New insights into the metal specificity of thePseudomonas aeruginosa pyoverdine–iron uptakepathway

Armelle Braud,1† Françoise Hoegy,1†

Karine Jezequel,2 Thierry Lebeau2 andIsabelle J. Schalk1*1Métaux et Microorganismes, Chimie, Biologie etApplications, UMR 7175-LC1, CNRS-Université LouisPasteur, ESBS, Blvd Sébastien Brant, F-67413 Illkirch,Strasbourg, France.2Equipe Dépollution Biologique des Sols (EDBS),Université de Haute Alsace, 29, rue de Herrlisheim, BP50568 Colmar Cedex, France.

Summary

Pyoverdine (PvdI) is the major siderophore secretedby Pseudomonas aeruginosa PAOI in order to getaccess to iron. After being loaded with iron in theextracellular medium, PvdI is transported across thebacterial outer membrane by the transporter, FpvAI.We used the spectral properties of PvdI to show thatin addition to Fe3+, this siderophore also chelates, butwith lower efficiencies, all the 16 metals used in ourscreening. Afterwards, FpvAI at the cell surface bindsAg+, Al3+, Cd2+, Co2+, Cu2+, Fe3+, Ga3+, Hg2+, Mn2+, Ni2+ orZn2+ in complex with PvdI. We used InductivelyCoupled Plasma-Atomic Emission Spectrometry tomonitor metal uptake in P. aeruginosa: TonB-dependent uptake, in the presence of PvdI, was onlyefficient for Fe3+. Cu2+, Ga3+, Mn2+ and Ni2+ were alsotransported into the cell but with lower uptake rates.The presence of Al3+, Cu2+, Ga3+, Mn2+, Ni2+ and Zn2+ inthe extracellular medium induced PvdI production inP. aeruginosa. All these data allow a better under-standing of the behaviour of the PvdI uptake pathwayin the presence of metals other than iron: FpvAI at thecell surface has broad metal specificity at the bindingstage and it is highly selective for Fe3+ only during theuptake process.

Introduction

Bacteria produce low-molecular-weight ligands, calledsiderophores, under iron stress (Boukhalfa and Crumb-liss, 2002; Winkelmann, 2002). Siderophores bind ferriciron and transport it back into the bacterial cells, via iron–siderophore outer membrane receptors (Braun, 2003;Braun and Endriss, 2007). Siderophores play an impor-tant role in the extracellular solubilization of iron fromminerals or organic substances. There is a wide range ofsiderophore concentrations in soil, from tens of micro-moles to a few millimoles per litre (Hersman et al., 1995).These molecules are characterized by a high affinityfor ferric iron [Ka = 1043 and 1032 M-1 for enterobactinand pyoverdine respectively (Albrecht-Gary et al., 1994;Raymond et al., 2003)], but they will also form complexeswith metals other than Fe3+, although with lower affinity(Hernlem et al., 1996; Neubauer et al., 2000). Forexample, the formation constants for complexes ofhydroxamate siderophore desferrioxamine B with Ga3+,Al3+ and In3+ are between 1020 and 1028 M-1 (Evers et al.,1989) and those for pyoverdine with Zn2+, Cu2+ and Mn2+

are between 1017 and 1022 M-1 (Chen et al., 1994). More-over, desferrioxamine B possesses an extraordinaryability to coordinate plutonium(IV), a much larger andstructurally diverse metal center than that of Fe3+ (Neuet al., 2000). Industry and agriculture distribute heavymetals and other metals in soil, sediment, waste andwastewater. Therefore, high concentrations of heavymetals are observed in the environment and may ofcourse cause environmental stress conditions for micro-organisms. Some metals, like cadmium, lead andmercury, are simply toxic for bacteria. Other metals, likezinc and copper, are required in trace concentrations forthe regular metabolism of the cell, but high intracellularconcentrations of these metal ions interfere with the func-tion of essential proteins and are toxic. High heavy metalconcentrations in the environment may also interfere withmicroorganism siderophore–iron uptake pathways andheavy metal toxicity may be modulated by the presence orthe absence of siderophores. Heavy metals enter theperiplasm of Gram-negative bacteria mostly by diffusion;thus, the binding of heavy metal to siderophore dramati-cally changes the free metal concentration, affecting this

Received 10 June, 2008; accepted 5 November, 2008. *Forcorrespondence. E-mail schalk@esbs.u-strasbg.fr; Tel. (+33)390 24 47 19; Fax (+33) 390 24 48 29. †Both authors contributedequally to this work.

Environmental Microbiology (2009) 11(5), 1079–1091 doi:10.1111/j.1462-2920.2008.01838.x

© 2008 The AuthorsJournal compilation © 2008 Society for Applied Microbiology and Blackwell Publishing Ltd

diffusion process. Moreover, previous studies in a numberof bacteria and fungi show that metals other than ironstimulate siderophore production (Huyer and Page, 1988;Hofte et al., 1993; Hu and Boyer, 1996). This stimulationof siderophore production may protect the bacteriaagainst metal toxicity: the metal is sequestered in theextracellular medium by the siderophore and is thus nolonger able to enter the bacterium by diffusion.

Fluorescent Pseudomonas species, common inhabit-ants of soil and water, are characterized by the overpro-duction under iron-deficient conditions of yellow-green,water-soluble siderophores, called pyoverdines (Pvds)(Meyer et al., 2002). More than 100 Pvds have beenidentified, forming a large class of mixed catecholate-hydroxamate siderophores characterized by a conserveddihydroxyquinoline-derived chromophore to which apeptide chain of variable length and composition isattached (Budzikiewicz, 1997; Meyer et al., 2002). Thesize and amino acid composition of Pvds are unique toeach species of Pseudomonas, indicating specializationby each bacterium of its Pvd (Meyer et al., 1997). InPseudomonas aeruginosa, three structurally differentPvds (with different peptide chains) have been identified(Pvd types I, II and III) (Cornelis et al., 1989; Meyeret al., 1997; De Vos et al., 2001). Each strain ofP. aeruginosa produces one of these three types of Pvd.Once iron-loaded, Pvds are recognized at thePseudomonas cell surface by a specific outermembranetransporter. Only four Pvd transporters have beencloned, all from P. aeruginosa: FpvAI and FpvB, involvedin the uptake of PvdI (Fig. 1) (Ghysels et al., 2004) andFpvAII and FpvAIII (de Chial et al., 2003) for PvdII andPvdIII uptake respectively. The P. aeruginosa FpvAI isthe best characterized. Its structure has been solved indifferent loading states (FpvAI, FpvAI–PvdI and FpvAI–PvdI–Fe) and its interactions with PvdI have beenstudied using the fluorescent properties of this sidero-phore (Schalk et al., 1999; 2001; Clément et al., 2004;

Greenwald et al., 2006). FpvAI of P. aeruginosa PAO1,like all outer membrane transporters, consists of aC-terminal b-barrel domain and an N-terminal plugdomain filling the barrel (Cobessi et al., 2005). Thebinding site for the ferric siderophore is located abovethe plug, well outside the membrane, and is composed ofresidues from the plug and b-barrel domains. The FpvAIbinding site consists mostly of aromatic residues, includ-ing six Tyr residues and two Trp residues, and only threehydrophilic residues (Cobessi et al., 2005). In the PvdI/FpvAI system it has been shown that FpvAI is alsoinvolved in a signalling cascade controlling the expres-sion of fpvA and genes related to PvdI biosynthesis [fora review see (Visca et al., 2002)]. This cascade involvesa transmembrane signalling system induced by thebinding of the siderophore to the outer membrane trans-porter, and the extracytoplasmic function (ECF) sigmafactor/antisigma factor pairs, FpvI/FpvR and PvdS/FpvR(Wilson and Lamont, 2000; Wilson et al., 2001; Lamontet al., 2002; Shen et al., 2002; Visca et al., 2002; Beareet al., 2003; Redly and Poole, 2003; 2005). Severalstudies have been published on the mechanisms of inter-action between FpvAI and PvdI or PvdI–Fe (for a reviewsee Schalk, 2008) and the siderophore recognition speci-ficity of this transporter in binding different Pvds or modi-fied Pvd (Hohnadel and Meyer, 1988; Meyer et al., 1999;Schons et al., 2005). Previous studies have also shownthat PvdI is able to chelates Ga3+, Al3+, Cr3+ and V4+

(Baysse et al., 2000; Folschweiller et al., 2002; Green-wald et al., 2006), and Pvd produced by P. fluorescens(CCUG 32456), curium (Moll et al., 2008), and that thePvdI–Ga, PvdI–Alu and PvdI–Cr complexes are able tobind to FpvAI at the cell surface (Folschweiller et al.,2002; Greenwald et al., 2006; 2008). Nevertheless, thereis a total lack of information in the way the FpvAI/PvdIsystem and siderophore uptake pathways in generalhandle the presence of metals other than iron in theirenvironment. However, as discussed above, bacteria

Fig. 1. Structure of PvdI.

1080 A. Braud et al.

© 2008 The AuthorsJournal compilation © 2008 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 11, 1079–1091

and the siderophores they produce are in the presenceof various metals in their natural environment.

In this study, we investigated the behaviour of the PvdI/FpvAI uptake pathway of P. aeruginosa PAO1 in the pres-ence of 16 metals. PvdI chelated all the metals tested inthis study: Ag+, Al3+, Cd2+, Co2+, Cr2+, Cu2+, Eu3+, Fe3+, Ga3+,Hg2+, Mn2+, Ni2+, Pb2+, Sn2+, Tb3+, Tl+ and Zn2+. Eight ofthese PvdI–metal complexes inhibited PvdI–55Fe uptakeand six induced PvdI production by P. aeruginosa PAO1.In the presence of PvdI, an efficient proton-motive depen-dent uptake was only observed for Fe3+ and with clearlylower uptake rates for Ga3+, Cu2+, Mn2+ and Ni2+. All thesedata indicate that the FpvAI binding site is able to interactwith PvdI in complex with different metals, but the highspecificity of the uptake mechanism allows only an effi-cient accumulation of Fe3+ inside P. aeruginosa cells. Thetoxicity of the metals tested were not linked to uptake bythe PvdI/FpvAI system. All these data allow a betterunderstanding of the behaviour of the PvdI uptakepathway in the presence of metals other than iron.

Results

Ability of PvdI to chelate metals other than Fe

PvdI is characterized by an emission of fluorescence at450 nm, by absorbances at 360 and 380 nm at pH 5.0(Table 1; Fig. 2A and B; Albrecht-Gary et al., 1994), andby a unique absorbance band at 400 nm at pH 8.0(Albrecht-Gary et al., 1994). If loaded with Ga3+, the PvdIfluorescence is markedly increased and the absorbanceslightly modified (Folschweiller et al., 2002). By contrast, ifloaded with iron, the UV spectrum is modified and showstwo bands at 400 and 450 nm, at pH 5.0 and 8.0 (Table 1;Albrecht-Gary et al., 1994) and the fluorescence isquenched by the metal (Albrecht-Gary et al., 1994; Schalket al., 1999). This is not surprising as spectral variationsof siderophores, synthetic chelators or proteins in thepresence of metals have been observed and are usedclassically to determine the kinetic and thermodynamicparameters of ligand–metal interactions (Carrano et al.,1996; Albrecht-Gary and Crumbliss, 1998; Palancheet al., 1999; 2004; Dhungana et al., 2004). In this study,these properties were used to investigate whether PvdIchelates other metals more or less efficiently than iron.The buffer choosen was 50 mM pyridin AcOH pH 5.0. Atthis pH the variation between the UV and fluorescencespectra of Pvd–Ga, Pvd–Al complexes (both describedpreviously) and metal-free PvdI, were more significantthan at neutral pH (Albrecht-Gary et al., 1994; Folsch-weiller et al., 2002). PvdI was incubated overnight in thepresence of each metal, as some metal–siderophoreinteractions display slow kinetics. None of the spectralchanges observed in UV and/or in fluorescence (Fig. 2

and Table 1) were observed if each metal was in solutionin the buffer in the absence of PvdI. Spectral changeswere observed only in the presence of both entities,indicating that they were due to PvdI–metal complexformation.

In pyridin AcOH pH 5.0 buffer, we observed significantchanges in the UV profile of PvdI in the presence of Al3+,Cr2+, Cu2+, Eu3+, Ga3+, Pb2+, Sn2+ and Tb3+, indicating thatthe siderophore is able to form complexes with thesemetals (Fig. 2A and B, Table 1). For most of these metals,new absorbance bands appeared between 380 and450 nm. These changes in the UV spectra are large andvery specific for each PvdI–metal complex when thespectra are carried out in the pyridin AcOH pH 5.0 buffer.As the absorbance of metal-free PvdI at pH 7.0 is shiftedto 400 nm (Albrecht-Gary et al., 1994), these changes inthe UV spectra are less important at neutral pH. For theeight other metals (Ag+, Cd2+, Co2+, Hg2+, Mn2+, Ni2+, Tl+

and Zn2+), no real change in the UV spectra was observedcompared with that of metal-free PvdI (Table 1). For somemetals, including Ag2+, Cd2+ and Tl+, the spectra weresimilar to that of metal-free PvdI, except that a smalldecrease (10–20%) of the absorbance level wasobserved at 360 and 380 nm. However, a lack of changein the UV spectra does not prove that PvdI is not able tochelate these metals.

Table 1. Fluorescent properties of PvdI in the presence of variousmetals.

Metals Absorbance (nm) Fluorescence

No metal 360, 380 ReferenceFe3+ 400, 450 QuenchingAg+ Not modified ¥0.3Al3+ 400 ¥18.5Cd2+ Not modified ¥22.4Co2+ Not modified ¥1.26Cr2+ 388, 410 ¥6.6Cu2+ 407 QuenchingEu3+ 365, 378 ¥1.9Ga3+ 402 ¥2.9Hg2+ Not modified ¥18.1Mn2+ Not modified ¥10.4Ni2+ Not modified ¥12.4Pb2+ 363, 380, 405 ¥21.9Sn2+ 362, 380, 400 ¥3.7Tb3+ 363, 380 ¥11.9Tl+ Not modified ¥1.9Zn2+ Not modified ¥2.3

The UV spectra corresponding to these data are represented inFig. 2A and B and the fluorescence spectra in Fig. 2C. PvdI of 50 mMwas incubated overnight in the presence of 500 mM of metal in 50 mMpyridin-AcOH pH 5.0, and the absorbance was monitored. For thefluorescence spectra, the mixtures were diluted to a concentration of5 mM PvdI and 50 mM of metal. For both the UV and the fluorescencecolumns, the spectra of PvdI in the presence of each metal arecompared with the UV or fluorescence spectra of metal-free PvdI. Notmodified: the absorbance is not modified compared with the metal-free PvdI UV spectrum. As a control, for each metal the UV and thefluorescence spectra were monitored in the same buffer but in theabsence of siderophore.

Metal specificity of the FpvA/Pvd uptake system 1081

© 2008 The AuthorsJournal compilation © 2008 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 11, 1079–1091

We observed maximal fluorescence similar to that offree PvdI, between 445 and 460 nm, for all metals tested(Fig. 2C); exceptions included Fe3+ and Cu2+. In these twocases, a high fluorescence quenching was observed, indi-cating the presence of PvdI–Fe and PvdI–Cu complexes.Partial quenching was observed in the case of Ag+, indi-cating that for this metal, under the experimental condi-tions used, we may have a mixture of PvdI–Ag and PvdI.

Fluorescence of PvdI incubated in the presence of allother metals was greater than that of metal-free PvdI in50 mM pyridin AcOH pH 5.0 (Fig. 2C, Table 1). This canonly be explained by the presence of PvdI–metal com-plexes or the presence of a mixture of PvdI and PvdI–metal. Thus, we can conclude that this siderophore formscomplexes with all the metals tested. When the experi-ment was repeated in a buffer at neutral pH, theseincreases of fluorescence in the presence of metals wereless important, indicating that 50 mM pyridin AcOH pH 5.0buffer is a suitable buffer for the detection of metal tracesby PvdI.

Competition between Fe3+ and other metals for PvdI

PvdI interacted with several metals in addition to iron(Fig. 2 and Table 1). Therefore, we investigated the abilityof PvdI to chelate its natural ligand Fe3+ in the presence ofan excess amount of another metal. PvdI at 50 mM wasincubated simultaneously with 1 equivalent of Fe3+ and 1or 10 equivalents of another metal for 1 h. PvdI–Fe com-plexes generate a particular absorbance pattern at450 nm (Table 1; Albrecht-Gary et al., 1994), which is notpresent for the PvdI–metal complexes studied in Table 1,except for PvdI–Sn. This absorbance is specific toPvdI–Fe and was thus used to quantify the amount ofPvdI–Fe present in each experimental conditiondescribed in Fig. 3 (the reference, 100%, was PvdI incu-bated in the presence of only Fe3+). Absorbance at450 nm was lower in the presence of Ga3+, indicating thatless PvdI–Fe was formed if Fe3+ is in competition with thismetal for PvdI. Ga3+ was the only metal able to competewith Fe3+ for PvdI. All the other metals interacted with PvdIwith lower affinities than Fe3+ and Ga3+.

Ability of the PvdI–metal complexes to inhibitPvdI–55Fe uptake

PvdI–Fe is transported into P. aeruginosa by the outermembrane transporter, FpvAI (Poole et al., 1993; Schalket al., 2001). This is the only process by which this ferric–siderophore complex can enter this bacteria (Schalket al., 2001). To determine if PvdI–metal complexes otherthan PvdI–Fe bind to the FpvAI binding site, we monitoredthe ability of various PvdI–metal complexes to inhibitPvdI–55Fe uptake in a Pvd- and Pch-deficient strain (Pchbeing pyochelin, another siderophore produced byP. aeruginosa). An inhibition of PvdI–55Fe uptake does notnecessarily imply an uptake of PvdI–metal in place ofPvdI–55Fe by FpvAI. However, it means at least that thePvdI–metal complex is able to take the place of PvdI–55Feon the transporter. Therefore, PvdI- and Pch-deficientcells (PAD07) were first incubated in the presence of100 nM PvdI–55Fe to evaluate the amount of 55Fe trans-

A

B

C

Fig. 2. UV (A and B) and fluorescent (C) spectra of PvdI incubatedin the presence of various metals. We incubated 50 mM PvdIovernight in the presence of 500 mM metals in 50 mM pyridin-AcOHpH 5.0, and monitored the absorbance. Metals that caused achange in the PvdI spectrum are represented in (A) and (B). Themaxima of absorbance are reported in Table 1. For thefluorescence spectra (C), the mixtures were diluted, at aconcentration of 5 mM PvdI and 50 mM metals. As a control, the UVand the fluorescence spectra were monitored for each metal inidentical buffer but in the absence of siderophore.

1082 A. Braud et al.

© 2008 The AuthorsJournal compilation © 2008 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 11, 1079–1091

ported in the absence of any competition (Fig. 4), during45 min. During this period, the cell concentration wasstable, no growth of P. aeruginosa occurred when 50 mMTris-HCl pH 8.0 buffer was used. The experiment wasrepeated with an excess of metal-free PvdI, as a control.Indeed, metal-free PvdI can bind FpvAI, but is not trans-ported and has no efficient inhibitory effect on PvdI–55Feuptake (Schalk et al., 2001). This was confirmed in thisstudy with a 500-fold excess of metal-free PvdI comparedwith PvdI–55Fe. We repeated the experiment in the pres-ence of an excess of each PvdI–metal complex. Eachcomplex was prepared by incubating the metal overnightwith a fourfold excess of PvdI.

A very low PvdI–55Fe uptake was observed only with500 equivalents of PvdI loaded with Ag+, Ga3+ or Hg2+.Under similar experimental conditions, Al3+, Zn2+, Mn2+,Cd2+ and Co2+ inhibited PvdI–55Fe uptake, displaying inhi-bition rates between 15% and 36%. In the presence of 50equivalents of PvdI–metal, only Ga3+, Hg2+, Co2+ and Cd2+

were able to inhibit PvdI–55Fe uptake, displaying inhibitionrates between 15% and 50%. To check that the observedPvdI–55Fe uptake inhibitions were not the result of celldeath due to metal toxicity, bacterial cells were plated onagar LB medium after incubation (45 min) in the presence

of PvdI–Fe and PvdI–metal complexes. None of themetals was toxic (data not shown) and the inhibition ofuptake (Fig. 4) was not due to cell death. Thus, inhibitionof PvdI–55Fe uptake was observed for Ag+, Al3+, Cd2+, Co2+,Ga3+, Hg2+, Mn2+ and Zn2+. These metals in complex withPvdI were able to bind to FpvAI at the cell surface.

PvdI production

The presence of metals other than iron is known to stimu-late siderophore production in a number of bacteria andfungi (Chakrabarty and Roy, 1964; Huyer and Page, 1988;Hofte et al., 1993). This property was investigated for allthe metals tested in our screening (Fig. 5). Pseudomonasaeruginosa PAO1 cells were grown in the presence andabsence of various metals (10 mM). The absorbanceswere monitored at 600 nm and at 400 nm to follow cellgrowth and PvdI production respectively. Metal concen-trations except for Ag+ and Hg3+ were not toxic for cells(Fig. 5A). The concentration of PvdI produced was esti-

Fig. 3. Competition between Fe3+ and different metals for PvdI.PvdI, FeCl3 and each tested metal, all at a concentration of100 mM, were incubated for 1 h in 50 mM pyridin-AcOH pH 5.0 andthe absorbance at 450 nm was monitored (dark grey bars). Theexperiments were also repeated in the presence of a 10-foldexcess of each tested metal, in relation to the PvdI and FeCl3concentrations (light grey bars). As a control, PvdI was alsoincubated without metal. To monitor the absorbance at 450 nmcorresponding to 100 mM PvdI–Fe (this absorbance at 450 nmrepresents 100%), PvdI was incubated in the presence of onlyFeCl3. All the data are the average of triplicate experiments.

Fig. 4. Ability of PvdI–metal complexes to inhibit PvdI–55Fe uptake.PvdI- and Pch-deficient PAD07 cells at an OD600 of 1 wereincubated for 45 min in the presence of 100 nM PvdI–55Fe at 37°C.We then filtered the mixtures and determined the radioactivitycounts. The amount of 55Fe transported in this experiment wasused as the reference (100% uptake). This experiment wasrepeated in the presence of an excess of each PvdI–metal complextested (dark grey and light grey bars correspond to concentrationsof 5 and 50 mM of the metals tested respectively), and in thepresence of an excess of metal-free PvdI. PvdI–metal complexeswere prepared by preincubating each metal in the presence of afourfold excess of PvdI overnight. As a control, the experiment wasrepeated for each metal with no cells (data not shown). All the dataare the average of three experiments.

Metal specificity of the FpvA/Pvd uptake system 1083

© 2008 The AuthorsJournal compilation © 2008 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 11, 1079–1091

mated in the extracellular medium using characteristic UVabsorbance of this siderophore at 400 nm, an approachcommonly used by scientists working in the field of PvdI(Hohnadel and Meyer, 1988; McMorran et al., 2001;Nader et al., 2007). In the absence of metal, the concen-tration of PvdI after an overnight culture of P. aeruginosa

PAO1 is about 120 mM (represents the 100% in Fig. 5B).Surprisingly, six metals increased PvdI production(Fig. 5B). Cu2+ and Mn2+ were more efficient than the othermetals, increasing PvdI production by 184% and 171%respectively. Zn2+, Ga3+, Al3+ and Ni2+ also increased PvdIproduction, but to a lower extent (120–140%). None of themetals except for Fe3+ decreased PvdI production.

Metals transported into P. aeruginosa cells in thepresence of PvdI

Our data have shown that PvdI chelates metals other thanFe3+, and that some of these siderophore–metal com-plexes are able to bind to FpvAI at the cell surface. Thus,we investigated which of these tested metals were trans-ported into the P. aeruginosa periplasm, by the PvdI/FpvAI system. We studied only the PvdI–metal complexesthat inhibit PvdI–55Fe uptake or induce PvdI production.Only these metals, if complexed with PvdI, may interactwith FpvAI. Bacteria were incubated in the presence ofthe PvdI–metal complex. After incubation, the cells wereharvested and metal levels inside the bacteria and/orbound at the cell surface were determined by ICP-AES(Inductively Coupled Plasma-Atomic Emission Spectrom-etry), a suitable spectrometric technique for determiningtrace amounts of elements. It is based on measuring lightemission from excited atoms and ions, in argon plasma.Each element is determined due to its specific emissionwavelength. To avoid any metal uptake by the Pchpathway, Pch- and Pvd-deficient PAD07 strains wereused for this experiment. PvdI–metal complexes wereprepared by incubating each metal with a fourfold excessof PvdI overnight. We carried out the experiment in thepresence and in the absence of the protonophore CCCP(carbonyl cyanide m-chlorophenylhydrazone) for tworeasons: first to discriminate between metals transportedin a TonB-dependent way from metals diffusing across theporines; and second to discriminate between metalsadsorbed on the cell surface to those transported insidethe cells. CCCP inhibits the proton-motive force of theinner membrane and therefore any TonB-dependentuptake in bacteria (Schalk et al., 2001; Clément et al.,2004). Under such experimental conditions, FpvAI is stillable to bind PvdI–Fe, but is no longer able to transport itinto the periplasm of P. aeruginosa (Schalk et al., 2001;Clément et al., 2004).

We observed differences in the uptake of Fe3+, Ga3+,Mn2+, Cu2+ and Ni2+ in the presence and absence ofCCCP (Fig. 6). Only these metals, if chelated by PvdI,were transported in a proton-motive force-dependentprocess. However, the uptake rates of Mn2+, Ni2+, Ga3+

and Cu2+ were not of the same order of magnitude asthat of Fe3+: 2.8 (mmol metal) g-1 of cells for Fe3+,0.4 (mmol metal) g-1 for Ga3+, 0.2 (mmol metal) g-1 for

A

B

Fig. 5. Effect of metals on P. aeruginosa PAO1 growth (A) andPvdI production (B). Cells were grown in succinate medium at29°C, in the presence or absence of each of the different metalstested (10 mM). The absorbances at 600 nm (A) and at 400 nm (B)were monitored after 12 h culture. The amount of PvdI producedwas determined for a similar amount of cells (OD600 of 1) for all theexperimental conditions studied. In both representations (A and B),the experiment without metal has been used as the reference. Thedata are the average of triplicate experiments.

1084 A. Braud et al.

© 2008 The AuthorsJournal compilation © 2008 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 11, 1079–1091

Mn2+, 0.076 (mmol metal) g-1 for Cu2+, and 0.066 (mmolmetal) g-1 for Ni2+. In the case of Al3+, Co2+ and Zn2+, theerror bars calculated for the concentrations preclude anyconclusions on their uptake. This experiment could notbe repeated with a FpvA- and Pch-deficient mutant todemonstrate clearly that FpvA is involved in the uptakeprocess observed for Ga3+, Mn2+, Cu2+ and Ni2+. Thismutant had difficulty growing in the succinate medium.Surprisingly, a large amount of Hg2+ [2.4 (mmol metal) g-1

of cells] and Ag+ [2.6 (mmol metal) g-1 of cells] weretransported into P. aeruginosa PAD07 cells or adsorbedon the cell surface. Siderophore outer membrane trans-porters do not transport these two metals, as CCCP hadno effect on these values. Thus, only Ga3+, Mn2+, Cu2+

and Ni2+ appeared to be transported by the PvdI/FpvAIsystem, but with efficiencies clearly lower than those forFe3+ uptake.

Toxicity of the metals for P. aeruginosa

We investigated whether the toxicity of a metal and itsability to be transported into P. aeruginosa via the PvdIpathway are related. Thus, PAO1 and DfpvAI cells weregrown in the presence of each metal (100 mM), in aniron-free medium. Hg2+, Ag+, Co2+ and Ga3+ were the most

toxic for both strains, inhibiting growth by 83%, 78%, 75%and 46% respectively. Ni2+, Cu2+, Zn2+ and Cd2+ were lesstoxic, with growth inhibition rates between 9% and 27%.The toxicity of all these metals was the same whetherFpvAI or FpvB were expressed or not. This indicates, forthe experimental conditions used here, that there is nocorrelation between the toxicity of the metal and its abilityto be transported into the cells by the PvdI/FpvAI system.For metals transported by the PvdI uptake pathway (Ni2+,Ga3+ and Cu2 in Fig. 6), the observed toxicity must be dueto metal diffusion through porines.

Discussion

The mechanisms of iron acquisition by siderophores andtheir corresponding outer membrane transporters havebeen the subject of several studies (Braun and Endriss,2007; Schalk, 2008). Nevertheless, no data on the metalspecificity of these transporters and whether they are ableto transport metals other than iron under certain circum-stances are currently available. In their environment,microorganisms are often in the presence of elevatedconcentrations of heavy metals and it is ecologicallyimportant to understand how bacterial iron uptake path-ways interact with other heavy metals. The aim of thisstudy was to determine if the PvdI uptake pathway inP. aeruginosa can transport metals other than iron, evenat very low concentrations, and to investigate the molecu-lar mechanisms that regulate metal specificity in this sid-erophore uptake pathway.

Specificity of the siderophore PvdI

PvdI fluorescence allowed us to show that this sidero-phore can chelate all of the 16 heavy metals tested(Table 1, Fig. 2). The fluorescence was either increasedor quenched for all metals; thus, this siderophore may beused as an efficient probe in 50 mM pyridin AcOH pH 5.0buffer for heavy metal detection. Indeed, fluorescence asa result of binding Al3+, Cd2+, Hg2+ and Pb2+ was 20-foldgreater than that observed with metal-free PvdI.

The competition experiments between PvdI, stoichio-metric concentrations of Fe3+, various other metals inexcess (Fig. 3) demonstrate that PvdI has been appropri-ately identified as a ‘siderophore’. Only Ga3+, a metal withsimilar coordination and size as Fe3+, was able to competeefficiently with iron for PvdI. PvdI affinity for Ga3+ must beequivalent to that for Fe3+ however, the PvdI affinity for allthe other metals is clearly lower, as these other metals areunable to dissociate PvdI–Fe3+. Therefore, even if a metalpollutes an environment, PvdI will always efficientlychelate Fe3+, even though it is present in much lowerconcentrations.

Fig. 6. Pmf-dependent metal incorporated in P. aeruginosa cellsby the PvdI uptake pathway. Pvd- and Pch-deficient PAD07 at anOD600 of 1 were incubated at 37°C for 45 min, in the presence of5 mM of each PvdI–metal complex, and in the presence (light greybars) or absence (dark grey bars) of 200 mM CCCP. The cells wereharvested, washed, and the metal content was determined byICP-AES. PvdI–metal complexes were prepared by incubating eachmetal in the presence of a fourfold excess of PvdI overnight. Thedata are the average of three experiments.

Metal specificity of the FpvA/Pvd uptake system 1085

© 2008 The AuthorsJournal compilation © 2008 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 11, 1079–1091

Specificity of the FpvAI binding site

In addition to iron, PvdI chelates several other metals withcertainly different affinities; thus, a logical question is howFpvAI, the PvdI–Fe outer membrane transporter, pro-cesses PvdI in complex with metals other than iron at thecell surface. Eight metals (Ag+, Al3+, Cd+, Co2+, Ga3+, Hg2+,Mn2+ and Zn2+) chelated by PvdI inhibited PvdI–55Feuptake in a Pch- and PvdI-deficient strain (Fig. 4). Thesemetals in complex with PvdI compete with PvdI–55Fe forthe FpvAI binding site at the bacterial cell surface. Thisexperiment was carried out with an excess of PvdI (four-fold excess of PvdI in relation to total metal concentration,including 55Fe and the tested metal), and thus thereshould have been no competition for PvdI between 55Feand the other metals.

The presence of metals other than iron is known tostimulate siderophore production in a number of bacteriaand fungi (Huyer and Page, 1988; Hofte et al., 1993; Huand Boyer, 1996). Our data have shown that the presenceof Al3+, Cu2+, Ga3+, Mn2+, Ni2+ and Zn2+ in the culturemedium induced PvdI production (Fig. 5B). One explana-tion may be that the metal in complex with PvdI, likePvdI–Fe after binding to FpvAI (Visca et al., 2002), acti-vates the signalling cascade regulating siderophore pro-duction. Alternatively, in the presence of other metals, thefree siderophore concentration in the medium will bereduced as a result of complex formation. This decreasein the PvdI concentration may be sufficient to cause addi-tional siderophore secretion into the medium. However,the concentration of PvdI after an overnight culture isabout 120 mM under the experimental conditions in thisstudy, a concentration much greater than the 10 mM ofmetal added to the culture. Moreover, all the metals thatincrease PvdI production either inhibit PvdI–55Fe uptake(Fig. 4) or are transported (Fig. 6). Thus, the stimulation ofPvdI production as a result of the presence of Al3+, Cu2+,Ga3+, Mn2+, Ni2+ and Zn2+ must be due to the signallingcascade being activated by the PvdI–metal complexbinding to FpvAI.

The FpvAI binding site binds PvdI in complex withAg+, Al3+, Cd+, Cu2+, Co2+, Ga3+, Hg2+, Mn2+, Ni2+ or Zn2+,if we consider the metals that inhibit 55Fe uptake andinduce PvdI production. This indicates that the FpvAI–siderophore binding site has broad metal specificity, whichappears to contradict the high siderophore specificity.Indeed, pseudomonads produce at least 100 sidero-phores with different peptide sequences. Each Pvd isrecognized at the level of the outer membrane by a spe-cific transporter, indicating a high specificity of this familyof transporters for the corresponding siderophore(Hohnadel and Meyer, 1988; Cornelis et al., 1989; Meyeret al., 1999; de Chial et al., 2003). According to our data,in the mechanism for PvdI–metal recognition by FpvAI,

the peptide section of the siderophore appears to be moreimportant than the nature of the metal chelated or theconformation of the PvdI–metal complex. This alsoexplains why FpvAI is able to interact with metal-freePvdI (Schalk et al., 1999; 2001; Clément et al., 2004).However, if loaded with Cr2+, Eu3+, Pb2+, Sn2+, Tb3+ and Tl+,the PvdI–metal complexes must either have had no affin-ity for FpvAI or an affinity too low to detect any PvdI–55Feuptake inhibition or increase in PvdI production. Anotherpossibility is that these metals are chelated by PvdI withless efficacy (only a small amount of PvdI is present asthe PvdI–metal form).

Metal specificity of the FpvAI uptake process and theactivation of the signalling cascade by the transporter

It is not surprising that some metals in complex with sid-erophores increase siderophore synthesis and its releaseinto the environment, as it is something that is beneficialto the bacteria. In the presence of another metal than iron,a certain amount of PvdI is already chelated to this metal;thus, producing greater quantities of siderophore helpsthe bacteria to access iron despite the presence of analternate metal interacting with the siderophore. More-over, siderophore production is probably used in situ bythe bacteria as a protective mechanism against heavymetal toxicity: siderophore–metal complexes being lesstoxic than free metal if not transported into the bacteria bythe siderophore uptake pathways. Therefore, the uptakeprocess by FpvAI must likely be more selective thansolely due to the recognition of the PvdI–metal complexby the binding site. TonB-dependent uptake was only effi-cient for Fe3+, among the 10 metals that interact withFpvAI. Cu2+, Ga3+, Mn2+ and Ni2+ were also imported intothe cell but with lower uptake rates: TonB-dependentuptake rates was seven-fold lower for PvdI–Ga, 14-foldlower for PvdI–Mn, 37-fold lower for PvdI–Cu and 42-foldlower PvdI–Ni. We were unable to draw conclusions onuptake from data for Al3+, Co2+ and Zn2+ (Fig. 6). TheTonB-dependent uptake pathway used by the metalsmust be PvdI/FpvAI, as the strain used was PvdI- andPch-deficient and the metals were pre-incubated in thepresence of a large excess of PvdI. All these resultsclearly show that the FpvAI binding site has broad metalspecificity and the selectivity occurs at the level of theuptake process. Interestingly, the metals transported werethose that induced greater PvdI production and did notinclude those with the highest inhibition rates for PvdI–55Fe uptake.

Structures of FpvAI in various loading states haveallowed the development of a sequential mechanismlinking transport and signalling (Brillet et al., 2007). In thismechanism, the starting point for both iron uptake andPvdI production (signalling cascade) is the binding of

1086 A. Braud et al.

© 2008 The AuthorsJournal compilation © 2008 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 11, 1079–1091

PvdI–Fe to FpvAI. A conformational change is then trans-mitted along FpvAI, from the binding site to the periplas-mic N-terminal region altering the interaction between theFpvAI signalling domain and the TonB box. In this newconformation, TonB is able to interact with the TonB box,displacing the signalling domain. The FpvAI–TonB inter-action allows the energy transduction necessary formovement of the plug domain (formation of a channel)and translocation of PvdI–Fe into the periplasm. Displace-ment of the signalling domain from the TonB box allows itto interact with FpvR, leading to gene expression. Metalsin complex with PvdI that bind to FpvAI can be divided intothree groups (Fig. 7). The first group is composed of Fe3+,Ga3+ and Mn2+. PvdI–Ga and PvdI–Mn were transported

by FpvAI, inhibited PvdI–55Fe uptake and induced PvdIproduction; thus, both complexes, like PvdI–Fe, mustinduce adequate conformational change in FpvAI foruptake and induction of PvdI production to occur. Al3+ andZn2+ probably also belong to this family, but we wereunable to draw conclusions on uptake from the data(Fig. 6). The second metal group consists of Cu2+ andNi2+. PvdI–Cu and PvdI–Ni were transported, inducedPvdI production, but were unable to inhibit PvdI–55Feuptake. Binding of PvdI–Ni and PvdI–Cu to FpvAI mustinduce similar conformational change on the transporteras PvdI–Fe, PvdI–Ga and PvdI–Mn; however, their affini-ties for FpvAI must be lower and thus uptake of thesecomplexes cannot compete with that of PvdI–55Fe. In the

Fig. 7. Metal specificity of the FpvAI/PvdI system in P. aeruginosa. PvdI when secreted into the extracellular medium chelates Fe3+, Ag+, Al3+,Cd2+, Co2+, Cr2+, Cu2+, Eu3+, Ga3+, Hg2+, Mn2+, Ni2+, Pb2+, Sn2+, Tb3+, Tl+ and Zn2+. FpvAI at the cell surface binds PvdI in complex with Fe3+, Ag+,Al3+, Cd2+, Co2+, Ga3+, Hg2+, Mn2+ and Zn2+. The presence of Al3+, Cu2+, Ga3+, Mn2+, Ni2+ and Zn2+ in the extracellular medium induces PvdIproduction by may be activating the signalling cascade: PvdI in complex with these metals binds to FpvAI and induces the activation of thesignalling cascade. TonB-dependent metal uptake, in the presence of PvdI, was only efficient for Fe3+ Cu2+, Ga3+, Mn2+ and Ni2+ were alsotransported into the cell, but with a 5- to 40-fold lower efficiency. The metals interacting with FpvAI could be divided into three groups: (i) Ga3+

and Mn2+, which induced production of PvdI, inhibited PvdI–55Fe uptake and were transported, but with much lower efficiencies than Fe3+ (ii)Cu2+ and Ni2+, which induced production of PvdI, were also slightly transported, but were not able to inhibit PvdI–55Fe uptake; and (iii) Ag+,Cd2+, Co2+ and Hg3+, which only inhibited PvdI–55Fe uptake. All these data indicate that FpvAI at the cell surface has broad metal specificity atthe binding stage and that it is highly selective only for the metal uptake process.

Metal specificity of the FpvA/Pvd uptake system 1087

© 2008 The AuthorsJournal compilation © 2008 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 11, 1079–1091

third group, PvdI–Ag, PvdI–Cd, PvdI–Co and PvdI–Hgbind to the FpvAI binding site, but the uptake process(formation of a channel in FpvAI) and activation of FpvR(signalling cascade for PvdI production) are apparently notinitiated. These FpvAI–PvdI–metal complexes may have aconformation that is unable to start these two biologicalprocesses. Cr2+, Eu3+, Pb2+, Sn2+, Tb3+ and Tl+ in complexwith PvdI are unable to interact with FpvAI, at least notunder the experimental conditions used in this study.

In conclusion, our data have shown that the PvdI uptakepathway becomes progressively more metal restrictive asthe siderophore penetrates deeper into the transportsystem. PvdI chelates several metals and may be a veryefficient probe for detecting heavy metals in a medium(nanomolar concentrations of metal can be detectedusing the fluorescent properties of PvdI). Surprisingly,FpvAI displayed broad PvdI–metal specificity, indicatingthat FpvAI transporters at the cell surface may be pollutedby undesirable PvdI–metal complexes; these complexesmay inhibit PvdI–Fe uptake and/or induce PvdI produc-tion. Despite the broad metal binding specificity of theFpvAI binding site, only Fe3+ was transported efficiently bythis transporter into the periplasm. Ga3+, Cu2+, Mn2+ andNi2+ were transported with lower uptake rates. This pre-liminary study investigates whether FpvAI recognizes andbinds metals other than iron in complex with PvdI. Furtherstudies are now underway to understand better themechanism of interaction of all these partners, and howPvdI–metal uptake by FpvAI may be more selective thanPvdI–metal binding.

Experimental procedures

Chemicals

PvdI was prepared as described previously (Schalk et al.,1999). AgNO3 (Sigma), AlCl3 (Sigma), CdCl2.2.5H2O(Probalo), CoCl2.6H20 (Strem), CrCl2 (Strem), CuCl2.2H2O(Strem), EuCl3 (Aldrich), FeCl3 (Prolabo), Ga(NO3)3.24H2O(Strem), HgCl2 (Sigma), MnSO4 (Prolabo), NiCl2.4H2O(Strem), Pb(NO3)2 (Aldrich), SnCl2 (Prolabo), TbCl3.6H2O(Strem), ZnCl2 (Alfa Aesar) and Tl2SO4 (Fluka) are the metalsused in this study. All the metals were prepared at a concen-tration of 50 mM in 0.5 N HCl or 0.5 N HNO3.

Bacterial strains and growth media

Wild-type P. aeruginosa strain PAO1 (Royle et al., 1981) andthree mutants were used. All the three mutants have beendescribed previously: the Pch- and PvdI-deficient P. aerugi-nosa strain, PAD07 (Takase et al., 2000), the Pch-deficientmutant, PAO6297 (Serino et al., 1995) and the FpvA-deficientmutant K691 (Schalk et al., 1999). All strains were grownovernight in succinate medium (Demange et al., 1990) in thepresence of 50 mg ml-1 tetracycline for K691, and in addition,100 mg ml-1 streptomycin for strain PAD07.

Fluorescence and UV analyses of PvdI–metalcomplexes

Fluorescence experiments were performed with a PTI(Photon Technology International TimeMaster, Bioritech)spectrofluorometer. For all experiments, the samples wereexcited at 400 nm. The UV measurements were performedon a Uvikon spectrophotometer 930 (Serlabo Technologies).To follow metal complexation by PvdI, 50 mM PvdI was incu-bated overnight at room temperature in the presence of 10equivalents of metal in 1 ml of 50 mM pyridin-AcOH pH 5.0buffer. For the fluorescent measurements, the mixture wasdiluted 10-fold in the same buffer.

Competition between Fe3+ and other metals forPvd is monitored by UV

PvdI was prepared at a concentration of 100 mM in 50 mMpyridin-AcOH pH 5.0 buffer. All the metals were prepared at5 mM or 50 mM in 0.5 N HCl or 0.5 N HNO3. For each metalto be tested, 960 ml of the PvdI solution was added to 20 ml ofFe3+ (5 mM) and 20 ml of metal (5 or 50 mM). The final con-centration of PvdI and Fe3+ in the experiment was 100 mMand that of the tested metal, 100 mM or 1 mM. We incubatedthe mixtures overnight at room temperature and monitoredthe absorbance at 450 nm. Two controls were performed withno metal and only Fe3+ to monitor the absorbance at 450 nmof 100 mM PvdI–Fe.

55Fe uptake in P. aeruginosa

Iron uptake assays with the various PvdI–metal complexeswere carried out with the PvdI- and Pch-deficient strainPAD07 and with the filtration assay previously described(Schalk et al., 2001). After overnight growth in iron-limitedmedium, bacteria were prepared at an OD600 of 1 in 50 mMTris-HCl (pH 8.0) and were incubated in the presence orabsence of 5 mM or 50 mM PvdI–metal at 37°C. Transportassays were initiated by adding 100 nM of 55Fe in complexwith PvdI. 55FeCl3 was isotopically diluted with 10 volumes ofnon-radioactive FeCl3 and PvdI–metal or PvdI–55Fe was pre-pared using a ratio of 4:1 for PvdI : metal. After a 45 minincubation, the mixtures were filtered and the retained radio-activity was counted. As controls, the experiment wasrepeated in the presence of (i) cell culture alone, PvdI–55Feand unloaded PvdI, or (ii) in the presence of cell culture andPvdI–55Fe (used as the reference -100% of 55Fe transported– in Fig. 4) or (iii) PvdI–55Fe alone (without cells).

Pvd production

Pseudomonas aeruginosa PAO1 cells were grown in thepresence or absence of metals at 10 mM. The absorbances at600 and 400 nm were monitored after 12 h of culture todetermine the amount of cells and the levels of PvdI pro-duced respectively (Nader et al., 2007). The amount of PvdIproduced was calculated for an OD600 of 1, for all the experi-mental conditions studied.

1088 A. Braud et al.

© 2008 The AuthorsJournal compilation © 2008 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 11, 1079–1091

Toxicity of PvdI–metal

Pseudomonas aeruginosa cells PAO1 and K691 were grownovernight (12 h) at 29°C in the presence of metals at a con-centration of 100 mM. The absorbance was monitored at600 nm.

Metal incorporation in PAD07 cells

An overnight culture of PAO6297, grown in iron-limitedmedium, was centrifuged. The supernatant containing PvdIwas filtered through a 0.22 mM nitrocellulose membraneand the final PvdI concentration was determined at OD400

(e = 19 000 M-1 cm-1). We adjusted the PvdI concentration to25 mM and incubated the solution overnight in the presenceof metals (6.25 mM) in a final volume of 8 ml of succinatemedium (solution A).

An overnight culture of PAD07, grown in iron-limitedmedium, was pelleted and re-suspended at an OD600 of 5 insuccinate medium. Half of the PAD07 bacterial suspensionwas incubated for 15 min at 4°C in the presence of 200 mMCCCP (20 mM in ethanol). The bacterial cells with or withoutCCCP were added to the PvdI solutions pre-incubated withmetals (solution A) at a final concentration of OD600 of 1 andin a final volume of 10 ml (final concentration of metal: 5 mM).The tubes were then incubated under shaking at 200 r.p.m.for 45 min at 37°C. Afterwards, the samples were centrifuged,and the bacterial pellets were washed once with ultra-purewater and dried at 50°C for 48 h in glass tubes (pre-washedwith HNO3 20%). Cells were mineralized in 68% (v/v) HNO3

for 24 h at room temperature. The volume was brought to10 ml with ultra-pure water, and the samples were filtered ona 0.45 mm membrane. We analysed the samples with anICP-AES apparatus (Jobin-Yvon, Ultima) at the followingwavelengths (in nm): Ag+ (328.068), Al3+ (396.152), Cd2+

(214.438), Co2+ (238.892), Cu2+ (324.754), Fe3+ (259.94),Ga3+ (294.364), Hg2+ (194.163), Mn2+ (257.61), Ni+ (231.604),Zn2+ (213.856). As a control, the amount of metals weredetermined in succinate medium and gave the following data:0.12 mM for Ag+, 2.36 mM for Al3+, 0.05 mM for Cd2+, 0.13 mMfor Co2+, 0.68 mM for Cu2+, 0.21 mM for Fe3+, 0.10 mM for Ga3+,0.11 mM for Hg2+, 0.02 mM for Mn2+, 0.03 mM for Ni2+ and0.07 mM for Zn2+.

Acknowledgements

This work was partly funded by the Centre National de laRecherche Scientifique, a grant from the ANR (Agence Natio-nale de Recherche, ANR-05-JCJC-0181–01) and by the Pro-gramme de Cooperation Franco-Libanais CEDRE. We thankProf Cornelia Reinmann for the gift of strain PAO6297.

References

Albrecht-Gary, A.M., and Crumbliss, A.L. (1998) Coordina-tion chemistry of siderophores: thermodynamics and kinet-ics of iron chelation and release. Met Ions Biol Syst 35:239–327.

Albrecht-Gary, A.M., Blanc, S., Rochel, N., Ocacktan, A.Z.,and Abdallah, M.A. (1994) Bacterial iron transport: coordi-

nation properties of pyoverdin PaA, a peptidic siderophoreof Pseudomonas aeruginosa. Inorg Chem 33: 6391–6402.

Baysse, C., De Vos, D., Naudet, Y., Vandermonde, A.,Ochsner, U., Meyer, J.M., et al. (2000) Vanadium interfereswith siderophore-mediated iron uptake in Pseudomonasaeruginosa. Microbiology 146 (Part 10): 2425–2434.

Beare, P.A., For, R.J., Martin, L.W., and Lamont, I.L. (2003)Siderophore-mediated cell signalling in Pseudomonasaeruginosa: divergent pathways regulate virulence factorproduction and siderophore receptor synthesis. Mol Micro-biol 47: 195–207.

Boukhalfa, H., and Crumbliss, A.L. (2002) Chemical aspectsof siderophore mediated iron transport. Biometals 15: 325–339.

Braun, V. (2003) Iron uptake by Escherichia coli. Front Biosci8: 1409–1421.

Braun, V., and Endriss, F. (2007) Energy-coupled outermembrane transport proteins and regulatory proteins. Bio-metals 20: 219–231.

Brillet, K., Journet, L., Celia, H., Paulus, L., Stahl, A., Pattus,F., and Cobessi, D. (2007) A b-strand lock-exchange forsignal transduction in TonB-dependent transducers on thebasis of a common structural motif. Structure 15: 1383–1391.

Budzikiewicz, H. (1997) Siderophores of fluorescentpseudomonads. Z Naturforsch [C] 52: 713–720.

Carrano, C.J., Drechsel, H., Kaiser, D., Jung, G., Matzanke,B., Winkelmann, G., et al. (1996) Coordination chemistry ofthe carboxylate type siderophore rhizoferrin: the Iron(III)complex and its metal analogs. Inorg Chem 35: 6429–6436.

Chakrabarty, A.M., and Roy, S.C. (1964) Effect of trace ele-ments on the production of pigments by a pseudomonad.Biochem J 93: 228–231.

Chen, Y., Jurkewitch, E., Bar-Ness, E., and Hadar, Y. (1994)Stability constants of pseudobactin complexes with transi-tion metals. Soil Sci Soc Am J 58: 390–396.

de Chial, M., Ghysels, B., Beatson, S.A., Geoffroy, V., Meyer,J.M., Pattery, T., et al. (2003) Identification of type II andtype III pyoverdine receptors from Pseudomonas aerugi-nosa. Microbiology 149: 821–831.

Clément, E., Mesini, P.J., Pattus, F., Abdallah, M.A., andSchalk, I.J. (2004) The binding mechanism of pyoverdinwith the outer membrane receptor FpvA in Pseudomonasaeruginosa is dependent on its iron-loaded status. Bio-chemistry 43: 7954–7965.

Cobessi, D., Célia, H., Folschweiller, N., Schalk, I.J., Abdal-lah, M.A., and Pattus, F. (2005) The crystal structure of thepyoverdin outer membrane receptor FpvA from Pseudomo-nas aeruginosa at 3.6 Å resolution. J Mol Biol 34: 121–134.

Cornelis, P., Hohnadel, D., and Meyer, J.M. (1989) Evidencefor different pyoverdine-mediated iron uptake systemsamong Pseudomonas aeruginosa strains. Infect Immun57: 3491–3497.

De Vos, D., De Chial, M., Cochez, C., Jansen, S., Tummler,B., Meyer, J.M., and Cornelis, P. (2001) Study of pyover-dine type and production by Pseudomonas aeruginosaisolated from cystic fibrosis patients: prevalence of type IIpyoverdine isolates and accumulation of pyoverdine-negative mutations. Arch Microbiol 175: 384–388.

Demange, P., Wendenbaum, S., Linget, C., Mertz, C., Cung,

Metal specificity of the FpvA/Pvd uptake system 1089

© 2008 The AuthorsJournal compilation © 2008 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 11, 1079–1091

M.T., Dell, A., and Abdallah, M.A. (1990) Bacterial sidero-phores: structure and NMR assigment of pyoverdins PaA,siderophores of Pseudomonas aeruginosa ATCC 15692.Biol Metals 3: 155–170.

Dhungana, S., Ratledge, C., and Crumbliss, A.L. (2004) Ironchelation properties of an extracellular siderophore exoch-elin MS. Inorg Chem 43: 6274–6283.

Evers, A., Hancock, R., Martell, A., and Motekaitis, R. (1989)Metal ion recognition in ligands with negatively chargedoxygen donor groups. Complexation of Fe(III), Ga(III),In(III), Al(III) and other highly charged metal ions. InorgChem 28: 2189–2195.

Folschweiller, N., Gallay, J., Vincent, M., Abdallah, M.A.,Pattus, F., and Schalk, I.J. (2002) The interaction betweenpyoverdin and its outer membrane receptor in Pseudomo-nas aeruginosa leads to different conformers: a time-resolved fluorescence study. Biochemistry 41: 14591–14601.

Ghysels, B., Dieu, B.T., Beatson, S.A., Pirnay, J.P., Ochsner,U.A., Vasil, M.L., and Cornelis, P. (2004) FpvB, an alterna-tive type I ferripyoverdine receptor of Pseudomonasaeruginosa. Microbiology 150: 1671–1680.

Greenwald, J., Hoegy, F., Nader, M., Journet, L., Mislin,G.L.A., Graumann, P.L., and Schalk, I.J. (2006) Real-timeFRET visualization of ferric-pyoverdine uptake inPseudomonas aeruginosa: a role for ferrous iron. J BiolChem 282: 2987–2995.

Greenwald, J., Zeder-Lutz, G., Hagege, A., Celia, H., andPattus, F. (2008) The metal dependence of Pyoverdineinteractions with its outer membrane receptor FpvA.J Bacteriol 190: 6548–6558.

Hernlem, B.J., Vane, L.M., and Sayles, G.D. (1996) Stabilityconstants for complexes of the siderophore desferrioxam-ine B with selected heavy metal cations. Inorg Chim Acta244: 179–184.

Hersman, L., Lloyd, T., and Sposito, G. (1995) Siderophore-promoted dissolution of hematite. Geochim CosmochimActa 59: 3327–3330.

Hofte, M., Buysens, S., Koedam, N., and Cornelis, P. (1993)Zinc affects siderophore-mediated high affinity iron uptakesystems in the rhizosphere Pseudomonas aeruginosa7NSK2. Biometals 6: 85–91.

Hohnadel, D., and Meyer, J.M. (1988) Specificity ofpyoverdine-mediated iron uptake among fluorescentPseudomonas strains. J Bacteriol 170: 4865–4873.

Hu, X., and Boyer, G.L. (1996) Siderophore-mediated alumi-num uptake by Bacillus megaterium ATCC 19213. ApplEnviron Microbiol 62: 4044–4048.

Huyer, M., and Page, W.J. (1988) Zn Increases siderophoreproduction in Azotobacter vinelandii. Appl Environ Micro-biol 54: 2625–2631.

Lamont, I.L., Beare, P.A., Ochsner, U., Vasil, A.I., and Vasil,M.L. (2002) Siderophore-mediated signaling regulatesvirulence factor production in Pseudomonas aeruginosa.Proc Natl Acad Sci USA 99: 7072–7077.

McMorran, B.J., Kumara, H.M., Sullivan, K., and Lamont, I.L.(2001) Involvement of a transformylase enzyme in sidero-phore synthesis in Pseudomonas aeruginosa. Microbiol-ogy 147: 1517–1524.

Meyer, J.M., Stintzi, A., De Vos, D., Cornelis, P., Tappe, R.,Taraz, K., and Budzikiewicz, H. (1997) Use of siderophores

to type pseudomonads: the three Pseudomonas aerugi-nosa pyoverdine systems. Microbiology 143 (Part 1):35–43.

Meyer, J.M., Stintzi, A., and Poole, K. (1999) The ferripyover-dine receptor FpvA of Pseudomonas aeruginosa PAO1recognizes the ferripyoverdines of P. aeruginosa PAO1 andP. fluorescens ATCC 13525. FEMS Microbiol Lett 170:145–150.

Meyer, J.M., Geoffroy, V.A., Baida, N., Gardan, L., Izard, D.,Lemanceau, P., et al. (2002) Siderophore typing, a powerfultool for the identification of fluorescent and nonfluorescentpseudomonads. Appl Environ Microbiol 68: 2745–2753.

Moll, H., Johnsson, A., Schafer, M., Pedersen, K., Budzik-iewicz, H., and Bernhard, G. (2008) Curium(III) complex-ation with pyoverdins secreted by a groundwater strain ofPseudomonas fluorescens. Biometals 21: 219–228.

Nader, M., Dobbelaere, W., Vincent, M., Journet, L., Adams,C., Cobessi, D., et al. (2007) Identification of residues ofFpvA involved in the different steps of Pvd-Fe uptake inPseudomonas aeruginosa. Biochemistry 46: 11707–11717.

Neu, M.P., Matonic, J.H., Ruggiero, C.E., and Scott, B.L.(2000) Structural characterization of a plutonium(IV)siderophore complex: single-crystal structure ofPu-desferrioxamine E. Angew Chem Int Ed Engl 39: 1442–1444.

Neubauer, U., Nowack, B., Furrer, G., and Schulin, R. (2000)Heavy metal sorption on clay minerals affected by thesiderophore desferrioxamine B. Environ Sci Technol 34:2749–2755.

Palanche, T., Marmolle, F., Abdallah, M.A., Shanzer, A., andAlbrecht-Gary, A.M. (1999) Fluorescent siderophore-based chemosensors: iron(III) quantitative determinations.J Biol Inorg Chem 4: 188–198.

Palanche, T., Blanc, S., Hennard, C., Abdallah, M.A., andAlbrecht-Gary, A.M. (2004) Bacterial iron transport: coor-dination properties of azotobactin, the highly fluorescentsiderophore of Azotobacter vinelandii. Inorg Chem 43:1137–1152.

Poole, K., Neshat, S., Krebes, K., and Heinrichs, D.E. (1993)Cloning and nucleotide sequence analysis of the ferripy-overdine receptor gene fpvA of Pseudomonas aeruginosa.J Bacteriol 175: 4597–4604.

Raymond, K.N., Dertz, E.A., and Kim, S.S. (2003) Enterobac-tin: an archetype for microbial iron transport. Proc NatlAcad Sci USA 100: 3584–3588.

Redly, G.A., and Poole, K. (2003) Pyoverdine-mediated regu-lation of FpvA synthesis in Pseudomonas aeruginosa:involvement of a probable extracytoplasmic-function sigmafactor, FpvI. J Bacteriol 185: 1261–1265.

Redly, G.A., and Poole, K. (2005) FpvIR control of fpvA ferricpyoverdine receptor gene expression in Pseudomonasaeruginosa: demonstration of an interaction between FpvIand FpvR and identification of mutations in each compro-mising this interaction. J Bacteriol 187: 5648–5657.

Royle, P.L., Matsumoto, H., and Holloway, B.W. (1981)Genetic circularity of the Pseudomonas aeruginosa PAOchromosome. J Bacteriol 145: 145–155.

Schalk, I.J. (2008) Metal trafficking via siderophores inGram-negative bacteria: specificities and characteristicsof the pyoverdine pathway. J Inorg Biochemi 102: 1159–1169.

1090 A. Braud et al.

© 2008 The AuthorsJournal compilation © 2008 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 11, 1079–1091

Schalk, I.J., Kyslik, P., Prome, D., van Dorsselaer, A.,Poole, K., Abdallah, M.A., and Pattus, F. (1999) Copuri-fication of the FpvA ferric pyoverdin receptor ofPseudomonas aeruginosa with its iron-free ligand: impli-cations for siderophore-mediated iron transport. Biochem-istry 38: 9357–9365.

Schalk, I.J., Hennard, C., Dugave, C., Poole, K., Abdallah,M.A., and Pattus, F. (2001) Iron-free pyoverdin binds to itsouter membrane receptor FpvA in Pseudomonas aerugi-nosa: a new mechanism for membrane iron transport. MolMicrobiol 39: 351–360.

Schons, V., Atkinson, R.A., Dugave, C., Graff, R., Mislin,G.L., Rochet, L., et al. (2005) The structure-activity rela-tionship of ferric pyoverdine bound to its outer membranetransporter: implications for the mechanism of iron uptake.Biochemistry 44: 14069–14079.

Serino, L., Reimmann, C., Baur, H., Beyeler, M., Visca, P.,and Haas, D. (1995) Structural genes for salicylatebiosynthesis from chorismate in Pseudomonas aerugi-nosa. Mol Gen Genet 249: 217–228.

Shen, J., Meldrum, A., and Poole, K. (2002) FpvA receptorinvolvement in pyoverdine biosynthesis in Pseudomonasaeruginosa. J Bacteriol 184: 3268–3275.

Takase, H., Nitanai, H., Hoshino, K., and Otani, T. (2000)Impact of siderophore production on Pseudomonas aerugi-nosa infections in immunosuppressed mice. Infect Immun68: 1834–1839.

Visca, P., Leoni, L., Wilson, M.J., and Lamont, I.L. (2002) Irontransport and regulation, cell signalling and genomics:lessons from Escherichia coli and Pseudomonas. MolMicrobiol 45: 1177–1190.

Wilson, M.J., and Lamont, I.L. (2000) Characterization of anECF sigma factor protein from Pseudomonas aeruginosa.Biochem Biophys Res Commun 273: 578–583.

Wilson, M.J., McMorran, B.J., and Lamont, I.L. (2001) Analy-sis of promoters recognized by PvdS, an extracytoplasmic-function sigma factor protein from Pseudomonasaeruginosa. J Bacteriol 183: 2151–2155.

Winkelmann, G. (2002) Microbial siderophore-mediatedtransport. Biochem Soc Trans 30: 691–696.

Metal specificity of the FpvA/Pvd uptake system 1091

© 2008 The AuthorsJournal compilation © 2008 Society for Applied Microbiology and Blackwell Publishing Ltd, Environmental Microbiology, 11, 1079–1091