The pyoverdin receptor FpvA, a TonB-dependent receptor involved in iron uptake byPseudomonas aeruginosa(Review)

Molecular Membrane Biology ISSN: 0968-7688 (Print) 1464-5203 (Online) Journal homepage: https://www.tandfonline.com/loi/imbc20 The pyoverdin receptor FpvA, a TonB-dependent receptor involved in iron uptake by Pseudomonas aeruginosa (Review) Nicolas Folschweiller, Isabelle J. Schalk, Hervé Celia, Mohamed A. Abdallah, Franc Pattus, Bruno Kieffer To cite this article: Nicolas Folschweiller, Isabelle J. Schalk, Hervé Celia, Mohamed A. Abdallah, Franc Pattus, Bruno Kieffer (2000) The pyoverdin receptor FpvA, a TonB-dependent receptor involved in iron uptake by Pseudomonas�aeruginosa (Review), Molecular Membrane Biology, 17:3, 123-133, DOI: 10.1080/09687680050197356 To link to this article: https://doi.org/10.1080/09687680050197356 Published online: 09 Jul 2009. Submit your article to this journal Article views: 220 View related articles Citing articles: 24 View citing articles Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=imbc20 Molecular Membrane Biology, 2000, 17, 123±133 The pyoverdin receptor FpvA, a TonB-dependent receptor involved in iron uptake by Ps eudomonas aeruginosa (Review) Nicolas Folschweiller², Isabelle J. Schalk², Herve Celia², Bruno Kieffer³ , Mohamed A. Abdallah² and Franc Pattus²* ²DeÂpartement des ReÂcepteurs et ProteÂines Membranaires, UPR 9050 CNRS; and ³ UPR 9004, E SBS, Boulevard SeÂbastien Brant, F-67400 Illkirch, Strasbourg, France. Summary Iron is an important element, essential for the growth of almost all living cells. Because of the high insolubility of iron(III) in aerobic conditions, many gram-negative bacteria produce, under iron limitation, small iron-chelating compounds called siderophores, together with new outer-membrane proteins, which function as receptors for the ferrisiderophores. Pseudomonas aeruginosa, an important human opportunistic pathogen, produces at least three known siderophores when grown in irondeficient conditions: pyochelin, salicylate and pyoverdin. This review focuses on pyoverdin and on the ability of FpvA to bind iron-free and ferric-PaA pyoverdin, in the light of recent information gained from biochemical and biophysical studies and of the recently solved 3D-structures of the related ferrichrome FhuA and enterobactin FepA receptors in Escherichia coli. Keywords: Siderophore, outer membrane receptor, FpvA, sequence alignment, fluorescence resonance energy transfer. Abbreviations: PaA, pyoverdin; OHAsp, b-hydroxyaspartic acid; OHOrn, Nd-hydroxyornithine; Dab, 2,4-diaminobutyric acid; CCCP, carbonyl cyanide 3-chlorophenyl hydrazone; FRET, fluorescence resonance energy transfer; Kdapp , apparent dissociation constant; Kia pp , apparent inhibition constant; a.u., arbitrary units. Introduction Iron is an essential element for many microorganisms . It catalyz es electron transfer in reactions as critical and different as oxygen transport or storage, and metabolism of hydrogen, oxygen or nitrogen (Braun 1997). Acquisition of iron is complicated by the low solubility of ferric iron in nature (Neilands et al. 1987) and, for pathogenic microbes , by the iron-limiting nature of the host (Sawatzki 1987). In vertebrate hosts , iron is chelated by transferrin and lactoferrin and internaliz ed by means of transferrin-receptors (Guerinot 1994), whereas a number of adaptiv e responses are employed by bacteria to obtain iron under limiting conditions, one of the most common involving the production and releas e of siderophores (Neilands 1981). Siderophores are low-molecular-mass , high-affinity iron chelators capable of delivering iron to bacterial cells via specific receptor proteins *To whom correspondence should be addressed. e-mail: pattus@ esbs.u-strasbg.fr on the cell surface (Neilands 1982). Since in vivo growth of bacteria depends upon their ability to acquire iron from the host, siderophore-mediated iron uptake is considered as an important component of pathogenesis , and its regulation is coupled to the expression of virulence determinants (Cox 1982, Crosa 1984, Griffiths 1987, Litwin and Calderwood 1993). Ps eudomona s a eruginos a is an opportunistic human pathogen, which caus es severe and often fatal infections , particularly affecting immuno-compromis ed patients (Fichtenbaum et al. 1994, Ali et al. 1995). These bacteria produce three known siderophores during growth under iron-limiting conditions: pyoverdin, which is the major siderophore, salicylate and pyochelin (Cox and Graham 1979, Cox and Adams 1985, Ankenbauer et al. 1988, Visca et al. 1993). Excess iron repress es the synthesis of the three siderophores (Cox and Graham 1979, Cox and Adams 1985, Visca et al. 1993). Repression is mediated by the protein Fur (Prince et al. 1993, Barton et al. 1996), which, in the pres ence of iron, binds to the promoters of the pvdS and pchR genes (Ochsner et al. 1995, Leoni et al. 1996). PvdS, a putativ e sigma factor, and PchR, a member of the AraC family of transcriptiona l activators , positively regulate the synthesis of, respectively, pyoverdin, salicylate and pyochelin (Heinrich and Poole 1993, 1996, Cunliffe et al. 1995, Miyazaki et al. 1995). In general, in gram-negativ e bacteria, the first step propos ed for entry of ferric siderophores into bacteria is mediated by specific outer membrane receptors . The transport into the periplas m requires the proton motive force (pmf) of the cytoplasmic membrane and an energy transduction complex compos ed of three proteins TonB, ExbB and ExbD (Kadner 1990, Postle 1993, Lars en et al. 1996, Stojiljkovic and Srinivasan 1997). Transport across the cytoplasmic membrane is carried out by an ABC transport system (reviewed in Braun 1997). The TonB-E xbB-E xbD complex has been identified in many gram-negativ e bacteria and probably represents a cons erved mechanis m for energy transduction to high-affinity transporters within the outer membrane (Lars en et al. 1996, Stojiljkovic and Srinivasan 1997). TonB is anchored in the cytoplasmic membrane by an uncleaved leader sequenc e (Postle and Skare 1988), and spans the periplasmic space to interact directly with outer membrane components (Skare et al. 1993). Many studies emphasiz e the importanc e of a Nterminal region of the outer membrane receptors, called the TonB-box, that mediates physical interactions with TonB (Kadner 1990, Lars en et al. 1997). In vivo cross-linking experiments (Skare et al. 1993, Cadieux and Kadner 1999) have demonstrated the direct interaction between TonB and outer membrane receptors . However, it is still unknown how TonB is coupled to the electrochemica l gradient of the cytoplasmic membrane, and how the energy is transferred to the outer membrane receptor. However, Larsen et al. (1999) Molecular Membrane Biology ISSN 0968-7688 print/ISSN 1464-5203 online Ó 2000 Taylor & Francis Ltd http://www.tandf.co.uk/journals N. Folschweiller et al. 124 have shown that the protonmotive force, ExbB and ligandbound receptor drive conformationa l changes in TonB, suggesting a dynamic model of energy transduction in which TonB cycles through a s et of conformations that differ in potential energy. Recent evidenc e has expanded the role of siderophore receptors to communicator s of extracellula r information via interactions in the periplasm. The binding of ferric citrate and ps eudobactin BN7/8 to their cognate receptors in the outer membrane, respectively FecA in E. coli and PupB in Ps eudomona s putida, induc e the expression of the genes responsible for siderophor e uptake (reviewed in Braun 1997), a process that can be separated from the transport of siderophore across the outer membrane. For these two receptors, a N-termina l domain called the FecR box, in front of the TonB box, plays a role in signa l transduction by interacting with the signa l transducer, a protein with simila r topology as TonB (FecR for E. coli and its analogue in P. putida). Previously (Schalk et al. 1999), it was pointed out that FpvA poss ess es, like FecA and PupB, a periplasmic Nterminal end with a signal transduction function, and that a homologu e of FecR is found in the P. a eruginos a genome. In addition, there is some evidence that pyoverdin positively regulates the expression of the fpva gene (Gensberg et al. 1992). Structures of pyoverdins Primary structures of pyoverdins Pyoverdins are chromopeptidic siderophores excreted by fluorescent Ps eudomona s. They possess a highly fluorescent chromophore derived from 2,3-diamino-6,7-dihydroxyquinoline bound to a peptidic moiety identical for all the pyoverdins occurring from a given strain, but generally with strong variation from strain to strain (s ee figures 1 and 3 for a few examples ). The chromophore, with its catecholate bidentate ligand, repres ents the common feature between Figure 1. Pseudomona s pyoverdins , probably indicating a biosynthetic pathway conserved through evolution, and has the same configura tion (S) for all the pyoverdins so far investigated. For a given strain, the pyoverdins differ only in the acyl substituent bound to the amino group on C-3 of the chromophore, which can in many cases derive from succinic acid, malic acid, a-ketoglutaric acid. The variability of the acid or amide substituent on C-3 of the chromophor e suggests that it has little importance beyond protection from hydrolysis of an aromatic free amine group to a hydroxyl group, and plays no role in binding of the complexed pyoverdin to its receptor. The peptide chain that provides the two additiona l bidentate ligands to complex Fe(III) is unique to each species of Ps eudomona s, indicating a specialization for each bacterial strain in its pyoverdin which confers , on complexation, specificity for its own receptor. Cross-feeding experiments for seven fluorescent Ps eudomona s strains have shown that the pyoverdin-mediated iron transport systems are strictly strain-specific for most of them (Cornelis et al. 1989, Hohnadel and Meyer 1988). However, in very few cas es this specificity seems to be less strict for pyoverdins from Ps eudomona s a eruginos a ATCC 15692, Ps eudomona s fluorescens ATCC 13525 and Ps eudomona s chlororaphis ATCC 9446. These strains produce pyoverdins which are structurally identical or very clos ely related (Briskot et al. 1989, Demange et al. 1990, Linget et al. 1992). Cons equently, thes e siderophores are interchangeable . To date, this is a unique case. Three dimensional structure of pyoverdins Although the primary structures of ~ 30 pyoverdins have been determined so far, and their physico-chemica l properties described (Abdalla h 1991, Pattus and Abdallah 2000), there are only three for which the three-dimensional structur e is known. The first and only example of an X-ray structur e elucidation of a pyoverdin is Ps eudobactin B10 from The primary structure of pyoverdin PaA, siderophore of Pseudomonas aeruginosa ATCC 15692. Pyoverdin outer membrane receptor FpvA Ps eudomona s B10 (Teintz e et al. 1981). The two others are NMR solution structures of the gallium(III) complexes of pyoverdin GM(II) and pyoverdin G4R, respectively, from Ps eudomona s fluorescens II and Ps eudomona s putida G4R (Mohn et al. 1994, Atkinson et al. 1998). Attempts to elucidate the solution structure of the aluminum complex of pyoverdin PaA from Ps eudomona a eruginos a ATCC 15692 showed the pres ence of two conformers in slow exchange which could not be reliably assigned (Mertz et al. 1991). Thes e three pyoverdin±metal(III) complexes (i. e. B10, GM II and G4R) actually poss ess the same chelating bidentate groups, but vary in the length of their peptidic moieties, which are, respectively, Lys-OHAsp-Ala-allo-Thr-Ala-cycloOHOrn for ps eudobactin B10, Ala-Lys-Gly-Gly-OHAs p-Gln-Ser-AlaAla-Ala-Ala-cycloOHOrn for pyoverdin GM(II) and Asp-OrnOHAsp/Dab-Gly-Ser-cycloOHOrn for pyoverdin G4R (In pyoverdin G4R, the two amino groups of 2,4-diaminobutyric acid (Dab) are condensed to the a-carboxyl of OHAsp, forming a carboxytetrahydropyrim idine ring.) Therefore, it is interesting to compare thes e three structures to understand how the peptidic moiety affects the structure of the same chelating groups (figure 2). The three bidentate groups which are complexing the metal(III) form a propeller-shaped octahedra l complex. The chirality of this propeller arrangement is either left-handed (K-configuration) or right-handed (D-configuration). The total number of possible stereoisomers is, thus, 16: eight K and eight D, numbered from I ± VIII in each series (for the axioms of nomenclature of pyoverdins , see Teintze et al. 1981). Due to the chirality of the ligand and to its large siz e, not all of thes e possible isomers are found in solution or in crystal structures, and generally only one main isomer is present for each structurally different pyoverdin. The comparison of the three structures shows that their coordination around the metal is not identical: while ps eudobactin B10-F e(III) and pyoverdin GMII-Ga(III) have K configurations around the metal ion (KI and KII, respectively), pyoverdin G4R-Ga (III) has a D configura tion. It has to be pointed out that, while the X-ray structure provides one with an unambiguous binding configura tion, the determination of the precis e configuration from NMR data is more difficult due to the non-chira l nature of NMR-derived distanc e constraints . Moreover, the gallium us ed in the solution studies is a diamagnetic metal and cannot be directly obs erved. The metal binding chirality differenc e between pyoverdin G4R and the two others is probably due to the siz e of its peptidic part, which is one of the shortest ever obs erved among pyoverdin peptides. The structure of pyoverdin G4R provides one with the minimal structura l scaffold that is necessary to ensure the metal binding function. Among the residues that are present between those complexing the metal ion (OHAsp-3 and OHOrn-7), Gly-5 and Dab-4 (condens ed to OHAsp-3) have an important structura l role in reversing the direction of the peptide chain. The only common features of the overall structure of the three pyoverdins shown in figure 3 is that the position of the metal ion is always located at an edge of the structure, defining an open access side for each pyoverdin. However, the orientation of this open access side with respect to the chromophor e is different for the three known structures as 125 well as the peptide fold. This lack of structural similarity explains the specificty of recognition and transport by the receptors and strongly suggests that the determinants of this specificity are located on the peptidic moiety and differ from one pyoverdin to the other. The availability of three-dimensiona l structures of pyoverdins may help in identifying the residues required for the receptor recognition on an individual basis . For example, the short length of the peptidic moiety in pyoverdin G4R leaves only three residues where functional groups may be recogniz ed by the receptor, namely Asp-1, Orn-2 and Ser-6. The three residues between the two peptidic coordinating residues of ps eudobactin B10 (namely OHAsp-2 and cycloOHOr n-6) form a more flexible sequenc e (L)-Ala-(D)-allo-Thr-(L)-Ala, which could modulate the specificity. Pyoverdin GM-II, on the other hand, has longer, more flexible s egments, (D)-Ala-(D)-Lys-Gly-Gly between the chromophore and the coordinating OHAsp-5, and (D)-Gln(D)-Ser-(L)-Ala-(D)-Ala-(D)-Ala-(L)-Ala, between the two coordinating OHAsp-5 and cyclo-OHOrn-12 residues. Moreover, the specificity of the determinants for iron uptake inside the bacteria may be even more complex, since the pyoverdins not only bind to the receptor at the cell surface but also need to carry iron inside the cell (transport). The specificity for binding to the receptor may be different from the specificity for transport. For example, in the cas e of Neurospor a crassa, the cyclic nature of the peptide is not required for the binding of the iron-coordinated region of the siderophore complex to the receptor, but is necessary for the subs equent transport (Huschka et al. 1985, Winkelmann and Huschka, 1987). 3D-structure of the pyoverdin receptor FpvA as deduced from s equence analysis and from the 3D-structure of FhuA and FepA from Escherichia coli The 3D structur e of FpvA, the pyoverdin receptor, is not yet known, but the 3D structures of two E. coli receptors have been reported recently: FhuA and FepA, the ferrichrom e and the enterobactin receptors, respectively (Ferguson et al. 1998, Locher et al. 1998, Buchanan et al. 1999, figure 4). Although thes e two proteins do not share extensiv e s equenc e similarity, they show a simila r fold. The FpvA receptor has ~ 42% similarity with the ferric-ps eudobactin receptor PupA of P. putida and the coprogen/rhodotorulic acid receptor FhuE of E. coli, and less than 20% similarity with other members of the siderophor e outer membrane receptors family (so far, ~ 50 characteriz ed members) including FhuA and FepA. Despite this low level of similarity, multiple sequenc e alignment of the receptors s equences within sub-families and then with the whole superfamily, strongly suggests that all the siderophor e outer membrane receptors have a common fold consisting of three domains: the b-barr el ( ~ 550 residues ), the `plug’ (129 residues for FhuA) and the N-terminal TonB box (five to s even residues ) arranged in the primary s equence, as shown in figure 5. Multiple sequence alignment reveals that the FpvA receptor belongs to a subgroup of seven receptors containing an additiona l N-termina l domain, the function of which is well documented for FecA, the E. coli 126 N. Folschweiller et al. Figure 3. Comparison of the three known three-dimensional pyoverdin structures. Ball and stick and planar representations of ferric pseudobactin (a); pyoverdin GMII-Ga(III) complex (b); and pyoverdin G4R-Ga(III) complex (c). Only the heavy atoms of the residues in the vicinity of the chelation centre are detailed: the chromophore, b-hydroxyaspartic acid and cyclo-N8-hydroxyornithine. The main peptide chain is represented by a ribbon going through the Ca positions. The succinate residue on carbon C-3 of the chromophore is not represented in the ball and stick model. ferric citrate receptor, and which is called `FecR’ box thereafter (s ee below, figure 5). The b-barrel domain The b-barrel domain forms the transmembrane part of the receptor in contact with the lipids . It is elliptical in shape and consists of 22 antiparallel strands , with large loops extending towards the extracellula r side, and short periplasmic turns, which is a typical feature of bacterial outer membrane proteins such as porins , OmpA and E. coli phospholipas e A2. The 22 b-strands are the best cons erved regions in the barrel, and the core of the FpvA barrel can be easily constructed by homology modelling from the FhuA and FepA Pyoverdin outer membrane receptor FpvA 3D structures. The b-strand siz es can vary from eight to 30 residues , and their s equences are characterized by an alternance of hydrophobic and hydrophilic residues , interacting, respectively, with the lipids and water molecules or plug residues located within the barrel. The b-barrel is also characteriz ed by a common outer membrane aromatic anchoring ring compos ed of Trp and Figure 2. Comparison of the coordination at the metal binding site of the three known three-dimensional pyoverdin-metal(III) complexes structures. Ball and stick representations of ferric pseudobactin (a); pyoverdin GMII-Ga(III) complex (a); and pyoverdin G4R-Ga(III) complex (c). Only the heavy atoms of the residues in the vicinity of the chelation centre are detailed: the chromophore, b-hydroxyaspartic acid and cyclo-N8-hydroxyornithine. The stars show the Ca atoms binding the peptide chains. 127 Phe, whereas Tyr are mostly located in the extra-cellula r portion of the b-strands and the extracellula r loops. The s equence alignment analysis suggests the existenc e of an interestingly conserved pattern in strand 21 and the following extra cellula r loop. This strand shows the only strictly conserved residue (Asn 738 in FpvA, figure 6) of the barrel among all siderophor e outer membrane receptors . This residue, together with the two next well cons erved hydrophobic residues (Ala/Val/Leu and Phe, figure 6), is located at the beginning of loop 11, forms a sharp bend sticking out from the b-barrel, positioning the Phe side chain well outside of the barrel. Interestingly, in the 3D structures of FhuA and FepA (Ferguson et al. 1998, Locher et al. 1998, Buchanan et al. Figure 4. Ribbon representation of the FhuA three-dimensional structure loaded with ferrichrom e (Ferguson et al. 1998, Locher et al. 1998). Residues 484 ± 612 have been removed to allow a clear view of the plug domain. The molecule is viewed as in the outer membrane plane, with the periplasmic loops at the bottom of the figure and the extracellula r one at the top. The plug domain is shown in green, while the beta barrel and loops are in blue. The ferrichrom eiron is represented as CPK model and coloured in red. The figure was prepared with the program MOLSCRIPT (Kraulis 1991), and using the Protein Data Bank coordinate file 1BY5 (Locher et al. 1998). Figure 5. Organization of the domains in the siderophore outer membrane receptor sequence alignment. FpvA predicted domains: the FecR box lies from residues 1 ± 62, the TonB box from 87 ± 91, the plug from 10 ± 228, and the barrel from 229 ± 770. In the FhuA protein, the FecR box is missing, the TonB box lies from residues 7 ± 14, the plug from 28 ± 156, and the barrel from 157 ± 714. Unassigned residues between FecRbox, TonB box and plug domains are represented with two black lines. 128 N. Folschweiller et al. Figure 6. C-terminal sequences alignments of enterobactin (BfeA, PirA, PfeA, FepA), ferrichrom e (FhuA), citrate (FecA) and pyoverdin (FpvA) outer membrane receptors. Identities are shown in outline bold and dark shade. while similarities are in outline and light shade (Using MacVector). The horizontal arrow shows the conserved Asn and the conserved part of the external loop between strands 21 and 22. The dotted arrow shows the full external loop between strands 21 and 22 (based on the FhuA and FepA 3D structures ). The black dot indicates the Cterminal conserved Phe. The numbering uses mature protein sequences except for BfeA and PirA, where the data is not available (reminded with an asterisk). This figure shows only a small number of sequences for illustration (ClustalW alignment, Higgins et al. 1996). 1999), this same loop 11 contains two antiparallel b-sheets. Apart from the first hydrophobic residues, the sequenc e of the whole loop is not cons erved, except for the FhuA family receptors. The turn between the two strands is located at the edge of the loop, close to the siderophor e binding site on FhuA, but there is not yet evidenc e of any functional role of this region, neither in ligand binding nor folding and/or outer membrane insertion of the barrel. The extracellula r loops show very poor sequenc e similarity, and their structures in the 3D models of FhuA and FepA are quite different. No information could be gained on the structur e of the FpvA extracellula r loops from the s equenc e and structura l analysis . As for other outer membrane proteins like porins , the last C-termina l residue is a Phe (or exceptionally a Trp, figure 6). It has been shown for PhoE (Struyve et al. 1991) that this last residue is essential for proper membrane ins ertion of the barrel. The plug domain The plug domain is a globula r domain located at the Nterminal part of the receptors . According to the 3D structure of FhuA and FepA, the plug domain folds into the barrel and completely clogs it. Topologica lly, the plug crosses the membrane, but without having any contact with lipids . According to the FhuA structure, three residues located on the external side participate to the ferrichrom e binding site, whereas its N-terminal end and the TonB box, face the periplasmic space. The similarity between the plug sequences is significant, in agreement with the common idea that this unique folding should be conserved in outer membrane siderophore receptors because of the constriction of the barrel and the suppos edly identica l ferrisiderophor e translocation pathway. As expected, residues of the plug surrounding ligand binding sites are in non-cons erved short loop regions providing the necessary variability to specific ligand binding. Thes e residues are part of what are called the two `s ensor loops ’ (Buchanan et al. 1999), which are supposed to participate in the signalling to the periplasmic side of the iron-siderophor e loaded status of the receptor. The TonB box In FpvA, the TonB box domain, which interacts with the TonB protein, lies between the plug and the FecR box domains (figure 5). In the crystal structur e of FepA, the TonB box domain comprizing residues 12 ± 18, extends towards the periplas m and does not show any s econdary structur e elements , whereas this domain is too disordered in the FhuA crystal to be s een. The cons ensus s equenc e of the TonB box for a number of siderophor e receptors is the (D,E)TIVV(T,S) pattern, right before the beginning of the plug, but there is an obvious lack of cons ervation Pyoverdin outer membrane receptor FpvA within the whole family of siderophore receptors. For FpvA, the NAITIS pattern starting at residue 68 has been proposed as the TonB box (Poole et al. 1993), but this sequenc e analysis points more towards a location of the TonB box at residue 86 (ATMITS). The FecR box FepA belongs to the subgroup of receptors which contains an additiona l domain at the N-terminus , called the FecR box (~ 60 residues long). This domain plays a role in the signalization cascade triggered by the binding of the siderophor e to its receptor, which leads to transcription of the siderophor e and receptor genes (Braun 1997). The s equence alignments of the FecR domains are shown in figure 7. Apart from the N-termina l extension of the FauA protein (Bordetella pertussis), which may not be related to the same function, the extent of similarity between the s equences is significa nt enough to predict a simila r fold and most probably a simila r function (interaction of the domain with the FecR signal transducer or its analogues ) for this domain. No structura l information is available on this domain, since it is abs ent in FhuA and FepA. A s econdary structur e prediction was performed using the PHD programm e (Rost et al. 1994), which obtained a prediction for two a-helices with a high degree of confidenc e and one possible b-sheet (shown in figur e 7). Mechanis m of signalling of the loaded status of the receptor and the unsolved translocation pathway of the siderophore In the FhuA-ferrichrom e X-ray structur e (Ferguson et al. 1998, Locher et al. 1998), the ferrichrome is bound above the plug, far from the membrane, by hydrogen and van der Waals contact. The binding site is formed by three residues of the plug domain and two residues of the barrel domain. 129 Comparison of the ligand-free and siderophor e-bound structure of FhuA highlights the mechanis m of signalling of the loaded status of the siderophor e to the TonB protein across the outer membrane. The binding of ferrichrom e to FhuA induces only small conformationa l changes in the binding pocket. However, thes e changes are amplified towards the periplasmic side, and result in the unwinding Ê of the a-carbon of an alpha helix and a displacement of 17 A of residue W-22 (Locher et al. 1998). These large conformationa l changes probably indicate the ligand loaded status of the receptor, by making the TonB box of FpvA ready to interact with the TonB protein. Although the loops of the FpvA plug located at the expected site for pyoverdin binding are totally different from the corresponding FhuA loops involved in ferrichrome binding, the s equenc e cons ervation of the core of the plug and of its sites of interaction with the b-barrel inner wall strongly suggest that this mechanis m of signal transduction applies to FpvA and to pyoverdin. The most unexpected structural feature of the FhuA and FepA crystal structures is this globula r domain, which blocks the passage of small molecules through the barrel (figure 4). Given the extensiv e surface of interaction between this domain and the inner-wall of the barrel, it s eems unlikely that TonB may provide sufficient energy to break more than 40 hydrogen bonds and s everal ion pairs to `unplug’ the barrel, in order to allow the passage of the siderophor e into the periplasm. Analysis of the cavities of the FhuA and FepA 3D structures and of cons ervation among receptors of zones of weakness in the interactions between the plug and the barrel did not provide clear clues on a possible common pathway of translocation for the iron-siderophor e complex. Clearly, more structura l and functional work is needed to understand how the TonB protein interacts with the receptor, and what are the conformationa l changes that allow the siderophore to travel through the protein. Figure 7. FecR box sequences alignment. Identities are shown in outline, bold and dark shade while similarities are in outline and light shade (Using MacVector). Outer membrane targeting sequences were removed from N-terminal. The two helices predicted by PHD are shown with an arrow (ClustalW alignment, Higgins et al. 1996). 130 N. Folschweiller et al. Fluorescence energy transfer as a tool to study the dynamics of pyoverdin to its receptor in vivo and in vitro The dynamics of siderophore binding to the bacterial envelope and the different steps leading to its translocation across the outer membrane are not yet well understood, due ess entially to the unavailability of proper experimenta l tools . Fluorescence resonanc e energy transfer (FRET) is one of the most powerful techniques to follow protein ligand interactions with time, provided that a suitable pair of fluorescenc e donors /acceptors can be us ed for a given biological system. Generally, chemical modifications of the ligand and/or the protein are required to achieve thes e goals . With the FpvA/pyoverdin system, one is in a unique situation among the family of siderophor e receptors where the siderophor e contains a chromophore with spectral properties suitable to carry out FRET experiments. According to thes e two spectra (figure 8), there is a good overlap between the Tryptopha n emission spectra of FpvA and the absorption spectra of iron-free PaA. The pyoverdin chromophor e exhibits a large Stokes’ shift, leading to a small overlap between absorption and emission spectra and, hence, to minima l problems of self-absorption. Thes e spectral characteristics are almost ideal to observe FRET between iron-free PaA and Trp of FpvA, if they are at a short distanc e from each other, i.e. when pyoverdin is bound to its receptor. Indeed, the characterization by fluorescence spectroscopy of purified FpvA produced by a PaA-producing Ps eudomonas a eruginos a strain (Schalk et al. 1999), revealed an unexpected fluorescence peak at 447 nm upon excitation of the Trp at 290 nm, demonstrating a copurifica tion of FpvA with iron-free PaA. Interestingly, in none of the various steps of the purification process , apo-PaA was releas ed from FpvA: it was tightly bound to the receptor and could be removed from it only using harsh conditions (e.g. urea, strong acidic media). After complexation with iron(III), the fluorescenc e of PaA is quenched by the metal: the ferric-PaA complex is no more fluorescent and no FRET can be obs erved for the FpvA-PaA-F e complex (figure 9). For these reasons , FRET is a very us eful tool to distinguis h between both forms of FpvA: FpvA-PaA and FpvA-PaA-F e. In vivo, FRET between Trp of FpvA and iron-free PaA can also be measured In vivo FRET was only obs erved in the P. a eruginos a strain producing both FpvA and PaA and not with the parent strain unable to synthesiz e FpvA but which s ecretes large amounts of PaA in the medium. Similarly, with a P. a eruginos a strain which is a FpvA-overproducing and a PaA-deficient mutant, a typical tryptophan emission spectra without any emission at 447 nm was obtained (I. Schalk, persona l communication). All these data show that FpvA-PaA complexes exist in P. a eruginos a cells, and that PaA loaded FpvA may be the normal state of the receptor. To the authors ’ knowledge, no other data are available on the binding of metal-free siderophor e to cells, outer membranes or purified receptors . However, transport assays on P. aeruginos a, using iron-free [3 H]PaA, have shown that, although iron-free PaA binds to FpvA, it is not transported inside the cell, in contrast to ironloaded PaA. Kinetics of FpvA-PaA and FpvA-PaA-Fe complexes FRET experiments were also us ed to study the kinetics of formation of FpvA-PaA or FpvA-PaA-F e complexes in vivo and in vitro. The kinetics of formation of FpvA-PaA-F e or Figure 8. UV spectrum of iron free PaA (thin line) and fluorescence spectra of purified FpvA (dotted line) and iron-free PaA (thick line). For the UV spectrum of iron-free PaA (thin line), PaA was dissolved in 50 mM pyridin-acetic acid buffer pH 5.0. For the fluorescence spectra of purified FpvA (dotted line) and iron-free PaA (thick line) the excitation wavelength was set respectively at 290 nm and 380 nm. The protein and iron-free PaA were dissolved in 50 mM Tris-HCl buffer pH 8.0 and 1 % (v/v) octyl-POE. Pyoverdin outer membrane receptor FpvA 131 Figure 9. Fluorescence spectra of purified FpvA (thick line) and FpvA-PaA (thin line) (Schalk et al. 1999). The excitation wavelength was set at 290 nm. Proteins were dissolved in 50 mM Tris-HCl buffer pH 8.0 and 1% (v/v) octyl-POE. Note that the emission maximum at 447 nm with FpvA-PaA (thin line) is absent with FpvA (thick line). FpvA-PaA complexes when PaA-free FpvA is incubated in the pres ence of PaA-F e or iron-free PaA are fast in vivo and in vitro. The saturation plateau of the complexes formation were reached in all four cas es in less than 15 min (t1/2 < 7 min, table 1). Thes e kinetics have been determined by FRET and also independently using labelled PaA-[55Fe]. Indeed, since PaA-F e is not fluorescent, FRET is obs erved only for the FpvA-PaA complex, and the formation of the complex could be monitored by the increas e of the emission peak at 447 nm. Surprisingly, a very slow FpvA-PaA-F e complex formation was obs erved, when purified FpvA-PaA was incubated in the presence of PaA-F e (Schalk et al. 1999). The plateau of the protein-siderophor e-iron(III) complex formation was reached only after 50 h incubation at room temperature (t1/2 = 24 h, table 1). The kinetics were determined by monitoring the loss of fluorescenc e energy transfer. However, the rate of formation of the FpvA-PaA-F e complex in vivo, when the FpvA- and PaA-producing P. a eruginos a cells were incubated in the presence of PaA-F e, was considerably faster (t1/2 = 4 min, table 1) than with the purified receptor in vitro, and was of the same order of magnitude as the rate of iron transport into the cells (Schalk, personal communication). Thes e findings sugges t that, during iron transport, a triggering mechanis m is activated, which speeds up the exchange between iron-free PaA and ferric-PaA on FpvA. It had been shown previously in E. coli, that alterations in the receptors TonB box (Reynolds et al. 1980, Gundmundsdot tir et al. 1989) or abs ence of the TonB protein (Frost and Ros enberg 1975) s eem to affect the kinetics of ferricsiderophor e binding or releas e, as well as the displacement Table 1. Half life of formation of FPvA-PaA and FpvA-PaA-Fe complexes in vivo and in vitro. Receptors incubated in presence of PaA or PaA-Fe Complexes studied FpvA in vitro FpvA-PaA t1/2 < 7 mina FpvA-PaA-Fe t1/2 Ç < 7 minb a b in vivo FpvA-PaA in vitro in vivo t1/2Ç < 7 mina t1/2Ç < 7 minb t1/2 = 24 ha t1/2 = 4 mina The kinetics were determined by FRET. The kinetics were determined using radioactive PaA-[55Fe] complex. of the ferric-siderophor e by phages. Genetic and biochemica l evidences demonstrated that the TonB-dependent outer membrane transporters contact TonB directly (Braun et al. 1991, Skare and Postle 1991, Skare et al. 1993, Cadieux and Kadner 1999). Ferrichrom e greatly enhanced the formation of the FhuA-TonB complex in vivo (Moeck et al. 1997). TonB levels have not been quantified, they are likely to be much less than the numbers of transport proteins (Postle and Reznikoff 1979, Postle 1990) and compete for TonBdependent transporters (Kadner and Heller 1995). According to the fluorescenc e data, TonB may also play a role in regulating the binding and exchange of ferric-PaA and ironfree PaA at the cell surface of P. a eruginos a. Affinities of iron-free PaA and ferric-PaA to FpvA and efficiency of transport in vivo The binding of iron-free PaA to FpvA in vivo raises, however, a fundamenta l question. In the medium, under iron limitation, the free form of the siderophor e is always in excess with 132 N. Folschweiller et al. respect to the iron-loaded siderophore. How, then, can one explain an efficient transport mechanis m for iron, since ironloaded PaA should compete with a large excess of iron-free PaA? The affinity of PaA-Fe cannot be directly measured in energiz ed cells, because PaA-F e is transported into the cell and, therefore, the relative affinities of iron-free and ferricPaA to FpvA cannot be compared. Preliminary experiments with TonB( Ð ) strains (no coupling and no transport) indicate a 17-fold difference for the affinities of PaA and PaA-F e (Kda pp = 1.3 nM and Kda pp = 19 nM, respectively). At first sight, this difference in the affinities for FpvA in favour of iron-free PaA may not be enough to explain an efficient transport mechanis m for iron. However, one may speculate that cell energization may enhanc e discrimination by the receptor between the two ligands . Moreover, efficient uptake of iron may not require large differences in affinity between the two forms of the siderophore for their receptor for the following reasons : · · The better affinity of PaA-F e to FpvA in energis ed cells is combined with a fast exchange of apo-PaA by ferric-PaA on FpvA in vivo, and Only ferric-PaA is transported and not apo-PaA. The combination of these two effects may be compatible with an efficient iron transport mechanis m in P. a eruginos a. Acknowledgements This paper is based on a contribution to a Joint Meeting of the Britis h Biophysical Society and Unilever Research, Port Sunlight, on `Structure, Dynamics and Perturbations of Membranes’, held 19 ± 20 April 2000 in Chester, UK. References Abdallah, M. A., 1991, Pyoverdins and Pseudobactins. In Handbook of Microbial Iron Chelates, G. Winkelmann, ed. (Boca Raton, FL: CRC Press Inc.), pp. 139 ± 153. Ali, N. J., Kessel, D. and Miller, R. F., 1995, Broncho-pulmonary infection with Pseudomonas aeruginosa in patients infected with human immunodeficiency virus. Genito-urinary Medicine, 71, 73 ± 77. Ankenbauer, R. G., Tokoyni, T., Staley, A., Rinehart, K. L. and Cox, C. D., 1988, Synthesis and biological activity of pyochelin, a siderophore of Pseudomonas aeruginosa. Journal of Bacteriology, 170, 5344± 5351. Atkinson, R. A., Salah El Din, A. M., Kieffer, B., LefeÁ vre, J.-F. and Abdallah, M. A., 1998, Bacteria l iron-transport: 1H NMR determination of the three-dimensional structure of the gallium complex of pyoverdin G4R, the peptidic siderophore of Pseudomonas putida G4R. Biochemistry, 37, 15965 ± 15973. Barton, H. A., Johnson, Z., Cox, C. D., Vasil, A. I. and Vasil, M. L., 1996, Ferric uptake regulator mutants of Pseudomonas aeruginosa with distinct alternations in the iron-dependent repression of exotoxin A and siderophores in aerobic and microaerobic environments. Molecular Microbiology, 21, 1001 ± 1017. Braun, V., 1997, Avoidance of iron toxicity through regulation of bacterial iron transport. Biological Chemistry, 378, 779 ± 786. Braun, V., Gunter, K. and Hantke, K., 1991, Transport of iron across the outer membranes. Biology of Metals, 4, 14 ± 22. Briskot G., Taraz K. and Budzikiewicz H., 1989, Bacterial constituents. Pyoverdin-type siderophores from Pseudomonas aeruginosa. Liebigs Annalen der Chemie, 4, 375 ± 384. Buchanan, S. K., Smith, B. S., Venkatramani, L., Xia, D., E ss er, L., Palnitkar, M., Chakraborty, R., van der Helm, D. and Deisenhofer, J., 1999, Crystal structure of the outer membrane active transporter FepA from Escherichia coli. Nature Structural Biology, 6, 56 ± 63. Cadieux, N. and Kadner, R. J., 1999, Site-directed disulfide bonding reveals an interaction site between energy-coupling protein TonB and BtuB, the outer membrane cobalamin transporter. Proceedings of the National Academy of Sciences (USA), 96, 10673± 10678. Cornelis, P., Hohnadel, D. andMeyer, J. M., 1989, Evidence for different pyoverdine-mediated iron uptake systems among Pseudomonas aeruginosa strains. Infection and Immunity, 57, 3491± 3497. Cox, C. D., 1982, Effect of pyochelin on the virulence of Pseudomonas aeruginosa. Infection and Immunity, 36, 17 ± 23. Cox, C. D. and Adams, P., 1985, Siderophore activity of pyoverdin for Pseudomonas aeruginosa. Infection and Immunity, 48, 130 ± 138. Cox, C. D. and Graham, R., 1979, Isolation of an iron-binding compound from Pseudomonas aeruginosa. Journal of Bacteriology, 137, 357 ± 364. Crosa, J. H., 1984, The relationship of plasmid-mediated iron transport and bacterial virulence. Annual Reviews of Microbiology, 38, 69 ± 89. Cunliffe, H. E., Merriman, T. R. and Lamont, I. L., 1995, Cloning and characterization of pvdS, a gene required for pyoverdine synthesis in Pseudomonas aeruginosa: PvdS is probably an alternative sigma factor. Journal of Bacteriology, 177, 2744 ± 2750. Demange, P., Wendenbaum, S., Linget, C., Mertz, C., Cung, M. T., Dell, A. and Abdallah, M. A., 1990, Bacteria l siderophores: structure and NMR assignment of Pyoverdins Pa, siderophores of Pseudomonas aeruginosa ATCC 15692. Biology of Metals, 3, 155 ± 170. Ferguson, A. D., Hofman, E., Coulton, J. W., Diederichs, K. and Welte, W., 1998, Siderophore-mediated iron transport: crystal structure of FhuA with bound lipopolysaccharide. Science, 282, 2215 ± 2220. Fichtenbaum, C. J., Woeltfe, K. F. and Powderly, W. G., 1994, Serious Pseudomonas aeruginosa infections in patients infected with human immunodeficiency virus: a case control study. Clinical Infectious Diseases, 19, 417 ± 422. Frost, G. E. and Rosenberg, H., 1975, Relationship between the TonB locus and iron transport in Escherichia coli. Journal of Bacteriology, 124, 704 ± 712. Gensberg, K., Hughes, K. and Smith, A. W., 1992, Siderophorespecific induction of iron uptake in Pseudomonas aeruginosa. Journal of General Microbiology, 138, 2381± 2387. Griffiths, E., 1987, The iron-uptake systems of pathogenic bacteria. In Iron and Infection, J. J. Bullen and E. Griffiths, eds (New York: John Wiley), pp. 69 ± 137. Gundmundsdottir, A., Bell, P. E., Lundrigan, M. D., Bradbeer, C. and Kadner, R. J., 1989, Point mutations in a conserved region (TonB box) of Escherichia coli outer membrane protein BtuB affect vitamin B12 transport. Journal of Bacteriology, 171, 6526 ± 6533. Guerinot, M. L., 1994, Microbia l iron transport. Annual Reviews of Microbiology, 48, 743 ± 772. Heinrich, D. E. and Poole, K., 1993, Cloning and sequence analysis of a gene (pchR) encoding an AraC family activator of pyochelin and ferripyochelin receptor synthesis in Pseudomonas aeruginosa. Journal of Bacteriology, 175, 5882 ± 5889. Heinrich, D. E. and Poole, K., 1996, PchR, a regulator of ferripyochelin receptor gene (fptA) expression in Pseudomonas aeruginosa, functions both as an activator and as a repressor. Journal of Bacteriology, 178, 2586 ± 2592. Higgins, D. G., Thompson, J. D. and Gibson, T. J., 1996, Using Clustal for multiple sequence alignments. Methods in E nzymology, 266, 383 ± 402. Hohnadel, D. and Meyer, J.-M., 1988, Specificity of pyoverdinmediated iron uptake among fluorescent Pseudomonas strains. Journal of Bacteriology, 170, 4865 ± 4873. Pyoverdin outer membrane receptor FpvA Huschka, H.-G., Naegeli, H. U., Leuenberger-Ryf, H., KellerSchierlein, W. and Winkelmann, G., 1985, Evidence for a common siderophore transport system but different siderophore receptors in Neurospora crassa. Journal of Bacteriology, 162, 715 ± 721. Kadner, R. J., 1990, Vitamin B12 transport in E scherichia coli energy coupling between membranes. Molecular Microbiology, 4, 2027 ± 2033. Kadner, R. J. and Heller, K. J., 1995, Mutual inhibition of cobalamin and siderophore uptake systems suggests their competition for TonB function. Journal of Bacteriology, 177, 4829± 4835. Kraulis, P. J., 1991, MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures . Journal of Applied Crystallography, 24, 946 ± 950. Larsen, R. A., Foster-Hartnett, D., McIntosh, M. A. and Postle, K., 1997, Regions of Escherichia coli TonB and FepA proteins essential for in vivo physical interactions. Journal of Bacteriology, 179, 3213 ± 3221. Larsen, R. A., Myers, P. S., Skare, J. T., Seachord, C. L., Darveau, R. P. and Postle, K., 1996, Identifica tion of TonB homologs in the family E nterobacteriacea e and evidence for conservation of TonBdependent energy transduction complexes. Journal of Bacteriology, 178, 1363± 1373. Larsen, R. A., Thomas, M. G. and Postle, K., 1999 Protonmotive force, ExbB and ligand-bound FepA drive conformational changes in TonB. Molcular Microbiology, 31, 1809± 1824. Leoni, L., Ciervo, A., Orsi, N. and Visca, P., 1996, Iron-regulated transcription of the pvdA gene in Pseudomonas aeruginosa: effect of Fur and PvdS on promoter activity. Journal of Bacteriology, 178, 2299 ± 2313. Linget, C., Azadi, P., MacLeod, J. K., Dell, A. and Abdallah, M. A., 1992, Bacterial siderophores: structures of the pyoverdins of Pseudomonas fluorescens ATCC 13525. Tetrahedron Letters, 33, 1737 ± 1740. Litwin, C. M. and Calderwood, S. B., 1993, Role of iron in regulation of virulence genes. Clinical Microbiological Reviews, 6, 137 ± 149. Locher, K. P., Rees, B., Koebnik, R., Mitschler, A., Moulinier, L., Rosenbush, J. P. and Moras, D., 1998, Transmembrane signaling across the ligand-gated FhuA receptor: crysta l structures of free and ferrichrome-bound states reveal allosteric changes. Cell, 95, 771 ± 778. Mertz, C., Demange, P., Marraud, M., Dell, A., Linget, C, Abdallah, M. A. and Cung, M. T., 1991, Conformational study of the Al(III)Pyoverdin complex from Pseudomonas aeruginosa by 1 H 2D NMR techniques. In Peptides 1990, E.Giralt and D. Andeu, eds (Leiden: ESCOM), pp. 543 ± 544. Miyazaki, H., Kato, H., Nakazawa, T. and Tsuda, M., 1995, A positive regulatory gene, pvdS, for expression of pyoverdin biosynthesis genes in Pseudomonas aeruginosa PAO. Molecular and General Genetics, 248, 17 ± 24. Moeck, G. S., Coulton, J. W. and Postle, K., 1997, Cell envelope signaling in Escherichia coli. Ligand binding to the ferrichrom e-iron receptor Fhua promotes interaction with the energy-transducing protein TonB. Journal of Biological Chemistry, 272, 28391± 28397. Mohn, G., Koehl, P., Budzikiewicz, H. and LefeÁvre, J.-F., 1994, Solution structure of pyoverdin GMII, Biochemistry, 33, 2843 ± 2851. Neilands, J. B., 1981, Iron absorption and transport in microorganisms. Annual Reviews of Nutrition, 1, 27 ± 46. Neilands, J. B., 1982, Microbia l envelope proteins related to iron. Annual Reviews of Microbiology, 36, 285 ± 309. Neilands, J. B., Konopka, K., Schwyn, B., Coy, M., Francis, R. T., Paw, B. H. and Bagg, A., 1987, Comparative biochemistry of microbial iron assimilation. In Iron transport in Microbes, Plants and Animals, G. Winkelmann, D. van der Helm and J. B. Neilands, eds (Weinheim, Germany: VCH Verlagsgesellschaft), pp. 3 ± 33. Ochsner, U. A., Vasil, A. I. and Vasil, M. L., 1995, Role of ferric uptake regulator of Pseudomonas aeruginosa in the regulatuion of siderophores and exotoxin A expression: purification and activity on iron-regulated promoters. Journal of Bacteriology, 177, 7194 ± 7201. 133 Pattus, F. and Abdallah, M. A., 2000, Siderophores and irontransport in microorganisms. Journal of the Chinese Chemical Society, 47, 1 ± 20. Poole, K., Neshat, S., Krebes, K. and Heinrichs, D. E., 1993, Cloning and nucleotide sequence analysis of the ferripyoverdine receptor gene FpvA of Pseudomonas aeruginosa. Journal of Bacteriology, 175, 4597 ± 4604. Postle K., 1990, Aerobic regulation of the Escherichia coli tonB gene by changes in iron availability and the fur locus. Journal of Bacteriology, 172, 2287± 2293. Postle, K., 1993, TonB protein and energy transduction between membranes. Journal of Bioenergetics and Biomembranes, 25, 591 ± 601. Postle, K. and Reznikoff, W. S., 1979, Identifica tion of the Escherichia coli tonB gene product in minicells containing tonB hybrid plasmids. Journal Molecular Biology, 131, 619 ± 636 Postle, K. and Skare, J. T., 1988, Escherichia coli TonB protein is exported from the cytoplasm without proteolytic cleavage of its amino terminus. Journal of Bioenergetics and Biomembranes, 263, 11000 ± 11007. Prince, R. W., Cox, C. D. and Vasil, M. L., 1993, Coordinate regulation of siderophore and exotoxin A production: molecular cloning and sequencing of the Pseudomonas aeruginosa fur gene. Journal of Bacteriology, 175, 2589± 2598. Reynolds, P. R., Mottur, G. P. and Bradbeer, C., 1980, Transport of vitamin B12 in E scherichia coli. Some observations on the roles of the gene products of BtuC and TonB. Journal of Bioenergetics and Biomembranes, 255, 4313 ± 4319. Rost, B., Sander, C. and Schneider, R., 1994, PHD, an automatic mail server for protein secondary structure prediction. Computer Applications in the Biosciences, 10, 53 ± 60. Sawatzki, G., 1987, The role of iron binding proteins in bacterial infections. In Iron transport in Microbes, Plants and Animals, G. Winkelmann, D. van der Helm and J. B. Neilands, eds (Weinheim, Germany: VCH Verlagsgesellschaft), pp. 477 ± 489. Schalk, I. J., Kyslik, P., Prome, D., van Dorsselaer, A., Poole, K., Abdallah, M. A. and Pattus, F., 1999, Copurification of the FpvA ferric-pyoverdin receptor of Pseudomonas aeruginosa with its iron-free ligand: Implications for siderophore-mediated iron uptake. Biochemistry, 38, 9357 ± 9365. Skare, J. T. and Postle, K., 1991, Evidence for a TonB-dependent energy transduction complex in Escherichia coli. Molecular Microbiology, 5, 2883± 2890. Skare, J. T., Ahmer, B. M., Seachord, C. L., Darveau, R. P. and Postle, K., 1993, Energy transduction between membranes. TonB, a cytoplasmic membrane protein, can be chemically cross-linked in vivo to the outer membrane receptor FepA. Journal of Bioenergetics and Biomembranes, 268, 16302 ± 16308. Stojiljkovic, I. and Srinivasan, N., 1997, Neisseria meningitidis tonB, exbB, and exbD genes: TonB-dependent utilization of proteinbound iron in Neiss eriae. Journal of Bacteriology, 179, 805 ± 812. StruyveÂ, M., Moons, M. and Tommassen, J., 1991, Carboxy-terminal phenylalanine is essential for the correct assembly of a bacterial outer membrane protein. Journal of Molecular Biology, 218, 141 ± 148. Teintz e, M., Hossain, M. B., Barnes, C. L., Leong, J. and van Der Helm, D., 1981, Structure of ferric pseudobactin, a siderophore from a plant growth promoting Pseudomonas. Biochemistry, 20, 6446± 6457. Visca, P., Ciervo, A., Sanfilippo, V. and Orsi, N., 1993, Iron-regulated salicylate synthesis by Pseudomonas aeruginosa spp. Journal of General Microbiology, 139, 1995 ± 2001. Winkelmann, G. and Huschka, H.-G., 1987, Molecular recognition and transport of siderophores in fungi. In Iron transport in Microbes, Plants and Animals, G. Winkelmann, D. van der Helm, and J. B. Neilands, eds (Weinheim, Germany: VCH Verlagsgesellschaft), pp. 317 ± 336. Received 19 May 2000, and in revised form 12 July 2000.