No.1 - SEQ ID No.18 and shown hereinafter.

Preparation of Oligonucleotides

The oligonucleotide sequences shown hereinafter were prepared on an Applied Biosystems 38OA DNA synthesiser from 5′-dimethoxytrityl base-protected nucleoside-2-cyanoethyl-N,N-diisopropylphosphoramidites and protected nucleosides linked to controlled-pore glass supports on a O.2 micro mol scale, according to protocols supplied by Applied Biosystems Inc.

Alternatively, the oligonucleotide sequences may be prepared by manual methods as described by Atkinson and Smith in 'Oligonucleotide Synthesis, a Practical Approach' (M. T. Gait, Editor, IRL Press, Oxford, Washington DC, pages 35-81).

In detail, the preparation of the oligonucleotide sequences by use of the Applied Biosystems 38OA DNA synthesiser was effected as follows:-

Each oligonucleotide, after cleavage from the solid support and removal of all protecting groups, was dissolved in water (1ml). A solution of 3M sodium acetate (pH5.6; 4Oµl) and ethanol (1ml) was added to the oligonucleotide solutions (4OOµl) and the mixtures stored at -7O°C for 2O hours. The resulting precipitates were collected by centrifugation (13,OOOrpm for 1O minutes) and the pellets washed with ethanol:water (7:3) (2OOµl) then dried briefly <u style="single">in vacuo</u> and dissolved in water (15µl) and 1Oµl of a formamide/dye mix. (1OmM NaOH, O.5mM EDTA, O.O1% Bromophenol Blue, O.O1% xylene cyanol, 8O% formamide.

The oligonucleotides were purified on a 1O% polyacrylamide gel in 5OmM Tris-borate (pH8.3) containing 8.3M urea. Oligonucleotides of correct length were identified by UV shadowing (Narang et al, 1979 in Methods in Enzymology Vol 68, 9O-98) - normally the most prominent band - excised from the gel and electroeluted in 5mM tris-borate (pH 8.3) at 3OOmV for 3-4 hours. The aqueous solutions were concentrated to about 2OOµl by treatment with n-butanol (mix, spin and removal of the upper organic layer). The purified oligonucleotides were precipitated at -7O°C for 2O hours from a O.3M sodium acetate solution by addition of ethanol (2.5 volumes).

Assembly of gene

Oligonucleotides SEQ ID No2 - SEQ ID No 17 (4OOpM of each) [as defined hereinafter] were phosphorylated with T4 polynucleotide kinase (3.6 units) for 2 hours at 37°C in 25µl of a solution containing ATP (8OOpM containing 25pM gamma- ³²P ATP), 1OOµM spermidine, 2OmM MgCl₂, 5OmM Tris-HCl (pH9.O) and O.1mM EDTA. The solutions were heated at 1OO°C for 5 minutes to terminate the reactions, then mixed in pairs as shown in Table 1 to give duplexes A to I (Oligonucleotides SEQ ID No 1 and SEQ ID No 18 (4OOmM in 25µl) were used unphosphorylated). O.3M Sodium acetate (pH5.6, 2OOµl) and ethanol (85Oµl) were added and the duplexes precipitated at -2O°C for 2O hours. The resulting precipitates were collected by centrifugation and washed with ethanol:water (7:3) then dissolved in water (5Oµl). The pairs of oligonucleotides were annealed together by first heating the solutions to 1OO°C for 2 minutes in a boiling water bath. The bath was then allowed to cool slowly to 4O°C (about 4 hours). Solutions containing 3 pairs of duplexes were combined as shown (see Table 1), to give groups I to III lyophilised and dissolved in 3Oµl of a solution containing T4 DNA ligase (1 unit; BRL), 5OmM Tris (pH7.6), 1OmM magnesium chloride, 5% (w/v) PEG 8OOO, 1mm ATP, 1mm DTT. (BRL, Focus Vol 8 no 1 Winter 1986) and the DNA ligated at 3O°C for 5 minutes followed by 2O hours at 16°C. 3M Sodium acetate (2Oµl) and water (15Oµl) was added and the product precipitated by addition of ethanol (75Oµl) and cooling to -2O°C for 2O hours. The precipitate was collected by centrifugation and washed with ethanol (1ml) then dissolved in water (15µl) and formamide/dye mix (1Oµl) and purified on a 1O% polyacrylamide gel in 5OmM Tris-borate (pH8.3), 1mM EDTA and 8.3M urea. Bands for strands of appropriate lengths (173-186 bases) were identified by autoradiography and isolated together by electroelution from a single gel slice as described above for individual oligonucleotide sequences. The DNA strands were annealed by first heating an aqueous solution (5Oµl) at 1OO°C for 2 minutes, then allowing it to cool to 4O°C over 4 hours.

Groups I, II and III were ligated together essentially as described for the group preparation to give as the product, the gene sequence shown in Figure 2. After precipitation, the gene was phosphorylated with T4 polynucleotide kinase as described previously for individual oligonucleotides, then dissolved in water (2Oµl). <tables id="tabl0008" num="0008"><img file="EP0459630A2_D0008.tif" /></tables>

b) Cloning of the synthetic gene for human G-CSF

The synthetic gene described above, was cloned into the plasmid vector, pSTP1 (Windass et al, Nucleic Acids Research (1983) Vol 1O, p6639.

For vector preparation, 1Oµg of STP1 was dissolved in water (37.5µl) and 1O x B restriction buffer (4.5µl) (BCL). the restriction endonuclease SalI (3µl) (BCL, 8 units/µl) was added and the mixture incubated at 37°C for 1 hour until linearised plasmid was predominant over supercoiled and nicked circular forms. The DNA was precipitated with ethanol at 4°C for 3O minutes, washed with ethanol:water (7:3) then dissolved in water (39.5µl), 1OX H buffer (4.5µl) (BCL). The restriction endonuclease EcoRI (1µl) (BCL, 9O units/µl) was added and the mixture incubated at 37°C for 1 hour until the large EcoRI-SalI fragment was predominant. The DNA was precipitated at -2O°C for 2O hours, washed with ethanol:water (7:3) then dissolved in water (2Oµl)

The large EcoRI - SalI fragment was purified on a 1% preparative agarose gel and electroeluted and precipitated as described previously, then dissolved in water (2Oµl). For ligation of the synthetic gene, a mixture of vector DNA (2µl of the EcoRI - SalI fragment solution), synthetic gene (5µl of the aqueous solution described previously, 5X ligase buffer (6µl -25OmM Tris pH7.6 5OmM MgCl₂, 25% W/V PEG8OOO, 5MM ATP, 5mM DTT exBRL) water (15µl) and T4 DNA ligase (2µl, lU/µl) was incubated at 16°C for 4 hours. The DNA mix was used directly (either 1µl of neat ligation mix or 2µl of ligation mix diluted 5X with water) to transform E. coli strain HB1O1. The DNA mixture (1 or 2µl) was added to competent E. coli HB1O1 cells (2Oµl, BRL) on ice and the mixture incubated on ice for 45 min then heat shocked at 42°C for 45 seconds. After 2 min on ice, 1OOµl of SOC buffer (Bactotryptone 2%; Yeast Extract O.5%; NaCl 1OmM; KCl 2.5mm; MgCl₂, MgSO₄ 2Omm (1Omm each); glucose 2Omm) was added and the mixture incubated at 37°c for 1 hour. aliquots of suspensions were plated onto L plates with 5Oµl/ml ampicillin. transformants were screened for the presence of cloned synthetic gene by colony hybridisation analysis using standard methods described in "Molecular Cloning: A Laboratory Manual" by Maniatis <u style="single">et</u><u style="single">al</u> (Cold Spring Harbor) and in UK Patent Application No 85O26O5. A total of 1OO colonies were streaked onto filters (Schleicher and Schuell), grown at 37°C for 2O hours, lysed and baked. The filter was hybridised at 65°C for 2O hours with a radioactive probe prepared from oligonucleotide sequence SEQ ID No 1 by use of a random-label kit (Pharmacia). Five colonies 1-5 giving a positive hybridisation signal were grown up in L broth at 37°C for 2O hours on a small scale (1OOml) and plasmid DNA prepared by centrifugation in a caesium chloride gradient essentially as described in "Molecular Cloning; A Laboratory Manual" by Maniatas <u style="single">et</u><u style="single">al</u> (Cold Spring Harbor).

The DNA was sequenced by the standard dideoxy chain-termination method as described by Sanger et al in Proc. Nat. Acad Sci. USA <u style="single">74</u>, 5463-5467 (1977) using a Sequenase (Trade Mark) kit (United States Biochemical Corporation). Oligonucleotides SEQ 1D No 19 to SEQ 1D No 23 (as defined hereinafter and see Table 2) were used as sequencing primers. <tables id="tabl0009" num="0009"><img file="EP0459630A2_D0009.tif" /></tables>

The plasmid DNA from clone 5 contained the DNA sequence shown in Figure 2. The plasmid (pAG88) was used to transform competent cells of the following E.coli strains by standard procedures:-

HB1O1

CGSC 63OO (hereinafter also referred to as MSD 522)

The E. coli strains HB1O1 and MSD522 (CGSC 63OO) are freely available. Thus for example they may be obtained from the E. coli Genetic Stock Centre, Yale University, USA. Moreover E. coli HB1O1 may additionally be obtained from for example BRL supplied by GIBCO Limited Unit 4, Cowley Mill Trading Estate, Longbridge Way, Uxbridge, UB8 2YG, Middlesex, England or GIBCO Laboratories, Life Technologies Inc., 3175 Staley Road, Grand Island, NY 14O72, USA. The genotype of strain HB1O1 is described in the aforementioned "Molecular Cloning - A Laboratory Manual" as Sup E44 hsd S2O (r_B⁻ m_B⁻ )rec A 13 ara-14 F⁻leu 6 thi-1 proA2 lac Y1 gal K2 rps L2O xyl⁻5 mtl⁻1. The genotype of MSD 522 (CGSC 63OO) is set out in Example 13.

c)

Cloning of the gene for human G-CSF into an expression vector

The gene described above was cloned in the plasmid pICI OO2O as described in Example 1(c) to yield the expression plasmid pICI 1O56.

d)

Fermentation

The plasmid pICI 1O56 was transformed and fermentation effected as described in Example 1(e) to achieve expression of human G-CSF.

e)

Purification

Purification was effected as described in the second purification procedure developed to yield larger quantities of hu G-CSF set out on pages 48 and 49 of PCT Patent Publication No. WO 87/O1132 with final dialysis being effected against phosphate buffered saline.

Reference Example 2

Preparation of genes for derivatives of human G-CSF by site-directed mutagenesis

The phosphorothioate method of Eckstein and co-workers was used:    Taylor, J W et al Nucleic Acids Research (1985) Vol pp 8749-8764    Taylor, J W et al Nucleic Acids Research (1985) Vol pp 8765-8785    Nakamaye, K et al Nucleic Acids Research (1986) Vol pp 9679-9698    Sayers, J R et al Nucleic Acids Research (1988) Vol pp 791-8O2 The procedure can be carried out using a kit supplied by Amersham International. The method is outlined below and incorporates changes to the original method with regard to the use of more than one mutagenic oligonucleotide and the incubation temperature for oligonucleotides of greater than 3O bases in length.

1. Annealing mutant oligonucleotide to single stranded DNA template: <tables id="tabl0010" num="0010"><img file="EP0459630A2_D0010.tif" /></tables> (Where two mutagenic oligonucleotides were used simultaneously, 2.5µl (1.6pmole/1µl) of each phosporylated oligonucleotide was added to 5µl single stranded DNA template (1µg/µl) in 3.5µl Buffer 1 and 3.5µl water. Where 3 mutagenic oligonucleotides were used 2.5µl (1.6pmol/µl) of each phosporylated oligonucleotide was added to 5µl single stranded DNA (1µg/µl in 3.5µl Buffer 1 and 1µl water). The above ingredients were placed in a capped tube in a 7O°C water bath for 3 minutes if the oligonucleotide was &lt;3Obases in length or in a boiling water bath for 3 minutes if the oligonucleotide was &gt; 3O bases in length. The tube was then placed in a 37°C water bath for 3O minutes.

2. Synthesis and ligation of mutant DNA strand:

To the annealing reaction were added <dl id="dl0003"><dt>MgCl₂ solution</dt><dd>5µl</dd><dt>Nucleotide mix 1</dt><dd>19µl</dd><dt>(contains dCTP alpha S)</dt><dd> </dd><dt>water</dt><dd>6µl</dd><dt>Klenow fragment (6 units)</dt><dd>1.5µl</dd><dt>T4 DNA ligase (5 units)</dt><dd>2µl</dd></dl> The above ingredients were placed in a 16°C water-bath and left overnight.

3. Removal of single stranded (non-mutant) DNA using disposable centrifugal filter units.

To the reaction from Step 2 the following ingredients were added:- <tables id="tabl0011" num="0011"><img file="EP0459630A2_D0011.tif" /></tables>

The 25Oµl sample was added to the top half of the filter unit and centrifuged at 15OO rpm for 1O minutes at room temperature in a SORVALL RT6OOOB bench top centrifuge using a SORVALL H1OOOB swing out rotor. Sample passes through two nitrocellulose membranes which bind the single stranded DNA leaving the double stranded DNA to pass through to the collection tube below. 1OOµl of 5OO mM NaCl were added and respun for 1O minutes to wash through any remaining RF DNA.

The following ingredients were added to the filtrate:- <dl id="dl0004"><dt>3M Sodium Acetate (pH6.O)</dt><dd>28µl</dd><dt>Cold Ethanol (-2O°C)</dt><dd>7OOµl</dd></dl>

The mixture was placed in a dry ice and ethanol bath for 2O minutes and centrifuged in an Eppendorf microfuge for 15 minutes. The pellet was then resuspended in 1Oµl buffer 2.

4. Nicking of the non-mutant strand using Nci I. To the reaction mix from step 3, was added 65µl Buffer 3 and 8 units Nci I (1µl). The mixture was placed in a 37°C water bath for 9O minutes.

5. Digestion of non-mutant strand using exonuclease III

To the reaction mix from step 4 was added <dl id="dl0005"><dt>5OO mM NaCl</dt><dd>12µl</dd><dt>Buffer 4</dt><dd>1Oµl</dd><dt>Exonuclease III (5Ounits)</dt><dd>2µl</dd></dl>

The mixture was placed in a 37°C water bath and incubated for 3O minutes at 37°C, 5O units of exonuclease III will digest approximately 3,OOO bases in 3O minutes). The mixture was then placed in a 7O°C water bath for 15 minutes to inactivate the enzymes. 6. Repolymerisation and ligation of the gapped DNA.

To the reaction mix from step 5 was added <dl id="dl0006"><dt>nucleotide mix 2</dt><dd>13µl</dd><dt>MgCl₂ solution</dt><dd>5µl</dd><dt>DNA polymerase 1 (4 units)</dt><dd>1µl</dd><dt>T4 DNA ligase (2.5 units)</dt><dd>1µl</dd></dl>

The mixture was placed in a 16°C bath for 3 hours.

7. Transformation of competent host <u style="single">E.</u><u style="single">coli</u> TG1 cells with the DNA: 3OOµl of freshly prepared competent <u style="single">E.</u><u style="single">coli</u> TG1 cells (prepared following the method of Mandel and Higa) were transformed with 2Oµl of the reaction mix from step 6 (in duplicate). The transformants were plated out in a lawn of log phase TG1 cells in TY Top agar on TY plates and incubated overnight at 37°C.

The E. coli strain TG1 is freely available from for example the E. coli Genetic Stock Centre, Yale University, USA and from Amersham International plc, Amersham Place, Little Chalfont, Amersham, Buckinghamshire HP7 9NA, England as supplied in their "<u style="single">in</u><u style="single">vitro</u>" mutagenesis system, oligonucleotide directed kit (Product code RPN 1523).

Reference Example 3

G-CSF Bioassay

A factor dependent cell line, Paterson - G-CSF (FDCP-G), obtained from the Paterson Institute, Manchester, England was cloned by limiting dilution in the presence of G-CSF. A G-CSF responsive clone, designated clone E7, was used to determine human recombinant G-CSF activity. 2.5 x 1O³ FDCP-G clone E7 cells in 1OOµl of RPMI 164O + 1O% FCS was added to an equal volume of RPMI 164O + 1O% FCS containing G-CSF. Each G-CSF sample was measured over 1O doubling dilutions. The final volume of RPMI 164O (see Moore GE et al (1967) JAMA, <u style="single">199</u>, 519) + 1O% FCS (foetal calf serum) in each well of 96-well microtitre plate was 2OOµl. The microtitre plate was incubated at 37°C in 5% CO₂ in a humidified incubator for 4 days. 1.OµCi of titrated thymidine was added per well and incubated over the final 6 hours. Cells were harvested onto glass fibre filter papers and the level of radioactivity determined by liquid scintillation counting. The level of tritiated thymidine incorporation was found to be directly proportional to the amount of G-CSF present. The FDCP-G clone E7 assay was calibrated using recombinant human G-CSF obtained from Amersham International with a declared specific activity of 1O⁸ units/mg of protein.

The potencies of G-CSF samples were determined by comparision to a standard of known activity.

The units of G-CSF activity per ml were calculated according to the following formula:- <tables id="tabl0012" num="0012"><img file="EP0459630A2_D0012.tif" /></tables>

Reference Example 4

Solution Stability of G-CSF and derivatives thereof

Appropriate dilutions of the stock solution of G-CSF and derivatives in phosphate buffered saline (PBS) at 4°C described in Example 9 were tested for solution stability. Solutions of 1mg/ml, 5mg/ml and sometimes 1Omg/ml of protein in PBS were incubated at 37°C for 14 days. Solutions were inspected visually at regular intervals for signs of precipitation. After 14 days each solution was centrifuged at 14,OOOrpm for 2O minutes, the supernatant removed by decantation and the pellet re-dissolved in PBS containing 1% w/v N-lauroyl sarcosine. The total protein content in each supernatant and re-dissolved precipitate was estimated by A_28O measurements and the monomer content in each was estimated by reverse phase HPLC. These were expressed as a percentage of the corresponding data given by solutions at the start of incubation and by a 1mg/ml solution incubated at 4°C for 14 days. Variations between total protein and monomer estimates were observed only in some of the re-dissolved pellets. The percentage protein remaining in solution in the supernatants from each starting concentration is summarised in the Table.

The specific activity of the product in each supernatant after incubation was shown to be the same as in the starting solution, and no differences were observed on PAGE-SDS under reducing or non-reducing conditions.

The following results were obtained: <tables id="tabl0013" num="0013"><img file="EP0459630A2_D0013.tif" /></tables><tables id="tabl0014" num="0014"><img file="EP0459630A2_D0014.tif" /></tables> nd means not done

[Met⁻¹,Ser¹⁷]hu G-CSF may be obtained as described in Reference Example 5.

The above results demonstrate that modifications of the present invention improve solution stability without loss of G-CSF activity, [Met⁻¹,Ser¹⁷]hu G-CSF in a concentration of 5mg/ml starting to precipitate out within 3 hours.

Reference Example 5

Preparation of [Ser¹⁷] hu G-CSF

The procedure described in Example 2 for the preparation of [Met⁻¹, Ser^17,27] hu G-CSF was repeated except as follows:- <ul id="ul0012" list-style="none"><li>1) The duplex for phosphorylation was prepared from oligonucleotide sequences SEQ ID Nos 24, 25, 3 and 4, the sequences SEQ ID Nos 3 and 4 respectively replacing sequences SEQ ID Nos 26 and 27 employed in Examples 1 and 2.</li><li>2) The duplex referred to in (1) was phosphorylated with T4 polynucleotide kinase, but was digested with SnaBI (1O units) in 1 x M buffer (BCL; 3Oµl) for 2 hours at 37°C.</li><li>3) Following purification with ethanol, the 72bp EcoRI-SnaBI fragment was purified as opposed to the 143 bp EcoRI-MstII fragment.</li><li>4) The synthetic EcoRI-SnaBI fragment was cloned into the plasmid vector pAG88 as described in Reference Example 1 and for vector preparation pAG88 was digested with SnaBI (2O units; BCL) in 1 x M buffer (BCL; 1OO µl) for 2 hours at 37°C instead of Mst II in 1 x H buffer.</li><li>5) Following precipitation with ethanol, the large EcoRI-SnaBI fragment was purified on a 1% agarose gel as opposed to the large EcoRI-MstII fragment.</li><li>6) The plasmid containing the gene for [Ser¹⁷] hu G-CSF was designated pICI 11O5.</li></ul>

Reference Example 6

Construction of pICI OO8O

a) Construction of pTB357 (also referred to herein as pLB OO4

Plasmid pTB357 utilises a repressed tetracycline resistance determinant, as found on the naturally-occurring plasmid RP4. This repressed system shuts off expression of the <u style="single">tetA</u> gene in the absence of tetracycline whereas most drug resistant mechanisms have constitutive expression.

The <u style="single">tet</u> locus was first mapped on RP4 by Barth and Grinter (<u style="single">J.Mol. Biol.113</u>: 455-474, 1977). This was shown to consist of adjacent genes: <u style="single">tetA</u>, the structural resistance gene and <u style="single">tetR</u>, the repressor gene and this region has been sequenced (Klock <u style="single">et al</u>, <u style="single">J. Bacteriol</u>: 161:326-332, 1985). These genes are located on adjacent <u style="single">Bg1II-SmaI</u> and <u style="single">SmaI-SmaI</u> fragments. The <u style="single">Bg1II</u> site is unique in RP4 but there are five <u style="single">SmaI</u> sites (Lanka, Lurz and Furste, <u style="single">Plasmid 1O</u>: 3O3-3O7, 1983).

i) Cloning the tetA + tetR genes

The plasmid RP4 is well documented (Datta <u style="single">et al</u>, <u style="single">J. Bacteriol</u> 1O8: 1244, 1971) and is freely available. Furthermore the plasmid RP4 has been deposited with the National Collection of Type Cultures, 61 Colindale Avenue, London, NW9 5HT under accession nos. 5OO78 and 5O437. <u style="single">E. coli</u> strains containing this plasmid were grown in selective broth cultures and plasmid DNA was isolated a scale-up of the Holmes and Quigley method (Holmes and Quigley, <u style="single">Anal. Biochem</u><u style="single">114</u>: 193-197, 1981). It was deproteinized by treatment with 2.5M ammonium acetate and reprecipitated with isopropanol. This plasmid DNA was treated, according to the supplier's recommended conditions, with restriction enzyme <u style="single">Bg1II</u> and cut to completion. It was then partially cut by <u style="single">XmaI</u> by using diluted enzyme and short incubation times. <u style="single">XmaI</u> is an isoschizomer of <u style="single">SmaI</u> but which produces 4-nucleotide cohesive ends at its cut sites.

The vector plasmid pUC8 (Yanisch-Perron, Vieira and Messing, <u style="single">Gene</u><u style="single">33</u>: 1O3-119, 1985) was similarly prepared and cut with <u style="single">BamHI</u> and <u style="single">XmaI</u> to completion. The RP4 fragments were cloned into this vector by ligation with T4 ligase at 12°C for 16 hours. This was used to transform <u style="single">E.</u><u style="single">coli</u> C6OO made competent by the calcium chloride method (Maniatis <u style="single">et al</u>, Cold Spring Harbor Laboratory, 1982). Cultures were then plated onto medium which selected for tetracycline resistance.

E. coli C6OO is freely available from numerous sources including many culture collections such as the E.coli Genetic Stock Centre, Yale University, USA under accession No GCSC 3OO4. The genotype of E.coli C6OO is K12 thr-1 leuB6 thi-1 hsdS1 lacY1 tonA21 λ⁻ supE44.

Several colonies with this resistance were checked for the expected phenotype (ampicillin and tetracycline resistance but not the kanamycin resistance indicative of RP4 itself). Colonies with the correct resistances were subjected to clone analysis by isolating plasmid DNA (Holmes and Quigley method). These preparations were cut with <u style="single">EcoRI</u> and <u style="single">HindIII</u> and analysed by gel electrophoresis. This established the size of the cloned insert which was found to be the 2.45 kb predicted for the <u style="single">Bg1II</u> - <u style="single">XmaI</u> - <u style="single">XmaI</u> fragment from RP4. A clone carrying this fragment containing the <u style="single">tetA</u> and <u style="single">tetR</u> genes was designated pTB344.

ii) Removal of the tet gene from pAT153

It was necessary to remove the <u style="single">tet</u> gene from the vector plasmid pAT153 before inserting the <u style="single">tetA</u> + <u style="single">tetR</u> cassette from RP4 to prevent gene duplication which can be a source of genetic instability. Also the <u style="single">tet</u> gene may not be effectively suppressed byt he non-cognate <u style="single">tetR</u>. The removal was done by isolating plasmid pAT153 DNA and cutting it with <u style="single">EcoRI</u> and <u style="single">AvaI</u>. Between these sites, synthetic olignucleotides with the sequence SEQ ID No.59:- <chemistry id="chem0001" num="0001"><img file="EP0459630A2_D0015.tif" /></chemistry> were clonded. These fit the <u style="single">EcoRI</u> and <u style="single">AvaI</u> cohesive ends and contain <u style="single">SphI</u><u style="single">BamHI</u> and <u style="single">ClaI</u> sites in addition. After transformation and selected, colonies were tested for the loss of the teracycline resistance determinant. Plasmid DNA from one clone was sequenced to confirm that the predicted sequence was correct. This plasmid was designated pCH19.

iii) Insertion of the tetA + tetR genes

The <u style="single">tetA</u> and <u style="single">tetR</u> genes were isolated from pTB344 on an <u style="single">EcoRI</u> to <u style="single">PstI</u> fragment. The pUC8 vector was destroyed by curring with <u style="single">SspI</u> because it carries the same selection determinant (ampicillin resistance) as pCH19. Plasmid pCH19 DNA was cut with <u style="single">EcoRI</u> and <u style="single">PstI</u> and then ligated with the 2.45 kb fragment carrying the <u style="single">tet</u> genes. This was used to transform <u style="single">E.coli</u> C6OO, the culture being plated out under selection for tetracycline reistant colonies. The insertion of the <u style="single">tet</u> genes was designed to replace most of the <u style="single">bla</u> genes in pCH19 which should thus lose its ampicillin resistance determinant. Loss of ampicillin resistance from the transformants was confirmed. A few clones were then used to isolate plasmid DNA which was subjected to restriction analysis. This confirmed that the constructed plasmid had the intended structure. It was designated pTB351.

iv) Insertion of the cer sequence

The naturally-occuring plasmid ColEI is very stably maintained in <u style="single">E.coli</u>, whereas its derivatives pBR322 and pAT153 are not. Summers and Sherratt (<u style="single">Cell, 36</u>: 1O97-11O3, 1984) demonstrated that this was due to the derivatives not containing a short (283 bp) sequence called <u style="single">cer</u> which is present in the parent plasmid. This sequence contains a site-specific plasmid multimer-resolution system which prevents the accumulation of plasmid multimers formed by homologous recombination. Such multimers have a deleterious effect on the process of partition which normally ensures stable inheritance of daughter plasmids during bacterial cell division.

The <u style="single">cer</u> sequence (Summers, D et al MGG, <u style="single">2O1</u>, p334-338, 1985) was isolated from plasmid pKS492 (provided by D. Sherratt) as a 289 bp fragment by cutting with <u style="single">BamHI</u> and <u style="single">TaqI</u>. The plasmid pTB351 was isolated as DNA from a <u style="single">dam</u> strain of <u style="single">E. coli</u> to prevent its <u style="single">ClaI</u> site being blocked by the <u style="single">dam</u>+ methylation system. This DNA was cut with <u style="single">BamHI</u> and <u style="single">ClaI</u> (both these sites having been introduced on the synthetic oligonucleotide for this cloning). The <u style="single">cer</u> fragment was ligated with the cut vector and then used to transform <u style="single">E. coli</u> C6OO, selection being made for tetracycline reisistance. Transformant colonies were subjected to clone analysis by <u style="single">AvaI</u> restriction and gel electrophoresis. The presence of an extra DNA band of about 3OO bp indicated the acquisition of the <u style="single">cer</u> fragment. Further restriction analyses were used to confirm that resultant plasmids had the correct structure. One of these was designated pTB357 (Figure 5) and also designated pLBOO4.

b) Plasmid pCH1O1

The plasmid pCH1O1 corresponds to pICI OO2O (see Example 1c) except that the EcoRI-SalI fragment (see Figure 1) is replaced by a fragment consisting of the SEQ ID No 53 (see Figure 6 also) and the interferon α₂ gene sequence as described by Edge M.D. <u style="single">et</u><u style="single">al</u>, Nucleic Acids Research 1983, Vol11, p6419-6435. In this regard the 3′-terminal ATG codon of SEQ ID No 53 immediately precedes the TGT codon which codes for cysteine (amino acid 1) in the interferon α₂ sequence of the above-mentioned Edge M.D. <u style="single">et</u><u style="single">al</u> Nucleic Acids Research reference. The 5′ nucleotide sequence GATCCATG and the complementary 3′ nucleotide sequence GTAC are thus omitted from the nucleotide sequence of the aforementioned reference.

c) Insertion of an Expression Cassette into pTB357

An expression cassette consisting of the <u style="single">trp</u> promoter, a ribosome binding site and the interferon α₂ gene was isolated from plasmid pCH1O1 (see b above) on an <u style="single">EcoRI</u> to <u style="single">SphI</u> restriction fragment. This was ligated into the production vector (pTB357) (see (a) above) similarly cut with <u style="single">EcoRI</u> and <u style="single">SphI</u>. This DNA was used to transform a competent culture of <u style="single">E. coli</u> C6OO and tetracycline resistant colonies were isolated. A few of these were tested by DNA clone analysis for the acquisition of the <u style="single">SstI</u> restriction site carried on the expression cassette. Clones positive in this respect were further tested by restriction mapping to check that the expected construct was correct. They were also checked for the conferred capacity to produce interferon α₂ protein as analysed on a polyacrylamide-SDS gel stained with Coomassie blue. One such confirmed clone was designated pLBOO5.

d) Insertion of T4 transcription terminator into pTB 244

The T4 transcription terminator sequence in the form of the <u style="single">SalI</u> to <u style="single">HindIII</u> fragment (67 bases pairs long) (see SEQ ID No. 51 and Figure 4a) was inserted into the multicloning site of an intermediate vector pTB 244 (described in European Patent Publication No. 237,269) between its <u style="single">SalI</u> and <u style="single">HindIII</u> sites. Clone analysis was used to confirm the structure of this construct (pTB244. T4 ter). From this vector, an <u style="single">SstI</u> to <u style="single">SphI</u> fragment containing most of the multicloning site and the T4 terminator was then isolated. This was inserted into pLBOO5 similarly cut with <u style="single">SstI</u> and <u style="single">SphI</u> thereby substituting the interferon α₂ gene but leaving a cassette consisting of the <u style="single">trp</u> promoter, multicloning site and T4 terminator. This construct was confirmed by clone analysis and the plasmid designated pLBO13.

e) Substitution of the multicloning site

The multicloning site in pLBO13 is not ideal for this vector in several respects: the <u style="single">SalI</u><u style="single">BamHI</u> and <u style="single">SmaI</u> sites are not unique but exist elsewhere on the plasmid. This fragment was therefore excised by cutting with <u style="single">SstI</u> and <u style="single">XbaI</u> (both unique) and synthetic oligonucleotides with the sequence of SEQ ID No. 54:- <chemistry id="chem0002" num="0002"><img file="EP0459630A2_D0016.tif" /></chemistry> were inserted in its place. Clones were analysed for acquisition of the new restriction sites and then confirmed by sequencing. One such plasmid was designated pLBO14. The new cloning sites inserted in this way are: <u style="single">NdeI</u>, <u style="single">KpnI</u>, <u style="single">BglII</u>, <u style="single">XhoI</u>, and <u style="single">ScaI</u> with the previous XbaI and SalI following them.

f) Further modification

It was discovered that the adjacent <u style="single">SstI</u> and <u style="single">NdeI</u> sites in pLBO14 could not be cut by both these restriction enzymes either simultaneously or sequentially presumably because of their close proximity. Anadditional sequence was therefore inserted between them. This was done by cutting pLBO14 with <u style="single">SstI</u> and <u style="single">KpnI</u> and then inserting the synthetic oligonucleotide of SEQ ID No. 55. <chemistry id="chem0003" num="0003"><img file="EP0459630A2_D0017.tif" /></chemistry> Clones were analysed for acquisition of an extra <u style="single">PvuII</u> or <u style="single">PstI</u> site and then confirmed by sequencing. One such plasmid was designated pLBO15 (=pICI OO8O) (see Figure 7). This plasmid, unlike pLBO14, is efficiently cut by <u style="single">SstI</u> and <u style="single">NdeI</u>. This is to provide a place to insert a variety of ribosome binding site sequences correctly positioned with respect to the upstream <u style="single">trp</u> promoter and with <u style="single">NdeI</u> designed to provide the ATG start codon of the gene to be expressed.

Reference Example 7

Construction of plasmid pICI 1295 (also referred to as pCG3OO

a) Production of pCG54 from pICI1O79

pICI1O79 is an ampicillin resistant, pAT153-derived plasmid containing the following elements between the EcoRI and StylI restriction sites:- <ul id="ul0013" list-style="none"><li>(i) a CI857 from phage λ;</li><li>(ii) a λP_L promoter;</li><li>(iii) a synthetic ribosome binding site;</li><li>(iv) a synthetic interferon α₂ gene sequence;</li><li>(v) a synthetic transcription terminator sequence, derived from phage T4, between the SalI and StyI restriction sites. The DNA sequence of this transcription terminator is shown in Figure 4 and SEQ ID No. 56.</li></ul>

pICI1O79 is illustrated in Figure 8.

pICI1O79 has been deposited under the Budapest Treaty, at the National Collections of Industrial and Marine Bacteria Limited (NCIMB), 23 St. Machar Drive, Aberdeen, AB2 1RY, Scotland, UK. (NCIMB No 4O37O, date of deposit 19 February 1991).

pCG54 was constructed in order to make available an expression vector containing the same promoter, ribosome binding site and transcription terminator sequences as above, ie: λp_L, RBS7 and T4, but lacking gene sequence encoding for production of a specific protein. Such a construct would provide the facility of a basic expression vector containing essential elements allowing transcription and translation for production of any protein of interest which could be introduced into this vector by subsequent cloning events.

Construction of the vector was initiated by restriction endonuclease cleavage of pICI1O79 at its respective EcoRI and SalI sites. This cleavage step released a vector fragment containing the pICI1O79 backbone complete with genes for plasmid replication and antibiotic resistance functions, plus the T4 transcription terminator sequence. The fragment was isolated by agarose gel purification steps using Geneclean for final purification of the DNA fragment.

To this vector fragment a second smaller DNA fragment of approximately 1.2Kb in size was introduced. This second fragment may be obtained, for example by DNA synthesis or by site directed or PCR mutagenesis of the small EcoRI-SalI restriction fragment obtained from pICI1O79 as described above. This second fragment contained exactly equivalent promoter and ribosome binding site sequences as originally present in pICI1O79 and additionally had EcoRI and SalI sites available at its 5′ and 3′ termini respectively, so providing compatible termini for ligation to the pICI1O79 fragment. A ligation reaction in the presence of Gibco-BRL enzyme T4 DNA ligase and its respective buffer, resulted in the formation of the construct pCG54.

Clones containing this construct were originally isolated following transformation of an aliquot of the ligation reaction mixture into <u style="single">E.coli</u> competent cells of strain HB1O1.

The construct pCG54 recovered was 3.682Kb in size and contained essential features as outlined on the map featured in Figure 9.

b) Production of pCG61 from pCG54 (also referred to as pICI54)

Synthetic oligonucleotide sequences were designed so as to include both the natural sequence for the T7A3 promoter and also a sequence which would provide an effective translation initiation region to enable correct processing of any polypeptide gene sequence cloned adjacent to it. A suitable candidate sequence for this latter region was identified as RBS1, the trp ribosome binding sequence. Therefore two complimentary oligonucleotides identified as SEQ ID No.57 and SEQ ID No.58 were synthesized to generate a double stranded DNA linker incorporating the T7A3 promoter and RBS1 sequences .

Oligonucleotides were prepared as 84mers by the standard protocol using an ABI gene synthesizer. They were designed so that in the double stranded form the synthetic fragments would have restriction endonuclease sites EcoRI and KpnI at the 5′ and 3′ ends respectively. Due to their length the oligomers could not be purified by means of HPLC and purification was undertaken by means of acrylamide gel electrophoresis using a 1O% acrylamide: 7M Urea gel.

Prior to purification, the oligomers were first checked on a sizing gel to ensure that not only are they of the correct size but that also the samples prepared contain as their greatest proportion the oligomers required and not a high contaminating proportion of smaller secondary oligonucleotides which result as by-products of synthesis.

The acrylamide gels were prepared by standard methods with ammonium persulphate and N,N,N′,N′-tetramethylethylenediamine used as catalysts for gel polymerisation.

Sizing of the oligonucleotides required that they could be visualized after electropohoresis. It was therefore necessary to radioactively label the samples using ³²P. This made it possible to assess sample quality following electrophoresis by way of autoradiography.

Oligonucleotide samples were supplied in a crude form unphosphorylated. This factor was made use of for radiolabelling purposes in that the samples could be 'hot' labelled at the 5′ termini by phosphorylation using the enzyme T4 polynucleotide kinase.

Oligomers were provided from synthesis in an unphosphorylated form and so after purification each oligomer was individually subjected to a phosphorylation reaction in which ATP was used to phosphorylate the 5′ end of each molecule in the presence of T4 polynucleotide kinase. (see Molecular Cloning: A Laboratory manual 2nd Edition, Sambrook, Fristch and Maniatis, p 5.68-5.71). Once phosphorylated the two complimentary oligonucleotides were annealed together to form the double strand DNA duplex containing the T7A3 promoter and the RBS1 sequence.

The vector molecule pCG54 was cleaved with restriction enzymes EcoRI and KpnI. On restriction digestion 2.3kb vector fragment and a 1.1kb fragment containing the λ_PL promoter and RBS1 sequence are generated. This cloning step is planned to replace the λ_PL -RBS1 sequence by EcoRI to Kpn1 synthetic fragment comprising the T7A3-RBS1 sequence. The 2.3kb vector fragment resulting from digestion of pCG54 was purified by the usual protocol using agarose gel electrophoresis and Geneclean methodology for removal of DNA from agarose fragments.

The 84bp EcoRI-KpnI synthetic fragment was ligated into the vector molecule prepared above and the ligated DNA used to transform <u style="single">E.coli</u> HB1O1 cells. Selection of positive recombinant clones was by ampicillin resistance. Following transformation a number of colonies containing recombinant plasmid were selected for screening purposes.

The synthetic fragment incorporated into the vector during cloning was of a size (84 mer) such as to make restriction analysis of recombinant plasmid DNA samples inappropriate as a simple screening method. Inserts of such a small size are not readily apparent on agarose gel electrophoresis. The fragment itself contains no internal restriction endonuclease cleavage site which could be diagnostic of its presence. Initial screening of recombinant clones was therefore by the method of colony hybridisation (see Grunstein and Hogness Proc. Natl Acad. Sci <u style="single">72</u>, 3961 (1975)). Nitrocellulose filters containing immobilized plasmid DNA from the recombinant clones were hybridised against a probe prepared by random radiolabelling of the synthetic annealed oligonucleotide SEQ ID No. 57 and SEQ ID No.58 . The DNA was labelled using α³²P-dCTP and incubation with Klenow polymerase at 37°C for 2 hours. Recombinant colonies which generated a positive hybridisation reaction were selected for plasmid DNA preparation. Plasmid DNA was prepared in each case by a relatively large scale method incorporating CsCl gradient density centrifugation to ensure purity see " Molecular Cloning - A laboratory manual "second edition, Sambrook Fritsch and Maniatis (Cold Spring Harbor Laboratory, 1989) p1.42-1.52. Preparation of DNA by such a method ensures high quality material suitable for use in subsequent cloning manipulations and sequence analysis.

All plasmid DNA isolated from recombinant clones was included in a secondary screen by sequence analysis, to ensure that the oligonucleotide sequence at the cloning junctions and of the T7A3-RBS1 fragment itself was absolutely correct. The sequencing protocol used was that of Sequenase and the sequencing primer selected for use was for example pBR322 UP (pBR322 universal primer). Sequencing was effected using the Sanger dideoxy chain termination sequencing technique. Clones having the correct sequence were designated as the new expression construct pCG61, and contained the T7A3 promoter, RBS1 sequence and the T4 terminator sequence (see Figure 1O).

c) Production of pCG3OO (also referred to as pICI 1295) from pCG61

The sequence and synthesis steps involved in construction of the G-CSF analogue [Ser^17,27]hu G-CSF are as described in Example 1 (see Figure 3). This G-CSF analogue sequence was isolated from a construct in which the gene had been incorporated into the plasmid pSTP1 to give pICI11O7 (see Example 2). pICI11O7 was digested with ScaI and the large fragment isolated following agarose gel electrophoresis and Geneclean purification. This fragment was then digested with the restriction endonuclease SalI to generate a [Ser^17,27]hu G-CSF gene on a ScaI to SalI restriction fragment suitable for cloning into pCG61 (see Figure 1O).

Following restriction with SalI the required fragment was isolated using agarose gel purification techniques once again.

The vector molecule pCG61 was digested with restriction enzyme Kpn1. Cleavage with this enzyme creates a 3′ overhang which was then blunt-ended using the enzyme T4 polymerase see "Molecular Cloning - a Laboratory manual", Second Edition Sambrook, Fritsch and Maniatis, p5.44 - 5.47. T4 polymerase activity was heat inactivated by incubation at 7O°C for 3O minutes and the DNA was recovered by ethanol precipitation. The pellet was dissolved in sterile distilled water and the solubilized DNA cleaved with SalI. The KpnI (now blunt-ended) to SalI vector fragment was recovered by means of ethanol precipitation followed by agarose gel electrophoresis and purification techniques.

The ScaI to SalI [Ser^17,27]hu G-CSF fragment was then ligated into the blunt-ended KpnI to SalI vector. Ligated DNA was transformed into <u style="single">E.coli</u> strain HB1O1. Selection of recombinant clones was for ampicillin resistance.

Initial screening of potential recombinant clones was by means of hybridisation. A radiolabelled probe was prepared by random labelling of an EcoRI to SalI fragment (containing the [Ser^17,27]hu G-CSF gene sequence) prepared from the plasmid pICI11O7. This was used in hybridisation against colonies whose DNA had been immobilized onto the surface of nitrocellulose filters. Subsequently plasmid DNA was prepared from 24 clones which had been hybridised in this screen. All DNA preparations were by the rapid mini-prep method see Birnboim and Doly, Nucleic Acids Research, <u style="single">7</u>, 1513, 1979. These recombinant DNA preparations were subjected to a secondary screen by way of restriction analysis. Linearization of the DNA with BamHI, which is a unique site within the expression cassette, is indicative of the presence of the [Ser^17,27]hu G-CSF sequence.

Sequence analysis was performed to confirm the presence of the [Ser^17,27]hu G-CSF gene and to verify that the base sequence at the cloning junctions and throughout the [Ser^17,27]hu G-CSF gene was correct. For this purpose large scale plasmid DNA samples were prepared from 16 recombinant clones using the CsCl gradient density centrifugation technique to ensure purity. Sequencing steps were performed in accordance with the sequence protocol and the sequencing primer selected was the pBR322 universal primer (EcoRI). Two of the recombinant clones contained the correct sequence at the ScaI end of the [Ser^17,27]hu G-CSF fragment and throughout the G-CSF peptide sequence itself. The clones were identified as expression construct pCG3OO (see Figure 12). <tables id="tabl0015" num="0015"><img file="EP0459630A2_D0018.tif" /></tables><tables id="tabl0016" num="0016"><img file="EP0459630A2_D0019.tif" /></tables><tables id="tabl0017" num="0017"><img file="EP0459630A2_D0020.tif" /></tables><tables id="tabl0018" num="0018"><img file="EP0459630A2_D0021.tif" /></tables><tables id="tabl0019" num="0019"><img file="EP0459630A2_D0022.tif" /></tables><tables id="tabl0020" num="0020"><img file="EP0459630A2_D0023.tif" /></tables><tables id="tabl0021" num="0021"><img file="EP0459630A2_D0024.tif" /></tables><tables id="tabl0022" num="0022"><img file="EP0459630A2_D0025.tif" /></tables><tables id="tabl0023" num="0023"><img file="EP0459630A2_D0026.tif" /></tables><tables id="tabl0024" num="0024"><img file="EP0459630A2_D0027.tif" /></tables><tables id="tabl0025" num="0025"><img file="EP0459630A2_D0028.tif" /></tables><img file="EP0459630A2_D0029.tif" /> Pro (where m is O or 1). <tables id="tabl0026" num="0026"><img file="EP0459630A2_D0030.tif" /></tables><img file="EP0459630A2_D0031.tif" /><tables id="tabl0027" num="0027"><img file="EP0459630A2_D0032.tif" /></tables><img file="EP0459630A2_D0033.tif" /><img file="EP0459630A2_D0034.tif" /><tables id="tabl0028" num="0028"><img file="EP0459630A2_D0035.tif" /></tables><img file="EP0459630A2_D0036.tif" /><img file="EP0459630A2_D0037.tif" />