Ferric-salicylaldehyde isonicotinoyl hydrazone, a synthetic iron chelate, alleviates defective iron utilization by reticulocytes of the Belgrade rat

The Belgrade rat has a hypochromic, microcytic anemia inherited as an autosomal recessive mutation. Although transferrin binds normally to reticulocytes and internalizes normally, iron accumulation into cells and heme is much slower than normal. We have investigated the role of the transferrin cycle in this mutant by bypassing transferrin iron delivery with the iron chelate ferric salicylaldehyde isonicotinoyl hydrazone (Fe-SIH). Fe-SIH increases iron uptake into heme by Belgrade reticulocytes, restoring it almost to normal levels. This increase indicates that Fe-SIH delivers iron to a step in iron utilization that is after the Belgrade defect. Depleting reticulocytes of transferrin did not alter these observations. Failure to achieve above normal rates of iron incorporation could indicate damage due to chronic intracellular iron deficiency. Also, iron delivery by Fe-SIH restored globin synthesis to near-normal levels in Belgrade reticulocytes. The rates of glycine incorporation into porphyrin and heme in Belgrade reticulocytes incubated with Fe2-transferrin or Fe-SIH paralleled the rates of iron incorporation into heme. These data are consistent with the concept that iron availability limits protoporphyrin formation in rat reticulocytes. The protoporphyrin used for heme synthesis is provided by de novo synthesis and not by a pool of pre-existing protoporphyrin. The Belgrade defect occurs in the movement of iron from transferrin to a step prior to the ferrous state and insertion into heme. This defect diminishes the synthesis of heme and, consequently, that of protoporphyrin and globin.     

Evidence for and consequences of chronic heme deficiency in Belgrade rat reticulocytes

The Belgrade rat has a microcytic, hypochromic anemia inherited as an autosomal recessive trait (gene symbol b). Transferrin-dependent iron uptake is defective because of a mutation in Nramp2 (now DMT1, also called DCT1), the protein responsible for endosomal iron efflux. Hence, Belgrade reticulocytes are iron deficient. We show that a chromatographic method is able to measure the amount of 'free' heme in reticulocytes. Most of the 'free' heme is the result of biosynthesis. Succinylacetone, an inhibitor of heme synthesis, decreases the level of 'free' heme and cycloheximide, an inhibitor of globin synthesis, increases the 'free' heme level. In a pulse-chase experiment with 59Fe-transferrin, the 'free' heme pool behaves as an intermediate, with a half-life of just over 2 h. Belgrade reticulocytes contain about 40% as much 'free' heme as do heterozygous or homozygous reticulocytes. This deficiency of 'free' heme slows initiation of translation in Belgrade reticulocytes by increasing the level of an inhibitor of initiation. Thus the Belgrade rat makes a whole animal model available with chronic heme deficiency.     

Inhibition of peptide chain initiation by a nonhemin-regulated translational repressor from Friend leukemia cells.

A nonhemin-regulated translational repressor protein has been purified partially from the postribosomal supernatant fraction of Friend leukemia cells grown in the absence of dimethylsulfoxide. This repressor inhibits protein synthesis in lysates from rabbit reticulocytes or Friend leukemia cells and in a fractionated system using Artemia salina ribosomes, reticulocyte mRNA, and soluble components from reticulocytes. In contrast, the hemin-controlled repressor from reticulocytes does not inhibit protein synthesis in lysates from Friend leukemia cells. The repressor from Friend leukemia cells has no effect on poly(U)-directed synthesis of polyphenylalanine using reticulocyte ribosomes nor on the extension and release of nascent globin chains that were initiated in intact reticulocytes. It does not block completion of peptides on ribosomes isolated from reticulocytes incubated with NaF nor does it inhibit initiation factor-dependent formation of methionylpuromycin, but it inhibits globin mRNA-dependent methionylvaline synthesis. The Friend leukemia cell repressor promotes peptide synthesis-dependent breakdown of polysomes in reticulocyte lysates that appears to involve inhibition of ribosome reattachment to mRNA during peptide chain initiation. It is concluded that the Friend leukemia cell repressor blocks peptide initiation at a point between the addition of methionyl-tRNA^fMet to the ribosomal initiation complex and the NaF-sensitive reaction.     

Heat inhibition of reticulocyte protein synthesis. Evidence for a mechanism independent of the hemin-controlled repressor

Incubation of rabbit reticulocytes at 45 degrees C results in a prompt but reversible decrease in protein synthesis and a concomitant conversion of polyribosomes to smaller aggregates. These effects occur even in the presence of 100 micrometer hemin in the incubation medium. There is also inhibition of heme synthesis but this occurs at a later time than the effect on protein synthesis. The inhibtion of heme synthesis results from a decrease in activity of beta-aminolevulinic acid synthetase. This decrease of heme synthesis appears to be secondary to the inhibition of protein synthesis with resultant accumulation of intramitochondrial heme (which will decrease beta-aminolevulinic acid synthetase activity). An inhibitor of reticulocyte cell-free protein synthesis formed in the postribosomal supernatants of cells incubated at both 45 and 37 degrees C but not at 0 degrees C. No temporal or quantitative differences in the amount of this inhibitor from cells treated at either 37 or 45 degrees C was apparent. The inhibitor was not found in the fraction where the hemin-controlled repressor is isolated. It is concluded that heat inactivation of intact reticulocyte protein synthesis does not depend upon a decrease in heme synthesis, heme concentration or generation of the hemin-controlled repressor. Furthermore, it appears that the inhibitor formed in the post-ribosomal supernatant cannot be the sole cause of the heat inhibition of protein synthesis.     

Effect of polyamines on hemoglobin and globin chain synthesis

Exposure of reticulocytes to 25 mM spermidine and spermine stimulates the incorporation of 14C leucine into hemoglobin and globin polypeptide chains. These two polyamines stimulate beta chain synthesis to a greater extent than they stimulate alpha chain synthesis, thus resulting in a decreased synthetic alpha/beta globin ratio. Because of the low permeability of the red cell membrane to these polycations, less than 1.0% of extracellular spermidine and spermine accumulate within the erythrocyte, thus these effects occur at intracellular polyamine levels that are physiological. Putrescine progressively decreases hemoglobin synthesis at extracellular concentrations greater than 10 mM without affecting the synthetic alpha/beta ratio. These findings suggest a role for polyamines in the fine tuning of the alpha/beta globin ratio at the translational level.     

Iron utilization in rabbit reticulocytes: A study using succinylacetone as an inhibitor of heme synthesis

We have investigated the effect of succinylacetone (4,6-dioxoheptanoic acid) on hemoglobin synthesis and iron metabolism in reticulocytes. Succinylacetone, 0.1 and 1 mM, inhibited [2-¹⁴C]glycine incorporation into heme by 91.2 and 96.4%, respectively, and into globin by 85 and 90.2%, respectively. 60 μM hemin completely prevented the inhibition of globin synthesis by succinylacetone, indicating that succinylacetone inhibits specifically the synthesis of heme. Added porphobilinogen, but not δ-aminolevulinic acid, partly overcame the inhibition of ⁵⁹Fe incorporation into heme caused by succinylacetone suggesting that the drug inhibits δ-aminolevulinic acid dehydratase in reticulocytes. Succinylacetone, 10 μM, 0.1 and 1 mM, inhibited ⁵⁹Fe incorporation into heme by 50, 90 and 93%, respectively, but stimulated reticulocyte ⁵⁹Fe uptake by about 25–30%. In succinylacetone-treated cells ⁵⁹Fe accumulates in a fraction containing plasma membranes and mitochondria as well as cytosol ferritin and an unidentified low molecular weight fraction obtained by Sephacryl S-200 chromatography. Reincubation of washed succinylacetone- and ⁵⁹Fe-transferrin-pretreated reticulocytes results in the transfer of ⁵⁹Fe from the particulate fraction (plasma membrane plus mitochondria) into hemoglobin and this process is considerably stimulated by added protoporphyrin. Although the nature of the iron accumulated in the membrane-mitochondria fraction in succinylacetone-treated cells is unknown some of it is utilizable for hemoglobin synthesis, while cytosolic ferritin iron would appear to be mostly unavailable for incorporation into heme.     

Control of heme synthesis during Friend cell differentiation: role of iron and transferrin

In many types of cells the synthesis of delta-aminolevulinic acid (ALA) limits the rate of heme formation. However, results from our laboratory with reticulocytes suggest that the rate of iron uptake from transferrin (Tf), rather than ALA synthase activity, limits the rate of heme synthesis in erythroid cells. To determine whether changes occur in iron metabolism and the control of heme synthesis during erythroid cell development Friend erythroleukemia cells induced to erythroid differentiation by dimethylsulfoxide (DMSO) were studied. While added ALA stimulated heme synthesis in uninduced Friend cells (suggesting ALA synthase is limiting) it did not do so in induced cells. Therefore the possibility was investigated that, in induced cells, iron uptake from Tf limits and controls heme synthesis. Several aspects of iron metabolism were investigated using the synthetic iron chelator salicylaldehyde isonicotinoyl hydrazone (SIH). Both induced and uninduced Friend cells take up and utilize Fe for heme synthesis directly from Fe-SIH without the involvement of transferrin and transferrin receptors and to a much greater extent than from saturating levels of Fe-Tf (20 microM). Furthermore, in induced Friend cells 100 microM Fe-SIH stimulated 2-14C-glycine incorporation into heme up to 3.6-fold as compared to the incorporation observed with saturating concentrations of Fe-Tf. In contrast, Fe-SIH, even when added in high concentrations, did not stimulate heme synthesis in uninduced Friend cells but was able to do so as early as 24 to 48 h following induction. In addition, contrary to previous results with rabbit reticulocytes, Fe-SIH also stimulated globin synthesis in induced Friend cells above the level seen with saturating concentrations of transferrin. These results indicate that some step(s) in the pathway of iron from extracellular Tf to protoporphyrin, rather than the activity of ALA synthase, limits and controls the overall rate of heme and possibly hemoglobin synthesis in differentiating Friend erythroleukemia cells.     

Biosynthesis of heme in immature erythroid cells. The regulatory step for heme formation in the human erythron

Heme formation in reticulocytes from rabbits and rodents is subject to end product negative feedback regulation: intracellular "free" heme has been shown to control acquisition of transferrin iron for heme synthesis. To identify the site of control of heme biosynthesis in the human erythron, immature erythroid cells were obtained from peripheral blood and aspirated bone marrow. After incubation with human 59Fe transferrin, 2-[14C]glycine, or 4-[14C]delta-aminolevulinate, isotopic incorporation into extracted heme was determined. Addition of cycloheximide to increase endogenous free heme, reduced incorporation of labeled glycine and iron but not delta-aminolevulinate into cell heme. Incorporation of glycine and iron was also sensitive to inhibition by exogenous hematin (Ki, 30 and 45 microM, respectively) i.e. at concentrations in the range which affect cell-free protein synthesis in reticulocyte lysates. Hematin treatment rapidly diminished incorporation of intracellular 59Fe into heme by human erythroid cells but assimilation of 4-[14C]delta-aminolevulinate into heme was insensitive to inhibition by hematin (Ki greater than 100 microM). In human reticulocytes (unlike those from rabbits), addition of ferric salicylaldehyde isonicotinoylhydrazone, to increase the pre-heme iron pool independently of the transferrin cycle, failed to promote heme synthesis or modify feedback inhibition induced by hematin. In human erythroid cells (but not rabbit reticulocytes) pre-incubation with unlabeled delta-aminolevulinate or protoporphyrin IX greatly stimulated utilization of cell 59Fe for heme synthesis and also attenuated end product inhibition. In human erythroid cells heme biosynthesis is thus primarily regulated by feedback inhibition at one or more steps which lead to delta-aminolevulinate formation. Hence in man the regulatory process affects generation of the first committed precursor of porphyrin biosynthesis by delta-aminolevulinate synthetase, whereas in the rabbit separate regulatory mechanisms exist which control the incorporation of iron into protoporphyrin IX.     

Control of haemoglobin synthesis. The effects of iron deprivation, cobalt and temperature on the rate and extent of globin synthesis in reticulocytes.

A detailed examination of the kinetics of protein synthesis in rabbit reticulocytes in the presence of the iron chelating agent 2,2'-dipyridyl showed that between 30 degrees C and 42 degrees C there were characteristically two distinct phases of protein synthesis. An initial phase (I), in which no inhibition of protein synthesis was apparent, was followed by a gradual decline in the rate of protein synthesis leading to the second phase (II) in which protein synthesis occurred at a linear but inhibited rate for extended periods. In contrast, below 30 degrees C, incubation in the presence of dipyridyl caused no inhibition of protein synthesis. Between 30 degrees C and 42 degrees C the duration and amount of protein synthesis occurring in phase I before the onset of inhibition were inversely related of the inhibition as was the final rate of incorporation in phase II. During phase II, a partial reversal of the inhibition caused by dipyridyl was obtained by lowering the incubation temperature. This resulted in a burst of protein synthesis at the uninhibited rate until the amount of protein synthesis reached the same level as that in reticulocytes maintained continuously with dipyridyl at the lower incubation temperature. This burst of synthesis was observed in reticulocytes which had been held in phase II for as long as 90 min. It was also possible to reverse the inhibition by addition of haemin to cells in phase II. At any particular incubation temperature, a fixed number of rounds of protein synthesis had to occur before the onset of phase II became apparent. By the use of puromycin we showed that this was not a requirement for the synthesis of globin or of any other protein. We believe that this critical amount of protein synthesis reflects the residual ability of reticulocytes to initiate new protein chains in the absence of concurrent haem synthesis. Reticulocytes preincubated in the presence of cobaltous ions showed almost no inhibition of protein synthesis upon subsequent incubation with dipyridyl. The results are compared to those obtained in reticulocyte lysates and are discussed in terms of current theories to account for control of protein chain initiation by haemin.     

Deficient heme synthesis as the cause of noninducibility of hemoglobin synthesis in a Friend erythroleukemia cell line.

Friend cells of the line Fw are not induced to accumulate substantial amounts of hemoglobin and to become benzidine-positive when treated with butyric acid or other inducers, except in the presence of exogenous hemin. The cells are shown to have a deficiency in heme synthesis since they require exogenous hemin during the period of maximal hemoglobin synthesis; since endogenous heme synthesis cannot be induced to the level found in normal inducible Friend cells, even after hemoglobin synthesis has been induced by hemin and butyric acid and the hemin has then been withdrawn; since they are not inducible for ferrochelatase (heme synthetase) activity; and since they accumulate free globin chains after stimulation with butyric acid in the absence of hemlin. Comparison of globin synthesis and globin mRNA content of the cells shows that globin synthesis is not controlled by the hemin-controlled repressor of protein synthesis (HCR) nor by any specific translational control of globin synthesis by hemlin.     

Regulation of translation in rabbit reticulocytes and mouse L-cells; comparison of the effects of temperature.

Various parameters of protein synthesis were analyzed in rabbit reticulocytes exposed to various temperatures for up to five hours. Between 10 degrees C and 40 degrees C total protein synthesis exhibited two different apparent activation energies (36 kcal/mole, 10-24 degrees C; 22 kcal/mole, 24-40 degrees C), as did protein elongation and release (35 kcal/mole, 10-25 degrees C; 12 kcal/mole, 25-40 degrees C). However, the level of polysomes remained essentially unchanged between 0 degrees C and 42 degrees C which implies that the activation energy for polypeptide initiation is quite similar to that for elongation and is also biphasic. This situation is different from that in cultured mouse L-cells where the polysome level is dependent on temperatures. Nevertheless, reticulocytes and L-cells appear to be similar in their temperature dependence of initiation and in their rate of elongation (5-6 amino acids/second at 36 degrees C.     

Non-transferrin donors of iron for heme synthesis in immature erythroid cells

The mechanism of iron uptake from several iron-containing compounds by transferrin-depleted rabbit reticulocytes and mouse spleen erythroid cells was investigated. Iron complexes of DL-penicillamine, citrate and six different aroyl hydrazones may be utilized by immature erythroid cells for hemoglobin synthesis, although less efficiently than iron from transferrin. HTF-14, a monoclonal antibody against human transferrin, reacts with rabbit transferrin and inhibits iron uptake and heme synthesis by rabbit reticulocytes. HTF-14 had no significant effect on iron uptake and heme synthesis when non-transferrin donors of iron were examined. Ammonium chloride (NH4Cl) increases intracellular pH and blocks the release or utilization of iron from the internalized transferrin. NH4Cl only slightly affected iron incorporation and heme synthesis from non-transferrin donors of iron. Hemin inhibited transferrin iron uptake and heme synthesis, but had a much lesser effect on iron incorporation and heme synthesis from non-transferrin donors of iron. These results allow us to conclude that transferrin-depleted reticulocytes take up iron from all of the examined non-transferrin iron donors without the involvement of the transferrin/transferrin receptor pathway.     

Maturation-associated loss and incomplete de novo synthesis of the transferrin receptor in peripheral sheep reticulocytes: response to heme and iron

Hemin, but not iron, in the culture medium stimulates the maturation-associated loss of the transferrin receptor from sheep reticulocytes (t1/2 for loss approximately 6 hr) and its appearance in a population of externalized vesicles. A similar pattern is seen with nucleoside binding (a measure of the nucleoside transporter), where hemin increases the loss of binding activity from the cells during culture, concomitant with an increase in nucleoside binding in the externalized vesicles. Sheep reticulocytes retain the ability to synthesize the transferrin receptor, but the 35S-labeled receptors are not detected in released vesicles. Whereas hemin stimulates the loss of 35S-labeled transferrin receptors from the cell (t1/2 for loss approximately 20 hr), nonheme iron is more effective than heme. This difference in response of native and 35S-labeled receptor to hemin and iron supplements appears to be related to the differences in the two classes of receptors. Although the 35S-labeled receptor binds transferrin and both native and 35S-labeled peptides comigrate after chemical deglycosylation, the 35S-receptor is approximately 2 kD smaller than the native receptor and fails to acquire its complete size even when chased for up to 24 hr. Moreover, the 35S-labeled receptor is not expressed at the cell surface, but is retained in a nonrecycling compartment, where it is insensitive to digestion by trypsin at both 0 degrees C and 37 degrees C.     

The incorporation of N15 glycine by avian erythrocytes and reticulocytes in vitro

A study of the incorporation of glycine-N¹⁵ by chicken red cells in vitro has shown that: 1. There is no detectable nitrogen turnover in the histone or desoxyribonucleic acid of erythrocytes or reticulocytes. 2. Hemoglobin synthesis in the nucleated reticulocyte proceeds at 2 to 3 times the rate observed in the mature erythrocyte. 3. The uptake of glycine-N¹⁵ by heme is 9 to 14 times the corresponding uptake into hemoglobin, and 12 to 20 times the calculated uptake into globin. 4. Maturation of the red cell results in a decline in the rate of synthesis of both heme and globin, but the deceleration is much more marked in globin. synthesis. 5. No significant differences could be detected in the low N¹⁵ incorporations of nuclear and cytoplasmic hemoglobins.     

Stabilization of apoglobin by low temperature increases yield of soluble recombinant hemoglobin in Escherichia coli.

Accumulation of soluble recombinant hemoglobin (rHb1.1) in Escherichia coli requires proper protein folding, prosthetic group (heme) addition, and subunit assembly. This served as a new model system for the study of the effects of temperature, protein synthesis rates, and protein accumulation rates on protein solubility in E. coli. Fermentation expression of rHb1.1 at 30 degrees C from cultures containing a medium or high globin gene dosage (pBR-based or pUC-based plasmids with rHb1.1 genes under the control of the tac promoter) was compared. A medium gene dosage resulted in rHb1.1 accumulating to approximately 7% of the soluble cell protein, of which 78% was soluble. A high globin gene dosage resulted in a > or = 3-fold increase in total globin to 23 to 24% of the soluble cell protein, but 70% was insoluble. Accumulation of insoluble rHb1.1 began immediately upon induction. The proportion of rHb1.1 from the high globin gene dosage that accumulated as insoluble globin was affected by reducing (i) the inducer concentration and (ii) the temperature. Reducing the inducer concentration reduced globin synthesis up to eightfold but increased the proportion of soluble rHb1.1 to 93%. In contrast, total globin protein synthesis was barely affected by reducing the temperature from 30 to 26 degrees C, while soluble globin accumulation increased > 2-fold to approximately 15% of the soluble cell protein. The contrast between the effects of reducing rates of protein synthesis and accumulation and those of reducing temperature suggests that lower temperature stabilizes one or more folding intermediates. We propose a simplified physical model which integrates protein synthesis, folding, and heme association. This model shows that temperature-dependent apoglobin stability is the most critical factor in soluble rHb1.1 accumulation.     

The effects of inhibition of heme synthesis on the intracellular localization of iron in rat reticulocytes

These studies assessed the fate and localization of incoming iron in 6-8-day rat reticulocytes during inhibition of heme synthesis by succinylacetone. Succinylacetone inhibition of heme synthesis increased iron uptake by increasing the rate of receptor recycling without affecting receptor KD for transferrin, transferrin uptake, or total receptor number. Its net effect was to amplify the number of surface transferrin receptors by recruitment of receptors from an intracellular pool. Despite increased iron influx in inhibited cells, only 2-4% of total incoming iron was diverted into ferritin. The majority of incoming iron (65-80%) in succinylacetone-inhibited cells was recovered in the stroma, where ultrastructural and enzymic analyses revealed it to be accumulated mainly in mitochondria. Intramitochondrial iron (70-75%) was localized mainly in the inner membrane fraction. Removal of succinylacetone restored heme synthesis, utilizing iron accumulated within mitochondria for its support. Thus, inhibition of heme synthesis in rat reticulocytes results in accumulation of incoming iron in a functional mobile intramitochondrial precursor iron pool used directly for heme synthesis. Under normal conditions, there is no significant intracellular or intramitochondrial iron pool in reticulocytes, which are therefore dependent upon continuous delivery of transferrin-bound iron to maintain heme synthesis. Ferritin plays an insignificant role in iron metabolism of reticulocytes.     

The effect of various chelating agents on the mobilization of iron from reticulocytes in the presence and absence of pyridoxal isonicotinoyl hydrazone

The chelating agent pyridoxal isonicotinoyl hydrazone (PIH) has recently been shown to mobilize ⁵⁹Fe from reticulocytes loaded with non-heme ⁵⁹Fe. In this study, various chelating agents were tested for their ability to effect the mobilization of iron from reticulocytes by PIH. They fall into several groups. The largest group includes chelators such as citrate, ethylenediaminetetracetic acid and desferrioxamine, which fail to affect PIH-induced iron mobilization and do not mobilize iron per se. Either these chelators do not enter reticulocytes or they do not take up iron from PIH-Fe complexes. The second group includes chelators such as 2,2′-bipyridine, 1,10-phenanthroline, bathophenanthroline sulfonate and N,N′-ethylenebis(o-hydroxyphenylglycine) which inhibit PIH-induced iron mobilization from reticulocytes and, when added together with PIH, induce radioiron accumulation in an alcohol-soluble fraction of reticulocytes. It appears that these chelators enter the cell and compete with PIH for ⁵⁹Fe(II), but having bound iron are unable to cross the cell membrane. Spectral analysis suggests that Fe(II) chelators such as 2,2′-bipyridine and 1,10-phenanthroline remove iron from Fe(II)PIH but are not able to do so from Fe(III)PIH. Then there are compounds such as 2,3-dihydroxybenzoic acid and catechol which potentiate PIH-induced iron mobilization although they are unable to mobilize iron from reticulocytes by themselves. Lastly, there is a group of miscellaneous compounds which include chelators that either potentiate the iron-mobilizing effect of PIH as well as mobilizing iron from reticulocytes by themselves (tropolone), or that reduce PIH-induced iron mobilization while themselves having an iron-mobilizing effect (N,N′-bis(2,3-dihydroxybenzoyl)-1,6-diaminohexane). In further experiments, heme was found to stimulate globin synthesis in reticulocytes, the heme synthesis of which was inhibited by PIH, suggesting that PIH is probably not toxic to the cells.     

Inhibition of heme synthesis in bone marrow cells by succinylacetone: effect on globin synthesis

The effects of 4,6-dioxoheptanoic acid (succinylacetone, SA), an inhibitor of delta-aminolevulinic acid dehydratase, on total iron uptake, heme synthesis, and globin synthesis were studied in rat marrow cells in culture in order to examine the coordination of heme and globin synthesis. SA inhibited heme synthesis in both control and erythropoietin-stimulated cells in a dose-dependent fashion; at 10(-3) M, inhibition was complete, whereas at 10(-7) M, there was no significant effect. Inhibition of total iron uptake was also dose-dependent although, at 10(-3) M, it was not complete. The inhibition of heme synthesis by SA was partially overcome by addition of 10(-4) M porphobilinogen or protoporphyrin IX. SA caused an almost complete suppression of globin formation in both erythropoietin-stimulated and unstimulated cells as early as five hours after the addition of the inhibitor. When inhibition of heme synthesis was incomplete, globin synthesis was partially inhibited. These results indicate that heme synthesis is required for erythropoietin-mediated induction of globin synthesis in cultured bone marrow cells.