iron mobilization from ferritin by chelating agents

Iron Mobilization from Ferritin Iby Chelating Agents

Robert R. Crichton, Francoise Roman, and Fran&e Roland Unite de Biochimie. Universite Catholique de Louvain

ABSTRACT

The release of iron from horse spleen ferritin by the chelating agents desfenioxamine B, rhodotoruhc acid, 2,3+iihydroxybenzoate, 2,2’-bipyridyl and pyridme2aldehyde-2- pyridyl hydrazone (Paphy) has been studied in vitro. Fenitin prepared by classical procedures involving thermal denaturation releases its iron less effectively than ferritin isolated by a modified procedure that avoids this step. Desferrioxamiue B and rhodotorulic acid are the most effective in releasing iron from both preparations of ferritin. When FMN is added, iron release by desferrioxamine B, rhodotorulic acid, and 2,3dihydroxy- benzoate was effectiveIy blocked, whereas both bipyridyl and Paphy showed a marked simuIation. A substantial increase in iron release was also observed for bipyridyl and Paphy with ascorbate; a less important increase was noted for rhodotorulic acid. EDTA exerted a marked inhibition of iron release from ferritin with rhodotorulic acid, 2,3di- hydroxybenzoate, bipyridyl, and Paphy. The effects of citrate and oxalate on iron release by the chelators was small. The effect of the concentration of flavin on iron release from ferritin by bipyridyl and desferrioxamine B have been studied. Desferrioxamine is unable to mobilize Fe11 from ferritin following reduction by reduced FMN, whereas bipyridyl can rapidly complex the ferrous iron. The results are discussed in the context of our current concepts of storage iron mobilization in the treatment of iron overload.

The storage of iron within cells in a soIubIe, nontoxic form is assured by ferritin in a large number of eukaryotic organisms (reviewed in Refs. 1 and 2). The iron is de- posited in the interior of an approximately spherical protein shell, apofenitin, predominantly as ferric oxyhydroxide FeO-OH together with some ferric phosphate_ Storage iron is released from ferritin and made available within the cell for the synthesis of hem and nonhem iron-containing proteins. It can also be released for incor-

poration into the extracelhdar iron pool of transferrin for distribution to other cells. The mechanisms involved in storage iron release have been studied both in simple

Ad&m reprinf requests fo: R. R. ~richton, unite de BhJchimie, i&river&e &thO~qtre de ~ouvain, ace L.

Pasteur, 1, B-1348 Louvain-la-Neuve, Belgium.

Journal of Inorganic Biochemirtry 13,305-3 16 (1980) 0 Elsevier North Holland, Inc., 1980 52 Vanderbilt Ave., New York, New York 10017

305 0162-0134/‘80/080305-12502.25

306 Robert R. Crichton et al.

well-defined biochemical systems using ferritin together with putative iron mobilizing systems, and in cellular and whole animal systems (reviewed in Refs. 2 and 3).

A number of alternative molecuku mechanisms can be envisaged for fenitin iron re- lease_ The iron could be leached out of the molecule by low-molecular-weight cheiitiug agents (specific for ferric iron), of which a number of potential candidates exist in the cell. Such systems have in general been found to release rather small amounts of iron in model studies [4-6] and do not seem likely to play a major roleiuvivo. Alternative pathways of iron release involving enzymes such as xanthine oxidase [7] or a ferriduc- tase [8] have been proposed, but at present it seems most likely that ferritin iron is reIeased by reduction of ferric iron to the ferrous form, in a reaction mediated by reduced flavins followed by subsequent complex&on of the ferrous iron by an as yet unidentified biological complexing agent [3,9-l l] _

However, the cheIation of ferritin iron by appropriate compIexants has a considerable potential in the treatment of disorders of iron metabolism characterized by secondary iron overIoad, such as Cooley’s anemia. The consequence of proIonged blood transfusion are an accumulation of excessive amounts of iron in the organism, initially ‘J the liver and subsequently in other tissues (reviewed in Refs. 12 and 13). The initial phase of iron accumulation is in ferritin; thereafter hemosiderin deposits are found, predominantly in lysosomes [ 13]_

It is of considerable therapeutic interest that this iron be released from the tissues in a form that can be readily eliminated via the bile or the kidneys. For this reason a number of potential iron chelators have been tested in whole animal studies and in cell culture with a view to their use in clinical practice to reduce the finally toxic effects of iron overload (reviewed in Refs. 14-20).

Since the ultimate goal of these studies is the mobilization of ferritin (and hemosiderin) iron, we have focused our attention on the capacity of a number of iron chelators currently in clinical use, or in clinical trials, namely desferrioxamine B and rhodotoruhc acid, which are both trihydroxamic acids and 2,3dihydroxybeuzoate, as well as 2,2’-bipyridyl and pyridine-2-aldehyde-2-pyridyl hydrazone (Paphy) to release iron from ferritin in vitro. Our rationale was that if a potential iron chelator could mobilize ferritin iron directly, it would be a good candidate for further study in a suitable cellular system, such as that described in [21] _

MATERIALS AND METHODS

Mops (morpholine propane sulphonic acid) was from Serva, Heidelberg, BRD. Des- ferrioxamine B was a gift from CIBA-GEIGY, Basle, Switzerland, 2,3dihydroxyben- zoate was obtained from Aldrich, Beer-se, Belgium, 2,2’-bipyridyl was from Fluka, Buchs, Switzerland, Paphy (pyridine-2-aldehyde-2-pyridyl hydrazone) was synthesized by A_ Cordy, C. Wiaux-Zamir, and L Ghosez (Unite de Chimie Organique de Synthese, Universite Catholique de Louvain) or obtained from Aidrich, Beerse, Belgium, and rhodotoruhc acia was a gift from Dr. A. Cerami, Rockefeller University, New York, USA FMN was from Serva, Heidelberg, BRD, and NADH was from Boehringer, Mann- heim, BRD.

Horse spleen ferritin was prepared by two procedures. The first was as described in Ref. 22 invohing thermal denaturation at 70°-80°C, followed by ammonium sulphate precipitation and crystallization from CdS04. The second method invoIved acidifica- tion of minced tissue at pH 4.75 with acetic acid, followed by incubation at 4°C for

Iron Mobilization from Ferritin 307

12 hr and centrifugation at 5000 g to clarify the suspension. Thereafter ammonium sulphate precipitation (35% w/v) was carried out and the purification was completed by three successive CdS04 crystallizations. The purity of both fenitin preparations was established by conventional gel electrophoresis in denaturing (SDS) and nonde- naturing media and by amino acid analysis. Samples were hydrolyzed with 6N HCI at 1 10°C for 16 hr and their amino acid concentration determined using a Locarte amino acid analyzer (Locarte Co., London, UK). The protein concentration was determined from amino acid analysis using a value of 10 residues of glycine/polypeptide chain (molecular weight 18,500). The iron content of the fenitin was determined by the method of Hill [23] as modified by Drysdaie and Munro [24] _ To I ml of a solution of ferritin (diluted in H20) 0.8 ml of 1 N HCI and 3.0 ml Hz0 were added and iucu- bated 20 min at 100°C. One milliliter of a 1.5 M solution of Na2S03 and 0.5 ml of a 0.1% w/v solution of 2,2’-bipyridyl in 60% v/v acetic acid were added, the volume adjusted to 10 ml and incubated for 30 min at 100°C. After cooling, the optical density at 520 nm was measured_ A calibration curve was established using a standard solution of 5Opg of FeCls/ml in 60% acetic acid. The ferritin preparations used had an iron content of 2230 g atoms/molecule.

The h maximum for the various iron chelators was determined using ferrous ammonium sulphate (British Drug Houses, Poole, UK) and ferric chloride (Merck, Darm- stadt, BBD) in 200 mM Mops buffer, pH 7.4, with an excess of chelator in the range 350-600 nm using a Beckmarm model 25 spectrophotometer (Analis, Namur, Belgium). The molar extinction coefficients for Fe II and Fe III were determined at the appropriate h,,, for each chelator_ TabIe 1 presents these data together with the complexation constants of the different chelators for Fe HI and Fe II taken from the literature_

Iron release from ferritin was followed by incubation of ferritin in 200 mM Mops buffer, pH 7.4, at a final concentration of IO- 6 M in protein together with the chelator in 200 mM Mops buffer, pH 7.4, at a final concentration of 1 or 20 mM at 37°C in a water bath. The samples, in a final volume of 1.0 ml in disposable cuvettes @&tell, Milan, Italy) were removed from the water bath at appropriate time intervals and the amount of iron released determined spectrometricahy at the appropriate h,,,. Five samples were measured at each time interval and the mean + standard error was de-

TABLE 1. Extinction Coefficients and Compelxation Constants for Ferrous and Ferric Iron of the Chelators Employed in This Study. The h,,, and E Were Determined at pH 7.4, As Described in the TextO

Chelator Fe II Fe III

h max E h max E

nm 1 mol--l cm-1 IogK nm 1 mol-1 cm-l Iog K

Desferrioxamine B 428 2,770 11.0 428 2,770 30.6

2,2’-Bipyridyl 520 8,400 17.0 - 16.3

Rhodotorulic acid - - - 480 1,800 28.3 Paphy 520 4,900 12.3 - - -

2.3-Dihydroxybenzoate 540 2,185 - 540 1,890 3.6 EDTA - - 14.3 - - 25.1

4 The complexation constants are taken from Refs. 36-41_


termined. In a second series of experiments, ferritin was incubated with the chelators - as described above, alone or in the presence of a final concentration of 1 mM in citrate,

EDTA, FMN, ascorbate, or oxalate_ The results in all cases are expressed as atoms of iron released/molecule of ferritin. Appropriate controls were carried out using FMN in buffer, FMN, buffer and chelator and FMN, buffer and ferritin, in the case of bipyridyl and Paphy, since the reduced flavin absorbed significantly at 520 nm. The effect of FMN concentration on iron release by desferrioxamine B was studied as described above in the range (final concentration) of l-0.01 mM in 200 mM Mops buffer, pH 7-4. For bipyridyl, the effect of FMN was examined in the range l-0.001 mM using200 rnM sodium phosphate buffer, pH 7.4. The values obtained for iron release in this buffer for control samples containing bipyridyl alone were similar to those in the Mops buffer.

RESULTS AND DISCU!ISION

In all of the experiments presented here, we have used horse spleen ferritm prepared by the classical procedure of Granick [22] and by a modification of this procedure (IMaterials and Methods) in which the thermal denaturation step is replaced by acid& cation of the minced tissue to pH 4.75 followed directly by ammonium sulphate precipitation- The difference between these two preparations is that in the modified procedure a ferritin preparation is obtained that, in contrast to that obtained by the classic Granick procedure [22] does not contain a fluorescent, UV absorbing material that has the characteristics of a lumichrome [25]. We assume, on the basis of preliminary characterization of this substance, that it is a degradation product of a flavin derivative (J_ C. Mareschal and R. R Crichton, unpublished observations). The results obtained with both preparations were qualitatively the same; however, in almost aii cases, the amount of iron released from ferritin prepared by the modified procedure was substantially greater than from the fenitin prepared by the Granick method [22]. Figures Ia and lb, which present the results for the five chelators studied with the two ferritin preparations, illustrate this point. The amount of iron released after 6 hr from the ferritin prepared by the modified procedure, was respectively, three and four times greater with desferrioxamine B and rhodotorulic acid than from that prepared by the standard method. For dihydroxybenzoate, Paphy, and bipyridyl, the differences between the two preparations were less pronounced (and indeed for 1 mM bipyridyl, the difference was within the experimental error)_ Rhodotorulic acid and desferrioxamine B were the most effective complexants of fenitin iron, followed by Paphy and 2,3-dihydroxybenzoate; bipyridyl at 1 and 20 mM mobilized very small amounts of iron in both samples of ferritin. This point is further substantiated in Table 2, which shows that desferrioxamine B and rhodotorulic acid released three times as much iron after 24 hr incubation than Paphy, and six times as much iron as 2,3dihydroxybenzoate at the same concentration (1 mM). In the case of rhodotorulic acid and desferrioxamine B, there is apparently a much more rapid release of iron in the course of the first hour (Figure lb) compared to subsequent time intervals. For the two hydroxamic derivatives this may correspond to the readily mobilizable iron described previously [26,27] , which was attributed to Fe II located at surface sites on the ferritin molecule.

With 2,3-dihydro=ybenzoate, we note that a plateau is rapidly attained with ferritin prepared by the modified procedure; after 6 hr 355 atoms/molecule have been mobiied (Figure 1 b) and the value shows very little increase after 24 hr (Table 2).

In the study of iron release by desferrioxamine B from transferrin 128,291 it was found that a number of low-molecular-weight mediators could increase the rate of iron


1x

1x

1oc

70

Lc

10

(a) (b)

FIGURE 1. iron mobilization from ferritin by chelating agents. Ferritin at a final concentration of 1.14 X 10-S M was incubated at 37°C in Mops buffer, 0.2 M, pH 7.4 containing the chelators at the final concentration given: 1. Paphy, 1 MM; 2. Bipyridyl, 20 mM; 3. Bipyridyl, 1 mM; 4. Des- ferrioxamine B, 1 mM; 5. Rhodotorulic acid, 1 mM; 6.2,3-Dihydroxybenzoate, 1 m!!. (a) gives the re.sults obtained with the ferritin prepared by the Granick procedure and (b) for the fen-&in prepared by the modifii method The results, as for Figures 24 are the mean of tive independent determinations; in no case was the standard error greater than t6% of the mean.

TABLE 2_ Iron Release from Ferritin (Merck) after 24 hr Incubation_ The Amount of Iron Release After 24 hr Incubation at 37OC of 1.14 X 10-6 M Ferritin (Merck) in 0.2 M Mops Buffer, pH 7.4, by the Different Chelators at the Final Concentration Given in the Table Was Measured at the Appropriate Maximum of the Chelator for Five Samples. The Results Are Expressed as atoms of Iron Released/Molecule of Ferritin 2 the Standard Error of the Mean

CheIator Concentration

(df) atoms of Fe Released/_Molecule

Paphy 1 85.4 i 10.7 Jiipyridyl 1 17.6 f 3.9 Bipyridyl 20 41.8 c 5-9 Desferrioxamine B 1 225.3 i 6.8 Rhodotorulic acid 1 2495 * 10.4 2,3_Dihydroxybenzoate 1 41.7 2 3.6


transfer_ It was suggested that the mediators displaced the bicarbonate bound adjacent to the iron and thus facilitated its transfer to desferrioxamine B. In order to establish whether such mediators could influence iron release from ferritin, we studied the effects of 1 mM (final concentration ) citrate, EDTA, FMN, ascorbate, and oxalate on iron mobilization by the five chelators described above.

For desferrioxamfne B (Figure 2) a slight positive effect was observed with ascorbate, and a more important increase with EDTA corresponding to 30% after 6 hr. With FMN a marked inhibition was observed_ After release of 11-12 g atomsfmolecule, no further iron was released even after 24 hr incubation (Table 3). The effects of citrate and oxalate were not significant. The effect of FMN was light-dependent, and was almost totally eliminated at concentrations of FMN of 0.05 mM or less (Table 3).

Iron release by rhodotorulic acid (Figure 3) as also stimulated by ascorbate, and again citrate and oxalate had no effect. However, EDTA exerted a marked inhibition of iron release; after 6 hr iron mobilization was reduced to 50% of the control. Once again FMN blocked iron release. Some 20-21 atoms/molecule were rapidly mobilized and thereafter no increase was observed_

With bipyricl;rl (Figure 4) no effect was observed with citrate and oxalate. EDTA inhibited iron release completely and ascorbate increased iron_mobilization considerably (sevenfold in 6 Ix). With FMN we also observed a large positive effect that was light dependent. Tbis effect was difficult to quantify, since photoreduction of FMN con- tributed to the optical density measured at 520 run. We were therefore led to use a series of controls that included the buffer alone with FMN, the buffer with FMN and bipyridyl, and the buffer with FMN and ferritin (Table 4). We were able to reduce the vatiations in absorption of the controls by carrying out the experiments in 200 mM sodium phosphate buffer, pH 7.4. The effect of FMN concentration on the mobilization of ferritin iron by bipyridyl was therefore studied in phosphate buffer (Table 3). Comparable results were obtained in Mops. Iron mobilization was similar for FMN concentrations of l-0.01 mM. Below this value iron release was less rapid, and at 0.001 mM the amount of iron released differed only slightly from that observed in the ab- sence of FMN_

The results obtained width Paphy were qualitatively very similar to those obtained with bipyridyl; citrate and oxalate had little or no effect, ascorbate and FMN increased iron release by 120% and %, respectively, and EDTA completely abolished iron release. For 2,3dihydroxybenzoate, a slight inhibition was observed with citrate, FMN, and ascorbate; oxaloacetate had no effect and EDTA inhibited iron release by 55%.

According to the chelator used, the results obtained fall into three general classes. The hydroxamic acid derivatives, desferrioxamine B, and rhodotorulic acid are considerably inhibited by FMN at a concentration of 1 mM. For desferrioxamine B the effect of FMN is light dependent and is almost completely eliminated at concentrations of 0.05 mM and below- To explain this effect we might suggest that reduced FMN competes with these bulky chelators for the sites on the protein from which the iron can be mobilized,which by analogy with our model for iron deposition in ferritin would be the peroxo-bridged iron at the sites of iron oxidation 1301. However, equi- librium dialysis studies Ill, 3 l] do not support the view that FMN or FMNHa binds to ferritin to an appreciable extent. The effects of ascorbate and EDTA suggest that desferrioxamin e B and rhodotorulic acid do not have easy access to ferritin iron. The stimulation by ascorbate could be explained by reduction of the ferritin iron and its release as an Fe II-ascorbate complex, which subsequently transfers its iron to the


140-

120-

loo-

80-

60-

40-

atoms Fe released / / ferritin molecule

6

2 3 4 5 6 t(h)

FIGURE 2. Ferritin iron release by desferrioxamine B-effect of mediators. Thk cbnditions were the same as those of Figure lb, i.e., ferritin (l-14 X 10-s M) in 0.2 M Mops buffer, PI-I 7.4, with desferrioxaroine B (1 r&f). 1. No addition; 2. Citrate, 1 mM; 3. EDTA, 1 r&f; 4. FMN, 1 mhi; 5. Ascorbate, 1 mM; 6. Oxalate, 1 mM.

TABLE 3. Effect of FMN Concentration on the Release of Iron from Ferritin by Desferrioxamine B and Bipyridyl, The Rest&s, Expressed as atoms of Iron Released/Molecule of Fenitin Are the Mean of Five Independent Measurements. The Experiments Were Carried Out at 37OC in 200 mM

Mops Buffer, pH 7.4, for Desferrioxamine B and in 200 mM Sodium Phosphate Buffer for Bipyridyl.

Iron Released (atoms/molecule) FIMN Concentration after

Chelator (nlM) 2hr shr 24 hr

Desferrioxamine B 1.0 11-7 12.3 12.6 0.5 10.0 10.3 - 0.05 61.3 112.0 - 0.01 50.9 93.0 - 0.00 62.7 105.6 225.3

Bipyridyl 1.0 15.8 41.3 116.5 0.1 17.6 38.9 118.2 0.01 12.0 25.3 70.3 0.001 5.0 12.9 46.9 0.00 3.4 8.9 41.8


6C

4c

2a

atoms Fe released/

ferritin molecule

1 2 3 -4 5 6 t(h)

FIGURE 3. Fenitin iron release by rhodotorulic acid-effect of mediators. The conditions were the same as for Figure 2 with 1 mM rhodotorulic acid in place of desferrioxamiue B.

hydroxamic acid; in view of the greater stability of the Fe III complex, it is possible that the iron is subsequently oxidized_ EDTA could exert its effect by chelating the ferritin iron and thereafter the transfer to the hydroxamic acid would be determined by the relative stability constants of the Fe-EDTA and Fe-hydroxamate, and by the exchange kinetics. In the case of desferrioxamin e B the exchange occurs and an enhanced iron release is observed, whereas EDTA can effectively compete with rhodotorulic acid for the ferritin iron and the exchange is sufficiently slow to result in a net inhibition of iron release by the rhodotorulic acid. An alternative explanation for the effect of ascorbate would be that it more readily complexes ferritin iron and then ex- changes its ircn with the hydroxamate derivatives. The enhancement of storage iron mobilization by desferrioxamin e B in the presence of ascorbate has been reported in whole animal and in clinical studies [32, 331 and the present results may serve to explain at least in part this effect at the level of fen-&in iron mobilization.

The second group of chelators comprises bipyridyl and Paphy. Both are chelators of Fe II and, in the case of bipyridyl, to a lesser extent of Fe III, although only the pink-colored complex of bipyridyl with Fe II could be detected. As might be ex- pected, both ascorbate and photoreduced FMN increase the amount of iron released,


atoms Fe released / ferritin molecule

loo-

60-

40-

20-

FIGURE 4. Ferritin iron release by bipyridyl-effect of mediators_ The conditions are the same as for Figure 2 with 20 mM bipyridyl in place of desferrioxamine B.

although the effect is much more marked for bipyridyl than for Paphy. The effect of FMN is entirely dependent on photoreduction and, for bipyridyl, remains demon- strable down to concentrations that approach those found, for example, in rat liver, namely 15-18 m (34). However, most of the flavins in tissues are protein-bound and no not readily diffuse; further, while a number of bacterial NAD(P)H/FMN oxido- reductases are known [ 1 l] their equivalent in mammalian systems is less well established.

TABLE 4. Fen-&in Iron Release by Bipyridyl in the Presence of FMN. The Composition of Each Sample and theAsP Ilm Values After 30 min and 6 hr Incubation at 37OC Are Given. The Final Concentrations Were 200 mM in Mops or Sodium Phosphate Buffer, pH 7.4,l mM in FMN, 5 mM in Bipyridyl, and 1 ,YM in Ferritin.

A520 nm Sample 30 min 6 hr

I Mops, FMN 0.146 f 0.002 0.311 f 0.002 II Mops, bipyridyl, FMN 0.196 f 0.008 0.215 f 0.015 III Mops, ferritin, FMN 0.277 f 0.007 0.477 f 0.005 IV Mops, ferritin, bipyridyl, FMN 0.403 f 0.013 0.602 r 0.013 V Phosphate, FMN 0.170 + 0.013 0.173 f 0.014 VI Phosphate, bipyridyl, FMN 0.207 f 0.006 0.232 i 0.014 VII Phosphate, fenitin, FMN 0.354 f 0.014 0.364 i 0.012 VIlI Phosphate, ferritin, bipyridyl, FMN 0.396 k 0.005 0612 i 0.013


The mobilization of ferritin iron by 2,3dihydroxybenzoate is inhiiited by all of the mediators with the exception of oxalate. The greatest inhibition is observed with EDTA, which in view of its much greater complexation constant can presumably compete not only for the ferritin iron, but also can remove iron from its complex with 2,3- dihydroxybenzoate. Both citrate and ascorbate diminkh the total amount of iron mobilized by 2,3dihydroxybenzoate, but do not appear to alter the rate at which the iron is released. The most surprising observation is with Fh4N, where, after an initial burst of iron release, corresponding to 23 g atoms/molecuIe, there is a total inhibition of subsequent iron mobilization. We concludeYentatively that the effect of ascorbate, EDTA, and citrate are due to exchange of iron from the 2,3dihydroxybenzoate complex to the mediators, whereas the effect of Fh4N is analogous to that observed with desferrioxamine B and rhodotorulic acid.

One explanation for the effect of FhIN is that the reduced flavin reacts with the peroxo bridge of the catalytic site to generate a flavin hydroperoxide and reduced iron on the sites. If the chelator used it too bulky to have access to the Fe II (due to ob- struction by the flavin) it will not be able to release the iron (desferrioxamine B and rhodotorulic acid). If the complexation constant for Fe II is not in excess of that of the sites (2,3dihydrozybenzoate) it will again be unable to remove the iron. In the

FIGURE 5. Iron mobilization from ferritin by desferrioxamiue B in the presence of FMN/NADH. Ferritin at a final concentration of 0.24 X Ifi M, containing 2230 atoms of iron/molecule was incubated at 20°C in tris-HCl buffer, 0.2 M, pH 7.4 containing 1 mM desferrioxamine B, 4 mM FMN, and 5 mhf NADH. After 1 hr bipyridyl was added at a final concentration of 5 mhf. iron release by desferrioxamine B was measured at 428 nm and bipyridyl at 520 mn.

SQa

MO

300

200

100


case of bipyridyl and Paphy, the chelators can not only complex the Fe II, but may also be oxidized by the flavin hydroperoxide to generate the reduced flavin and thus enable a subsequent catalytic cycle to begin. That ferritin in the presence of flavins is able to oxidize both bipyridyl to its N-oxide and triphenylphosphine to its oxide has been established [30, and A. Crutzen, H. Bonnemann, and R R Crichton, in preparation).

We have previously shown that NADH can reduce FMN and readily release ferritin iron as Fe II, which can be complexed by a suitable chelator such as bipyridyl [30,10, 25]. FMN, NADH, and bipyridyl release ferritin iron as the Fe II-bipyridyl complex after a lag phase which corresponds to the consumption of the dissolved oxygen in the medium [25]. When ferritin is incubated under the same conditions, but with desferrioxamine B in place of bipyridyl (Figure 5), very little iron is released over the period of 1 hr. If we now add bipyridyl at a final concentration of 5 mM we observe an immediate release of iron, which corresponds to somewhat more than is released during a 15-min incubation of fenitin with FMN, NADH, and bipyridyl alone. Des- ferrioxamine B is a chelator of Fe III, but also of Fe II (the values for the complexation constants are 1031 and loll, respectively); when incubated with Fe II it chelates it immediately to give a stable colored complex that does not exchange its iron with bi-

pyridyl even after 1 week of incubation at 20°C. The conclusion seems inescapable that in the presence of reduced flavin the desferrioxamine does not have access to the fenitin iron. In contrast, bipyridyl can readily remove the iron as Fe II from the ferritin molecule in the presence of reduced flavin.

In conclusion, we have found that of the iron chelators used in clinical practice, desferrioxamine B and rhodotorulic acid are effective in iron mobilization from fer-

&in, but are unable to release iron from ferritin in the presence of 1 mM FMN. In contrast, 2,3dihydroxybenzoate is a poor chelator of ferritin iron and is also inhibited by FMN. The Fe II chelators, bipyridyl, and Paphy are not very effective in ferritin iron release but show greatly enhanced iron mobilization in the presence of photoreduced FMN, ascorbate also increases their capacity to release ferritin iron_ We hope that these preliminary in vitro studies, which we are presently extending to cell culture systems [21], will help to establish a rational screening system, which can be used to test potential iron chelators for use in the treatment of iron overload.

We thank Professor Tony Cerami for the generous gift of rhodotondic acid, CIBA-GEIGY (Basle) fi- the gift of desferrioxamine 8. end Professor Leon Ghosez, Dr. Alexir cordy and Dr. Chantul wiaux.Zamar for the synthesis of Puphy and for valuable discussions We thankI_RS.lA. f.nstitut pour rEncouragement de la Recherche Scientifiue dans I’Indusm~e et I’Agriculture) for a Bourse de Specialisation to F. Rommr

REFERENCES

l_ R. R. Crichton. Stnrcture and Bonding 17,67 (1973). 2. P. M. Harrison, Sern. Haematol. 14,55 (1977). 3. R. R. Crichton, CTBA Sympotim 65,57 (1979). 4. H. J. BieJ.ig and E. Bayer, iVaturwiksenrchaf?en 42,125 (1955). 5. L. Pape, J. S. Multani, C. Stitt, and P. Saltman, Biochemm 7,613 (1965). 6. J_ Dognin and R. R. Crichton, FEBS Lett. 54,234 (1975). 7. A. Mazur. S. Green, A. Saha, and A. Carleton,1 clin. Invest. 37,1809 (1958). 8. S. Osaki and S. Sirivech, Fad. Proc. 30,1292 (abstract, 1971).

316 Robert R_ Crichton et al.

9. S. Sir&&, E. Frieden, and S. OsaJci, Biochem J. 143,311(1974). 10. R. R. CrIchton, D. CoIIetCassar& Y. Ponce-Oritz. M. Wauters, F. Roman, and E. Paques, in

Proteins of Iron Metabolism, E_ B. Brown, Ed-, Grune & Stratton, Inc, New York, 1977, pp. 13-22.

11. T_ Jones, R. Spencer and C. Walsh, Biochemistry 17,4011(1978). 12. A. Jacobs, CIBA Symposium 51,91(1977). 13. T. C. Iancu, H. B. Neustein, and N. H. Landing, CiWA Symposium 51,293 (1977). 14. R. W. Grady, J. H. Graziano, H. A- Akers, and A- Cerami, J. PhmmaroL Exp. Therap. 196,

478 (1976). 15_ R_ W_ Grady, J. H. G raziano, G_ P_ White, A. Jacobs, and A. Cerami, J. PhmnzacoL Exp.

Therap- 205,757 (1978). 16. G. P. White, R. Bailey-Wood, and A. Jacobs, C7in_ Sci MoL Med- 50.145 (1976). 17. C. Hers&o, Blood 51,415 (1978). 18. C. Hershko, J. Lab_ C7i.n Med_ 85,913 (1975). 19_ R_ W. Grady, C. M. Peterson, R. L. Jones, J- H- Graziano, K- K- Bhargara, V- A- Berdoukas,

G. Kokkini, D. LoukopouIos, and A. Cerarni, J. PhannacoL Exp- Z%erap- 209,342 (1979). 20_ C_ G. Pitt, G. Gupta, W. E- Estes. H- Rosencrantz, J- J- MetterviLle, A- L- Crumbliss, R- k

Palmer, K. W. Nordquest, K. A. Sprinkle-Hardy, D. R- Whitcomb, B- R- Byers, J. E- L- Arcen- eaux, C. G- Graines, and C. V. Sciortino, J. PharmacoL Exp. Therap- 208,12 (1979).

21. J. N. Octave, Y. J. Schneider, P. Hoffman, A. Trouet, and R. R. Crichton, FEBS Lett- 108, 127 (1979).

22. S. Granick, J. BioL Chem. 149,157 (1943). 23. R- Hiil,Ploc. Roy. Sot B_ 107,205 (1931). 24. J. W. DrysdaIe and H. N. Munro,J. B&L Chem. 241,363O (1966). 25. M. Wauters. Doctoral thesis, Universite Catholique de Louvain (1977). 26. A- Mazur, S- Baez, and E- Shorr,J_ BioZ. Chem- 213,147 (1955). 27. J- Dognin, J- L- Girardet, and Y- Chapron, B&him. Biophys Acta 297,276 (1973). 28. S. PollarJr. P- Aisen, F. Lasky, and G. Vanderhoff, Brit_ .L HaematoL 34.235 (1976)_ 29. S- Pollack, G- Vanderhoff and F- Lasky, Biochim Biophys. Acta 497,481(1977). 30. R- R- Crichton and F- Roman, Biochem. Sot_ Z’kam 5,1126 (1977). 31. E. P- Paques, Doctoral thesis, Universite CathoIique de Louvain (19?9). - 32. D- A- Lips&&z. T- H- Bothwell, H. C. Seftel, A. A. Wapnick, and R. W. Char&on, Bit_ L

HaematoL 20,155 (1971).

33. A- W- Nienhuis, C. DeIea, R. Aamodt. F. Bartter and W. F. Anderson, Birth Defects 12,177 (1956).

34. K- Yagi, innfethods in Enzymology 18B. 290 (1971). 3.5. R- R- Crichton, M. Wauters, and F. Roman, inProteins of Iron Storage and l’kansport in Bio-

.&em&y and Medicine, R- R- Crichton, Ed. North Holland, Amsterdam, 1975, pp_ 287-294.

36. G- Anderegg, F- L’Eplattenier, and G. Schwartzenbach,ffeZv_ Chim. Acta 46, 1400 (1963). 37. R- W- Green, P- S- Hallman, and F- Lions, Inorg. aem_ 3, 376 (1964). 38. A- E- Martell, Spec. PubL 17. The Chemical Society, Section IL 39. S- Moeschlin and Schnider, U-, in Iron Metabob-sm, F_ Gross, Ed. Springer Verlag. Berlin,

1964. pp- 525-550_ 40. J- Schuberf in Iron Metabolkn, F- Gross, Ed- Springer Verlag. Berlin, 1964. pp_ 466-493. 41. J- Tsin-jao. L- Sommer, and A- Okac, CoZZ_ Czech Chem Commun 27, 1171 (1962).

Received Febmary 8.1980; revised ApriZ 4.1980

iron mobilization from ferritin by chelating agents

Documents