JOURNAL OF BACrERIOLOGY, June 1994, p. 3455-3459
Vol. 176, No. 12
0021-9193/94/$04.00+0 Copyright © 1994, American Society for Microbiology
Siderophore Activity of myo-Inositol Hexakisphosphate in Pseudomonas
aeruginosa
ANTHONY W. SMITH,* DAVID R. POYNER, H. KEVIN HUGHES, AND PETER A. LAMBERT Department of Pharmaceutical and Biological Sciences, Aston University, Aston Triangle, Birmingham B4 7ET, United Kingdom Received 18 January 1994/Accepted 18 March 1994
myo-Inositol hexakisphosphate (InsP6), which is found in soil and most, if not all, plant and animal cells, has been estimated to have an affinity for Fe3' in the range of 1025 to 1030 M-1. In this report, we demonstrate that the Fe-InsP6 complex has siderophore activity and is able to reverse the iron-restricted growth inhibition of Pseudomonas aeruginosa by ethylene diamine di(o-hydroxyphenyl)acetic acid. With 55Fe-InsP6 in transport studies, iron uptake is strongly iron regulated, being repressed after growth in iron-replete conditions and inhibited by treatment with potassium cyanide and carbonyl cyanide m-chlorophenylhydrazone. The kinetics of iron transport revealed a Km of 100 nM. Self-displacement of binding of [3HJInsP6 to isolated membranes by InsP6 revealed a single class of binding sites (Kd = 143 + 6 nM; Hill coefficient, 1.1 0.1). The binding of [3H]InsP6 to membranes was not dependent on whether cells had been grown under conditions of high or low iron concentrations. We believe that this is the first report of inositol polyphosphate activity in prokaryotic cells. In an oxygen-containing environment, iron is readily oxidized from the ferrous to the ferric state, forming at a neutral pH essentially insoluble oxyhydroxide polymers which are not bioavailable to bacteria. This iron deficiency is recognized to have an impact on many bacterial systems, including virulence in human, animal, and plant hosts (11, 22, 25, 34). Most bacterial species require iron for growth and have developed distinct mechanisms to solubilize and transport environmental iron (20). Many gram-negative species synthesize iron chelators (siderophores), which are based on catechols and hydroxamates and which solubilize iron and transport it into the cell via specific receptor proteins in the outer membrane. A feature of this system is its regulation by iron; under iron-deprived conditions, it is derepressed, whereas under iron-replete conditions, it is repressed (4). In addition, some organisms also utilize siderophores produced by other bacterial and fungal species. Examples include Escherichia coli, which can transport iron by using the fungal siderophores ferrichrome, coprogen, and rhodotorulic acid (reviewed in reference 6), and Pseudomonas aeruginosa, which can utilize ferrioxamine B, a siderophore produced by Streptomyces spp. (7), and enterobactin, one of the major siderophores produced by E. coli (23). Strains of the plant growth-promoting bacterium Pseudomonas putida can also utilize siderophores from a wide variety of other root-colonizing pseudomonads (15), and it has been postulated that this ability may contribute to improving its capacity to compete for iron in the rhizosphere. Other pathogens have adapted mechanisms to use sources of iron in vivo, including transferrin and lactoferrin (19, 26, 27). Other adaptive mechanisms include the utilization of iron from ferric dicitrate, notably the inducible system in E. coli (32). At present, it is not clear whether there are alternative sources of complexed iron, other than those of fungal or bacterial origin, which species can utilize to improve their survival prospects.
One candidate
source
is iron complexed to myo-inositol
hexakisphosphate (InsP6).
InsP6 (phytic acid) is found in soil and most, if not all, plant and animal cells (8, 29). In plants, large amounts of InsP6 accumulate in seeds and other storage tissues in organelles identified as protein bodies. The role of InsP6 and the various inositol pentakisphosphate isomers in mammalian cells is less clear, despite the finding that they are present in larger amounts than other inositol polyphosphates (5, 17). InsP6 binding activity has been demonstrated on many neuronal structures in the brain, including the cerebellum (13, 21). More recently, it was demonstrated that an InsP6 receptor is related to clathrin assembly protein AP-2 (31, 33), suggesting a possible link with receptor-mediated endocytosis. It has been recognized for many years that InsP6 is a potent metal ion chelator, and recently attention was drawn to its ability to bind Fe3+, which in turn potentiates its binding to rat cerebellar membranes (24). Competition experiments with other chelators suggest that InsP6 has an affinity for Fe3+ at a neutral pH in the range of 1025_103o M-1 (12), a value comparable to those of many fungal and bacterial siderophores. Given the possibility that iron in the environment could be chelated by InsP6, we were prompted to determine whether bacterial species, notably the soil organism P. aeruginosa, could utilize iron from this source. In this report, we characterize the siderophore activity of InsP6 in P. aeruginosa. MATERIALS AND METHODS Bacterial strains and culture conditions. P. aeruginosa PAO1 (ATCC 15692) was from our laboratory collection. IAl, a siderophore-deficient mutant of PAO1 (3), was obtained from C. D. Cox (University of Iowa). Strains were cultured in iron-deficient succinate medium (18). Iron-replete cells were grown in succinate medium supplemented with 100 ,uM FeCl3. 55Fe-InsP6 transport studies. Cells were grown in succinate medium in an orbital shaking incubator at 37°C for 15 h and harvested by centrifugation at 4°C. The cells were washed twice in 100 mM N-morpholinepropanesulfonic acid (MOPS) (pH
* Corresponding author. Present address: School of Pharmacy and Pharmacology, University of Bath, Claverton Down, Bath BA2 7AY, United Kingdom. Phone: 44 225 826826. Fax: 44 225 826114. Electronic mail address:
[email protected].
3455
3456
SMITH ET AL.
J. BACTERIOL.
30
G) U
20
0
10
0 FIG. 1. Disk bioassay for siderophore activity of InsP6 in P. aeruginosa IAl inhibited by 400 pRM EDDHA. Disks: 1, 20 ,ul of 10 mM InsP6-20 1iM FeCl3; 2, 20 RIl of 10 mM InsP6; 3, 20 RI of 20 ,uM FeCl3.
7.0) containing 60 p,M glucose, suspended to an optical density at 470 nm of 2.0 (corresponding to 2 x 109 cells per ml), and equilibrated at 37°C for 15 min prior to transport studies. No observable increase in the optical density at 470 nm was detected during this period. Potassium cyanide and carbonyl cyanide m-chlorophenylhydrazone (CCCP) were added as needed to the cell suspension to 20 and 2 mM, resulting in final concentrations in the uptake assay of 10 and 1 mM, respectively. The Fe-InsP6 complex was formed at 37°C 15 min prior to uptake as 200 p,M myo-InsP6 (Aldrich Chemical Co., Gillingham, Dorset, United Kingdom; purity was confirmed by 3"P nuclear magnetic resonance spectroscopy) and 400 nM 55FeC13 (1.91 GBq mg-'; Amersham) in 1 ml of MOPS buffer. The addition of this labelled complex to the cell suspension resulted in final concentrations of 100 p.M InsP6 and 200 nM 55FeC13 (500:1 ratio). Iron transport was assayed by withdraw-
ing 200-plI samples and filtering them through 0.2-pum-poresize cellulose acetate membrane filters. The membrane filters were washed twice with 10-ml volumes of saline and allowed to air dry. The activity retained on the membrane filters was determined by scintillation counting on the 3H channel. [3H]InsP6 binding assays. Cell membranes were prepared from cells grown for 15 h in succinate medium, broken by sonication, and sedimented by centrifugation at 100,000 x g for 1 h. Binding was carried out at 4°C for 30 min with 100 mM MOPS (pH 7.0) and terminated by centrifugation (13). Routine assays were performed with a final reaction volume of 0.5 ml containing 1.0 nM [3H]InsP6 (902.8 GBq/mmol; 0.37 MBq/ ml; NEN; approximately 30 000 dpm per assay) and approximately 1 mg of membrane protein per ml. RESULTS Growth promotion by Fe-InsP6. Preliminary experiments indicated that InsP6 could mediate iron uptake into P. aeruginosa; however, for siderophore activity it was necessary to demonstrate a reversal of the growth inhibition imposed by iron limitation. P. aeruginosa IAl, a mutant of type strain PAO1 which is deficient in siderophore production (3) and susceptible to the growth inhibition imposed by nonutilizable chelators, was selected for further study. Figure 1 shows a surface-seeded succinate agar plate supplemented with 400 puM ethylene diamine di(o-hydroxyphenyl)acetic acid (ED-
10
20
30
Time (min) FIG. 2. 55Fe transport from Fe-InsP6 by P. aeruginosa after growth in iron-depleted succinate medium (0) and succinate medium supplemented with 100 ,uM FeCl3 (0). The energy dependence of the process was determined by performing experiments with iron-depleted cells in the presence of 1 mM CCCP (A) and 10 mM KCN (A). The data are the mean + the standard deviation for four experiments.
DHA). A zone of growth occurred around a disk supplemented with 20 RI of 10 mM InsP6-20 1iM FeCl3 (Fig. 1, disk 1, 10 mm), whereas there was negligible growth around a disk supplemented with FeCl3 only (Fig. 1, disk 3, 1 mm). InsP6 alone resulted in a small zone of growth (Fig. 1, disk 2, 3 mm). Uptake time course. Iron supplied to P. aeruginosa in the form of Fe-InsP6 was very efficiently taken up by the cells (Fig. 2). The highest rate of transport was seen in cells grown in an iron-deficient medium. When this medium was supplemented with 100 ,uM ferric chloride, there was an approximately 80% decrease in the initial rate of transport, indicating that the transport process was iron repressible. The rate of transport was also decreased by reducing the temperature of the uptake experiment. No transport was seen when the cells were incubated on ice, although low counts accumulated on the cells at 30 s and then remained constant (data not shown). To investigate further the presence of a carrier-mediated process, the kinetics of assimilation were studied. A rapid initial rate of uptake over 30 s, which also occurred with cells incubated on ice (see above), was followed by a slower rate of iron accumulation, probably reflecting true transport. The rate of transport over the range of 5 to 600 nM is shown in Fig. 3. The initial transport rates were determined between 30 s and 7 min; the Vman was 0.58 pmol/109 cells per min. The Km was found to be approximately 100 nM. Binding of InsP6 to cell membranes. Experiments carried out with a buffer of 100 mM MOPS (pH 7.0) at 37°C (as in the "5Fe uptake studies) indicated that binding was unstable, possibly because of degradation of the radiolabelled ligand. However, at 4°C, binding was essentially complete by 5 min (the fastest time measurable with a microcentrifuge binding assay) and stable for at least 30 min; consequently, these conditions were used for subsequent experiments. Self-displacement of InsP6 revealed a single class of binding sites (Kd = 143 + 6 nM; Hill coefficient, 1.1 + 0.1; n = 3) (Fig. 4). The binding was heat sensitive, as it was reduced by approximately 50% on boiling of the membranes for 10 min before the binding assay was carried out.
SIDEROPHORE ACTIVITY OF myo-INOSITOL HEXAKISPHOSPHATE
VOL. 176, 1994
~0.6I
0.4
>
0.2
0.0 100
0
Fe(EI) FIG. 3.
300
200
400
600
500
concentration
700
(nM)
Concentration-dependent kinetics of "5Fe uptake by P. in iron-depleted succinate medium at an Fe/InsP6
aeruginosa grown
ratio of 1:500. The data are the mean
experiments. V,
rate of
the standard deviation for four
±
transport.
DISCUSSION Our results show that
InsP6
has
siderophore activity
in P.
aeruginosa. The Fe-InsP6 complex was able to reverse the growth inhibition imposed by the iron(III) chelator EDDHA; iron alone at the same concentration did
not
support growth.
growth around a disk supplemented with InsP6 alone was not as large as that around a disk containing the iron complex, a result which we attribute either to an inability of InsP6 to compete with EDDHA for bound iron or to very slow The zone of
,0 0
*Ob
;9 In
1
100
10
InsP6
1000
10000
(nM)
Competition [3H]InsP6 binding by unlabelled point represents the mean + the standard deviation for three determinations. Binding in the absence of unlabelled InsP6 was approximately 2,200 dpm per incubation. Nonspecific binding was defined in the presence of 1 mM unlabelled InsP6 (approximately 150 dpm). The curve represents fit to a single site following mass-action kinetics, with a Kd of 143 nM. FIG.
InsP6.
4.
Each
with
1 nM
3457
kinetics of exchange between the ligands. Simple in vitro competition experiments with InsP6 and EDDHA support this conclusion. The characteristics of iron(III) uptake resemble those of siderophore-dependent systems, including repression after growth under conditions of high iron availability and inhibition by the metabolic inhibitors KCN and CCCP. The uptake system did not appear to be inducible by growth in the presence of InsP6, nor did InsP6 induce changes in outer membrane protein expression (data not shown). The Km was found to be approximately 100 nM, a value which falls within the range of values reported for siderophore-mediated iron transport. These extend from the value for the extremely high-affinity system for ferric schizokinen transport (Km = 0.04 ,uM) in the cyanobacterium Anabaena sp. (16) to those for comparatively low-affinity systems, such as that for ferric coprogen transport (Km = 5 ,iM) in Neurospora crassa (14). Binding studies with radiolabelled InsP6 and whole cells, as in the iron transport experiments, at either 37 or 4°C yielded very low counts which were poorly reproducible. Therefore, to increase the potential number of binding sites, cell membranes, comprising both cytoplasmic and outer membranes, were used. Each incubation represented binding sites from approximately 100-fold more cells than in the iron transport assays. The characteristics of the binding sites at 4°C were similar to those observed for mammalian tissues (24), for which it has been argued that in the absence of EDTA and particularly in the presence of Mg2+, InsP6 can associate with membrane components (perhaps phospholipids) in a metal ion-dependent fashion. As with many mammalian membranes, Mg2+ potentiated and 5 mM EDTA inhibited InsP6 binding (Mg2+/EDTA ratio, 3 + 0.4; n = 3). However, even in the presence of EDTA, an InsP6 binding site with a capacity of 1.45 pmol/mg of membrane protein could still be identified, with an affinity not significantly different from that seen in the absence of EDTA (data not shown). It was previously argued that in the presence of EDTA, InsP6 binding is due to membrane proteins (24), and it is likely that a component of this binding represents an InsP6 recognition protein involved in Fe3+ uptake. The whole-cell transport studies indicated a very rapid association of the Fe-InsP6 complex with the cells and then a slower, energydependent accumulation of Fe3+. The binding site could be involved in the first phase. Interestingly, no difference in InsP6 binding was seen with membranes from cells grown under either iron-replete or iron-deficient conditions, indicating that the iron-dependent transport of Fe is not due to a variation in the interaction with the cell surface. In contrast, Fe uptake from InsP6 by whole cells was partially repressed by growth in iron-rich media. Therefore, some components are presumably regulated by iron, while the membrane-associated binding proteins are not. At the present time, it is not clear how iron is released or transported from the Fe-InsP6 complex. Our InsP6 binding studies with isolated membranes at 37°C indicated that there may be some degradation of the radiolabelled ligand (data not shown). There have been some reports of phytases from microorganisms (9), and we are currently designing phytaseresistant InsP6 analogs to determine whether breakdown is part of the iron release mechanism. P. aeruginosa does produce a phosphatase, but it is unlikely that it is strongly expressed under the phosphate-rich growth conditions used in this study (30), nor is it known how, if released, iron would enter the cell. One candidate mechanism could be a system similar to the SfuABC iron transport system of Serratia marcescens, which accepts Fe3+ solubilized with oxaloacetate, sodium PPi, and citrate (35). When E. coli sfiABC transformants mutated in the
3458
SMITH ET AL.
fecBCDE genes, which are required for citrate-mediated iron transport across the cytoplasmic membrane, were used, it was shown that under conditions of mild iron deficiency, citrate carried iron across the outer membrane and the Sfu system carried it across the cytoplasmic membrane. Under conditions of severe iron deficiency, the TonB-dependent outer membrane receptor FecA was required (2). Three SfuABC proteins exhibited typical features of a periplasmic binding proteindependent mechanism; these included a periplasmic polypeptide (SfuA), a very hydrophobic protein (SfuB), and a hydrophilic protein with nucleotide-binding domains (1). If such a system does exist in P. aeruginosa, it is not clear at this stage how iron or the complex crosses the outer membrane. A polyphosphate-selective porin, OprO, has been identified in P. aeruginosa, but it was expressed only in stationary phase under conditions of phosphate starvation (28). In this work, phosphate-rich growth conditions were used, arguing against a role for OprO in the transport process. Moreover, cells grown in phosphate-depleted medium did not show increased accumulation of Fe (data not shown). Previous reports of inositol activity in bacteria have been limited to myo-inositol transport in Klebsiella aerogenes and a single Pseudomonas sp. isolated from soil, in which cyclitol can be used as the sole carbon and energy source (10). myo-Inositol penetrates K aerogenes by a system which is linked to proton symport and in which ATP is not directly involved, whereas in the Pseudomonas sp., penetration is dependent on a periplasmic binding protein. We have not been able to determine whether InsP6 enters the cell in our experimental system. However, it is clear that sufficient iron accumulates intracellularly to reverse the iron restriction imposed by EDDHA. The accumulation of iron from Fe-InsP6 exhibits many of the characteristics of a siderophore-mediated process. Given the ubiquity of InsP6 and its affinity for iron(III), its utilization as a siderophore is an attractive strategy for cells to adopt, obviating the demand imposed on cells by siderophore biosynthesis. ACKNOWLEDGMENTS We thank P. T. Hawkins for helpful discussions and Steve Howitt for technical assistance. This work was supported by a grant to A.W.S. from the Royal Society.
1.
2. 3. 4. 5. 6. 7.
8.
REFERENCES Angerer, A., S. Gaisser, and V. Braun. 1990. Nucleotide sequences of the sjfuA, sfuB, and sfuC genes of Serratia marcescens suggest a periplasmic-binding-protein-dependent iron transport mechanism. J. Bacteriol. 172:572-578. Angerer, A., B. Klupp, and V. Braun. 1992. Iron transport systems of Serratia marcescens. J. Bacteriol. 174:1378-1387. Ankenbauer, R., S. Sriyosachati, and C. D. Cox. 1985. Effects of siderophores on the growth of Pseudomonas aeruginosa in human serum and transferrin. Infect. Immun. 46:132-140. Bagg, A., and J. B. Neilands. 1987. Molecular mechanism of regulation of siderophore-mediated iron assimilation. Microbiol. Rev. 51:509-518. Berridge, M. J., and R. F. Irvine. 1989. Inositol trisphosphate and calcium signalling. Nature (London) 341:197-205. Braun, V., K. Gunter, and K. Hantke. 1991. Transport of iron across the outer membrane. Biol. Metals 4:14-22. Cornelis, P., N. Moguilevsky, J. F. Jacques, and P. L. Masson. 1987. Studies of the siderophores and receptors in different clinical isolates of Pseudomonas aeruginosa, p. 290-306. In G. Doring, I. A. Holder, and K. Botzenhart (ed.), Basic research and clinical aspects of Pseudomonas aeruginosa. S. Karger, Basel. Cosgrove, D. J. 1980. Inositol phosphates: their chemistry, bio-
J. BACTERIOL.
chemistry and physiology. Elsevier, Amsterdam. 9. Cosgrove, D. J., D. C. J. Irving, and S. M. Bromfield. 1970. Inositol phosphate phosphatases of microbiological origin. The isolation of soil bacteria having inositol phosphate phosphatase activity. Aust. J. Biol. Sci. 23:339-343. 10. Deshusses, J. P. M. 1985. myo-Inositol transport in bacteria: H+ symport and periplasmic binding protein dependence. Ann. N. Y. Acad. Sci. 456:351-360. 11. Griffiths, E. 1991. Iron and bacterial virulence-a brief overview. Biol. Metals 4:7-13. 12. Hawkins, P. T., D. R. Poyner, T. R. Jackson, A. J. Letcher, D. A. Lander, and R. F. Irvine. 1993. Inhibition of iron-catalysed hydroxyl radical formation by inositol polyphosphates-a possible physiological function for myo-inositol hexakisphosphate. Biochem. J. 294:929-934. 13. Hawkins, P. T., J. M. Reynolds, D. R. Poyner, and M. R. Hanley. 1990. Identification of a novel inositol phosphate recognition site-specific [3H] inositol hexakisphosphate binding to brain regions and cerebellar membranes. Biochem. Biophys. Res. Commun. 167:819-827. 14. Huschka, H., H. U. Naegeli, H. Leuenberger-Ryf, W. KellerSchierlein, and G. Winkelmann. 1985. Evidence for a common siderophore transport system but different siderophore receptors in Neurospora crassa. J. Bacteriol. 162:715-721. 15. Koster, M., J. van de Vossenberg, J. Leong, and P. J. Weisbeek. 1993. Identification and characterisation of the pupB gene encoding an inducible ferric-pseudobactin receptor in Pseudomonas putida WCS358. Mol. Microbiol. 8:591-601. 16. Lammers, P. J., and J. Sanders-Loehr. 1982. Active transport of ferric schizokinen in Anabaena sp. J. Bacteriol. 151:288-294. 17. Martin, J., M. Foray, G. Keen, and M. Satre. 1987. Identification of inositol hexaphosphate in 31P-NMR spectra of Dictyostelium discoideum amebas-relevance to intracellular pH determination. Biochim. Biophys. Acta 931:16-25. 18. Meyer, J. M., and M. A. Abdallah. 1978. The fluorescent pigment of Pseudomonasfluorescens: biosynthesis, purification and physicochemical properties. J. Gen. Microbiol. 107:319-328. 19. Morton, D. J., and P. Williams. 1989. Utilisation of transferrinbound iron by Haemophilus species of human and porcine origins. FEMS Microbiol. Lett. 65:123-128. 20. Neilands, J. B. (ed.). 1974. Microbial iron metabolism. Academic Press, Inc., New York. 21. Nicoletti, F., V. Bruno, S. Cavallora, A. Copari, M. A. Sortino, and P. L. Canonico. 1990. Specific binding sites for inositol hexakisphosphate in brain and anterior pituitary. Mol. Pharmacol. 37:689-693. 22. Payne, S. M. 1988. Iron and virulence in the family Enterobacteriaceae. Crit. Rev. Microbiol. 16:81-111. 23. Poole, K., L. Young, and S. Neshat. 1990. Enterobactin transport in Pseudomonas aeruginosa. J. Bacteriol. 172:6991-6996. 24. Poyner, D. R., F. Cooke, M. R. Hanley, D. J. M. Reynolds, and P. T. Hawkins. 1993. Characterization of metal ion-induced [3H] inositol hexakisphosphate binding to rat cerebellar membranes. J. Biol. Chem. 268:1032-1038. 25. Schippers, B., A. W. Bakker, and P. A. H. M. Bakker. 1987. Interactions of deleterious and beneficial rhizosphere microorganisms and the effect of cropping practices. Annu. Rev. Phytopathol. 25:339-358. 26. Schryvers, A. B., and L. J. Morris. 1988. Identification and characterization of the transferrin receptor from Neisseria meningitidis. Mol. Microbiol. 2:281-288. 27. Schryvers, A. B., and L. J. Morris. 1988. Identification and characterization of the human lactoferrin binding protein from Neisseria meningitidis. Infect. Immun. 56:1144-1149. 28. Siehnel, R. J., C. Egli, and R. E. W. Hancock. 1992. Polyphosphate-selective porin OprO of Pseudomonas aeruginosa: expression, purification and sequence. Mol. Microbiol. 6:2319-2326. 29. Stephens, L. R., P. T. Hawkins, A. F. Stanley, T. Moore, D. R. Poyner, P. J. Morris, M. R. Hanley, R. R. Kay, and R. F. Irvine. 1991. myo-Inositol pentakisphosphates-structure, biological occurrence and phosphorylation to myo-inositol hexakisphosphate. Biochem. J. 275:485-499. 30. Tan, A. S. P., and E. A. Worobec. 1993. Isolation and character-
VOL. 176, 1994
SIDEROPHORE ACTIVITY OF myo-INOSITOL HEXAKISPHOSPHATE
ization of two immunochemically distinct alkaline phosphatases from Pseudomonas aeruginosa. FEMS Microbiol. Lett. 106:281286. 31. Timerman, A. P., M. M. Mayrleitner, T. J. Lukas, C. C. Chadwick, A. Saito, D. M. Watterson, H. Schindler, and S. Fleischer. 1992. Inositol polyphosphate receptor and clathrin assembly protein AP-2 are related proteins that form potassium selective channels in planar lipid bilayers. Proc. Natl. Acad. Sci. USA 89:89768980. 32. Van Hove, B., H. Staudenmaier, and V. Braun. 1990. Novel two-component transmembrane transcription control: regulation
3459
of iron dicitrate transport in Escherichia coli K-12. J. Bacteriol. 172:6749-6758. 33. Voglmaier, S. M., J. H. Keen, J.-E. Murphy, C. D. Ferris, G. D. Prestwich, S. H. Snyder, and A. B. Theibert. 1992. Inositol hexakisphosphate receptor identified as the clathrin assembly protein AP-2. Biochem. Biophys. Res. Commun. 187:158-163. 34. Weinberg, E. D. 1984. Iron withholding-a defense against infection and neoplasia. Physiol. Rev. 64:65-101. 35. Zimmermann, L., A. Angerer, and V. Braun. 1989. Mechanistically novel iron(III) transport system in Serratia marcescens. J. Bacteriol. 171:238-243.
