H emel is a complex of iron with protoporphyrin. Cell Biology of Heme PREM PONKA, MD, PHD

Transcription

1 PREM PONKA, MD, PHD ABSTRACT: Heme is a complex of iron with protoporphyrin IX that is essential for the function of all aerobic cells. Heme serves as the prosthetic group of numerous hemoproteins (eg, hemoglobin, myoglobin, cytochromes, guanylate cyclase, and nitric oxide synthase) and plays an important role in controlling protein synthesis and cell differentiation. Cellular heme levels are tightly controlled; this is achieved by a fine balance between heme biosynthesis and catabolism by the enzyme heme oxygenase. On a per-cell basis, the rate of heme synthesis in the developing erythroid cells is at least 1 order of magnitude higher than in the liver, which is in turn the second most active heme producer in the organism. Differences in iron metabolism and in genes for 5-aminolevulinic acid synthase (ALA-5, the first enzyme in heme biosynthesis) are responsible for the differences in regulation and rates of heme synthesis in erythroid and nonerythroid cells. There are 2 different genes for ALA-5, one of which is expressed ubiquitously (ALA-51), whereas the expression of the other (ALA-52) is specific to erythroid cells. Because the 5' -untranslated region of the erythroid-specific ALA-52 mrna contains the iron-responsive element, a cis-acting sequence responsible for translational induction of erythroid ALA-52 by iron, the availability of iron controls protoporphyrin IX levels in hemoglobin-synthesizing cells. In noneryth- H emel is a complex of iron with protoporphyrin IX and serves as the prosthetic moiety of numerous hemoproteins that are essential for the function of all aerobic cells. Hemoproteins are involved in a remarkable array of crucial biologic functions including oxygen binding (hemoglobin, myoglobin), oxygen metabolism (oxidases, peroxidases, and catalases) and electron transfer (cytochromes). Moreover, heme is the prosthetic group of numerous he- 1 Heme is ferroprotoporphyrin IX; hemin is ferric protoporphyrin IX In this article, the term heme is used as a generic expression denoting no particular iron valence state. From the Lady Davis Institute for Medical Research, Jewish General Hospital, and Departments of Physiology and Medicine, McGill University, Montreal, Quebec, Canada This work was supported by grants from the Medical Research Council of Canada. Correspondence: Dr. Prem Ponka, Lady Davis Institute for Medical Research, Jewish General Hospital, 3755 Cote Ste-Catherine Road, Montreal, QC, Canada H3T le2 ( mdpp@musica.mcgill.ca). THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES roid cells, the rate-limiting step of heme production is catalyzed by ALA-51, whose synthesis is feedback-inhibited by heme. On the other hand, in erythroid cells, heme does not inhibit either the activity or the synthesis of ALA-5 but does inhibit cellular iron acquisition from transferrin without affecting its utilization for heme synthesis. This negative feedback is likely to explain the mechanism by which the availability of transferrin iron limits heme synthesis rate. Moreover, in erythroid cells heme seems to enhance globin gene transcription, is essential for globin translation, and supplies the prosthetic group for hemoglobin assembly. Heme may also be involved in the expression of other erythroid-specific proteins. Furthermore, heme seems to playa role in regulating either transcription, translation, processing, assembly, or stability of hemoproteins in nonerythroid cells. Heme oxygenase, which catalyzes heme degradation, seems to be an important enzymatic antioxidant system, probably by providing biliverdin, which is an antioxidant agent. KEY INDEXING TERMS: Heme synthesis; Heme oxygenase; Heme-regulated inhibitor; 5-Aminolevulinic acid synthase; Erythroid cells; Iron metabolism; Transferrin receptor; Iron-regulatory proteins (I RPs); Iron-regulatory elements (IREs); Carbon monoxide; Nitric oxide. [Am J Med Sci 1999;318(4): ] moproteins that synthesize important regulatory or signaling molecules including cyclic GMP (guanylate cyclase), steroid hormones (hydroxylases) and nitric oxide (nitric oxide synthase). Furthermore, heme plays an important role in controlling the expression of numerous proteins (globin, heme biosynthetic enzymes, cytochromes, myeloperoxidase, heme oxygenase-l, transferrin receptor) and by providing carbon monoxide (CO), which may have a regulatory role akin to that of nitric oxide (NO). Cellular heme levels seem to be tightly controlled; this is achieved by a fine balance between heme biosynthesis and catabolism by the enzyme heme oxygenase. Heme requirements vary significantly among various cells and tissues; the most rapid rates of heme biosynthesis occur in erythroid cells and in liver. Even between these 2 tissues there is a dramatic difference in synthetic rates, because 85% of organismal heme is synthesized in immature erythroid cells whose total number is considerably lower than that of hepatocytes. Hence, on a per-cell 241

2 CH, CH, SUCCINYL-(;oA coo I CH, I CH, I CoAS/~"O ~ B, \ \ H CoASH CO2 I p, p, HEME.----c=----, p, PAOTOPORPHYRINOGEN III H-C-NH2 I 000- GlYCINE coo I COO- CH2 I I CH2 CH2 NH2-CH2~H H PORPHOBILINOGEN ~I PSG-~OEAM~I!NASillE I l i-' p, '"' Pr Pr UROPORPHYAINOGEN III Pr Pr COPROPOAPHYRINOGEN III Figure 1. The pathway of heme biosynthesis. B6, pyridoxal-5' phosphate; URO'GEN, uroporphyrinogen; COPRO'GEN, coproporphyrinogen; PROTO'GEN, protoporphyrinogen; Ac, acetate; Pr, propionate; Vi, vinyl. [Reprinted from Cox TM, Ponka P, Schulman HM. Erythroid cell iron metabolism and heme synthesis. In Ponka P, Schulman HM, Woodworth RD, editors. Iron transport and storage. Boca Raton (FL): CRC Press; p Copyright CRC Press. Used with permission.] basis, the rate of heme synthesis in the developing erythroid cells is at least 1 order of magnitude higher than that in the liver. Consequently, hemoglobin provides most of the organismal heme for catabolism by heme oxygenase in macrophages. However, all cells, with the exception of mature erythrocytes and perhaps some other highly differentiated cells, can degrade heme. Our knowledge regarding intracellular heme trafficking and assembly of hemoproteins is far less than adequate, although some information on the formation of mitochondrial cytochromes has been obtained from experiments with yeast. These studies revealed that heme, which is produced in mitochondria (see Figure 1), is covalently attached to cytochrome c by the enzyme cytochrome cherne lyase, localized in the mitochondrial intermembrane space. Apocytochrome c is synthesized in the cytoplasm, transported to the mitochondrial intermembrane space, and subsequently attached to heme by heme lyase. Failure to attach heme to cytochrome c, which can be produced by a heme analog, mutation of cytochrome c, or by a lack of heme lyase, leads to accumulation of the precursor apocytochrome c in the cytoplasm.1-3 Additional experiments have suggested that the synthesis of cytochrome c6 precursor may be coupled to heme availability by a control mechanism operating at the translational level. 4 Information on the assembly of many important extramitochondrial hemoproteins (eg, hemoglobin, myoglobin, cytochrome P450) is incomplete, although the role of heme in the processing of a myeloperoxidase is reasonably well understood (see below). Hepatocytes are known to contain a "free" or "uncommitted" heme pool consisting of newly synthesized heme that serves both precursor and regulatory functions. In primary cultures of adult rat hepatocytes, 20% of newly formed heme is directly converted to bile pigments, whereas 80% is used for the formation ofhemoproteins.5 Hence, it seems that in the hepatocytes, heme is formed in slight excess over its metabolic needs. In erythroid cells, "uncommitted" heme is even more enigmatic and was shown to increase in mitochondria when the cells were treated with protein synthesis inhibitors.6 This suggests that globin chains are probably needed for the release of heme from mitochondria. However, it is also possible that specific heme-binding proteins are involved in the intracellular transport of heme and in its targeting to appropriate intracellular locations. Candidates for these functions include liver fatty acid-binding protein (L-FABP),7 glutathione S-transferase,8 and a heme-binding protein designated HBP23.9 HBP23 is a particularly attractive candidate because it binds heme with an affinity (K d = 55 nm) that is much higher than that of the other 2 proteins.9 Interestingly, HBP23 is highly homologous to the mouse macrophage 23-kDa stress protein, which is inducible by oxidant stress in peritoneal macrophages.10,1l These results suggest that HBP23 may also have an antioxidant function. More recently, Taketani et a11 2 purified a novel heme binding protein with a molecular mass of 22 kda (termed p22 HBP) and identified its cdna, the sequence of which revealed that p22 HBP comprises 190 amino acid residues. p22 HBP is ubiquitously expressed in various tissues and is extremely abundant in the liver. It binds heme (K d ~ 25 nm), protoporphyrin, and coproporphyrin with relatively high affinities. p22 HBP (both mrna and protein) is induced during erythroid differentiation, and antisense oligonucleotides to p22 HBP decrease heme biosynthesis in dimethyl sulfoxide (DMSO)-induced murine erythroleukemia (MEL) cells. Hence, p22 HBP may be involved in heme utilization for hemoprotein synthesis and may even be coupled to hemoglobin synthesis during erythroid differentiation.12 As mentioned, heme is not only a precursor of hemoproteins but also serves as a regulator of numerous metabolic functions; one of those functions is regulation of gene expression (see below). Such a function would seem to require heme entry into 242 October 1999 Volume 318 Number 4

3 Ponko nuclei, but to the best of my knowledge, only one report13 documents the accumulation of nonhemoglobin heme in the nuclei of differentiating MEL cells. Many detailed reviews on various aspects of heme biosynthesis,14-23 catabolism,24-26 transport,27 and function28,29 are available. In the present review, I will discuss regulation of heme biosynthesis as well as various controlling roles heme plays in different cell types, focusing primarily on erythroid cells. Heme Biosynthesis and Its Control by Heme Overview of Heme Synthesis Heme biosynthesis involves 8 enzymes, 4 of which are cytoplasmic and 4 of which are mitochondrial (Figure 1). The first step occurs in the mitochondria and involves the condensation of succinyl coenzyme A (CoA) and glycine to form 5-aminolevulinic acid (ALA), catalyzed by ALA synthase (ALA-S). The next 4 biosynthetic steps take place in the cytosol. ALA dehydratase (ALA-D) converts 2 molecules of ALA to a monopyrrole porphobilinogen (PBG). Two subsequent enzymatic steps convert 4 molecules of PBG into the cyclic tetrapyrrole uroporphyrinogen III, which is then decarboxylated to form coproporphyrinogen III. The final 3 steps of the biosynthetic pathway, including the insertion of ferrous iron into protoporphyrin IX by ferrochelatase, occur in the mitochondria (Figure 1). Physiological reasons or possible advantages for the mitochondrial localization of the initial and last 3 enzymes, as opposed to the cytoplasmic location of the remaining enzymes, are not known. Iron Pathway for Heme Synthesis It is not known whether heme synthesis in nonerythroid cells depends on an extracellular source of iron (transferrin, an iron transporter in the circulation), whether ferrochelatase can use iron from intracellular sources (heme catabolism and/or ferritin, a protein of iron storage), or whether both extra- and intracellular iron sources can support heme synthesis. Physiologically, however, heme synthesis in erythroid cells is stringently dependent on transferrin iron;23 hence, a brief discussion on iron acquisition by hemoglobin-synthesizing cells is warranted. Cellular iron uptake is also discussed in the article by Dr. Conrad elsewhere in this issue. An absolute requirement for transferrin by erythroid cells in vivo is demonstrated by the observations that both humans and mice with hereditary atransferrinemia have severe hypochromic microcytic anemias associated with generalized iron overload.3o,31 Moreover, mice homozygous for atransferrinemia die shortly after birth unless transfused with erythrocytes or administered transferrin,31 which indicates that nontransferrin iron sources cannot be used for hemoglobin syn- THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES thesis. 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.23 The current view on iron acquisition via transferrin-receptor-mediated endocytosis is schematically depicted in Figure 2.32 In the first step, transferrin attaches to specific transferrin receptors on the cell surface by a physicochemical interaction that does not require temperature and energy. By a temperature- and energy-dependent process, the cells then internalize the transferrin-receptor complexes, enclosed within endocytic vesicles. Iron is released from the transferrin within the endocytic vesicles by a temperature- and energy-dependent process that involves endosomal acidification. An influx of protons into endosomes probably occurs via an adenosine triphosphate-dependent H+ pump that is, however, poorly defined. The minimum ph in endosomes (~ 5.3) is not low enough to remove iron from both Fe binding sites of transferrin, yet cells are capable of removing iron from transferrin with remarkable efficacy. This paradox has recently been explained by demonstrating that binding to cellular receptors promotes more efficient iron release from transferrin at mildly acidic ph.33 Although our know ledge of the mechanism by which iron traverses the endosomal membrane is sparse, a possible candidate for the endosomal iron transporter is the natural resistance-associated macrophage protein 2 (Nramp2).34,35 It seems likely that ferric iron liberated from transferrin is reduced to the more ductile and soluble ferrous form; this conclusion is endorsed by recent findings that Nramp2 [also known as divalent cation transporter 1 (DCT1)] can transport only ferrous iron.36 Metabolism of iron after its release from the endosomes, including the nature of the elusive intermediary pool of iron and its cellular trafficking, remains totally enigmatic. Only in erythroid cells does some evidence exist23,32,37 for a specific targeting of iron toward mitochondria [sites of heme production by ferrochelatase (see Figure 1)]. 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.32 The above-described mechanism of iron uptake via transferrin-receptor mediated endocytosis is likely to be identical in all cell types. However, iron metabolism in erythroid cells is characterized by certain unique features that should be highlighted, because they are highly relevant to heme synthesis regulation. First, the rate of iron acquisition from transferrin limits and therefore controls the rate of heme synthesis in erythroid cells.23 This was dem- 243

4 Extracellular... '.'.'.'.'.'.'.'..... '.....,..... '... { Clathrin-Coated Pit!f)/.~ ri.~~ r"'gf$$ ~eceptor-mediated Endocytosis Proton ~ H+ Pump :ii ~ Iron Transport 1 Intracellular IRP1 and IRP2 Regulation of Intracellular Iron Metabolism :: Nramp2? i / r i Eo' Intracellular Labile Iron E ~o Fe --+: PooI[Fe(lI) I Fe(lII)?]., ~ T f' Low Molecular Weight Fe ~ k rans ernn Complexes?,...;-: Receptor Diferric» ~::::. Transferrin J ~ Hemeand ~ ~#,., "'" -.~ t.~.r.' Apo Transferrin ~.'.'" \~ ~ ~ \ -:.\ ~ Non-Heme Fe : Containing ~ Proteins High Molecular Weight Intermediates?... t '1'" ':r~~ '~~t~~~" I Fe ~ Iron into Ferritin Storage Ferritin Figure 2. Schematic representa~ tion of iron uptake from trans~ ferrin via receptor-mediated endocytosis in mammalian cells.32 Extracellular diferric transferrin (Tf) is bound by the membranebound Tf receptor and internalized via receptor-mediated endocytosis into an endosome. Iron is released from Tf probably by a decrease in ph and is then transported through the membrane by an uncharacterized transporter. A likely candidate for endosomal iron transport is Nramp Once the iron has passed through the membrane, it then enters a very poorly characterized compartment known as the intracellular labile iron pool. Apotransferrin remains bound to the transferrin receptor and is then released by exocytosis. Iron that enters the cell can be used for metabolic functioning or can be stored in ferritin. It is thought that iron in the intracellular iron pool modulates the activity of iron regulatory proteins 1 and 2 (IRP1 and IRP2). (Adapted from Richardson DR, Ponka P. The molecular mechanisms of the metabolism and transport of iron in normal and neoplastic cells. Biochim Biophys Acta 1997;1331:1-40. Copyright 1997 Elsevier Science. Used with permission.) onstrated by observations that the cellular uptake of iron through a transferrin independent pathway, and in excess of the maximum obtainable from transferrin, stimulated the synthesis of heme in erythroid cells but was without effect in nonerythroid cells Second, the expression of transferrin receptors, the gatekeepers for physiological iron acquisition, is controlled differently in erythroid cells than in nonerythroid cells. In proliferating nonerythroid cells, transferrin receptor synthesis is regulated by interplay between iron-responsive elements [IREs, present in the 3' untranslated region (UTR) 244 of transferrin receptor mrnal and iron-regulatory proteins (IRP-l and IRP-2) that "sense" iron levels in the labile pool. Upon cellular iron depletion in such cells, IRP-1 binds to IREs, thereby stabilizing transferrin receptor mrna, leading to an increased expression of transferrin receptors. On the other hand, the expansion of the labile iron pool inactivates (in terms of RNA binding) IRP-1 and leads to a degradation of IRP-2, resulting in a rapid degradation of transferrin receptor mrna [reviewed in references 23, 32, 40, 41, and in the article by Dr. Haile in this issue (41a)1. However, in hemoglobin- October 1999 Volume 318 Number 4

5 Ponko Figure 3. A model for the terminal mitochondria-associated steps TfR Endosome in the heme biosynthesis pathway.23 Three terminal enzymes, coproporphyrinogen III oxidase (CPO), protoporphyrinogen III oxidase (PPO), and ferrochelatase (FC), are associated with Heme the inner mitochondrial membrane. Coproporphyrinogen III Cytosol Copro'gen (Copro'gen) uptake by mitochondria may be mediated by peripheral-type benzodiazepine recep O.M. tors (PBR).141 The ligand carrying iron toward mitochondria is unknown. It has been recently proposed that, in erythroid cells, there may be direct interaction of I.M. the endosome with the mitochondrion followed by the transfer of iron through membrane-bound transporters to ferrochelatase.37 Obviously, much further work is needed to provide direct evidence for a transient mitochondria-endosome interaction, a model that Matrix has recently been referred to as "the kiss-and-run" hypothesis of cellular iron transfer.23 Only ferrous iron (Fe2 +) can be used by ferrochelatase but neither the site where iron reduction occurs nor the source of electrons (e-) are known. This model proposes that an iron "channel" is involved in iron transport from endosomes toward ferrochelatase and that a specific transporter is involved in the exit of heme from mitochondria. a.m., outer mitochodrial membrane; l.m., inner mitochondrial membranes. (Reprinted from Ponka P. Tissue-specific regulation of heme synthesis: distinct control mechanisms in erythroid cells. Blood 1997;89:1-25. Copyright 1997 W.B. Saunders. Used with permission.) synthesizing cells, transferrin receptor mrna levels are only slightly affected by high iron concentrations, which suggests that transferrin receptor expression is regulated differently in erythroid cells than in nonerythroid cells. This relative insensitivity of transferrin receptor mrna to destabilizing effects of iron is probably caused by the high rate of transferrin receptor gene transcription in erythroid cells.42,43 Third, as already mentioned, in erythroid cells, iron seems to be specifically targeted toward mitochondria. This is documented by experiments showing that in hemoglobin-synthesizing cells, iron acquired from transferrin continues to flow into mitochondria even when the synthesis of protoporphyrin IX is markedly suppressed.37 Based on this observation it has been proposed23,32,37 that after iron is released from transferrin in the endosome of erythroid cells, it is passed directly from protein to protein until it reaches ferrochelatase in the mitochondrion (Figure 3). This hypothesis has been supported by experiments showing that the inhibition of endosome mobility decreases the rate of 59Fe incorporation from 59Fe-Iabeled endosomes into heme.44 This, together with confocal microscopy studies44 (using transferrin and mitochondria labeled by distinct fluorescent markers), suggests that in erythroid cells, a transient mitochondria-endosome interaction may be involved in iron translocation to ferrochelatase. Fourth, in erythroid cells, "uncommitted" heme THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES inhibits the acquisition of iron from transferrin45-51 and consequently regulates the rate of heme synthesis. This guarantees that the uptake of iron by immature erythroid cells is tightly coupled with the use of iron for heme synthesis. Fifth, in erythroid cells, protoporphyrin IX levels are coupled to the availability of iron as a result of translational induction of erythroid-specific ALA-S (denoted ALA-S2) by iron. The mrna for ALA-S2 contains an IRE in its 5' UTR,52-54 which is similar to that present in the 5' UTR of ferritin mrna. As discussed in an accompanying article,41a this position of the IRE explains that the expression of ALA-S2 is translationally stimulated by iron. Regulation of Heme Biosynthesis All mammalian heme pathway enzymes have been cloned and the genes encoding these enzymes reside on different chromosomes. There are 2 different genes for ALA-S, one of which is expressed ubiquitously and the other is specific to erythroid cells (see below). These 2 genes are responsible for the occurrence of ubiquitous ("housekeeping") and erythroid-specific mrnas for ALA-S and, consequently, 2 corresponding isoforms ofthe enzyme. No tissue-specific isozyme is known for ALA-D, but there are subtle differences in 5' UTRs in "housekeeping" and erythroid ALA-D mrnas. PBD de aminase (Figure 1) exists in 2 isoforms; one is present in all cells and the other is expressed only in erythroid cells. However, these isoforms are translated from 2 245

6 mrnas that 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 2 polyadenylation signals, have been reported for coproporphyrinogen oxidase and ferrochelatase (reviewed in reference 23). It seems unlikely that any enzyme beyond ALA-S would be involved in regulating the overall rate of heme synthesis. ALA synthase, which condenses glycine with succinate upon decarboxylation, plays an important regulatory role in heme biosynthesis. However, the mechanisms controlling ALA-S expression in the liver, and perhaps in other nonerythroid cells, is dramatically different from those occurring in hemoglobin-synthesizing cells. In the early 1980s, Bishop et a156 described significant differences in kinetic and ligand-binding properties between the erythroid and non-erythroid forms of ALA-S of guinea pig, suggesting that there may be tissue-specific isozymes of ALA-S. More recently, Riddle et al57 isolated and sequenced ALA-S cdnas from erythroid cells and livers of chicken and unequivocally showed that 2 separate genes encode the erythroid and the hepatic ALA-S isozymes. The ubiquitous or ''housekeeping'' ALA-S (ALA-Sl or ALAS-N) gene has been assigned to chromosome 3p21 and the erythroid-specific CALA-S2 or ALAS-E) gene to a distal subregion ofxpll.21. The Control of Heme Synthesis in Nonerythroid Cells: Role of ALA-Sl. The promoter of the ALA-Sl gene contains a TATA box and 2 control elements located immediately upstream of the TATA box that are homologous to the binding site for the transcription factor nuclear respiratory factor-i (NRF-l), which seems to be involved in the expression of some proteins involved in oxidative phosphorylation.21 NRF-l binding sites have been identified in the promoters of cytochrome c, cytochrome c oxidase subunit VIc, and ATP synthase ')I-subunit, which suggests that NRF-l may coordinate the expression of ALA-Sl (and consequently the supply of mitochondrial heme) with the synthesis of some respiratory chain subunits.21 The seminal studies of Granick58 showed that the druginduced synthesis of ALA-S in chick embryo hepatocytes was inhibited by hemin. Although it is commonly thought that heme suppresses ALA-Sl transcription, current evidence indicates that heme primarily regulates ALA-Sl by decreasing the halflife of its mrna and by blocking translocation of ALA-Sl precursor protein into mitochondria Importantly, heme-mediated repression of ALA-Sl is responsible for rendering this enzyme the ratelimiting step in nonerythroid heme biosynthetic pathway. The Control of Heme Synthesis in Erythroid Cells: Role of Iron. In contrast to ALA-Sl, the promoter of the human erythroid-specific ALA-S2 gene contains several putative, erythroid-specific, 246 cis-acting elements, including GATA-l, the CACCC box, and the nuclear factor-erythroid-2 (NF-E2) binding sites.21 Interestingly, intron 8 of the human ALA-S2 gene also contains a GATA-l binding site and CACCC boxes and these sequences are critical for the erythroid-specific enhancer activity.62 It is highly likely that erythroid-specific transcription factors, such as GATA-l and NF-E2, control and induce ALA-S2 transcription in concert with the induction of other erythroid-specific genes. As discussed previously, erythroid-specific ALA-S2 mrna contains an IRE at its 5' UTR;52-54 this localization of the IRE dictates that the translation of erythroid ALA-S mrna depends on the availability of iron.23,32,40,41a Hence, in erythroid cells, iron acquisition, rather than ALA production, is by definition the rate-limiting step in heme synthesis. When discussing ALA-S2, it is pertinent to mention that the IREs are present in all vertebrate erythroid forms of ALA-S mrna.63 This indicates that in vertebrates, expression of the erythroid form of the enzyme and, hence, hemoglobin production can be influenced by the intracellular content of iron. Another important feature of ALA-S2, with possible implications for heme synthesis regulation in erythroid cells, is that this enzyme associates with succinyl CoA synthase.64 It seems that succinyl CoA serves as an anchor protein for erythroid-specific ALA-S and may provide succinyl CoA more efficiently to ALA-S.64 In erythroid cells, heme does not inhibit the synthesis of ALA-S2. In fact, hemin treatment of MEL cells significantly increased radiolabeled glycine incorporation into heme, which suggests that all enzymes in the heme pathway, including ALA-S2, were increased.29 Hemin was also shown to increase ALA-S2 mrna levels in MEL cells, but it is not yet known whether this increase is transcriptional,29 Interestingly, heme has also been reported to stimulate the translation of ALA-S2 in MEL cells.65 However, observations indicate that heme deficiency leads to enhanced synthesis of ALA-S2. Beaumont et a166 reported that succinylacetone (SA), a potent inhibitor of ALA-D that decreases cellular heme content, potentiates the induction of ALA-S by DMSO in MEL cells. This observation was confirmed by Elferink et al,67 who also showed that heme deficiency in DMSO-treated MEL cells did not potentiate the transcription of the ALA-S2 gene but did increase ALA-S2 protein levels. In erythroid cells, heme does not inhibit ALA-S2 activity;23 rather, it blocks the uptake of iron from transferrin Hence, although hemoglobin-synthesizing cells remain faithful to the basic principle whereby heme feedback inhibits heme synthesis,68 this feedback regulation is executed differently in erythroid cells than in nonerythroid cells. Whether heme inhibits transferrin endocytosis48,49 or iron release from transferrin46,47,5o is still controversial. Regardless of this uncertainty, the control at the October 1999 Volume 318 Number 4

7 Ponka level of iron acquisition from transferrin (ie, at the last step in the ''heme pathway"), together with the translational induction of ALA-S2 by iron (see above), guarantees very high rates of heme biosynthesis in erythroid cells. Moreover, high heme synthesis rates are attained by a positive feedback mechanism in which heme is needed for maintaining a high rate of transferrin receptor synthesis in erythroid cells. Cox et a169 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 et apo 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 mrna42,71,72 and protein42 levels in nucleated erythroid cells but had little effect on the expression of the transferrin receptor in cells that did not synthesize hemoglobin.42 More recently,73 the treatment of MEL cells with ALA was demonstrated to increase transferrin receptor mrna levels; this was accompanied by an enhanced uptake of transferrin by the cells. In addition, SA was shown to prevent the ALA-mediated enhancement of transferrin receptor mrna,73 which indicates that the effect required the conversion of ALA into heme. Interestingly, control and ALA-treated MEL cells contained identical levels ofirp-1,73 which suggests that endogenous heme (derived from added ALA) may stimulate transferrin receptor expression by a transcriptional mechanism. It is likely that the enhanced expression of transferrin receptors by heme is related to a general dependence of the erythroid differentiation program on the availability of heme. Moreover, it is possible that a heme-regulated elf -2a kinase (HRI) may be one of the factors involved in maintaining high transferrin receptor levels in erythroid cells (see below). In summary, in erythroid cells (Figure 4A), several regulatory mechanisms have apparently evolved into a cooperative system that seems to keep heme and overall hemoglobin synthesis at "maximum" levels. On the other hand, in nonerythroid cells, the feedback repression of ALA-S1 (Figure 4B) permits only very low rates of the synthesis of heme, which prevents its accumulation yet ensures adequate supply of heme for the formation of essential hemoproteins. It should be pointed out that in many nonhepatic tissues (eg, heart, adrenal gland, and testes), ALA-S is probably not suppressed by hemin.20 This suggests that the regulation of heme biosynthesis in nonhepatic tissues is distinct from that in the liver, but the nature of this difference is unknown. Interestingly, Ingi et ap4 recently reported that although hemin did not affect ALA-S activity in olfactory receptor neurons, exogenous heme clearly suppressed heme synthesis in these cells. Hence, these authors have proposed that the regulatory mechanism of heme biosynthesis by olfactory receptor neurons may be similar to that in the erythroid cells. Pathology of Heme Biosynthesis Disturbances of porphyrin metabolism are known clinically as the porphyrias, which can be caused by defects in any of the heme biosynthetic enzymes (except ALA_S1).19,20 In general, porphyrias are not associated with decreased production of heme in erythroid cells. On the other hand, defects in the ALA-S2 gene75 lead to hypochromic microcytic anemias caused by decreased heme synthesis in erythroblasts. The hallmark of this anemia is the ring sideroblast, a pathologic erythroid precursor containing excessive deposits of nonheme nonferritin iron within mitochondria, which shows perinuclear distribution accounting for the ringed appearance. It has been proposed23 that a combination of 4 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) this iron cannot be used because of the lack of protoporphyrin IX; (3) there is a lack of heme, the negative regulator ofiron uptake (see above); and (4) iron can leave mitochondria only after being inserted into protoporphyrin IX. However, in some sideroblastic anemias, in particular the primary acquired ones, there is little evidence for inhibited protoporphyrin formation. Recently, a point mutation in mitochondrial DNA coding for one of the subunits of cytochrome c oxidase has been demonstrated in hematopoietic cells of a patient with primary acquired sideroblastic anemia.76 It has been suggested that defective cytochrome c oxidase is responsible for the reduced availability of ferrous iron for ferrochelatase, leading to Fe(lII) accumulation in mitochondria. Decreased heme production causing hypochromic microcytic anemias, but without sideroblast formation, can be brought about by a genetic defect in intracellular iron translocation. Fleming et al recently identified a point mutation in the gene encoding Nramp2, a putative endosomal iron transporter (Figure 2), that causes decreased iron uptake by erythroid cells of mice with microcytic anemia (mklmk)34 and of anemic Belgrade (bib) rats.35 There are also genetically based hypochromic microcytic anemias in humans;77 one attractive possibility is that they are caused by a mutation in Nramp2 gene as well. Regulatory Role of Heme in Protein Synthesis and Cell Differentiation Role of Heme in Erythroid Differentiation MEL cells, arising in mice infected with Friend virus, can undergo erythroid differentiation when THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES 247

8 A ERYTHROID HEME SYNTHESIS REGULATION ApoTf B "HOUSEKEEPING" HEME SYNTHESIS REGULATION Fe \ I Hal (ER) \ Fe ~ Gly +SucCoA,... JI' : Gene for: ALA-S1 -<, : j. ~"'- Ferritin h,,~, l.~~~~~.1..: I.. \ l'..: ~ ~\ i - (repression) [K]-PROTO IX \1 ji 1 HEME / 1 HEMOPROTEINS they are treated with chemical inducers, such as DMSO.78 MEL cells, after their induction by DMSO, express increasing amounts of globin mrna 79 and hemoglobin.78 Moreover, differentiating MEL cells accumulate nonhemoglobin heme in their nuclei,13 and hemin treatment of MEL cells leads to a rapid accumulation of globin mrna, whereas heme deficiency leads to a decrease in globin mrna levels.29,79-82 These effects can be probably explained by heme-mediated up-regulation of NF-E2 binding activity.29,83,84 NF-E2 is an important erythroid-specific transcription factor that is involved in activation of globin genes and several other erythroid- 248 Figure 4. Distinct aspects of heme synthesis regulation in ery throid and nonerythroid cells. 23 Heme synthesis involves 8 dis tinct enzymes (see Figure 1). Differences in Fe metabolism and in genes for ALA-S probably account for the differences in the rate of heme synthesis and its regulation in erythroid cells (A) compared with that in nonerythroid cells (B). In nonerythroid cells, the rate-limiting step of heme production is that catalyzed by the ALA-Sl. The synthesis of ALA-S1 is subject to feedback inhibition by heme; this is probably one of the factors making ALA-S1 the rate-limiting enzyme of the heme pathway. On the other hand, in erythroid cells, heme inhibits neither the activity nor the synthesis of ALA-S2, but does inhibit cellular iron acquisition from transferrin without affecting its utilization for heme synthesis. This negative feedback is likely to explain the mechanism by which the availability of transferrin iron limits heme synthesis rate. It also clarifies why the system transporting iron from transferrin to ferrochelatase operates so efficiently, leaving normally mature erythrocytes with negligible amounts of nonheme iron. The different responses of ALA-S to heme probably occur because there are both ubiquitous (ALA-S1) and erythroid-specific (ALA-S2) enzymes coded for by separate genes. Because the 5' untranslated region of ALA-S2 mrna contains an IRE, the availability of iron controls ALA-S2 translation, the rate of ALA formation, and consequently the overall rate of heme synthesis in hemoglobin-synthesizing cells. Heme is also essential for globin translation (see Figure 5) and may also be involved in globin gene transcription. In contrast, in nonerythroid cells, the removal of iron from transferrin is not regulated by intracellular heme and the ubiquitous ALA-S1 mrna does not contain the IRE, so Fe availability does not control the overall heme synthesis rate. Tf, transferrin; Gly, glycine; PROTO IX, protoporphyrin IX; HO, heme oxygenase; Fe, ferrochelatase; ER, endoplasmic reticulum). (Reprinted from Ponka P. Tissue-specific regulation of heme synthesis: distinct control mechanisms in erythroid cells. Blood 1997;89:1-25. Copyright 1997 W.B. Saunders. Used with permission.) specific genes.85 Hemin has also been shown to enhance the growth of murine erythroid colonies in vitro;86,87 collectively, these observations suggest that heme promotes globin gene transcription and some other aspects of erythroid differentiation. Hemin was also shown to greatly accelerate hemoglobin accumulation in human erythroid cells whose differentiation was induced by erythropoietin,88 a physiological regulator of erythropoiesis. Interestingly, hemin increased preferentially the production of fetal hemoglobin (F) compared with adult hemoglobin; this increase was associated with a selective 2-fold elevation in -y-globin mrna levels.88 These observations may possibly be exploited clinically, because a selective increase in hemoglobin F would be of benefit for patients with genetic diseases of the f3-g1obin chain. Embryonic stem (ES) cells that lack the ALA-S2 gene89,90 have been highly valuable experimental tools in dissecting the steps at which heme is required during erythroid differentiation. Although normal ES cells can, under the proper conditions, differentiate along the erythroid pathway, ALA-S2 knockout ES cells are unable to undergo erythroid differentiation. Moreover, the induction of normal ES cells leads to the expression of mrnas for globin genes as well as for the genes in the heme synthetic October 1999 Volume 318 Number 4

9 Ponko pathway, whereas the ALA-S2 knockout ES cells fail to significantly increase the level of expression of these enzymes.89,90 These studies have provided convincing evidence that heme plays an essential role in gene expression of numerous proteins necessary for erythroid development, including the enzymes in the heme synthetic pathway. Interestingly, in contrast to the development of the iron-overloaded mitochondria in the erythroblasts of patients with defective ALA-S2 (see above), ring sideroblasts were not seen in the ALA-S2 knockout ES cells.90 Therefore it is tempting to speculate that erythroid-specific ALA-S-protein, which is totally lacking in ALA-S2 knockout ES cells, is somehow involved in mitochondrial iron uptake. Recently, Zhu et al91 reported that the role of heme in erythroid differentiation is remarkably complex. These investigators carried out differential display of K562 cells treated with heme for 4 or 8 hours versus untreated cells. They identified 11 distinct heme-regulated genes. Seven of these genes were induced by heme, 3 were repressed, and 1 was repressed at 4 h and induced at 8 h after heme treatment. Genes induced by heme included those encoding the growth-associated protein p62 (involved in Ras signaling), chaperonin Tcp20, histone H2A.Z, and a subunit of the small nuclear ribonucleoprotein complex (involved in splicing). On the other hand, heme-repressed genes included the gene for H+ -ATPase proton channel subunit and a cellular immediate-early-response gene. It is likely, as these investigators point out,91 that heme may affect a wide spectrum of regulatory factors and that heme-induced erythroid cell differentiation involves changes not only in transcription but also in cellular signaling pathways and in splicing. Role of Heme in the Translation of Erythroid Specific Proteins In addition to the effects described above, which suggest that heme may be involved in the transcription of globin and probably other erythroid-specific proteins, it has long been known that the translation of proteins in intact reticulocytes and their lysates is dependent on the availability of heme Heme deficiency inhibits protein synthesis in part through activation of the heme-regulated inhibitor (HR!). HRI is a cyclic AMP-independent protein kinase that specifically phosphorylates the a-subunit of eukaryotic initiation factor 2 (eif-2). The phosphorylation of elf -2a blocks initiation of protein synthesis (Figure 5). By contrast, hemin binds to HRI; this association inhibits the enzyme by promoting formation of disulfide bonds, perhaps between 2 HRI subunits.95 Disulfide bond formation reverses the inhibition of protein synthesis seen during heme deficiency. Interestingly, HRI contains 2 sequences that are similar to the heme regulatory motif (HRM) found in numerous other proteins whose functions e1f-2 Heme deprivation + HRI e1f-2 (P) 1 Exchange factor: eif-2b eif-2-gdp "-- Translation / initiation eif-2-gtp Figure 5. Role of heme regulated inhibitor (HR!) kinase in globin translation.97 During translation initiation, elf -2 GTP associated with Met-tRNAMet binds to 40S subunit and participates in the recognition of the initiation codon. After translation initiation, elf-2-gtp is hydrolyzed to elf-2-gdp. Because elf-2 has a much greater affinity for GDP, a guanine nucleotide exhange factor, elf-2b, is required to recycle elf-2 to the GTP-bound form. Phosphorylation of elf-a at Ser51 blocks the activity of elf-2b, reducing the level of elf-2-gtp. Heme binding to HRl inhibits the phosphorylation of elf-2a by HRl, resulting in an efficient translation of globin and probably other proteins in erythroid cells (Adapted from Wek RC. elf-2 kinases: regulators of general and gene-specific translation initiation. Trends Biochem Sci 1994;19: Copyright 1994 Elsevier Science. Used with permission.) are regulated by heme (see below). Importantly, a recent study by Chefalo et al96 indicates that there are, in fact, 2 distinct types of heme-binding sites in the HRI homodimer. The binding of heme to the first site is stable (ie, HRI is a hemoprotein), whereas the binding of heme to the second site is responsible for the rapid down-regulation ofhri activity by heme.96 The mrna for HRI is also present in uninduced MEL cells and is increased after the induction of erythroid differentiation. The accumulation of HRI mrna in differentiating MEL cells is dependent upon the presence of heme, because an inhibitor of heme synthesis markedly reduces HRI mrna accumulation.98 Hence, heme regulates HRI mrna levels and HRI activity. The expression ofhri seems to be confined to erythroid cells98; hence, HRI plays an important physiological role in the translation of globin and probably other proteins synthesized in erythroid cells. In conclusion, in erythroid cells, heme seems to enhance globin gene transcription, is essential for globin translation, and supplies the prosthetic group for hemoglobin assembly. Moreover, heme may be THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES 249

10 involved in the expression (at transcriptional as well as translational levels) of numerous other erythroidspecific proteins. Hemin was shown to inhibit activity of numerous enzymes in mature red blood cells. These include enzymes of glycolysis (aldolase), pentose phosphate shunt (6-phosphogluconate dehydrogenase), and nucleotide metabolism (adenylate kinase, pyrimidine nucleoside monophosphate kinase, and pyrimidine 5'-nucleotidase).99 Because heme content increases in both membranesloo and CytOSOllOl of red blood cells in sickle cell disease, heme-inflicted inhibition of the aforementioned enzymes may play a role in erythrocyte hemolysis in patients with hemoglobinopathies.99 Effects of Heme on Hemoprotein Synthesis in Nonerythroid Cells Because heme plays such a crucial role in controlling hemoglobin synthesis, it can be predicted that heme may regulate the synthesis of some other hemoproteins. However, our knowledge on the regulatory roles of heme in nonerythroid cells is incomplete and rather fragmented. The following text provides several interesting examples illustrating the involvement of heme in transcriptional as well as post-translational regulation of some hemoproteins. Transcriptional Effects. The involvement of heme in the transcriptional regulation of specific proteins in yeast has been amply documented. Heme levels in yeast cells are proportional to oxygen concentrations in media, and the yeast transcriptional activator heme-activating protein 1 (HAPl) activates transcription of genes that encode cytochromes in response to oxygen and/or heme HAPl seems to be tightly regulated by heme. In the absence of heme, HAPl forms a high-molecularweight complex that contains HAPl as well as 4 other cellular proteins, including heat shock proteins Hsp82 and Ydjl.105 When heme is not available, the formation of this complex is responsible for the function of HAPl as a repressor. On the other hand, heme disrupts this high-molecular-weight complex and permits HAPl to bind with high affinity to DNA as a dimer, resulting in transcriptional activation. l05 Detailed investigation of HAPl revealed 3 domains required for heme regulation: the dimerization domain, the heme domain, and the HRM7 (heme-responsive motif 7) domain that cooperate to form the high-order complex and mediate heme regulation.105 Similarly, heme seems to be a positive modulator of drug-mediated activation of the transcription of cytochrome P450 genes in rat liver. A phenobarbitone-induced, heme-modulated binding of a transcription factor to the upstream region of the cytochrome P450 gene has been demonstrated. This binding seems to correlate well with the transcriptional status of the cytochrome P450 gene in liver 250 under various conditions.106 Recently, a 65-kDa protein was isolated and shown to mediate the constitutive requirement of heme for the transcription of 1 cytochrome P450 isoform.107 However, it should be pointed out one report failed to document that heme is required for the cytochrome P450 transcription. los Although numerous known inducers of cytochrome P450 induced its mrna in rat liver, the levels of these induced mrnas were not affected by SA, an inhibitor of heme synthesis.1 s Post-Translational Effects. Because HRI is expressed only in erythroid cells, it is unlikely that heme would stimulate the translation of hemoproteins in other cell types. However, heme may be involved in the post-translational regulation of at least some hemoproteins in nonerythroid cells. One such example is myeloperoxidase, synthesized by neutrophils and monocytes. Myeloperoxidase catalyzes the H20 2-dependent, 2-electron oxidation of chloride to yield the potent oxidizing and chlorinating agent hypochlorous acid (HOCI), which plays a critical role in microbicidal activity. The primary translation product undergoes cotranslational N-linked glycosylation with subsequent insertion of heme into the peptide backbone, converting the enzymatically inactive apopromyeloperoxidase into the peroxidatively active precursor, promyeloperoxidase, which undergoes proteolytic processing into native, lysosomal myeloperoxidase (Figure 6). The inhibition of heme synthesis with SA was shown to reduce peroxidase activity and profoundly block processing of promyeloperoxidase to mature myeloperoxidase.109 Moreover, in the presence of SA, myeloperoxidase precursor remained in the endoplasmic reticulum and, compared with control cells, about 50% was degraded. The disruption in myeloperoxidase processing caused by SA was reversible by the addition of exogenous heme. 110 Hence, the availability of heme is important in the complex maturation of myeloperoxidase in the endoplasmic reticulum. More-recent studies have revealed that both calreticulin and calnexin function as molecular chaperones during myeloperoxidase biosynthesis. Interestingly, myeloperoxidase mutants unable to incorporate heme have prolonged association with calnexin, which suggests that calnexin may facilitate protein maturation in the endoplasmic reticulum of myeloid cells.1 11 It may be pertinent to mention that the treatment ofhl-60 cells with SA was demonstrated to decrease the activity of another hemoprotein, catalase,l1 but the mechanism of this decrease has not been defined. In conclusion, it can be expected the lack of heme will affect virtually every hemoprotein. Further work is needed to identify steps at which heme exerts its role in regulating transcription, translation, processing and/or transport, assembly or stability of specific hemoproteins. More recently, heme was also shown to play an important role in intracellular trafficking of thyroperoxidase, a hemoprotein involved in thyroid hormone synthesis. Faydat et al112 investigated the role October 1999 Volume 318 Number 4

11 Ponka Primary translation product apoprompo light heavy propeptide subunit (~) subunit (a), + ; ~ I signal peptide t T '-. oligosaccharides /- t heme A~~~ B[ T rl1~rll~ t F mature MPO: ~_-s-sl heavy subunit '1 (a) '1 light subunit (~) Y~~~a ~_-s-sl & ~~~~a Figure 6. The processing steps of myeloperoxidase_142 The primary translation product undergoes cotranslational cleavage of the signal peptide followed by N-linked glycosylation to generate apopromyeloperoxidase_ Initial processing also includes acquisition of heme, which yields enzymatically active promyeloperoxidase. The stepwise processing into mature dimeric myeloperoxidase includes removal of the propeptide. (Reprinted from Andersson E, Hellman L, Gullberg U, et al. The role of the propeptide for processing and sorting of human myeloperoxidase. J BioI Chem 1998;273: Copyright 1998 by the American Society for Biochemistry and Molecular Biology. Used with permission.) of heme moiety insertion in the exit of thyroperoxidase from the endoplasmic reticulum, and demonstrated that cell surface expression of thyroperoxidase was inhibited by SA. On the other hand, Fetransferrin, ALA, and (even more significantly) hemin increased cell surface activity of thyroperoxidase. Moreover, hydrogen peroxide, generated at the apical pole of thyroid cells, was shown to play an essential role in the autocatalytic covalent heme binding to the thyroperoxidase molecule.112 Other Effects of Heme The regulatory role of heme is probably not restricted to biochemical events occurring in erythroid cells or to the control of hemoprotein synthesis. Heme was shown to stimulate neurite outgrowth in neuroblastoma cells,113 increase the activities of the neurotransmitter biosynthetic enzymes tyrosine hydroxylase and choline acetyltransferase,114 enhance differentiation of adipocytes, 115 and promote the differentiation and maturation of regenerated skeletal myotubes116 as well as avian muscle cells.117 More THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES 1 recently, hemin was demonstrated to transcriptionally induce thioredoxin,118 a protein that is known to regulate intracellular redox-dependent processes as well as cell proliferation and apoptosis.118 It is difficult to judge whether these observations are physiologically relevant or merely reflect a chemical property of heme responsible for promoting cell differentiation and/or transcription of some genes. In another set of experiments, hemin was also shown to inhibit the transcription of the gene for tartrate-resistant acid phosphatase (TRAP), an ironcontaining protein encoded by the same gene that codes for uteroferrin, a placental iron transport protein. Gel shift assays showed a DNA binding protein in nuclear extracts of hemin-treated cells that was termed hemin-response element binding protein (HREBP). HREBP binds to GAGGC sequences present in the TRAP gene 5' flanking region. Interestingly, binding of HREBP to the mouse TRAP DNA sequence was inhibited by anti-hap 1 antibodies, indicating homology between the hemin-responsive factor and HAP More recently, HREBP was shown to be a heterogeneous complex composed of Ku antigen (a heterodimeric DNA binding protein playing a role in DNA repair), redox factor protein (refl, a nuclear protein that facilitates activator protein 1 DNA-binding activity) and a unique 133- kda protein whose function is unknown.12o It has been proposed that heme promotes interactions among these components, thereby facilitating hemeresponsive element binding and down-regulation of heme responsive genes.120 Heme Catabolism Although the major proportion of cellular heme is probably associated with hemoproteins, it is generally assumed that cells contain low levels of heme in an "uncommitted" pool. It is likely that they are equipped with a "sensing" system to monitor changes in the size of this pool. Recently, a number of diverse proteins known to be regulated by heme were shown to contain a short cysteine-proline-rich sequence, termed HRM, that binds heme.104 These proteins include the yeast transcriptional activator HAP1,104,105 the leader targeting sequence of ALA S,121 heme lyase (which attaches heme to cytochromes),2 HRI,95 Escherichia coli catalase (which degrades H ),122 yeast heme lyase,2 HBP23,9 and macrophage 23-kDa stress protein,lo as well as heme oxygenase It should be pointed out that the HRMs are present in almost all eukaryotic species of ALA-S, which suggests their fundamental role in the regulation of this enzyme.63 The HRM is different from sequences found in cytochromes and globins (usually histidine-methionine pair or bishistidine) that bind heme very tightly, sometimes covalently (cytochromes). It is conceivable that the presence of HMRs on the enzymes involved in heme 251

12 metabolism represents a mechanism by which "heme sensing" is accomplished. One marvels at the remarkable evolutionary forces that conveyed HMRs into the various genes coding for these enzymes. The only physiological mechanism of heme degradation is via heme oxygenase. Mammalian heme oxygenase catabolizes cellular heme to biliverdin, CO, and iron and is represented in 2 isoforms (1 and 2) encoded by separate genes In 1991, Marks et al124 drew attention to the chemical similarities between CO and NO, and proposed that CO may be a physiological regulator akin to NO. The hypothesis proposed by Marks et al has now been supported by numerous studies showing that CO can act as a neuronal125,126 or endothelial127,128 messenger, presumably by binding to iron in the heme moiety of guanylate cyclase to produce cyclic GMP. It is believed that the source of the "regulatory" CO is heme oxygenase-2. Although the concept that heme-derived CO may control physiological processes is interesting, several intriguing questions remain to be answered. Compared with erythroblasts and hepatocytes, other cell types synthesize heme with infinitesimal rates and are unlikely to contain conspicuous "free heme" pools. Hence, the identity of the substrate for heme oxygenase-2 is undefined. In addition, it is unclear how heme oxygenase is regulated to fulfill its role in producing CO for regulatory purposes. Is there a specific heme-binding protein that serves to supply heme for the aforementioned regulatory purposes? What regulates the heme biosynthetic pathway to replenish the "regulatory" pool of heme after its catabolism by heme oxygenase is unknown. This is a particularly intriguing question in neural tissues, because the heme pathway enzymes are found predominantly in Schwann cells rather than in neuronal cells.129 Moreover, some investigators are skeptical about the physiological role of CO in activating guanylate cyclase because CO was unable to break the bond between the heme Fe and proximal histidine of the enzyme, a cleavage that is considered a prerequisite for guanylate cyclase activation.130 However, it is possible that CO is turned into a potent activator of soluble guanylate cyclase by the substance YC-l, a benzyl indazole derivative.131 In the presence of YC-l, maximal activation of guanylate cyclase by CO can be achieved by formation of a 6-coordinate complex between CO and the heme, which indicates that the cleavage of the iron-histidine bond is not required for maximal activation of the enzyme.132 The heme oxygenase-l isoform is thought to provide an antioxidant defense mechanism, because it is activated in virtually all cell types, not only by heme but also by hypoxia, inflammatory cytokines, and many types of "oxidative stress."24,26,133 Recently, Poss and Tonegawa 134 provided direct support for the idea that heme oxygenase-l is an important enzymatic antioxi- 252 dant system. These investigators generated mice lacking functional heme oxygenase-l and demonstrated that, compared with normal fibroblasts, embryonic fibroblasts isolated from heme oxygenase-i-deficient pups produced high levels of oxygen free radicals in response to treatments by hemin, hydrogen peroxide, paraquat, or cadmium chloride. Moreover, young adult mice deficient in heme oxygenase-l were highly vulnerable to mortality and hepatic necrosis when challenged with endotoxin.134 It seems likely that the protective effects of heme oxygenase-l are related to this enzyme's capacity to provide biliverdin and bilirubin, both of which are antioxidant agents.135 Conversely, it can be argued that heme oxygenase also increases levels of cellular iron that, in its ferrous form, can catalyze the formation of free radicals. Although this may represent a hazard shortly after heme oxygenase induction, experimental up-regulation of heme oxygenase-i by treatment with heme affords protection against subsequent oxidative challenges.136 This can be explained by an iron-mediated increase in translation of ferritin,32,40,41 which has antioxidant effects by sequestering iron. In this connection it is pertinent to mention that heme oxygenase-l deficient adult mice develop an anemia associated with abnormally low serum iron, indicating that heme oxygenase-l is crucial for the recycling of hemoglobin iron and the release of iron from tissue stores.137 Interestingly, Yachie et ap38 recently identified the first human case of heme oxygenase-l deficiency. Mutational analysis revealed abnormalities in both heme oxygenase-l alleles. Complete deletion of exon 2 was found in the maternal allele and a 2-nucleotide deletion was present within exon 3 of the paternal allele. The patient, a 6-year-old boy, suffered marked growth retardation and developmental delay associated with erythrocyte fragmentation, persistent intravascular hemolysis, and marked abnormality of the coagulation/fibrinolysis system. Some of these changes may be related to extremely high serum heme concentration (490 p.m, normally not detectable) found in this patient. Nonheme iron deposition was found in renal and hepatic tissues, and a lymphoblastoid cell line derived from the patient was extremely sensitive to hemininduced cell injury. Hence, numerous phenotypic changes found in this patient,138 such as growth retardation, tissue iron overload, and vulnerability to oxidative stress, are similar to those found in heme oxygenase-i deficient mice.134,137 Recent studies by Dennery et al139 have revealed that heme oxygenase-2 also plays an important role in protection against oxidative injury. These investigators demonstrated that mutant mice lacking heme oxygenase-2 had substantially increased mortality with chronic hyperoxic exposure. Moreover, they had significantly increased markers of oxidative injury even before hyperoxic exposure. Furthermore, during hyperoxia, lung hemoproteins and iron content were significantly increased without in- October 1999 Volume 318 Number 4

13 Ponko creased ferritin, a situation that suggests an accumulation of redox-active iron. Interestingly, the absence of hemoxygenase-2 was associated with hemoxygenase-l induction, which, however, was not sufficient to protect the mutant mice against hyperoxia-induced oxidative injury.l39 It can be expected that further work exploiting hemoxygenase-l- and -2 deficient mice will provide exciting new information on heme and iron metabolism as well as on the pathogenesis of oxidative injury. Juckett et ap40 clearly document that nitric oxide (NO) readily nitrosylates intracellular free heme and prevents its degradation by heme oxygenase. It seems that the nitrosylation of heme diminishes both the induction of heme oxygenase synthesis and its enzymatic activity.l40 Hence, this study documented another intriguing link between heme and NO, in addition to those that have been known before: NO synthase is a cytochrome P450 type hemoprotein and NO elicits many of its physiological actions by activating guanylate cyclase via its binding to the heme present in this enzyme. Recently, Garrick et al. l43 developed a chromatographic method that enables the measurement of the amount of "free" heme in reticulocytes (immature blood cells still capable of hemoglobin synthesis). These investigators showed that most of the "free" heme is the result of biosynthesis and that the "free" heme pool behaves as an intermediate, with a half-life of about 2 h. Acknowledgments I thank Michael Parniak for reading the manuscript and for helpful discussions and suggestions and Andrew Dancis for useful advice. Special thanks are due to David Boldt, editor of this Symposium, for meticulous reading of the manuscript and helpful suggestions. I am indebted to Sandy Fraiberg and Rhona Rosenzweig for excellent editorial assistance. Because of space restrictions, I had to select the references and sincerely apologize to those colleagues whose work I did not cite. References 1. Dumont ME, Cardillo TS, Hayes MK, et al. Role of cytochrome c heme lyase in mitochondrial import and accumulation of cytochrome c in Saccharomyces cerevisiae. Mol Cell BioI 1991;11: Steiner H, Kispal G, Zollner A, et al. Heme binding to a conserved Cys-Pro-Val motif is crucial for the catalytic function of mitochondrial heme lyases. J BioI Chern 1996;271: Wang X, Dumont ME, Sherman F. Sequence requirements for mitochondrial import of yeast cytochrome c. J BioI Chern 1996;271: Howe G, Merchant S. Role of heme in the biosynthesis of cytochrome C6' J BioI Chern 1994;269: Grandchamp B, Bissel DM, Licko V, et al. Formation and disposition of newly synthesized heme in adult rat hepatocytes in primary culture. J BioI Chern 1981;256: THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES 6. Ponka P, Borova J, Neuwirt J. Accumulation of heme in mitochondria from rabbit reticulocytes with inhibited globin synthesis. Biochim Biophys Acta 1973;304: Vincent SH, Muller-Eberhard U. A protein of the Z class of liver cytosolic proteins in the rat that preferentially binds heme. J BioI Chern 1985;260: Harvey JW, Beutler E. Binding of heme by glutathione S-transferase: a possible role of the erythrocyte enzyme. Blood 1982;60: Iwahara S-I, Satoh H, Song D-X, et al. Purification, characterization, and cloning of a heme-binding protein (23 kda) in rat liver cytosol. Biochemistry 1995;34: Ishii T, Yamada M, Sato H, et al. Cloning and characterization of a 23-kDa stress-induced mouse peritoneal macrophage protein. J BioI Chern 1993;268: Ishii T, Kawane T, Taketani S, et al. Inhibition of the thiol-specific antioxidant activity of rat liver MSP23 protein by hemin. Biochem Biophys Res Commun 1995;216: Taketani S, Adachi Y, Kohno H, et al. Molecular characterization of a newly identified heme-binding protein induced during differentiation of murine erythroleukemia cells. J BioI Chern 1998;273: Lo SC, Aft R, Mueller GC. Role of nonhemoglobin heme accumulation in the terminal differentiation of Friend erythroleukemia cells. Cancer Res 1981;41: Bottomley SS, Muller-Eberhard U. Pathophysiology of heme synthesis. Semin Hematol 1988;25: Dailey HA, editor. Biosynthesis of heme and chlorophylls. New York: McGraw-Hill; May BK, Bhasker CR, Bawden MJ, et al. Molecular regulation of 5-aminolevulinate synthase. Mol BioI Med 1990;7: Bonkovsky HL. Key relationships of hepatic heme and iron metabolism. Gastroenterol Intern 1991;4: Moore MR. Biochemistry of porphyria. Int J Biochem 1993; 25: Wyckoff EE, Kushner JP. Heme biosynthesis, the porphyrias and the liver. In: Arias 1M, Boyer JL, Fausto N, et ai, editors. The liver: biology and pathobiology. 3rd ed. New York: Raven; p Kappas A, Sassa S, Galbraith RA, et al. The porphyrias. In: Scriver CR, Beaudet AL, Sly WS, et ai, editors. The metabolic and molecular bases of inherited diseases. New York: McGraw-Hill; p May BK, Dogra SC, Sadlon TJ, et al. Molecular regulation of heme biosynthesis in higher vertebrates. Prog Nucleic Acid Res Mol BioI 1995;51: Dailey HA. Ferrochelatase. In: Hausinger RP, Eichhorn GL, Marzilli LG, editors. Mechanisms ofmetallocenter assembly. New York: VCH; p Ponka P. Tissue-specific regulation of heme synthesis: distinct control mechanisms in erythroid cells. Blood 1997;89: Maines MD. Heme oxygenase. Clinical applications and functions. Boca Raton (FL): CRC Press; Maines MD. The heme oxygenase system: A regulator of second messenger gases. Annu Rev Pharmacol Toxicol1997; 37: Abraham NG, Drummond GS, Lutton JD, et al. The biological significance and physiological role of heme oxygenase. Cell Physiol Biochem 1996;6: Smith A. Transport of tetrapyrroles: mechanisms and biological and regulatory consequences. In: Dailey HA, ed. Biosynthesis of heme and chlorophylls. New York: McGraw-Hill Publishing Co.; p Lutton JD, Levere RD, Abraham NG. Physiologic role of heme and cytochrome P450 in hemetopoietic cells. Proc Soc Exp BioI Med 1991;196: Sassa S, Nagai T. The role of heme in gene expression. Int J Hematol 1996;63: Goya N, Miyazalo S, Kodate S, et al. A family of congenital atransferrinemia. Blood 1972;40:

14 31. Bernstein SE. Hereditary hypotransferrinemia with hemosiderosis, a murine disorder resembling human atransferrinemia. J Lab Clin Med 1987;110: Richardson DR, Ponka P. The molecular mechanisms of the metabolism and transport of iron in normal and neoplastic cells. Biochim Biophys Acta 1997;1331: Bali PK, Zak 0, Aisen P. A new role for the transferrin receptor in the release of iron from transferrin. Biochemistry 1991;30: Fleming MD, Trenor CC III, Su MA, et al. Microcytic anaemia mice have a mutation in Nramp2, a candidate iron transporter gene. Nat Genet 1997;16: Fleming MD, Romano MA, Su MA, et al. Nramp2 is mutated in the anemic Belgrade (b) rat: evidence of a role for Nramp2 in endosomal iron transport. Proc Nat! Acad Sci U S A 1998;95: Gunshin H, Mackenzie B, Berger UV, et al. Cloning and characterization of a mammalian proton-coupled metal-ion transporter. Nature 1997;388: Richardson DR, Ponka P, Vyoral D. Distribution of iron in reticulocytes after inhibition of heme synthesis with succinylacetone. Examination of cytoplasmic and mitochondrial intermediates involved in iron metabolism. Blood 1996;87: Ponka P, Schulman HM. Acquisition of iron from transferrin regulates reticulocyte heme synthesis. J Bioi Chern 1985;260: Laskey JD, Ponka P, Schulman HM. Control of heme synthesis during Friend cell differentiation: role of iron and transferrin. J Cell Physiol 1986;129: Klausner RD, Rouault TA, Harford JB. Regulating the fate of mrna: the control of cellular iron metabolism. Cell 1993;72: Hentze MW, Kiihn LC. Molecular control of vertebrate iron metabolism: mrna-based regulatory circuits operated by iron, nitric oxide, and oxidative stress. Proc Nat! Acad Sci U S A 1996;93: ' 41a.Haile DJ Regulation of genes of iron metabolism by the iron-response proteins. Am J Med Sci 1999;318: Chan RYY, Seiser C, Schulman HM, et al. Regulation of transferrin receptor mrna expression: distinct regulatory features in erythroid cells. Eur J Biochem 1994;220: Chan L-NL, Gerhardt EM. Transferrin receptor is hyperexpressed and transcriptionally regulated in differentiating erythroid cells. J Bioi Chern 1992;267: Zhang A-S, Sheftel AD, Hajek J, et al. Delivery of iron into erythroid cell mitochondria requires endosome motility. Blood 1997;90(Suppl 1):6a. 45. Ponka P, Neuwirt J. Regulation of iron entry into reticulocytes. I. Feedback inhibitory effect of heme on iron entry into reticulocytes and on heme synthesis. Blood 1969;33: Ponka P, Neuwirt J, Borova J. The role of heme in the release of iron from transferrin in reticulocytes. Enzyme 1974;17: Schulman HM, Martinez-Medellin J, Sidloi R. The reticulocyte-mediated release of iron and bicarbonate from transferrin: effect of metabolic inhibitors. Biochem Biophys Acta 1974;343: Iacopetta B, Morgan EH. Hemin inhibits transferrin endocytosis in immature erythroid cells. Biochem Biophys Acta 1984;805: Cox TM, O'Donnell MW, Aisen P, et al. Hemin inhibits internalization of transferrin by reticulocytes and promotes phosphorylation of the membrane transferrin receptor. Proc Natl Acad Sci USA 1985;82: Ponka P, Schulman HM, Martinez-Medellin J. Haem inhibits iron uptake subsequent to endocytosis of transferrin in reticulocytes. Biochem J 1988;251: Hradilek A, Neuwirt J. Inhibition of cellular iron uptake by haem in mouse erythroleukaemia cells. Br J Haematol 1989; 73: Dierks P. Molecular biology of eukaryotic /l-aminolevulinate synthase. In: Dailey HA, ed. Biosynthesis of heme and chlorophylls. New York, McGraw-Hill, pp Cox TC, Bawden MJ, Martin A, et al. Human erythroid /l-aminolevulinate synthase: promoter analysis and identification of an iron-responsive element in the mrna. EMBO J 1991;10: Dandekar T, Stripecke R, Gray NK, et al. Identification of a novel iron-responsive element in murine and human erythroid /l-aminolevulinic acid synthase mrna. EMBO J 1991; 10: Deleted in proof. 56. Bishop DF, Kitchen H, Wood WA. Evidence for erythroid and nonerythroid forms of /l-aminolevulinate synthase. Arch Biochem Biophys 1981;206: Riddle RD, Yamamoto M, Engel JD. Expression of /l-aminolevulinate synthase in avian cells: separate genes encode erythroid-specific and non-specific isozymes. Proc Nat! Acad Sci USA 1989;86: Granick S. The induction in vitro of the synthesis of /l-aminolevulinic acid synthetase in chemical porphyria: a response to cetain drugs, sex hormones and foreign chemicals. J Bioi Chern 1966;241: Tyrrell DL, Marks GS. Drug-induced porphyrin biosynthesis. V. Effect of protohemin on the transcriptional and posttranscriptional phases of /l-aminolevulinic acid synthetase induction. Biochem Pharmacol 1972;21: Kikuchi G, Hayashi N. Regulation by heme of synthesis and intracellular translocation of delta-aminolevulinate synthase in the liver. Mol Cell Biochem 1981;37: Hamilton JW, Bement WJ, Sinclair PR, et al. Heme regulates hepatic 5-aminolevulinate synthase mrna expression by decreasing mrna half-life and not by altering its rate of transcription. Arch Biochem Biophys 1991;289: Surinya KH, Cox TC, May BK. Identification and characterization of a conserved erythroid-specific enhancer located in intron 8 of the human 5-aminolevulinate synthase 2 gene. J Bioi Chern 1998;273: Duncan R, Faggart MA, Roger AJ, et al. Phylogenic analysis of the 5-aminolevulinate synthase gene. Mol Bioi Evol 1999;16: Furuyama K, Sassa S. The erythroid-specific /l-aminolevulinate synthase associates with succinyl CoA synthase: an anchoring mechanism to the inner mitochondrial membrane. Blood 1998;92(Suppl 1,Part 1):669a. 65. Melefors 0, Goossen B, Johansson HE, et al. Translational control of 5-aminolevulinate synthase mrna by ironresponsive elements in erythroid cells. J Bioi Chern 1993;268: Beaumont C, Dyebach J-C, Grandchamp B, et al. Effects of succinylacetone on dimethylsufoxide-mediated induction of heme pathway enzymes in mouse Friend virus-transformed erythroleukemia cells. Exp Cell Res 1984;154: Elferink CJ, Sassa S, May BK. Regulation of5-aminolevulinate synthase in mouse erythroleukemic cells is different from that in liver. J Bioi Chern 1988;263: Burnham BF, Lascelles J. Control of porphyrin biosynthesis through a negative feedback mechanism. Studies with preparation of /l-aminolevulinic acid synthetase and /l-aminolevulinic acid dehydratase from Rhodopseudomonas spheroides. Biochem J 1963;87: Cox TM, O'Donnell MW, Aisen P, et al. Biosynthesis of the transferrin receptor in rabbit reticulocytes. J Clin Invest 1985;76: Grdisa M, Mathew A, Johnstone RM. Expression and loss of the transferrin receptor in growing and differentiating HD3 cells. J Cell PhysioI1993;155: Battistini A, Coccia E-M, Marziali G, et al. Intracellular heme coordinately modulates globin chain synthesis, transferrin receptor numbers, and ferritin content in differentiating Friend erythroleukemia cells. Blood 1991;78: Hradilek A, Fuchs A, Neuwirt J. Inhibition of heme syn- 254 October 1999 Volume 318 Number 4

15 Ponko thesis decreases transferrin receptor expression in mouse erythroleukemia cells. J Cell Physiol 1992;150: Lok NC, Ponka P. Role of heme in transferrin receptor expression in erythroid cells. Blood 1997;90(Suppl1):6a-7a. 74. Ingi T, Chiang G, Ronnett GV. The regulation of heme turnover and carbon monoxide biosynthesis in cultured primary rat olfactory receptor neurons. J Neurosci 1966;16: May A, Fitzsimons E. Sideroblastic anaemia. Baillieres Clin Haematol 1994;7: Gattermann N, Retzlaff S, Wang YL, et al. Heteroplasmic point mutations of mitochondrial DNA affecting subunit I of cytochrome c oxidase in two patients with acquired idiopathic sideroblastic anemia. Blood 1997;90: Hartman KR, Baker JA. Microcytic anemia with iron malabsorption: an inherited disorder of iron metabolism. Am J Hematol 1996;51: Friend C, Scher W, Holland JG, Sato T. Hemoglobin synthesis in murine virus-induced leukemic cells in vitro: stimulation of erythroid differentiation by dimethyl sulfoxide. Proc Natl Acad Sci USA 1971;68: Ross J, Ikawa Y, Leder P. Globin messenger-rna induction during erythroid differentiation of cultured leukemia cells. Proc Natl Acad Sci USA 1972;69: Ross J, Sautner D. Induction of globin mrna accumulation by hemin in cultured erythroleukemia cells. Cell 1976;8: Dabney BJ, Beaudet AL. Increase in globin chains and globin mrna in erythroleukemia cells in response to hemin. Arch Biochem Biophys 1977;179: Fuchs 0, Ponka P, Borova J, et al. Effect of heme on globin messenger RNA synthesis in spleen erythroid cells. J Supramol Struct Cell Biochem 1981;15: Solomon WB, Lin C-H, Palma J, et al. Suppression of a cellular differentiation program by phorbol esters coincides with inhibition of binding a cell-specific transcription factor (NF -E2) to an enhancer element required for expression of an erythroid-specific gene. J BioI Chern 1993;268: Nagai T, Igarashi K, Furuyama K, et al. Upregulation of the erythroid transcription factor, NF-E2, by DMSO and hemin: distinctive mechanisms. Blood 1995;86:676a. 85. Andrews NC. Erythroid transcription factor NF-E2 coordinates hemoglobin synthesis. Pediatr Res 1994;36: Porter PN, Meints PH, Mesner K. Enhancement of erythroid colony growth in culture by hemin. Exp Hematol 1979;7: Monette FC, Holden SA. Hemin enhances the in vitro growth of primitive erythroid progenitor cells. Blood 1982;60: Fibach E, Kollia P, Schechter AN, et al. Hemin-induced acceleration of hemoglobin production in immature cultured erythroid cells: preferential enhancement of fetal hemoglobin. Blood 1995;85: Yin X, Dailey HA. Erythroid 5-aminolevulinate synthase is required for erythroid differentiation in mouse embryonic stem cells. Blood Cells Mol Dis 1998;24: Harigae H, Suwabe N, Weinstock PH, et al. Deficient heme and globin synthesis in embryonic stem cells lacking the erythroid-specific 8-aminolevulinate synthase gene. Blood 1998;91: Zhu Y, Hon T, Zhang L. Heme initiates changes in the expression of a wide array of genes during the early erythroid differentiation stage. Biochem Biophys Res Commun 1999; 258: Bruns GP, London 1M. The effect of hemin on the synthesis of globin. Biochem Biophys Res Commun 1965;18: Zucker WV, Schulman HM. Stimulation of globin chain initiation by hemin in the reticulocyte cell-free system. Proc Natl Acad Sci USA 1968;59: London 1M, Levin DH, Matts RL, et al. Regulation of protein synthesis. In: Boyer PD, editor. The enzymes. Vol 18. New York: Academic Press; p THE AMERICAN JOURNAL OF THE MEDICAL SCIENCES 95. Chen J-J, London 1M. Regulation of protein synthesis by heme-regulated eif-2 alpha kinase. Trends Biochem Sci 1995;20: Chefalo PJ, Oh J, Rafie-Kolpin M, et al. Heme-regulated eif-2alpha kinase purifies as a hemoprotein. Eur J Biochem 1998;258: Wek RC. eif-2 kinases: regulators of general and genespecific translation initiation. Trends Biochem Sci 1994;19: Crosby JS, Lee K, London 1M, et al. Erythroid expression of the heme-regulated eif-2a kinase. Mol Cell BioI 1994;14: Zerez CR, Hseih JW, Tanaka KR. Inhibition of red blood cell enzymes by hemin: a mechanism for hemolysis in hemoglobinopathies. Trans Assoc Am Physicians 1987;100: Shaklai N, Shuiro Y, Rabizadek E, et al. Accumulation and drainage of hemin in red cell membrane. Biochim Biophys Acta 1985;821: Liu SC, Zhai S, Palek J. Detection of hemin release during hemoglobin S denaturation. Blood 1988;71: Pfeifer K, Prezant T, Guarente L. Yeast HAP1 activator binds to two upstream activation sites of different sequence. Cell 1987;49: Creusot F, Verediere J, Gaisne M, et al. CYP1 (HAP!) regulator of oxygen-dependent gene expression in yeast. I. Overall organization of the protein sequence displays several novel structural domains. J Mol BioI 1988;204: Zhang L, Guarente L. Heme binds to a short sequence that serves a regulatory function in diverse proteins. EMBO J 1995;14: Zhang L, Hach A, Wang C. Molecular mechanism governing heme signaling in yeast: a higher-order complex mediates heme regulation of the transcriptional activator HAP1. Mol Cell BioI 1998;18: Rangarajan PN, Padmanaban G. Regulation of cytochrome P-450b/e gene expression by a heme- and phenobarbitone-modulated transcription factor. Proc Natl Acad Sci U SA 1989;86: Sultana S, Nirodi CS, Ram N, et al. A 65-kDa protein mediates the positive role of heme in regulating the transcription of CYP2B1IB2 gene in rat liver. J BioI Chern 1997;272: Srivastava G, Bawden MJ, Hansen AJ, et al. Heme may not be a positive regulator of cytochrome-p450 gene expression. Eur J Biochem 1989;178: NauseefWM, McCormick S, Yi H. Roles of heme insertion and the mannose-6-phosphate receptor in processing of the human myeloid lysosomal enzyme, myeloperoxidase. Blood 1992;80: Pinnix IB, Guzman GS, Bonkovsky HL, et al. The posttranslational processing of myeloperoxidase is regulated by the availability of heme. Arch Biochem Biophys 1994;312: NauseefWM, McCormick SJ, Goedken M. Coordinated participation of calreticulin and calnexin in the biosynthesis of my eloper oxidase. J BioI Chern 1998;273: Fayadat L, Niccoli-Sire P, Lanet J, et al. Role of heme in intracellular trafficking ofthyroperoxidase and involvement of H 20 2 generated at the apical surface of thyroid cells in autocatalytic covalent heme binding. J BioI Chern 1999;274: Ishii DN, Maniatis GM. Haemin promotes rapid neurite outgrowth in cultured mouse neuroblastoma cells. Nature 1978;274: Bonyhady RD, Hendry la, Hill CE, et al. Effects of haemin on neurones derived from the neural crest. Dev Neurosci 1982;5: Chen JJ, London 1M. Hemin enhances the differentiation of mouse 3T3 cells to adipocytes. Cell 1981;26: Funanage VL, Schroedl NA, Moses PA, et al. Hemin 255

16 enhances differentiation and maturation of cultured regenerated skeletal myotubes. J Cell PhysioI1989;141: Smith SM, Schroedl NA. Heme-containing compounds replace chick embryo extract and enhance differentiation in avian muscle cell culture. In Vitro Cell Dev BioI 1992;28A: Leppa S, Pirkkala L, Chow SC, et al. Thioredoxin is transcriptionally induced upon activation of heat shock factor 2. J BioI Chern 1997;272: Reddy SV, Alcantara 0, Roodman GD, et al. Inhibition of tartrate-resistant acid phosphatase gene expression by hemin and protoporphyrin IX. Identification of a heminresponsive inhibitor of transcription. Blood 1996;88: Reddy SV, Alcantara 0, Boldt OH. Analysis of DNA binding proteins associated with hemin-induced transcriptional inhibition. The hemin response element binding protein is a heterogeneous complex that includes the Ku protein. Blood 1998;91: Lathorp JT, Timko MP. Regulation by heme ofmitochondrial protein transport through a conserved amino acid motif. Science 1993;259: Triggs-Raine BL, Doble BW, Mulvey MR, et al. Nucleotide sequence of katg, encoding catalase HPI of Escherichia coli. J Bacteriol 1988;170: McCoubrey WK Jr, Huang TJ, et al. Heme oxygenase-2 is a hemoprotein and binds heme through heme regulatory motifs that are not involved in heme catalysis. J BioI Chern 1997 ;272: Marks GS, Brien JF, Nakatsu K, et al. Does carbon monoxide have a physiological function? Trends Pharmacol Sci 1991;12: Verma A, Hirsch DJ, Glatt CE, et al. Carbon monoxide: a putative neural messenger. Science 1993;259: Glaum SR, Miller RJ. Zinc protoporphyrin-ix: the effects of metabotropic glutamate receptor activation in the rat nucleus tractus solitarii. Mol Pharmacol 1993;43: Cook MN, Nakatsu K, Marks GS, et al. Heme oxygenase activity in the adult rat aorta and liver as measured by carbon monoxide formation. Can J Physiol Pharmacol 1995; 73: Zakhary R, Gaine SP, Dinerman JL, et al. Heme oxygenase 2: endothelial and neuronal localization and role in endothelium-dependent relaxation. Proc Natl Acad Sci U S A 1996;93: Whetsell WO Jr, Sassa S, Bickers D, et al. Studies on porphyrin-heme biosynthesis in organotypic cultures of chick dorsal root ganglion. 1. Observations on neuronal and non-neuronal elements. J Neuropathol Exp NeuroI1978;37: Deinum G, Stone JR, Babcock GT, et al. Binding of nitric oxide and carbon monoxide to soluble guanylate cyclase as observed with resonance Raman spectroscopy. Biochemistry 1996;35: Friebe A, Schultz G, Koesling D. Sensitizing soluble guanylyl cyclase to become a highly CO-sensitive enzyme. EMBO J 1996;15: Stone JR, MarIetta MA. Synergistic activation of soluble guanylate cyclase by YC-1 and carbon monoxide: implications for the role of cleavage of the iron-histidine bond during activation by nitric oxide. Chern BioI 1998;5: Applegate LA, Luscher P, Tyrrell RM. Induction of heme oxygenase: a general response to oxidant stress in cultured mammalian cells. Cancer Res 1991;51: Poss KD, Tonegawa S. Reduced stress defense in heme oxygenase 1-deficient cells. Proc Nat! Acad Sci USA 1997; 94: Stocker R, Yamamoto Y, McDonagh AF, et al. Bilirubin is an antioxidant of possible physiological importance. Science 1987;235: Balla J, Jacob HS, Balla G, et al. Endothelial-cell heme uptake from heme proteins: induction of sensitization and desensitization to oxidant damage. Proc Nat! Acad Sci USA 1993;90: Poss KD, Tonegawa S. Heme oxygenase 1 is required for mammalian iron reutilization. Proc N atl Acad Sci USA 1997;94: Yachie A, Niida Y, Wada T, et al. Oxidative stress causes enhanced endothelial cell injury in human heme oxygenase-1 deficiency. J Clin Invest 1999;103: Dennery PA, Spitz DR, Yang G, et al. Oxygen toxicity and iron accumulation in the lungs of mice lacking heme oxygenase-2. J Clin Invest 1998;101: Juckett M, Zheng Y, Yuan H, et al. Heme and the endothelium: effects of nitric oxide on catalytic iron and heme degradation by heme oxygenase. J BioI Chern 1998;273: Taketani S, Kohno H, Furukawa T, et al. Involvement of peripheral-type benzodiazepine reporters in the intracellular transport of heme and porphyrins. J Biochem 1995;117: Andersson E, Hellman L, Gullberg U, et al. The role of the propeptide for processing and sorting of human myeloperoxidase. J BioI Chern 1998;273: Garrick MD, Scott D, Romano MA, et al. Evidence for and consequences of chronic heme deficiency in Belgrade rat reticulocytes. Biochim Biophys Acta 1999;1449: October 1999 Volume 318 Number 4