Albomycin uptake via a ferric hydroxamate transport system of Streptococcus pneumoniae R6

The antibiotic albomycin is highly effective against Streptococcus pneumoniae, with an MIC of 10 ng/ml. The reason for the high efficacy was studied by measuring the uptake of albomycin into S. pneumoniae. Albomycin was transported via the system that transports the ferric hydroxamates ferrichrome and ferrioxamine B. These two ferric hydroxamates antagonized the growth inhibition by albomycin and salmycin. Cross-inhibition of the structurally different ferric hydroxamates to both antibiotics can be explained by the similar iron coordination centers of the four compounds. [(55)Fe(3+)]ferrichrome and [(55)Fe(3+)]ferrioxamine B were taken up by the same transport system into S. pneumoniae. Mutants in the adjacent fhuD, fhuB, and fhuG genes were transport inactive and resistant to the antibiotics. Albomycin, ferrichrome, ferrioxamine B, and salmycin bound to the isolated FhuD protein and prevented degradation by proteinase K. The fhu locus consisting of the fhuD, fhuB, fhuG, and fhuC genes determines a predicted ABC transporter composed of the FhuD binding lipoprotein, the FhuB and FhuG transport proteins, and the FhuC ATPase. It is concluded that active transport of albomycin mediates the high antibiotic efficacy in S. pneumoniae.     

Sideromycins: tools and antibiotics

Sideromycins are antibiotics covalently linked to siderophores. They are actively transported into gram-positive and gram-negative bacteria. Energy-coupled transport across the outer membrane and the cytoplasmic membrane strongly increases their antibiotic efficiency; their minimal inhibitory concentration is at least 100-fold lower than that of antibiotics that enter cells by diffusion. This is particularly relevant for gram-negative bacteria because the outer membrane, which usually forms a permeability barrier, in this case actively contributes to the uptake of sideromycins. Sideromycin-resistant mutants can be used to identify siderophore transport systems since the mutations are usually in transport genes. Two sideromycins, albomycin and salmycin, are discussed here. Albomycin, a derivative of ferrichrome with a bound thioribosyl-pyrimidine moiety, inhibts seryl-t-RNA synthetase. Salmycin, a ferrioxamine derivative with a bound aminodisaccharide, presumably inhibts protein synthesis. Crystal structures of albomycin bound to the outer membrane transporter FhuA and the periplasmic binding protein FhuD have been determined. Albomycin and salmycin have been used to characterize the transport systems of Escherichia coli and Streptococcus pneumoniae and of Staphylococcus aureus, respectively. The in vivo efficacy of albomycin and salmycin has been examined in a mouse model using Yersinia enterocolitica, S. pneumoniae, and S. aureus infections. Albomycin is effective in clearing infections, whereas salmycin is too unstable to lead to a large reduction in bacterial numbers. The recovery rate of albomycin-resistant mutants is lower than that of the wild-type, which suggests a reduced fitness of the mutants. Albomycin could be a useful antibiotic provided sufficient quantities can be isolated from streptomycetes or synthesized chemically.     

Ferrichrome transport in Escherichia coli K-12: altered substrate specificity of mutated periplasmic FhuD and interaction of FhuD with the integral membrane protein FhuB.

FhuD is the periplasmic binding protein of the ferric hydroxamate transport system of Escherichia coli. FhuD was isolated and purified as a His-tag-labeled derivative on a Ni-chelate resin. The dissociation constants for ferric hydroxamates were estimated from the concentration-dependent decrease in the intrinsic fluorescence intensity of His-tag-FhuD and were found to be 0.4 microM for ferric aerobactin, 1.0 microM for ferrichrome, 0.3 microM for ferric coprogen, and 5.4 microM for the antibiotic albomycin. Ferrichrome A, ferrioxamine B, and ferrioxamine E, which are poorly taken up via the Fhu system, displayed dissociation constants of 79, 36, and 42 microM, respectively. These are the first estimated dissociation constants reported for a binding protein of a microbial iron transport system. Mutants impaired in the interaction of ferric hydroxamates with FhuD were isolated. One mutated FhuD, with a W-to-L mutation at position 68 [FhuD(W68L)], differed from wild-type FhuD in transport activity in that ferric coprogen supported promotion of growth of the mutant on iron-limited medium, while ferrichrome was nearly inactive. The dissociation constants of ferric hydroxamates were higher for FhuD(W68L) than for wild-type FhuD and lower for ferric coprogen (2.2 microM) than for ferrichrome (156 microM). Another mutated FhuD, FhuD(A150S, P175L), showed a weak response to ferrichrome and albomycin and exhibited dissociation constants two- to threefold higher than that of wild-type FhuD. Interaction of FhuD with the cytoplasmic membrane transport protein FhuB was studied by determining protection of FhuB degradation by trypsin and proteinase K and by cross-linking experiments. His-tag-FhuD and His-tag-FhuD loaded with aerobactin specifically prevented degradation of FhuB and were cross-linked to FhuB. FhuD loaded with substrate and also FhuD free of substrate were able to interact with FhuB.     

A lipoprotein of Yersinia enterocolitica facilitates ferrioxamine uptake in Escherichia coli.

A cloned fragment of Yersinia enterocolitica DNA complemented the defect in ferrioxamine B uptake of an Escherichia coli fhuE mutant lacking the outer membrane high-affinity transport protein FhuE. Subcloning revealed that a 13.7-kDa outer membrane protein was required for complementation. The amino acid sequence deduced from the nucleotide sequence showed extensive homology to PCPHi, an outer membrane lipoprotein of Haemophilus influenzae. We therefore termed this protein PCPYe. Plasmid-encoded pcpY mediated a low-affinity uptake of ferrioxamine B which may be caused by changes in the permeability of the outer membrane due to an overexpression of this outer membrane protein. A transposon insertion mutant in the plasmid-encoded pcpY gene was transferred into the chromosome of Y. enterocolitica. The resulting mutation had no effect on the high-affinity uptake of ferrioxamine B in Yersinia cells. Using the antibiotic ferrimycin we were able to isolate a Y. enterocolitica mutant lacking the high-affinity outer membrane receptor for ferrioxamine uptake, termed FoxA.     

The TonB protein of Yersinia enterocolitica and its interactions with TonB-box proteins

Summary The tonB gene is required for energy-dependent transport processes across the outer membrane of gram-negative bacteria. Using the antibiotics albomycin and ferrimycin, a tonB mutant of Yersinia enterocolitica was isolated. Comparison of the tonB mutant with the parent strain revealed that in Y. enterocolitica the uptake of ferrioxamine, ferrichrome, pesticin and heme is TonB-dependent. The tonB gene from Y. enterocolitica was sequenced and found to be similar to those of other Enterobacteria. The Y. enterocolitica tonB gene complemented a Y. enterocolitica tonB mutant. In contrast, some Tong functions of an Escherichia coli tonB mutant were not restored by the tonB gene of Y. enterocolitica. The observed differences in the ability to complement E. coli Tong functions correlated with the degree to which the Tong boxes of the receptors and colicins differed from the TonB box consensus sequence. Furthermore, the N-terminal membrane anchor of the TonB proteins and the TolA protein are likely to form an -helix with an identical sequence motif (SHLS) located at one face of the a-helix, suggesting this region to be involved in the functional cross-talk between the TonB-ExbBD-and TolABQR-dependent transport systems across the outer membrane.     

Iron transport in Streptomyces pilosus mediated by ferrichrome siderophores, rhodotorulic acid, and enantio-rhodotorulic acid

Streptomyces pilosus is one of several microbes which produce ferrioxamine siderophores. In the accompanying paper (G. Müller and K. Raymond, J. Bacteriol. 160:304-312), the mechanism of iron uptake mediated by the endogenous ferrioxamines B, D1, D2, and E was examined. Here we report iron transport behavior in S. pilosus as mediated by the exogenous siderophores ferrichrome, ferrichrysin, rhodotorulic acid (RA), and synthetic enantio-RA. In each case iron acquisition depended on metabolic energy and had uptake rates comparable to that of [55Fe]ferrioxamine B. However, the synthetic ferric enantio-RA (which has the same preferred chirality at the metal center as ferrichrome) was twice as effective in supplying iron as was the natural ferric RA complex, suggesting that stereospecific recognition at the metal center is involved in the transport process. Iron uptake mediated by ferrichrome and ferric enantio-RA was strongly inhibited by kinetically inert chromic complexes of desferrioxamine B. These inhibition experiments indicate that iron from these exogenous siderophores is transported by the same uptake system as ferrioxamine B. Since the ligands have no structural similarity to ferrioxamine B except for the presence of three hydoxamate groups, we conclude that only the hydroxamate iron center and its direct surroundings are important for recognition and uptake. This hypothesis is supported by the fact that ferrichrome A and ferrirubin, which are both substituted at the hydroxamate carbonyl groups, were not (or were poorly) effective in supplying iron to S. pilosus.     

Identification of the ferrioxamine B receptor,FoxB, inEscherichia coli K12

The photoreactivep-azidobenzoyl analog of ferrioxamine B was used to show that ferrioxamine-B-mediated iron transport is separate and distinct from coprogen-mediated iron transport inEscherichia coli. Photolysis of this analog inhibited uptake of [⁵⁹Fe]ferrioxamine B but not [⁵⁹Fe]coprogen or [⁵⁹Fe]ferrichrome. Conversely, photolysis of thep-azidobenzoyl analog of coprogen B inhibited uptake of [⁵⁹Fe]coprogen but not [⁵⁹Fe]ferrioxamine B or [⁵⁹Fe]ferrichrome. Photolabeling of outer membranes withp-azidobenzoyl-[⁵⁹Fe]ferrioxamine B resulted in the labeling of two iron-regulated peptides with molecular masses of about 66 and 26 kDa. Expression of these peptides was increased when ferrioxamine B was the sole iron source. Both peptides were present in outer membrane preparations of thefhuF mutant H1717, but the 66 kDa peptide was not inducible. These results are evidence for an outer membrane receptor inE. coli unique for linear ferrioxamines.     

Conversion of the coprogen transport protein FhuE and the ferrioxamine B transport protein FoxA into ferrichrome transport proteins

The FhuA protein of Escherichia coli K-12 transports ferrichrome and the structurally related antibiotic albomycin across the outer membrane and serves as a receptor for the phages T1, T5, and φ80 and for colicin M. In this paper, we show that chimeric proteins consisting of the central part of FhuA and the N- and C-terminal parts of FhuE (coprogen receptor) or the N- and/or C-terminal parts of FoxA (ferrioxamine B receptor), function as ferrichrome transport proteins. Although the hybrid proteins contained the previously identified gating loop of FhuA, which is the principal binding site of the phages T5, T1, and φ80, only the hybrid protein consisting of the N-terminal third of FoxA and the C-terminal two thirds of FhuA conferred weak phage sensitivity to cells. Apparently, the gating loop is essential, but not sufficient for wild-type levels of ferrichrome transport and for phage sensitivity. The properties of FhuA-FoxA hybrids suggest different regions of the two receptors for ferric siderophore uptake.     

Evidence for common binding sites for ferrichrome compounds and bacteriophage phi 80 in the cell envelope of Escherichia coli.

Mutants ton A and ton B of Escherichia coli K12, known to be resistant to bacteriophage phi80, were found to be insensitive as well to albomycin, an analogue of the specific siderochrome ferrichrome. Ferrichrome at micromolar concentrations strongly inhibited plaque production by phi80. Preincubation with ferrichrome did not inactivate the phage. At a concentration at which ferrichrome allowed 90% inhibition of plaque formation, the chromium analogue of ferrichrome showed no detectable activity. Similarly, ethylenediaminetetraacetic acid, ferrichrome A, and certain siderochromes structurally distinct from ferrichrome, such as ferrioxamine B, schizokinen, citrate, and enterobactin, did not show detectable inhibitory activity. However, rhodotorulic acid showed moderate activity. A host range mutant of phi80, phi80h, was also inhibited by ferrichrome, as was a hybrid of phage lambda possessing the host range of phi80. However, phage lambdacI- and a hybrid of phi80 possessing the host range of lambda were not affected by ferrichrome. Finally, ferrichrome and chromic deferriferrichrome were shown to inhibit adsorption of phi80 to sensitive cells, ferrichrome giving 50% inhibition of adsorption at a minimal concentration of 8 nM. It is suggested that a component of the ferrichrome uptake system may reside in the outer membrane of E. coli K12 and may also function as a component of the receptor site for bacteriophage phi80, and that ferrichrome inhibition of the phage represents a competition for this common site.     

Thiamine uptake in Escherichia coli: III. Regulation of thiamine uptake in Escherichia coli

The metabolic regulation of thiamine uptake in Escherichia coli has been investigated. A thiamine regulatory mutant (PT-R1), which is three times higher in cellular thiamine concentration than the parent E. coli K12 and contains a normal level of the membrane thiamine kinase (ATP: thiamine pyrophosphotransferase, EC 2.7.6.2), showed the rate of thiamine uptake half that of the parent strain. This reduction in the rate of thiamine uptake in PT-R1 is not attributable to alterations in the activity and specificity of the thiamine transport system, to an increase in the exit rate of thiamine nor to feedback inhibition. The results obtained with PT-R1 suggest that formation of the transport system is repressed by the enhanced cellular thiamine in this mutant.     

Characterization of siderophore-mediated iron transport inGeotrichum candidum,a non-siderophore producer

Summary Geotrichum candidum is capable of utilizing iron from hydroxamate siderophores of different structural classes. The relative rates of iron transport for ferrichrome, ferrichrysin, ferrioxamine B, fusigen, ferrichrome A, rhodotorulic acid, coprogen B, dimerium acid and ferrirhodin were 100%, 98%, 74%, 59%, 49%, 35%, 24%, 12% and 11% respectively. Ferrichrome, ferrichrysine and ferrichrome A inhibited [⁵⁹Fe]ferrioxamine-B-mediated iron transport by 71%, 68% and 28% respectively when added at equimolar concentrations to the radioactive complex. The inhibitory mechanism of [⁵⁹Fe]ferrioxamine B uptake by ferrichrome was non-competitive (K _i 2.4 M), suggesting that the two siderophores do not share a common transport system. Uptake of [⁵⁹Fe]ferrichrome, [⁵⁹Fe]rhodotorulic acid and [⁵⁹Fe]fusigen was unaffected by competition with the other two siderophores or with ferrioxamine B. Thus,G. candidum may possess independent transport systems for siderophores of different structural classes. The uptake rates of [¹⁴C]ferrioxamine B and⁶⁷Ga-desferrioxamine B were 30% and 60% respectively, as compared to [⁵⁹Fe]ferrioxamine B. The specific ferrous chelates, dipyridyl and ferrozine at 6 mM, caused 65% and 35% inhibition of [⁵⁹Fe]ferrioxamine uptake. From these results we conclude that, although about 70% of the iron is apparently removed from the complex by reduction prior to being transported across the cellular membrane, a significant portion of the chelated ligand may enter the cell intact. The former and latter mechanisms seem not to be mutually exclusive.     

Uptake and conversion of the antibiotic albomycin by Escherichia coli K-12.

The antibiotic albomycin is transported into cells of Escherichia coli K-12 by the same uptake system as the iron-supplying ferrichrome complex. The iron-complexing hydroxamate moieties of albomycin and ferrichrome are structurally similar. During the phase of rapid iron uptake the chelators were not found in the cells. In order to understand the antibiotic activity of albomycin, it was labeled in the hydroxamate with tritium and in the presumed antibiotically active area with radioactive sulfur. While the tritium label was not retained by the cells, part of the sulfur label was taken up and concentrated 500-fold within the cell. The sulfur was not incorporated into proteins or nucleic acids since it was recovered as a low molecular weight component. Gel filtration on Bio-Gel P-2 revealed one tritium-labeled and two sulfur-labeled cleavage products in the incubation medium. We conclude that albomycin is actively transported via its ferrichrome-like portion into the cells and that the growth-inhibitory moiety is released by hydrolysis intracellularly and remains there.     

Transport activity of FhuA, FhuC, FhuD, and FhuB derivatives in a system free of polar effects, and stoichiometry of components involved in ferrichrome uptake

The Escherichia colifhu operon, composed of the fhuA, C, D, and B genes, is essential for the utilization of ferric siderophores of the hydroxamate type and for the uptake of the antibiotic albomycin. We have had difficulty studying the effects of missense mutations in individual plasmid-encoded transport genes because appropriate test strains were not found: all isolated chromosomal mutations in either one of the fhu genes (with a complete loss of function) negatively influenced the expression of other fhu genes in the operon. In order to analyze Fhu mutant proteins in a system free of polar effects, we constructed a plasmid-encoded gene cassette system by introducing unique restriction sites that allowed precise cloning of individual fhu genes. The fhu cassette operon expressed in a chromosomal fhu deletion mutant enabled us to evaluate the transport activity of mutated FhuA, FhuC, FhuD or FhuB derivatives. In addition, we found that transport across the outer membrane (via FhuA, TonB, ExbB, D) rather than transport across the cytoplasmic membrane (via FhuC, D, B) was rate limiting. The stoichiometry of the components involved in the uptake of iron(III) hydroxamates seems to be important for proper functioning. Received: 20 October 1997 / Accepted: 22 December 1997     

Ferrioxamine-mediated Iron(III) utilization by Salmonella enterica

Utilization of ferrioxamines as sole sources of iron distinguishes Salmonella enterica serotypes Typhimurium and Enteritidis from a number of related species, including Escherichia coli. Ferrioxamine supplements have therefore been used in preenrichment and selection media to increase the bacterial growth rate while selectivity is maintained. We characterized the determinants involved in utilization of ferrioxamines B, E, and G by S. enterica serotype Typhimurium by performing siderophore cross-feeding bioassays. Transport of all three ferric siderophores across the outer membrane was dependent on the FoxA receptor encoded by the Fur-repressible foxA gene. However, only the transport of ferrioxamine G was dependent on the energy-transducing protein TonB, since growth stimulation of a tonB strain by ferrioxamines B and E was observed, albeit at lower efficiencies than in the parental strain. Transport across the inner membrane was dependent on the periplasmic binding protein-dependent ABC transporter complex comprising FhuBCD, as has been reported for other hydroxamate siderophores of enteric bacteria. The distribution of the foxA gene in the genus Salmonella, as indicated by DNA hybridization studies and correlated with the ability to utilize ferrioxamine E, was restricted to subspecies I, II, and IIIb, and this gene was absent from subspecies IIIa, IV, VI, and VII (formerly subspecies IV) and Salmonella bongori (formerly subspecies V). S. enterica serotype Typhimurium mutants with either a transposon insertion or a defined nonpolar frameshift (+2) mutation in the foxA gene were not able to utilize any of the three ferrioxamines tested. A strain carrying the nonpolar foxA mutation exhibited a significantly reduced ability to colonize rabbit ileal loops compared to the foxA+ parent. In addition, a foxA mutant was markedly attenuated in mice inoculated by either the intragastric or intravenous route. Mice inoculated with the foxA mutant were protected against subsequent challenge by the foxA+ parent strain.     

The pyocin Sa receptor of Pseudomonas aeruginosa is associated with ferripyoverdin uptake.

We have used Tn5 mutagenesis to obtain a mutant resistant to pyocin Sa. When grown in iron-deficient succinate medium this mutant lacked an 85-kDa iron-regulated outer membrane protein (IROMP), and expression of a 75-kDa IROMP was increased compared with that in the parent strain. The mutant was deficient in pyoverdin biosynthesis and showed a 95% decrease in transport of ferripyoverdin purified from the parent strain, suggesting that the 85-kDa IROMP is the specific receptor for ferripyoverdin and pyocin Sa. The mutant compensated for the deficiency in pyoverdin biosynthesis and transport by exhibiting a fourfold increase in ferripyochelin transport. The low-level transport of ferripyoverdin in the Sa-resistant mutant, which extended to heterologous pyoverdins from other strains, suggests that Pseudomonas aeruginosa has a second ferripyoverdin uptake system of lower affinity and broader specificity.     

Functional characterization and metal ion specificity of the metal-citrate complex transporter from Streptomyces coelicolor

Secondary transporters of citrate in complex with metal ions belong to the bacterial CitMHS family, about which little is known. The transport of metal-citrate complexes in Streptomyces coelicolor has been investigated. The best cofactor for citrate uptake in Streptomyces coelicolor is Fe(3+), but uptake was also noted for Ca(2+), Pb(2+), Ba(2+), and Mn(2+). Uptake was not observed with the Mg(2+), Ni(2+), or Co(2+) cofactor. The transportation of iron- and calcium-citrate makes these systems unique among the CitMHS family members reported to date. No complementary uptake akin to that observed for the CitH (Ca(2+), Ba(2+), Sr(2+)) and CitM (Mg(2+), Ni(2+), Mn(2+), Co(2+), Zn(2+)) systems of Bacillus subtilis was noted. Competitive experiments using EGTA confirmed that metal-citrate complex formation promoted citrate uptake. Uptake of free citrate was not observed. The open reading frame postulated as being responsible for the metal-citrate transport observed in Streptomyces coelicolor was cloned and overexpressed in Escherichia coli strains with the primary Fe(3+)-citrate transport system (fecABCDE) removed. Functional expression was successful, with uptake of Ca(2+)-citrate, Fe(3+)-citrate, and Pb(2+)-citrate observed. No free-citrate transport was observed in IPTG (isopropyl-beta-d-thiogalactopyranoside)-induced or -uninduced E. coli. Metabolism of the Fe(3+)-citrate and Ca(2+)-citrate complexes, but not the Pb(2+)-citrate complex, was observed. Rationalization is based on the difference in metal-complex coordination upon binding of the metal by citrate.     

Properties of the FhuA channel in the Escherichia coli outer membrane after deletion of FhuA portions within and outside the predicted gating loop.

Escherichia coli transports Fe3+ as a ferrichrome complex through the outer membrane in an energy-dependent process mediated by the FhuA protein. A FhuA deletion derivative lacking residues 322 to 355 (FhuA delta322-355) forms a permanently open channel through which ferrichrome diffused. This finding led to the concept that the FhuA protein forms a closed channel that is opened by input of energy derived from the electrochemical potential across the cytoplasmic membrane, mediated by the Ton system. In this study, we constructed various FhuA derivatives containing deletions inside and outside the gating loop. FhuA delta322-336 bound ferrichrome and displayed a residual Ton-dependent ferrichrome transport activity. FhuA delta335-355 no longer bound ferrichrome but supported ferrichrome diffusion through the outer membrane in the absence of the Ton system. FhuA delta335-355 rendered cells sensitive to sodium dodecyl sulfate and supported diffusion of maltotetraose and maltopentaose in a lamB mutant lacking the maltodextrin-specific channel in the outer membrane. Cells expressing FhuA delta70-223, which has a large deletion outside the gating loop, were highly sensitive to sodium dodecyl sulfate and grew on maltodextrins but showed only weak ferrichrome uptake, suggesting formation of a nonspecific pore through the outer membrane. FhuA delta457-479 supported Ton-dependent uptake of ferrichrome. None of these FhuA deletion derivatives formed pores in black lipid membranes with a stable single-channel conductance. Rather, the conductance displayed a high degree of current noise, indicating a substantial influence of the deletions on the conformation of the FhuA protein. FhuA also supports infection by the phages T1, T5, and phi80 and renders cells sensitive to albomycin and colicin M. Cells expressing FhuA delta322-336 were sensitive to albomycin and colicin M but were only weakly sensitive to T5 and phi480 and insensitive to T1. Cells expressing FhuA delta335-355 were resistant to all FhuA ligands. These results indicate different structural requirements within the gating loop for the various FhuA ligands. Cells expressing FhuA delta457-479 displayed a strongly reduced sensitivity to all FhuA ligands, while cells expressing FhuA delta70-223 were rather sensitive to all FhuA ligands except albomycin, to which they were nearly resistant. It is concluded that residues 335 to 355 mainly determine the properties of the gate with regard to FhuA permeability and ligand binding.