Role of two siderophores in Ustilago sphaerogena

242

Biochimica etBiophysica Acta, 720 (1982) 242-249 Elsevier Biomedical Press

BBA 11024

ROLE OF TWO S I D E R O P H O R E S IN USTILAGO SPHAEROGENA REGULATION OF BIOSYNTHESIS AND UPTAKE M E C H A N I S M S D.J. ECKER, C.W. PASSAVANT and T. EMERY

Department of Chemistry and Biochemistry, Utah State University, Logan, UT 84322 (U.S.A.) (Received November 9th, 1981 )

Key words: Siderophore synthesis; Ferrichrome; Iron transport," ( Ustilago sphaerogena)

Under iron-deficient conditions the smut fungus Ustilago sphaerogena produces two kinds of siderophores, ferrichrome and ferrichrome A. Regulation of ligand biosyntheses and uptake mechanisms of the iron chelates were studied to determine the role of each chelate in U. sphaerogena. The biosynthesis of each ligand was differentially regulated. Ferrichrome A, the more effective chelate, was preferentially synthesized under more extreme conditions of iron stress, but completely repressed when the cell was supplied with sufficient iron. In contrast, biosynthesis of ferrichrome was strongly but not completely repressed by iron. The mechanism of repression was examined using a newly developed in vivo synthesis assay. Chromium and gallium-containing siderophore analogs had no effect on siderophore ligand biosynthesis. Iron, added as siderophores, resulted in increased oxygen uptake and amino acid transport, which was soon followed by decreased iigand biosynthesis, suggesting that regulation may he indirect and related to oxidative metabolism. Uptake experiments were used to rule out a ligand-exchange mechanism for ferrichrome A-iron transport. The data suggest that ferrichrome A-iron is taken up at a specific site that results in a rapid distribution of iron inside the cell.

Introduction

In order to sequester iron from the environment, many microorganisms excrete high affinity iron-chelating agents [I]. These compounds were originally named siderochromes [2], but more recently the term siderophore has become accepted [3]. Many organisms excrete multiple types of siderophore ligands. For example, Ustilago sphaerogena excretes deferriferrichrome and deferriferrichrome A [4]. Fusarium roseum excretes ester-type fusarinines as well as the peptide-type malonichrome ligand [5]. Three mechanisms for microbial iron transport Abbreviations: Mops, 4-morpholinepropanesulfonicacid. 0167-4889/82/0000-0000/$02.75 © 1982 Elsevier Biomedical Press

have been proposed [6]. In the first, the metal chelate is transported into the cell intact and cellular dissociation of the metal involves destruction of the ligand. Enterobactin transport in Salmonella is an example. In the second mechanism, the chelate is transported intact but the ligand is conserved and re-excreted for another cycle of transport. Ferrichrome transport in /_7. sphaerogena occurs by this mechanism. Finally, the chelate may merely bring iron to the cellular membrane where dissociation occurs. The metal enters the cell, but the ligand remains extracellular. This mechanism has been observed for ferrichrome A iron-uptake in U. sphaerogena. With few exceptions, the biosynthesis of these compounds is severely inhibited by trace amounts

243

of iron in the growth medium. This has been observed in numerous bacteria and fungi, for example, the biosynthesis of ferrichrome [4], ferrioxamines [2], fusarinines [7], mycobactin [8], rhodotorulic acid [9] and coprogen [10]. The mechanism of this inhibition is unknown. In this present paper we (a) develop a sensitive new assay for the in vivo synthesis of ferrichrome and ferrichrome A ligands, (b) utilize the assay to investigate the mechanism by which ferrichrome and its metal analogs inhibit hydroxamate synthesis, and (c) provide data which suggest a rational basis for production of two different siderophores by a single organism. Methods

Cell cultures. U. sphaerogena (ATCC 12421) was grown under iron-deficient conditions and uptake experiments were performed as previously described [11]. Growth medium was made copper deficient by passage through a column of Chelex 100 (Bio-Rad) resin, Na + form, before addition of other trace metals, and cells were transferred through at least three subcultures on this medium prior to use. Cells were disrupted either with a Braun MSK homogenizer [12], or by passing.a suspension twice through a French pressure cell at 15 • 103 1b . / i n 2. Fractionation of cell homogenates was by the method of Schatz et al. [13]. Respiration was determined with an 02 electrode on airsaturated, 4ml aliquots from a 400 ml irondeficient 4-day culture. Siderophores. Siderophores were isolated from cultures of U. sphaerogena and labelled with 59Fe as previously described [11]. Cr(III)deferriferrichrome was prepared by the method of Leong and Raymond [14]. Ga(III)deferriferrichrome was prepared by the method of Emery and Hoffer [15]. Radiopurity of the metal chelates was checked by p a p e r e l e c t r o p h o r e s i s on p y r i d i n e / a c e t i c acid/water, 14:10:930 (v/v), p H 5.2, and silica gel thin layer chromatography on two solvent systems (methanol/water, 80:20, (v/v), and chlorof o r m / m e t h a n o l / w a t e r , 35 : 60 : 5 (v/v)). Separation and quantitation of ferrichrome and ferrichrome A was done by paper electrophoresis, water elution from the dried paper, and absorbance measurements at 425 nm (c -- 2895 M - n . c m - ~) and

440 nm ( c = 3 3 6 0 M -~ . c m - n ) , respectively. Quantitation of total hydroxamates was by the method of Atkin et al. [16]. Amino acid determination. Cells were harvested by centrifugation, washed and lyophilized. Amino acids were extracted from the dry pellet with 14 ml 6% trichloroacetic a c i d / g cells for 1 h at 25°C. Following centrifugation, the supernatant was extracted three times with 2vol. diethyl ether to remove acid. Amino acids were purified on a A G 50W-X8, 100-200 mesh H + form column. The column was washed with 3 bed vol. water, and amino acids eluted with 1 M NH4OH. After taking to dryness, the sample was analyzed on a Beckman model 120B amino acid analyzer. Controis showed recovery of ornithine and all amino acid analyzer standards to be quantitative, except for arginine, which was recovered with approx. 50% efficiency. Iron exchange. Iron exchange from ferrichrome A into deferriferrichrome was followed by mixing [59Fe]ferrichrome A and deferriferrichrome in 0.01 M phosphate buffer, pH 7, at 30°C. Aliquots were removed and rapid separation of ferrichrome and ferrichrome A was accomplished by passing 0.1-ml aliquots of the reaction mixture through a 5 cm column of DEAE-cellulose (OH - form) in a pasteur pipette, followed by water elution of the ferrichrome and scintillation counting. Controls with [59Fe]ferrichrome and [59Fe]ferrichrome A resulted in quantitative recovery of ferrichrome in the eluate and quantitative retention of ferrichrome A on the column. Biosynthesis assay of deferriferrichromes. The principle of this assay is that labelled ornithine added to a culture of U. sphaerogena is rapidly taken up by the cells and incorporated with high yield into deferriferrichrome and deferriferrichrome A, which are excreted back into the medium. At any time, the total counts in the medium are contributed by three species: residual ornithine, deferriferrichrome, and deferriferrichrome A. At neutral pH, these compounds have + 1, 0 and - 3 charges, respectively, and are thus easily separated with ion exchange resins. To a 400 ml iron-deficient culture 150 /~mol [~4Cl]L-ornithine (1.3.107 cpm) was added. At various time intervals, 5-ml aliquots were centrifuged at 27000 × g for 10 min at 4°C. An 0.7-ml

244

aliquot of the supernatant was counted in a scintillation counter. Ornithine was removed from 2.0 ml supernatant by addition of 1.0 ml 2-propanol and 1.0 ml of a thick suspension of A G 50W-X8, 100-200 mesh, H ÷ form. After thorough mixing for 2 rain, the mixture was centrifuged at low speed for 1 min. An aliquot of supernatant was counted. The exact dilution introduced at this step was determined by controls using a solution of blue dextran of known absorbance at 625 nm. The dilution factor was calculated from the decrease in absorbance of the supernatant after treatment with the resin. Deferriferrichrome A was next removed from 0.7 ml of the ornithine-free supernatant by addition of 0.1 ml 1 M Mops (Sigma), pH 7.2, followed by 0.6 ml of a suspension of AG1-X8, 200-400 mesh pre-equilibrated with 0.2 M Mops buffer, pH 7.2. After mixing for 2 rain and centrifugation, 0.7 ml supernatant was counted. Ornithine is calculated from the counts lost in the first resin step,deferriferrichrome A from counts lost in the second resin step, and deferriferrichrome calculated by difference. Controls using added [14C1]deferriferrichrome and [ 14Ci]deferriferrichrome A showed recoveries to be 85-90% quantitative.

Results

Effect of iron on siderophore-ligand biosynthesis. It is generally accepted that iron represses the biosynthesis of the ferrichromes in U. sphaerogena [4], but the nature of repression has not been investigated. In batch culture, reduced growth and increased hydroxamate excretion caused by iron deficiency is detectable within 24 h. The rates of biosynthesis of deferriferrichrome and deferriferrichrome A were examined by accumulation in the medium (Fig. 1), and by the in vivo biosynthesis

2000

/r st i I

,z

1000

,/ 10

O~

I s

2OOO

E} ~000

Q5

04

O~ C

03

200C

Q2

1000

0

I"

i

0

2

4

6

8

h

O0

O

2

4

6

8

Days

Fig. 1. Siderophore-ligand concentration in the medium of U.

sphaerogena. Siderophores were extracted by standard methods [4]. O

O, ferrichrome; A

A , ferrichrome A.

Fig. 2. Relative rates of ferrichrome (O O) and ferrichrome A ( A A ) ligand biosynthesis by U. sphaerogena. The biosynthesis assay described in Methods was used to determine the relative rate of excretion of each ligand into the medium as a function of culture age. a, 6 days; b, 2 days: c, I day.

245 12

TABLE I E F F E C T OF I R O N ON I N T R A C E L L U L A R A M I N O A C I D POOLS

10

Cells were harvested when they reached a density of 2.0 m g / m l . Iron-grown cells reached this density in 33 h while iron-deficient cells required 48 h. A m i n o acid

/~mol amino a c i d / g dry wt. Iron-deficient cells

Iron-grown a cells

5,2 6,9 60.6 181,7 6,0 13,0 121,1 2.6 2.6 1.7 0.9 2.6 + 479.6 12.1 8.7 13.0

7.6 8.0 137.5 125.2 12.8 13.0 141.4 26.6 3.0 1.0 1.0 1.2 + 498.0 19.8 4.5 10.2

×

Asp Thr Set Glu Pro Gly Ala Cys Val Met Leu Ile Tyr Phe Orn Lys His Arg

.0

0

2

4

6

e,

10

h

Fig. 3. Effect of iron on total siderophore-ligand biosynthesis. Biosynthesis assay was used without separation of ferrichrome and ferrichrome A ligands. All effectors were added at 10/~M concentration where indicated by the arrow. O O, chromic deferriferrichrome; A - A , f e r r i c h r o m e A; + + , ferrichrome and × - × , control.

a 50/Lmol iron as FeSO 4. 06 Q5

assay (Fig. 2). Initially, the rates of deferriferrichrome and deferriferrichrome A synthesis were similar. However, as the culture progressed and became increasingly more iron deficient on a per cell basis, deferriferrichrome A was preferentially synthesized. The cellular concentrations of ferrichrome and ferrichrome A (both the iron-free ligand and iron chelate) were determined as a function of the iron concentration in the medium. Ferrichrome was recovered from all cultures at intracellular concentrations of 0.5-0.7 / x m o l / g dry wt. Ferrichrome A was recovered from iron-deficient cells at 0.7 /~mol/g dry wt. N o ferrichrome A was detected in cells grown with 5 # M or more iron

added to the medium and only a trace was detected in cells grown with I#M iron in the medium. A m i n o acid analysis of cell extracts from both

04 E

0.2 01

0

1

2

3 4 5 6 7 Days Fig. 4. Effect o f metals on total hydroxamate biosynthesis. Aliquots o f the medium were assayed for total hydroxamates by the method o f A t k i n et al. [16]. CuSO 4 (0.08 p.M) was added to the copper-deficient medium in the control. The curves for 10 # M added Ga(IIl)citrate, Ga(III)deferriferrichrome and Cr(llI)deferriferrichrome were identical to the control and are not shown. The flask with iron contained 10 a M FeSO 4. O - O, copper-deficient; /', A , control; + + , iron added.

246

iron-deficient and i r o n - s u p p l e m e n t e d cells (Table I) showed high cellular concentrations of the amino acids that are precursors of the ferrichromes. The relative changes of the concentrations of precursor amino acids are no greater than several non-precursors, such as cysteine or proline. Thus, the regulatory mechanism by which iron represses the biosynthesis of the ferrichromes appears not to occur at the level of the amino acids. The in vivo biosynthesis assay was used to determine the time course of repression of biosynthesis. When added either as ferrichrome or ferrichrome A, 10 # M iron inhibited the biosynthesis of the deferriferrichromes 2.5 h after addition (Fig. 3), and biosynthesis completely stopped after 4 h. Effect of other metals. Chromium can be substituted for iron in ferrichrome. The resulting kinetically-inert complex is structurally similar to ferrichrome and is transported into the cell at an identical rate [6]. However, the cell is unable to remove Cr(III) from the ligand, and the complex remains unchanged inside the cell. Chromic deferriferrichrome, at the same concentration as ferrichrome, was completely ineffective in repressing biosynthesis (Fig. 3) or affecting the rate of cell growth (data not shown). Similarly, Ga(III) can be substituted for ferrichrome iron, and the resulting complex is indistinguishable from ferrichrome by the U. sphaerogena transport system [15]. In contrast to chromium, the Ga(III) is kineticaUy labile and also has the thermodynamic stability to displace ferric ion in siderophores and in cellular proteins to a significant degree [15]. Neither Ga(III)deferriferrichrome nor Ga(III)citrate at concentrations of 10 /~M had any effect on siderophore biosynthesis (Fig. 4) or cell growth. Thus, it appears that neither ferrichrome itself nor a ferric ion complex is responsible for repressed biosynthesis. It is interesting that when copper was removed from the growth medium along with iron, the yield of ferrichromes was greatly increased over that from cells grown on normal medium, which contains only 0.08 # M copper (Fig. 4). Cell growth was not affected by copper deficiency. Copper is a constituent of cytochrome oxidase, and the observed effect on biosynthesis of the ferrichromes may be a result of inhibition of electron transport.

16

,2

y

t--

3~D 61D 9'0 1;~D 150 I~D 210 rain Fig. 5. Effect of 10 # M ferrichrome on oxygen uptake. An iron-free 4-day culture was diluted in the medium supernatant from an identical culture to a final absorbance of 0.75 at 650 nm, and a final volume of 400 ml. The culture was maintained under standard culturing conditions while 4 ml aliquots were removed at various intervals, saturated with air, and the rate of oxygen uptake was measured with an oxygen electrode equipped with a chart recorder. Slopes were taken during the linear portion of the curve and reported as • u p t a k e / m i n where 100% represents air-saturated medium.

lO

I

~

2 h

½

4

Fig. 6. Effect of siderophores on ornithine transport. [14C t ]Lornithine, (20/~mol, 2.9- 105 cpm) was added to 50 ml of a 24-h iron-deficient culture. At various intervals, 0.5-ml aliquots were removed, centrifuged and the medium supernatant counted in a scintillation counter. Ferrichrome and ferrichrome A at a final concentration of 10 #M were added where indicated by the arrow. ferrichrome; a A, ferrichrome A; + + , control.

0

O,

247 3O(

2O(

O0

8

16

24

h

Fig. 7. Exchange of iron from ferrichrome A into deferriferrichrome. At time zero, deferriferrichrome was added to an equimolar concentration of [59Fe]ferrichrome A 080/~M final concentration, 3 ml vol. specific activity 1.5. l05 epm//~mol) in 0.01 M phosphate buffer, pH 7 at 30°C. Aliquots were removed and ferrichrome and ferrichrome A were separated by DEAE-cellulose chromatography.

Effect of siderophores on respiration and ornithine uptake. Ferrichrome at a concentration of 10/~M significantly increased cellular respiration within 10 min of addition to an iron-deficient culture (Fig. 5). The rate of [14Cl]ornithine transport was also greatly increased within 30 min after the addition of 10/~M ferrichrome or ferrichrome A (Fig. 6). Uptake of ferrichrome A. It is well established that ferrichrome functions as a true ferric ionophore in U. sphaerogena, and that ferrichrome A only shuttles the iron to the cell membrane. The cell takes up the iron from ferrichrome A and the ligand remains extracellular [11]. The rate of uptake of iron from ferrichrome A is considerably slower than the rate of ferrichrome uptake. We investigated the possibility that the mechanism of iron uptake from ferrichrome A is by exchange of the iron into deferriferrichrome to yield ferrichrome, a true ionophore. The rate of in vitro exchange of iron from ferrichrome A into deferriferrichrome was measured. The initial rate of exchange was approx. 0.3 n m o l / m i n per ml (Fig. 7). The rate of iron uptake from ferrichrome A in a typical experiment, (17 m g / m l cell concentration; 36/~M ferrichrome A concentration) is approx. 0.3 n m o l / m i n per ml. Although the concentrations used in the in vitro exchange experiment were 5-times that of an uptake experiment, exchange could not be definitely ruled out for ferrichrome

A-iron uptake, provided that the cell excretes sufficient deferriferrichrome in short time periods. However, in control experiments the amount of total hydroxamates excreted in 1 h was less than 3 # M , which is far too low to account for a quantitative exchange mechanism. In addition, the rate of ferrichrome A-iron uptake was measured in the presence of a 5-fold excess of deferriferrichrome. No significant increase in the rate of ferrichrome A-iron uptake was observed. The rate of exchange of ferrichrome A-iron into deferriferrichrome was also measured in the presence of cytoplasm and membrane fractions. Under the same conditions as in an uptake experiment, no increase in the rate of exchange was observed over controls with boiled cytoplasm and membrane fractions. The ratio of binding constants for ferrichrome and ferrichrome A can be calculated from measurements of their respective concentrations at equilibrium (Fig. 7). With equimolar ferrichrome A and deferriferrichrome, 30% of the ferrichrome A-iron exchanged into deferriferrichrome as determined by DEAE chromatography and electrophoresis. When this value is substituted in the equation: Kfc _ (ferrichrome)(deferriferrichrome A) Kf~^ ( ferrichrome A) (deferriferrichrome) the calculated ratio of binding constants of 0.184 is obtained. The value of K~c as determined by Anderegg et al. [17] is 1.2. 10 29. Using this value and our exchange data, the calculated binding constant for ferrichrome A is 6.4- 10 29, which is in good agreement with the value determined by Neilands (personal communication).

Intracellular distribution of ferrichrome A -derived iron. Cells were incubated with [59Fe]ferrichrome A and subjected to homogenization and fractionation in the presence of ferrozine to trap nonsiderophore ferrous iron and prevent reoxidation and chelation by intracellular deferrisiderophores. Controls showed that the concentrations of ferrozine used did not result in removal of iron from ferrichrome or ferrichrome A nor did deferrisiderophores remove ferrozine iron. The following cell-free fractions were examined for the presence of ferrichrome A-derived iron: insoluble cellular

248 debris, soluble iron as siderophores, soluble ferrous iron (ferrozine complex), mitochondria and dialyzable iron. Of the total iron taken up by the cells, greater than 90% was recovered. 20 min after incubation and for the duration of the experiment, about 24% of the radioactive iron was found in cellular debris. Homogenization using a French pressure cell resulted in a value of 15 % associated with the cellular debris. Of the 71% of the supernatant counts, less than 5% were extractable into benzyl alcohol as ferrichrome A (1.1%) or ferrozine (3.4%). No counts were recovered as ferrichrome. Of the remaining supernatant counts, 36% were found associated with the mitochondrial fraction (probably as heme), 20% was dialyzable and 15% was non-dialyzable. The rate of incorporation into mitochondria was similar to that when the iron was supplied as ferrichrome [12]. Discussion Why should a microorganism excrete two ironchelating agents, only one of which serves as a true ionphore? The data presented in this paper suggest that a second, more powerful chelator is excreted only when the true ionophore fails to sequester sufficient iron. Ferrichrome A has a 2-fold advantage over the ionophore, ferrichrome. First, it is a thermodynamically stronger chelator of iron and is thus more effective in competing for iron with trihydroxamic acids produced by other microorganisms. Second, it is a tricarboxylic acid which lowers extracellular p H to aid in solubilization of the metal. Lowering of p H is a mechanism well known to occur in higher plants to solubilize iron [18], and we have observed the p H of fungal cultures drop as much as 2 p H units. This phenomenon of switching siderophores under increased iron stress is not unique to U. spaerogena. As cultures of F. roseum progress and become more iron deficient on a cellular basis, deferrimalonichrome replaces fusarinines in the medium [5]. Malonichrome is similar to ferrichrome A in that the three acyl groups are carboxylic acids, the binding constant for iron is very high, and the compound functions poorly as an ionophore in F. roseum. Acidic siderophores are far more effective in solubilizing ferric oxides than are neutral siderophores, such as ferrichrome or triacetylfusarinine.

Partial repression of siderophore biosynthesis by iron augments this phenomenon of differential synthesis. When 5/~M iron is added to U. sphaerogena growth medium, no detectable amount of ferrichrome A can be found either in the medium or intracellularly. However, the intracellular concentration of ferrichrome remains at 0.5-0.7 /~mol/g dry wt. even with the addition of 50 ~M iron to the growth medium. Regardless of extracellular iron concentration, a neutral ionophore such as ferrichrome is probably still necessary to carry ferric ions across lipid membranes. U. sphaerogena has an exaggerated iron requirement, and the importance of siderophore biosynthesis is reflected by the cellular ornithine pool, which is extraordinarily high and far exceeds the concentration of most other amino acids (Table I). Serine is a constituent of ferrichromeA and a precursor of glycine, a constituent of both ferrichrome and ferrichrome A. The cellular serine concentration is also notably high, as is glutamic acid, an ornithine precursor. The only other amino acid that is exceptionally high is alanine. Although alanine itself is not a constituent of the ferrichromes, it is in equilibrium with pyruvate, the latter being a precursor of acetate and methylglutaconate which serve as the acyl groups of ferrichrome and ferrichromeA, respectively. It should be noted that under iron-deficient conditions of growth, the total yield of the ferrichromes m a y exceed the dry weight of the cells, indicating that siderophore biosynthesis has become the major pathway of cellular metabolism. The mechanism by which iron represses siderophore biosynthesis is unknown. Neilands suggested that the iron is removed from the siderophore complex before acting as a repressor, either directly or indirectly [19]. Our results tend to support that hypothesis. Gallium and chromium form complexes with siderophore ligands that are so analogous to ferrichrome as to completely trick the highly specific active transport system, yet these derivatives of ferrichrome have no repressive effect on hydroxamate synthesis in U. sphaerogena. The repressive effect of iron may be indirect. Within minutes after the addition of ferrichrome, U. sphaerogena responds with physiological changes such as increased oxygen uptake (Fig. 5). The regulation of siderophore biosynthesis may be

249 closely related to cellular oxidative metabolism. A deficiency of iron m a y lead to increased levels of reduced pyridine nucleotides and increased partial pressure of oxygen, b o t h of which would enhance the committed step of siderophore biosynthesis, the hydroxylation of ornithine. This is compatible with our observation that copper deficiency also results in greatly increased siderophore synthesis since copper is a constitutent of c y t o c h r o m e oxidase. Although the rate of in vitro exchange of iron f r o m ferrichrome A into equimolar deferriferrichr o m e is quite rapid (Fig. 7), our data indicate that this is not the mechanism by which ferrichrome A-iron enters the cell. The a m o u n t of deferriferrichrome excreted by cells in the course of an uptake experiment is far too low to allow significant exchange, nor does addition of deferriferrichrome increase the rate of ferrichrome A-iron uptake. Furthermore, iron added as ferrichrome A is distributed m u c h more rapidly a m o n g nonsiderophore cellular fractions than is an equivalent a m o u n t of metal entering the cell as ferrichrome [12]. W h e n iron is added as ferrichromeA, no detectable intracellular ferrichrome is found, and less than 5% is contained in ferrichrome A. This is not unexpected if the metal is removed from the strong chelate prior to its entry into the cell. We are currently investigating the mechanism of ferrichrome A-iron transport. Preliminary kinetic studies indicate that ferrichrome A-iron transport exhibits saturation kinetics, which is indicative of a specific transport system.

Acknowledgment This work was supported by N I H Research G r a n t AI09580.

References I Neilands, J.B. (1974) in Microbial Iron Metabolism (Neilands, J.B., ed.), pp. 3-34, Academic Press, New York 2 Keller-Schierlein, W., Prelog, V. and Z~ihner, H. (1964) Prog. Chem. Org. Natl. Prod. 22, 279-322 3 Lankford, C.E. (1973) Crit. Rev. Microbiol. 2, 273-331 4 Emery, T. (1966) Biochemistry 5, 3694-3701 5 Emery, T. (1980) Biochim. Biophys. Acta 629, 382-390 6 Raymond, K.N. and Carrano, C.J. (1979) Acc. Chem. Res. 12, 183-190 7 Emery, T.F. (1965) Biochemistry 4, 1410-1417 8 Snow, G.A. (1965) Biochem. J. 97, 166-175 9 Atkin, C.L. and Neilands, J.B. (1968) Biochemistry 7, 3734 -3739 10 Muthukrishnan, S., Padmanaban, G. and Sarma, P.S. (1969) J. Biol. Chem. 244, 4241-4246 11 Emery, T. (1971) Biochemistry 10, 1483-1488 12 Straka, J.G. (1978) Dissertation, Utah State University, Logan, UT 13 Schatz, G., Penefsky, H. and Racker, E. (1967) J. Biol. Chem. 242, 2552-2560 14 Leong, J. and Raymond, K.N. (1974) J. Am. Chem. Soc. 96, 6628-6630 15 Emery,T. and Hoffer, P.B. (1980) J. Nucl. Med. 21,935-939 16 Atkin, C.L., Neilands, J.B. and Phaff, H.J. (1970) J. Bacteriol. 103, 722-733 17 Anderegg, G., L'Eplatterfier, F. and Schwarzenbach, G., (1963) Helv. Chim. Acta 46, 1409-1422 18 Olsen, R.A. and Brown, J.C. (I 980) J. Plant Nutr. 2, 629-645 19 Neilands, J.B. (1977) in Adv. in Chemistry Series (Raymond, K.N., ed.), p. 5, Am. Chem. Soc., Washington, D.C.