Involvement of enterobactin in norepinephrine-mediated iron supply from transferrin to enterohaemorrhagic Escherichia coli

FEMS Microbiology Letters 222 (2003) 39^43

www.fems-microbiology.org

Involvement of enterobactin in norepinephrine-mediated iron supply from transferrin to enterohaemorrhagic Escherichia coli Primrose P.E. Freestone a , Richard D. Haigh a , Peter H. Williams a , Mark Lyte b

b;c;


a Department of Microbiology and Immunology, University of Leicester, Leicester, UK Department of Surgery, Minneapolis Medical Research Foundation, 914 South 8th Street, D3, Minneapolis, MN 55404, USA c Department of Biological Sciences, Minnesota State University at Mankato, Mankato, MN 56001, USA

Received 4 October 2002; accepted 11 February 2003 First published online 11 April 2003

Abstract Exposure of bacteria to members of the stress-associated family of catecholamine hormones, principally norepinephrine, has been demonstrated to increase both growth and production of virulence-related factors. Mutation of genes for enterobactin synthesis and uptake revealed an absolute requirement for enterobactin in norepinephrine-stimulated growth of Escherichia coli O157:H7. The autoinducer produced by norepinephrine-stimulated E. coli could not substitute for enterobactin. We also demonstrate that norepinephrine promotes iron shuttling between transferrin molecules, thereby enabling the bacterial siderophore enterobactin to more readily acquire iron for growth. These results suggest one of the possible mechanisms by which the hormonal output of stress may affect enterohaemorrhagic E. coli pathogenicity. 2 2003 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Norepinephrine ; Enterobactin ; Escherichia coli O157:H7; Iron; Autoinducer

1. Introduction The catecholamine stress hormone norepinephrine (noradrenaline, NE) stimulates the growth of a range of clinically important bacteria in a serum-containing minimal salts medium designed to mimic the stressful conditions pathogenic organisms are likely to encounter during infection [1^7]. NE also stimulates the in vitro expression of virulence determinants such as the adhesive pili of a strain of enterotoxigenic Escherichia coli [8], and the shiga toxins of enterohaemorrhagic E. coli (EHEC) [9]. NE £uxes during physiological stress, such as trauma, have been proposed as a means by which bacteria may sense their environment and initiate pathogenic processes [10]. For example, in an in vivo model of trauma, the release of NE from neurotoxin-damaged nerves within the body as a whole and particularly the adrenergic nerves comprising

* Corresponding author. Tel. : +1 (612) 347 6815; Fax : +1 (612) 347 5169. E-mail address : [email protected] (M. Lyte).

the enteric nervous system resulted in the overgrowth of E. coli within the caecum which returned to pre-trauma levels following neuronal repair [11]. Recent work aimed at identifying the mechanism(s) by which NE induces the growth of E. coli, demonstrated that stimulation by NE is due to its ability to facilitate iron removal from the host iron-binding proteins transferrin (Tf) and lactoferrin (Lf) [12]. Although we showed that Tf-derived iron and NE were internalised by strains of E. coli, and that Tf associated with the bacterial envelope, we were unable to demonstrate the existence of iron^ NE complexes free of Tf. Thus, it was not clear whether the interaction of NE with Tf or Lf alone was su⁄cient to result in increased provision of iron (through NE directly acting as an iron-delivery vehicle) or if alternative mechanisms, such as actual contact of the Tf protein, or the presence of an intermediary acceptor of the iron released by NE, such as a siderophore, were involved. In order to determine whether contact between bacterial cells and Tf^NE complexes is necessary for NE-mediated enhancement of growth, iron uptake experiments were performed in the presence or absence of NE, in which 55 Fe-labelled Tf was physically separated from the bacteria.

0378-1097 / 03 / $22.00 2 2003 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. doi:10.1016/S0378-1097(03)00243-X

FEMSLE 10950 6-5-03

40

P.P.E. Freestone et al. / FEMS Microbiology Letters 222 (2003) 39^43

2. Materials and methods 2.1. Media and chemicals Sterile SAPI minimal salts medium containing 0.2% (w/v) was prepared as previously described [5] and bu¡ered with 50 mM Tris^HCl (pH 7.5). Puri¢ed human iron-free and iron-saturated Tf, the (3)-isomer of NE and Fe(NO3 )3 were obtained from Sigma (Poole, UK). 55 Fe^Tf was prepared as described previously [12], and E. coli autoinducer (AI) [4] puri¢ed as described in [13].

D-glucose

2.2. Bacterial strains EHEC strain NCTC12900 (O157 :H7) mutants defective either in the synthesis of enterobactin (mutant RDH10, having an in-frame deletion in the entA gene) or in active siderophore uptake (mutant RDH11, with a kanamycin resistance cassette in the tonB gene) were constructed by standard PCR techniques and introduced into the EHEC background by homologous recombination on the sacBdependent positive selection vector pRDH10 [14]. The entA mutation was complemented (strain RDH12, pEntA) by transforming RDH10 with pUC18 containing entA [14]. 2.3. Bacterial growth and incorporation of of NE

55

Fe in presence

Analyses of growth-responsiveness to NE were performed in serum-supplemented SAPI medium [8,9,12]. Wild-type and mutant EHEC strain NCTC12900 (O157:H7) were grown overnight in Luria broth, washed and resuspended in phosphate-bu¡ered saline (PBS), and diluted into 1-ml aliquots of serum^SAPI medium to give an initial inoculum of approximately 102 colony-forming units (CFU)/ml [8,9,12]. Assays (supplemented with 50 WM NE, or an equivalent volume of water) were incubated statically for 16 h in a humidi¢ed 37‡C incubator in air supplemented with 5% v/v CO2 . Cultures were then mixed and aliquots withdrawn for estimation of growth by pour plate analysis as described previously [8,9,12]. Growth assays and plate counts were carried out in triplicate, and all experiments were performed on at least two separate occasions. To test the ability of the E. coli O157:H7 strains to acquire iron from Tf, 5 ml sterile SAPI medium bu¡ered with 50 mM Tris^HCl, pH 7.5, was supplemented with 50 WM NE or an equivalent volume of water, and with 2U105 cpm ml31 of ¢lter-sterilised 55 Fe^Tf added either directly into the medium or enclosed within 1-cm diameter dialysis membrane (4-kDa cut-o¡, Scienti¢c Industries International Inc., Loughborough, UK). Bacteria were grown as described above, washed twice with warm (37‡C) PBS, and added directly to uptake assay mixtures at 2^4U108 CFU ml31 . For siderophore cross-feeding ex-

periments, siderophore-donating wild-type NCTC12900 bacteria and the enterobactin-de¢cient mutant strain (as a negative control) were prepared as described above, enclosed within dialysis tubing, and added to entA 55 Fe uptake assays in equivalent numbers to the entA strain. Analyses of 55 Fe incorporation by the entA mutant in the presence of AI were carried out using 400 units of AI produced by E. coli grown in NE-supplemented serum^SAPI medium [4], and puri¢ed as described in [13]. A unit of AI activity is the minimum amount of AI that will stimulate the growth of E. coli strain E2348/69 in serum^SAPI medium from an initial inoculum of approximately 102 ^107 CFU ml31 at 37‡C in a 5% CO2 /air humidi¢ed incubator for 17 h [13]. All 55 Fe uptake assays were incubated at 37‡C in a 5% CO2 atmosphere for 4 h, during which time there was essentially no additional growth. Bacteria were then harvested, washed in PBS and assayed for cell numbers and 55 Fe incorporation as described previously [12]. 2.4. E¡ect of NE on iron-binding capacity of Tf Iron-free Tf (1 ml of a 1 mg ml31 solution) was enclosed in dialysis tubing and incubated for 18 h at 37‡C in 10 ml of Tris-bu¡ered SAPI medium (pH 7.5) containing 1 mg of iron-saturated Tf ml31 in the presence or absence of 50 WM NE. Following incubation, protein samples from within the tubing and in the surrounding bu¡er were analysed by electrophoresis in polyacrylamide gels containing 6 M urea as previously described [12]. 2.5. Statistical analysis Results were analysed by unpaired t-test in which a twotailed P value was calculated (Instat program, GraphPad Software). Statistical signi¢cance was de¢ned as a P value of 6 0.05.

3. Results and discussion 3.1. Enterobactin is required for NE-mediated growth enhancement The results shown in Fig. 1 suggest that enterobactin synthesis is required for E. coli O157:H7 to be able to respond to NE with enhanced growth when incubated at low initial inoculum in serum-supplemented SAPI medium. Complementation of the entA mutation restored the ability to respond to NE and allowed a 4-log increase in growth over unsupplemented control cultures (P 6 0.0001), indicating that the lack of responsiveness to NE was due to a defect in siderophore production, and not to an increase in serum-sensitivity of the entA mutant. The response to NE of E. coli enterobactin uptake mutants was also investigated. Since there is likely to be

FEMSLE 10950 6-5-03

P.P.E. Freestone et al. / FEMS Microbiology Letters 222 (2003) 39^43

Fig. 1. NE growth-responsiveness of EHEC NCTC12900 siderophore mutants. Growth of siderophore mutants of EHEC NCTC12900 inoculated at approximately 102 CFU ml31 into serum^SAPI medium containing no additions (control) or supplemented with 50 WM NE. Growth was enumerated by serial dilution of the cultures in PBS and plating onto Luria agar, as described in Section 2. The results shown (P 6 0.0001) are representative data from three separate experiments; within experiments data points typically showed variation of less than 5%.

more than a single receptor for enterobactin in EHEC we decided to inactivate TonB, the protein that energises outer membrane receptors. The tonB mutant showed no response to NE (P 6 0.0001) (Fig. 1), indicating that both enterobactin synthesis and uptake are required for NEmediated growth enhancement of E. coli O157:H7. 3.2. Enterobactin is required for NE-mediated Tf-derived Fe uptake Previous reports from our laboratories using low-density populations of E. coli inoculated into serum^SAPI Table 1 E¡ect of NE on incorporation of

55

B

C D

medium have shown a positive correlation between growth stimulation by NE (and other catecholamines) and bacterial acquisition of iron from Tf [2,12,15]. However, in these experiments, bacteria were in direct contact with the Tf. Table 1, part A, suggests that the presence of NE can also facilitate the acquisition of iron by wildtype E. coli O157:H7 by a non-contact-dependent route. When the bacteria were in direct contact with 55 Fe^Tf, the inclusion of NE in the dialysis bu¡er resulted in a 3.7-fold increase in iron uptake as compared to control (P 6 0.0001). Separation of bacteria and the Tf by the dialysis membrane resulted in an overall reduced e⁄ciency of iron acquisition ; however, supplementation with NE still resulted in a nearly 30-fold enhancement of iron uptake over unsupplemented control cultures (P 6 0.0001). These data clearly show that NE can mediate incorporation of iron from 55 Fe^Tf via a contact-independent mechanism, although we do not exclude the possibility of contact-dependent iron acquisition processes when bacteria are in close physical association with the Tf. Previous reports have established that microbial growth stimulation by NE is a widespread phenomenon [2^7,9, 12,15,16]; however, it is unlikely that the ¢nal stages of NE-mediated iron uptake will be the same in all bacterial species. For example, NE may act directly in a siderophore-like fashion for some bacteria due to its catecholate structure, but for others it may function indirectly by releasing iron that can subsequently be sequestered by endogenous siderophores. In the case of NCTC12900, addition of excess iron to the Luria broth culture used to prepare the inoculum for 55 Fe uptake experiments resulted in signi¢cant reduction in 55 Fe incorporation (Table 1, part A), suggesting that the NE-mediated Fe uptake process is iron regulated (P = 0.0028). Moreover, incubation of bacteria in SAPI medium containing sodium azide prior

Fe from Tf by EHEC strain NCTC12900

Strain/proximity to Tf

A

41

55

Treatment

Fe incorporation (cpm)a

3NE

+NE

P

Tf in the bu¡er: NCTC12900 (wild-type) Tf partitioned in dialysis tubing:

None

24 872 [783]

92 737 [1916]

6 0.0001

NCTC12900 NCTC12900 NCTC12900 RDH10 (entA) RDH11 (tonB) RDH12 (pEntA) RDH10 (entA) RDH10 (entA) RDH10 (entA)

None High iron mediumb Sodium azidec None None None Cross-fed with RDH10 (entA) Cross-fed with NCTC12900 (wild-type) Puri¢ed E. coli AI

642 [6] 26 [4] 29 [2] 871 [64] 1 041 [29] 853 [55] 1 003 [49] 516 [55] 948 [39]

19 356 [268] 158 [11] 108 [12] 361 [22] 525 [38] 24 771 [554] 839 [79] 13 144 [346] 1 059 [52]

6 0.0001 0.0028 0.0072 0.0059 0.0004 6 0.0001 0.0565 0.0004 0.0604

55

Fe uptake assays were performed as described in Section 2. Standard deviations are shown in square brackets; statistical signi¢cance is de¢ned as a P value of 6 0.05. a Representative data obtained in experiments performed on at least two separate occasions. b Bacteria grown in Luria broth containing 200 WM Fe(NO3 )3 prior to 55 Fe uptake assays. c Bacteria incubated at 37‡C for 30 min with the metabolic poison sodium azide (6 mM) prior to 55 Fe uptake assays.

FEMSLE 10950 6-5-03

42

P.P.E. Freestone et al. / FEMS Microbiology Letters 222 (2003) 39^43

to addition of 55 Fe^Tf also resulted in marked reduction in 55 Fe uptake, indicating that the uptake process is also dependent upon the ability to utilise ATP (P = 0.007). Both these observations are consistent with the requirement for a high-a⁄nity active iron transport mechanism. The major siderophore of E. coli, and the sole siderophore of NCTC12900, is enterobactin [17,18]. To test the hypothesis that EHEC uses this molecule as the ¢nal acceptor of iron released from Tf in the presence of NE, iron uptake was assayed in the entA and tonB mutants. When NE was present in the uptake medium, the ability of both mutants to acquire iron from 55 Fe^Tf partitioned within the dialysis tubing was very much reduced in comparison to the parental strain (Table 1, part B). Moreover, for the mutant strains the presence of NE resulted in a reduction in 55 Fe uptake rather than an enhancement, possibly due to the iron-chelating capacity of NE further reducing levels of any Tf-derived 55 Fe spontaneously released into the assay medium (P = 0.0059). Complementation of the entA mutant with a plasmid-encoded copy of the entA gene (RDH12, pEntA), restored its ability to accept the 55 Fe from partitioned Tf in the presence of NE (P 6 0.0001) (Table 1, part B), as did the inclusion into entA uptake assays of wild-type (i.e. enterobactin-producing) NCTC12900 cells enclosed within dialysis tubing (P = 0.0004), but not of the entA mutant, RDH10 (P = 0.065) (Table 1, part C). These results indicate that for E. coli O157: H7 enterobactin is essential for NE-mediated iron uptake. In addition to directly enhancing bacterial growth through the provision of iron from host iron-binding proteins such as Tf and Lf, growth of Gram-negative bacteria in serum^SAPI medium in the presence of NE also induces the synthesis of a novel, heat-stable bacterial growth stimulator, or AI [1,4,13]. A recent report by Burton et al. [19] suggested that AI may be enterobactin or its breakdown product 2,3-dihydroxybenzoylserine. If this were indeed the case, AI should be able to act as a substitute siderophore for 55 Fe uptake by an entA mutant. Table 1, part D, shows that addition of excess puri¢ed AI (400 units) [13] did not result in a signi¢cant increase in 55 Fe uptake by entA mutant strain RDH10 in the presence of NE (P = 0.0604). This shows that AI, unlike enterobactin, is not the acceptor of the iron released by the interaction of NE with Tf. The actual mechanism by which AI induces bacterial growth is currently under investigation in our laboratories. 3.3. Norepinephrine can mobilise Tf-bound Fe In order to demonstrate that NE can modulate the ironbinding capacity of Tf, thereby providing a more accessible iron source for acquisition by enterobactin, iron-free Tf was enclosed within dialysis tubing and incubated overnight in a bu¡er containing iron-saturated Tf as described in Section 2. In the absence of NE, the Tf in the dialysis membrane underwent mobility shifts consistent with par-

Fig. 2. Urea polyacrylamide gel electrophoresis of Tf either contained within dialysis tubing (lanes 1^3) or free in the surrounding bu¡er (lanes 4^6), incubated in the presence or absence of NE. Lanes 1 and 4 show the iron-free Tf within the membrane and essentially fully saturated Tf in the bu¡er, respectively, at the start of the experiment. Lanes 2 and 5 illustrate the e¡ect of incubation for 18 h at 37‡C in the absence of NE. Lanes 3 and 6 show the e¡ect of incubation in the presence of 50 WM NE. Lane M contains iron-free (Tf), monoferric with iron in the N-terminal or C-terminal domains (Fe^Tf and Tf^Fe, respectively), and saturated (Fe2 ^Tf) isoforms as markers.

tial conversion from iron-free to the two monoferric isoforms, probably due to scavenging of iron spontaneously released from the diferric Tf in the bu¡er (Fig. 2). In the presence of NE, however, most of the iron-free Tf within the dialysis membrane had become iron saturated and a signi¢cant proportion of the Tf in the external bu¡er (originally iron saturated) was in the monoferric and iron-free forms. These results demonstrate that NE stimulates redistribution, or shuttling, of Tf-bound iron between available binding sites on either side of the dialysis membrane. Whether this occurs as free ferric ions or as complexes with NE is not known, but it should be noted that the iron chelators citrate and nitrilotriacetate have similar e¡ects [20,21]. Our previous study, demonstrating the role of Tf^NE complexes in the ability of NE to stimulate bacterial growth, utilised low inoculum levels (100^1000 CFU ml31 ) in the restrictive environment of serum^SAPI medium for 16^18-h growth periods [12]. As such, it was not possible to distinguish if the NE-dependent iron uptake from 55 Fe^Tf we observed was due to NE acting directly in a siderophore-like manner, or NE^Tf interactions were releasing iron for subsequent uptake by bacterial iron acquisition systems, or these uptake systems could acquire Tf iron directly once bacterial growth had been induced by NE. By contrast, the experiments described in the present report examined iron uptake over reduced incubation times (4 h) in high-density populations ( s 108 CFU ml31 ) in a nutritionally poor salts medium lacking serum supplementation. As bacterial growth is no longer an important factor under these conditions we were able to make direct comparison of NE-dependent iron uptake in wild-type and growth-de¢cient siderophore mutants. The results of this study have shown that enterobactin plays an important role in the ability of NE to stimulate

FEMSLE 10950 6-5-03

P.P.E. Freestone et al. / FEMS Microbiology Letters 222 (2003) 39^43

EHEC growth since mutants defective in siderophore synthesis or uptake showed only basal levels of incorporation of Tf-derived iron. The importance of NE is that it apparently permits shuttling of iron between binding sites in Tf and also between Tf and enterobactin (Fig. 2) under serum-based conditions. It has been recognised for many years that stress can increase susceptibility to infectious diseases [22]. The vast majority of reports have attributed this increased risk to the deleterious e¡ects of stress on the immune competence of the host [23]. However, the results of the present study suggest a possible mechanism by which the release of stress hormones, such as NE, may interact with host defence mechanisms in an unexpected fashion and thereby in£uence bacterial pathogenicity.

Acknowledgements This work was supported by grant 064488/Z/01/Z from the Wellcome Trust (to P.P.E.F. and P.H.W.) grant F/212/ W from the Leverhulme Trust (to P.H.W.), and by Public Health Service grant AI-44918 from the National Institute of Allergy and Infectious Diseases (to M.L.).

References [1] Lyte, M., Frank, C.D. and Green, B.T. (1996) Production of an autoinducer of growth by norepinephrine cultured Escherichia coli O157:H7. FEMS Microbiol. Lett. 139, 155^159. [2] Neal, C.P., Freestone, P.P., Maggs, A.F., Haigh, R.D., Williams, P.H. and Lyte, M. (2001) Catecholamine inotropes as growth factors for Staphylococcus epidermidis and other coagulase-negative staphylococci. FEMS Microbiol. Lett. 194, 163^169. [3] Alverdy, J., Holbrook, C., Rocha, F., Seiden, L., Wu, R.L., Musch, M., Chang, E., Ohman, D. and Suh, S. (2000) Gut-derived sepsis occurs when the right pathogen with the right virulence genes meets the right host: evidence for in vivo virulence expression in Pseudomonas aeruginosa. Ann. Surg. 232, 480^489. [4] Freestone, P.P., Haigh, R.D., Williams, P.H. and Lyte, M. (1999) Stimulation of bacterial growth by heat-stable, norepinephrine-induced autoinducers. FEMS Microbiol. Lett. 172, 53^60. [5] Lyte, M. and Ernst, S. (1992) Catecholamine induced growth of Gram-negative bacteria. Life Sci. 50, 203^212. [6] Belay, T. and Sonnenfeld, G. (2002) Di¡erential e¡ects of catecholamines on in vitro growth of pathogenic bacteria. Life Sci. 71, 447^ 456.

43

[7] Kinney, K.S., Austin, C.E., Morton, D.S. and Sonnenfeld, G. (2000) Norepinephrine as a growth stimulating factor in bacteria-mechanistic studies. Life Sci. 67, 3075^3085. [8] Lyte, M., Erickson, A.K., Arulanandam, B.P., Frank, C.D., Crawford, M.A. and Francis, D.H. (1997) Norepinephrine-induced expression of the K99 pilus adhesin of enterotoxigenic Escherichia coli. Biochem. Biophys. Res. Commun. 232, 682^686. [9] Lyte, M., Arulanandam, B.P. and Frank, C.D. (1996) Production of Shiga-like toxins by Escherichia coli O157:H7 can be in£uenced by the neuroendocrine hormone norepinephrine. J. Lab. Clin. Med. 128, 392^398. [10] Lyte, M. (1993) The role of microbial endocrinology in infectious disease. J. Endocrinol. 137, 343^345. [11] Lyte, M. and Bailey, M.T. (1997) Neuroendocrine-bacterial interactions in a neurotoxin-induced model of trauma. J. Surg. Res. 70, 195^ 201. [12] Freestone, P.P., Lyte, M., Neal, C.P., Maggs, A.F., Haigh, R.D. and Williams, P.H. (2000) The mammalian neuroendocrine hormone norepinephrine supplies iron for bacterial growth in the presence of transferrin or lactoferrin. J. Bacteriol. 182, 6091^6098. [13] Freestone, P.P.E., Williams, P.H., Lyte, M. and Haigh, R.D. (2001) Escherichia coli, Salmonella or Hafnia autoinducers; Patent #6, 316, 244. United States Patent Trademark O⁄ce. [14] Haigh, R.D. (1999) Characterisation of phosphorylation-de¢cient mutants of enteropathogenic Escherichia coli. Ph.D. thesis. University of Leicester. [15] Freestone, P.P., Williams, P.H., Haigh, R.D., Maggs, A.F., Neal, C.P. and Lyte, M. (2002) Growth stimulation of intestinal commensal Escherichia coli by catecholamines: a possible contributory factor in trauma-induced sepsis. Shock 18, 465^470. [16] Coulanges, V., Andre, P., Ziegler, O., Buchheit, L. and Vidon, D.J. (1997) Utilization of iron-catecholamine complexes involving ferric reductase activity in Listeria monocytogenes. Infect. Immun. 65, 2778^2785. [17] Neilands, J.B. (1995) Siderophores: structure and function of microbial iron transport compounds. J. Biol. Chem. 270, 26723^26726. [18] Ratledge, C. and Dover, L.G. (2000) Iron metabolism in pathogenic bacteria. Annu. Rev. Microbiol. 54, 881^941. [19] Burton, C.L., Chhabra, S.R., Swift, S., Baldwin, T.J., Withers, H., Hill, S.J. and Williams, P. (2002) The growth response of Escherichia coli to neurotransmitters and related catecholamine drugs requires a functional enterobactin biosynthesis and uptake system. Infect. Immun. 70, 5913^5923. [20] Morgan, E.H. (1979) Studies on the mechanism of iron release from transferrin. Biochim. Biophys. Acta 580, 312^326. [21] Aisen, P. (1968) Citrate-mediated exchange of FE3þ among tranferrin molecules. Biochem. Biophys. Res. Commun. 32, 220^226. [22] Peterson, P.K., Chao, C.C., Molitor, T., Murtaugh, M., Strgar, F. and Sharp, B.M. (1991) Stress and pathogenesis of infectious disease. Rev. Infect. Dis. 13, 710^720. [23] Webster, J.I., Tonelli, L. and Sternberg, E.M. (2002) Neuroendocrine regulation of immunity. Annu. Rev. Immunol. 20, 125^163.

FEMSLE 10950 6-5-03