Modulation of cellular iron metabolism by hydrogen peroxide. Effects of H2O2 on the expression and function of iron-responsive element-containing mRNAs in B6 fibroblasts

Cellular iron uptake and storage are coordinately controlled by binding of iron-regulatory proteins (IRP), IRP1 and IRP2, to iron-responsive elements (IREs) within the mRNAs encoding transferrin receptor (TfR) and ferritin. Under conditions of iron starvation, both IRP1 and IRP2 bind with high affinity to cognate IREs, thus stabilizing TfR and inhibiting translation of ferritin mRNAs. The IRE/IRP regulatory system receives additional input by oxidative stress in the form of H(2)O(2) that leads to rapid activation of IRP1. Here we show that treating murine B6 fibroblasts with a pulse of 100 microm H(2)O(2) for 1 h is sufficient to alter critical parameters of iron homeostasis in a time-dependent manner. First, this stimulus inhibits ferritin synthesis for at least 8 h, leading to a significant (50%) reduction of cellular ferritin content. Second, treatment with H(2)O(2) induces a approximately 4-fold increase in TfR mRNA levels within 2-6 h, and subsequent accumulation of newly synthesized protein after 4 h. This is associated with a profound increase in the cell surface expression of TfR, enhanced binding to fluorescein-tagged transferrin, and stimulation of transferrin-mediated iron uptake into cells. Under these conditions, no significant alterations are observed in the levels of mitochondrial aconitase and the Divalent Metal Transporter DMT1, although both are encoded by two as yet lesser characterized IRE-containing mRNAs. Finally, H(2)O(2)-treated cells display an increased capacity to sequester (59)Fe in ferritin, despite a reduction in the ferritin pool, which results in a rearrangement of (59)Fe intracellular distribution. Our data suggest that H(2)O(2) regulates cellular iron acquisition and intracellular iron distribution by both IRP1-dependent and -independent mechanisms.     

Multiple red cell ferritin mRNAs, which code for an abundant protein in the embryonic cell type, analyzed by cDNA sequence and by primer extension of the 5'-untranslated regions

Ferritin maintains iron in a bioavailable, nontoxic form for vertebrates and invertebrates, higher plants, fungi, and bacteria; the protein is formed from two classes of subunits (H and L) in ratios which vary in different cell types. Ferritin may be an abundant, differentiation-specific protein or a "housekeeping" protein. The red cells of embryos are specialized for iron storage and have abundant ferritin; iron regulates the synthesis of ferritin in such cells translationally by recruitment of stored, ferritin mRNA and by translational competition. To characterize mRNA regulated in such a manner, we prepared cDNA from reticulocytes of bullfrog tadpoles, a readily available source of embryonic red cells; moreover, no protein sequence information was available for nonmammalian ferritin. An almost full-length (817 base pairs) cDNA (pJD5F12) was isolated and sequenced, the 5' end was analyzed by primer extension, and the cloned DNA was used as a hybridization probe. We have shown that ferritin mRNA is stored in the cytoplasm and that the 5' end of the mRNA is heterogeneous. The 5'-untranslated region of ferritin mRNA consisted of 143 nucleotides in the major (65%) species and 146 or 152 in the minor species (approximately 17% each). (Heterogeneity is characteristic of some other abundant mRNAs, e.g. globin, which is also translationally regulated.) Since excess iron had no detectable effect on the heterogeneity of the 5' end of ferritin mRNA, the feature is more likely associated with mRNA abundance and/or cell specialization than translational control. In the bullfrog, as in humans and rats, ferritin is encoded by multiple genomic sequences (four to eight) which specify proteins of considerable homology. For example, 75 of the 81 amino acids present in all mammalian ferritins sequenced are also present in the frog; the overall homology between frogs and humans or rats is 59-66%. Ferritin H and L subunits in humans are distinct (overall homology 56%) and appear to have diverged from a common precursor relatively recently. In contrast, ferritin H and L subunits have high homology in tadpole red cells, determined by hybrid select translation, which suggests that bullfrog red cell ferritin may be close to the primordial sequence.     

A ferritin-responsive internal ribosome entry site regulates folate metabolism

Cytoplasmic serine hydroxymethyltransferase (cSHMT) enzyme levels are elevated by the expression of the heavy chain ferritin (H ferritin) cDNA in cultured cells without corresponding changes in mRNA levels, resulting in enhanced folate-dependent de novo thymidylate biosynthesis and impaired homocysteine remethylation. In this study, the mechanism whereby H ferritin regulates cSHMT expression was determined. cSHMT translation is shown to be regulated by an H ferritin-responsive internal ribosome entry site (IRES) located within the cSHMT mRNA 5'-untranslated region (5'-UTR). The cSHMT 5'-UTR exhibited IRES activity during in vitro translation of bicistronic mRNA templates, and in MCF-7 and HeLa cells transfected with bicistronic mRNAs. IRES activity was depressed in H ferritin-deficient mouse embryonic fibroblasts and elevated in cells expressing the H ferritin cDNA. H ferritin was shown to interact with the mRNA-binding protein CUGBP1, a protein known to interact with the alpha and beta subunits of eukaryotic initiation factor eIF2. Small interference RNA-mediated depletion of CUGBP1 decreased IRES activity from bicistronic templates that included the cSHMT 3'-UTR in the bicistronic construct. The identification of this H ferritin-responsive IRES represents a mechanism that accounts for previous observations that H ferritin regulates folate metabolism.     

A universal oligonucleotide probe for acetylcholine receptor genes. Selection and sequencing of cDNA clones for the mouse muscle beta subunit

A region of 25 nucleotides is highly conserved in genes coding for the alpha, beta, gamma, and delta subunits of the nicotinic acetylcholine receptor (AChR) of human, mouse, calf, chicken, and Torpedo. Based on this observation, a 2-fold degenerate oligonucleotide was synthesized and used as a probe to screen a cDNA library made from a mouse myogenic cell line. Clones coding for the beta, gamma, and delta subunits were identified by the probe. The protein sequence deduced from the beta subunit clones codes for a precursor polypeptide of 501 amino acids with a calculated molecular weight of 56,930 daltons, which includes a signal peptide of 23 amino acids. The protein sequence and structural features of the beta subunits of mouse, calf, and Torpedo are conserved. A clone coding for the mouse gamma subunit was isolated, and its identity was confirmed by alignment of its sequence to previously published cDNA sequences for the mouse and calf gamma subunits. The clone contained approximately 200 nucleotides more at its 3' end untranslated region than a mouse gamma clone recently described. Northern blot analysis, utilizing as probes these beta and gamma subunit cDNAs and previously characterized alpha and delta subunit cDNAs, shows that the steady-state levels of the four AChR mRNAs increase coordinately during terminal differentiation of cultured C2 and C2i mouse myoblasts. The increase in mRNA levels can account for the rise of cell surface receptors during myogenesis and suggests that the muscle AChR genes may be regulated during development by a common mechanism. Utilization of this oligonucleotide probe should prove useful for screening a variety of libraries made from different species and tissues which are known to express AChRs.     

Ferritin H and L mRNAs in human neoplastic tissues

The typical tissue isoferritin pattern varies during neoplastic transformation, usually shifting toward a more acidic profile. To investigate the molecular basis of this phenomenon, we have analyzed the steady-state levels of the H and L mRNAs in several neoplastic tissues. By using specific probes for the two ferritin subunits, we have found, in three different adenocarcinomas and in a case of Hodgkin lymphoma, a two- to four-fold increase of the H and L mRNA levels compared to those found in normal human liver.     

Transferrin receptors,iron transport and ferritin metabolism in friend erythroleukemia cells

Four aspects of iron metabolism were studied in cultured Friend erythroleukemia cells before and after induction of erythroid differentiation by dimethyl sulfoxide. (1) The binding of ¹²⁵I-labeled transferrin was determined over a range of transferrin concentrations from 0.5 to 15 μM. Scatchard analysis of the binding curves demonstrated equivalent numbers of transferrin binding sites per cell: 7.78 ± 2.41 · 10⁵ in non-induced cells and 9.28 ± 1.57 · 10⁵ after 4 days of exposure to dimethyl sulfoxide. (2) The rate of iron transport was determined by measuring iron uptake from ⁵⁹Fe-labeled transferrin. Iron uptake in non-induced cells was approx. 17 000 molecules of iron/cell per min; 24 h after addition of dimethyl sulfoxide it increased to 38 000, and it rose to maximal levels of approx. 130 000 at 72 h. (3) Heme synthesis, assayed qualitatively by benzidine staining and measured quantitatively by incorporation of ⁵⁹Fe or [2-¹⁴C]glycine into cyclohexanone-extracted or crystallized heme, was not detected until 3 days after addition of dimethyl sulfoxide, when 12% of the cells were stained by benzidine and 6 pmol ⁵⁹Fe and 32 pmol [2-¹⁴C]glycine were incorporated into heme per 10⁸ cells/h. After 4 days, 60% of the cells were benzidine positive and 34 pmol ⁵⁹Fe and 90 pmol [2-¹⁴C]glycine were incorporated into heme per 10⁸ cells/h. (4) The rate of incorporation of ⁵⁹Fe into ferritin, measured by immunoprecipitation of ferritin by specific antimouse ferritin immunoglobulin G, rose from 4.4 ± 0.6 cells to 18.4 ± 1.3 pmol ⁵⁹Fe/h per 10⁸ cells 3 days after addition of dimethyl sulfoxide, and then fell to 11.6 ± 3.1 pmol 4 days after dimethyl sulfoxide when heme synthesis was maximal. These studies indicate that one or more steps in cellular iron transport distal to transferrin binding is induced early by dimethyl sulfoxide and that ferritin may play an active role in iron delivery for heme synthesis.     

Ferritin is a translationally regulated heat shock protein of avian reticulocytes

Heat-shock avian reticulocytes exhibit enhanced synthesis of a greater than 450-kDa protein. Biochemical, immunochemical, and visual criteria were used to identify this protein as the iron storage protein ferritin. The 21-kDa ferritin subunits synthesized during heat shock are similar in size and pI to the subunits that are constitutively synthesized. The 2-6-fold heat shock-induced increase in ferritin synthesis appears to be regulated at the translational level as it is insensitive to actinomycin D. Northern and dot-blot hybridization analyses of cytoplasmic RNAs with avian H-ferritin cDNA fragments support the contention that the heat shock stimulation of ferritin synthesis is translationally regulated. These latter studies demonstrate that the heat shock-induced synthesis of ferritin does not involve a change in the amount of total cytoplasmic ferritin mRNAs, but rather appears to entail a translocation of cytoplasmic H-ferritin mRNAs from a polyribosome-free, translationally repressed state to a polyribosome-associated, translationally active state. These results suggest that thermally stressed avian reticulocytes have a critical and functional need for the synthesis of additional ferritin and that its enhanced synthesis, unlike the new and/or enhanced synthesis of the well-established avian heat shock proteins, is regulated wholly at the translational level.     

Tumor necrosis factor-alpha and interleukin 1-alpha regulate transferrin receptor in human diploid fibroblasts. Relationship to the induction of ferritin heavy chain

We have studied transferrin receptor expression in MRC5 human fibroblasts in response to tumor necrosis factor-alpha (TNF, cachectin) or interleukin 1-alpha (IL-1). Treatment of exponentially growing MRC5 cells with these cytokines led to a 3-4-fold increase in transferrin receptor mRNA and a coordinate increase in transferrin receptor protein by 24 h. Under these conditions, stimulation of [3H]thymidine incorporation was minimal, suggesting that the induction of transferrin receptor by TNF and IL-1 is mediated by a growth-independent regulatory mechanism. A study of the time course of this response showed that cytokine-mediated increases in transferrin receptor mRNA and protein proceeded after a lag of 12-24 h. A simultaneous analysis of the effects of TNF and IL-1 on ferritin in MRC5 cells was also performed. Ferritin L mRNA levels were unchanged. However, induction of ferritin H mRNA was seen within 4 h, preceding the induction of the transferrin receptor. The synthesis of ferritin H (but not ferritin L) protein peaked at 8 h after TNF or IL-1 treatment, followed by a rapid decrease in both ferritin H and L protein synthesis. As ferritin H synthesis declined, levels of transferrin receptor protein increased, reaching a maximum by 24 h. These results suggest that the cytokine-dependent induction of ferritin H and subsequent increase in the transferrin receptor are related and possibly interdependent events. This study demonstrates that the complex role of TNF and IL-1 in iron homeostasis includes modulation of the transferrin receptor.     

Modulation of cytokeratin and actin gene expression and fibrillar organization in cultured rat hepatocytes.

Intermediate filaments of rat hepatocytes are composed of cytokeratins 8 and 18 (CK8 and CK18, respectively). Recent work from our laboratory has indicated a close relationship between the synthesis of these cytokeratins, their organization into intermediate filaments, and the promotion of growth and differentiation of cultured rat hepatocytes by insulin, epidermal growth factor, and dexamethasone. In the present study, we examined the mRNA expression, level of protein synthesis, and fibrillar distribution of cytokeratins 8 and 18 and actin in hepatocytes, isolated from normal and dexamethasone-injected rats and cultured as monolayers or spheroids in the presence of insulin, or from normal rat hepatocytes, cultured as monolayers in the presence of dexamethasone, insulin, and dimethyl sulfoxide. The CK8 mRNA level was lower in hepatocytes isolated from noninjected rats and cultured as either monolayers or spheroids, than in those from dexamethasone-injected rats. However, the CK18 mRNA level varied in a manner that was different from that of CK8 mRNA, showing that the modes of expression of the two genes were independent. The various changes in hepatocyte culture conditions led to variations in albumin mRNA levels that largely followed those observed in CK8 mRNA levels. In the case of actin, the amount of mRNAs varied from relatively high levels in hepatocyte monolayers to extremely low levels in hepatocyte spheroids, even though in both cases the cells were isolated from dexamethasone-injected rats. These changes in mRNA levels did not necessarily correlate with changes in the synthesis of cytokeratins 8 and 18, and actin. Changes in culture conditions induced a major reorganization in the distribution of cytokeratin intermediate filaments and actin filament between the region near the surface membrane and the cytoplasm. The most divergent patterns in cytokeratin intermediate filaments and actin filament distributions were observed between hepatocytes cultured as spheroidal aggregates and as monolayers in the presence of dimethyl sulfoxide. The former condition resulted in patterns of cytokeratin and actin gene expression and fibrillar organization that best matched those in situ. In the latter condition, inappropriate patterns were obtained, in spite of the fact that dimethyl sulfoxide treated hepatocytes are known to exhibit survival and functional activities equivalent to that of hepatocyte spheroids. These results demonstrate for the first time that the survival and functional activity (i.e., albumin production) of rat hepatocytes in vitro is not necessarily correlated with a particular pattern of cytokeratin and actin gene expression and fibrillar arrangement.     

Both subunits of rat liver ferritin are regulated at a translational level by iron induction.

A few hours after administering iron to rats, liver ferritin synthesis increases several fold. However, Northern blot analysis with cDNA probes for ferritin light (L) and heavy (H) subunit mRNAs failed to show an increase in total population of either messenger. Cytoplasmic distribution of ferritin messages was therefore investigated in control and iron administered rats killed at 3.5 hours. The liver post-mitochondrial supernatant was fractionated on a sucrose gradient to separate polyribosomes, monosomes, ribosomal subunits and cell sap. RNA extracted from each fraction and analyzed using Northern blotting showed that 65% of the total mRNA population for each subunit was present in the cell sap of control rats, presumably as mRNP particles since ribosomal RNA was absent from this fraction. After iron administration, these reserves of free mRNA were recruited onto the polysomes, reducing the free mRNA pool to 15% of the total. We interpret this to be due to activation of blocked ferritin messages on entry of iron into the cell.     

Rat preprocarboxypeptidase H. Cloning, characterization, and sequence of the cDNA and regulation of the mRNA by corticotropin-releasing factor

Carboxypeptidase H is a putative post-translational processing enzyme which removes basic amino acid residues from intermediates during protein hormone biosynthesis. A 2.2-kilobase pair cDNA was shown to contain the complete amino acid sequence of rat carboxypeptidase H. The deduced amino acid sequence revealed that the enzyme was synthesized as preprocarboxypeptidase H, a precursor form of 476 amino acid residues. Preprocarboxypeptidase H contained a putative hydrophobic signal peptide and a short propeptide which contained 5 adjacent Arg residues at its C terminus. Northern blot analysis identified a single carboxypeptidase H mRNA of approximately 2.3 kilobases in brain, pituitary, and heart, as well as in mouse AtT20 cells. No carboxypeptidase H mRNA was detected in rat liver, spleen, kidney, lung, and mammary gland. Sequence analysis of cDNAs obtained from different rat tissues suggested that a single mRNA encodes an identical carboxypeptidase in several tissues. Treatment of AtT20 cells with dexamethasone decreased the levels of both carboxypeptidase H and preproopiomelanocortin (POMC) mRNAs by approximately 30%. Exposure of the dexamethasone-treated cells to corticotropin-releasing factor effected a 2- to 3-fold increase in the carboxypeptidase H and POMC mRNA levels relative to those of dexamethasone-treated cells exposed to control medium. This suggests that the mRNA levels of POMC and one of its putative post-translational processing enzymes, carboxypeptidase H, are co-regulated by corticotropin-releasing factor and steroid hormones.