GB2483075A - Derivatives of the calcium dependant antibiotic comprising altered acyl side chains - Google Patents

Abstract

The invention relates to derivatives of the Calcium Dependant Antibiotic variants CDA 1 b, CDA2a, CDA3a, CDA3b, CDA3b CDA4a, CDA4b, CDA2d, CDA2fb, CDA2fa, CDA5a and CDA6a which comprise either a 2,3-epoxybutanoyl fatty acid side-chain or a butanoyl fatty acid side-chain, in place of the wild-type CDA trans 2,3 epoxyhexanoyl fatty acid side-chain. The invention also relates to a method of synthesizing derivatives of the Calcium Dependant Antibiotic variants which comprise either a 2,3- epoxybutanoyl fatty acid side-chain or a butanoyl fatty acid side-chain in place of the wild-type CDA fatty acid trans 2,3 epoxyhexanoyl, comprising mutating a bacterial strain comprising the CDA biosynthetic gene cluster and so as to mutate fabF3 of the CDA biosynthetic gene cluster so that residue Phe107 of FabF3 is altered to either an isoleucine, a leucine or a valine residue.

Description

I

Engineered biosynthesis of novel CPA derivatives comprising modified fatty acid side-chains.

[001] The present invention was the result of a multidisciplinary project combining techniques and approaches from two laboratories operating in the fields of molecular microbiology and natural product chemistry.

[002] The present invention relates to compounds which are derivatives of the Calcium Dependant Antibiotic (CDA) of Streptomyces coelicolor. The present invention also relates to a method of making these compounds by genetically modifying Streptomyces coelicolor so that the mutant strains synthesize modified forms of CDA.

[003] The model streptomycete, Streptomyces coelicolor, makes several antibiotics. The first of these to be described was actinorhodin, a red-blue acid base indicator pigment first described in the Gottingen strain of S. coelicolor [Brockmann and Hieronimus, 1955] and later studies in Streptomyces coelicolor A3(2) mutants [Wright and Hopwood, 1 976a; Rudd and Hopwood, 1979]. Undecyprodiginine, a red pigmented antibiotic, first described in other strains was also described in S. coelicolor A3(2) (Rudd and Hopwood, 1980). The antibiotic Methyleneomycin first described in S. violaceoruber (a close relative of the A3(2) strain) was also identified in A3(2) derivatives carrying the SCP1 plasmid either autonomously (SCP1+) or chromosomally integrated (the NF state) [Wright and Hopwood, 1 976b] and was shown to be specified by the plasmid.

[004] In 1983 a fourth antibiotic produced by Streptomyces coelicolor was discovered, known as the Calcium Dependant Antibiotic on account of it requiring the presence of Ca2 ions for antibiotic activity [Hopwood and Wright, 1983; Lakey et ai., 1983]. On the basis of similarities between the action of CDA and amphomycin and daptomycin (LY146032) on biological membranes [Lakey et al., 1983; Lakey and Ptak, 1988] it was thought that CDA may be a lipopeptide comprising a non-ribosomally synthesized peptide moiety [Hopwood, et al., 1995]. This was proven later by the chemical characterization of CDA [Kempter et al., 1997]. It has antibiotic activity against Gram-positive bacteria, for example Micrococcus luteus and Bacillus mycoides both of which are commonly used in bioassays for CDA, but is ineffective against Gram-negative species. Unlike actinorhodin and undecyprodiginine the Calcium Dependant Antibiotic is unpigmented and its synthesis can only be determined by bioassay, or by analytical techniques e, g. mass-spectrometry.

[005] The Calcium Dependant Antibiotic (CDA) of Streptomyces coelicolor [Hopwood and Wright, 1983; Kempter et al., 1997] belongs to the non-ribosomally biosynthesized acidic lipopeptide group of antibiotics which includes the A54 145 group of antibiotics which are produced by Streptomycesfradiae [Boeck, et al., 1990; Fukuda et al., 1990] the A21978C group of antibiotics, including daptomycin, (Debono et al., 1987) which are produced by Streptomyces roseosporus, molecules of the friulimicin and the amphomycin families [Vertsey et al., 2000] which are produced by Actinoplanesfriuliensis, and the laspartomycin/glycinocin group [Borders et al., 2001; Kong & Carter, 2003]. Table 1 provides further details regarding the lipopeptide antibiotics and the Streptomyces species which produce them.

Table 1

Compound Producer Daptomycin (A2 1 978C) Streptomyces roseosporus NRRL 11379 A54 145 Streptomycesfradiae NRRL1 8158 __________________________ Streptomyces refuineus ssp thernoto1erans CDA Streptomyces coelicolor A3(2) _________________________ Streptomyces viuolczceoruber Kutner 673 ________________________ Streptomyces lividans Amphomycin Sireptoinyces canus ATCC 12237 Zaomycin Streptomyces zaomyceticus ATCC 27482 (lshida N-187) Crystallomycin Streptornyces griseoflavus DSM40709 Aspartocin (A-8999) Streptomyces griseus var. spiralis ATCC 13733 ________________________ Streptoniyces violaceus var. aspartocinius ATCC 13734 Glumamycin Streptomyces zaomyceticus ATCC 13876 Tsushimycin Streptoniyces pseudogriseolus ATCC 211939 ______________________ Streptomyces pseudogriseolus ATCC 211940 Parvu line Streptomyces parvulus var. parvuli NRRL 5740 Friulimicin Actinoplanesfriuliensis DSM 7358 A-1437 Actinoplanes sp( Actinoplanesfriuliensis DSM 7358) Laspartomycin Streptomyces viridochrornogenes var. komebensis ATCC 29814 (M307-M5) Glycinocin Actinomycete AW998 [006] The members of the acidic lipopeptide group of antibiotics all have the common feature of requiring the presence of calcium ions to varying degrees for their antibiotic activity. The structures of the amphomycins, friulimicins, laspartomyins/glycinocins, A21978C, A54 145 and CDA also share many points of similarity (Figure 1). They all comprise cyclic decapeptide structures (amphomycins, friulimicins, & laspartomyinlglycinocins) or cyclic decadepsipeptide structures (A21978C, A54145 and CDA) with exocyclic tails of between 1 and 3 amino acids in length, coupled to a fatty acid side-chain (Figure 1). All of the antibiotics are produced as complex mixtures of slightly differing variant forms of the antibiotics. The source of this variation are the differing lipid side-chains present for the amphomycins, friulimicins, laspartomyins/glycinocins, A21978C and A54 145. A source of variation for the amphomycins, friulimicins, A54 145 and CDA are differences in the amino acid composition (Figure 1).

[007] The decapeptide/depsipeptide rings all have the same, or very similar, amino acid residues at a number of common positions. Using the A54145/ daptomycin amino acid residue numbering system as a common reference system these are Asp at position 7, Asp (or a modified Asp) at position 9, Gly in position 10, followed by a D-amino acid at position 11 and D or achiral residues at the fifth and eigth positions. Furthermore, the depsipeptides A54 145, CDA and A21978C also contain Threonine (Thr), and either 3-methyiglutamic acid, or glutamic acid, at equivalent positions.

Similarly, the amphomycins, friulimicins, laspartomyins/glycinocins all comprise D-Pip and proline residues in equivalent positions. The exocyclic residues ofA2l978C and A54145 are also similar there being a D-Tryptophan at position 1 in both molecules in addition to an acidic residue, and either asparagine (Asn), or hydroxyasparagine (hAsn).

[008] The lipopeptide antibiotics in addition to comprising proteogenic amino acids also comprise unusual, non-proteogenic amino acids. For example, in addition to comprising the D form of tryptophan at position 3 CDA contains four unusual amino acids:-D-4-hydroxyphenylglycine (position 6), D-3 -phosphohydroxyasparagine (position 9), L-3 -methyglutamate (position 10) and Z-2,3 dehydrotryptophan (position 11) [Kempter et al., 1997; Hojati eta!., 2002]. The residues present at positions 9, 10 and 11 are not fixed and a number of differingly modified residues may be substituted at each of these positions (Figure 2). CDA may be therefore be considered to be a heterogeneous group of molecules comprising variants of the same basic core molecule which differ in the degree to which the amino acids at positions 9, 10 and 11 are derivatized.

There are seven naturally occurring forms of CDA:-CDA1b, CDA2a, CDA3a, CDA3b, CDA4a and CDA4b (Figure 3).

[009] CDA1b, CDA2b, CDA3b and CDA4b differ in their substitution patterns at position 9, (either D-3-phosphohydroxyasparagine or D-3- hydroxyasparagine) and at position 10, (either L-glutamic acid or L-3-methylgiutamic acid) [Kempter eta!., 1997] (Figure 3). CDA2a, CDA3a and CDA4a comprise Z-2, 3 dehydrotryptophan instead of L-tryptophan which is found at position 11 in the other forms of CDA [Hojati et al., 2002] (Figure 3).

[0101 Different strains of Streptomyces coelico!or produce differing amounts of each of the CDA variants [Hoj ati et al., 2002]. Additionally, the nature of the CDA forms synthesized also depends on the growth media used for culturing the CDA producing strain. For example, the Streptomyces coelicolor strain 2377 produces CDA2a and CDA4a only in small amounts in liquid culture whereas when grown on solid ONA media [Kieser et al., 2000], they are the major forms of CDA synthesized [Hojati et al., 2002]. It is likely that CDA2a which represents the most structurally modified form of the molecule constitutes the ultimate product of CDA biosynthesis.

[011] In addition to the naturally occurring forms of CDA several non-naturally occurring CDA derivatives have been produced using a mutasynthesis approach. This involved the deletion of the 4-hydroxymandelate synthase encoding gene hmaS from the CDA biosynthetic gene cluster to produce a mutant strain which was unable to synthesize the non-proteogenic amino acid L-hydroxyphenyl glycine (L-HPG). The hrnaS mutant was therefore unable to incorporate the D-isomer of HPG at position 6 and CDA was not produced. However, the mutant was complemented by the feeding of racemic 4-hydroxymandelate, 4.-hydroxyphenylglyoxylate or L-HPG which restored CDA biosynthesis and CDA2b was produced, indicating that the deletion of hrnaS did not affect the expression of the surrounding genes, which are also involved in CDA biosynthesis. CDA production was however not restored by the feeding of D-HPG indicating that L-FIPG is first activated and added to the growing CDA peptide before being epimerized to the D isomer. A series of synthetic mandelate, aryiglyoxylate and aryiglycine analogues were also fed to the hmaS mutant strain and production of CDA derivatives were screened. Following the feeding of 4-fluoro and 4-dehydroxy forms of arylglycines three novel forms of CDA were produced. These comprised CDA2d which comprises a phenyiglycine residue at position 6 and major and minor products CDA2fb and CDA2fa which both have a 4-fluorophenyiglycine residue at position 6, and differ in that the minor product CDA2fa has a dehydrotryptophan residue at position 11, and the major product, CDA2fb, has a tryptophan residue at this position [Hojati et a!., 2002] (Figure 4).

[012] Two further non-natural forms of CDA were made by Streptomyces coelicolor strains MT1 110 and 2377 comprising deletions of the asnO gene (SC03236) [Neary et al., 2007]. This gene encodes an asparagine oxygenase which catalyzes the 3-hydroxylation of Asn to generate the 3 -OHAsn found in the naturally occurring forms of CDA, CDA3a, CDA3b, CDA4a and CDA4b. The S. coelicolor 2377 AasnO strain made a novel form of CDA, CDA5a, and the S. coelicolor MT1 110 JasnO strain made a novel form of CDA, CDA6a. CDA5a is a non-hydroxylated version of CDA3a, and CDA6a is a non-hydroxylated Asn containing variant of CDA4a, to which it is related by loss of the oxygen [Neary et al., 2007] (Figure 5). Deletion of the 3-hydroxyasparaginyl phosphotransferase gene, hasP (SC03234) in S. coelicolor MT 1110 and S. coelicolor 2377 did not result in them generating any novel forms of CDA. However, the deletants did not produce any phosphorylated forms of CDA, so confirming the function of HasP. These studies indicated that phosphorylation of the CDA Asn residue at position 9 by HasP occurs after it is incorporated into the peptide ring. Other studies have shown that Asn is hydroxylated prior to it being incorporated into CDA [Strieker et al., 2007], and Neary and co-workers also suggested that both Asn, and OH-Asn (generated by AsnO), are acceptable substrates for incorporation into the peptide ring [Neary et a!., 2007].

[013] The cyclic decapeptide portions of the lipopeptide antibiotics exhibit conservation of amino acid composition and position. For example, there are conserved residues in the common 10 amino acid residue lactone rings of CDA and daptomycin. These conserved amino acid residues include acidic amino acids and in particular aspartic acidlaspartate residues. The presence in all of the acidic lipopeptide antibiotics of a conserved aspartate residue at position 7, and a second conserved aspartate, or a modified Asp at position 9 (using the A54145/A21978C amino acid residue numbering system) is due to these residues being necessary for the antimicrobial action of the molecules. Deletion of these residues has been shown to abolish the antimicrobial activity of lipopeptide derivatives which lack these residues [Grunewald et al., 2004b].

[014] Two nuclear magnetic resonance (n.m.r.) structures of daptomycin (a specific form ofA21978C comprising the constant A21978C peptide core coupled to a decanoic acid (C10) side-chain) have been determined [Ball et al, 2004; Jung et al., 2004; reviewed in Micklefield, 2004] (Figure 6).

Structures were obtained for the molecule in the absence and presence of calcium ions and both report that daptomycin binds Ca2 in a 1.1 ratio. Ball et al., do not report a conformational change in daptomycin on binding Ca2 however, they do report Ca2 binding induces aggregation of the daptomycin molecules. The results of Jung et al., are also consistent with the oligomerization of daptomycin on binding Ca2. Moreover, Jung et al., also report a conformational change occurs in daptomycin on it binding Ca2, with Gly-5 functioning as a flexible hinge allowing the decapeptide core to draw inwards. It is likely that this movement allows Asp.-3 and Asp-.7 to coordinate Ca2. The result of this conformational change is, in addition to reducing the net charge of the molecule and constraining movement of the depsipeptide core, to increase the amphipathicity and area of solvent exposed hydrophobic regions. These effects, induced on binding Ca2 result in interaction of, and penetration of daptornycin into neutral or acidic phospholipid membranes leading to an increase in the flip-flop of phospholipids between leaflets. The deeper penetrative effect which occurs on binding Ca2 has been noted by investigators prior to Jung et al., 2007 [Lakey and Ptak, 1988]. Given the location of the fatty acid within the structure of daptomycin it is likely that the function of the lipid side-chain of the lipopeptide antibiotics is to mediate the interaction/penetration of cell membranes. Jung et al., also suggest a second conformational change occurs that promotes deeper penetration into, and perturbation of, the membrane. A specific interaction ofA2l978C with polar phospholipid headgroups has previously been suggested [Lakey et al., 1989]. The precise role of daptomycin oligomerization on binding Ca2 is at present unclear.

Amphomycin has also been shown by circular dichroism (C.D.) analysis to undergo a conformational change on binding Ca2 which could involve formation of a beta-hairpin structure which could enhance lipid bilayer solubility [Lakey et a!., 1988].

[015] An X-ray crystal structure of tsushimycin complexed with Ca2 has been determined to lÀ resolution [Bunkoczi, et al., 2005]. This indicates that the tsushimycin peptide core forms a saddle-like structure equipped with a long acyl side-chain with a Ca2 ion bound in the centre of the saddle by four backbone carbonyl groups and the side-chains of two Asp residues (Asp-i & Asp-5). A second Ca2 ion is bound to the surface of each tsushimycin molecule and together with reciprocal hydrogen bonds between the Asp-7 the Dab-9 residues of adjacent tsushimycin molecules participates in the formation of tsushimycin dimers. It is at present unclear whether daptomycin binds Ca2 in a similar fashion to tsushimycin. The dimers form a tunnel like structure which is large enough to contain a substrate and this is thought to be the biologically active form of the molecule. As the dimers expose highly polar surfaces to the solvent they are thought to interact with bacterial cell membranes via their fatty acid side-chains [Bunkoczi, et a!., 2005] [016] The precise nature of the mechanism, or mechanisms, by which the lipopeptide antibiotics exert their killing effect/s is/are unknown. Most studies have focused upon investigation of the mechanism of action of the clinically used antibiotic daptomycin. It was found that at high concentrations (i.e. 25pg/ml-25-50 times M.I.C.) daptomycin inhibited peptidoglycan synthesis in Staphylococcus aureus and similar concentrations prevented uptake of peptidoglycan precursors in Bacillus megaterium (Allen et al., 1987). It was suggested that daptomycins target was the plasma membrane and the dissipation of the cell membrane potential inhibited the uptake of cell wall precursors and so inhibited peptidoglycan synthesis [Albom et al, 1991; Allen eta!., 1991]. At the M.I.C daptomyin bound to the cell wall and membrane in the presence of Ca2 and inhibited partial blockage of DNA, RNA, protein and peptidoglycan synthesis in E. faeciurn [Canepari et a!., 1990] in addition to inhibiting incorporation of phosphate into teichoic acid and glycerol into lipoteichoic acid (LTA) [Canepari et a!., 1990]. Daptomycin was also shown to cause 82% inhibition of LTA synthesis in protoplasts, 40% inhibition of lipid biosynthesis and 14-23% inhibition of DNA, RNA and protein biosynthesis and to tightly bind the cell membrane in the presence of Ca2 [Boaretti et a!., 1993]. It was found that at the M.LC daptomycin bound to five proteins and the authors suggested that LTA biosynthesis was the target of daptomycin [Boaretti and Canepari, 1995]. Four of the five daptomycin binding proteins have subsequently been purified but their functions are unknown [Boaretti and Canepari, 2000]. The addition of purified LTA did not inhibit the activity of daptomycin against S. aureus or E. faecalis indicating the killing effect of daptomycin is not due to it binding LTA [Laganas et al., 2003].

[017] It was shown by Silverman and coworkers [Silverman et al., 2003] that the kinetics of daptomycin induced dissipation of cell membrane potential and loss of cell viability are closely correlated and were coupled to the loss of K ions from the cell. It was suggested that binding of Ca2 induced daptomycin to insert into membranes and oligomerize so forming pores in the bacterial plasma membrane leading to loss of IC ions, membrane depolarization and ultimately cell death (Silverman, et al., 2003).

However, it is clear from the work of Jung et al., that any depolarization of the plasma membrane that occurs due to daptomycin treatment of bacterial cells occurs after cell death. It is likely that the killing action of the Ca2 -bound from of daptomycin is due to its insertion into, and non-specific general disruption of membrane function, as opposed to inhibition of a single specific molecular target [Jung et al., 2004; Hancock, 2005].

[018] The mechanism of action of other lipopeptide antibiotics have been less well researched than that of daptomycin, although given their similar structures and common requirement for Ca2 for antibiotic activity it seems likely that they exert their killing effects through similar/identical mechanisms. Daptomycin and A54 145 exhibit little or no antimicrobial activity in the absence of Ca2 but exhibit their maximum antibacterial activity at a concentration of 50mg/i [Elipoulos et al., 1985; Elipoulos et al., 1986; Counter et al., 1990] i.e. the concentration of Ca2 normally found in human serum [Barry et a!., 2001]. Friulimicin B is less Ca2 dependant than is daptomycin [Bemer et al., 2003] and CDA has a shallower dependence curve and 16mM Ca2 is required for maximum potency.

[019] Given the role of the fatty acid side-chains of the lipopeptide antibiotics in the penetration of and/or associationlinteraction with cell membranes it is interesting to consider the various lipid moieties which the lipopeptide antibiotics may comprise, together with their natural variability.

[20] Although amphomycin was first described over 50 years ago [Heinemann et al., 1953] its structure and relationship to a large number of obviously related compounds with similar properties which have been variously reported under different names, i.e. zaomycin [Hinuma, 1954], crystallomycin [Guaze eta!., 1957], tsushimycin [Shoji et al., 1968; Shoji and Otsuka, 1969], aspartocin [Shay et al., 1960; Hausmann eta!., 1963; Hausmann et al., 1969], glumamycin [Shibata eta!., 1962; Inoue, 1962; Fujino, 1965, Fujino eta!., 1965], laspartomycin [Naganawa et al., 1968; Naganawa et a!., 1970] and parvuline [US3798 129] were not determined until relatively recently. By varying the fermentation conditions it was found that Actinoplanesfriuliensis produced two distinct groups of lipopeptide antibiotics, one of which was the amphomycins. The amphomycin group of compounds -i.e. A-1437 A20, B21, E24 and G26 share a common decapeptide ring structure linked via the N-terminal diarninobutyrate to an exocyclic Asp residue to the fatty acid side-chains A3-iso-tridecenoic, A3-iso-tetradecenoic A3 -anteiso-tridecenoic and \3 -anteiso pentadecenoic acid, respectively [Vertsey et a!., 2000]. It was found by comparing results with the prior art literature that A-1437 E24 was identical to amphomycin, A- 1437 B21 was identical to the major component of tsushimycin and A-1347 G26 was identical to the major component of aspartocin whilst A-1437 A20 was identified with parvuline which comprises L\3-iso-tridecenoyl.

[0211 The second group of lipopeptides produced by Streptoinyces friu!iensis is the friulimicins (A-D, 22, 23, 25 and 27) [Vertsey et al., 2000] which are structurally very similar to the amphomycins possessing the same fatty acid side-chains and differing only in the peptide core in that they possess an Asn residue instead of an Asp as the exocyclic amino acid. Production of particular forms of the amphomycins and friulimicins may be enhanced by feeding different amino acid precursors of fatty acids. For example, feeding leucine favours the production of A-1437 A and friulimicin A, feeding valine favours the production of A-1437 B and friulimicin B and feeding isoleucine favours the production ofA-1437 E and G, and friulimicins C and D. [022] It was shown that the structure of laspartomycin C 29 placed this compound in the glycinocin 29-32 group of lipopeptides [Kong and Carter, 2003] which are related to the amphomycins, although there differences in both amino acid composition and in fatty acid side-chains. Glycinocins 20- 32 comprising 2-iso-pentadecenoy1, i.2-iso-hexadecenoyl, t2 -iso-tetradecenoyl and, 2-iso-pentadecenoyl, respectively. The position of unsaturation in the fatty acid side-chains shifts to create, a, 3, unsaturated amides.

[023] The structures of the three major components of the A21978C complex produced by Streptomyces roseosporus were determined by incubating the purified compounds with Actinoplanes utahensis which produces a deacylase capable of removing fatly acid groups and then characterizing the peptide cores by amino acid composition analysis, followed by Edman degradation and GC-MS analysis of derivatized peptides generated by partial hydrolysis. The peptide cores of all three compounds were found to be identical and the branched chain fatty acyl groups attached to the N-terminus of the peptide core were found to be anteiso-undecanoyl, iso dodecanoyl and antesio tridecanoyl for A2 1 978C 1-3 respectively. The minor constituents A21978C4 and A21978C5 both comprise C12 fatty acid side-chains whilst A2 1 978C0 is a mixture comprising both linear and branched C10 fatty acids. The proportions of the various forms of A2 1 978C may be influenced by media composition and the addition of extra valine to the medium causes an increase in the amount ofA21978C2 synthesized whilst the addition of isoleucine leads to increases in A21978C 1 & 3. The additiOn of leucine resuhs in the fonation f isomers of A21978C1 & A2 1 978C3 comprising iso-undecanoyl, and iso-tridecanoyl side-chains respectively.

[024] A54 145 from Streptomycesfradiae is also a complex mixture of lipopeptides which comprise three different fatty acid side-chains-iso-decanoyl, n-decanoyl and anteiso-undecanoyl [Boeck et al., 1990]. Feeding S. fradiae with valine or isoleucine resulted in greater synthesis of forms comprising branched chain fatty acids [Boeck and Wetzel, 1990].

[025] CDA consists of a cyclic lactone undecapeptide comprising a 10 residue membered ring structure with a C6 fatty acid side-chain, trans 2,3 epoxyhexanoic acid, attached to the N-terminus of the exocyclic serine residue [Kempter et al., 1997]. Uniquely amongst the lipopeptide antibiotics all of natural occurring forms of CDA (CDA1b, CDA2a, CDA3a, CDA3b, CDA4a and CDA4b) comprise a single, invariant species of fatty acid moiety, which at six carbon atoms in length is considerably shorter than the acyl side-chains of the other lipopeptide antibiotics. The three modified CDA derivatives (CDA2d, CDA2fa & CDA2fb) synthesized as a result of the mutasynthesis experiments described in paragraph 011 [Hojati et al., 2002] also comprise the standard trans 2,3 epoxyhexanoic acid side-chain.

Moreover, the two modified CDA derivatives (CDA5a & CDA6a) synthesized as a result of the experiments described in paragraph 012 [Neary et al., 2007] also comprise the standard trans 2,3 epoxyhexanoic acid side-chain.

[026] The chromosomal location of the cluster of genes responsible for the biosynthesis of CDA was first identified by using degenerate probes against conserved motifs of non-ribosomal peptide synthetases and targeted disruption of the NRPS genes demonstrated their role in CDA biosynthesis [Chong et al., 1998]. The gene cluster was partially sequenced [Chong et al., 1998] and subsequently the sequence of the whole cluster (and indeed the entire Streptomyces coelicolor A3(2) M145 genome) was determined [Bentley et al., 2002] with the CDA biosynthetic cluster being located on cosmids E8-E64-E29 [Redenbach et al., 1996]. The CDA biosynthetic cluster is in the "core" region of the Streptomyces chromosome and comprises -.82kb of DNA (approximately 1% of the entire genome). It consists of at least 40 ORFs from SC03210 to SC03249 (Figure 7) although it is unclear how many of these ORFS are actually involved in CDA biosynthesis. The majority of the genes present in the CDA cluster have either been shown to play specific roles in CDA biosynthesis or have putative roles assigned to them based on homology studies with other natural product biosynthetic gene clusters [Hojati et a!., 2002] (Figure 8).

However, it has been difficult to envisage or assign putative roles for the gene products of SC03237-SC03244 in CDA biosynthesis. The limits of the gene cluster have been confirmed by microarray analysis of the genes expressed during CDA production [Huang et al., 20011 although the involvement of proteins encoded by genes outside of the gene cluster in CDA biosynthesis has not been ruled out.

[027] Within the CDA biosynthetic gene cluster, at the 3' end, is an operon of five genes (SC03245-3249) which have been shown to be involved in synthesis of the CDA specific fatty acid, trans 2,3 epoxyhexanoyl [Hojati et al., 2002] [028] SC03249 encodes a protein with sequence similarities to acyl-carrier proteins (ACPs) from fatty acid synthases (FAS) and polyketide synthases (PKS) and is thought to transfer the acyl chain between the enzymes involved in synthesis of trans 2,3 epoxyhexanoyl [Hojati et a!., 2002; Powell et a!., 2007; Kopp et al., 2008].

[029] SCO3 246 (fiibH4) encodes a protein similar to 1 -ketoacyl-ACP synthase KAS-IJI enzymes, and is responsible for initiation of fatty acid synthesis {Lai, and Cronan, 2003]. FabH4 catalyzes a Claisen-type condensation of acetyl and malonyl units to generate acetoacyl-S-ACP which is presumably processed by the primary metabolic fatty acid synthase (FAS) to give butanoyl-S-ACP. OtherfabH KAS-ilI type genes are found in Streptomyces coelicolor [Florova, and Reynolds, 2005], however, fabH4 is specific for the synthesis of trans 2,3 epoxyhexanoyl.

[030] The 1 -ketoacyl-ACP synthase KAS-JI enzyme encoded by SC03248 (fabF3) is responsible for the subsequent elongation of the fatty acid chain by catalyzing an additional condensation reaction between the butanoyl-S-ACP and malonyl-CoA to generate 3 -ketohexanoyl-ACP. This C6 lipid is then modified by oxidation i. e. introduction of an unsaturated double bond, followed by its epoxidation to form the final product trans 2,3 epoxyhexanoyl. -The-oxidation and expoxidation reactions are effected by the proteins encoded by the two remaining genes of thefab operon, SC03245 & SC03247.

[031] The hxcO gene (SC03247) encodes a protein which has sequence similarity to FAD-dependant acyl.-CoA oxidases and dehydrogenases, enzymes responsible for the desaturation of their thioether substrates in fatty acid catabolism. Streptomyces coelicolor mutant strains comprising a deletion of the hxcO gene have been shown to produce modified forms of CDA (hCDA4a & hCDA4b) comprising a non-native hexanoyl fatty acid side-chain (Figure 9). This suggested the hxcO encodes a hexanoyl-CoA oxidase responsible for producing the desaturated trans-2,3-hexenoyl -CoA or ACP prior to alkene epoxidation, and in its absence the hexarioyl precursor accumulates and is directly transferred to the serinyl substrate tethered to the peptidyl carrier protein (PCP) domain of the first peptide synthases (CDAPSI) [Powell, eta!., 2007]. The fact that purified HxcO protein was unable to oxidize hexanoyl-CoA suggested that its preferred substrate was the ACP bound form of the fatty acid [Powell, et a!., 2007].

[032] This surmise was shown to be true by other researchers [Kopp et al., 2008]. In vitro studies using purified HxcO protein showed that that HxcO is unable to accept as substrate free fatty acids and only accepts those bound to an acyl carrier protein. The loading of free fatty acids onto the ACP in vitro was achieved using the 4' phosphopantetheine transferase (PPTase) (Sf) from Bacillus subtilis. These studies also indicated that in addition to being a FAD dependant acyl-ACP oxidase HxcO possesses an intrinsic epoxidase activity as HxcO was able to catalyze the synthesis of 2,3-hexanoyl-S-ACP as a major product, producing only minor amounts of hex-2-enoyl-S-ACP. It seems that HxcO generates FADH2 in situ by oxidation of the fatty acid substrate and that the reduced cofactor then used as the reactive species in the epoxidation reaction. Further in vitro investigation of the substrate specificity of HxcO indicated it was able to accept as substrates a range of ACP bound substrate analogues, namely linear fatty acids of 4-10 carbon atoms, and produce the corresponding enoyl species [Kopp et al., 2008].

[033] The hmcO gene (SC03245) encodes a protein which has sequence similarity to a diverse class of flavin dependant monooxygenases, particularly salicylate hydroxylases and zeanthin epoxidases. p-hydroxybenzoate and salicylate hydroxylases. In vivo gene deletion studies of hmcO indicated that the Streptomyces coelicolor thmcO mutant strain did not produce any CDA [Powell et a!., 2007]. In vitro studies using the purified HcniO protein indicated that it was able to epoxidate the presumed natural hexenoyl substrate bound to ACP by Sfp and like HxcO it was unable to accept substrates not bound to the ACP as HmcO was unable to epoxidate hexenoyl-CoA [Kopp et al., 2008]. HmcO was also unable, in vitro, to epoxidate an unsaturated hexanyol moiety bound to ACP and in studies designed to assess the substrate specificity of HcmO with various enoyl-ACP substrates although variations in the position and configuration of the double bond were not tolerated a crotonyl-S-ACP with a chain length shorter than that of the physiological C6 substrate was accepted [Kopp et al., 2008].

[034] HcmO was unable to epoxidate chemoenzymatically generated hexa- 2-enoyl -CDA and HxcO was unable to oxidize chemoenzymatically generated hexanoyl-CDA indicating that the mature CDA molecule is not a substrate for these enzymes. Presumably the trans 2,3 hexanoyl fatty acid is loaded onto CDA, rather than hexanoyl-CDA being synthesized and later oxidized and epoxidated by HxcO and HcmO respectively.

[035] There are differences between the in vivo and in vitro results regarding HxcO and HcmO and it is difficult on the basis of the data available at present to conclusively define the roles and precise substrates of these enzymes in trans 2,3 epoxyhexanoyl biosynthesis. In particular the relevance of the in vitro observations regarding HxcO epoxidation activity are questionable and it may be that under physiological conditions where all of the of trans 2,3 epoxyhexanoyl biosynthetic enzymes are present in a complex with other members of the FAS complex this does reaction not occur with the oxidized hexenoyl-ACP produced by HxcO being passed onto HcmO for epoxidation (Figure 10). However, various other scenarios have been envisaged to account for the in vitro results [Kopp et al., 2008].

[036] In contrast to CDA the other lipopeptide antibiotics, for example daptomycin and A54145, do not comprise a constant fatty acid side chain but instead the fatty acid is a variable substituent. In the case ofA2l978C (daptomycin) the group of lipopeptide compounds initially identified were found to possess a common cyclic peptide core with different fatty acid acyl groups attached by an N-acetyl bond [Debono et al., 1980; Debono et al., 1984].

[037] The reason for the heterogeneity of A54145 and A21978C with regard to their fatty acid moieties may be explained by reference to their mechanisms of biosynthetic mechanisms which differ from that of CDA in this regard. In contrast to thefab operon present in the CDA biosynthetic cluster the biosynthetic gene clusters responsible for the synthesis of daptomycin and A54 145 do not contain genes encoding enzymes responsible for the biosynthesis of fatty acids. Instead these clusters comprise genes encoding proteins which are responsible for the activation of free fatty acids present in the cytoplasm, and their loading onto the growing peptide moiety by an ACP. In the daptomycin biosynthetic gene cluster the fatty acid activating enzyme is encoded by dptE and the fatty acid loading protein, known as the ACP (acyl carrier protein) is encoded by dptF [Wittmann et al., 2008]. dptE and dptF form an operon and are co regulated and co-expressed as a di-.cistronic transcript from the dptE promoter. DptE shares -20% sequence identity with several members of the acyl AMIP/CoASH (coenzyme A) ligase superfamily of enzymes which catalyze the formation of fatty acyl AMP/CoASH from a fatty acid substrate, ATP and CoASH in a Mg2 dependent two-step reaction. A fatty acyl adenylate intermediate is formed in the first step, followed by conversion of the fatty acyl adenylate to fatty acid-CoA with the release of AMP [Wittmann et al., 2008]. The specificity of the DptE-DptF interaction was also demonstrated with DptE being unable to productively interact with the LipD (ACP) from the friulimicin biosynthetic gene cluster of A ctionoplanes friuliensis [Wittmann et a!., 2008].

[038] However, it was shown that DptE first activates a fatty acid by adenylation so forming a fatty acyl AMP and then DptE acts as an acyl ACP synthetase and the fatty acyl AMP is transferred onto DptF for loading onto the peptide core [Wittmann et al., 2008]. In the S. fradiae A54 145 biosynthetic gene cluster [Miao et a!., 2006] the iptEF gene encodes a bi-domain protein comprising an adenylating enzyme domain coupled to an ACP (acyl carrier protein) domain which fulfils a role similar to that played by both DptE and DptF in S. roseosporus.

[039] DptE/F and LptEF are fairly "promiscuous" enzymes which are able to adenylate a number of free fatty acids present in the bacterial cytoplasm and load them onto the peptide moieties of the lipopeptide antibiotics. The substrate specificity of DptE has been investigated and unsurprisingly it has a preference for fatty acids found naturally in the A21978C lipopeptides i.e. linear fatty acids with chain lengths of between 8 and 14 carbon units, particularly iso/anteiso branched chain fatty acids and decanoic acid, with those fatty acids not naturally present, for example octanoic acid, tetradecanoic acid and the 3-hydroxy fatty acid, being poor substrates [Wittmann eta!., 2008]. However, it was found that long chain (16 carbon units or longer) and short-chain fatty acids (6 carbon units or shorter) are not accepted by DptE. This relaxed-specificity of DptE/F and LptEF with regard to substrate explains why the A54145 and A21978C antibiotic complexes each comprise a group of compounds which possess different fatty acid side-chains.

[040] In Actinoplanesfriuliensis an acyl-CoA dehydrogenase encoded by lipB is involved in the formation of a double bond in the fatty acid side-chain of Friulimicin [Heinzelmann, et al., 2005] and a small ORF (orfl2) which overlaps with the downstream terminus of lipB also has similarity to acyl-CoA dehydrogenases. It has been suggested that the putative acyl-CoA synthase LipA may be able to activate branched, and long chain fatty acids which are then unsaturated by LipB and then transferred via the putative ACP, LipD onto the peptide core [Heinzelmann, et a!., 2005].

[041] The N-terminal condensation (C') domains of all NRPS molecules involved in synthesis of lipopeptide antibiotics are different from the other C domains within the NRPS molecules as they perform the specific function of attaching the acyl side-chain onto the amino acid bound to the initiating NRPS module. It is likely that the substrate specificity of the initial C' domain is partly responsible for determining which species of fatty acid side-chains are incorporated into the lipopeptide antibiotics. It is likely that C' domain substrate specificity is responsible for the finding that when the daptomycin dpt, biosynthetic gene cluster was heterologously expressed in Streptomyces lividans the core A2 1 978C peptide was acylated with the usual complement of A2 1 978C fatty acids and not with the trans 2,3 hexanoyl fatty acid endogenously produced by the CDA biosynthetic enzymes [Penn etal., 2005].

[042] The substrate specificity of the CDAPSI C' domain was investigated using a mutasynthesis approach [Powell et al., 2007]. The active site serine (Serl 122) of the PCP domain of Module 1 of CDAPSI was altered using site directed mutagenesis (SDM), to an alanine residue and it was predicted that this Sen 122Ala amino acid substitution would abolish phosphopantetheinylation of the CDA domain, and so CDA production. In the absence of synthesis of the N-epoxyhexanoyl-L-serinyl-S-PCP intermediate on Module 1 exogenously produced N-acyl-L-seninyl-NAC analogues were fed to the mutant strain to investigate which, if any, compounds were able to complement the mutation and restore CDA production [Powell et al,, 2007].

[043] Feeding of N-hexanoyl-L-serinyl-NAC resulted in it being successfully incorporated into CDA to produce the same hCDA4b produced by the hxcO mutant strain discussed above in paragraph 031 [Powell et al., 2007]. Similarly, feeding of N-pentanoyl-L-serinyl-NAC resulted in it being successfully incorporated into CDA to produce a novel pentanoyl lipopeptide derivative-pCDA4b [Powell et al., 2007]. However, feeding of the N-decanoyl-L-serinyl-NAC and N-heptanoyl-L-serinyl-NAC did not result in the production of new derivatives of CDA and it was concluded that the substrate specificity of the C' domain was for acyl chains comprising C6 and shorter i.e. the CDA biosynthetic system had evolved to conserve the specificity for the CDA specific fatty acid [Powell et al., 2007].

[044] The heterogeneity of the A21978C with regard to the fatty acid moiety has proved to be a problem in the industrial production and purification of a single, well defined and characterized form of the antibiotic molecule for use in the manufacture of pharmaceuticals. Moreover, the finding that subtle changes in the fatty acid side-chain have considerable effects on the human toxicity of the various A21978C forms made it important that a single non-toxic form of A21978C comprising a single fatty acid side-chain could be produced.

[045] Following the discovery ofA21978C it was found that the form of A21978C comprising a decanoic acid side-chain resulted in a compound (LY146032: daptomycin) with superior biological activity to other compounds in the group possessing other fatty acids side-chains [Fukada et al., 1984; Counter et a!., 19841. The initial process for the synthesis and purification of daptomycin was lengthy and complex and involved two biological and three different recovery processes. The process basically involved the enzymatic removal of the natural acyl chains from the peptide nucleus which could then be re-acylated with a single fatty acid by chemical means. A group of Actinoplanes cultures were known to have penicillin deacylase activity [US3 150059] and an Actinoplanaceae enzyme was found to be capable of deacylating the A-30912 cyclic peptide antibiotic [US4293482, US4299764, US4299762, US4304716, US4293490]. One strain in particular, Actinoplanes utahensis NIRRL 12052, was found to be capable of deacylating A2 I 978C (US4399067).

[046] The process involved culturing Streptoinyces roseosporus in primary, secondary and tertiary inoculum development stages for 96 hours arid then in the producing stage for 140 hours. The microbes were separated from the soluble portion of the culture liquid by filtration and the antibiotic complex was adsorbed from the culture liquid, eluted from a resin and concentrated.

The free amino function of the complex was "blocked" with di-tert-butyl dicarbonate to give mono-ten'-butyloxycarbonyl (rBOC) and the mixture again concentrated. Actinoplanes utahensis was cultured for 120 hours in primary and secondary inoculum development stages and then grown for a further 72 hours in a stirred reactor before the "blocked" antibiotic complex was added to the medium for deacylation. After 24 hours the spent medium was filtered and the "blocked" nucleus was adsorbed and eluted from a resin column. The eluate was concentrated and the "blocked" nucleus was acylated with either the anhydride or halide of decanoic acid and then the protecting group removed by hydrolysis. The final product was adsorbed and eluted from a resin column and subjected to final purification. This process resulted in a low yield of daptomycin at a low concentration [US4399067; Huber eta!., 1988].

[047] A similar process where the de-acylated nucleus was re-acylated using activated esters of 2,4,5-trichiorophenol (TCP) and de-blocking of the Orn group using trifluoroacetic acid (TFA) has been used to make a range of A2 1 978C derivatives including fatty acyl derivatives, extended peptide derivatives and amino aroyl derivatives [Debono et al., 1988]. For fatty acid derivatives in vitro antibacterial activity increased with acyl side-chains up to C12-C13. In vivo antibacterial activity similarly increased but LD50 data indicated that there was increased toxicity when the length of the fatty acid was increased beyond C11. More detailed studies on the decanoyl, undecanoyl, N-decanoylphenylalanyl and N-dodecanoyl-p-aminophenacetyl derivatives ultimately led to selection of the decanoyl derivative, which had the best therapeutic index in mice, for clinical trial.

[048] It was found that the fatty acid decanoic acid could be fed to cultures of Streptomyces roseosporus and incorporated into A21978C so that approximately 34% of the complex comprised the decanoic acid form of the antibiotic i.e. daptomycin [Huber eta!., 1988]. As decanoic acid is toxic to Streptomyces roseosporus and forms a waxy solid at the cultivation temperature (30°C), carefully controlled, dispersed feeding [Huber et al., 1987] of a 1.1 formulation of decanoic acid and methyl oleate was necessary to achieve good results [Huber et a!., 1988]. This biosynthetic approach has greatly facilitated and simplified the industrial production of daptomycin.

[049] A number of derivatives ofA54145 have been synthesized using a similar approach to that described in paragraphs 046-047 [Fukada et al., 1990]. N-Lys-tBOC-protected A54 145 complex was de-acylated using A. utahensis and the protected nuclei of the A, B and F forms ofA54145 were isolated, the C form not being present in useable quantities. Acyl chain derivatives ofA54145 using fatty acids from n-hexanoyl to n-tetradecanoyl were produced by acylation of the Trp N-terminus with TCP esters of the appropriate fatty acids, or acyl chlorides. As for A21978C the length of the fatty acid side-chain was a significant determinant of biological activity [Counter et al., 1990] with the peptide nucleus and the hexanoyl derivative being inactive, and the in vitro and in vivo antibacterial effect increasing with increasing acyl chain length up to n-tetradecanoyl. It was found that for different A54 145 nuclei bearing identical fatty acid moieties the B form was the most active and the F form the least. In vivo studies of a number of forms of A54 1 45A containing different fatty acid side-chains have been conducted and they indicate the C11 side-chain was the most active. However, as its antibacterial activity and therapeutic index were 40 fold less than that of daptomycin this compound has not been pursued clinically.

[050] In a similar fashion to A21978C the incorporation of fatty acids into the lipopeptide antibiotics during feeding experiments was examined and toxicity was a problem necessitating slow, controlled feeding of lipids to bioreactor cultures. Ethyl caprate increased the proportion of decanoyl containing factors from the natural level of-' 14% to 80% and new factors containing C6, C8 and C9 resulted from feeding with hexanoate, caprylate and nonanoate respectively.

[051] A number of derivatives of amphomycin have been synthesized using a similar approach to that described in paragraphs 046-047. Protection of the free amine group of Dab-9 was followed by removal of the lipid side-chain using A. utahensis deacylase and replacement of fatty acid side-chains via N-succinimidyl esters [Sgarbi et al., 2005]. There was good correlation between antibacterial potency and chain length with C16 being the optimum, and amino acids such as Gly, and aryl groups, such as substituted benzoic acids, could be used as spacers between the fatty acid and peptide core.

[052] The deacylation of laspartomycin by A. utahensis deacylase [US651 1962; US6737403; US6750 199] resulted in two products:-the peptide core possessing and lacking the exocyclic Asp residue. Derivatives of both were made using aromatic acyl side-chains or varying the number and nature of the amino acids in the linking region. The presence of an aspartate residue in the linker was found to be critical and the most active compound produced had a pentadecanoyl-aspartyl or alkoxybenzoyl-aspartyl group coupled to the cyclic peptide nucleus. Further studies have indicated that the optimum length of the fatty acid side-chain is 15-17 carbon units.

[053] Generalizing from all of the available data relating to the correlation between lipopeptide antibiotic activity and acyl chain length it seems that for the depsipeptide antibiotics A21978C and A54 145 that potency increased up to an acyl chain length of C12-13 (A21978C) or C14 (A54145), although toxicity increased above C11. For amphomycin and laspartomycin the optimum acyl chain length for antibacterial potency was [054] There are no published studies of CDA activity in vivo and the molecule has poor in vitro activity. The antibacterial activities of the CDA derivatives comprising hexanoyl and pentanoyl side-chains [Powell et al., 2007], as opposed to the usual fatty acid moiety have not yet been investigated.

[055] There is a continual need for new antibiotics in view of the increase in infections caused by Gram-positive pathogens and the rise of antibiotic resistant strains. Of particular concern are the prevalence of VRSA, (vancomycin resistant strains of Methicillin resistant Staphylococcus aureus (MIRSA)) and vancomycin resistant Enterococcus species.

[056] Daptomycin was discovered by Eli Lilley in the early 1980's however, concerns relating to skeletal muscle toxicity at doses as low as 6mg/kg per day in clinical trials led to its abandonment. In 1997 Cubist Pharmaceuticals licensed the antibiotic and following clinical trials in September 2003 the US FDA approved daptomycin under the trade name "CubicinTM" at 4mg/kg per day for the treatment of complicated skin and skin structure infections (SSSI) caused by susceptible strains of Staphylococcus aureus (including MRSA), Streptococcus pyo genes, Streptococcus agalactiae, Streptococcus dysgalactiae subspecies equim ilis and Enterococcus faecalis (vancomycin susceptible strains only). It should be noted that although a successful antibiotic daptomycin is only currently used against a relatively small number of pathogens in a small number of clinical situations and that its high cost compared to vancomycin has also limited its general use.

[057] Following the introduction of daptomycin, the first clinical use of a lipopeptide antibiotic, it seemed to us that there was a need for further antibiotics of this class with superior properties to daptomycin. As discussed at length above in paragraphs 047-052 the structure of the fatty acid side-chain of lipopeptide antibiotic is a major factor in determining both the toxicity and antibacterial potency of the molecules and we therefore determined to generate CDA derivatives possessing altered acyl side-chains which may confer improved functional properties.

[0581 There are several reasons why we selected CDA as an object of study/research, as opposed to other lipopeptide antibiotics. Unique amongst the lipopeptide antibiotics CDA comprises a single, specific, fatty acid moiety and the specificity of fatty acid biosynthesis/loading onto CDA means that if we were able to alter this mechanism so as to synthesize/load a fatty acid different from trans 2,3 epoxyhexanoyl we would expect to generate a novel CDA derivatives comprising a single type of fatty acid as the side-chain. This approach would make synthesis/purification of the lipopeptide much more straightforward than the arduous, expensive and complicated approaches detailed in paragraphs 046-052 used to generate lipopeptides derived from A2 1 978C, A54 145 and the amphomycins and laspartomycin with defined fatty acid side chains by chemical modification/feeding methods. CDA was also chosen for this approach as it seemed likely to us that its poor antibiotic activity was due to it comprising an unusually short lipid side-chain, compared to the other lipopeptide antibiotics, and that if we were to increase the length of the acyl chain that antibacterial potency would likewise be increased.

[059] To investigate the reason for the unusually short C6 fatty acid side-chain of CDA the amino acid sequence of the putative KAS-Il enzyme encoded byfabF3 (SCO3 248) from the CDA biosynthetic gene cluster was aligned with two other KAS-Il paralogues from S. coelicolor, and other bacterial KAS-Il enzymes [Powell et al., 2007]. This revealed an acyl binding pocket similar to that of KAS-Il enzymes from Escherichia coli and Synechocystis sp whose structures had been elucidated by X-ray crystallography [Huang et al., 1998; Moche et al., 1999; Moche et a!., 20011. However, it was noted that uniquely amongst the enzymes examined fabF3 possessed a Phenylalanine at amino acid position 108 (E. coli residue numbering system) which equates to amino acid 107 of FabF3 [Powell et al., 20071. The other KAS-IT enzymes possess residues at this position which are smaller than phenylalanine and are commonly isoleucine or leucine (Figure 11). Interestingly, the X-ray crystal structures show that in the E. coli enzyme lie 108 is the middle of the hydrophobic acyl substrate binding pocket. Moreover, mutation of theE. coil KAS-Il lie 108 to Phe abolishes the specificity for longer chain acyl-ACP (>C8) whilst retaining specificity for the shorter chain hexanoyl-ACP i.e. the residue at position 108 has been shown to be involved in determining the size of the binding pocket available for fatty acid elongation [Val et al., 2000].

[060] On the basis of this we developed the hypothesis that Phe 107 in FabF3 serves to limit the size of the acyl binding pocket of the enzyme resulting in termination of fatty acid synthase (FAS) activity at the hexanoyl-ACP intermediate [Powell et al., 2007]. It seemed to us that alteration ofPhel07 of FabF3 to amino acids possessing smaller hydrophobic R-groups by site directed mutagenesis may serve to increase the size of the acyl binding pocket of FabF3 and so permit synthesis of fatty acid/s longer than C6.

Consequently, the >C6 fatty acid would presumably be modified by HxcO and HcmO and loaded on the CDA peptide core by the ACP to generate a novel CDA derivative comprising an altered fatty acid side-chain.

[061] The research described herein was designed to determine the validity of this hypothesis and involved substituting smaller amino acids found in other KAS-Il homologues for Phe 108 offabF3. These studies were designed with several aims in mind: (I) to demonstrate that it is possible to generate, novel forms of CDA through specific bioengineering of a fatty acid side-chain biosynthetic enzyme so as to specifically modify its activity (II) to prove that FabF3 is a KAS-Il enzyme responsible for elongation of the fatty acid chain and (III) that FabF3 PhelO7 is involved in determining the size of the acyl binding pocket, and so controls the length of the fatty acid synthesized by FabF3.

[062] We decided to alter Phe 107 to isoleucine, leucine and alanine. These amino acids were chosen as they possess smaller side-groups than phenylalanine, but are similar to Phe in that they are still hydrophobic amino acids. Moreover, based on alignments with other KAS-Il enzymes (Figure 11) lie, Leu and Ala are represented at position 108 (E. coli enzyme numbering) in other KAS-Il enzymes and therefore we predicted that these amino acids may be able to successfully functionally substitute for Phe at the equivalent position.

Closest prior art

[063] Our survey of the scientific and patent literature relating to the area of lipopeptide antibiotic acyl side-chains is summarized above in paragraphs 0 19-053. However, we acknowledge several pieces of prior art to be particularly close to the present invention.

[0641 The first of these is the work described by Powell and co-workers [Powell et al., 2007] which was conducted in the laboratory of the present applicants. The deletion of hxcO, and the synthesis by the mutant Streptomyces coelicolor LthxcO strain of novel CDA derivatives (hCDA4a & hCDA4b) comprising a hexanoyl fatty acid side-chain is described. This work differs significantly from the present invention in several respects.

Firstly, the length of the fatty acid side-chain of the CDA derivatives described in Powell et al., 2007 was not modified from the normal wild-type fatty acid length, and remained constant at 6 carbon units. Secondly, although a CDA fatty acid biosynthetic gene (hxcO) was mutated it was largely deleted (1206 bp deleted from 1 802bp) and its activity therefore completely abolished, rather than employing a specific, targeted SDM approach designed to modify, rather than abolish, the activity of fabF3, to which the present invention relates.

[065] Powell et al., 2007 also describe the synthesis of a novel CDA derivative (pCDA4b) comprising a pentanoyl side-chain using a mutasynthesis approach, as described above in paragraphs 042-043. This work differs significantly from the present invention in that the lipid moiety of pCDA4b is a C5 fatty acid, and not a C4 lipid described herein. Moreover, the mutasynthesis approach significantly differs from the present invention in that the SDM described by Powell and co-workers was designed to abolish the activity of the PCP domain of Module 1 of CDAPSI rather than modify the substrate activity of a CDA fatty acid biosynthetic gene fabF3, which the present invention relates to.

[066] Powell et aL, 2007 also does not describe the actual modification of fabF3 and this document does not suggest any specific residues which Phe 107 could be replaced by. The replacement of PhelO7 by smaller hydrophobic amino acids residues is not mentioned and the potential result of generating a fatty acid shorter than C6 is also not suggested. The skilled man reading Powell et al., 2007 in combination with Val et al., 2000 (which is cited in Powell et al., 2007) would as a result of the information provided in these documents predict that modification of S. coelicolor FabF3 Phe 107 to smaller hydrophobic amino acids would lead to enlargement of the acyl binding pocket of FabF3 and consequent synthesis of>C6 lipid derivatives of CDA.

[067] However, the results of the present document, which comprise and relate to the present invention, show that modification of S. coelicolor FabF3 Phe 107 to smaller hydrophobic amino acids leads to synthesis of a shorter fatty acid than the wt, native fatty acid, and consequently to the synthesis of C4 lipid derivatives of CDA. Therefore, what the skilled man would predict through reading Powell et aL, 2007 in combination with Val et aL, 2000 is completely different to, and indeed is the opposite of the result described in the present document. Therefore, it is believed that the surprising result described herein, which relates to/comprises part of the present invention, confers an inventive step upon the present invention, as the results of the work which led to the present invention could not have been predicated from the prior art which in fact teach towards a different, and completely opposite, result.

[068] Another document which we acknowledge as relevant prior art is Kopp et al., 2008. As summarized above in paragraphs 032-034 this document describes the in vitro activities of HxcO and HmcO using as substrates a variety of fatty acids (C4-C10) coupled to the ACP. Both enzymes are described as being able to accept and modify C4 fatty acids.

However, Kopp et al., 2008 do not describe the synthesis of novel forms of CDA comprising C4 fatty acids, nor do they teach towards synthesis of such compounds.

[069] Kopp et al., 2008 do not mention modification of the fatty acid synthase enzymesfabH4 and fabF3, and indeed hardly mention these enzymes at all. Where they do mention them they cast doubt on their exclusive role in the synthesis of the hex-2-enoyl intermediate and suggest that other FAS enzymes involved in primary fatty acid metabolism may supply hex-2-enoyl-ACP directly to HcmO. The totally in vitro approach used in the experiments of Kopp et al., 2007 is also very different from the very much in vivo approach which the present invention relates to.

Statements of Invention

[070] The present invention relates to derivatives of the Calcium Dependant Antibiotic variants CDA1b, CDA2a, CDA3a, CDA3b, CDA4a, CDA4b, CDA2d, CDA2fb, CDA2fa, CDA5a and CDA6a which comprise either a 2,3-epoxybutanoyl fatty acid side-chain or a butanoyl fatty acid side-chain in place of the wild-type CDA trans 2,3 epoxyhexanoyl fatty acid side-chain.

[0711 The present invention relates to a method of synthesizing derivatives of the Calcium Dependant Antibiotic variants CDA1b, CDA2a, CDA3a, CDA3b, CDA4a, CDA4b, CDA2d, CDA2flD, CDA2fa, CDA5a and CDA6a which comprise either a 2,3-epoxybutanoyl fatty acid side-chain or a butanoyl fatty acid side-chain in place of the wild-type CDA trans 2,3 epoxyhexanoyl fatty acid side-chain comprising mutating a bacterial strain comprising the CDA biosynthetic gene cluster and which is capable of producing CDA so as to mutatefabF3 of the CDA biosynthetic gene cluster so that residue Phe 107 of FabF3 is altered to either an isoleucine, a leucine or a valine residue.

[072] The present invention also relates to derivatives of the Calcium Dependant Antibiotic variants according to the present invention as hereinbefore described.

[073] The present invention also relates to a method of synthesizing derivatives of the Calcium Dependant Antibiotic variants comprising mutating a bacterial strain comprising the CDA biosynthetic gene cluster and which is capable of producing CDA so as to mutate fabF3 of the CDA biosynthetic gene cluster as hereinbefore described.

Description of the present invention.

[074] It should be noted that the present invention relates to derivatives of CDA variants. For the purposes of the present invention we define "CDA variants" as any naturally occurring or man-made forms of CDA known at the present date which comprise a wild-type trans 2,3 epoxyhexanoyl acyl side-chain. The term "CDA variants" therefore encompasses the following naturally occurring molecules which are referred to in paragraph 008:-variants CDA1b, CDA2a, CDA3a, CDA3b, CDA4a and CDA4b.

Additionally, the term "CDA variants" also encompasses the following non-naturally occurring molecules which are referred to in paragraphs 011 & 012;-CDA2d, CDA2fb, CDA2fa, CDA5a and CDA6a.

[075] The term "CDA variants" does not encompass forms of CDA which possess non-wild-type acyl side-chains i.e. forms of CDA which possess fatty acid side-chains other than trans 2,3 epoxyhexanoyl. Therefore, the term "CDA variants" does not encompass the non-natural forms of CDA, hCDA4a, hCDA4b and pCDA4b described in paragraphs 031 & 043 which possess hexanoyl and pentanoyl acyl side-chains [Powell et al., 2007]. The reason for this exclusion is that if the fatty acid side-chains of these molecules were replaced by 2,3-epoxybutanoyl fatty acid side-chain, or a butanoyl fatty acid side-chain, (which is the nub of the present invention) then the resultant molecules would be indistinguishable from the 2,3-epoxybutanoyl fatty acid side-chain, or butanoyl fatty acid side-chain derivatives, of the CDA variants CDA4a and CDA4b which are already within the scope of the present invention. Therefore, to avoid patenting the same material twice the present invention does not relate to derivatives of forms of CDA which differ from CDA1b, CDA2a, CDA3a, CDA3b, CDA4a, CDA4b, CDA2d, CDA2fb, CDA2fa, CDA5a and CDA6a by virtue of them possessing non-wild-type acyl side-chains.

[076] The present invention relates to derivatives of CDA variants Derivatives of CDA variants are CDA variants which instead of possessing the naturally occurring wild-type fatty acid side-chain trans 2,3 epoxyhexanoyl instead possess a 2,3-epoxybutanoyl fatty acid side-chain or a butanoyl fatty acid side-chain.

[077] It may seem prima facie that the present invention relates to two types of molecules i.e. firstly derivatives of CDA variants comprising a 2,3-epoxybutanoyl side-chain and secondly derivatives of CDA variants comprising a butanoyl side-chain. However, as both fatty acid molecules are straight chain C4 fatty acids differing only in their epoxidation status and moreover as the 2,3-epoxybutanoyl is derived by the modification of butanoyl it is considered reasonable to assume unity of invention exists.

[078] Additionally, as derivatives of CDA variants comprising a 2,3-epoxybutanoyl side-chain and derivatives of CDA variants comprising a butanoyl side-chain are co-synthesized by the same mutant bacterial strains as a result of the same genetic modification, unity of invention is also conferred by their shared method of synthesis.

[079] The present invention also relates to a method of synthesizing derivatives of the Calcium Dependant Antibiotic variants CDA 1 b, CDA2a, CDA3a, CDA3b, CDA4a, CDA4b, CDA2d, CDA2fb, CDA2fa, CDA5a and CDA6a which comprise either a 2,3-epoxybutanoyl fatty acid side-chain or a butanoyl fatty acid side-chain in place of the wild-type CDA trans 2,3 epoxyhexanoyl fatty acid side-chain comprising mutating a bacterial strain comprising the CDA biosynthetic gene cluster and which is capable of producing CDA so as to mutatefabF3 of the CDA biosynthetic gene cluster so that residue Phe 107 of FabF3 is altered to either an isoleucine, a leucine or a valine residue.

[080] The bacterial strain comprising the CDA biosynthetic gene cluster and which is capable of producing CDA may be Streptomyces coelicolor which naturally possesses the CDA biosynthetic gene cluster and which is capable of producing CDA. In the most preferred embodiment of the present invention the bacterial strain comprising the CDA biosynthetic gene cluster and which is capable of producing CDA is Streptomyces coelicolor. Any strain of Streptomyces coelicolor may be used which comprises the CDA biosynthetic gene cluster e.g. MT111O, 2377, M600, M145. Additionally, the bacterial strain comprising the CDA biosynthetic gene cluster and which is capable of producing CDA may be Streptomyces lividans which naturally possesses the CDA biosynthetic gene cluster and which is capable of producing CDA. In a preferred embodiment of the present invention the bacterial strain comprising the CDA biosynthetic gene cluster and which is capable of producing CDA is Streptomyces lividans. Any strain of Streptomyces lividans may be used which comprises the CDA biosynthetic gene cluster e.g. 1326, TK21, TK24.

[08 1] The bacterial strain may be a strain/species which does not naturally comprise the CDA biosynthetic gene cluster, or produce CDA, but which has been genetically modified by means well known in the art to the skilled man (Penn et al., 2005; Mootz et al., 2002; Miao, et al., 2006; Nguyen et al., 2006; Hahn & Stachelhaus, 2006; Mootz et al., 2000; Doekel & Marahiel, 2000; Stacheihaus et al., 1999; Uguru et al., 2004; Sambrook & Russell, 2001; Kieser et al., 2000) so as to comprise the heterologous CDA biosynthetic gene cluster and produce CDA. Preferably the bacterial strain is a member of the Streptomyces genus, excepting Streptomyces coelicolor and Streptoinyces lividans. In an embodiment of the present invention the bacterial strain comprising the CDA biosynthetic gene cluster and which is capable of producing CDA is a member of the Streptomyces genus, excepting Streptomyces coelicolor and Streptomyces lividans.

[082] In especially preferred embodiments of the present inventionfabF3 of the CDA biosynthetic gene cluster is mutated so that residue PhelO7 of FabF3 is altered to either an isoleucine or a leucine residue. The worked examples provided below relate to a mutant strain where residue Phe 107 of FabF3 is altered to an isoleucine residue and to a mutant strain where Phe 107 of FabF3 is altered to a leucine residue. It is clear that altering Phe 107 to both isoleucine and leucine gives the same result for both mutant strains in terms of CDA phenotype. Based on these results it is entirely reasonable to extrapolate from them the general teaching that alteration of Phe 107 to amino acids which possess smaller hydrophobic R groups results in similar identical CDA phenotypes i.e. production of a mixture of wt CDA and two CDA variants which instead of possessing the naturally occurring wild-type fatty acid side-chain trans 2,3 epoxyhexanoyl instead possess 2,3-epoxybutanoyl or butanoyl fatty acid side-chains.

[083] On the basis of this general teaching we strongly believe that if Phe 107 of FabF3 were replaced by a valine residue the mutant strain comprising the mutated F 1 O7VfabF3 would have a CDA phenotype consistent with those observed in the 11e107 and LeulO7fabF3 mutant strains and produce a mixture of wt CDA and two CDA variants which instead of possessing the naturally occurring wild-type fatty acid side-chain trans 2,3 epoxyhexanoyl instead possess 2,3-epoxybutanoyl or butanoyl fatty acid side-chains. Therefore, although we do not have irrefutable experimental evidence to prove that a Phe 1 O7Val FabF3 mutant would produce derivatives of CDA variants comprising 2,3-epoxybutanoyl or butanoyl fatty acid side-chains the skilled man would on the basis of the general teaching presented herein conclude that a VallO7fabF3 mutant would be able to synthesize the derivatives of the Calcium Dependant Antibiotic variants which are the subject of the present invention and conclude that the present application is fully supported in this regard.

[084] How the present invention may be worked is described in the following example in paragraphs 087-095. It should be noted that the worked example of the invention presented here is by way of illustration only and should not be construed as limitation of the scope of the present invention, as defined in the claims.

[0851 It would be obvious to the skilled man that the invention as defined in the claims could readily be achieved by methods other than those described in the worked example. For example, although the example below describes the use of the suicide/shuttle vector pMAH in a classic double-crossover method to introduce the desired mutation by homologous recombination the skilled man would be aware that other vectors could be used, for example pWl-1M3 [Kieser et al., 2000], to achieve the same result. Alternatively, the REDIRECT system may be used to introduce mutations of interest [Gust et a!., 2003].

[086] Similarly, although the mutant codons for Ile, Leu and Ala used to replace the Phe codon in the example below were chosen so as to minimize the number of necessary nucleotide changes and be consistent with Streptomyces codon usage the skilled man would appreciate that due to the redundancy of the gentic code other codons which specify Ile, Leu and Ala could be used as alternatives. Additionally, there are many site directed mutagenesis methods, and commercial kits for effecting such methods available, and although the Stratagene QuikChange kit was used in the example below the skilled man would appreciate that other methods/kits could readily be used to achieve the same result.

Bacterial strains, microbiological methods and genetic manipulations.

[087] All methods used for growth and transformation of E. coli and S. coelicolor are as described previously [Sambrook and Russell, 2000: Kieser et a!., 2000]. A -2.kb DNA fragment spanningfabF3 was amplified by polymerase chain reaction (PCR) using S. coelicolor MT 1110 genomic DNA as template and the oligonucleotides fabF3Up and fabF3Down as PCR primers (see Table 2) and the KODRTM Hot Start polymerase PCR kit purchased from Novagen&TM. Each PCR (50p1) mix contained 3O.5tl water, 5.i,l lox KOD polymerase Buffer, 5j.tl 2mM dNTP's, 3tl 25mM MgSO4, 2.5p1 DMSO, 1.d (4Opmoles) of upstream primer, ipi (4Opmoles) of downstream primer, 1 111 of template DNA, 1 t1 KOD polymerase.

The PCR was effected in a GeneAmpRTM PCR System 9700" thermocycler from PE Applied BiosystemsRTM The touchdown PCR program was as follows:-1 cycle of:-(95°C, 2 mm); 20 cycles of:-(95°C, 30°sec; 85-75°C ("touchdown" with 0.5°C drop per cycle), 40 sec; 70°C, 100 sec); 20 cycles of (95°C, 3Osec; 75°C, 40 sec; 70°C, 100 sec); 1 cycle (70°C, 15 mm); 4°C (indefinite hold).

Table 2

Primer name Sequence: 5'-3' 1 fabF3Up CTGGTCAGAICTGACCTCGACAACCG fabF3Down CCGACCTCIAGACGGGCCAGCACC fabF3IleUp GGGCAGCGCCTCGGGGGCGTGACC fabF3lleDown GGTCACGCCCCCGATGGCGCTGCCC fabF3LeuUp GGGCAGCGCCTCGGGGGCGTGACC fabF3LeuDown GGTCACGCCCCCGAGGGCGCTGCCC fabF3AlaUp GGGCAGCGCCCGGGGGCGTGACC fabF3AlaDown GGTCACGCCCCCGGCGGCGCTGCCC fabF3Ml 3 forward CGCCAGGGTTTTCCCAGTCACGACG

_________________ CGATGTGCGCAGCCCGGTCCC

fabF3Ml 3reverse AGCGGATAACAATTTCACACAGGAG

CGTCTTGCCGAGCCCGAACGCC

Table 2 provides the sequences of all primers used in the present study. Non-native residues are underlined. Bold type indicates codon 107 (fabF3 numbering).

[088] The PCR product was digested with XbaI (Roche'TM) and cloned into similarly cut pHSG398, [Takeshita et al., 1983] treated with Antarctic phosphastase (N.E.B. RTM), to yield pRAL75. This construct was used in the QuikChangeRTM protocol (StratageneRTM) using a series of mutagenic primer pairs (Table 2) designed to substitute lie (FabF3IleUp & fabF3lleDown), Leu (FabF3LeuUp & fabF3LeuDown) and Ala (FabF3AlaUp & fabF3AlaDown) residues for PhelO7.The PCR was effected in a "GeneAmp PCR System 9700" thermocycler from PE Applied Biosysterns'TM The QuikChangeRlM PCR program was as follows:-1 cycle of:-(95°C, 1 mm); 18 cycles of:-(95°C, 50sec; 60°C, 50 sec; 68°C, 600 sec); 1 cycle (68°C, 7 mm); 4°C (indefinite hold).

[089] The presence of the desired mutations was verified by DNA sequencing (EurofinsMWGOperonRTM) using forward and reverse universal M13 primers (Ml2uni(-43); Ml3rev(-49)), and the mutated inserts were each cloned into the S. coelicolor suicide vector pMAH as XbaI fragments to generate the deletion constructs pRAL759 1 (F-I), pRAL7592 (F-L) and pRAL7593 (F-A) respectively. After passaging through the DNA demethylating E. coli strain ET12567 these constructs were used to transform Streptomyces coelicolor MT 1110 protoplasts and transformants selected for using hygromycin (InvitrogenRTM). Following two rounds of unselective growth hygromycin sensitive colonies, identified by replica plating, were screened, using DNA sequencing (EurofinsMWGOperoflRTM) of PCR products generated using primers fabf3 Ml 3 foward & faBf3Ml3reverse (Table 2) spanning the mutated region of fabF3, to identify mutants. The PCR was effected in a "GeneAmp'TM PCR System 9700" thermocycler from PE Applied Biosystems'TM The touchdown PCR program was as follows:-1 cycle of:-(95°C, 2 mm); 20 cycles of:-(95°C, 3Osec; 85-75C ("touchdown" with 0.5°C drop per cycle), 40 sec; 70°C, 30 sec); 20 cycles of (95°C, 30sec; 75°C, 40 sec; 70°C, 30 sec); 1 cycle (70°C, mm); 4°C (indefinite hold).

CDA plate bioassays [090] CDA plate bioassays were effected as described previously [Chong et al., 1998] Standardized numbers of spores (5 x 10) for each strain tested were washed with 2 x imi of sterile distilled water to remove glycerol and spotted out onto duplicate OxoidRTM Nutrient Agar (ONA) plates containing 25m1 of agar. The plates were incubated at 30°C for 24 hours and then one plate (Ca2 -ye) was overlaid with 5m1 of molten (-40-50°C) soft agar overlay (containing 1 ml of dense overnight culture of Bacillus inycoides cultured in ONB medium (OxoidRTM) and the other duplicate plate (Ca2 +ve) was overlaid with 5m1 of soft agar overlay (containing imi of dense, overnight culture of Bacillus mycoides cultured in ONB medium (OxoidTM) and 120p1 of 3M Ca(No3). The Ca2 diffuses from the soft agar overlay through the whole plate to give a final concentration of 12mM Ca2t The plates were incubated for a further 24 hours and then photographed using a PanasonicRTM Lumix digital camera.

Cultivation of S. coelicolor, Isolation and Detection of CDA variants [091] Wild-type MT1 110 and mutant strains fabF3 F-I, F-L, F-S were grown on solid Mannitol-Soya (MS) media [Kieser et al., 2000], then inoculated into SV2 liquid media [Kieser et al., 2000] and incubated for a further 5 days at 28°C and 180 rpm. Liquid media was spun down at 4500 rpm for 10 minutes; the supernatant was filtered, adjusted to pH 2.0 with 1M HC1 and then re-filtered before being loaded onto a VarianRTM C-18 Bond EluteTM SPE cartridge. Crude CDA was extracted using a methanol gradient (30-100%), evaporated under reduced pressure and analyzed using LC-MS [Powell et al., 2007].

Purification of CDA variants [092] CDA extracts were purified on a Varian ProstarRT4 instrument, equipped with a ProstarRTM 335 photodiode array detector. Purification was achieved using semipreparative reversed phase HPLC: PhenomenexRlM C- 18 5mm, 250 x 10 mm column. Solvent A was H20 containing 0.1% HCO2H and solvent B was acetonitrile also containing 0.1% HCO2H.

At a flow rate of 3 mL min' B was increased from 5% to 100% over a 30 mm linear gradient, held at 100% for 5 mm before being brought back down to 5% (pre-equilibration for next run); detection by UV at 280 nm and 350 nm. Fractions were analysed by LC-MS [Powell et al., 2007].

Analysis of CPA Extracts Using LC-MS [093] General LC-MS analysis was carried out on a MicromassTM LCT orthogonal acceleration time of flight mass spectrometer, equipped with an electrospray ionisation source run in positive mode (scanning from 700 to 1700 m/z) combined with a WatersRTM 2790 Separation module. A reversed phase C-18 150 x 4.6 mm 3 1um PhenomenexRTM column was used. Solvent A was H20 with 1 % acetonitrile and 0.1 % HCO2H and solvent B was acetonitrile with 1 % 1120 and 0.1 % HCO2H. At a flow rate of 1 ml min with a gradient of 80 % A and 20 % B, increasing to 70 % B over 10 mm and increased to 100 % B over the 1 mm and held for a further 4 mm [Powell, 2007].

Accurate Mass and Tandem MS analyses of CPA using LTQ-Orbitrap XL-nUPLC [094] Samples were initially diluted in water, and subsequently in 95 % water! 5 % acetonitrile supplemented with 0.1 % HCO2H (hereafter referred to as Buffer Al). Data were acquired using a nanoacquity chromatograph (WatersRTM MS technologies) coupled to an LTQOrbitrapRTM XL mass spectrometer (ThermoFisher Scientific'TM) equipped with the manufacturer's dynamic nanospray source, fitted with a PicoTipRTM emitter (20-10 tm diameter, New Objective). The sample temperature in the autosampler was maintained at 10 °C, and 4 tl of sample was injected initially onto a trapping colunin (Waters C-18 180 jtm x 20 mm), with a sample loading time of 1 mm, using the partial loop mode of injection, at a flow rate of 18 p1/mm (Buffer Al). Following loading, the valve position was switched, such that the trapping and analytical columns were in-line with respect to each other, and on-line with the mass spectrometer. The analytical column (nanoACQUTTY UPLCRTM BEH C-18 75 jtm x 150 mm 1.7 p.m column) was maintained at a temperature of 35 °C, and at a constant flow rate of 300 nI/mm. The gradient conditions were as follows: 0.33 mm - % acetonitrile/0. 1 % HCO2H, up to 50 % acetonitrile/ 0.1 % HCO2H over a 30 mm linear gradient; 31 mm -85 % acetonitrile/ 0.1 % HCO2H, 35 mm, % acetonitrile/ 0.1 % HCO2H (pre-equilibration for the next run). The total length of an analytical run was 50 mm.

[095] For high mass accuracy, all spectra were acquired in an LTQ-OrbitrapRTM XL with the OrbitrapRTM operating at a resolution of 30,000 (defined at mlz 400). The OrbitrapRlM was set to acquire alternate full MS scans, and tandem MS data from the linear ion trap. Data were acquired in positive ion mode using XcaliburRlM version 2.0.5 and TuneplusRTM version 2.4 SP1, and in data-dependent mode, with the most intense peak from full MS spectra directing collision induced dissociation (CID), at normalized collision energy of 35 %, and an activation q of 0.25. Dynamic exclusion was applied for a duration of 30 sec, with a repeat count of 2, and an exclude duration width of 180 sec to avoid repeated selection of the same mlz precursors for MSMS.

Results and discussion.

[096] After consideration of the repertoire of amino acids used at position 108 (E. coli enzyme numbering) of a large number of KAS-Il enzymes, and previous studies [Val et al., 2000] it was decided to introduce F-I, F-L and F-A changes into fabF3, as these amino acids are commonly found in this position in many species in which the fatty acid synthesized is >C6. It was considered these changes were conservative enough to maintain the overall structure of the enzyme, whilst the slightly smaller size of the hydrophobic amino acid R groups it was hoped would increase the size of the lipid binding pocket and lead to elongation of the lipid chain past C6.

[097] The standard double-cross-over technique, using the insertion, and then excision of the suicide vector pMAH by homologous recombination to replace the wild-type fabF3 sequence with the mutant copy of the fabF3 region enabled us to successfully obtain the F-I and F-L mutants. However, instead of the expected F-A change mutants generated by double crossover events, from two independent pRAL7593 transformants contained F-S mutations (TTC-TCC) only. The reason for this phenomenon is unknown, however a possible explanation is that the two base mismatch between the F-A mutant construct and wt MT1 110 chromosome acts as a recombination hotspot and that one or other of the crossover events occurred between the mutagenized and wild-type (wt) codons to generate a recombinant serine codon (TCC) from the wt Phe (TIC) and mutant Ala (GCC) codons. As the serine R group is similar in size and charge to alanine this unexpected result was not seriously deleterious to the work. Also we note that serine is a residue which is also present at position 108 (E. coil enzyme numbering) of KAS-Il enzymes.

[098] CDA plate bioassays were done of the F1071, F1O7L and F1O7S mutants together with the wt parent strain Streptomyces coeiicoior A3(2) MT1 110 as a positive control and as a negative control Streptomyces coeiicolor A3(2) MT1 110 comprising a deletion of the cda biosynthetic gene cluster. The CDA bioassays (Figure 12) indicated that when compared to wt MT1 110 all three mutants produced smaller zones of inhibition in the B. myco ides lawn surrounding the S. coelicolor patches, indicating, either production of less CDA than the parent strain, or production of a less bioactive forrnls of the antibiotic, or both (Figure 12). The zones of inhibition surrounding the F 1071 and F 1 07L mutants were similar in size whereas the zone of inhibition surrounding the Fl 07S mutant was smaller than the zones of inhibition surrounding the F 1071 and Fl 07L mutants (Figure 12).

[099] The mutants were then grown in liquid culture under conditions that favour CDA production and the supematants were analyzed by LC-MS. This demonstrated that all three of the mutant strains produced wild-type CDA4a as the major product (Figure 13A). The PhelO7Ile and PhelO7Leu mutants both produce smaller amounts of two products, not present in extracts from the parental MT1 110 strain with retention times of 8.10 and 8.27 mm that exhibited protonated, sodiated, and potassiated singly charged as well as doubly ions, which are consistent with products of mw 1466 and 1452 Da respectively (Figure 13B) These products were purified by RP-HPLC and UV analysis showed both products have UV spectra typical of a-series CDAs, possessing Z-dehydrotryptophan residues (lm, 350 nm). The products were further subjected to high-resolutions tandem MS, using a method previously developed, which enables the trans-2, 3 -epoxyhexanoyl-L-Ser tail of CDAs to be fragmented from the decapeptide lactone core [Powell et al., 2007] (Figure 14 & Figure 15). This indicated that both of the new products possess the same peptide cores (Y ions) as CDA4a, with fatty acid-L-Ser tails that are 14 and 28 mass units lower. This is consistent with CDA variants possessing shorter (C4) 2,3-epoxybutanoyl (ebCDA4a) and butanoyl (bCDA4a) fatty acid side chains (Figure 16 and Figure 1).

[100] These results are in agreement with the proposal that fabH4 encodes a KAS-Ill enzyme that catalyses a Claisen-type condensation of acetyl and malonyl units to generate acetoacyl-S-ACP which is presumably processed by the primary metabolic fatty acid synthase (FAS) to give butanoyl-S-ACP 1 (Figure 18). The KAS-Il enzyme FabF3 catalyzes the subsequent condensation of butanoyl-S-ACP 1 with a malonyl unit leading to CDA4a via the known natural C6 fatty acid precursors 2-4 (Figure 18). FabF3 is thus responsible for determining the chain length of the lipid synthesized. The PhelO7Ile and PhelO7Leu FabF3 mutants are presumably less efficient than the wt enzyme at catalyzing this reaction as, in addition to reduced synthesis of hexanoyl-S-ACP, the unextended C4 precursor, butanoyl-S-ACP, is also subsequently passed to the downstream fatty acid processing enzymes.

[101] The results demonstrate the resilience of the CDA fatty acid synthesis/loading enzymes, and that despite the fact that usually their sole substrate is a specific molecule, trans, 2-3, epoxyhexanyol, they also have the ability to accommodate altered substrates. This is supported by the fact that an S. coelicolor strain L\hxcO, which lacks the hxcO gene, was shown to produce CDA variants with hexanoyl fatty acid side chains [Powell et al., 2007]. It is interesting to note that despite their shortened chains the C4 lipid intermediates (2,3-epoxybutanoyl and butanoyl) are still capable of being used productively, at least to some extent, in the synthesis of CDA, i.e. they must be capable of being bound by the ACP -essential for their participation in downstream reactions. This finding is consistent with previous findings that the ACP is only capable of binding lipids of C6 and smaller [Kopp et al., 2008]. It is also interesting to note that the butanoyl-S-ACP 1 (Figure 18) which accumulates in the PhelO7Ile and PhelO7Leu FabF3 mutant stains, is also transformed by HxcO and HcmO to give 2,3-epoxybutanoyl-S-ACP 6 leading to ebCDA4a. This is supported by earlier in vitro studies which show that both butanoyl-S-ACP 1 (Figure 18) and butenoyl-S-ACP 5 (Figure 18) are viable substrates for recombinant FixcO and HcmO respectively [Kopp et al., 2008]. Whilst, the exact mechanism of the lipidation is not known, it is suggested that the N-terminal C-domain of the first NRPS (cdaPS 1) mediates the transfer of the fatty acid from the ACP to a Ser residue tethered to an adjacent PCP domain [Powell et al., 2007; Wittman et al., 2008]. The fact that 2,3-epoxybutanoyl-S-ACP 6, and butanoyl-S-ACP, 1 (Figure 18) carl be used to initiate lipopeptide assembly indicates that lipidation of CDA is reasonably promiscuous. Furthermore, the linear CDA peptides comprising the C4 lipids seem to be acceptable substrates capable of participating in an accurate cyclization reaction by the TE domain of CDAPSIII [Grunewald et al., 2004a].

[102] The small quantities of the new lipopeptides produced prevented detailed studies of their biological activity which at present are unknown. In view of the fact that the F1071 and F1O7L mutants strains produce a mixture of mostly wt CDA with smaller amounts of the novel CDA derivatives which are the subject of the present invention it is difficult to determine the antibiotic effect of the novel CDA derivatives from this.

[103] In summary the results presented here provide further evidence to support the hypothesis that fabH4 and fabF3 encode KAS-IlI and KAS-Il enzymes responsible for effecting the first and second Claisen-type condensation reaction in the biosynthesis of the CDA fatty acid moiety and determining the length of the fatty acid chain at C6. It is perhaps surprising based on complementary studies with E. coli KAS-Il enzyme [Val et a!., 2000], that the mutagenesis which generated the FabF3 mutants do not appear to increase the depth of acyl binding pocket allowing further extensions leading to CDAs with longer fatty acid moieties. A precise explanation of the results obtained in terms of structure of the binding pocket is beyond the scope of the current study, nevertheless, our results are consistent with our hypothesized role of Phe 107 playing a key role in the architecture of the lipid binding pocket. We can not rule out the possibility that longer fatty acid precursors are produced, by the FabF3 mutants, which are not transferred to or processed by the downstream NRPS. Nevertheless, our results provide the first example of how mutagenesis of a FAS system associated with an NRPS assembly line can be used to engineer new lipopeptides with modified lipid groups.

Figure legends:-Figure 1 legend Schematic diagrams illustrating the amino acid composition of the depsipeptide lipopeptide antibiotics (A) A54145, (B) A21978, (C) CDA.

Figure 2 legend Schematic chemical structure diagram of CDA indicating the position of amino acids and the various possible amino acid variants which may be present at positions 9, 10 and 11.

Figure 3 legend (A) Schematic chemical structure diagram of CDA, (B) Table providing a non-exhaustive list of naturally, and non-naturally, occurring CDA variants, their respective molecular weights and details as to the different chemical structures which differentiate them at residue positions R6, R9, RiO and Ru.

Figure 4 legend (A) Schematic chemical structure diagram of CDA, (B) Table providing a details of the chemical differences which differentiate the non-naturally, occurring CDA variants CDA2d, CDA2fa and CDA2th generated as described in Powell et al., 2007.

Figure 5 legend (A) Schematic chemical structure diagram of CDA, (B) Table providing a details of the chemical differences which differentiate the non-naturally, occurring CDA variants CDA5b, CDA6b and CDA6a generated as described in Neary et al., 2007.

Figure 6 legend Schematic chemical structure diagram of daptomycin Figure 7 legend Schematic diagram illustrating the genetic composition and organization of the CDA biosynthetic gene cluster. Gene names ("sco numbers") are shown under the respective genes and the gene arrows are patterned according the proposed/proved function of the gene product as shown in the inset key. This Figure is adapted from Hojati et a!., 2002.

Figure 8 legend Table listing the genes of the CDA biosynthetic gene cluster and providing details concerning them This Figure is adapted from Hojati et a!., 2002.

Figure 9 legend (A) Schematic chemical structure diagram of CDA, (B) Table providing a details of the chemical differences which differentiate the non-naturally, occurring CDA variants hCDA4b, hCDA4a and hCDA3b generated as described in Powell et al., 2007.

Figure 10 legend (A) Schematic chemical structure diagram of CDA, (B) Table providing a details of the chemical differences which differentiate the non-naturally, occurring CDA variants CDA5b, CDA6b and CDA6a generated as described in Neary et a!., 2007.

Figure 10 legend Schematic chemical structure diagram illustrating the proposed mechanism for the biosynthesis of trans 2,3, epoxyhexanoyl.

Figure 11 legend Alignment of the known amino acid sequences of the acyl-substrate binding pockets of KAS-Il enzymes from E. co/i and Synechocystis sp with the putative substrate binding pockets from the KAS-Il paralogues (FabF 1-3) from S. coelicolor. The numbering system corresponds to the positions relative to the E. coli KAS-JI enzyme. Residues marked with an asterisk (*) are known, or assumed, to come from a second protein subunit. Residue 108 (E. coli numbering system) is marked with an arrow.

Figure 12 legend.

CDA plate bioassays of MT111O wt and fabF3 mutant strains using B. 2+ myco ides as an indicator strain in the presence (A) and absence (B) of Ca (a) MT1 110; (b) MT1 110 cdaPSI-IIIL\; (c) MT1 1 lOfabF3 F1O8S; (d) MT1 1 lOfabF3 F1O8L; (e) MT1 1 lOfabF3 F 1081.

Figure 1.3 legend LC-MS analysis of CDAs from MT11 10 and the mutant strain (PhelO7Ile) was carried out on a Micromass LCT orthogonal acceleration time of flight mass spectrometer, equipped with an electrospray ionisation source run in positive mode (scanning from 700 to 1700 rn/z) combined with a Waters 2790 Separation module. (A) The LC-MS chromatogram from MT1 110 shows a product with retention time 9.01 minutes, which exhibits protonated, sodiated and potassiated ions in the mass spectrum consistent with CDA4a. (B) The LC-MS chromatogram of mutant strain Phe 1 0711e shows peaks with a retention time of 8.10 minutes, 8.27 minutes and 8.95 minutes that exhibit protonated, sodiated and potassiated ions, which correlate to ebCDA4a, bCDA4a and CDA4a respectively.

Figure 14 legend High resolution LC-MS and MS-MS analysis was also carried out using LTQ-Orbitrap XL-Nuplc. (A) An extracted ion chromatogram of ebCDA4a.

(B) Mass spectrum for ebCDA4a with accurate masses that are consistent with the proposed formula for protonated, sodiated, and potassiated ions of ebCDA4a. (C) MS-MS for ebCDA4a shows the Y ion consistent with formula for the decapeptide core of CDA4a.

Figure 115 legend High resolution LC-MS and MS-MS analysis was also carried out using LTQ-Orbitrap XL-Nupic. (A) An extracted ion chromatogram of bCDA4a.

(B) Mass spectrum for ebCDA4a with accurate masses that are consistent with the proposed formula for protonated, sodiated, and potassiated ions of bCDA4a. (C) MS-MS for bCDA4a shows the Y ion consistent with formula for the decapeptide core of CDA4a.

Figure 16 legend Accurate mass and tandem MS data for CDA4a, bCDA4a and ebCDA4a obtained using the LTQ-Orbitrap XL-nUPLC.

Figure 17 legend Schematic chemical structure diagram of the peptide core of CDA4a, and the trans 2,3 epoxyhexanoyl, butanoyl and epoxybutanoyl fatty acid side-chains of CDA4a, bCDA4a and ebCDA4a respectively, shown alongside.

Figure 18 legend Schematic diagram illustrating the biosynthesis of the CDA fatty acid moiety. FAS primary metabolic fatty acid synthase ketoreductase, dehydratase and enoylreductase enzymes; FabH4 3-ketoacyl-S-ACP synthase (KAS-IJI); FabF3 = -ketoacyl-S-ACP synthase (KAS-Il); HxcO FAD-dependant hexanoyl-S-ACP oxidase; HcmO trans-hexanoyl-S-ACP monoxygenase activity respectively. 1 = butanoyl-S-ACP; 2= hexanoyl-S - ACP; 3 hexenoyl-S-ACP; 4=2,3 -epoxyhexanoyl-S-ACP; 5= butenoyl-S-ACP; 62,3 -epoxybutanoyl-S-ACP.

References All references herein cited are to be treated as being incorporated into the present document in their entirety.

Albom Jr, W.E.A., Allen, N.E., & Preston, D.A. (1991). Daptomycin disrupts membrane potential in growing Staphylococcus aureus. Antimicrob.

Agents Chemother. 35: 2282-2287.

Allen, N.E., Alborn Jr, W.E. & Hobbs, J.N. (1991). Inhibition of membrane potential-dependent amino acid transport by daptomycin. Antimicrob.

Agents Chemother. 35: 263 9-2642.

Allen, N.E., Hobbs, J.N. and Alborn, W.E. (1987). Inhibition of peptidoglycan biosynthesis in gram-positive bacteria by LY 146032.

Antimicrob. Agents & Chemother. 31: 1093-.1099.

Aretz, W., Meiwes, J., Seibert, G., Vobis, E & Wink, J. (2000). Friulimicins: novel lipopeptide antibiotics with peptidoglycan synthesis inhibiting activity from Actinoplanesfriuliensis sp. nov. I. Taxonomic studies of the producing microorganism and fermentation. J. Antibiot. 53: 807-8 15.

Ball, L-J., Goult, C.M., Donarski, J.A., Micklefield, J., and Rarnesh, V. (2004). NMR structure determination of and calcium binding effects of lipopeptide antibiotic daptomycin. Org. Biomol. Chem. 2, 1872-1878.

Barry, A.L., Fuchs, P.C., & Brown, S.D. (2001). In vitro activities of daptomycin against 2,789 clinical isolates from 11 North American medical centers. Antimicrob. Agents Chemother 45: 1919-1922.

Bemer, P., Juvin, M.E., Bryskier, A., & Brugeon, H., (2003). In vitro activities of a new lipopeptide, HMIR 1043, against susceptible and resistant gram-positive isolates. Antimicrob. Agents Chemother. 47: 3025-3029.

Boaretti, M., Canepari, P (1995). Identification of daptomycin-binding proteins in the membrane of Enterococcus hirae. Antimicrob. Agents Cheinother 39: 2068-2072.

Boaretti, M., Canepari, P., Lleo, M.M., & Satta, G. (1993). The activity of daptomycin on Enterococcusfaecium protoplasts: indirect evidence supporting a novel mode of action on lipoteichoic acid synthesis. J. Antimicrob Chemother 31: 227-335.

Boaretti, M., Canepari, P (2000). Purification of daptomycin binding proteins (DBPs) from the membrane of Enterococcus hirae.

New Microbiol. 23: 305-3 17.

Boeck, L.D., Papiska, H.R., Wetzel, R.W., Mynderse, J.S., Fukuda, D.S., Mertz, F.P., and Berry, D.M. (1990). A54 145, a new lipopeptide antibiotic complex: discovery, taxonomy, fermentation and HPLC. Journal of Antibiotics, 43: 587-593.

Boeck, L.D., & Wetzel. R.W. (1990). A54145, a new lipopeptide antibiotic complex: factor control through precursor directed biosynthesis. J. Antibiot.

43: 607-615.

Borders, D.B., Leese, R.A., Jarolman, H., Francis, N.D. Fantani, A.A. Falla, T., and Aumelas, A. (2001). Interscience Conference on Antimicrobial Agents & Chemotherapy, Poster/Abstract \1 154 Chicago, Il, December 16-Brockmann H., & Hieronimus E., (1955). Uber Actinomycetenfarbstoffe, V Mitteil.: Zur Konstitution des Actinorhodins, III. Mitteil. Chem. Ber. 88: 1379-1390.

Bunkoczi, G., Vertsey, L. and Sheidrick, G.M. (2005). Structure of the lipopeptide antibiotic tsushimycin. Acta Crystallogr. Sect. D.: Biol Crystallogr. 61(Pt 8): 1160-1164.

Canepari, P., Boaretti, M., Lleo, M.M., & Satta, G. (1990). Lipoteichoic acid as a new target for activity of antibiotics: mode of action of daptomycin (LY146032). Antimicrob. Agents Chemother 34: 1220-1226.

Challis, G.L., Ravel, J., and Townsend, C.A. (2000). Predictive, structure based model of amino acid recognition by nonribosomal peptide synthetase adenylation domains. Chemistry & Biology. 7: 2 11-224.

Chong, P.P., Podmore, S.M., Lieser, H.M., Redenbach, M., Turgay, K., Marahiel, M.A., Hopwood, D.A., and Smith, C.P. (1998). Physical identification of a chromosomal locus encoding biosynthetic genes for the lipopeptide calcium-dependant antibiotic (CDA) of Streptomyces coelicolor A3 (2). Microbiology 144, 193-199.

Counter, F.T., Allen, N.E., Kukuda, D.S., Hobbs, J.N., Ott, J., Ensminger, P.W., Mynderse, J.S., Preston, D.A., & Wu, C.Y. (1990). A54145 a new lipopeptide antibiotic complex: microbiological evaluation. J. Antibiot. 43: 6 16-622.

Debono, L.D, Abbott, B.J., Molloy, R.M., Fukuda, D.S., Hunt, A.H., Daupert, V.M., Counter, FT., Ott, J.L., Carrel!, C.B., Howard, L.C., Boeck, L.D., & Hamill, R.L. (1988). Enzymatic and chemical modifications of lipopeptide antibiotic A2 I 978C: the synthesis and evaluation of daptomycin (LY146032). J. Antibiot. 41, 1093-1105.

Debono, M., Barnhart, M., Carrell, C.B., Hoffman, J.A., Occolowitz, J.L., Abbott, B.J., Fukada, D.S., and Hamill, R.L. (1987). A21978C, a complex of new acidic peptide antibiotics: isolation, chemistry, and mass spectral structural elucidation. J. Antibiot. 40, 761-777.

Doekel, S., and Marahiel, M.A. (2000). Dipeptide formation on engineered hybrid peptide synthases. Chemistry & Biology. 7: 343-384.

Eliopoulos, G.M., Thauvin, C., Gerson, B., & Moellering, R.C. (1985). In vitro activity and mechanism of action ofA21978C1, a novel cyclic lipopeptide antibiotic. Antimicrob. Agents Chemother., 27: 357-362.

Eliopoulos, Willey, S., Reiszner, E., Spitzer, P.G., Caputo, G., & Moellering, R.C. (1986). In vitro and in vivo activity of LY 146032, a new cyclic lipopeptide antibiotic. Antimicrob. Agents Chemother. 30: 532-535.

Florova, Y.L.G., and K.A. Reynolds. (2005). Alteration of the fatty acid profile of Streptomyces coelicolor by replacement of the initiation enzyme 3-keto acyl carrier protein synthase III (FabH). Journal of Bacteriology, 187: 3795-3799.

Fujino, M. (1965). Bull. Chem. Soc. Jpn. 38: 517.

Fujino, M. Inoue, M., Ueyanagi, J., & Miyake, A. (1965). Bull. Chem. Soc. Jpn. 38: 515.

Fukuda, D.S., Du Bus R.H., Baker, P.J., Berry, D.M., and Mynderse, J.S.

(1990). A54145, a new lipopeptide antibiotic complex: isolation and characterization. J. Antibiot. 43, 594-600.

Fukuda, D.S., Debono, L.D, Molloy, Mynderse, J.S., (1990). A54145, a new lipopeptide antibiotic complex: microbial and chemical modification. J. Antibiot. 43: 601-606.

Gauze, G.F., Preobrazhenskaya, T.P., Kovalenova, V.K., Il"icheva, N.P., Brazhnikova, M.G., Lomakina, N.N., Kouscharova, I.N., Shorin, U.A., Kunrat, l.A., Shapolova, S.P. (1957) Antibiotiki, 2: 9.

Grunewald, J., Sieber, S. A., and Marahiel, M, (2004a). Chemo-and regioselective peptide cyclization triggered by the N-terminal fatty acid chain length: the recombinant cyclase of the Calcium dependant antibiotic from Streptomyces coelicolor. Biochemistry. 43. 2915-2925.

Grunewald, J., Sieber, S. A., Mahiert, C., Linne, U. and Marahiel, M, (2004b). Synthesis and derivatization of daptomycin: A chemoenzymatic route to acidic lipopeptide antibiotics. J. Am. Chem. Soc, 126: 17025-17031.

Gust, B., Challis, G.L., Fowler, K., Kieser, T., and Chater, K.F. (2003).

PCR-targeted Streptomyces gene replacement identifies a protein domain needed for biosynthesis of the sesquiterpene soil odor geosmin. P.N.A.S.

100: 1541-1546.

Hahn, M., and Stachelhaus, T. (2006). Harnessing the potential of communication-mediating domains for the combinatorial synthesis of nonribosomal peptides, P.N.A. S. 103, 275-280 Hancock, R.E (2005). Mechanisms of action of newer antibiotics for Gram-positive pathogens. Lancet Infect. Dis. 5: 209-218.

Haneishi, T., Terahara, A., Arai, M., Hata, T., & Tarnura, C. (1974). New antibiotics, methylenomycins A & B. 2 structures of methylenomycin A & B. J. Antibiot. 27: 393-399.

Huang, W., Jia, J., Edwards, P., Dehesh, K Schneider, G., and Lindqvist Y, (1998). The crystal structure of 3-ketoacyl-acyl carrier protein synthase II from E. coil reveals the molecular architecture of condensing enzymes. The EMBO Journal 17: 1193-1191..

Hausmann, W.K., Borders D.B., & Lancaster, J.E. (1969). Alpha, beta-diaminobutyric acid obtained from aspartocin. J. Antibiot. 22: 207-210.

Hausmann, W.K. Struck, A.H., Martin, J.H., Barritt, R.H., & Bohonos, N. (1963). Structure determination of fatty acids from the antibiotic aspartocin.

Antimicrob. Agents Chemother. 161: 352-359.

Heinemann, B., Kaplan, A.A., Muir, R.D., Hooper, IR. (1953) Antibiot.

Chemother 3: 1239.

Hinuma, Y., (1954). Zaomycin, a new antibiotic from a Streptomyces sp.

Journal of Antibiotics, 7: 134-136.

Hopwood, D.A., and Wright, H.M. (1983). CDA is a new chromosomally determined antibiotic from Streptomyces coelicolor A3(2). J. Gen.

Microbiol. 129, 3575-3579.

Hojati., Z., Milne, C., Harvey, B., Gordon, L., Borg, M., Flett, F., Wilkinson, B., Sidebottom, P.J., Rudd, B.A.M., Hayes, Smith, C.P and Micklefield J. (2002). Structure, biosynthetic origin and engineered biosynthesis of calcium-dependant antibiotics from Streptomyces coelicolor.

Chemistry & Biology. 9, 1175-1187.

Huber, F.M., Pieper, R.L., & Tietz, A.J. in "Biotechnology Processes ", ed C.S. Ho & J.Y. Oldshue, American Institute of Chemical Engineers Publications, New York, 1987, p249.

Huber, F.M. Pieper, R.L. and Tietz, A.J. (1988). The formation of daptomycin by supplying decanoic acid to Streptomyces roseosporus cultures producing the antibiotic complex A21978C. J. Biotechnol. 7: 283-292.

moe, M (1962). Bull. Chem. Soc. Jpn. 35: 1256.

Jung, D., Rozek, A., Okon, M., and Hancock, R.E.W. (2004). Structural transitions as determinants of the action of the calcium-dependant antibiotic daptomycin. Chemistry & Biology, 11, 949-957.

Kempter, C. Kaiser, D., Haag, S., Nicholson, G., Gnau, V., Walk, T, Gierling,K.H., Decker, H., Zahener, H., Jung, G & Metzger, J.W. (1997).

CDA: calcium dependant antibiotics from Streptomyces coelicolor A3(2) containing unusual residues. Angewandte Chemie International English Edition, 36: 498-501.

Kieser, T., Bibb, M.J., Buttner, M.J., Chater, K.F., and Hopwood, D.A.

(2000). Practical Streptomyces Genetics (Norwich, UK: The John Innes Foundation).

Kong F., and Carter, G.T. (2003). Structure determination of glycinocins A to D, further evidence for the cyclic structure of the amphomycin antibiotics.

Journal of Antibiotics, 56: 557-564.

Kopp, F., Linne, U., Oberthr, M., Marahiel, M.A. (2008). Harnessing the chemical activation inherent to carrier protein-bound thioesters for the characterization of lipopeptide fatty acid tailoring enzymes. J.A.C.S. 130: 2656-2666.

Laganas V., Alder, J., and Silverman, J.A. (2003). In vitro bactericidal activities of daptomycin against Staphylococcus aureus and Enterococcus faecalis are not mediated by inhibition of lipoteichoic acid biosynthesis.

Antimicrob. Agents Chemother. 47: 2682-2684.

Lai, C-Y., and J.E. Cronan. (2003). 3-ketoacyl carrier protein synthase III (FabH) is essential for bacterial fatty acid synthesis. Journal of Biological Chemistry 278: 51494-5 1503.

Lakey, J.H. Maget-Dana, R., and Ptak, M. (1988). Conformational change on binding by the lipopeptide antibiotic amphomycin. A C.D. and monolayer study. Biochem. Biophys. Res. Comm. 150: 384-390.

Lakey, J.H., Lea, E.J., Rudd, B.A., Wright, H.M., & Hopwood, D.A. (1983).

A new channel-forming antibiotic from Streptomyces coelicolor A3(2) which requires calcium for its activity. J. Gen. Microbiol, 129: 3565-3673.

Lakey, J.H. Maget-Dana, R., and Ptak, M. (1989) The lipopeptide antibiotic A21978C has a specific interaction with DMPC only in the presence of calcium ions. Biochem. Biophys. Acta. 985: 60-66.

Lakey, J.H. and Ptak, M. (1988). Fluorescence indicates a calcium-dependant interaction between the lipopeptide antibiotic LY146032 and phospholipid membranes. Biochemistry 27: 4639-4645.

Maget-Dana, R., Lakey, J.H and Ptak, M. (1988). A comparative monomolecular film study of antibiotic A21978C homologues of various lipid chain length. Biochem. Biophys. Acta. 962: 2 10-207.

Miao, V., Brost, R., Chapple, J., She, K., Ceoffet-Le Gal., M-F., and Baltz, R.H. (2006). The lipopeptide antibiotic A54145 biosynthetic gene cluster from Streptomycesfradiae. J. md. Microbiol. Biotechnol. 33: 129-140.

Miao, V., Coeffet-Le Gal., M-F., Nguyen, K., Brian, P., Penn, J., Whiting, A., Steele, J., Kau, D., Martin, S., Ford, R., Gibson, T., Bouchard, M., Wrigley, S.K., and Baltz, R.H. (2006). Genetic engineering in Streptomyces roseosporus to produce hybrid lipopeptide antibiotics. Chemistry & Biology.

13, 269-276.

Micklefield, J (2004). Daptomycin structure and mechanism of action revealed. Chemistry & Biology, 11: 887-888.

Moche, M, Dehesh, K., Edwards, P., and Lindqvist Y, (2001). The crystal structure of J3-ketoacyl-carrier protein synthase II from Synechocystsis sp. At 1.54A resolution and its relationship to other condensing enzymes. J. Mol Biol. 305: 49 1-503.

Mochet, M, Schneider, G., Edwards, P., Dehesh, K., and Lindqvist Y, (1999). Structure of the complex between the antibiotic cerulenin and its target, 3-ketoacyl carrier protein synthase. J. Biol. Chem. 274. 6031-6034.

Mootz, H.D., Schwarzer, D., and Marahiel, M.A. (2000). Construction of hybrid peptide synthetases by module and domain fusions. P.N.A.S. 97, 5848-5853.

Mootz, H.D., Schwarzer, D., and Marahiel, M.A. (2002). Ways of assembling complex natural products on molecular nonribosomal peptide synthetases. ChemBioChem. 3, 490-504.

Naganawa, H., Hamada, M., Maeda, K., Okami, Y., Takeushi, T., & Umezawa, H. (1968). Laspartomycin, a new anti-staphylococcal peptide. J. Antibiot. 21: 55-62.

Naganawa, H., Takita, T., Maeda, K., & Umezawa, H. (1970). A novel fatty acid from laspartomycin. J. Antibiot. 23: 423-424.

Neaiy, J.M., Powell, A., Gordon, L., Mime, C., Flett, F., Wilkinson, B., Smith, C.P. and Micklefield, J. (2007). An asparagine oxygenase (AsnO) and a 3-hydroxyasparaginyl phosphotransferase (HasP) are involved in the biosynthesis of calcium-dependant lipopeptide antibiotics. Microbiology 153: 768-776.

Nguyen K.T., Ritz, D., Gu, J-Q., Alexander, D., Chu, M., Miao, V., Brian, P., and Baltz, R.H. (2006). Combinatorial biosynthesis of novel antibiotics related to daptomycin. P.N.A.S. 103, 17462-17467.

Penn, J., Li, X., Whiting, A., Wrigley, S.K. Latif, M. Gibson, T., Silva, C., Davies, J.E., Miao, V., Brian, P., and Baltz, R.H. (2005). Heterologous production of daptomycin in Streptoniyces lividans. Journal of Industrial Microbiology, 33: 12 1-128.

Powell, A. J. (2007).PhD thesis. University of Manchester, (Manchester).

Powell, A., Borg, M., Amir-Heidari, B., Neary, J.M., Thiriway, J., Wilkinson, B., Smith, C.P., and Micklefield, J. (2007). Engineered biosynthesis of nonribosomal lipopeptides with modified fatty acid side chains. J.A.C.S. 129, 15182-15 191.

Rudd, B.A.M., & Hopwood, D.A. (1980). A pigmented mycelial antibiotic in Streptomyces coelicolor A3(2). J. Gen. Microbiol, 96: 289-297.

Sambrook, J. and Russell, D.W. (2001). Molecular Cloning: A Laboratory Manual. (Cold Spring Harbor, N.Y: Cold Spring Harbor Laboratory).

Sgarbi, P.W.M., Boyd, V.A., Cameron, D.R., Chen, Y., Jia, Q., Nodwell., M, Siu, R., Sun, J., Wacowich-Sgarbi, S.A. Wang, L., Borders, D.B., Curran, W.V., & Leese, R.A. (2005). American Chemical Society Meeting, San Diego, CA, March 1317th Shay, A.J., Adam, J.A., Martin, J.H., Hausmann, W.K., Shu, P., and Bohonos, N. (1960). Aspartocin. I. Production, isolation, and characteristics Antibiot. Annu. 7: 194-198.

Shibata, M., Kanzaki, T., Nakazawa, K., Inoue, M., Hitomi, H., Mizuno, K., Fujino, M., & Akira, M. (1962). On glumamycin, a new antibiotic. J. Antibiot. 15: 1-6.

Shoji, J.I., Kozuki, S., Okamoto, S., Sakazaki, R., & Otsuka, H. (1968).

Studies on tsushimycin. I. Isolation and characterization of an acidic acylpeptide containing a new fatty acid. J. Antibiot. 21: 439-443.

Shoji, J., & Otsuka, H. (1969). Studies on tsushimycin. II. The structures of constituent fatty acids. J. Antibiot. 22: 473-439.

Silverman, J.A., Perimutter, N.G., & Shapiro, H.M., (2003). Correlation of daptomycin bactericidal activity and membrane depolarization in Staphylococcus aureus. Antimicrob. Agents Chemother 47: 2538-2544.

Stacheihaus, T., Mootz, H.D. and Marahiel, M.A. (1999). The specificity-conferring code of adenylation domains in nonribosomal peptide synthetases. Chemistry & Biology. 6, 493-505.

Strieker, M., Kopp, F., Mahlert, C., Essen, L-O., and Marahiel, M.A. (2007).

Mechanistic and structural basis of stereospecific 3-hydroxylation in Calcium Dependant antibiotic, a daptomycin-type lipopeptide. ACS Chemical Biology, 2: 187-196.

Takeshita, S., Sata, M., Matsahashi, W., & Hashimoto-Gotoh, T. (1987).

High-copy-number and low-copy-number plasmid vectors for lacZ alpha complementation and chloramphenicol or kanamycin resistance. Gene, 61: 63-74.

Uguru, G.C. Mime, C., Borg, M., Flett, F., Smith, C.P., and Micklefield, J. (2004). Active-site modifications of adenylation domains lead to hydrolysis of upstream ribosomal peptidyl thioester intermediates, J.A.C.S. 126, 5032-5033.

Va!, D., Banu, G., Sahadi, K., Lindqvist, Y., and Dehesh, K. (2000). Re-engineering ketoacyl synthase specificity. Structure. 8. 565-566.

Vertsey L., Ehiers, E., Kogler, H., Kurz, M., Meiwes, J., Seibert, G., Vogel, M., and Hammaann, P. (2000) Friulimicins: Novel lipopeptide antibiotic with peptidoglycan synthesis inhibiting activity from Actinoplanes friuliensis sp.nov. J. Antibiot. (Tokyo) 53, 8 16-827.

Wright, L.F., & Hopwood, D.A. (1976a). Actinorhodin, is a chromosomally-determined antibiotic in Streptomyces coelicolor A3(2). J. Gen. Microbiol, 96: 289-297.

Wright, L.F., & Hopwood, D.A. (1976b). Identification of the antibiotic determined by the SCPI plasmid of Streptomyces coelicolor A3(2). J. Gen.

Microbiol, 95: 96-106.

US65 11962 (Borders, D.B., Curran, W.V., Fantini, A.A., Francis, N.D., Jarolman, H., & Leese, R.A.) US6737403 (Borders, D.B., Curran, W.V., Fantini, A.A., Francis, N.D., Jarolman, H., & Leese, R.A.) US6750 199 (Borders, D.B., Curran, W.V., Fantini, A.A., Francis, N.D., Jarolman, H., & Leese, R.A.).

US3798 129 (Toth-Sarudy, E., Horvath, I., Gyimesi, J., Ott, I., Alfoldi, L., Berdy, J., Koczka, I., Scholtz, V., Szell, V., & Laszio E.).