|
|
Next Article 
Blood, Vol. 89 No. 1 (January 1), 1997:
pp. 1-25
REVIEW ARTICLE
Tissue-Specific Regulation of Iron Metabolism and Heme Synthesis: Distinct Control Mechanisms in Erythroid Cells
By
Prem Ponka
From Lady Davis Institute for Medical Research of the Jewish General Hospital, and the Departments of Physiology and Medicine, McGill University, Montreal, Quebec, Canada.
 |
INTRODUCTION |
IRON IS INDISPENSABLE for life, serving as metal cofactor for many enzymes, either nonheme or hemoproteins. In the latter, iron is inserted like a gem in the center of the heme prosthetic group. Hemoproteins are involved in a broad spectrum of crucial biologic functions including oxygen binding (hemoglobins), oxygen metabolism (oxidases, peroxidases, catalases, and hydroxylases), and electron transfer (cytochromes).1 Therefore, heme is formed in almost all living systems, except for a few obligatory anaerobes and certain unicellular organisms auxotrophic for porphyrins and/or heme. Interestingly, metal-bound porphyrin rings were present at the time that photosynthesis evolved2 (and S.I. Beale, personal communication, July 1996), indicating that some organisms synthesized heme before oxygen appeared in the earth's atmosphere. All animal cells can synthesize heme, with the exception of mature erythrocytes and perhaps some other cells at the very end of their differentiation pathways. In the late 1940s and early 1950s, the laboratories of David Shemin and Albert Neuberger elucidated the basic aspects of heme biosynthesis and showed that glycine and succinyl CoA were the source of all the heme.3 At the same time, in vitro heme synthesis was demonstrated in red blood cells of birds4 and human reticulocytes.5 In 1963 Burnham and Lascelles6 demonstrated that heme biosynthesis in the bacterium Rhodobacter spheroides is subject to a negative feedback control by heme, the end-product of biosynthetic pathway. Higher organisms remain faithful to this basic principle of heme synthesis control, although the step at which feedback regulation is executed differs in various tissues.
The heme biosynthetic pathway and its subcellular compartmentation are probably identical in all mammalian cells. Heme biosynthesis involves 8 enzymes, 4 of which are cytoplasmic and 4 which are localized in the mitochondria (Fig 1). The first step occurs in the mitochondria and involves the condensation of succinyl CoA and glycine to form 5-aminolevulinic acid (ALA), catalyzed by ALA synthase (ALA-S). The next four biosynthetic steps take place in the cytosol. ALA dehydratase (ALA-D) converts two molecules of ALA to a monopyrrol porphobilinogen (PBG). Two subsequent enzymatic steps convert four molecules of PBG into the cyclic tetrapyrrole uroporphyrinogen III, which is then decarboxylated to form coproporphyrinogen III. The final three steps of the biosynthetic pathway, including the insertion of ferrous iron into protoporphyrin IX by ferrochelatase, occur in the mitochondria (Fig 1). Free porphyrins have no biologic utility in humans, and are generally produced only as accidental side-products of heme biosynthesis. Heme biosynthesis is normally remarkably efficient, with near-complete utilization of porphyrin intermediates.
View larger version (34K):
[in this window]
[in a new window]
| Fig 1.
The pathway of heme biosynthesis. B6, pyridoxal-5'-phosphate; ALA, 5-aminolevulinic acid; PBG, porphobilinogen; URO'GEN, uroporphyrinogen; COPRO'GEN, coproporphyrinogen; PROTO'GEN, protoporphyrinogen; Ac, acetate; Pr, propionate; Vi, vinyl. (Reprinted with permission from Cox TM, Ponka P, Schulman HM: Erythroid cell iron metabolism and heme synthesis, in Ponka P, Schulman HM, Woodworth RD (eds): Iron Transport and Storage. Boca Raton, FL, CRC, 1990, p 263. Copyright CRC Press, Boca Raton, Florida. ©1990.7 )
|
|
As discussed above, heme plays a fundamental role in many crucial biochemical reactions and its biosynthesis is finely tuned to these requirements which vary significantly among various cells and tissues. Compared with other cells in the organism, rapid rates of heme biosynthesis occur in liver and erythroid cells where large amounts of heme are needed not only for mitochondrial cytochromes but also as prosthetic groups for cytochrome P450 and hemoglobin, respectively. Even between these two tissues there is a dramatic difference in synthetic rates because 85% of organismal heme is synthesized in erythroid cells whose total number is considerably lower than that of hepatocytes. Hence, on a per-cell basis the rate of heme synthesis in the erythron is at least one order of magnitude higher than that in the liver. This is not unexpected because hemoglobin is the most abundant hemoprotein containing as much as 70% of the total iron content of a normal adult. Thus, it is reasonable to expect that both iron metabolism and the regulation of heme synthesis are different in hemoglobin-synthesizing as compared with nonerythroid cells. Many detailed reviews on various aspects of heme synthesis are available.8-18
Iron is transported between sites of absorbtion, storage, and utilization by transferrin,7,19,20 which has a specific relationship to hemoglobin-synthesizing cells since both in vitro and in vivo studies indicate that transferrin is the only physiologic source of iron for erythroid cell heme synthesis. In vitro studies with erythroid cells have shown that the only physiologically active chelate that can provide iron for their hemoglobin synthesis is transferrin. Even when these cells incorporate iron bound to some low-molecular-weight chelators (eg, citrate), this uptake is mediated by cell associated transferrin.21 An absolute requirement for transferrin by erythroid precursors in vivo is demonstrated by the observations that both humans and mice with hereditary atransferrinemia have severe hypochromic microcytic anemias22-25 that can be explained only by the stringent dependency of hemoglobin synthesis on transferrin iron. The fact that all iron for hemoglobin synthesis comes from transferrin and that this delivery system operates so efficiently, leaving mature erythrocytes with negligible amounts of nonheme iron, suggests that the iron transport machinery in erythroid cells is an integral part of the heme biosynthesis pathway. The importance of iron for erythroid heme synthesis is generally not appreciated and, to rectify the situation, this review will pay considerable attention to iron metabolism in hemoglobin-synthesizing cells. However, to fully comprehend the peculiar regulation of iron metabolism in erythroid cells, its unique aspects will be compared with those encountered in other tissues. With this background and relevant discussion of heme pathway enzymes, it should become obvious that while in nonerythroid cells heme-synthesis rate depends on the rate of ALA production in the first and rate-limiting porphyrin biosynthetic enzyme, ALA-S, in erythroid cells it is determined by the availability of iron for ferrochelatase.
| |
ENZYMES OF HEME BIOSYNTHESIS |
All mammalian heme pathway enzymes have been cloned. The genes encoding these enzymes reside on different chromosomes. There are two different genes for ALA-S, one of which is expressed ubiquitously while the expression of the other is specific to erythroid cells. These two genes are responsible for the occurrence of ubiquitous ("housekeeping") and erythroid-specific mRNAs for ALA-S and, consequently, two corresponding isoforms of the enzyme. No tissue-specific isozyme is known for ALA-D but there are subtle differences in 5' untranslated regions (UTRs) in "housekeeping" and erythroid ALA-D mRNAs. PBG deaminase (PBG-D) (see Fig 1) exists in two isoforms, one being present in all cells whereas the other is expressed only in erythroid cells. However, these isoforms are translated from two mRNAs which differ solely in their 5' ends. There is no evidence that the ubiquitous and the erythroid enzymes would be different in the rest of the pathway but variations in mRNAs, caused by the alternative use of the two polyadenylation signals, have been reported for coproporphyrinogen oxidase and ferrochelatase. Thus, while numerous distinct tissue-specific features in the heme pathway enzymes have been identified, it is intriguing that there is no unifying model defining erythroid heme biosynthesis compared with nonerythroid heme biosynthesis at the gene, mRNA, or regulatory levels.
1. ALA-S
ALA is the first intermediate unique to the biosynthesis of tetrapyrroles, and in higher eukaryotes and some bacteria (eg, R spheroides) it is produced by ALA-S which condenses glycine with succinate upon decarboxylation. The enzyme ALA-S was identified simultaneously by Shemin's group26 in bacterial extracts and by Neuberger and coworkers27 in chicken erythrocytes. In plants, algae, and some photosynthetic bacteria, ALA is formed from the intact carbon skeleton of glutamate in a process requiring three enzymatic reactions and glutamyl-tRNA. Interestingly, the latter pathway provides the major source of ALA in the biosphere and likely predates the glycine pathway by more than 1 billion years.28
ALA-S is a homodimer residing on the matrix side of the inner mitochondrial membrane with absolute specificity for glycine. Even so, this amino acid binds to the enzyme with low affinity (Km in the millimolar range). The second substrate, succinyl-CoA, is an intermediate in the tricarboxylic acid cycle and provides the major source of energy for the entire porphyrin biosynthesis pathway.29 The enzyme requires pyridoxal 5'-phosphate (PLP) as cofactor, and Ferreira et al30 recently showed that  -amino group of lysine 313 (in murine ALA-S) is involved in the Schiff base linkage with PLP. Interestingly, this lysine is present in the conserved C-domain (catalytic domain) of all known ALA-S. Aminolevulinate synthase, whose lysine-313 was replaced by alanine, histidine, or glycine using site-directed mutagenesis, can still bind PLP but the mutant enzymes bind the cofactor noncovalently. Moreover, the addition of glycine to the mutant enzymes leads to the formation of external aldimines.31 These observations indicate that the linkage between PLP and ALA-S, which is also referred to as an internal aldimine, facilitates the transfer of the cofactor (PLP) to the substrate (glycine) to form an external aldimine.31 The fact that ALA-S requires PLP as cofactor has obvious clinical implications.
Experiments published in the 1960s showed different responses of hepatic and erythroid ALA-S to heme and suggested that there may be tissue-specific isozymes of ALA-S. However, the first direct evidence supporting this idea was provided by Bishop et al,32 who described significant differences in kinetic and ligand-binding properties between the erythroid and nonerythroid forms of ALA-S of guinea pigs. Subsequent immunochemical studies confirmed the occurrence of liver and erythroid cell ALA-S isozymes in chickens33 and rats.34 More recently, Riddle et al35 isolated and sequenced ALA-S cDNAs from erythroid cells and livers of chicken and unequivocally showed that two separate genes encode the erythroid and the hepatic ALA-S isozymes. ALA-S genes and/or cDNAs have also been isolated and sequenced from mouse erythroid cells, rat liver, and erythroid cells36-39 and human liver and erythroid cells.40,41 Ubiquitous or "housekeeping" ALA-S (ALA-S1, also referred to as ALAS-N) gene has been assigned to chromosome 3p21 and the erythroid-specific (ALA-S2, also referred to as ALAS-E or e-ALAS) gene to a distal subregion of Xp11.21.42-44 The two ALA-S isozymes show extensive similarities but only in the C-terminal 75% of the mature protein. The C-terminal region probably represents the catalytic domain of all ALA-S proteins. Interestingly, most of the ALA-S2 defects, which are responsible for X-linked sideroblastic anemias,15 have been mapped to the C-terminal region. The N-terminal domain of ALA-S does not seem to be required for enzymatic activity but is involved in the targeting of the enzyme to the mitochondria followed by the removal of the "presequence." Erythroid cells contain equal amounts of two ALAS-2 mRNAs generated by an alternative splicing mechanism leading to the absence of exon 4 in one of the transcripts.45 Because this alternative splicing does not occur in mouse and dog erythroid cells and because the relative levels of two ALA-S2 transcripts do not change during erythroid development,45 the splicing is unlikely to play an important regulatory role in erythroid heme synthesis.
The promoter of ALA-S1 gene contains a TATA box and two control elements located immediately upstream of the TATA box that are homologous to the binding site for the transcription factor NRF-1,46 which seems to be involved in the expression of some proteins involved in oxidative phosphorylation.47 On the other hand the promoter of the human erythroid-specific ALA-S2 gene contains several putative erythroid-specific cis-acting elements, including GATA-1, the CACCC box, and the NF-E2 binding sites.41 It is highly likely that erythroid-specific transcription factors such as GATA-1 and NF-E2 control and induce ALA-S2 transcription in concert with the induction of other erythroid-specific genes.
Interestingly, erythroid-specific ALA-S2 mRNA contains an iron-responsive element (IRE) at its 5' UTR that was first identified by Dierks,48 a finding subsequently confirmed by others.41,49 As will be explained hereafter, this localization of the IRE dictates that the translation of erythroid ALA-S mRNA depends on the availability of iron. Hence, by definition, in erythroid cells iron acquisition, rather than ALA production, is the rate-limiting step in heme synthesis.
Role of Heme in the Control of ALA Synthesis
Nonerythroid cells.
Studies conducted mainly on hepatocytes indicate that ALA-S1 is subjected to a negative feedback control that involves heme-dependent inhibition of a variety of processes. First, based on the seminal studies of Granick,50 who showed that the drug-induced synthesis of ALA-S in chick embryo hepatocytes was inhibited by hemin,* as well as on numerous studies that followed,51 it is generally believed that heme exerts feedback repression on the synthesis of ALA-S1. Although it is commonly thought that heme suppresses ALA-S1 transcription,52 the evidence for such an effect is not convincing and in fact may suggest the opposite.53 First, heme, when added in physiologically relevant concentrations to cultures of chick embryo hepatocytes, failed to inhibit the transcription rate of ALA-S gene.53 Second, heme inhibits the synthesis of ALA-S at the posttranscriptional level in avian embryonic cells.54 Third, heme significantly decreases the stability of ALA-S1 mRNA in chicken embryo hepatocytes.53,55 The heme-mediated mRNA instability can be prevented by cycloheximide, suggesting that heme effect is mediated by a labile protein. Fourth, Kikuchi and his colleagues56 were the first to show that heme blocks the transfer of a precursor form of ALA-S1 protein from the cytoplasm to its final destination in mitochondria (reviewed in ref 58). Moreover, when the translocation and consequently maturation of pre-ALA-S1 is blocked by hemin, the enzyme's precursor is rapidly degraded, with a half-life of 30 minutes.57 Finally, heme was shown to directly inhibit the activity of ALA-S1 (Ki = 20 µmol/L),58 but this inhibition may not be physiologically relevant, because the aforementioned effects of heme are achieved at much lower heme concentrations. Moreover, ALA-S1 activity does not seem to be altered by heme produced in mitochondria with rates at least 75 times higher than those occurring in vivo.59 In conclusion, current evidence indicates that heme primarily regulates ALA-S1 by decreasing the half-life of its mRNA and by blocking translocation of ALA-S1 precursor protein into mitochondria,53,55,56 but the relative importance of one or the other mechanism in the overall downregulation by heme is unknown. Despite this uncertainty, heme-mediated repression of ALA-S1 is responsible for rendering this enzyme the rate-limiting step in the nonerythroid heme biosynthetic pathway.
Erythroid cells.
Although there is no doubt that heme controls erythroid heme synthesis in a negative feedback manner, there has been remarkable controversy as to both the mechanism and the level at which this regulation is executed.7,20,60,61 Karibian and London62 were the first investigators to document an end-product negative feedback mechanism that regulates heme formation in erythroid cells. They showed that hemin reduced incorporation of radiolabeled glycine into heme of reticulocytes incubated in vitro. Because hemin had little effect on the incorporation of radiolabeled ALA into heme, it has been concluded that hemin limited the formation of ALA. However, later experiments showed that ALA-S activity in isolated reticulocyte mitochondria was not affected by heme63 added at concentrations sufficient to significantly inhibit heme synthesis in intact cells64 and revealed, together with other studies,65-76 that heme blocks the uptake of iron from transferrin by erythroid cells. Moreover, a heme-induced block of heme synthesis in erythroid cells could not be restored by added ALA which, however, alleviated heme synthesis inhibition caused by the inhibitor of ALA-S, isonicotinic acid hydrazide (INH).77 On the other hand, heme was unable to inhibit heme synthesis in erythroid cells when Fe-transferrin was replaced by lipophilic Fe-chelates that deliver iron directly to mitochondria.74 However, when the physiologic donor of iron, transferrin, was used, heme inhibited radiolabeled glycine incorporation into reticulocyte heme but not into protoporphyrin.74 Collectively, these experiments have documented that heme does not inhibit ALA-S2 activity but does inhibit cellular iron acquisition from transferrin (see also below). The more pronounced effect of heme on glycine, compared with ALA, incorporation62,77,78 can be explained by the fact that heme inhibits the transport of glycine into reticulocytes.74 Although the physiologic significance of this has not been established, it is pertinent to mention that glycine availability may regulate ALA production due to the very high Km of ALA-S2 for glycine (14 mmol/L).32 Concentrations of glycine in plasma are between 0.15 and 0.37 mmol/L (1.1 to 2.8 mg/100 mL, ref 79) and those in tissues about 5- to 10-fold higher79 and, hence, the aforementioned regulation by substrate availability may be physiologically relevant.
In erythroid cells heme does not inhibit the synthesis of ALA-S2. In fact, hemin treatment of murine erythroleukemia (MEL) cells significantly increased radiolabeled glycine incorporation into heme, suggesting that all enzymes in the heme pathway, including ALA-S2, were increased.69 More recently hemin was shown to increase ALA-S2 mRNA levels in MEL cells80 but it is not yet known whether this increase is transcriptional. Interestingly, in MEL cells heme has also been reported81 to stimulate the translation of ALA-S2. However, there are observations that may be interpreted as indicating that heme exerts a negative feedback on ALA-S2 synthesis. Beaumont et al82 reported that succinylacetone (SA), a potent inhibitor of ALA-D that decreases cellular heme content, potentiates the induction of ALA-S by dimethyl sulfoxide (DMSO) in MEL cells. This observation was confirmed by Elferink et al,83 who also showed that heme deficiency in DMSO-treated MEL cells did not potentiate the transcription of ALA-S2 gene but did increase ALA-S2 protein levels. One possible interpretation is that intracellular accumulation of nonheme iron following SA (see below) stimulates ALA-S2 synthesis via IRE/iron regulatory protein (IRP) mechanism; if this proves to be the case, then ALA-S2 induction following SA would not be related to heme depletion.
Human ALA-S2 synthesized in the cytoplasm contains N-terminal leader peptide (49 amino acids) required for the transport to the mitochondria. During transport and processing, the leader peptide is cleaved off yielding a mature enzyme of 538 amino acids (molecular weight ~59.5 kD).41 The leader targeting sequence of ALA-S2 contains three cysteine-proline-rich motifs84 similar to the heme regulatory motifs (HRMs) of the yeast transcriptional activator, HAP1, which binds heme via the HMRs.85 In vitro studies of Lathorp and Timko84 indicate that the HRM in the pre-ALA-S2 confers sensitivity to feedback inhibition by hemin of mitochondrial import of this precursor, but it is unknown whether such a heme-mediated inhibition of ALA-S2 precursor into the mitochondria occurs in intact cells. If it does, it will be difficult to reconcile this effect with the aforemented results showing that the synthesis of ALA-S2 is stimulated by heme.
2. ALA-D
ALA-D is composed of eight identical 36-kD subunits and contains four catalytic sites.14 The enzyme requires Zn2+ (1 atom/subunit) and intact sulphydryl groups for activity. It is the most abundant heme pathway enzyme and is, therefore, unlikely to play a regulatory role. Surprisingly, ALA-D was recently shown to be identical to the 240-kD proteasome inhibitor (CF-2),86 but the potential role of ALA-D in the ubiquitin-dependent protein degradation pathway has not been clarified. It has recently been proposed87 that this function of ALA-D may play a critical role in the rapid hemoglobin accumulation occurring during erythropoiesis.
No tissue-specific isozyme is known for ALA-D. Bishop et al87 recently examined the structure of ALA-D gene and showed that it contains two first exons (named 1A and 1B), which are alternatively spliced to exon 2, where the coding region begins. They also showed that each exon has its own promoter. Although the promoter driving exon 1A expression is TATA-less and contains many GC boxes, the exon 1B promoter contains GATA-1 binding sites. Tissue distribution studies showed that ALA-D mRNA containing exon 1A is ubiquitous, whereas the mRNA containing exon 1B is expressed only in erythroid cells.87 Somewhat surprisingly, CFU-E (colony-forming unit, erythroid) maturation and DMSO-induction of MEL cells was associated with increases of both 1B and 1A ALA-D mRNA levels. This can probably be explained by the fact that the region 2.3 kb upstream of exon 1A bears a site that may bind the erythroid transcription factor NF-E2. The human ALA-D gene also contains two alternative noncoding exons resulting in the production of the "housekeeping" and erythroid-specific transcripts. The promoter region upstream of housekeeping exon is GC-rich and contains three potential Sp1 elements and a CCAAT box. The promoter region upstream of erythroid-specific exon has several CACCC boxes and two potential GATA-1 binding sites.88
3. PBG-D
PBG-D catalyzes the stepwise deamination and condensation of four molecules of PBG to yield an extremely unstable tetrapyrrole intermediate, preuroporphyrinogen (hydroxymethylbilane). The enzyme (molecular weight 35 to 40 kD) contains at its catalytic site a dipyrrolomethane cofactor assembled via deamination and polymerization of two molecules of PBG. The cofactor is covalently linked to the enzyme, does not turn over, and serves to anchor the substrate molecules at the catalytic site.
In humans the PBG-D gene has been assigned to chromosome 11 (11q23-11qter, ref 89). PBG-D exists in two isoforms, ubiquitous (44 kD) and erythroid-specific (42 kD) ones90 encoded by two distinct mRNAs, one being exclusive to erythroid cells. The two mRNAs arise from two overlapping transcription units in a single PBG-D gene that spreads over 10 kb of DNA and is split into 15 exons.91 In the first (ubiquitous) mRNA exon 1 is spliced to exon 3 while the second transcript (erythroid-specific) initiates at exon 2. Exon 2 does not contain an AUG and, hence, the translation of erythroid specific mRNA starts at an AUG located in exon 3. On the other hand, exon 1 (present in the "housekeeping" mRNA) contains an AUG that is spliced into the same reading frame as the AUG in exon 3. Translation of this mRNA yields a protein that contains 17 additional amino acids at its NH2 -terminus, providing an explanation for the larger molecular mass of the ubiquitous isoform.90,91
The upstream promoter of the PBG-D gene is active in all tissues and has some structural features of a "housekeeping" promoter. The erythroid promoter, located 3 kb downstream in intron 1, is active only in erythroid cells and contains several erythroid-specific cis-acting sequences, including GATA-1, NF-E2, and the CACCC motif.92-94 Undoubtedly, these sequences are involved in PBG-D transcription during erythroid differentiation.
A dominantly inherited partial deficiency of the PBG-D causes acute intermittent porphyria (AIP) characterized by acute, life-threatening attacks of abdominal pain, motor and sensory neurological deficits, and psychiatric symptoms.14 Lindberg et al95 recently generated PBG-D-deficient mice by gene targeting. These mice exhibit the typical biochemical characteristics of human AIP, including decreased hepatic PBG-D activity, increased hepatic ALA-S activity, and dramatically increased urinary excretion of ALA after treatment with phenobarbital. Moreover, behavioral tests revealed decreased motor function, and histological examination showed axonal neuropathy and neurologic atrophy.95 To the best of this author's knowledge this is the first report of a "knock-out" of the heme biosynthetic enzyme in whole animal.
4. Uroporphyrinogen III Synthase (URO-S) and 5. URO Decarboxylase (URO-D)
URO-S catalyzes a unique reaction in which the ring d of preuroporphyrinogen is inverted and the chain is cyclized to furnish the isomer III type that is characteristic of all living systems. The enzyme has not been well characterized but the cDNA for human URO-S has been cloned.96
URO-D catalyzes the sequential decarboxylation of the four acetate side chains of uroporphyrinogen III to yield coproporphyrinogen III. URO-D does not require the cofactor but free thiols are important for enzyme activity.14 Human URO-D gene has been cloned and characterized.97 There are two transcriptional start sites but they are used in the same proportion in erythroid and nonerythroid cells. Because there is no evidence for tissue-specific promoters, it is difficult to explain increased transcription of URO-D during erythroid differentiation. Subnormal activity of URO-D is responsible for porphyria cutanea tarda, the most common porphyria in humans.
Terminal Three Enzymes
The terminal three enzymes of the heme biosynthetic pathway are located in mitochondrion in association with the inner membrane, and this arrangement led to suggestions that these enzymes may form a multienzyme complex with accompanying substrate chanelling.98,99 Although more recent experiments failed to show that the obligatory chanelling complex exists for the terminal three enzymes, their results are still compatible with the conclusion that the enzymes function in a manner that prevents significant accumulation of intermediates.100 It is equally appealing to envisage that a channel or transporter for iron is localized in close proximity to ferrochelatase (see Fig 3).
View larger version (21K):
[in this window]
[in a new window]
View larger version (21K):
[in this window]
[in a new window]
| Fig 3.
Models for the terminal mitochondria-associated steps in the heme biosynthesis pathway. Three terminal enzymes, CPO, PPO, and ferrochelatase (FC), are associated with the inner mitochondrial membrane.227 Coproporphyrinogen III (Copro'gen) uptake by mitochondria may be mediated by peripheral-type benzodiazepine receptors (PBR).102 The ligand carrying iron toward mitochondria is unknown, and it has been recently proposed that in erythroid cells there may be direct interaction of the endosome with the mitochondrion followed by the transfer of iron through membrane-bound transporters to FC.166 Obviously, much further work is needed to provide direct evidence for a transient mitochondria-endosome interaction, a model that has recently been referred to as "the kiss-and-run" hypothesis of cellular iron transfer.284 Only ferrous iron (Fe2+) can be used by ferrochelatase169 but neither the site where iron reduction occurs nor the source of electrons (e-) are known. (A) Model proposing an iron "channel" involved in iron transport from endosomes toward FC and a specific transporter involved in the exit of heme from mitochondria. (B) "Reductionistic" model proposing that the flexibility of the outer (O.M.) and inner (I.M.) mitochondrial membranes allows an association of a putative "iron-binding protein" (FeBP) with FC as well as an interaction of the PBR with PPO. The PBRs may be involved not only in the transport of coproporphyrinogen and protoporphyrinogen but also in the exit of heme from mitochondria since various porphyrins appear to bind to these receptors.228
|
|
6. Coproporphyrinogen Oxidase (CPO)
CPO catalyzes oxidative decarboxylation of the two propionate groups at positions 2 and 4 of coproporphyrinogen III to two vinyl groups. CPO is localized in the intermembrane space of mitochondria, probably in a loose association with the outer face of the inner membrane but little information on the mechanism by which coproporphyrinogen III crosses outer mitochondrial membrane is available. Interestingly, peripheral-type benzodiazepine receptors, located on the outer membrane of mitochondria, were recently shown to increase during MEL cell differentiation.101 Moreover, Taketani et al102 reported that benzodiazepine ligands (eg, diazepam and isoquinoline carboxamide) inhibit the binding of coproporphyrinogen III to mitochondria, as well as its conversion to protoporphyrinogen IX. Hence, peripheral type benzodiazepine receptors may play a role in mitochondrial uptake of coproporphyrinogen III.
Recently human cDNA103,104 as well as the gene105 encoding CPO have been cloned. The human genome contains a single CPO gene with multiple transcription activation sites. It appears that a single promoter is active but differentially regulated in erythroid and nonerythroid cells.105 There are six Sp1 binding sites located upstream of the major initiation sites. Moreover, four GATA sites and the CACCC boxes are present in the promoter region. Interestingly, CPO transcripts increase during the erythroid differentiation of MEL cells in culture.106 The CPO gene contains two polyadenylation sites, and an alternative use of different polyadenylation signals seems to be responsible for the production of CPO mRNA with different 3' ends identified in placenta and fibroblasts.
7. Protoporphyrinogen Oxidase (PPO)
PPO is an integral protein of the inner mitochondrial membrane with the active site facing the intermembrane space.14 PPO catalyzes the penultimate step in the heme pathway during which six hydrogen atoms are removed from the porphyrinogen ring. Hemin (50 µmol/L) was reported107 to inhibit PPO activity by approximately 50%, but the concentration of heme was probably too high to consider the effect physiologically relevant. Recently, the cDNA for mice and human108-110 PPO have been cloned and sequenced. Based on the derived amino acid sequences, mouse and human PPO share certain structural characteristics such as molecular size (~51,000) and the presence of a dinucleotide binding motif in the NH2 -terminal region, similar to the one present in many flavin-containing proteins. The deduced amino acid sequences gives rise to a protein without a typical mitochondrial targeting sequence, a finding concurring with earlier observations that PPO is located on the cytoplasmic side of the inner mitochondrial membrane.99,111 Roberts et al112 recently cloned human PPO gene and showed that it is localized on chromosome 1q23. The promoter region contains multiple Sp1 elements, CCAAT boxes, and potential GATA-1 binding sites.113 Northern blot analysis showed that mouse tissues contain two transcripts with lengths of 1.8 and 3.5 kb, the longer one probably being derived from a second downstream polyadenylation site.108 Somewhat surprisingly, while induced and uninduced MEL cells appear to contain similar amounts of PPO mRNA,108 the induction of erythroid differentiation is associated with a significant increase in PPO enzyme activity.114,115 In contrast to mouse, human tissues contain only a single PPO transcript whose size is approximately 1.8 kb.110
8. Ferrochelatase
Goldberg et al116 were the first investigators to identify "heme synthetase," currently known as ferrochelatase, in extracts of chicken erythrocytes. The enzyme activity was subsequently identified in a wide variety of organisms and tissues18,117 but several reports, demonstrating nonenzymatic insertion of ferrous iron into protoporphyrin, challenged the view of an absolute requirement for ferrochelatase. However, Dailey and Lascelles118 described a bacterial heme auxotroph lacking ferrochelatase activity that exhibited an absolute requirement for heme, and this report unequivocally established the biologic importance for ferrochelatase.
Ferrochelatase has two substrates, protoporphyrin IX and ferrous iron, and two products, heme and two protons. De novo synthesized ferrochelatase contains an NH2 -terminal signal peptide that is proteolytically processed during transport into the mitochondria. The mature protein is bound to the inner mitochondrial membrane with the active site located on the matrix side. Ferrochelatases purified from several mammalian species have similar properties, and active enzyme is probably a monomer although there have been suggestions that it functions as a homodimer with a molecular weight of about 80,000.117
Solubilized ferrochelatase is stimulated by fatty acids and phospholipids,117 suggesting that local mitochondrial environment plays a role in the overall enzyme activity. The sulphydryl groups are reportedly essential for enzyme activity.119 Although this latter observation was initially interpreted as indicating that cysteine(s) may be involved in ferrous iron binding, this is probably not the case. The conversion of highly conserved histidine residues to alanine, using site-directed mutagenesis, resulted in a decrease of mutant ferrochelatase activity, suggesting the involvement of histidine in Fe(II) binding.120 Unfortunately, these investigators did not show that this mutation did not have an effect on [Fe-S] cluster, which is important for the enzyme activity (see below). Nevertheless, recent Mössbauer spectroscophy studies support the view that histidine residues of ferrochelatase are ligands of the ferrous substrate but do not rule out the involvement of other ionic ligands (eg, aspartate, glutamate, tyrosinate).121 Although the mechanism of iron reduction is not fully understood, Taketani et al122 proposed that ferrochelatase is associated with complex I of the mitochondrial electron transport chain and suggested that Fe(II) is produced via an NADH-dependent ferric-iron-reducing system.
Recently murine123 and human124 ferrochelatase were shown to contain a [2Fe-2S] cluster which may be essential for enzyme activity or play either a redox or a regulatory role. Because Escherichia coli ferrochelatase is devoid of the [2Fe-2S] cluster,124 the involvement of the cluster in the regulation of mammalian ferochelatase would seem more plausible.
Heme inhibits ferrochelatase125-127 in a noncompetitive manner and in concentrations that may be physiologically relevant (ie,  10 µmol/L), but it is unclear whether overall heme synthesis or cellular iron metabolism are modulated by this effect of heme.
Human ferrochelatase gene, located on chromosome 18q21.3,128 spans approximately 45 kb and consists of 11 exons.129 The proximal promoter of the gene is contained within a region that structurally resembles a CpG island and is devoid of general cis elements such as TATA and CAAT boxes.130 The human ferrochelatase gene promoter contains cis elements recognized by the erythroid-specific factors NF-E2 and GATA-1 and by the Spl family of transcription factors.130 It seems likely that both NF-E2 and GATA-1 play a role in the induction of ferrochelatase during erythroid differentiation, whereas the GC box is responsible for the maintaining of the "housekeeping" expressions of the gene.
Both human131 and mouse132,133 ferrochelatase genes encode two mRNAs of different lengths, produced by the use of two polyadenylation sites. Although both mRNA species have been found in extracts of mouse liver, kidney, brain, muscle, and spleen, the 2.9-kb transcript is more abundant in nonerythroid tissues whereas the 2.2-kb transcript is more predominant in spleen, the site of erythropoiesis in mice.133 In MEL cells, the 2.9-kb ferrochelatase transcript is also more abundant; however, after the induction of erythroid differentiation there is a preferential increase in the 2.2-kb transcript which eventually predominates. With mouse reticulocytes, the purest immature erythroid cell population available, over 90% of the total ferrochelatase mRNA is present as the 2.2-kb transcript.133 Hence, it appears the preferential utilization of the upstream polyadenylation signal, that produces the 2.2-kb transcript, may be an erythroid-specific characteristic of ferrochelatase gene expression. Interestingly, both ferrochelatase mRNA transcripts increase in MEL cells following their treatment with hemin,133,134 but desferrioxamine has only a slight inhibitory effect on ferrochelatase mRNA levels in both induced and uninduced cells.133 Hence, it seems likely that hemin-induced MEL-cell differentiation per se causes an increase in ferrochelatase mRNA.
| |
REGULATION OF CELLULAR IRON METABOLISM IN ERYTHROID AND NONERYTHROID CELLS |
Cellular Iron Acquisition From Transferrin
As discussed earlier, erythroid cells can take up iron only via the transferrin receptor-pathway. Although nonerythroid cells can also take up iron from nontransferrin sources,135,136 under normal conditions they acquire iron from transferrin. Efforts in numerous laboratories have contributed to our current understanding of the mechanism of cellular iron uptake but some seminal contributions to the field should be singled out. In 1949 Finch and coworkers137 first showed that reticulocytes take up iron from plasma and incorporate it into hemoglobin in vitro. In 1958 Paoletti et al138 demonstrated that reticulocytes remove iron from transferrin without catabolizing the protein. The well-known and now classical experiments of Jandl, Katz and coworkers, published in the late 1950s and early 1960s,139,140 first suggested the existence of a membrane transferrin receptor. Somewhat later Morgan et al141,142 obtained the first evidence for the internalization of transferrin by cells, probably via endocytic vesicles. Schulman and coworkers143,144 then showed that apotransferrin-receptor complexes remained stable at pH 5.0, and Morgan145 found that the inhibitors of intravesicular acidification decreased the release of iron from transferrin within reticulocytes. Thus, virtually all aspects of the transferrin cycle were reasonably well worked out before 1980, but around that time the concept of transferrin endocytosis was still highly controversial. The eventual support for transferrin internalization and movement to an acidic intracellular compartment came from studies of Klausner,146 Lodish,147 and their coworkers that were conducted on cell lines which did not synthesize hemoglobin. The currently accepted model of how cells acquire iron from transferrin is shown in Fig 2.
View larger version (17K):
[in this window]
[in a new window]
| Fig 2.
Fe-transferrin cycle. (1) Transferrin attaches to specific receptors (for review see refs 7, 19, 20, 135) on the cell surface by a physicochemical interaction, not requiring temperature and energy. (2) By a temperature- and energy-dependent process, the transferrin-receptor complexes are internalized by the cells enclosed within endocytic vesicles. (3) Iron is released from the transferrin within the endocytic vesicles by a temperature- and energy-dependent process which involves endosomal acidification. However, the minimum pH in endosomes (~5.3) is not low enough to remove iron from both Fe binding sites of transferrin, and yet cells are capable of doing so. Thus, an additional factor(s) may be involved in vivo, and it has recently been documented that binding to cellular receptors promotes more efficient iron release from transferrin at mildly acidic pH.148,149 (4) Iron is transported to intracellular sites of utilization and/or storage in ferritin but this aspect of iron metabolism, including the nature of the elusive intermediary pool of iron and its cellular trafficking, remains enigmatic. Only in erythroid cells does evidence exist for a specific targeting of iron toward mitochondria (see below). (5) The iron-free apotransferrin, which remains attached to the receptor at pH ~5.5, returns to the cell surface, where the apotransferrin is released from the cells. (Reprinted with permission from Cox TM, Ponka P, Schulman HM: Erythroid cell iron metabolism and heme synthesis, in Ponka P, Schulman HM, Woodworth RD (eds): Iron Transport and Storage. Boca Raton, FL, CRC, 1990, p 263. Copyright CRC Press, Boca Raton, Florida. ©1990.7 )
|
|
Our knowledge on the mechanism by which iron traverses the endosomal membrane is elusive, and an attractive possibility is that released ferric iron is reduced to the more malleable and soluble ferrous form. Membrane carriers for ferrous iron, analogous to those found in lower organisms, may be involved in the transmembrane transport of Fe(II).135,136,150,151 As already pointed out, reticulocytes, physiologically, acquire iron only from transferrin.7,20 However, under artificial conditions they can take up nontransferrin iron but only when it is in ferrous forms.152,153 Hence, such an iron uptake may represent the transmembrane transport system responsible for transporting Fe(II) across the endosomal membrane once it is released from transferrin. One possibility for the reticulocyte-endosomal iron transporter may be the proton adenosine triphosphatase (ATPase), which is responsible for endosomal acidification.154,155 Iron, after its release from endosomes, very likely enters the labile intermediate pool from which it is available for mitochondrial heme synthesis, for the insertion into iron-dependent proteins and enzymes, and for storage in ferritin. It has been proposed that iron in this pool may be complexed to citrate, sugars, some amino acids, pyridoxal and nucleotides,135,156 but the real chemical nature of this metabolically and kinetically active pool remains as ill-defined as when its concept was first proposed157,158 and then refined.159 The only abiding certainty seems to be that this iron is available for being intercepted by strong chelators. However, this does not necessarily mean that the iron is complexed to low-molecular-weight ligands; it can equally well be bound to yet unidentified membrane-associated carriers.
When iron reaches the outer mitochondrial membrane, it is entrapped by yet-unidentified ligands and transferred across the inner membrane to ferrochelatase. It has been suggested that the mitochondrial uptake of iron is coupled to the efflux of heme,160 but such a mechanism can hardly operate in erythroid cells where inhibited heme synthesis leads to iron accumulation within mitochondria.161-166 On the other hand, Romslo's160 hypothesis that the mitochondrial uptake of iron is linked to an antiport for heme may be valid in nonerythroid cells where the mitochondrial iron pool remains constant when ferrochelatase is inhibited.167 Ferrochelatase is located on the inner mitochondrial membrane with its active site facing the mitochondrial matrix.168 Because only the reduced form of iron can be processed by ferrochelatase,169 reduction of iron must occur at some point after its release from transferrin. As already mentioned, Taketani et al122 suggested that complex I of the mitochondrial electron transport system is associated with ferrochelatase and is responsible for the production of ferrous iron for this enzyme.
Regulation of Cellular Iron Levels in Nonerythroid Cells
Iron is indispensable for life, but if not appropriately shielded it may become very toxic because of its catalytic action in one-electron redox reactions that produce harmful oxygen radicals which ultimately cause peroxidative damage to vital cell structures.170 Thus, organisms were compelled to solve one of the many paradoxes of life, ie, to keep "free iron" at the lowest possible level and yet in concentrations allowing adequate supply of the essential element for the synthesis of hemoproteins and other iron-containing molecules. In nonerythroid cells enlargement of the intracellular transit iron pool stimulates the synthesis of ferritin171-173 and decreases the expression of transferrin receptors,174,175 and the opposite scenario develops when this pool is depleted of iron. Recent efforts in numerous laboratories have dramatically enhanced our understanding of a feedback control that maintains the size of the transit iron pool at appropriate levels. A remarkable regulation system has emerged that coordinately regulates cellular iron uptake and storage and, in erythroid cells, its utilization for heme synthesis.
Research conducted on nonerythroid cells cultured in vitro showed that iron-dependent regulation of both ferritin and transferrin receptor occurs posttranscriptionally, and is mediated by virtually identical iron-responsive elements (IREs). IREs were first identified in the 5' UTRs of ferritin H- and L-chain mRNAs176-186 and documented to mediate inhibition of ferritin mRNA translation in iron-deprived cells. Somewhat later, five similar IRE motifs (contrary to single IRE in ferritin mRNAs) were identified within the 2.7-kb 3' UTR of transferrin rceptor mRNA.187 These IREs are located in two regions, about 200 bases each, that confer differential stability to transferrin receptor mRNAs as a function of cellular iron levels.188-190 The IRE is also present in the 5' UTR of mRNA for ALA-S2 whose expression in hemoglobin-synthesizing cells depends on the availability of iron (see below). IREs are cis-acting nucleotide sequences, forming stem-loop structures, that contain an unpaired cytidine 6 bases 5' of a six-membered loop whose sequence is CAGUGN. These hairpin structures are recognized by trans-acting cytosolic RNA-binding proteins191,192 known as iron-regulatory proteins (IRPs, formerly known as IRE-binding proteins, IRE-BPs, iron regulatory factor [IRF], or ferritin repressor protein [FRP]).181-183,186
Two closely related IRPs, designated IRP-1192-195 and IRP-2,194,196-198 have been purified and cloned from a variety of mammalian tissues and cells. IRP-1 (Mr = 98 kD) shares homology with mitochondrial aconitase,199 a[4Fe-4S] cluster-containing enzyme of the citric acid cycle. In iron-replete cells IRP-1 also contains a cubane [4Fe-4S] cluster and in this form possesses aconitase activity200 and binds RNA with low affinity. In contrast, when iron is scarce IRP-1 lacks a [4Fe-4S] cluster and aconitase activity and binds to IREs with high affinity. The transition between the aconitase and RNA-binding form of IRP-1 is regulated posttranslationally and occurs without changes in IRP-1 protein levels.195,201 IRP-2 shares 61% overall amino acid identity with IRP-1,198,202 binds to IREs with similar affinities as IRP-1,197 and upon binding represses translation of IRE-containing mRNAs.197,203 However, in contrast to IRP-1, IRP-2 functions solely as an RNA-binding protein because it lacks aconitase activity, and regulation of IRP-2 by iron is mediated by specific proteolysis.197,204,205
Plentiful evidence collected in the above studies indicates that the interactions of IRPs with IREs control iron metabolism in nonerythroid cells in the following manner: When cellular iron becomes limiting, the IRP-1 is recruited into the high-affinity binding state. The binding of IRP-1 to the IRE in the 5' UTR of the ferritin mRNA represses the translation of ferritin, whereas an association of IRP-1 with IREs in the 3' UTR of transferrin receptor mRNA stabilizes this transcript against as yet undefined ribonucleases. On the other hand, the expansion of the labile iron pool inactivates IRE-P1 and leads to a degradation of IRP-2, resulting in an efficient translation of ferritin mRNA and rapid degradation of transferrin receptor mRNA.
Much has recently been written on the involvement of this fascinating regulatory system in a "fine tuning of iron metabolism" and "iron homeostasis." However, a generalization that IRE/IRP-dependent mechanism is involved in the coordinated regulation of iron uptake and storage by all cells, implicitly inherent in most recent discussions, may be incorrect. One likely exception is erythroid cells that maintain very high transferrin receptor levels despite the gluttonous appetite with which they take up iron. Could the explanation be found in the overproduction of protoporphyrin IX that efficiently "mops up" iron, making it virtually undetectable by the IRP-sensing system? This appears unlikely, because iron flux into mitochondria is not prevented in the absence of protoporphyrin,161-166 suggesting that heme synthesis is not the driving force for the uptake of iron by the organelle. Moreover, levels of active IRP-1 remain virtually unchanged when MEL cells are induced to synthesize hemoglobin,206 indicating that IRP-1 probably does not play a role in the increase of transferrin receptor mRNA during erythroid differentiation. Hence, erythroid cells appear to possess distinct control mechanisms to satisfy their needs for iron and to specifically handle the metal, and further evidence supporting this concept is presented below.
Erythroid-Specific Iron Metabolism
Iron Supply for Heme Synthesis in Erythroid Cells: A Rate-Limiting Step
The possibility that the rate of iron acquisition from transferrin may play a crucial role in heme synthesis regulation in erythroid cells was suggested by the finding that heme reduced not only assimilation of transferrin iron by erythroid cells but also inhibited their heme synthesis.63-66 The hypothesis that "the supply of iron to the critical sites of haem synthesis may be a limiting factor controlling the rate of haem synthesis" was first proposed more than 25 years ago,207 but its confirmation had to wait for another 15 years. Until relatively recently, a classical experimental approach to determine whether excessive iron supply will increase the rate of heme synthesis in erythroid cells was not feasible because no chelate could bypass transferrin receptors in the delivery of iron to the hemoglobin-synthesizing apparatus of the cell. The identification of acyl hydrazones such as pyridoxal isonicotinoyl hydrazone (PIH) and salicylaldehyde isonicotinoyl hydrazone (SIH) as lipophilic iron chelators,208,209 and the demonstration that Fe-PIH or Fe-SIH can shunt iron to the heme-synthesizing apparatus of mitochondria without involving transferrin and transferrin receptors,210-213 allowed testing of the above hypothesis.
When compared with saturating levels of Fe-transferrin, Fe-SIH in high concentrations was shown to stimulate the incorporation of radiolabeled glycine into reticulocyte heme. On the other hand, the addition of ALA, the product of the supposedly limiting step in heme synthesis, did not stimulate 59Fe incorporation into heme from either 59Fe-transferrin or 59Fe-SIH.211-213 These studies have provided evidence that in reticulocytes, a step in the pathway of iron from extracellular transferrin to ferrochelatase rather than formation of ALA limits, and thereby controls, the overall rate of heme synthesis. A similar mechanism responsible for heme synthesis control appears to operate also in less mature erythroid cells212 but not in nonerythroid cells.
The generality of this proposal was challenged when it was shown that the formation of ALA is rate-limiting for heme synthesis in human reticulocytes, and it was concluded that the control of erythroid heme synthesis is different in humans and rabbits.214 However, it is unlikely that unique regulatory mechanisms control heme synthesis in human erythroid cells. First, the rates of heme synthesis in human reticulocytes, which are released from the bone marrow very mature,215 are approximately 10- to 20-fold lower than the rates in reticulocytes from anemic rabbits.214 Because Fe-SIH fails to stimulate heme synthesis in mature rabbit reticulocytes (H.M. Schulman and P. Ponka, unpublished data, 1990), the described difference between rabbit and human reticulocytes214 is probably related to differences in maturity of the cells rather than to intrinsic differences in the modes of regulation of heme synthesis. Second, the IRE element is present not only in experimental animal but also in human ALA-S2 mRNA,41 indicating that the formation of ALA-S in human erythroid cells is limited, and hence controlled, by the availability of iron.
Transcriptional Control of Transferrin Receptor Expression
Iron modulates transferrin receptor expression in nonerythroid cells in a negative feedback manner at the level of mRNA stabilization that affects steady-state transferrin receptor mRNA levels (see above). However, in hemoglobin-synthesizing cells transferrin receptor mRNA levels are only slightly affected by high iron concentrations,206 suggesting that transferrin receptor expression is regulated differently in erythroid compared with nonerythroid cells. This relative insensitivity of transferrin receptor mRNA to the destabilizing effects of iron is probably due to the high rate of transferrin receptor gene transcription,206,216 and it is of considerable interest that GATA-binding sequences have been recently identified in regions upstream and flanking the first and second exons of chicken and mouse transferrin receptor genes.217 Hence, it is likely that GATA transcription factor may play a role in the induction of transferrin receptor gene transcription during erythroid differentiation. The transferrin receptor gene promoter also contains a DNA sequence with some similarity to the AP1 recognition site,188,218 and it has been reported that the NF-E2 can bind to some AP-1 sites.219 Considering this, it may be that NF-E2 is not only involved in hemoglobin synthesis, but may also take part in the enhanced transcription of the transferrin receptor that occurs during erythroid differentiation. It is also possible that erythroid cells may express a unique transferrin receptor isoform220 which may be subject to different control mechanisms.
Iron Targeting Into Mitochondria
In hemoglobin-synthesizing cells iron appears to be specifically targeted toward mitochondria.20 In erythroid cells, the vast majority of iron released from endosomes must cross both the outer and the inner mitochondrial membranes to reach ferrochelatase. Interestingly, in hemoglobin-synthesizing cells iron acquired from transferrin continues to flow into mitochondria, even when the synthesis of protoporphyrin IX is markedly suppressed (eg, by INH or SA).161-166 The chemical nature of iron which accumulates in mitochondria is unknown but iron does not appear to be in the form of ferritin. This is remarkable considering almost ubiquitous occurrence of ferritin, whose function is tailored to prevent iron toxicity. To the best of this author's knowledge the only other living system capable of accumulating excessive iron in a nonferritin form is yeast. A certain proportion of this nonheme iron accumulated in mitochondria is in a form readily available for heme synthesis when protoporphyrin IX formation is restored.161,162,164,166 Under these conditions, heme is quickly transported out of the mitochondria to combine with globin chains in the cytosol, suggesting that the majority of iron taken up by erythroid mitochondria can leave the organelle only after the iron is inserted into protoporphyrin IX. Interestingly, when heme synthesis is inhibited in erythroid cells, very little162,166 or no164 iron accumulates in cytosolic ferritin. In contrast, it is well established that, in nonerythroid cells, iron in excess of metabolic needs ends up in ferritin. Thus, it seems highly likely that some specific mechanisms and controls are involved in the transport of iron into mitochondria in erythroid cells,7,20 but the nature of these processes has yet to be identified.
The unique nature of iron metabolism in the mitochondria of erythroid cells is further emphasized by the fact that mitochondria in nonerythroid cells do not accumulate nonheme iron, even in severely iron overloaded individuals. Interestingly, when protoporphyrin synthesis is inhibited in reticulocytes obtained from copper-deficient animals, it does not result in an accumulation of iron in the mitochondrion.221 Therefore, either copper, or more likely, a copper-dependent protein, seems to be involved in mitochondrial Fe uptake, and it is tempting to speculate that such an uptake system may be homologous to one of the iron transport systems recently described in yeasts.136,150
A defect in heme synthesis provides a likely explanation for ringed sideroblast formation in at least some sideroblastic anemia patients, for example those with a defect in the ALA-S2 gene.15,222,223 The ringed sideroblast is a pathologic erythroid precursor containing excessive deposits of nonheme nonferritin iron within mitochondria, which shows perinuclear distribution accounting for the ringed appearance. Based on the above discussion, it can be proposed that the combination of four factors plays a role in the pathogenesis of mitochondrial iron accumulation in those sideroblastic anemias in which the defect in heme synthesis has been established: (1) iron is specifically targeted toward erythroid mitochondria; (2) iron cannot be used because of the lack of protoporphyrin IX; (3) there is a lack of heme  the negative regulator of iron uptake (see below); (4) iron can leave mitochondria only after being inserted into protoporphyrin IX.
However, in some sideroblastic anemias, in particular primary acquired ones, there is little evidence for inhibited protoporphyrin formation. Perhaps, here a lesson can be learned from sideroblastic anemias that accompany rare Pearson's syndrome, also known as marrow-pancreas syndrome.15 In these patients no defect in heme synthesis has been shown but they have a congenital disorder caused by mitochondrial DNA deletions.15 It is conceivable that in erythroblasts of children with Pearson's syndrome, a respiratory enzyme defect is responsible for the reduced availability of ferrous iron for ferrochelatase, leading to Fe(III) accumulation in mitochondria. Hence, it is conceivable that some forms of primary acquired sideroblastic anemia could be caused by age-related mitochondrial DNA deletions or rearrangements. Because primary acquired sideroblastic anemia is a clonal disorder originating in a pluripotent stem cell,224 an additional nuclear mutation conferring growth advantage may be needed.
Based on the fact that transferrin-bound iron is extremely efficiently used for hemoglobin synthesis, that iron is targeted into erythroid mitochondria, and that no cytoplasmic iron transport intermediate can be identified in rabbit reticulocytes, a new hypothesis of intracellular iron transport has been suggested.135,166 This model proposes that after iron is released from transferrin in the endosome, it is passed directly from protein to protein until it reaches ferrochelatase in the mitochondrion. Such a transfer would bypass the cytosol, as the transfer of iron between proteins could be mediated by the direct interaction of the endosome with the mitochondrion (Fig 3). Although much further work is required to provide direct evidence for this model, it is a reasonable working hypothesis that can explain at least some erythroid-specific aspects of intracellular iron trafficking.
Control of Iron Uptake in Erythroid Cells
Mature erythrocytes contain virtually all their iron in hemoglobin, indicating that the uptake of iron by immature erythroid cells is tightly coupled with the use of iron for heme synthesis. The fact that this iron delivery system operates so efficiently is likely to be due to a negative feedback control in which nonhemoglobin-bound heme ("uncommitted heme")225,226 inhibits iron acquisition from transferrin64-76 (Table 1). Unfortunately, it is still unresolved whether heme inhibits transferrin endocytosis72,73 or iron release from transferrin.66,67,75
A possible clue on the mechanism of heme action may be provided by results obtained with Belgrade rat (b/b) reticulocytes, which have a genetically-based defect in iron uptake.229,230 Heme does not inhibit iron uptake by b/b reticulocytes68 and, interestingly, the inhibition of heme synthesis does not lead to nonheme iron accumulation in their mitochondria.165 It can be suggested that the erythroid cells from b/b animals lack an erythroid-specific protein, perhaps mitochondrial iron transporter, that is controlled by heme. It is important to point out that this effect of heme is specific for hemoglobin-synthesizing cells, because heme does not inhibit iron uptake in nonerythroid cells.231,232 The aforementioned action of heme on erythroid cell iron uptake commences very rapidly (minutes), and should be distinguished from the much slower inhibitory effect of heme on transferrin receptor expression in cells that do not synthesize hemoglobin. In this second scenario, heme decreases transferrin receptor mRNA levels233,234 via the IRE/IRP system.235 However, current experimental evidence indicates that IRP-binding activity is not regulated by heme itself but rather by iron released from heme due to the action of heme oxygenase.236
Somewhat paradoxically, heme was shown to be essential for maintaining a normal rate of transferrin receptor synthesis in erythroid cells (see Table 1). Cox et al237 showed that inhibition of heme synthesis by SA depresses transferrin receptor synthesis in reticulocytes, and that this can be restored by the addition of heme. Similarly, Johnstone and coworkers238 found that heme was essential for optimal expression of transferrin receptors in chicken erythroid cells. On the other hand, heme synthesis inhibitors were shown to strongly inhibit transferrin receptor expression at both the mRNA206,239,240 and protein206 levels in nucleated erythroid cells, but had little effect on the expression of the transferrin receptor in cells that did not synthesize hemoglobin.206 The mechanism of heme action on transferrin receptor expression in erythroid cells remains to be determined, but it may be related to a general dependence of the erythroid differentiation program on the availability of heme. It is possible that a heme-regulated eIF-2 kinase (HRI) may be one of the factors involved in maintaining high transferrin receptor levels in erythroid cells. Heme binding to HRI inhibits the phosphorylation of eIF-2 by HRI, resulting in an efficient translation of globin and perhaps other proteins in erythroid cells (for review see ref 268).
Role of Iron in the Translation of ALA-S2
The theory that iron supply regulates heme formation in hemoglobin-synthesizing cells received additional support by the discovery of an erythroid-specific form of ALA-S. As already mentioned, the mRNA for ALA-S2 contains an IRE in its 5' UTR,41,48,49 which is similar to that present in the 5' UTR of ferritin mRNA. As expected, iron has been consequently shown to translationally control the expression of ALA-S281,241,242 and to stimulate ALA production in erythroid cells.241,242 Moreover, iron is also required to maintain the stability of ALA-S2242 and appears to play a role in the synthesis or stability of ALA-S2 mRNA.243 Results of these experiments can explain the mechanism by which iron deficiency decreases erythroblast ALA-S activity.244
| |
REGULATION OF HEME SYNTHESIS IN DIFFERENT TISSUES |
Erythroid Cells
The mechanisms of regulating heme synthesis in hemoglobin-synthesizing cells may be looked at from two different perspectives with respect to erythroid cell development. The first concerns the regulatory mechanisms during the induction of hemoglobinization, ie, during a period in which erythroid precursors acquire all the machinery for hemoglobin synthesis. Earlier reports with MEL cells suggested a sequential induction of the enzymes of the heme biosynthetic pathway, ALA-S being induced first and ferrochelatase last.245 However, Beaumont et al82 reported th |
Introduction
References