AU5208400A - Screening method for peptides - Google Patents

Description

WO 00/73433 PCT/DKOO/00287 1 SCREENING METHOD FOR PEPTIDES FIELD OF THE INVENTION The present invention relates to a method for screening a pool of nucleotide sequences to select a nucleotide sequence 5 encoding a peptide. BACKGROUND OF THE INVENTION Various bioactive peptides are known to kill or inhibit the growth of target cells, e.g. antimicrobial enzymes, anti tumor peptides and antimicrobial peptides. An improved screen 10 ing method for such peptides is desirable for the development of new bioactive peptides. SUMMARY OF THE INVENTION The object of the invention is to provide a method to identify novel or improve existing genes encoding bioactive 15 peptides that can kill or inhibit the growth of target cells. The inventor has developed a suicide expression system (SES) for such peptides. The rationale of the SES is to generate li braries encoding peptides in cells, induce expression of the individual peptides, and select/identify peptide-encoding se 20 quences according to their ability to kill or inhibit the growth of host cells as a result of synthesis of the peptide. Successive rounds of peptide induction, selection, plasmid am plification and mutagenesis can be used for the identification of peptides with improved bioactivity. However no protection or 25 scaffold peptide is needed in this method to protect the active peptide from digestion within the cell. Such peptide may be needed for recovering and purifying the active peptide, but not WO 00/73433 PCT/DKOO/00287 2 to identify the nucleotide sequence encoding the active peptide such as described in this invention. Accordingly, the invention provides a method for screen ing a pool of nucleotide sequences to select a nucleotide se 5 quence encoding a peptide, said method comprising: (a) ligating a plasmid with the pool of nucleotide sequences operably linked to an inducible promoter, so as to ex press a peptide, which is an enzyme or a mature peptide of less than 100 amino acid residues, optionally linked 10 to a signal peptide, (b) transforming host cells which are sensitive to the pep tide with the ligated plasmids, (c) screening the transformed host cells so as to select vi able cells, 15 (d) cultivating the viable cells in the presence of inducer so as to induce expression of said nucleotide sequence, (e) selecting cells according to the effect of the inducer on cell growth, and (f) recovering the nucleotide sequence encoding the peptide 20 from the selected cells. The rationale of the presented suicide expression system (SES) is to generate peptide libraries in microorganisms, induce ex pression of the individual peptides, and select/identify cells according to whether they are killed or severely growth inhib 25 ited as a result of sudden peptide synthesis. For the identification of novel gene-encoded antimicro bial activities, libraries of genes harboring putative antim icrobial activities are cloned into the relevant plasmid, syn thesis is induced, growth-inhibited or dead bacteria are iden 30 tified and the corresponding gene sequenced and analyzed.

WO 00/73433 PCT/DK00/00287 3 For the identification of variants of peptides with in creased bioactivity, mutant libraries of an existing peptide is generated and introduced into the target organism. Successive rounds of peptide induction (using stepwise lower amounts of 5 inducer) , selection, plasmid amplification and shuf fling/mutagenesis will allow the identification of peptides with improved bioactivity. BRIEF DESCRIPTION OF DRAWINGS Figure 1. 10 Shows a flow diagram of the method step in a SES system using Fluorescence Assisted Cell Sorting (FACS) equipment for identi fication of modified Anti Microbial Peptides (AMP' s) , wherein A is a library of mutant Anti Microbial Peptides (AMP's) in bac terial host cells; B is FACS-mediated removal of dead bacteria; 15 C is induction of transcription; D is FACS-mediated selection of non-viable bacteria and E is PCR amplification, shuffling of amplified genes, cloning and transformation. The following symbols are used: WO 00/73433 PCT/DKOO/00287 4 -AMP - Defective AMP - Truncated AMP - amp gene o - Mutation * - Viability-probe - Viability-probe - Dead bacteria p - Transcription Figure 2 Shows a flow diagram of the screening strategy for conventional agar plates using solid media for identification of modified 5 AMP's, wherein A is distribution of microbial clones on agar plates; B is making of a replica plate; C is induction of tran scription and D includes characterization of colonies, such as properties, AMP sequence, identification of dead or inhibited cell colonies, PCR amplification, gene shuffling and re cloning 10 Figure 3. Shows a flow diagram of the screening strategy for microtiter plates using liquid media for identification of modified AMP's, wherein A is distribution of microbial clones in micro titer 15 wells; B is making of a replica plate; C is induction of tran scription and D includes characterization of colonies, such as properties, AMP sequence, identification of dead or inhibited cell colonies, PCR amplification, gene shuffling and re clon ing.

WO 00/73433 PCT/DKOO/00287 5 Figure 4. Show the effect on E. coli transformed with DNA encoding AMP's, wherein the expression of the AMP's is inducible with an arabi 5 nose inducer. In the vertical direction levels of inducer are indicated. different AMP's are tested, wherein 1 is Andropin; 2 is Bac7; 3 is Bac5, 4 is StyelinD; 5 is StyelinC; 6 is PR39; 7 is ClavA; 8 is ClavAK; 9 is CAP18 and 10 is pBAD. the effects on the different E. coli colonies are visually detectable. 10 Figure 5. A: Growth-curves at different levels of arabinose inducer of E. coli having induced expression of Andropin (PHHA1000-Andropin), wherein the Andropin is kept in the cytoplasm. Growth is moni 15 tored by measuring OD at 450 nm of cell suspensions. Inducer levels are given in % w/w. B: Growth-curves at different levels of arabinose inducer of E. coli having induced expression of Andropin (PHH1000-Andropin), wherein the Andropin is secreted to the periplasmic space. 20 Growth is monitored by measuring OD at 450 nm of cell suspen sions. Inducer levels are given in % w/w. Figure 6. A: Growth-curves at different levels of arabinose inducer of E. 25 coli having induced expression of Bac7 (PHHA1100-Bac7), wherein the Bac7 is kept in the cytoplasm. Growth is monitored by meas uring OD at 450 nm of cell suspensions. Inducer levels are given in % w/w. B: Growth-curves at different levels of arabinose inducer of E. 30 coli having induced expression of Bac7 (PHH1100-Bac7), wherein WO 00/73433 PCT/DKOO/00287 6 the Bac7 is secreted to the periplasmic space. Growth is moni tored by measuring OD at 450 nm of cell suspensions. Inducer levels are given in % w/w. 5 Figure 7. A: Growth-curves at different levels of arabinose inducer of E. coli having induced expression of BacS (PHHA1200-Bac5), wherein the Bac5 is kept in the cytoplasm. Growth is monitored by meas uring OD at 450 nm of cell suspensions. Inducer levels are 10 given in % w/w. B: Growth-curves at different levels of arabinose inducer of E. coli having induced expression of Bac5 (PHH1200-Bac5), wherein the Bac5 is secreted to the periplasmic space. Growth is moni tored by measuring OD at 450 nm of cell suspensions. Inducer 15 levels are given in % w/w. Figure 8. A: Growth-curves at different levels of arabinose inducer of E. coli having induced expression of StyelinD (PHHA1300-StyelinD), 20 wherein the StyelinD is kept in the cytoplasm. Growth is moni tored by measuring OD at 450 nm of cell suspensions. Inducer levels are given in % w/w. B: Growth-curves at different levels of arabinose inducer of E. coli having induced expression of StyelinD (PHH1300-StyelinD), 25 wherein the StyelinD7 is secreted to the periplasmic space. Growth is monitored by measuring OD at 450 nm of cell suspen sions. Inducer levels are given in % w/w. Figure 9. 30 A: Growth-curves at different levels of arabinose inducer of E.

WO 00/73433 PCT/DKOO/00287 7 coli having induced expression of StyelinC (PHHA1400-StyelinC), wherein the StyelinC is kept in the cytoplasm. Growth is moni tored by measuring OD at 450 nm of cell suspensions. Inducer levels are given in % w/w. 5 B: Growth-curves at different levels of arabinose inducer of E. coli having induced expression of StyelinC (PHH1400-StyelinC), wherein the StyelinC is secreted to the periplasmic space. Growth is monitored by measuring OD at 450 nm of cell suspen sions. Inducer levels are given in % w/w. 10 Figure 10. A: Growth-curves at different levels of arabinose inducer of E. coli having induced expression of PR39 (PHHA1500- PR39), wherein the PR39 is kept in the cytoplasm. Growth is monitored 15 by measuring OD at 450 nm of cell suspensions. Inducer levels are given in % w/w. B: Growth-curves at different levels of arabinose inducer of E. coli having induced expression of PR39 (PHH1500- PR39), wherein the PR39 is secreted to the periplasmic space. Growth is moni 20 tored by measuring OD at 450 nm of cell suspensions. Inducer levels are given in % w/w. Figure 11. A: Growth-curves at different levels of arabinose inducer of E. 25 coli having induced expression of ClavaninA (PHHA1600 ClavaninA) , wherein the ClavaninA is kept in the cytoplasm. Growth is monitored by measuring OD at 450 nm of cell suspen sions. Inducer levels are given in % w/w. B: Growth-curves at different levels of arabinose inducer of E. 30 coli having induced expression of ClavaninA (PHH1600- WO 00/73433 PCT/DKOO/00287 8 ClavaninA), wherein the ClavaninA is secreted to the periplas mic space. Growth is monitored by measuring OD at 450 nm of cell suspensions. Inducer levels are given in % w/w. 5 Figure 12. A: Growth-curves at different levels of arabinose inducer of E. coli having induced expression of ClavaninAK (PHHA1700 ClavaninAK), wherein the ClavaninAK is kept in the cytoplasm. Growth is monitored by measuring OD at 450 nm of cell suspen 10 sions. Inducer levels are given in % w/w. B: Growth-curves at different levels of arabinose inducer of E. coli having induced expression of ClavaninAK (PHH1700 ClavaninAK), wherein the ClavaninAK is secreted to the perip lasmic space. Growth is monitored by measuring OD at 450 nm of 15 cell suspensions. Inducer levels are given in % w/w. Figure 13. A: Growth-curves at different levels of arabinose inducer of E. coli having induced expression of CAP18 (PHHA1800-CAP18), 20 wherein the CAP18 is kept in the cytoplasm. Growth is monitored by measuring OD at 450 nm of cell suspensions. Inducer levels are given in % w/w. B: Growth-curves at different levels of arabinose inducer of E. coli having induced expression of CAP18 (PHH1800-CAP18), 25 wherein the CAP18 is secreted to the periplasmic space. Growth is monitored by measuring OD at 450 nm of cell suspensions. In ducer levels are given in % w/w. Figure 14. 30 A: Growth-curves at different levels of arabinose inducer of E.

WO 00/73433 PCT/DKOO/00287 9 coli having induced expression of control peptide Myc/HIS6, wherein the Myc/HIS6 is kept in the cytoplasm. Growth is moni tored by measuring OD at 450 nm of cell suspensions. Inducer levels are given in % w/w. 5 B: Growth-curves at different levels of arabinose inducer of E. coli having induced expression of control peptide Myc/HIS6, wherein the Myc/HIS6 is secreted to the periplasmic space. Growth is monitored by measuring OD at 450 nm of cell suspen sions. Inducer levels are given in % w/w. 10 Figure 15. Growth curves of E. coli having induced expression of StyelinC variants from randomly picked mutants. The randomly picked clones of mutant StyelinC is numbered from #1-#10. levels of 15 inducer is given in %w/w. DETAILED DESCRIPTION OF THE INVENTION Peptide The method of the invention may be used to screen pep tides according to their bioactivity, i.e. their ability to 20 kill or inhibit the growth of target cells. Thus, the peptide may be a peptide compound interacting/binding/sequestering es sential cellular targets. The peptide of interest may be an an timicrobial enzyme or a short peptide (less than 100 amino acid residues), e.g., an anti-microbial peptide (AMP) or an anti 25 tumor peptide. The antimicrobial enzyme may be, e.g., a muramidase, a lysozyme , a protease, a lipase, a phospholipase, a chitinase, a glucanase, a cellulase, a peroxidase, or a laccase. Alterna- WO 00/73433 PCT/DK00/00287 10 tively, a consortium of enzymes synthesizing conventional anti biotics, e.g. polyketides or penicillins, can be employed. The antimicrobial peptide (AMP) may be, e.g., a membrane active antimicrobial peptide, or an antimicrobial peptide af 5 fecting/interacting with intracellular targets, e.g. binding to cell DNA. The AMP is generally a relatively short peptide, con sisting of less than 100 amino acid residues, typically 20-80 residues. The antimicrobial peptide has bactericidal and/or fungicidal effect, and it may also have antiviral or antitumour 10 effects. It generally has low cytotoxicity against normal mam malian cells. The antimicrobial peptide is generally highly cationic and hydrophobic. It typically contains several arginine and ly sine residues, and it may not contain a single glutamate or as 15 paratate. It usually contains a large proportion of hydrophobic residues. The peptide generally has an amphiphilic structure, with one surface being highly positive and the other hydropho bic. The bioactive peptide and the encoding nucleotide se 20 quence may be derived from plants, invertebrates, insects, am phibians and mammals, or from microorganisms such as bacteria and fungi. The antimicrobial peptide may act on cell membranes of target microorganisms, e.g. through nonspecific binding to the 25 membrane, usually in a membrane-parallel orientation, interact ing only with one face of the bilayer. The antimicrobial peptide typically has a structure be longing to one of five major classes: a helical, cystine-rich (defensin-like) , P-sheet, peptides with an unusual composition WO 00/73433 PCT/DKOO/00287 11 of regular amino acids, and peptides containing uncommon modi fied amino acids. Examples of alpha-helical peptides are Magainin 1 and 2; Cecropin A, B and Pl; CAP18; Andropin; Clavanin A or AK; Stye 5 lin D and C; and Buforin II. Examples of cystine-rich peptides are a-Defensin HNP-1 (human neutrophil peptide) HNP-2 and HNP 3; r-Defensin-12, Drosomycin, yl-purothionin, and Insect defen sin A. Examples of P-sheet peptides are Lactoferricin B, Tachy plesin I, and Protegrin PG1-5. Examples of peptides with an un 10 usual composition are Indolicidin; PR-39; Bactenicin Bac5 and Bac7; and Histatin 5. Examples of peptides with unusual amino acids are Nisin, Gramicidin A, and Alamethicin. Another example is the antifungal peptide (AFP) from As pergillus giganteus. 15 In a preferred embodiment the expressed peptide is free of any protecting scaffold proteins. Pool of nucleotide sequences The commercial utility of antimicrobial peptides as anti biotics or antimicrobial agents depends on their potency, spe 20 cies specificity and ability to perform under the appropriate conditions. More often than not, these conditions are quite different from those under which the peptide originally evolved. Most antimicrobial peptides have, for example, not been evolved to simultaneously target a broad range of differ 25 ent microbes, to work in a physiological salt range, to evade the human immune system or resist the clearing capacity of the mammalian circulatory system. For a given antimicrobial peptide, this dilemma can in principle be solved by either knowledge-based rational modifi- WO 00/73433 PCT/DKOO/00287 12 cations of the peptide or by directing further the evolution of the peptide, creating random variants of the parental sequence, and subsequently selecting mutants in which the desired combi nation of properties are found. Directed evolution, an itera 5 tive process by which large areas of sequence space are ex plored to create mutant proteins and peptides that possess par ticular desired characteristics, combined with powerful High Throughput assays allows large libraries of native or modified gene-encoded antimicrobial peptides to be generated and evalu 10 ated for the identification of lead candidates with the desired characteristics. These approaches are now being adopted widely by academics and the industry alike to create novel protein based activities at an unprecedented rate. A nucleotide sequence, which encodes the bioactive pep 15 tide, may be obtained from chromosomal DNA from on of the above-mentioned source organisms and/or it may be chemically synthesized. The nucleotide sequence may also be a cDNA derived from such source organisms. The screening method of the invention may be used to de 20 velop peptides with an improved bioactivity. Thus, starting with a known gene encoding a bioactive peptide, a DNA pool may be obtained, e.g., by random mutagenesis to produce a mutant library, by gene shuffling, or by synthesizing degenerate genes. 25 The sequences to be shuffled may be related sequences from different organisms (so-called "family shuffling"), or they may include a parent sequence and a variant thereof. In a preferred embodiment of the invention random mutagenesis is achieved by shuffling of homologous DNA se 30 quences in vitro such as described by Stemmer (Stemmer, 1994.

WO 00/73433 PCT/DKOO/00287 13 Proc. Natl. Acad. Sci. USA, 91:10747-10751; Stemmer, 1994. Na ture 370:389-391) and Crameri, A., et al., 1997. Nature Bio technology 15:436-438 all incorporated by reference. The method relates to shuffling homologous DNA sequences by using 5 in vitro PCR techniques. Positive recombinant genes containing shuffled DNA sequences are selected from a DNA library based on the improved function of the expressed proteins. The above method is also described in WO 95/22625, hereby incorporated by reference, in relation to a method for shuf 10 fling homologous DNA sequences. An important step in the method is to cleave the homologous template double-stranded polynucleotide into random fragments of a desired size followed by homologously reassembling the fragments into full-length genes. 15 In another preferred embodiment of the invention random mutagenesis is achieved by the method described in WO 98/41653, incorporated by reference, which discloses a method of DNA shuffling in which a library of recombined homologous polynu cleotides is constructed from a number of different input DNA 20 templates and primers by induced template shifts during in vi tro DNA synthesis. In this context especially the special ver sion of this in vitro recombination through induced template shifts during DNA synthesis, also described in WO 98/41653, is preferred. Here, small (>5 nucleotides) random DNA primers are 25 employed to randomly initiate DNA synthesis on the mutant DNA templates that are to be combined. Due to the small size of the genes encoding antimicrobial peptides, special attention has to be taken into consideration when using each of the above methods for generation and combi 30 nation of sequence diversity. Since most shuffling methods rely WO 00/73433 PCT/DK00/00287 14 on a substantial number of identical bp (20-100 bp) flanking the mutations that has to be recombined, the mutations in small genes are technically difficult to combine by the above de scribed methods. 5 Accordingly, other formats of directed evolution have to be employed on small genes. In a preferred embodiment involving the combination of variants of a given peptide of less than ap proximately 50 amino acids, one degenerate DNA primer harboring all the desired mutations would be synthesized. In a given po 10 sition in this degenerate primer, both the wt nucleotide as well as the mutant nucleotide should be present. The frequency of wt-to-mutant nucleotides can be adjusted as considered opti mal; rules and considerations are known in the art. By includ ing all desired mutations in one primer, the desired sequence 15 space could be completely sampled. This method allows for the sampling and combination of all desired mutations irrespec tively of how close they would be in the primary gene sequence. If peptides of more than approximately 50 amino acids are employed, two or more separate and degenerate primers would 20 have to be used. This is due to the constraints generally ex perienced when synthesizing DNA primers; only DNA primers of less than approximately 180-200 nucleotides can routinely be synthesized. In another embodiment where peptides longer than approxi 25 mately 50 amino acids are employed, the sequence diversity (the individual mutants) to be combined can individually be harbored in small oligonucleotides of 20-30 base pairs of length. In this approach, a specific DNA oligo is employed for each muta tion that should be included in the library. The mutations 30 should preferentially be located in the middle of the small WO 00/73433 PCT/DKOO/00287 15 oligo to optimize annealing. Spiking in several or numerous of these small oligoes in a PCR reaction using the wt peptide gene as backbone for the amplification, would allow for the combina tion of the desired mutations. By varying the amount of the in 5 dividual oligoes to be combined, desired ratios of individual variants to wt's can be created. As approximately 10 base pairs is required on each side of the sequence mismatch, this method cannot effiently combine mutations that are immediately adja cent. 10 The Suicide Expression System is not limited to the iden tification of improved variants of existing and already charac terized peptides. New genes encoding peptides that affects the growth of a given host cell can also be identified. Libraries of cDNA's or randomly generated whole-genome DNA fragments can 15 be employed as starting material and cloned into the Suicide Expression System. Host cell The host cell must be sensitive to the peptide, enzyme or 20 secondary metabolite of interest. Thus, in the case of screening for an antimicrobial pep tide (AMP) or an antimicrobial enzyme, the host cell can be a bacterium such as E. coli or Bacillus, e.g. B. subtilis, or the host cell can be a fungal cell, e.g. a filamentous fungus such 25 as Aspergillus or a yeast, such as Saccharomyces or Candida. It may be preferred to use a host cell related to the target mi croorganisms against which the antimicrobial peptide is in tended to be used.

WO 00/73433 PCT/DKOO/00287 16 In the case of screening for an anti-tumor peptide, the host cell is preferably a mammalian cell, particularly a tumor cell. The host cell should be capable of transporting the in 5 ducer across the membrane preferably without metabolizing or degrading it. This is advantageous for expression studies as the level of inducer will be constant inside the cell and not decrease over time. This can be achieved by selecting a "gra tuitous" inducer, or it can be achieved by deleting one or more 10 genes necessary for metabolism of the inducer. The host cell must be selected so as to be able to ex press the antimicrobial peptide. Thus, a fungal cell is pre ferred for peptides with disulfide bridges such as the cystine rich peptides mentioned above. 15 Plasmid The plasmid should be replicable in the host organism, and should be able to express the bioactive peptide (and op tionally signal peptide) under the control of the inducible promoter. It will usually contain a selectable marker such as 20 an antibiotic marker. A number of such plasmids are known in the art. Ligation The plasmid is ligated with the pool of nucleotide se quences so that these sequences may be operably linked to an 25 inducible promoter in the plasmid, enabling inducible expres sion of the peptide of interest, optionally linked to a signal peptide. The short peptide or enzyme of interest may be ex pressed without any extension (other than the optional signal WO 00/73433 PCT/DKOO/00287 17 peptide) , or it may be expressed with a short extension of, e.g., 1-5 amino acids. Expression in the form of a fusion pro tein is neither preferred nor necessary. Inducible promoters and inducers 5 The plasmid to be used according to the invention must comprise an inducible promoter regulating the expression of the inserted nucleotide sequence encoding the peptide. It is advan tageous for the applicability of the SES, that it allows a com plete shutdown of the synthesis of the encoded bioactive pep 10 tide. In addition, the induction of the encoded bioactive pep tides should be significant, since peptides are inherently un stable and easily degraded in the cytoplasm of microorganisms. The inducible promoter employed in the current examples is both positively and negatively regulated by two proteins. In the 15 presence of inducer, expression from the promoter is turned on, while in the absence of inducer, very low levels of expression occur from the promoter. Uninduced levels are repressed even further by growth in the presence of a secondary metabolite. By varying the activity of the two regulators, protein expression 20 levels can be manipulated to optimize expression of potentially toxic or essential genes. The promotor may be the Lac promotor as descibed in Ta guchi S., Nakagawa K. , Maeno M. and Momose H. ; "In Vivo Moni toring System for Structure-Function Relationship Analysis of 25 the antibacterial peptide Apidaecin"; Applied and Environmental Microbiology, 1994, pp. 3566-3572, which may be regulated by presence of the inducer lactose or by the synthetic non digestible lactose derivative IPTG. Other inducible promoters are known in the art such as trp promoters induced by trypto- WO 00/73433 PCT/DKOO/00287 18 phan or gal promoters induced by galactose for E. coli, gall promoter for S. cerevisiae, AOXl promoter for Pichia pastoris, pMT (metallothionein) promoter for Drosophila, MMTV LTR , pVgRXR or pIND promoters for mammalian expression. Using an in 5 ducer that is not metabolized or digested in the cell offers the advantage that the inducer concentration may be kept con stant throughout the screening process. However a drawback of the Lac promoter may be that it cannot be entirely switched off by the absence of the inducer. The promoter may also be the 10 pBAD promoter as used in the examples, vide infra. This promo tor is, inter alia, induced by the digestible inducer arabi nose. However to achieve the above mentioned advantage of hav ing a constant level of inducer, the host cells ability to di gest arabinose can be eliminated by deleting suitable genes 15 from the host cell genome (a description of the genotype may be found in the examples) . An important consideration selecting a suitable promotor is however that the corresponding inducer should be able to permeate the cell membrane(s) to gain access to the promoter. 20 The pBAD promoter is both positively and negatively regulated by two proteins, AraC and cAMP-CRP. In the presence of arabinose, expression from the promoter is turned on, while in the absence of arabinose, only very low levels of expression occur from the promoter. 25 Uninduced levels are repressed even further by growth in the presence of glucose. Glucose acts by lowering cAMP levels, which in turn decreases the binding of cAMP-CRP to the promoter region of pBAD. As cAMP levels are lowered, transcriptional ac tivation is decreased. This is ideal when the peptide of inter 30 est is extremely growth inhibitive or toxic to the host. In WO 00/73433 PCT/DKOO/00287 19 conclusion, by varying the activity of the two regulators, pro tein expression levels can be manipulated to optimize expres sion of potentially toxic or essential genes. Signal peptide 5 A DNA sequence encoding a signal peptide may optionally be inserted into the plasmid downstream of the inducible pro moter and upstream of the sequence encoding the peptide, so that the antimicrobial peptide will be expressed with the sig nal peptide attached. A suitable signal peptide for a given 10 host cell may be selected according to principles known in the art. Generally, the peptide of interest initially attacks or penetrates the target organism from the outside, so the success of the SES will in most cases require that the peptide in ques 15 tion is exported out of the cell. In general, the bioactive peptide will be secreted through the cell membrane, e.g. to the periplasmic space in gram-negative prokaryotes, and from there allowed to interact with its cellular target, e.g. the cellular membranes, components in the membranes or the periplasmic 20 space, or allowed to further diffuse through the outer mem brane. A signal peptide can be omitted if the peptide of inter est can exert its action when expressed and retained within the cell, e.g. peptides that bind to the cellular DNA or peptides 25 that do not depend on a trans-membrane potential or peptides with intracellular targets. An example is the family of proline-arginine-rich peptides Bac5, Bac7 and PR39, which in the literature have been suggested to interact and sequester nucleic acids.

WO 00/73433 PCT/DKOO/00287 20 Screening process As mentioned above the invention relates to a method for screening a pool of nucleotide sequences to select a nucleotide 5 sequence encoding an antimicrobial peptide which acts on cell membranes, cell walls or DNA of target microorganisms, said method comprising the steps of: (a) ligating a plasmid with the pool of nucleotide sequences operably linked to an inducible promoter, so as to ex 10 press a peptide which is an enzyme or a mature peptide of less than 100 amino acid residues, optionally linked to a signal peptide, (b) transforming host cells which are sensitive to the pep tide with the ligated plasmids, 15 (c) screening the transformed host cells so as to select vi able cells, (d) cultivating the viable cells in the presence of inducer so as to induce expression of said nucleotide sequence, (e) selecting cells according to the effect of the inducer on 20 cell growth, and (f) recovering the nucleotide sequence encoding the peptide from the selected cells. Prior to step a) of the screening process certain pre paratory steps may be necessary. A host, for which a killing 25 and/or growth inhibiting peptide is desired to be found and a suitable plasmid compatible with that host should be chosen. A library of nucleotide sequences, such as a pool of nucleotide sequences derived by mutating a known sequence encoding a known WO 00/73433 PCT/DKOO/00287 21 antimicrobial peptide, should be prepared, e.g. by conventional methods, such as described vide supra. In step a) the library is ligated to the suitable plasmid and transformed (step b) into the host cell culture by conven 5 tional methods. Step c) is a first screening or selection step, in which viable cells are separated from cells which died and/or became growth inhibited during the cause of the transformation proc ess. This step is an important one because the ultimate goal in 10 the screening process is to identify cells that dies and/or is growth inhibited by the induced expression of the inserted nu cleotide sequence producing an antimicrobial peptide. Cells which death and/or inhibition occurred before the screening process and thus is not caused by the antimicrobial peptide 15 would generate a false positive response in the screening if they were not separated from the viable cells. In step d) an inducer is introduced to the viable cells, which are cultivated so as to induce expression of the nucleotide sequence from the library comprised in the inserted plasmid. As the peptide is 20 produced by transcription the host cell will die and/or be growth inhibited if said peptide has antimicrobial effect against the host. In a preferred embodiment host cells which are dead and/or growth inhibited are selected. By selection of dead and/or growth inhibited host cells, cell comprising nu 25 cleotide sequences encoding peptides having antimicrobial ac tivity may be isolated. More than one level of inducer concen tration may be employed in parallel so that a graduated re sponse may be achieved and nucleotide sequences encoding pep tides having different antimicrobial effects or potency may be 30 identified. In step e) host cells which die and/or becomes WO 00/73433 PCT/DKOO/00287 22 growth inhibited are selected and separated from host cells which are unaffected by the peptide expressed from the plasmid nucleotide sequence which was inserted under the transformation of the host cells. One may select only cells which are greatly 5 affected by the induced expression of a peptide, e.g. which are affected by small concentrations of inducer, or one may select all affected cells, depending on the intended scope of the screening and/or the existing knowledge of the pool or library of nucleotide sequences. 10 The criterion on which a cell is selected may be chosen individually, e.g. a maximum inducer concentration may be set so that only cells which are affected by the presence of in ducer below this inducer concentration are selected, and/or de creasing levels of transcriptional induction using incremental 15 lower concentrations of inducer on replicas of the transformed host cells will allow the isolation of peptides with increased bioactivity (figure 4). The inserted nucleotide sequences from the selected host cells identified as encoding bioactive peptides may be recov 20 ered by conventional methods. The nucleotide sequences may be amplified by conventional methods, e.g. PCR amplification. From here an identified and amplified nucleotide sequence may be in serted into a production host and the corresponding peptide identified may be mass produced according known methods where a 25 peptide may be expressed through fusion to a bigger polypeptide which then may be exported by the host cell. Said polypeptide may have the function of protecting the peptide of interest from digestion within the cell and thereby inactivation by the host cell enzymes and/or the polypeptide may have the function 30 of lowering the effect of the peptide on the host cell so that WO 00/73433 PCT/DKOO/00287 23 the host may proliferate and continue expression of the peptide without being significantly affected by the expressed peptide, an effect which would occur if the peptide had not been incor porated into the polypeptide. The identified and amplified nu 5 cleotide sequence encoding the peptide may also be mutated as described, vide supra, e.g., by random mutagenesis, by gene shuffling, or by synthesizing degenerate genes. These new mu tated nucleotide sequences may then be screened again according to steps a) to f) to identify nucleotide sequences encoding new 10 peptides with an improved effect e.g. by lowering the concen tration of inducer in subsequent screenings. The screening or separation processes may in a specific embodiment be conducted by application of conventional plate assays, so that the transformed host cells are streaked out on 15 plates comprising a nutrient medium and optionally an antibi otic. If the transformation plasmid comprises a gene for resis tance to such an antibiotic untransformed host cells will die on such a medium while transformed host cells will survive. The plates are then incubated for a predetermined period of time to 20 enable colony formation of transformed host cells and from these plates cell samples are transferred to other plates fur ther comprising an inducer inducing expression and production of the peptide comprised in the inserted plasmid. If a trans formed host cell keeps growing and forms normal colonies in 25 this environment it may be deduced that the expressed and pro duced peptide do not kill and/or inhibit the host cell. If on the other hand the host cell does not form any colonies or re duced colonies as compared to normal growth, it is evident that the induced peptide has an antimicrobial effect on the host 30 cell. A depiction of the screening strategy for conventional WO 00/73433 PCT/DKOO/00287 24 agar plates is given in figure 2.Plate assays, however, in volves time consuming and tedious procedures and in a more pre ferred embodiment the screening or separation processes are performed in microtiter assays as described in the art. In this 5 type of assay the liquid host cell culture is placed with a single or only a few cells comprising different inserted nu cleotide sequences in each well of microtiter plates or other wise securing that each well comprise only a single or a few nucleotide sequences to be investigated (e.g. a large number of 10 host cells comprising the same inserted nucleotide sequence) . The host cells in each well may be cultivated by addition of a nutrient medium and a copy or replicas of the microtiter plates may be prepared by transferring subsamples to additional test ing plates. A medium containing an inducer may then be added to 15 each well and the proliferation of the host cell culture in each well upon cultivation may be monitored, e.g. by measuring the optical diffraction through the cell suspension in each well. If a host cell grows unaffected of the expressed peptide the number of cells in this well will increase normally and the 20 optical diffraction of the cell suspension measured through the well will increase. If, however, the host cell growth is af fected of the expressed peptide the number of cells in this well will be lowered as compared to normal growth and the change in optical diffraction of the cell suspension will also 25 be lowered. A depiction of the screening strategy for micro titer plates is given in figure 3.As a third and most preferred embodiment the screening or separation processes may be per formed by employing Fluorescence Assisted Cell Sorting (FACS) equipment such as described in Gant V.A., Warnes G., Phillips 30 1. and Savidge G.F.; "The application of flow cytometry to the WO 00/73433 PCT/DKOO/00287 25 study of bacterial responses to antibiotics"; J. Med. Micro biol.; 1993; 39; pp. 147-154. This type of equipment is exten sively described in the art, e.g. by the manufacturers of such equipment. Using this approach a viability probe, e.g. a fluo 5 rescent or colorimetric probe) is incorporated in the host cells, the probe being an indicator of the proliferation of each cell. Suitably the inducer and viability probe is added to an exponentially growing liquid culture of host cells, and the dead or growth-inhibited microorganism is identified and col 10 lected. Having a such probe incorporated the viability of a cell may be monitored by measuring the fluorescence of the probe in the cell by exposing the cell with excitation light of a wave length suitable for the probe, e.g. if fluorescence of the 15 probe can be measured the cell is alive or vice versa. With a FACS machine cells which exhibits the desired characteristics may be selected at a tremendous speed and accuracy also aided by the fact that fluorescence measurements are highly sensi tive. In the method of the invention FACS equipment may be em 20 played in the screening or selection step c) where viable and transformed host cells are selected and/or e) where dead and/or inhibited cells are selected after inducing expression of the peptide or the FACS equipment may be combined with plate and/or microtiter plate techniques as described supra. Many suitable 25 fluorescent probes are commercially available for this purpose, e.g. from Molecular Probes, Inc, Eugene, OR, USA. Using such probes it may be monitored whether e.g. the membrane-structure is compromised or deteriorated, whether the cross-membrane po tential is reduced or eliminated, or whether specific probes 30 are allowed to interact with intracellular targets (e.g. DNA).

WO 00/73433 PCT/DKOO/00287 26 Examples of such probes include but is not limited to SYTO(X)* nucleic acid stains from Molecular Probes, Inc. which are probes which are designed to either penetrate dead and/or dam aged cells and make nucleic acids within such cells fluorescent 5 (e.g. SYTOX* Green nucleic acid stain) or it may be designed to penetrate and make fluorescent living cells (e.g. SYTO® live cell nucleic acid stains). The procedures for using such probes are available from the manufacturer. Also fluorescent redox probes (sensitive towards the cross membrane potential may be 10 employed as described in Rodriguez, G.G., Phipps D., Ishiguro K. and Ridgway H.F.; "Use of a Fluorescent Redox Probe for Di rect Visualization of Actively Respiring Bacteria", Applied and Environmental Microbiology, 1992, pp. 1801-1808, wherein a 5 cuano-2,3-ditolyltetrazolium chloride probe is employed or in 15 Jepras R.I., Paul F.E., Pearson S.C. and Wilkinson M.J.; " Rapid Assessment of Antibiotic Effects on Escherichia coli by bis-(l,3-Dibutylbarbituric Acid) trimethine Oxonol and Flow cy tometry; Antimicrobial Agents and Chemotherapy; 1997; pp. 2001 2005, wherein a DiBAC 4 (3) probe available from Molecular Probes 20 is employed. FACS equipment or conventional luminiscence equipment may also be adapted to the use of bioluminiscence, a technique de scribed in Virta M, Akerman K.E.O., Saviranta P., Oker-Blom C. and Karp M.T.; " Real time measurement of cell permeabilization 25 with low-molecular-weight membranolytic agents", Journal of An timicrobial Chemotherapy; 1995; 36; pp. 303-315. In addition, colorimetric indicators may be applied such as described in Roslev P. and King G.M.; " Application of a Tetrazolium Salt with Water-Soluble Formazan as an Indicator of viability in Re 30 spiring Bacteria"; Applied and Environmental Microbiology, WO 00/73433 PCT/DKOO/00287 27 1993, pp. 2891-2896; although such indicators may more suitable be employed non-fluorescence cell sorting equipment designed for colorimetric measurements. A depiction of the screening strategy for FACS is given in figure 1. 5 Applying the method of the invention for screening for bioactive peptides in cells, in particular eukaryotic cells, e.g. mammalian cells, an alternative preferred screening or se lection strategy may be employed. Upon the induced expression of the inserted nucleotide sequence the resulting peptide may, 10 if it is bioactive against the host cell, disrupt and/or dete riorate the cell membrane and cells killed and/or affected cells may be separated from unaffected cells by centrifugating or filtering the cell suspension. If centrifugating the unaf fected living cells will precipitate while nucleic acid mate 15 rial from affected cells may be isolated in the supernatant. This separation may also be achieved by filtering off the unaf fected living cells. From this separation step amplification and mutation can be carried out as described vide supra. Once the best peptides has been identified other tests 20 may be performed, wherein the peptide sensitivity towards salt concentration, ionic strength, pH and/or and especially sensi tivity of normal mammalian cells toward the peptides or other relevant parameters and conditions are tested depending of the intended application of the peptides. 25 Use of antimicrobial peptide The antimicrobial peptides found from the method of the inven tion may be employed in many areas of application. One suitable area is preservation of e.g. of food/feed, paint formulations, detergent formulations, cosmetics or other personal care prod- WO 00/73433 PCT/DKOO/00287 28 ucts as an alternative to chemical preservatives. The peptides may also be used to preserve medical devices such as prosthetic implants, intravenous tubing e.g. by coating such materials with a coating comprising said peptides. The peptides may also 5 be actively applied to disinfect and/or kill and/or inhibit mi crobial cells on an object e.g. in the cleaning industry, e.g. as an disinfectant for treatment of biofilm. One preferred ap plication is the preparation of peptides for treating microbial infections and/or tumors in the human and/or animal body or on 10 the skin or mucous membranes. It is contemplated that the use of the screening method of the invention is a versatile tool for finding extremely bioactive peptides which is able of kill ing and/or inhibiting microbial and/or tumor cell, but which have little or no negative effect on normal mammalian and/or 15 eukaryotic cells. The peptides may be formulated for oral ad ministration or for intravenous or subcutaneous injection or as an ointment. EXAMPLES Example 1: Growth inhibiiton of E. coli upon expression of 20 various antimicrobial peptides Genes of model AMP's, i.e. CAP18, PR39, Andropin, Bac5, Bac7 ClavaninA, Clavanin AK (a Clavanin A variant) , Styelin D and Styelin C, were synthesized using DNA oligoes in a standard PCR reaction and cloned in the presented SES in order to evalu 25 ate its potential for identifying AMP's with improved bioactiv ity.

WO 00/73433 PCT/DKOO/00287 29 Plasmid Two series of experiments were made: Series pHH (using plasmid pBAD/gIIIA) allowed for the export of the AMP's to the periplasmic space of E. coli, from where the peptides were al 5 lowed to interact with the cellular membranes. In series pHHA (using plasmid pHHA) the peptides lacked a signal sequence and correspondingly were retained in the cytoplasm. One of the parental plasmids employed, pBAD/gIIIA, is commercially available from Invitrogen. It is a pUC-derived ex 10 pression vector designed for tightly regulated, recombinant protein expression in E. coli. This plasmid allows the cloning of peptides and proteins toxic to E. coli, as no expression of the recombinant peptides occurs in the absence of inducer in the growth medium. However, transcription and hence peptide 15 synthesis, can be extensively induced. Secretion As all AMP's initially attacks or penetrates the target organism from the outside, the success of the SES will in most cases require that the AMP in question is exported out of the 20 cell. In the current system, the AMP was secreted to the perip lasmic space, and from there allowed to interact with its cel lular target, e.g. the cellular membranes, components in the membranes or the periplasmic space, or allowed to further dif fuse through the outer membrane. 25 In series pHH, the gene III signal sequence in pBAD/gIIIA was located in front of the inducible promoter in order to me diate secretion of the peptide/protein in question. Gene III encodes pIII, one of the minor capsid proteins from the fila mentous phage fd. pIII is synthesized with an 18 amino acid, N- WO 00/73433 PCT/DKOO/00287 30 terminal signal sequence, and requires the bacterial Sec system for insertion into the membrane. The signal sequence was re moved after crossing the inner membrane, thus leaving the na tive peptide. A NcoI restriction site immediately succeeds the 5 signal sequence cleavage site. In the other parental plasmid series, pHHA, the peptides were synthesized without a signal sequence, and correspond ingly, were retained in the cytoplasm. The pHHA plasmid differs only from pBAD/gIIIA, in that the gene III has been deleted. 10 This deletion was produced by the introduction of an additional Ncol site overlapping the translation initiation site. Hence, the signal sequence was removed by restriction with NcoI, and the plasmid religated to produce pHHA. In both plasmid systems (pBAD/gIIIA and pHHA) , the AMP 15 genes are inserted as a NcoI-XbaI fragment. Regulation of the inducible promoter The inducible promoter employed, pBAD, is both positively and negatively regulated by two proteins, AraC and cAMP-CRP. In the presence of arabinose, expression from the promoter is 20 turned on, while in the absence of arabinose, very low levels of expression occur from the promoter. Uninduced levels are re pressed even further by growth in the presence of glucose. Glu cose acts by lowering cAMP levels, which in turn decreases the binding of cAMP-CRP to the promoter region of pBAD. As cAMP 25 levels are lowered, transcriptional activation is decreased. This is ideal when the peptide of interest is extremely growth inhibitive or toxic to the host. In conclusion, by varying the activity of the two regulators, protein expression levels can WO 00/73433 PCT/DKOO/00287 31 be manipulated to optimize expression of potentially toxic or essential genes. C-terminal myc epitope and 6xHis tag The various AMP genes mentioned above, were cloned in 5 front of an in-frame myc epitope and 6xHis tag. A TAG stop codon separated the AMP gene and the sequence encoding the two epitopes. This means that in normal E. coli cells, only the AMP is synthesized upon transcriptional induction. Translational termination at the TAG stop codon can, if the fusion protein is 10 non-toxic to E. coli, be suppressed in various strains, allow ing for synthesis of an AMP/myc/6xHis fusion protein. This fu sion protein is easily purified using affinity nickel (Ni") resins, and is easily detected using anti-Myc or anti-6xHis an tibody. Following purification, CnBr cleavage will separate the 15 two tags from the AMP by cleavage at a conveniently located me thionine. This system allows for an easy, convenient purifica tion of selected peptides allowing for confirmation, using tra ditional assays, of the antimicrobial activity. Host organism 20 The strain employed was E. coli TOP10 (commercially available from Invitrogen) . It is araBADC and araEFGH*. AMP genes The active fragments of the following model AMP's, i.e. CAP18, PR39, Andropin, Bac5, Bac7 ClavaninA, ClavaninAK (a Cla 25 vaninA variant) , StyelinD and StyelinC, were synthesized in a standard PCR reaction using the DNA oligoes shown below. The PCR amplified AMP genes were restricted with NcoI and XbaI, and ligated into the corresponding cloning sites in pBAD/gIIIA and WO 00/73433 PCT/DKOO/00287 32 pHHA. These ligation mixtures were transformed into chemically competent E. coli TOP10, and individual clones analyzed. The AMP genes were finally verified by DNA sequencing. Primers used for synthesis of the AMP genes. 5 Andropin c.cggccatggtatttattgatattcttgacaaagtggaaaacgcaatacacaatgctgctcaa gtgggaattggctttgctaagccctttgaaaaattgatcaatccgaagtagatggctctagac ggc 10 PR39 ccggccatggcgaggagacgtccccgacccccatatttgccaaggccaaggccacctccgttt ttcccaccaaggctcccaccaaggatcccaccagggttcccaccaaggttcccaccacggttc cccggaaaacggtagatggctctagacggc 15 Bac5 ccggccatggcgagatttcgtccaccaatccgtcgtccaccaatccgtccgccgttctatcca ccgttccgcccgccgatccgcccaccgatcttcccaccgatccggccaccgttccgtccaccc ttaggaccgtttcctggtagacggtagatggctctagacggc 20 Bac7-Forw ccggccatggcgaggagaattcgtccccggccaccacgtttgccaaggccaaggccaaggcca ttgccattcccacggcctgggccaaggccaattccaaggccactgccattcccacggcctggg ccaaggccaattccaaggccactg WO 00/73433 PCT/DKOO/00287 33 Bac7 -rev gccgtctagagccatctacaatggccttggaattggccttggcccaggccgtgggaatggcag tggccttggaattggccttggcccaggccgtgggaatggcagtggccttggaattgg 5 CAP 18 ccggccatggggctgcgcaagcgcttacgaaaatttagaaacaagattaaagaaaagcttaaa aaaattggtcagaaaatccagggtttcgtgccgaaacttgcacccaggacagattactagatg gctctagacggc 10 Clavanin A ccggccatggtattccaattccttggcaaaattattcatcatgttggcaattttgtacatggt tttagccacgtgttttagatggctctagacggc 15 Clavanin AK ccggccatggtattccaattccttggcaaaattattaagaaggttggcaattttgtaaagggt tttagcaaggtgttttagatggctctagacggc Styelin C 20 ccggccatgggctggtttggaaaagctttcagatcagtaagcaacttttacaaaaaacataaa acatacatccatgcaggactttcagctgctacattgcttggttagatggctctagacggc WO 00/73433 PCT/DKOO/00287 34 Styelin D ccggccatgggttggttgagaaaagctgccaaatctgtaggaaaattttactacaaacacaaa tattacatcaaagcagcctggcaaattggaaagcatgccttaggttagatggctctagacggc The corresponding amino acid sequences of the AMP genes 5 are shown below. The amino acids in lower case are not present in the native AMP's but were introduced as a result of the cloning strategy where NcoI is employed as the proximal cloning site (CCATGG; ATG encodes methionine). The natural codon usage of the genes have been retained. 10 CAP18 mGLRKRLRKFRNKIKEKLKKIGQKIQGFVPKLAPRTDY PR39 maRRRPRPPYLPRPRPPPFFPPRLPPRIPPGFPPRFPPRFPGKR 15 Andropin mVFIDILDKVENAIHNAAQVGIGFAKPFEKLINPK Bac5 maRFRPPIRRPPIRPPFYPPFRPPIRPPIFPPIRPPFRPPLGPFPGRR Bac7 maRRIRPRPPRLPRPRPRPLPFPRPGPRPIPRPLPFPRPGPRPIPRPLPF 20 PRPGPRPIPRPL ClavaninA mVFQFLGKIIHHVGNFVHGFSHVF WO 00/73433 PCT/DKOO/00287 35 ClavaninAK mVFQFLGKIIKKVGNFVKGFSKVF StyelinC mGWFGKAFRSVSNFYKKHKTYIHAGLSAATLLG 5 StyelinD mGWLRKAAKSVGKFYYKHKYYIKAAWQIGKHALG Growth Inhibition of E. coli upon Expression of Various AMP's The following experiment was conducted in order to evalu ate whether E. coli, upon induction of endogenous AMP expres 10 sion, was growth-inhibited in liquid media. On day one, cells harboring the various AMP-encoding plasmids, were inoculated in LB + 100 y ampicillin and grown under non-induced conditions at 37TC with vigorous shaking. These overnight cultures were on day two diluted 100-fold into 15 fresh, pre-warmed LB broth + 100 y ampicillin, in the presence of varying amounts of arabinose (0%; 0.001%, 0.01% and 0.1%). 100 pLl of these freshly diluted cultures were transferred to microtiter plates, and incubated at 37 C with vigorous shaking. Growth of the cultures were measured at regular intervals at 20 OD 450 using an ELISA reader. The corresponding growth-curves are shown in figures 5-14. In the pHH-series (using the parental plasmid pBAD/gIIIA) , where the peptides are secreted to the periplasmic space, expression of all AMP's except Andropin, significantly 25 inhibited the growth of the bacteria. Similar results were ob tained when other strains of E. coli with similar genotype WO 00/73433 PCT/DKOO/00287 36 (araBADC~ and araEFGH*) were employed (Data not shown) . An dropin has been reported to require substantial concentrations of salt in order to fold and exert antimicrobial activity; an observation that can explain the less potent inhibition seen. 5 No growth inhibition is evident in strains carrying the control plasmid pBAD/gIIIA expressing the 38 amino acid Myc/HIS6 con trol peptide fused to the pIII signal sequence. Hence, it can be concluded that expression of peptides fused to the pIII sig nal sequence per se does not significantly inhibit the growth 10 of E. coli. The results are shown in Figs. 5b - 14b. In the pHHA-series, where the peptides were retained in the cytoplasm, a different pattern of growth inhibition was ob served. Only a subset of the AMP exerted growth-inhibition in this context. This difference most likely reflects differences 15 in mode of action of the AMP's. Peptides that do not depend on a trans-membrane potential or peptides with intracellular tar gets (e.g. the family of proline-arginine-rich peptides Bac5, Bac7 and PR39 have in the literature been suggested to interact and sequester nucleic acids) would be expected to affect the 20 viability when expressed and retained within the cell. The re sults are shown in Figs. 5a - 14a. Growth Inhibition of E. coli on Solid Media upon Expression of Various AMP's The following experiment was conducted in order to evalu 25 ate whether E. coli, upon induction of endogenous AMP expres sion, was growth-inhibited on solid media. On day one, cells harboring AMP-encoding plasmids, were inoculated in LB + 100 y ampicillin and grown under non-induced conditions at 37)C with vigorous shaking. These overnight cul- WO 00/73433 PCT/DKOO/00287 37 tures were then on day two diluted 100-fold into fresh, pre warmed LB broth + 100 y ampicillin, and 3 ptl spotted onto LB agar plates containing 100 y ampicillin and varying amounts of arabinose (0%; 0.001%, 0.01% and 0.1%). These agar plates were 5 then incubated over night at 37 0 C. The growth bacterial clones were recorded the following day. Inhibition of growth was ob served to correlate with the amount of arabinose present. The pattern of inhibition reflected the results observed when grown in liquid broth. 10 Mutant libraries In order to examine whether the suicide expression system is able to distinguish among peptide variants displaying different antimicrobial activity, mutant libraries of the 9 AMP genes were generated using PCR and 0.5 mM MnCl 2 . For illustrative 15 purposes, randomly picked clones of mutant StyelinC clones have been selected and analyzed for inducer-dependent growth inhibition. Mutant AMP's have been identified with what appears to be altered bioactivity (Fig. 15) . A subset of the mutants were sequenced, and all clones displaying altered bioactivity 20 were found to be different from the wt (wild-type) in question.

Claims (18)

1. A method for screening a pool of nucleotide sequences to se lect a nucleotide sequence encoding a peptide, said method com prising: 5 a) ligating a plasmid with the pool of nucleotide se quences operably linked to an inducible promoter, so as to express a peptide which is an enzyme or a mature peptide of less than 100 amino acid residues, optionally linked to a signal peptide, 10 b) transforming host cells which are sensitive to the peptide with the ligated plasmids, c) screening the transformed host cells so as to select viable cells, d) cultivating the viable cells in the presence of in 15 ducer so as to induce expression of said nucleotide se quence, e) selecting cells according to the effect of the in ducer on cell growth, and f) recovering the nucleotide sequence encoding the pep 20 tide from the selected cells.

2. The method of claim 1 wherein the pool of nucleotide se quences is produced by random mutagenesis, by gene shuffling or by synthesizing degenerate genes.

3. The method of claim 1 or 2 wherein the peptide is an enzyme. WO 00/73433 PCT/DKOO/00287 39

4. The method of claim 1 or 2 wherein the peptide is a short peptide consisting of less than 100 amino acid residues, pref erably an antimicrobial peptide or an anti-tumor peptide.

5. The method of claim 4 wherein the peptide is an antimicro 5 bial peptide or an antimicrobial enzyme active on bacteria, and the host cell is a bacterial cell (particularly E. coli or Ba cillus) , a filamentous fungus (particularly Aspergillus) or a yeast cell (particularly Candida or Saccharomyces).

6. The method of claim 4 wherein the peptide is an anti-tumor 10 peptide, and the host cell is a mammalian cell, particularly a tumor cell.

7. The method of any of claims 4-6 wherein the peptide is a short peptide of less than 100 amino acid residues, and the ligation is such as to express the peptide without an extension 15 or with an N-terminal extension of 1-5 amino acids and option ally a signal peptide.

8. The method of any of claims 1-7 wherein the selection of steps c) and e) is performed using agar plates

9. The method of any of claims 1-7 wherein the selection of 20 steps c) and e) is performed in microtiter wells.

10. The method of any of claims 1-7 wherein the selection of steps c) and e) is performed using a FACS machine.

11. The method of any of claims 1-7 wherein the selection of steps c) and e) is performed by centrifugation or filtration. WO 00/73433 PCT/DKOO/00287 40

12. Use of the method of any preceding claim for finding and preparing a composition for treatment of the human or animal body.

13. The method of claim 2, wherein the gene shuffling includes 5 in vitro shuffling of homologous DNA comprising cleaving ho mologous template double-stranded polynucleotide into random fragments followed by homologously reassembling the fragments into full-length genes.

14. The method of claim 2, wherein the gene shuffling includes 10 formation of a library of recombined homologous polynucleotides constructed from input DNA templates and random DNA primers by induced template shifts during in vitro DNA synthesis.

15. The method of claim 14, wherein the random DNA primers are employed to randomly initiate DNA synthesis on the mutant DNA 15 templates that are to be combined.

16. The method of claim 2, wherein the gene shuffling includes using primers having less than 30 base pairs harboring the mu tations to be combined.

17. The method of claim 2, wherein the gene shuffling includes 20 synthesis of one or more degenerate DNA primers encoding an en tire gene.

18. The method of claim 1, wherein the gene sequence diversity includes libraries of cDNA's or randomly generated whole-genome DNA fragments as starting material. 25