Iron chelators from mycobacteria (1954–1999) and potential therapeutic applications Anne F. Vergne Andrew J. Walz and Marvin J. Miller* Department of Chemistry & Biochemistry University of Notre Dame 251 Nieuwland Science Hall Notre Dame IN46556-5670 USA 7 Received (in Cambridge UK) 11th August 1999 Covering 1954 to 1999 Previous review 1970 4 Isolation and characterization of exochelins and carboxymycobactins 8 Biosynthesis of mycobactins and exochelins 8.1 Isotopic labeling studies of mycobactin biosynthesis 8.2 Identification of gene clusters encoding for biosyntheses of mycobactin T and exochelins of Introduction Discovery of mycobactins Production extraction and purification of 123 mycobactins 3.1 Production 3.2 Extraction and purification 45 Structure elucidation of mycobactins M.smegmatis 9 Iron transport mechanisms in mycobacteria 9.1 Introduction 9.2 Mycobactin in the cell envelope 9.3 Iron sequestration and transport utilizing salicylate 9.4 Iron sequestration and transport utilizing citrate 9.5 Iron sequestration and transport utilizing exochelins 9.6 Iron sequestration and transport utilizing Possible therapeutic applications of mycobacterial carboxymycobactins 9.7 Iron regulated envelope proteins as possible receptors for extracellular ferrisiderophores 9.8 Iron release and storage in mycobacteria 10 and mycobacterial-like siderophores targeting the iron uptake mechanisms of mycobacteria 10.1 Introduction Physical and chemical properties of mycobactins 5.1 Ultraviolet absorption spectrum 5.2 Mass spectrometry 5.3 Chromatographic properties 5.4 NMR 6 Synthesis of mycobactins and mycobactin analogs 6.1 Synthesis of oxazoline derivative 4 6.2 Synthesis of hydroxamic acid units 5 and 7 6.3 Synthesis of mycobactic acid 2 6.4 Synthesis of cobactin 3 and analogs 6.5 Synthesis of mycobactin 1 and analogs Anne Vergne was born in France in 1975.She completed her undergraduate studies at CPE-Lyon France (School of Chemistry-Physics-Electronics). In 1997 she joined the research group of Professor Marvin J. Miller at the University of Notre- Dame IN. She is currently a third-year graduate student and her research interests focus on the synthesis of mycobactin analogues with potential antituberculosis and antitumor activities.Marvin Miller was born and raised in Dickinson ND. His interest in organic chemistry was inspired by the enthusiasm of Professor S. P. Pappas his undergraduate mentor at North Dakota State. Professor Marc Loudon introduced him to hydroxamic acid chemistry during Ph.D. studies at Cornell. Subsequent to an NIH postdoctoral fellowship with Professor Henry Rapoport at Berkeley Dr. Miller joined the faculty of Notre Dame in 1977 where he is now the George and Winifred Clark Professor of Chemistry. Current research interests include synthesis and studies of biologically important molecules. He is sustained by the love of his wife Patty and children Chris Katie Joe and Carl. Marvin Miller Andrew Walz received his B.S.degree in 1994 from Kent State University. His undergraduate research conducted under the direction of Dr. L.-C. Chien of the Liquid Crystal Institute dealt with the synthesis of polysiloxane based liquid crystals. He received his Ph.D. from the University of Virginia in 1999 under the direction of Dr. Richard J. Sundberg. His graduate research focussed on the synthesis of heterocyclic marine natural products and analogs as potential protein kinase C inhibitors. He is presently a postdoctoral research associate with Dr. Marvin J. Miller working on the synthesis of mycobactin analogs. He enjoys the loving support of his wife Elizabeth and two sons Jack and Frank. Anne Vergne Andrew Walz 99 Nat. Prod.Rep. 2000 17 99–116 This journal is © The Royal Society of Chemistry 2000 10.2 Mycobactins and mycobactin analogs as anti-mycobacterial agents Snow’s hypothesis 10.3 Other strategies targeting the iron uptake mechanisms of mycobacteria for possible therapeutic applications 10.4 Siderophore-mediated drug transport as a means for treating mycobacterial infections Other possible therapeutic applications of 11 mycobacterial siderophores Conclusion Acknowledgments References 12 13 14 1 Introduction Iron is an essential nutrient for virtually all forms of life. Indeed iron is utilized in oxygen transport electron transport and other fundamental and life sustaining biological functions. Thus the acquisition of iron by an organism is necessary for survival.Siderophores Greek for ‘iron carriers’ are low-molecular weight iron chelating compounds generated by plants microbes and even higher organisms for the acquisition of iron.1–3 Mycobacteria synthesize and utilize at least three types of siderophores the mycobactins exochelins and carboxymycobactins. Mycobactins were the first mycobacterial siderophores discovered and extensively studied. An excellent comprehensive review of these compounds was written by G. A. Snow in 1970.4 Since that time no comprehensive review has appeared in the literature covering all three known types of mycobacterial siderophores. Other reviews have been written concerning mycobacterial iron transport and metabolism involving these siderophores.5–7 This review will examine the past and current literature concerning the discovery isolation structure elucidation chemical properties synthesis biological function and therapeutic potential of these iron chelating natural products while highlighting important material covered in the existing reviews.Although this area of research is interesting as basic science the recent rise in mycobacterial infections worldwide most notably those due to Mycobacterium tuberculosis and M. avium adds significance to this review as will be discussed in the text. Table 1 General structure of mycobactins Nat. Prod. Rep. 2000 17 99–116 100 2 Discovery of mycobactins In the 1910’s M. paratuberculosis then called M. johnei was a microorganism of considerable interest and research because it had been determined to be the cause of cattle disease that resulted in considerable losses to farmers of Europe.However M. paratuberculosis and a bacillus found in the lesions of leprosy were the only two mycobacteria that could not be grown on laboratory media. In 1912 Twort and Ingram managed to grow M. paratuberculosis by adding killed human tubercle bacilli to the synthetic medium.8 So they suggested the existence of an essential growth factor from other mycobacteria that M. paratuberculosis was unable to produce itself. Subsequently extracts of M. phlei were added to synthetic media in order to promote growth of M. paratuberculosis but no attempt was made to isolate the unknown growth factor(s) until 1949 when Francis Macturk Madinaveitia and Snow isolated it from M.phlei.9 They managed to crystallize the growth factor as an aluminum complex but failed to isolate it in a metal-free form. A few years later in 1953 the same authors isolated from M. phlei the growth factor for M. paratuberculosis for which they suggested the name mycobactin.10 They described a new extraction method that allowed them to obtain pure mycobactin in a metal-free form. In 1954 the first structure of a mycobactin was determined by chemical degradation of mycobactin P and identification of its constituent fragments.11 Mycobactin P actually is a mixture of four similar chemical entities differing only in a fatty side chain. Two possible core structures were proposed,12 and it was only in 1965 that Snow13 determined the correct structure (See Table 1) established the absolute configuration of all of the stereocenters and obtained the active ferric complex of mycobactin P in crystalline form.In the same year the growth factor from M. tuberculosis was discovered by Marks.14,15 It was called mycobactin T in order to distinguish it from mycobactin P the growth factor from M. phlei.16 Mycobactin T was isolated by Snow from M. tuberculosis as an iron complex in 1965.17 As with mycobactin P it also was a mixture of at least four different components differing only in their fatty acid side chains; the structure of the main component was determined and appeared to be very similar to the structure of mycobactin P. In 1968 Snow and White utilized thin-layer chromatography of the ferric or aluminum complexes to separate and identify the different mycobactins isolated from different mycobacteria.18 They were able to determine that M.terrae M. marinum and M. smegmatis produce mycobactins that differ from each other and from mycobactins P and T. In order to distinguish the mycobactins a nomenclature was also suggested. Mycobactins possessing the same nucleus and forming a homologous series with various side-chain lengths form a family designated by a letter. Indication of the number of carbons in the side chain is used to differentiate the different members of a family. From these early studies it appeared that most species of mycobacteria grown in iron-deficient media produce mycobactins and that different strains of mycobacteria produce mycobactins with differing substructures.Subsequently mycobactins from several different strains of mycobacteria have been isolated and structurally compared to those previously characterized. 19 All the known structures of mycobactins are summarized in Table 1.19–22 M. paratuberculosis and other strains of pathogenic mycobacteria upon initial laboratory isolation require the addition of an external mycobactin growth factor for cultivation as mentioned previously. This ‘mycobactin dependence’ can be lost in subsequent generations following the initial cultivation resulting in mycobactin production by the isolants under appropriate conditions. This led to a controversy related to the discovery of the mycobactin produced by M.paratuberculosis23 since two different structures were proposed mycobactin J and mycobactin ‘J’.20,24,25 This controversy was resolved in 1992 by Barclay Furst and Smith26 who found that mutants of M. paratuberculosis were selected during growth under iron-deficient conditions suggesting that two different mycobactins could have been isolated by the two groups. 3 Production extraction and purification of mycobactins Isolation of mycobactins requires growth of mycobacteria followed by extraction and purification of the mycobactins. Various methods used for these three steps are only briefly summarized and updated in this section since complete early details were provided by Snow.4 3.1 Production After the first isolation of a mycobactin in 1953,10 satisfactory liquid media were developed for the growth of mycobacteria and improved18 by White and Snow based on the early observation that the microorganisms had to be grown under iron-deficient conditions.4 These media contained KH2PO4 Na2HPO4 glycerol and asparagine in water.18 In a few cases of mycobactin-dependent mycobacteria such as certain strains of M.paratuberculosis M. avium and M. leprae it was necessary to add some mycobactins to the growth medium. In order to insure iron deficiency the glassware had to be carefully treated with acid. In 1982 Hall and Ratledge proposed a convenient method for the production of mycobactins which did not require acid cleaning of the glassware.27 The usual glycerol– asparagine–salt medium was solidified using agar leading to a much more convenient procedure.With both the stirred liquid or solid synthetic media growth of the microorganisms was slow. However in 1993 a new method was developed to produce mycobactins in high iron concentration conditions in the presence of ethylenediaminodi(o-hydroxyphenylacetic acid) (EDDA) an iron chelator.28 The authors described a twostep method where the mycobacteria were initially grown in an iron-containing medium and then shifted to an EDDAcontaining medium for the mycobactin production phase. 3.2 Extraction and purification Francis Snow and co-workers developed the first method to extract and purify a mycobactin in a metal-free form10 but in 1965 the attempted isolation of mycobactin T with the same method failed because it was entirely present as the ironcomplex.17 In 1969 White and Snow described general methods applicable to most of the mycobactins that allowed obtention of either ferric mycobactins or mycobactins in a metal-free form.4,22 These methods are still the basis of procedures used today.24,29 Eventually simplified purification methods of the mycobactins were reported by Ratledge and Hall in 198227,30 and 1986.31 These consist in scraping the cells from the solid medium extracting with ethanol and after treatment with FeCl3 further extraction with chloroform and methanol. It should be re-emphasized that the criterion of purity of mycobactins is not the presence of a single chemical compound since all but mycobactins A and R are isolated as mixtures of homologues differing only in the length of the fatty acid side chain.The criterion of purity refers then to the absence of mycobactins with a different nucleus or any other organic molecule. Except for M. marinum which produces separable mycobactins M and N and M. fortuitum which produces inseparable mycobactins F and H mycobacteria produce only one major type of mycobactin. Traces of other types are removed by purification. Separation of homologous components has been effected for the major component of mycobactin P by counter-current distribution18 and more recently HPLC separation of homologous components of mycobactin S has been accomplished (See Section 5.3).4 Structure elucidation of mycobactins The first structure elucidation of a mycobactin was determined by Snow in 1954 through chemical degradation of mycobactin P (Scheme 1).11 Treatment of mycobactin P under mildly basic conditions cleaved the ester linkage affording two compounds mycobactic acid P and cobactin P. Subsequent acid hydrolysis of mycobactic acid P yielded (a) 6-methylsalicylic acid which underwent degradation to afford m-cresol and carbon dioxide; (b) a b-hydroxy amino acid (serine); (c) Ne-hydroxy-L-lysine and (d) a mixture of long chain fatty acids in most cases a,b unsaturated. Acid hydrolysis of cobactin P afforded two fragments (e) a b-hydroxy acid and (f) another Ne-hydroxy-llysine. From this chemical degradation two structures were possible.12 In 1965 Snow determined by oxidation of mycobactic acid P with periodate that the side-chain was attached by a hydroxamic acid linkage rather than an amide linkage.13 The stereochemistry of all asymmetric centers was also determined by comparison of the optical rotations UV and IR spectra and GC retention times of the degradation products with reference compounds.The same methods of chemical degradation were applied to determine the structures of mycobactin T17 and mycobactins S and H.22 A method to determine the structure of mycobactins by NMR,32 developed by Greatbanks and Bedford complemented the existing chemical degradation methods. Similarly in 1982 in order to elucidate the structure of mycobactin J,25 McCullough and Merkal used mass spectroscopy and NMR (proton and carbon) of both the intact molecule and its fragment cobactin J1.By comparing their results with the known data for mycobactin P the authors determined the structure of mycobactin J. Fragments (c) and (f) were as for the other mycobactins forms of Ne-hydroxy-l-lysine. The b-hydroxy fragment (e) 2,4-dimethyl-3-hydroxypentanoic acid differed from previously known cobactins because it contained an isopropyl group. In summary two types of mycobactins can be distinguished mycobactins P T A R F H S and J have the same core structure but differ only in details of substitution and absolute stereochemistry. However mycobactins M and N from M. marinum differ from the others by having only small acyl groups at the hydroxamic group of the mycobactic acid moiety.101 Nat. Prod. Rep. 2000 17 99–116 Scheme 1 Chemical degradation of mycobactin P (Reproduced by permission from G. A. Snow Bacteriol. Rev. 1970 34 99). In 1974 Hough and Rogers determined the X-ray structure of ferrimycobactin P which delineated the coordination of iron by mycobactins.33 Ferrimycobactin can be considered as a sphere of diameter between 11 and 14 Å. Six atoms five oxygens and one nitrogen coordinate the Fe(iii) atom and form a particularly strained octahedron. The metal atom lies in a ‘V-shaped cleft’. According to the authors the dimensions of the octahedron can explain both the ‘high stability of the iron-complex of mycobactins and the ease of release of the iron atom’.Studies on the scandium yttrium and lanthanum complexes of mycobactin S showed that these IIIb cations form 1+1 complexes with mycobactin S. Semi-empirical calculations predict nonoctahedral configuration of the siderophore about these metal cations contrary to the iron(iii) complexes.34 5 Physical and chemical properties of mycobactins This section will give a brief overview of the physical and chemical properties of mycobactins reviewed in detail by Snow4 and will present the results reported since then. When isolated pure in a metal-free form all known mycobactins are white powders whose melting points are definite occurring between 155 °C and 175 °C.10,17,21,22 Mycobactins have very poor solubility in water (5 to 15 mg mL21 at 20 °C).Their solubility in non-polar solvents such as ether benzene or aliphatic hydrocarbons is low but they are somewhat soluble in ethanol (2% at 20 °C) and even more soluble in chloroform. 5.1 Ultraviolet absorption spectra The main structural features accounting for the UV-adsorption of mycobactins are the hydroxyphenyl-oxazoline residue and the acyl-hydroxamic acid groups. Two general types of spectra are obtained one with lmax in methanol at 243 249 and 304 nm if the benzyl group is substituted and lmax at 250 and 311 nm if not.4 Nat. Prod. Rep. 2000 17 99–116 102 5.2 Mass spectrometry Mass spectra of aluminum complexes of mycobactins can be easily obtained and are very useful for identification of mycobactins.4,21 First the parent peaks allow a determination of the molecular weight.As indicated earlier mycobactins are in most cases mixtures of homologues differing only in their side chain. Since the components have similar volatility the relative abundance can be measured. Eventually the first degradation peak permits the classification of the mycobactin being studied as either the P-type or the M-type. For the P-type mycobactins the first fragmentation consists of the loss of the side chain beyond the a,b-double bond whereas the M-type mycobactins lose their cobactin fragment leaving the mycobactic acid part of the parent ion. 5.3 Chromatographic properties In 1968 White and Snow described a method to separate and identify the different mycobactins4,18,21 that utilized isopropan- 2-ol for elution during silica gel thin-layer chromatography of the corresponding aluminum and ferric complexes.The authors found that since the side-chain length does not affect the Rf values the migration depends on the nature of the mycobactin nucleus allowing a separation according to the source mycobacteria. The use methanol–water reversed phase HPLC by Ratledge and Ewing further facilitated characterization35 without requiring pure metal-free mycobactins as in the case of NMR. In contrast to the earlier TLC methods HPLC also allowed separation of mycobactins differing in the unsaturation or the length of the pendant acyl-chain and not only in the nucleus. The power of HPLC was demonstrated by the successful separation of the seven components of mycobactin S.5.4 NMR Although not routinely utilized for mycobactin characterization after extraction another convenient way of identifying myco-bactins is nuclear magnetic resonance of metal-free mycobactins. 32 For all of the mycobactins studied by Greatbanks and Bedford the proton peaks were divided into the same five regions of the spectrum. The aromatic protons absorb in the 7.7–6.5 ppm region; the olefinic protons the protons a to an oxygen atom and the protons a to two groups such as a nitrogen atom or a carbonyl group are between 6.0 ppm and 4.0 ppm. Protons a to a nitrogen atom appear at 4.0–3.0 ppm and allylic protons at 3.0–2.0 ppm. They established a simple scheme (Scheme 2) allowing for convenient characterization of several mycobactins based on well-defined peaks characteristic of the different mycobactins.Scheme 2 Identification of mycobactins by NMR. Mycobactins F S T R H A and P can be identified via their NMR spectra by simple observation of the presence or absence of well-defined peaks (Reproduced by permission from D. Greatbanks and G. R. Bedford Biochem. J. 1969 115 1047). Combination of all these methods constitutes a precise fingerprinting method for the identification of the mycobactins. 36 Combined with efficient ways of producing and purifying mycobactins,27,31 these techniques make it possible to apply Snow’s suggestion of utilizing mycobactins as chemotaxonomic markers for mycobacteria. Indeed as different strains of mycobacteria produce different mycobactins two strains of mycobacteria producing structurally close mycobactins will be considered taxonomically related.The combination of TLC and HPLC led to the development by Hall,31 of an efficient method of identification of mycobactins. Mycobactin analysis was successfully applied in 1985 to the classification of armadillo-derived mycobacteria37 and in 1992 to the identification by Bosne and Levy-Febrault of M. fortuitum and M. chelonae subspecies (Table 2).38 Table 2 Sources of known mycobactins Source Mycobactin M. phlei M. tuberculosis (and M. bovis M. africanum) M. smegmatis (and M. Clitzky M. Gerston) M. thermoresitible and M. fortuitum M. aureum M. terrae M. fortuitum (and M. farcinogenes M. senegalense) M.marinum M. marinum M. paratuberculosis M. kansaii P T S H A R F M N J and ‘J’ K The identification of mycobacteria through the types of mycobactins produced was also adapted to facilitate clinical identification of mycobacteria. In order to propose a medical treatment it is of particular importance to be able to quickly identify the specific mycobacterial cause of an infection since different strains of mycobacteria have different responses to antibiotics. The method developed by Barclay Furst and Smith was based on TLC of 55Fe labeled mycobactins.26 The characterization could be done from 1 to 5 days depending on the isolated strain facilitating clinical identification of mycobacteria. A non-mycobactin based methodology for routine identification of mycobacteria is the GC analysis of the microbial fatty acids and mycolic acid cleavage products.39–42 Therefore joint analyses of mycobactins and mycolic acids by TLC developed in 1995 by Leite et al.afforded a simple test for the clinical identification of mycobacteria.43 6 Synthesis of mycobactins and mycobactin analogs After detailed studies Snow suggested that alternate or modified forms of mycobactins might serve as antagonists of mycobacterial growth by competitively binding the iron and/or inhibiting its assimilation by targeted forms of mycobacteria and thus might be of therapeutic value.4 Additionally design and synthesis of mycobactin analogs that contain potential ‘drug linkers’ may provide a new class of therapeutic agents.Direct attachment of antimicrobial agents to the analogs may allow the pendant drug to be targeted to mycobacteria. The importance of all these possible therapeutic applications initiated considerable interest in the total syntheses of mycobactins and analogs. A retrosynthesis for mycobactins was proposed44 based on the products of their chemical degradation. As shown in Scheme 3 disconnection of the ester bond of mycobactins 1 gives two fragments mycobactic acids 2 containing the acyclic hydroxamate residue and cobactins 3 containing the cyclic hydroxamate residue. Further disconnections of the amide bonds in fragments 2 and 3 result in four corresponding components oxazolines 4 hydroxamic acids 5 b-hydroxy acids 6 and Nehydroxycyclo-l-lysine 7.This section will present methods for the preparation of these fragments and/or protected versions of them. 6.1 Synthesis of oxazoline derivative 4 All syntheses reported so far related to the oxazoline fragment 4 focused on the core structure represented by compound 11 i.e. residue 4 with R2 = R3 = H corresponding to the synthesized mycobactins T and S. In 1972 St. C. Black and Wade reported several methods for the preparation of protected forms of 2-(2A-hydroxyphenyl)- 4,5-dihydrooxazole-4-carboxylic acid 11.45 Because of the stability of the oxazoline ring to mild hydrolytic conditions the first method employed by the authors was the hydrolysis of a corresponding ester (14 in Scheme 4). They prepared esters of type 14 by reaction of serine ester hydrochlorides 13 with ethyl 2-hydroxybenzimidate 12.This sequence allowed for the synthesis of an optically active oxazoline but the overall yield was not high (55%) and starting material 12 was formed from 2-hydroxybenzonitrile via its hydrochloride salt in only 27% yield. An alternative route involving cyclization of hydroxy amide 17 was accomplished by the same authors (Scheme 5). Amide 17 was obtained in high yield by the reaction of 2-hydroxybenzoyl azide 15 with serine 16 followed by conversion to the isopropyl ester. Cyclization by reaction with thionyl chloride afforded oxazoline ester 14a. While this second route seemed suitable for a high-yield synthesis of oxazolines unfortunately the use of a strong base in the preparation of 17 led to racemization.Eventually a milder procedure was investigated involving the treatment of the bis-p-toluenesulfonyl (tosyl) derivative 18 with base to give the tosyl oxazoline 19 which afforded the desired oxazoline acid 11 in 19% yield after hydrolysis (Scheme 6). These results constituted a good first approach of the oxazoline synthesis but needed improvements in yield and stereocontrol. 103 Nat. Prod. Rep. 2000 17 99–116 Scheme 3 Retrosynthetic analysis of mycobactins. Scheme 6 Dehydrative cyclization of the corresponding b-hydroxyamide for the preparation of (S)-2-[2-(benzyloxy)phenyl]- 4,5-dihydrooxazol-4-carboxylic acid was reported by Maurer and Miller in 1983 (Scheme 7).46 Thus 2-benzyloxybenzoic acid 20 was coupled with the methyl ester of l-serine 21 using dicyclohexylcarbodiimide (DCC) or 2-ethoxy-N-(ethoxycarbonyl)-1,2-dihydroquinoline (EEDQ).The optically pure oxazoline 23 was formed by treating the hydroxy amide 22 with SOCl2 at 215 °C in THF. Scheme 4 Scheme 7 Scheme 5 Eventually Hu and Miller developed another method using Burgess’s reagent for the cyclization of amide 24 (Scheme 8).47 2-(Benzyloxy)benzoic acid (20) was coupled with l-serine benzyl ester 13 to afford amide 24 in 90% yield. Treatment of amide 24 with Burgess’s reagent provided oxazoline 25 in 66% Nat. Prod. Rep. 2000 17 99–116 104 yield. This last method appears to be satisfactory as far as convenience and yield are concerned although it requires the preparation of Burgess’s reagent.Scheme 8 6.2 Synthesis of hydroxamic acid units 5 and 7 Mycobactins contain two hydroxamic acid units which are the cobactin ring system 27 and the mycobactic fragment 28 both derived from l-lysine 29. The syntheses of 27 and 28 are closely linked and will be discussed in this section. Early methods reported the synthesis of Ne-hydroxy-dllysine. In 1963 Roger and Neilands achieved the preparation of Ne-hydroxy-dl-lysine in the form of its 2-nitroindane-1,3-dione salt.48 In 1969 reaction of 5-(4-bromobutyl)hydantoin with hydroxylamine by Lancini Lazzari and Diena afforded the synthesis of Ne-hydroxy-dl-lysine.49 Eventually in 1972 St. C. Black Brown and Wade,50 in an attempted formation of the cobactin ring system obtained hydroxylamine 30 by treatment of diethyl 5-bromo-1-phthalamidopentane-1,1-dicarboxylate with benzaldoxime and subsequent transformation of the resulting nitrone to the corresponding hydroxylamine.In 1974 Isowa and Ohmori reported the preparation of optically active Ne-hydroxy-l-lysine 35 by mild deprotection of Ne-tosyl-Ne-benzyloxy-L-lysine 34 (Scheme 9).51 Precursor 34 was obtained by enzymatic resolution of Ne-acetyl-Ne-tosyl-Nabenzyloxy-dl-lysine 33. In 1982 Maurer and Miller described the synthesis of 38 a protected form of Ne-acetyl-Ne-hydroxy-dl-lysine from N-Boce-bromonorleucine benzyl ester 36 (Scheme 10).52 Protected Ne-acetyl-Ne-hydroxy-dl-lysine 38 was prepared by direct alkylation of O-benzyl acetohydroxamate 37 with bromide 36 which gave predominant N-alkylation.In 1993 J. P. Genet et al. applied palladium-catalyzed amination of an allylic ester with (N,O)-bis-Boc protected Scheme 9 Scheme 10 hydroxylamine to synthesize Ne-hydroxy-l-lysine a component of mycobactin T.53 Allyl-l-glycine 39 was functionalized in six steps54 to give allylic carbonate 40 (Scheme 11). The palladium-catalyzed amination of 40 in the presence of (N,O)- bis-Boc-protected hydroxylamine and subsequent deprotection provided Ne-hydroxy-l-lysine dihydrochloride 42. Scheme 11 The first studies related to syntheses of the terminal cyclic hydroxamate of the mycobactins were by St. C. Black Brown and Wade who reported three different approaches reductive cyclization ring expansion and ring oxidation.55 The first route involved reductive cyclization of a hydroxyimino diester by treatment with Fe–HCl.This afforded seven-membered hydroxamic acid 31 but in impure form and very low yield (1.5%). Using a ring expansion strategy the authors synthesized a-bromo cyclic hydroxamic acid 32 in 3% yield from the reaction of the corresponding a-bromocyclohexanone with benzenesulfonohydroxamic acid. Eventually the third strategy involved a ring oxidation. Treatment of an a-bromo lactam with triethylammonium fluoroborate followed by mCPBA afforded pure a-bromo cyclic hydroxamic acid 32 in 3% overall yield. Despite the low yields 32 was considered a precursor to a cobactin. In 1981 Maurer and Miller reported another route to the cobactin ring.56 The carboxyl group of 43 was converted to an O-benzyl hydroxamate by coupling with O-benzylhydrox- 105 Nat.Prod. Rep. 2000 17 99–116 ylamine (Scheme 12). When 44 was treated with a slight excess of triphenylphosphine (PPh3) and diethyl azodicarboxylate (DEAD) intramolecular alkylation took place leading to the desired protected hydroxamic acid 45. Scheme 12 This method required the synthesis and enzymatic resolution of the l-e-hydroxynorleucine precursor of 43 and the subsequent cyclization was complicated by N- vs. O-selectivity. To circumvent these problems in 1995 Hu and Miller applied a direct oxidation57,58 of primary amines by dimethyldioxirane to the synthesis of both hydroxamic acid units of the mycobactins Ne-acetyl-Ne-hydroxy-l-lysine and the seven-membered cyclic hydroxamic acid.Thus esters of Na-Cbz-l-lysine (46) were oxidized by dimethyldioxirane (DMD) to provide nitrones 47 (Scheme 13). Scheme 13 From nitrones 47 both hydroxamic acid units could be prepared. For the synthesis of the linear Ne-acetyl-Ne-hydroxyl-lysine the crude nitrone product was treated with hydroxylamine hydrochloride salt and directly acetylated with acetic anhydride in pyridine in order to avoid isolation of unstable hydroxylamine 48 (Scheme 14). This acetylation afforded Scheme 14 desired hydroxamic acid 49a in a 51% yield along with N,Odiacetylated product and O-acetylated product. Use of palmitoyl chloride in the acylation step produced 50b a key ironbinding component of the mycobactins that contains the natural long acyl group and facilitated the total synthesis of mycobactin S.47 Since the acylation of hydroxylamine 48 also produced byproducts from N,O-diacylation and O-acylation a slight Nat.Prod. Rep. 2000 17 99–116 106 variation was reported by Xu and Miller in 1998.59 Treatment of hydroxylamine 48 with a large excess palmitoyl chloride in the presence of NaHCO3 resulted in complete N,O-diacylation of the hydroxylamine moiety. Subsequent selective removal of the O-palmitoyl hydroxamate group by treatment with Hünig’s base afforded hydroxamic acid 49b cleanly. Treatment of the related hydroxylamine 51 obtained from 47b with DCC DMAP (dimethylaminopyridine) and DMAP·HCl provided the desired cyclic hydroxamic acid 52 which was immediately protected to give compound 53 (Scheme 15).57 Scheme 15 6.3 Synthesis of mycobactic acid 2 With the preparation of oxazoline 4 and linear hydroxamic acid 5 the synthesis of mycobactic acid 2 one of the two major fragments of mycobactins could be realized.Different coupling agents have been used that allow for various functional group protection strategies. In the early synthesis of a non-iron binding mycobactin analog Carpenter and Moore60 coupled dl-2-phenyl-4,5-dihydrooxazoline-4-carboxylic acid 54 with the protected lysine ester 55 in the presence of DCC. Subsequent saponification of the product afforded the desired mycobactic acid analog 56 (Scheme 16). Scheme 16 In the paper by Maurer and Miller,46 the first intention of the authors was to couple the oxazoline fragment 26 (Scheme 8) with amines but all attempts with various coupling agents failed.This result contrary to the coupling realized by Carpenter and Moore was attributed to the presence of the bulky benzyloxy group on the phenyl ring. Thus instead of forming the oxazoline before coupling as originally suggested acyclic dipeptide 59 was assembled in high yield (92%) by coupling of 57 and 58 with EEDQ in chloroform (Scheme 17). Compound 59 was subsequently cyclized in the presence of SOCl2 at –15 °C to give the oxazoline peptide 60 as a pure (S,S)- diastereoisomer. In 1997 Hu and Miller removed the benzyl-protecting group of the oxazoline before the coupling with amine 63.47 The coupling of oxazoline acid 62 with hydroxamic acid 63 using [1-[3-(dimethylamino)propyl]-3-ethylcarbodiimide hydrochloride] (EDC) afforded mycobactic ester 65 in 59% yield (Scheme 18).This confirmed the hypothesis that the presence of a large benzyloxy group in acid 26 was a major factor in the failure of the initial coupling attempts. Conditions developed by Xu and Miller allowed the coupling of the oxazoline 62 in the presence of EDC and 1-hydroxy-Scheme 17 Scheme 18 7-azabenzatriazole (HOAt) with amine 64 containing a free hydroxamic acid.59 The desired coupling product 66 was formed in 91% yield confirming the author’s hypothesis that on the basis of pKa values coupling reactions can be carried out under finely tuned conditions without masking certain functional groups.Thus effective syntheses of mycobactic acids have been developed. 6.4 Synthesis of cobactin 3 and analogs Different preparations of cobactin the second major fragment of mycobactin are described in this section. Carpenter and Moore initiated studies with synthesis of 69,60 again a non-iron chelating analog of cobactin T by coupling of dl-3-aminohexanolactam 68 with dl-3-hydroxybutyric acid hydrazide 67 in the presence of HNO3 which generates an intermediate acylazide (Scheme 19). After isolation of azopine derivative 70 Maurer and Miller synthesized protected cobactin T 74 46 by EEDQ-mediated coupling of 70 and d-b-hydroxybutyric acid 73 (Scheme 20). Scheme 19 Scheme 20 Subsequently Hu and Miller utilized a modified route in which amine 71 was coupled with (R)-3-hydroxybutyric acid (73) in the presence of DCC DMAP DMAP·HCl to provide OTBDPS (tert-butyldiphenylsilyl) protected cobactin T 75 in 63% yield.47 For the synthesis of mycobactin analogs with synthetic variation of the normal b-hydroxybutyrate component,59 four different cobactin derivatives were targeted.First free cobactin analogs 78 and 79 were prepared in 95% yield by coupling cyclolysine 72 and Cbz-protected-l-serine 76 (or Cbz-protected l-threonine 77) in the presence of EDC and HOAt (Scheme 21). Using HOAt over other traditional coupling additives provided significant enhancement of the reaction rate and yield while minimizing side reactions. Scheme 21 The same coupling conditions were used to couple amine 71 with carboxylic acids 80 or 81 in order to prepare protected cobactins 82 in 97% yield or 83 in 91% (Scheme 22).Scheme 22 6.5 Synthesis of mycobactin 1 and analogs Coupling of the two main fragments mycobactic acid and cobactin by an ester linkage afforded the natural mycobactins. Alternate incorporation of an amide linkage provided novel mycobactin analogs. Both approaches have been investigated and are discussed here. The route chosen by Carpenter and Moore for the synthesis of the non-chelating analog of mycobactin actually differs from the usual coupling between a mycobactic acid and a cobactin 107 Nat. Prod. Rep. 2000 17 99–116 fragment.60 Indeed they realized the coupling of fragments 56 and 69 by an ester bond in the presence of the coupling agent carbonyl diimidazole (CDI) and then attached the lipophilic C18 side-chain (Scheme 23).This route afforded desired mycobactin analog 84. Scheme 23 In 1983 the first total synthesis of an iron chelating mycobactin mycobactin S2 was reported by Maurer and Miller (Scheme 24).46 The product differed from natural mycobactin S by replacement of the long acyl chain by an acetyl group. But all of the chiral centers of mycobactin S were incorporated. The synthesis was performed in a convergent manner by forming the ester linkage between mycobactic acid 85 and cobactin 87 using the PPh3–DEAD-mediated ester bond formation. This method allowed incorporation of the correct stereochemistry at the ester linkage by inversion of the hydroxyl group of the bhydroxybutyrate component during the coupling.The protected mycobactin S2 (89) was isolated in 50% yield with the correct (S,S,S,S)-configuration corresponding to mycobactin S. The same Mitsunobu conditions were utilized during the synthesis of mycobactin S by Hu and Miller,47 where protected mycobactin S 90 was obtained in 49% yield by coupling 86 and 88. Scheme 24 Eventually Xu and Miller developed an improved synthetic approach which increased the overall chemical yield by Nat. Prod. Rep. 2000 17 99–116 108 circumventing unproductive protection and deprotection steps.59 Both ester and amide linkages between mycobactic acid 91 and cobactin analogs 92–95 were separately generated using EDC and HOAt (and DMAP in the cases of 94 95) (Scheme 25).Under these conditions coupling reactions in the presence of the free hydroxamic acid and/or the free phenolic hydroxy group consistently proceeded in excellent yields. Subsequent deprotection provided the target compounds. Scheme 25 7 Isolation and characterization of exochelins and carboxymycobactins In 1975 the search for an extracellular iron-binding compound produced by mycobacteria able to solubilize iron under physiological conditions led to the discovery of a new siderophore differing from mycobactins. From the nonpathogenic M. smegmatis Macham and Ratledge isolated a water-soluble iron-binding agent,61,62 whose structure was not completely determined but was found to be a polypeptide.The hypothesis of the authors was that it was a new extracellular siderophore that they called exochelin MS. The same year the same authors isolated another extracellular siderophore from the pathogenic M. bovis.63 Comparing exochelin MB from M. bovis and exochelin MS from M. smegmatis they noticed differences in properties related to extraction by organic solvents. Subsequently various extracellular iron-binding agents were isolated from both pathogenic and non-pathogenic strains of mycobacteria. Extracellular siderophores were isolated from M. avium in 1977,64 from M. avium M. intracellulare M. scrofulaceum and M. paratuberculosis in 198365 and from M. vaccae in 1986.66 In 1988 exochelins were isolated from M.tuberculosis M. bovis and M. africanum67 but were found to be the same as those previously known. All of these siderophores were extracted from microorganisms grown under iron-deficient conditions and purified by combination of ion-exchange chromatography and HPLC. It was only in 1995 twenty years after the discovery of the first exochelin that the structures of these extracellular siderophores were elucidated. Sharman Williams Ewing and Ratledge characterized both exochelins MS 102,68 from M. smegmatis and exochelin MN 103,69 from M. neoaurum. They are both polypeptides (Scheme 26). Exochelin MS is a formylated pentapeptide with unusual linkages between the amino acid components of the molecule. This absence of conventional peptide bonds suggests potential evolutionary design of resistance to peptidase-mediated hydrolysis.Exochelin MN is a unique hexapeptide that contains a b-hydroxyhistidine residue found in only one other siderophore. Scheme 26 Exochelins MS 102 and MN 103. The structures of extracellular siderophores isolated from pathogenic mycobacteria M. avium70 and M. tuberculosis29 also were reported in 1995 and called carboxymycobactins. They were found as mixtures of compounds whose structures differ radically from the peptide exochelins isolated from nonpathogenic mycobacteria but share a common core with the structure of their intra-cellular counterparts the mycobactins Scheme 27 Structure of carboxymycobactins isolated from M. avium 104 M. tuberculosis 105 and M.smegmatis 106. (Scheme 27). However they are smaller than mycobactins because of a shorter saturated or unsaturated side chain (R1) which usually terminates in a carboxylic acid or an ester.71 (This apparently depends on the length of the culture of the microorganism.)72 The modified side chain induces greater hydrophilicity to carboxymycobactins and accounts for their occurrence in aqueous extracellular media. The structures of mycobactins and carboxymycobactins were so similar that Gibson Horwitz et al. suggested that their core structure is synthesized by the same set of enzymes and that only one of the final steps determines if a mycobactin or a carboxymycobactin is produced.29 All of the structure elucidations of extracellular siderophores were accomplished by a combined application of derivatization MS and GC analyses and modern NMR techniques which even allowed the determination of the threedimensional structures in some cases.Thus extracellular siderophores synthesized by mycobacteria appeared to be of two different types the water-soluble polypeptides synthesized by non-pathogenic strains and the chloroform-extractable mycobactin-like siderophores called carboxymycobactins from pathogenic strains of mycobacteria. However in 1996 a carboxymycobactin was isolated from nonpathogenic strains of M. smegmatis as a second extracellular siderophore.73 Its ‘mycobactin-like’ structure was confirmed in 1998. The proportion of carboxymycobactin to the total amount of siderophores produced by M.smegmatis was estimated to be 5–10%. It is also interesting to note that the production of any of the three types of siderophores was maximal at the same iron concentration in the media. Eventually this last discovery eliminated the possibility that mycobacterial virulence was directly correlated with the presence of carboxymycobactins. The respective role of these three different iron-binding agents occurring in mycobacteria will be discussed in a following section. It should be noted that the literature can be confusing with regards to the terms exochelins and carboxymycobactins. Prior to the structural elucidations of the carboxymycobactins all extracellular mycobacterial siderophores were referred to as exochelins independent of their solubility differences.Herein the term carboxymycobactin refers to the water and chloroform soluble extracellular siderophores while exochelin refers to the water only soluble extracellular siderophores. 8 Biosynthesis of mycobactins and exochelins 8.1 Isotopic labeling studies of mycobactin biosynthesis Early biosynthetic studies of the mycobactins were restricted to isotopic labeling experiments. Allen and co-workers demonstrated the incorporation of [U-14C]-l-lysine into the cobactin and mycobactic acid fragments of mycobactin P. Suprisingly 17% of the mycobactic acid radioactivity was found in the serine of the oxazoline ring.74 The authors offered no explanation for the radioactive serine except that complete degradation of the radiolabeled lysine was unlikely.Tateson further demonstrated nearly equal incorporation of [U-14C]-l-lysine into both cobactin and mycobactic acid by M. phlei and M. smegmatis.75 Ne-Hydroxy-[U-14C]-l-lysine was not taken up by the cells and no conclusion as to a common hydroxylamine intermediate for the cyclic and linear hydroxamic acids could be made. Incubation of M. phlei and M. smegmatis with [2-14C]propionate and [1-14C]acetate respectively yielded high radioactivity in the cobactin linker fragments of the two mycobactins. Isotopic labeling studies on the aromatic region of mycobactin S were performed using [G-14C]-shikimic acid which led to the radiolabeling of the mycobactic acid and also to the extracellular salicylic acid excreted by M.smegmatis.76,77 Treatment of the same strain of mycobacteria with [carboxy- 14C]salicylic acid led to the incorporation of a radioactive salicylic acid moiety into mycobactin S. This demonstrated salicylate uptake by mycobacteria and its intermediacy in the biosynthesis of salicylate-containing mycobactins.78 The salicylic acid incorporated into the mycobactins has its origin through a chorismate to isochorismate to shikimate pathway.79 The 6-methylsalicylic acid-containing mycobactins (e.g. mycobactin P) do not utilize this pathway as shown by the failure to incorporate radiolabeled salicylic acid into the lipophilic siderophore and are thought to employ a polyketidebased biosynthesis of the aromatic moiety.80 No isotopic labeling studies involving the biosynthesis of the exochelins have been reported.109 Nat. Prod. Rep. 2000 17 99–116 8.2 Identification of gene clusters encoding for biosyntheses of mycobactin T and exochelins of M. smegmatis With the genome of M. tuberculosis H37Rv sequenced,81 Cole et al. postulated that a 24 kilobase region contained the mycobactin biosynthetic (mbt) genes. Further studies on this mbt gene cluster by Walsh and co-workers82 allowed for a more detailed proposal for the biosynthesis of mycobactin T. The gene cluster designated mbtA-J contains ten open reading frames. The proposed gene products are isochorismate synthase acetyl hydrolase salicylate-AMP ligase polyketide synthase lysine-N-oxygenase phosphopantetheinyl (PPT) transferase and non-ribosomal peptide synthetase.The nonribosomal peptide bond biosyntheses are thought to be mediated by the post-translational modification of aryl and peptide carrier proteins with PPT transferase. This would generate thioesters as the activated intermediates necessary to form the mycobactin peptide framework. The presence of genes encoding for isochorismate synthase salicylate-AMP ligase and polyketide synthases confirmed the isotopic labeling studies mentioned previously. It is unknown whether lysine-N-oxygenase oxidizes the terminal amine of the lysine residues before or after incorporation into mycobactin T. A proposed linear biosynthesis is initiated with salicylic acid and terminated with an intramolecular lactamization forming the cyclic hydroxamic acid.The authors were able to express four of the genes in E. coli and reveal the formation of a salicylate-aryl carrier protein complex mediated by a PPT intermediate. The genetic basis for exochelin biosynthesis has also been studied in M. smegmatis.83 Deletion studies revealed the presence of a formyl transferase gene designated fxbA as necessary for exochelin biosynthesis. Two genes fxbB and fxbC homologous with known peptide synthetase genes were identified in the same cluster.84 The presence of an exiT gene encoding for a putative ABC transporter suggests that it assists the transport and excretion of the exochelin from the cell.85 Thus as with mycobactin T a non-ribosomal-mediated synthesis of the peptide-based exochelin was postulated.An iron dependent mycobacterial ideR protein may regulate exochelin biosynthesis in M. smegmatis.86 An ideR mutant demonstrated derepressed exochelin biosynthesis under iron sufficient conditions. 9 Iron transport mechanisms in mycobacteria 9.1 Introduction The following section will discuss the current understanding and proposed models of the complex iron transport mechanisms (ITMs) in mycobacteria. Microbial iron acquisition has been extensively studied and reviewed.3 Three general mechanisms have been recognized. Chelation of extracellular iron by siderophores is followed by recognition and transport of the ferrisiderophore complex. This process has been shown to involve receptor proteins. Siderophore biosynthesis and expression of the receptor proteins are regulated by the presence/absence of iron.The two other processes involve low-affinity ferric-transport directly from host iron sources and Fe(iii) reduction to Fe(ii) prior to transport. Chelation and low affinity transport mechanisms have been proposed for mycobacteria. No evidence for the reductive transport has been found. It is well known that biologically essential Fe(iii) is insoluble under physiological conditions. Acquisition of iron by mycobacteria depends on the presence of soluble Fe(iii) complexes generated from host iron sources. This solubilized Fe(iii)- complex must then be sequestered by the mycobacteria and transported across the cell envelope to be utilized immediately or stored for future metabolic need.As will be discussed below the ITM utilized by a particular strain of mycobacteria seems to Nat. Prod. Rep. 2000 17 99–116 110 be determined by the soluble iron chelator-Fe(iii) complex and the pH of the extracellular environment. ITMs have been proposed involving ferric salicylate ferric citrate and ferriexochelins/ carboxymycobactins. Additional involvement of membrane bound proteins also have been examined and will be discussed. Scheme 28 updated and adapted from the one developed by Ratledge,5 represents the proposed modes of iron transport employed by these microbes. The involvement of two structurally distinct peptide based siderophores is fairly unique for microbial iron transport. Many of the processes shown have not been definitively proven demonstrating the necessity for more intense research in this area.Scheme 28 Iron transport in mycobacteria involving siderophores. A potential pitfall in the study of these ITMs is the use of in vitro mycobacterial systems. In order to obtain detectable amounts of mycobacterial siderophores in vitro studies have been consistently carried out under iron deficient conditions. Any results obtained from these studies will be difficult to correlate to the ITM of in vivo mycobacterial infections. Mycobacterial infections occur in the macrophages of host tissue and the state of iron sufficiency/deficiency and pH in the macrophages are not definitively known and may diminish any correlation between in vitro/in vivo ITMs. 9.2 Mycobactin in the cell envelope Mycobactins lipophilic water insoluble siderophores (iron complex Ks ~ 1036) are membrane bound iron chelators as determined by electron microscopy.87 Ratledge demonstrated their involvement in an ITM of M.smegmatis through radioactive 55Fe-mycobactin uptake.88,89 Their water insolubility precludes their being extracellular iron chelators. The role of mycobactins in vivo is not certain. Mycobactins have not been isolated from infections of M. leprae in armadillo liver,90 M. tuberculosis in mouse spleen M. avium from flamingo liver and spleen and M. paratuberculosis in bovine ileum.91 This absence may be due to many factors. The pH of the phagocytic vacuoles lies somewhere between 4.5 and 6.0. Lambrecht proposed that this pH range may allow for the dissociation of iron from some host iron carrier proteins generating a state of iron sufficiency thus eliminating the need for mycobactins.The relationship between the stage of an infection and iron sufficiency/deficiency is not known and the lack of any detectable mycobactins does not necessarily mean they were never involved in iron transport. They may be produced at an early stage of the infection and then degraded when a state of iron sufficiency was reached. Furthermore the infecting cells may not have had homogenous access to iron in vivo and different pockets of mycobacteria could be either iron deficient or sufficient. 9.3 Iron sequestration and transport utilizing salicylate Prior to growth M. smegmatis releases salicylic acid into the extracellular environment.Investigation into its role as an iron chelator mediating the transfer of iron to the lipid bound mycobactin was then studied.92 It was discovered that there is no ferric salicylate in the extracellular medium in the presence of phosphate ions. A model system was developed in which mycobactin was suspended in n-octanol as a lipid surrogate. Transfer of iron was efficiently accomplished. Since the phosphate concentration in macrophages is not known the possibility exists that salicylate may be an extracellular ironsequestering agent that transfers iron to mycobactin in the cell envelope. 9.4 Iron sequestration and transport utilizing citrate Citric acid is known to exist in human serum and also participates in the transfer of iron to and from transferrin.Thus it was examined as a participant in the ITM of M. smegmatis.93 Iron deficient and sufficient washed cell suspensions took in iron from the citrate complex without incorporating the citrate itself as seen through radiolabeling studies. The uptake of iron was unaffected by the addition of ferriexochelin and ferric salicylate. Biosyntheses of the exochelin and mycobactin S were unaffected by the presence of ferric citrate. These data suggest a distinct ITM. M. tuberculosis also readily acquired iron from ferric ammonium citrate.94 Mössbauer and EPR studies on whole cells provided evidence for the transient formation of ferrimycobactin in the membrane of M. smegmatis upon treatment with ferric citrate.This provided the first direct evidence for transfer of iron from extracellular ferric citrate to the membrane bound mycobactin. 95 9.5 Iron sequestration and transport utilizing exochelins The discovery of extracellular water-soluble mycobacterial siderophores the exochelins (iron complex Ks ~ 1025) initiated research into ITMs involving sequestration of iron by these compounds. Participation of the exochelins in an ITM necessitates their ability to remove Fe(iii) from host iron sources. Growth of M. smegmatis in serum containing transferrin was enhanced by the addition of exochelins. Furthermore these compounds were able to remove iron from ferritin at rates comparable to ethylenediaminetetraacetic acid (EDTA) and desferrioxamine.63 Additional evidence for exochelin-mediated transport of iron is found in the ability of M. smegmatis to grow with the extracellular iron sources ferritin or ferric phosphate contained in diffusion capsules.62 Inoculation with exochelin MS increased the rate of iron uptake in M. smegmatis after the fourth day.96 When the exochelins were discovered an ITM was proposed in which the exochelins act as extracellular iron scavengers migrate to the cell wall and transfer iron to the mycobactins. Attempts to prove this model resulted in two ITM models one of which does not involve mycobactin.97,98 Under conditions of low exochelin concentration M. smegmatis utilizes an energycoupled transport process inhibited by electron transport inhibitors uncouplers of oxidative phosphorylation and thiol reagents.This high affinity process apparently does not require mycobactin and entails direct transport of the ferriexochelin complex into the cell envelope. This view is supported by the lack of ferric-salicylate inhibition of ferriexochelin iron uptake and the existence of M. vaccae R877R that excretes exochelins but does not produce mycobactins.66 Further indirect evidence is found in the absence of any signals corresponding to ferrimycobactin in the membrane of M. smegmatis when examined by Mössbauer and EPR spectroscopy after treatment with the xenosiderophore hexapeptide ferricrocin.95 With elevated ferriexochelin concentrations a low affinity passive ITM was postulated by Ratledge.This process was inhibited by the presence of ferric salicylate and thus a mechanism involving iron transfer from ferriexochelin to mycobactin was assumed. Morrison proposed an alternative role for the exochelins in 1995.99 Due to their release late in the growth phase of the mycobacteria when iron demand is diminished iron scavenging by the exochelins for iron transport may not be necessary. The exochelins as secondary metabolites were postulated to be iron stress exhaustion peptides that are synthesized and excreted only to be recycled by the mycobacteria to delay cell death. 9.6 Iron sequestration and transport utilizing carboxymycobactins Relatively little research has been conducted on an ITM utilizing the carboxymycobactins water and chloroform soluble siderophores.It has been shown that these extracellular siderophores promote the growth of three strains of M. tuberculosis albeit slower than the exochelin promoted growth rate of M. smegmatis.96 In 1996 a study by Gobin and Horwitz demonstrated the removal of iron by purified carboxymycobactin from human iron carriers.94 In either 4+1 or 1+1 ratios 95% and 40% saturated human transferrin irreversibly donated iron to the carboxymycobactins and the same was true with iron lactoferrin. Saturation occurred at a maximum time of 3 hours. Much slower rates of iron transfer were found with the incubation of the carboxymycobactins with horse ferritin which required 2 days for saturation to occur. The same report revealed that incubation of M.tuberculosis cultured under iron deficient conditions with ferricarboxymycobactin promoted eight times more iron uptake relative to the control. The cells also did not obtain iron from human transferrin without exogenous ferricarboxymycobactin. Different ITMs may be present in virulent as opposed to nonvirulent mycobacteria based on the exclusive production of exochelins by non-virulent species and the production of carboxymycobactins by virulent species with the exception M. smegmatis.72,73 While chloroform-soluble carboxymycobactins from M. intracellulare and M. bovis promoted the growth of M. smegmatis a lack of growth promotion was found in M. intracellulare and M. bovis by the water soluble exochelins of M. smegmatis.100 9.7 Iron regulated envelope proteins as possible receptors for extracellular ferrisiderophores In 1987 Ratledge and co-workers noted the expression of 180 kDa 84 kDa 29 kDa and 25 kDa proteins in M.smegmatis under iron deficient conditions but not under iron sufficient conditions.101 They were termed the iron regulated envelope proteins (IREPs). Antibodies raised against the 29 kDa IREP inhibited exochelin-mediated iron uptake by 70% in iron deficiently grown cells and had no effect against cells grown in iron sufficient conditions. The expression of the IREPs and the biosynthesis of exochelins and mycobactins in M. neoaurum were maximized with iron concentrations below 0.5 mg mL21.102 The 29 kDa IREP of M. smegmatis was shown to associate with exochelin MS by affinity chromatography but isolation of the complex was hindered by the concomitant denaturation of the protein.103 Virulent strains of mycobacteria also express IREPs under iron deficient conditions in vitro and in vivo suggesting a role for these proteins as receptors for ferricarboxymycobactins.104,105 The in vivo studies revealed the presence of proteins that were expressed only under iron sufficient conditions in 111 Nat. Prod. Rep. 2000 17 99–116 vitro. This led Ratledge to propose that the state of iron concentration during the course of an infection fluctuates between iron sufficiency/deficiency and the presence of both types of envelope proteins can then be expected. Further studies need to be carried out to fully elucidate the role these proteins play in iron transport especially the 29 kDa IREP found to associate with exochelins of M.smegmatis that also is expressed in vivo from M. avium and M. leprae. Matzanke and co-workers tested growth promotion in M. smegmatis and M. fortuitum with various non-mycobacterial xenosiderophores.95 Many xenosiderophores similar in structure and iron chelation ability differed markedly in their promotion of mycobacterial growth. The authors concluded that the uptake process is not diffusion controlled but mediated by protein receptors. 9.8 Iron release and storage in mycobacteria For ITMs that utilize mycobactins it is necessary to release the iron from the lipophilic siderophore for metabolic use or further storage.Ratledge and co-workers reported NADPH dependent ferrimycobactin reductase activity from the extracts of M. smegmatis.106,107 The authors proposed that the reductive release of Fe(ii) was followed by complexation with salicylate and then used for whatever need was present in the cell. Cell extracts from other organisms (E. coli and Candida utilitis) were able to perform the same transformation thus indicating that a single enzyme may not be responsible for the reduction in mycobacteria. The presence of non-mycobactin-mediated ITMs led Ratledge to propose an iron storage role for mycobactins.97,98 The recent evidence for and discovery of iron storage proteins such as bacterioferritin in three species of mycobacteria (M. leprae,108,109 M.smegmatis,110 and M. avium111) may require a revision of this model. Furthermore the previously mentioned Mössbauer and EPR studies on whole cells of M. smegmatis upon treatment with ferric citrate revealed a transient existence of the ferrimycobactin intermediate and the rise of spectroscopic signals consistent with the presence of bacterioferritin stored iron.95 The presence of iron storage proteins in mycobacteria may explain the failure to isolate mycobactins in vivo. If a state of iron sufficiency existed when the tissues were isolated for mycobactin extraction iron saturated bacterioferritin may have triggered degradation of the mycobactins and repressed their biosynthesis. The presence of bacterioferritin in mycobacteria led Morrison to propose a cell to cell transport of iron.99 Mycobacteria as close packed colonies and mats could be linked by linear networks of mycobactins.This network could mediate reversible non-aqueous intercellular iron transport from the bacterioferritin of one cell to another. 10 Possible therapeutic applications of mycobacterial and mycobacterial-like siderophores targeting the iron uptake mechanisms of mycobacteria 10.1 Introduction The previous section highlighted the complexity of the iron transport mechanism in mycobacteria. The plethora of mechanisms available for the acquisition and storage of iron both known and possibly unknown emphasizes the necessity of this metabolic nutrient for mycobacterial survival and reproduction. Furthermore the course of microbial infection depends on the competition for iron between the host and microbe.112 Any disruption of mycobacterial iron acquisition may prove to be lethal to the microbes and offer a possible therapeutic regimen for the treatment of mycobacterial infections.A resurgence in the occurrence of tuberculosis (TB) infections worldwide113 and the emergence of drug resistant strains114 re-emphasizes the Nat. Prod. Rep. 2000 17 99–116 112 constant need for research into novel anti-mycobacterial agents. TB is the world’s leading killer among infectious diseases causing 3 million fatalities annually and approximately 2 billion people are currently infected. Coinfection with HIV has contributed to the rapid growth in TB infection rates worldwide.115,116 Other mycobacterial infections also have been found in HIV immuno-compromised patients most notably M. avium.117 New modes of treatment of TB and other mycobacterial infections are clearly necessary. One problem in the treatment of microbial infections is the inability of drugs to diffuse across the cell envelope and reach their intended targets. With mycobacterial infections this difficulty is particularly acute due to the nature of the mycobacterial cell envelope.118,119 The cell membrane is coated with a polymeric complex consisting of peptidoglycans covalently attached to a layer of arabinogalactans terminating with a densely packed array of mycolic acids. Glycolipids have been found to associate with the mycolic acids forming a second bilayer.This overall structure exhibits extremely low fluidity and diffusion of drugs across the envelope can be significantly retarded. Utilizing the siderophores of the mycobacteria and their analogs as a basis for anti-mycobacterial agents may be a means overcoming the drug permeability problem. The so called ‘Achilles’ heel’ in mycobacteria may lie in their metabolism of iron.5 The complexity of iron transport in mycobacteria suggests that targeting a specific aspect of the iron transport mechanism may not be entirely successful because the microbe will adapt by using another means of acquiring iron. This may indeed prove true unless there is a potential therapeutic target common to all the various mechanisms that is absolutely necessary for the mycobacterium’s survival.10.2 Mycobactins and mycobactin analogs as anti-mycobacterial agents Snow’s hypothesis Detailed studies and the structural elucidation of the mycobactins led Snow to suggest and demonstrate that antagonists to the growth of one species of mycobacteria may be found in the naturally occurring mycobactins of another.4 Depressed growth rates of M. paratuberculosis M. kansasii and M. tuberculosis were found when inoculated with either mycobactins M or N. Furthermore treatment of M. paratuberculosis with combinations of mycobactins P and either M or N led to growth rates significantly slower than treatment of the microbes with any of them individually. Although not therapeutically useful the hypothesis was proven.A more recent study on the growth inhibition by natural mycobactins revealed the growth inhibition of M. aurum by ferrimycobactins J and S at concentrations above 5 mM.120 The most striking example validating Snow’s hypothesis is found in recent work by Miller and Hu.47 Synthetic mycobactin S was examined for its effect on the growth of M. tuberculosis H37Rv. Initial studies indicated 99% growth inhibition at 12.5 mg mL21 and subsequently a minimum inhibitory concentration (MIC) of 3.13 mg mL21 was determined! This inhibition is even more intriguing when viewed mechanistically. Mycobactin S and T differ only at the configuration of the methyl group in the butyrate fragment of the mycobactin framework. Assuming that mycobactin S serves as a surrogate for mycobactin T in the iron uptake process it is reasonable to suggest that a very specific recognition event involving ferrimycobactin occurs at some point during iron transport.This event may happen at protein-mediated transfer of extracellularly sequestered iron to the membrane bound mycobactin or at a reductase-mediated transfer of iron to bacterioferritin or salicylate. Metal-analogs of the ferrimycobactins have also received attention as potential inhibitors of mycobacterial growth. Chromic complexes of mycobactin P inhibited the growth of M. paratuberculosis.4 A later study by Barclay and Ratledge found that some inhibition of growth was found relative to ferrimycobactins/ exochelins in M. tuberculosis and M. avium with a variety of non-ferric metal complexes.121 According to Ratledge such an approach as a therapeutic weapon would most likely fail in the long term due to the thermodynamic preference for ferric iron by mycobacterial siderophores over other metal ions.Also the metal ions themselves may prove to be toxic. Synthetic analogs of the mycobactins have also been examined for possible anti-mycobacterial activity. The first example was an analog of mycobactin T lacking the hydroxamic acid and phenolic functionalities needed for iron chelation. 60 The analog demonstrated no appreciable inhibition against M. tuberculosis H37Rv. Impressive results were reported by Miller and Xu in 1998 concerning synthetic mycobactin analogs as inhibitors of M. tuberculosis H37Rv.59 Incorporation of a benzyloxycarbonyl protected serine or threonine in the place the butyrate linker of mycobactin S provided analogs 96 and 97 with 44% and 48% growth inhibition respectively at > 12.5 mg mL21.Analog 100 with b-alanine linker proved to be weakly inhibitory (25% growth inhibition at > 12.5 mg mL21). Suprisingly the Boc protected serine linker with replacement of the ester with an amide linkage provided analog 101 even more potent than mycobactin S with an MIC of 0.2 mg mL21! The enhanced activity of 101 relative to 100 may be attributed to the addition of the bulky NHBoc that induces increased hydrophobicity and/or restricted rotation about the amide bond replacing the naturally occurring ester. The role of hydrophobic groups may be crucial to the development of mycobactin analogs because of their location in the cell envelope of mycobacteria.Snow’s hypothesis has been dramatically proven in the case of the mycobactins by the growth inhibitory activities of mycobactin S and analog 101 against M. tuberculosis H37Rv. Application of Snow’s hypothesis to the exochelins has not been tested since the structures of these extracellular siderophores have only recently been elucidated and no analog syntheses have yet been reported. The literature suggests that use of naturally occurring exochelins of one species of mycobacteria against another may not be successful since mycobacteria recognize such a variety of extracellular iron solubilizers for iron uptake. Synthetic exochelin/ carboxymycobactin analogs may provide useful growth inhibitors although the growth promotion of some species of mycobacteria has been seen with synthetic and natural nonmycobacterial siderophores.95 Thus further research is needed to find more potent mycobacterial siderophore-based inhibitors.Effective synthetic methodology for the mycobactins has set the stage for the generation of a library of analogs. With regards to the exochelins/carboxymycobactins synthetic methodologies need to be developed. Analogs could then be elaborated which target and irreversibly bind to a receptor protein e.g. 29 kDa IREP possibly involved in the transport of the ferriexochelins/ carboxymycobactins. 10.3 Other strategies targeting the iron uptake mechanisms of mycobacteria for possible therapeutic applications The known anti-TB drug p-aminosalicylate (PAS) was originally proposed as an inhibitor of folic acid biosynthesis.Yet it is relatively inactive against other bacteria. Brown and Ratledge studied the mechanism of action of PAS on M. smegmatis.122,123 Their data showed PAS inhibition of iron uptake by 50% at 0.33 mM. Iron dependent enzymes such as glycerol dehydrogenase and NADH-cytochrome c reductase showed reduced activity as well. Finally PAS affects the biosynthesis of mycobactin S. These data suggest that a primary mode of action of this drug is a disruption of the iron acquisition and utilization. Since PAS may partially exert its inhibitory action through a disruption in mycobactin biosynthesis further exploration of agents designed to inhibit the biosynthesis of the mycobacterial siderophores may be a viable route for the development of antiinfectious drugs.Isoiazid a known TB drug forms a covalent complex with an acyl carrier protein AcpM and a b-ketoacyl carrier protein synthase,124 indicating that synthases and carrier proteins involved in the biosynthesis of the mycobacterial siderophores may be possible targets. The use of modified intermediates in the biosynthetic pathway or development of novel inhibitors of the biosynthetic enzymes may reduce the amount of the siderophores present for iron acquisition. Furthermore anti-sense oligonucleotides could also be seen as potential inhibitors by targeting the expression of the previously mentioned biosynthetic genes.All mechanisms proposed for iron transport involve the reduction of Fe(iii) to Fe(ii) for storage and metabolic use. Targeting this reduction would seem to allow for a general antimycobacterial agent that could conceivably be effective independent of the iron transport mechanism utilized. Although this reduction is universal alone it does not represent the ‘Achilles’ heel’ of the mycobacteria. Any inhibitor of the reductase enzyme(s) involved or any of the previously proposed non-siderophore-based inhibitors must still transverse the cell envelope a process that tends to be inefficient in mycobacteria. Siderophore-mediated drug transport targeting the reductive removal of iron may provide a potent and reliable means of treating mycobacterial infections.10.4 Siderophore-mediated drug transport as a means for treating mycobacterial infections Siderophore mediated drug transport has been demonstrated in these laboratories.125–127 Furthermore nature has provided examples of iron transporters used to deliver toxic substances to bacteria in the albomycins,128–132 ferrimycin A1,133,134 and the salmycins.135 Direct attachment of anti-mycobacterial agents to a siderophore is represented in Scheme 29. Scheme 29 General structure of a siderophore–drug conjugate. This conceptually novel approach for the treatment of mycobacterial infections will allow for the generation of conjugates containing drugs targeting the reductase and biosynthetic enzymes mentioned previously anti-sense oligonucleotides and any other agents targeting new gene expression products found from the sequenced genome of M.tuberculosis. Known antibiotics that alone can not diffuse through the dense cell envelope of the mycobacteria could be re-evaluated as siderophore–drug conjugates. Natural and synthetic growth promoting siderophores that participate in iron transport are ideal candidates for the proposed conjugates. These siderophores must contain functionality allowing for rapid and mild synthetic attachment of the linker and drug. Depending on the drug target the linker must be designed for appropriate release from the siderophore. Thus amide linkers could be designed along with chemically labile linkers capable of being non-enzymatically hydrolyzed such as oximes or activated esters.11 Other possible therapeutic applications of mycobacterial siderophores The Fenton reaction is the generation of hydroxyl radicals through the interaction of water and Fe(ii).136 Hydrogen 113 Nat. Prod. Rep. 2000 17 99–116 peroxide and superoxide radicals can initiate Fenton reactions in vivo catalytic in iron(ii).137–139 With no naturally occurring hydroxyl radical scavengers cellular damage can be quite acute and this chemistry is thought to be the cause of reperfusion injury to ischemic organs. Desferrioxamine140 and the synthetic siderophore analogs spermetaxins and spermetaxols,127 are able to inhibit this type of oxidative damage. Carboxymycobactins from M.tuberculosis were examined and patented as cardiac reperfusion injury inhibitors.141,142 Desferricarboxymycobactins were markedly more active than desferroxamine in their ability to preserve systolic and diastolic left ventricular function and blood flow after a period of ischemia in isolated rabbit hearts. Furthermore concentrations of hydroxyl radical produced metabolites and cardiac enzyme lactic dehydrogenase were decreased. Chelation of ferric iron by carboxymycobactins in the cardiac cellular lipid compartments may be the basis for the activity in that lipid oxidation is thought to be a cause of reperfusion injury.143 The use of desferricarboxymycobactins for the treatment of atherosclerosis and vascular injury by prevention of smooth muscle proliferation has been patented.144 Undesired growth of smooth muscle cells plays a key role in the onset of a variety of vascular diseases including atherosclerosis restenosis following angioplasty or other vascular surgeries.Inhibition of this undesired growth could be a means of treatment. The lipid solubility of the carboxymycobactins is thought to impart potent and rapid antiproliferative effects in cell and organ based assays. Another property of the desferricarboxymycobactins is the inhibition of low density lipoprotein cholesterol oxidation that may prevent atherosclerosis. The mycobactins as a structural class of siderophores were originally thought to be exclusively produced by mycobacteria. This has been shown to be false by the isolation of mycobactinlike structures from other microbes.Compounds exhibiting chromatographic properties similar to the mycobactins were isolated from the genus Rhodococcus.145 Ratledge and coworkers isolated mycobactin-like structures from nocardia structurally differing from the mycobactins by the presence of an oxazole rather than an oxazoline ring.146–148 Ratledge and coworkers isolated mycobactin-like structures termed the nocobactins from nocardia structurally differing from the mycobactins by the presence of an oxazole rather than an oxazoline ring.146–148 More recently formobactin from Nocardia sp. strain ND20 exhibited anti-lipid peroxidation activity.149 This mycobactin like compound exhibited an IC50 of 0.65 mM against lipid peroxidation in rat brain homogenate.It also reduced l-glutamate toxicity neuronal cells (EC50 0.017 mM) and suppressed apoptotic cell death by the oxygen radical producer buthione sulfoximine. (EC50 0.072 mM) Scheme 30 These studies suggest that the mycobactins carboxymycobactins and analogs may possess potent anti-oxidative activity due to their lipid solubility and ability to chelate iron(iii). Nat. Prod. Rep. 2000 17 99–116 114 Antitumor activity has been found in the BE-32030 A-E isolated from Nocardia sp. A32030150 and the amamistatins A and B from an actinomycete.151 Scheme 31 A patent reported the growth prevention and destruction of T47D-YB human breast cancer cells grown in vitro upon treatment with the carboxymycobactins of M.tuberculosis.152 50 mM of carboxymycobactin resulted in a seven-fold reduction in growth and 90% destruction of the cancer cells with substantial shift to a low percentage of cells in the growth phase after four days. Mechanistically this activity was not examined but the authors concluded that iron chelation was only partly responsible for the anti-cancer activity. This conclusion was reached due to activity exhibited by ferricarboxymycobactin although it was substantially less potent than the desferri form. The significance of iron acquisition in tumor growth promotion or inhibition is beginning to be explored.153–155 The antitumor activities of these mycobactin like compounds suggest that studies concerning antitumor activity of the siderophores of the mycobacteria and their synthetic analogs are merited.The siderophores of the mycobacteria and analogs may possess a myriad of useful biological activities that either promote or inhibit processes involving iron. They could also be useful probes for elucidating the role of ferric iron in biological systems. 12 Conclusion Mycobactins exochelins and carboxymycobactins represent a fascinating class of microbial siderophores. Much is known but much more needs to be uncovered concerning their role in mycobacterial iron transport. Their simple ability to tightly bind ferric iron could lead to numerous therapeutic applications requiring additional synthetic and biological study. 13 Acknowledgments We gratefully acknowledge grant GM-25845 from the National Institutes of Health and the University of Notre Dame for support of our siderophore research.Scheme 32 14 References 1 Iron Transport in Microbes Plants and Animals eds. G. Winkelmann D. van der Helm and J. B. Neilands VCH Weinheim FRG 1987 and references therein. 2 R. C. Hider Structure and Bonding 1984 58 25. 3 M. L. Guerinot Ann. Rev. Microbiol. 1994 48 743. 4 G. A. Snow Bacteriol. Rev. 1970 34 99. 5 P. R. Wheeler and C. Ratledge in Tuberculosis Pathogenesis Protection and Control ed. B. R. Bloom American Society of Microbiology Washington DC 1994 353. 6 C. Ratledge in Iron Transport in Microbes Plants and Animals eds. G. Winkelmann D. van der Helm and J. B. Neilands VCH Weinheim 1987 p. 207. 7 C.Ratledge Microbiol. Ser. 1984 15 603. 8 F. W. Twort and G. L. Y. Ingram Proc. R. Soc. London Ser. B Biol. Sci. 1912 84 517. 9 J Francis J. Madinaveitia H. M. Macturk and G. A. Snow Nature (London) 1949 163 365. 10 J. Francis H. M. Macturk J. Madinaveitia and G. A. Snow Biochem. J. 1953 55 596. 11 G. A. Snow J. Chem. Soc. 1954 2588. 12 G. A. Snow J. Chem. Soc. 1954 4080. 13 G. A. Snow Biochem. J. 1965 94 160. 14 J. Marks J. Pathol. Bacteriol. 1953 66 151. 15 J. Marks J. Pathol. Bacteriol. 1954 67 254. 16 G. A. Snow Biochem. J. 1961 81 4P. 17 G. A. Snow Biochem. J. 1965 97 166. 18 A. J. White and G. A. Snow Biochem. J. 1968 108 593. 19 R. M. Hall and C. Ratledge J. Gen. Microbiol. 1985 131 1691. 20 R. Barclay D. F. Ewing and C. Ratledge J.Bacteriol. 1985 164 896. 21 G. A. Snow and A. J. White Biochem. J. 1969 115 1031. 22 A. J. White and G. A. Snow Biochem. J. 1969 111 785. 23 R. S. Merkal and W. G. McCullough Curr. Microbiol. 1982 7 333. 24 R. S. Merkal W. G. McCullough and K. Takayama Bull. Inst. Pasteur (Paris) 1981 79 251. 25 W. G. McCullough and R. S. Merkal Curr. Microbiol. 1982 7 337. 26 R. Barclay V. Furst and I. Smith J. Med. Microbiol. 1992 37 286. 27 R. M. Hall and C. Ratledge FEMS Microbiol. Lett. 1982 15 133. 28 S. Bosne F. Papa S. Clavel-Sérès and N. Rastogi Curr. Microbiol. 1993 26 353. 29 J. Gobin C. H. Moore J. R. Reeve Jr. D. K. Wong B. W. Gibson and M. A. Horwitz Proc. Natl. Acad. Sci. USA 1995 92 5189. 30 C. Ratledge in CRC Handbook of Microbiology 2nd edn eds.A. I. Laskin and H. A. Lechevalier CRC Press Boca Raton 1982 Vol. IV p. 575. 31 R. M. Hall Actinomycetes 1986 19 92. 32 D. Greatbanks and G. R. Bedford Biochem. J. 1969 115 1047. 33 E. Hough and D. Rogers Biochem. Biophys. Res. Commun. 1974 57 73. 34 Y. Andres H. J. MacCordick and J. C. Hubert Biol. Met. 1991 4 207. 35 C. Ratledge and D. F. Ewing Biochem. J. 1978 175 853. 36 R. M. Hall and C. Ratledge J. Gen. Microbiol. 1984 130 1883. 37 R. M. Hall and C. Ratledge FEMS Microbiol. Lett. 1985 28 243. 38 S. Bosne and V. V. Levy-Febrault J. Clin. Microbiol. 1992 30 1225. 39 S. Chou P. Chedore and S. Kasatiya J. Clin. Microbiol. 1998 36 577. 40 V. J. Lévy-Fébrault K. S. Goh and H. L. David J. Clin. Microbiol. 1986 24 835.41 M. Daffé M. A. Lanéelle C. Asselineau V. Lévy-Fébrault and H. David Ann. Inst. Pasteur/Microbiol. 1983 134B 241. 42 M. Luquin V. Ausina F. L. Calahorra F. Belda M. G. Barcelo C. Celma and G. Prats J. Clin. Microbiol. 1991 29 120. 43 C. Q. F. Leite A. M. W. Barreto and S. R. A. Leite Rev. Microbiol. 1995 26 192. 44 J. Hu PhD Dissertation University of Notre Dame 1996. 45 D. St. C. Black and M. J. Wade Aust. J. Chem. 1972 25 1797. 46 P. J. Maurer and M. J. Miller J. Am. Chem. Soc. 1983 105 240. 47 J. Hu and M. J. Miller J. Am. Chem. Soc. 1997 119 3462. 48 S. Rogers and J. B. Neilands Biochemistry 1963 2 6. 49 G. C. Lancini E. Lazzari and A. Diena Farmaco Ed. Sci. 1969 24 169. 50 D. St. C. Black R. F. C. Brown and M. J. Wade Aust. J. Chem.1972 25 2155. 51 Y. Isowa and M. Ohmori Bull. Chem. Soc. Jpn. 1974 47 2672. 52 P. J. Maurer and M. J. Miller J. Am. Chem. Soc. 1982 104 3096. 53 J.-P. Genet S. Thorimbert and A.-M. Touzin Tetrahedron Lett. 1993 34 1159. 54 Y. Ohfune K. Hori and M. Sakaitini Tetrahedron Lett. 1986 27 6079. 55 D. St. C. Black R. F. C. Brown and M. J. Wade Aust. J. Chem. 1972 25 2429. 56 P. J. Maurer and M. J. Miller J. Org. Chem. 1981 46 2835. 57 J. Hu and M. J. Miller Tetrahedron Lett. 1995 36 6379. 58 J. Hu and M. J. Miller J. Org. Chem. 1994 59 4858. 59 Y. Xu and M. J. Miller J. Org. Chem. 1998 63 4314. 60 J. G. D. Carpenter and J. W. Moore J. Chem. Soc. (C) 1969 1610. 61 L. P. Macham and C. Ratledge J. Gen. Microbiol. 1975 89 379. 62 L. P. Macham M.C. Stephenson and C. Ratledge J. Gen. Microbiol. 1977 101 41. 63 L. P. Macham C. Ratledge and J. C. Nocton Infect. Immun. 1975 12 1242. 64 K. A. McCready and C. Ratledge Int. J. Syst. Bacteriol. 1977 27 288. 65 R. Barclay and C. Ratledge J. Bacteriol. 1983 153 1138. 66 A. J. M. Messenger R. M. Hall and C. Ratledge J. Gen. Microbiol. 1986 132 845. 67 R. Barclay and C. Ratledge J. Gen. Microbiol. 1988 134 771. 68 G. J. Sharman D. H. Williams D. F. Ewing and C. Ratledge Biochem. J. 1995 305 187. 69 G. J. Sharman D. H. Williams D. F. Ewing and C. Ratledge Chem. Biol. 1995 2 553. 70 S. J. Lane P. S. Marshall R. J. Upton C. Ratledge and M. Ewing Tetrahedron Lett. 1995 36 4129. 71 D. K. Wong J. Gobin M. A. Horwitz and B. W. Gibson J. Bacteriol.1996 178 6394. 72 S. J. Lane P. S. Marshall R. J. Upton and C. Ratledge BioMetals 1998 11 13. 73 C. Ratledge and M. Ewing Microbiology 1996 142 2207. 74 M. Allen A. J. Birch and A. R. Jones Aust. J. Chem. 1970 23 427. 75 J. E. Tateson Biochem. J. 1970 118 747. 76 A. T. Hudson and R. Bentley Tetrahedron Lett. 1970 11 2077. 77 A. T. Hudson and R. Bentley Biochemistry 1970 9 3984. 78 C. Ratledge and M. J. Hall FEBS Lett. 1970 10 309. 79 B. J. Marshall and C. Ratledge Biochem. Biophys. Acta 1972 264 106. 80 A. T. Hudson I. M. Campbell and R. Bentley Biochemistry 1970 9 3988. 81 S. T. Cole et al. Nature 1998 393 537. 82 L. E. N. Quadri J. Sello T. A. Keating P. H. Weinreb and C. T. Walsh Chem. Biol. 1998 5 631. 83 E. H. Fiss S. Yu and W.R. Jacobs Jr Mol. Microbiol. 1994 14 557. 84 S. Yu E. Fiss and W. R. Jacobs Jr. J. Bacteriol. 1998 180 4676. 85 W. Zhu J. E. L. Arceneaux M. L. Beggs B. R. Byers K. D. Eisenach and M. D. Lundrigan Mol. Microbiol. 1998 29 629. 86 O. Dussurget M. Rodriguez and I. Smith Mol. Microbiol. 1996 22 535. 87 C. Ratledge P. V. Patel and J. Mundy J. Gen. Microbiol. 1982 128 1559. 88 C. Ratledge Biochem. Biophys. Res. Commun. 1971 45 856. 89 C. Ratledge and B. J. Marshall Biochim. Biophys. Acta 1972 279 58. 90 L. Kato Int. J. Lepr. 1985 57 58. 91 R. S. Lambrecht and M. T. Collins Microb. Pathog. 1993 14 229. 92 C. Ratledge L. P. Macham K. A. Brown and B. J. Marshall Biochim. Biophys. Acta 1974 372 39. 115 Nat. Prod. Rep. 2000 17 99–116 93 A.J. M. Messenger and C. Ratledge J. Bacteriol. 1982 149 131. 94 J. Gobin and M. A. Horwitz J. Exp. Med. 1996 183 1527. 95 B. F. Matzanke R. Böhnke U. Möllmann R. Reissbrodt V. Schünemann and A. X. Trautwein BioMetals 1997 10 193. 96 B. Raghu G. R. Sarma and P. Venkatesan Med. Sci. Res. 1993 21 773. 97 M. C. Stephenson and C. Ratledge J. Gen. Microbiol. 1979 110 193. 98 M. C. Stephenson and C. Ratledge Biochem. Soc. Trans. 1978 6 423. 99 N. E. Morrison Int. J. Lepr. 1995 63 86. 100 M. C. Stephenson and C. Ratledge J. Gen. Microbiol. 1980 116 521. 101 R. M. Hall M. Sritharan A. J. M. Messenger and C. Ratledge J. Gen. Microbiol. 1987 133 2107. 102 M. Sritharan and C. Ratledge FEMS Microbiol. Lett. 1989 60 183. 103 L. G. Dover and C.Ratledge Microbiology 1996 142 1521. 104 M. Sritharan and C. Ratledge Biol. Met. 1990 2 203. 105 B. Raghu and G. R. Sarma Biochem. Mol. Biol. Int. 1993 31 333. 106 K. A. McCready and C. Ratledge J. Gen. Microbiol. 1979 113 67. 107 K. A. Brown and C. Ratledge FEBS Lett. 1975 53 262. 108 R. G. Deshpande M. B. Khan D. A. Bhat and R. G. Navalkar FEMS Immunol. Med. Microbiol. 1995 11 163. 109 M. C. V. Pessolani D. R. Smith B. Rivoire J. McCormick S. A. Hefta S. T. Cole and P. J. Brennan J. Exp. Med. 1994 180 319. 110 S. Kikuchi M. Fukumoto and H. Takahashi Biosci. Biotechnol. Biochem. 1994 58 885. 111 N. F. Inglis K. Stevenson A. H. F. Hosie and J. M. Sharp Gene 1994 150 205. 112 Iron and Infection eds. J. J. Bullen and E. Griffiths Wiley New York 1987.113 A. M. Rouhi Chem. Eng. News 1999 77 (20) 52 (May 17 1999). 114 A. Rattan A. Kalia and N. Ahmad Emerg. Infect. Dis. 1998 4 195. 115 D.E. Snider Jr. M. Raviglione and A. Kochi in Tuberculosis Pathogenesis Protection and Control ed. B. R. Bloom American Society of Microbiology Washington DC 1994 p. 3. 116 S. Lucas and A. M. Nelson in Tuberculosis Pathogenesis Protection and Control ed. B. R. Bloom American Society of Microbiologists Washington DC 1994 503. 117 C. A. Benson Curr. Opin. Infect. Dis. 1994 7 95. 118 G. S. Besra and D. Chatterjee in Tuberculosis Pathogenesis Protection and Control ed. B. R. Bloom American Society of Microbiologists Washington DC 1994 p. 285. 119 N. D. Connel and H. Nikaido in Tuberculosis Pathogenesis Protection and Control ed.B. R. Bloom American Society of Microbiologists Washington DC 1994 p. 333. 120 S. Bosne-David L. Bricard F. Ramiandrasoa A. DéRoussent G. Kunesh and A. Andremont Antimicrob. Agents Chemother. 1997 41 1837. 121 R. Barclay and C. Ratledge Zentralbl. Bakteriol. Mikrobiol. Hyg. Ser. A 1986 262 203. 122 K. A. Brown and C. Ratledge Biochim. Biophys. Acta 1975 385 207. 123 C. Ratledge and K. A. Brown Am. Rev. Resp. Dis. 1972 106 774. 124 K. Mdluli R. A. Slayden Y. Q. Zhu S. Ramaswamy X. Pan D. Mead D. D. Crane J. M. Musser and C. E. Barry Science 1998 280 1607. Nat. Prod. Rep. 2000 17 99–116 116 125 M. J. Miller and F. Malouin Acc. Chem. Res. 1993 26 241. 126 M. J. Miller F. Malouin E. K. Dolence C. M. Garsparski M. Ghosh P.R. Guzzo B. T. Lotz J. A. McKee A. Minnick and M. Teng in Recent Advances in the Chemistry of Anti-infective Agents ed. P. H. Bently and R. Ponsford R. Soc. Chem. Spec. Pub. 1993 119 135. 127 M. J. Miller and F. Malouin in The Development of Iron Chelators for Clinical Use eds. R. J. Bergeron and G. M. Brittenham CRC Boca Raton FL 1994 p. 275. 128 G. Benz T. Schroder J. Kurz C. Wunsche W. Karl G. Steffens J. Pfitzner and D. Schmidt Angew. Chem. Suppl. 1982 1322. 129 G. Benz L. Born M. Briedan R. Grosser J. Kurz H. Paulsen V. Sinnwell and B. Weber Leibigs Ann. Chem. 1984 1408. 130 G. Benz Leibigs Ann. Chem. 1984 1399. 131 G. Benz and D. Schmidt Liebigs Ann. Chem. 1984 1434. 132 H. Paulsen M. Briedan and G. Benz Liebigs Ann. Chem. 1987 565.133 H. J. Rogers in Iron Transport in Microbes Plants and Animals eds. G. Winkelmann D. van der Helm and J. B. Neilands VCH Weinheim FRG 1987 223. 134 J. B. Neilands and J. R. Valenta in Metal Ions in Biological Systems ed. H. Sigel Dekker New York 1985 vol. 19 p. 313 135 L. Vertesy W. Aretz H.-W. Fehlhaber and H. Kogler Helv. Chim. Acta 1995 78 46. 136 C. Walling Acc. Chem. Res. 1975 8 125. 137 J. L. Zweier J. Biol. Chem. 1988 263 1353. 138 E. Graf J. R. Mahoney R. G. Bryant and J. W. Eaton J. Biol. Chem. 1984 259 3620. 139 R. A. Floyd and C. A. Lewis Biochemistry 1983 22 2645. 140 J. M. C. Gutteridge R. Richmond and B. Halliwell Biochem. J. 1979 184 469. 141 L. D. Horwitz N. A. Sherman Y. Kong A. W. Pike J. Gobin P. V. Fennessey and M. A. Horwitz Proc. Natl. Acad. Sci. USA 1998 95 5263. 142 L. D. Horwitz M. A. Horwitz B. W. Gibson and J. Reeve US Patent 5 721 209 (Chem. Abstr. 1999 130 47472y). 143 Y. Kong E. J. Lesneefsky J. Ye and L. D. Horwitz Am. J. Physiol. 1994 267 H2371. 144 L. D. Horwitz US Patent 5 786 326 (Chem. Abstr. 1998 129 144856p). 145 R. M. Hall and C. Ratledge J. Gen. Microbiol. 1986 132 853. 146 C. Ratledge and P. V. Patel J. Gen. Microbiol. 1976 93 141. 147 C. Ratledge and G. A. Snow Biochem. J. 1974 139 407. 148 P. V. Patel and C. Ratledge Biochem. Soc. Trans. 1973 1 886. 149 Y. Murakami S. Kato M. Nakajima M. Matsuoka H. Kawai K. Shin-Ya and H. Seto J. Antibiot. 1996 49 839. 150 M. Tsukamoto K. Murooka S. Nakajima S. Abe H. Suzuki K. Hirano H. Kondo K. Kojira and H. Suda J. Antibiot. 1997 50 815. 151 K. Suenaga S. Kokubo C. Shinohara T. Tsuji and D. Uemura Tetrahedron Lett. 1999 40 1945. 152 L. D. Horwitz and K. B. Horwitz US Patent 5 837 677 (Cont.-in-part Chem. Abstr. 1999 130 47472y). 153 I. S. Towbridge and M. B. Omary Proc. Natl. Acad. Sci. USA 1981 78 3039. 154 T. Kameyama A. Takahashi S. Kurasawa M. Ishizuka Y. Okami T. Takeuchi and H. Umezawa J. Antibiot. 1987 40 1664. 155 R. J. Bergeron and J. S. McManis Tetrahedron 1989 45 4939. Review a809397k