The folate and thiamine transport proteins of lactobacillus casei

Two separate binding proteins, one specific for folate and the other for thiamine, have been isolated from membrane fragments of Lactobacillus casei. Purification to homogeneity was achieved by fractionation of the Triton-solubilized proteins with microgranular silica (Quso G-32) and Sephadex G-150. Amino acid analyses revealed that the folate (M_r = 25,000) and thiamine (M_r = 29,000) binders have unusually low polarity constants, 0.32 and 0.26, respectively. Evidence obtained with intact cells has established a direct role for these binding proteins in transport of the corresponding vitamins: (A) In each case, the processes of binding and transport showed similarities in substrate affinities and repression by excess vitamin in the growth medium. (B) Competition studies employing amethopterin, 5-formyl tetrahydrofolate, and 5-methyl tetrahydrofolate (for folate) and thiamine monophosphate and thiamine pyrophosphate (for thiamine) have shown that the ability of these compounds to inhibit the transport of the corresponding vitamins is paralleled by their ability to inhibit binding. (C) Amethopterin-resistant mutants which are defective in folate transport have a comparable defect in ability to bind folate. (D) Amethopterin-resistant cells which (compared with the parent cell line) contain folate transport systems with altered affinities for amethopterin also contain binding proteins whose affinities for amethopterin have changed by equivalent amounts. (E) Both the transport and binding of folate by one of the mutants were stimulated (approximately 3-fold) in parallel by the addition of mercaptoethanol.     

Identification of N5,N10-methylene tetrahydrofolate reductase as the enzyme involved in the 5-methyl tetrahydrofolate-dependent formation of a beta-carboline derivative of 5-hydroxytryptamine in human platelets.

An enzyme in human platelets or rat brain incubated with 5-methyl tetrahydrofolate (5MeH₄folate) yields formaldehyde (4, 13), which will combine with biogenic amines to form β-carbolines (5) or tetrahydroisoquinolines. This activity was purified 500-fold from human platelets which are the main storage site for 5-hydroxytryptamine in man. This enzyme was identical to N⁵, N¹⁰-methylene tetrahydrofolate (N⁵,N¹⁰-methylene H₄folate) reductase by the following criteria: (i) co-purification, (ii) heat denaturation, (iii) pH response, (iv) molecular weight, (5) cofactor requirements. A mechanism involving the enzymatic generation of formaldehyde followed by adduct formation with a biogenic amine is proposed.     

Resistance of Pediococcus cerevisiae to amethopterin as a consequence of changes in enzymatic activity and cell permeability I. Dihydrofolate reductase,thymidylate synthethase and formyltetrahydrofolate synthetase in amethopterin-resistant and -sensitive strains of Pediococcus cerevisiae

Pediococcus cerevisiae/AMr, resistant to amethopterin, possesses a higher dihydrofolate reductase (5, 6, 7, 8-tetrahydrofolate: NADP⁺ oxidoreductase, EC 1.5.1.3) activity than the parent, a folate-permeable and thus amethopterin-susceptible strain and than the wild-type. The properties of dihydrofolate reductase from the three strains have been compared. Temperature, pH optima, heat stability, as well amethopterin binding did not reveal significant differences between the enzymes from the susceptible and resistant strains. The enzyme from the wild-type was 10 times more sensitive to inhibition by amethopterin and more susceptible to heat denaturation. The apparent K_m values for dihydrofolate in enzymes from the three strains were in the range of 4.8–7.2 μM and for NADPH 6.5–8.0 μM. The amethopterin-resistant strain exhibited cross-resistance to trimethoprim and was about 40-fold more resistant to the latter than the sensitive parent and the wild-type. The resistance to trimethoprim appears to be a direct result of the increased dihydrofolate reductase activity. Inhibition of dihydrofolate reductase activity by this drug was similar in the three strains. 10–20 nmol caused 50% inhibition of 0.02 enzyme unit. Trimethoprim was about 10 000 times less effective inhibitor of dihydrofolate reductase than amethopterin. The cell extract of the AMr strain possessed a folate reductase activity three times higher than that of the sensitive strain. The activities of other folate-related enzymes like thymidylate synthethase and 10-formyltetra-hydrofolate synthetase (formate: tetrahydrofolate ligase (ADP)-forming), EC 6.3.4.3) were similar in the three strains studied.     

Synthesis of folate derivatives in aerated storage tissue disks

Microbiological assay of extracts prepared from carrot, potato, turnip and beet storage tissue disks revealed that folate derivatives were synthesized during a 48 hr aeration period in sterile distilled water. The composition of the folate pool in carrot was examined by DEAE—cellulose column chromatography, γ-glutamylcarboxypeptidase treatment and differential assay of individual derivatives using Lactobacillus casei and Streptococcus faecalis. The principal folates were polyglutamates of formyl and methyl tetrahydrofolate. Smaller quantities of the corresponding mono- and di-glutamates were also detected. The latter derivatives occurred in pools having a high degree of metabolic turnover. The specific activities of three enzymes catalyzing production of these derivatives from tetrahydrofolate increased during the first 12 hr of aeration. Amino acid analyses revealed that folate synthesis in carrot disks was accompanied by depletion of free serine and by net synthesis of free and protein methionine.     

Nonsense suppression in thymine-requiring strains of Escherichia coli is a consequence of altered folate metabolism

Thymine-requiring strains of Escherichia coli suppress nonsense and frameshift mutants of T4 phage. We proposed that these mutants make errors during translation because of an imbalance in folate metabolism. A thymine-requiring strain grown under suppressing conditions has elevated levels of reduced folates. We tested the effect of either mutational blocks or the inhibition of various steps in folate biosynthesis on suppression. Conditions which prevent the accumulation of 5-methyl tetrahydrofolate inhibit suppression, suggesting that elevated levels of this folate are required for suppression. Furthermore, conditions that result in an accumulation in dihydrofolate inhibit suppression.     

The utilization of formate by human erythrocytes

The incorporation of radioactive formate into an acid-stable non-volatile form by human erythrocytes is dependent upon the addition of 5-amino-4-imidazolecarboxamide riboside. The formate-incorporating activity of human erythrocytes varies widely among normal individuals and the values obtained are characteristic of the erythrocytes obtained from these individuals. The variation is unrelated to the total folate levels of the erythrocytes as measured by the growth response of Lactobacillus casei but is roughly correlated with the quantity of folate forms in the erythrocytes which support the growth of Steptococcus faecalis. The activities of several enzymes involved in the metabolism of the folate coenzymes has also been measured in extracts of erythrocytes. Extracts from all the individuals contained 10-formyltetrahydrofolate synthase, 5-amino-4-imidazolecarboxamide ribotide transformylase, and 5,10-methylenetetrahydrofolate dehydrogenase. None of the extracts contained detectable quantities of either 5,10-methylenetetrahydrofolate reductase or 5-methyltetrahydrofolate-homocysteine methyltransferase. These data support the conclusion that 5-methyltetrahydrofolate is not in metabolic equilibrium with the other forms of folate in the erythrocyte and the uptake of formate by intact erythrocytes is a function of those forms of the folate coenzymes which can be converted to tetrahydrofolate.     

Resistance of Pediococcus cerevisiae to amethopterin as a consequence of changes in enzymatic activity and cell permeability : II. Permeability changes to amethopterin and other folates in the drug-resistant mutant

Pediococcus cerevisiae/AMr, resistant to amethopterin, possesses a higher dihydrofolate reductase (5, 6, 7, 8-tetrahydrofolate: NADP⁺ oxidoreductase, EC 1.5.1.3) activity than the parent, a folate-permeable and thus amethopterin-susceptible strain and than the wild-type. The properties of dihydrofolate reductase from the three strains have been compared. Temperature, pH optima, heat stability, as well amethopterin binding did not reveal significant differences between the enzymes from the susceptible and resistant strains. The enzyme from the wild-type was 10 times more sensitive to inhibition by amethopterin and more susceptible to heat denaturation. The apparent K_m values for dihydrofolate in enzymes from the three strains were in the range of 4.8–7.2 μM and for NADPH 6.5–8.0 μM. The amethopterin-resistant strain exhibited cross-resistance to trimethoprim and was about 40-fold more resistant to the latter than the sensitive parent and the wild-type. The resistance to trimethoprim appears to be a direct result of the increased dihydrofolate reductase activity. Inhibition of dihydrofolate reductase activity by this drug was similar in the three strains. 10–20 nmol caused 50% inhibition of 0.02 enzyme unit. Trimethoprim was about 10 000 times less effective inhibitor of dihydrofolate reductase than amethopterin. The cell extract of the AMr strain possessed a folate reductase activity three times higher than that of the sensitive strain. The activities of other folate-related enzymes like thymidylate synthethase and 10-formyltetra-hydrofolate synthetase (formate: tetrahydrofolate ligase (ADP)-forming), EC 6.3.4.3) were similar in the three strains studied.     

The regulation of one-carbon oxidation in the rat by nitrous oxide and methionine

Formate is oxidized to CO₂ in the rat by folate-dependent reactions. Nitrous oxide treatment inhibited hepatic methionine synthetase activity, reduced hepatic S-adenosyl-l-methionine (Ado-Met) and tetrahydrofolate (H₄ folate) concentrations and decreased the rate of formate oxidation in the rat. The administration of methionine to nitrous oxide-treated rats increased hepatic Ado-Met concentrations and restored hepatic H₄folate levels and formate oxidation to control values but did not reverse the inhibition of methionine synthetase. Positive correlations were observed between hepatic Ado-Met levels and H₄folate concentrations and between hepatic H₄folate concentrations and formate oxidation. These results suggest that alterations in hepatic H₄folate concentrations may profoundly influence the oxidation of one-carbon compounds. They confirm the importance of the methionine synthetase reaction as a major source of regeneration of H₄folate. These findings also indicate that methionine acts at a site other than the methionine synthetase reaction to restore hepatic H₄folate concentrations and formate oxidation to control values in nitrous oxide-treated rats.     

Lack of dihydrofolate reductase activity in brain tissue of mammalian species: possible implications

IT IS becoming increasingly clear that folates play a vital, yet until recently an unrecognized, role in the development and function of the brain. Thus several groups of patients have been found with severe maldevelopment of the brain and mental retardation associated with inborn errors of folate metabolism resulting from congenital deficiency in one or more enzymes involved in folate metabolism (ARAKAWA et al., 1965; 1966; 1967; MUDD, LEVY and ABELES, 1969; ARAKAWA, 1970). The presence of folate coenzymes in brain tissue has been reported by several investigators (ALLEN and KLIPSTEIN, 1970; MCCLAIN and BRIDGERS, 1969). MCCLAIN and BRIDGERS (1969) showed that much less of the folates in brain are in the form of the N5-methyl derivatives than is the case for folates in plasma, red blood cells and liver. Appreciable activity of several folate interconverting enzymes have been demonstrated in brain tissue; for example, N5-methyl tetrahydrofolate homocysteine methyl transferase has been found to exist in higher levels in brain than in liver or kidney (MANGUM, 1972); N5-methyl FH,N-dimethyl-dopamine methyl transferase (LADURON, 1972) and serine transhydroxymethylase (EC 2.1.1; L-Serine: tetrahydrofolate 5, 10-hydroxymethyl transferase) (BRIDGERS, 1968) have recently been detected in brain. The last enzyme is known to catalyse a reaction responsible for the generation of a major portion of one-carbon units. In mouse brain, the activity of this enzyme declines during the first 2 weeks of extra-uterine life (BRIDGERS, 1968). The aim of the present study was to determine the levels of dihydrofolate reductase(5,6,7,8-tetrahydrofolate:NADP+ oxidoreductase; EC 1.5.1.3) in mammalian brain tissues in comparison to the levels in other tissues. This enzyme occupies the first and key position in folate metabolism, reducing the metabolically inert vitamin, folic acid, to tetrahydrofolate. This enzyme also functions in thymidylate synthesis to regenerate tetrahydrofolate from dihydrofolate, a product of the reaction (HWHREYS and GREENBERG, 1958). In this reduced state the molecule can accept one-carbon units from various sources to give rise to metabolically active coenzyme forms of folate. This communication reports the complete absence of dihydrofolate reductase in brain tissue of several mammalian species.     

The reaction of nitro-blue tetrazolium with d,l-L-tetrahydrofolate: The effect of pH,oxygen, formaldehyde and palladium(II)

d,l-L-Tetrahydrofolate (d,l-L-FH₄) transfers two electrons to nitro-blue tetrazolium (NBT) in oxygen-free buffers to form the highly coloured nitro-blue formazan and oxidized folate. Both the rate and extent of this reaction are affected by the pH, the nature of the buffer and the oxygen concentration. Inhibition of both the rate and extent of this reaction in air-saturated solutions by superoxide dismutase (SOD) indicates that the superoxide anion is an intermediate in the reaction so that formazan can be produced by both superoxide independent and superoxide dependent routes in air-saturated solutions.In oxygen-free solutions several lines of evidence can be interpreted to mean that the reduced pteridine ring of tetrahydrofolate is the electron donor in the reaction with NBT. Ionization of the amide hydrogen of the pteridine ring and subsequent increase in electron density of that ring might explain the large increases observed in the rate and extent of the reaction of tetrahydrofolate with NBT as the pH increases. Formaldehyde reacts non-enzymatically with tetrahydrofolate to form a methylene bridge between nitrogens 5 and 10 of methylenetetrahydrofolate. This molecule is much less reactive with nitro-blue tetrazolium than tetrahydrofolate. Complexes formed between tetrahydrofolate and palladium(II) ions are also less reactive with NBT than the tetrahydrofolate alone. This result provides added evidence that palladium(II) ions interact with tetrahydrofolate at the nitrogen 5, nitrogen 10 site of the molecule.     

Plasmodium berghei: folic acid levels in mouse erythrocytes

Folate levels of parasitized whole blood increased over fourfold above noninfected controls based on microbiological assays with Lactobacillus casei. Most of this increase is attributed to oxidized forms of folate (mono-, di-, and triglutamates) and/or N⁵-methyltetra-hydropteroyl mono-, di-, and triglutamates. Assays with Pediococcus cerevisiae showed that levels of tetrahydro-forms of folate polyglutamates doubled during parasitization of mouse erythrocytes. The folate activity for Lactobacillus casei after conjugase treatment was about 1.5 times as high for infected as for uninfected blood. During the malarial infection there was more a change in the form of folate than in its overall level.     

On the nature of folate derivatives in neurospora crassa and the effect of sulphanilamide

Summary The folic acid complex of Neurospora crassa has been separated by chromatography on DEAE-cellulose column and the nature of the individual derivatives ascertained from microbiological response. The functional groups of the polyglutamyl folates were ascertained.Major portion of Neurospora folates is composed of the polyglutamyl derivatives, as seen from the increase in activities for Streptococcus faecalis R, Pediococcus cerevisiae, and Lactobacillus casei after conjugase treatment.The presence of N¹⁰ and N⁵-formyl and N⁵-methyl tetrahydrofolic acid, N⁵-methyl diglutamyl tetrahydrofolic acid, two formylated derivatives of polyglutamates, and two N⁵-methyl polyglutamyl folates was ascertained in the folate complex of Neurospora crassa.Sulphanilamide growth inhibition results in the lowering of all the Neurospora folate derivatives.     

DISTRIBUTION OF FOLIC ACID COENZYMES AND FOLATE DEPENDENT ENZYMES IN MOUSE BRAIN1

—Folic acid coenzymes were found to be distributed equally between post-nuclear particulate and soluble fractions from whole Swiss mouse brain. Mitochondria isolated from the particulate fraction contained essentially only the N⁵-methyl derivative of folate, virtually all of which was in a polyglutamate form. Isolated synaptosomes contained significantly more folate than did mitochondria, with the greater proportion being non-N⁵-methyl derivatives. Osmotic lysis of synaptosomes released only a small portion of the folate; approximately 80 per cent remained with the particulate components of the synaptosome. The enzymes serine transhydroxymethylase and N⁵, N¹⁰-methylenetetrahydrofolate dehydrogenase were found in both the soluble and particulate fractions while formiminoglutamic acid:tetrahydrofolate formiminotransferase activity could not be detected. These findings may be of importance with respect to the synaptic functions of folate coenzymes, including the methylation of biogenic amines.     

Nonsense suppression in thymine-requiring strains of Escherichia coli is a consequence of altered folate metabolism

Thymine-requiring strains of Escherichia coli suppress nonsense and frameshift mutants of T4 phage. We proposed that these mutants make errors during translation because of an imbalance in folate metabolism. A thymine-requiring strain grown under suppressing conditions has elevated levels of reduced folates. We tested the effect of either mutational blocks or the inhibition of various steps in folate biosynthesis on suppression. Conditions which prevent the accumulation of 5-methyl tetrahydrofolate inhibit suppression, suggesting that elevated levels of this folate are required for suppression. Furthermore, conditions that result in an accumulation in dihydrofolate inhibit suppression.     

A simple procedure for the synthesis of high specific activity tritiated (6S)-5-formyltetrahydrofolate

The 5-position of tetrahydrofolate was found to be unusually reactive with low concentrations of formic acid in the presence of a water-soluble carbodiimide. The product of this reaction has neutral and acid ultraviolet spectra and chromatographic behavior consistent with its identity as 5-formyltetrahydrofolate (leucovorin). When enzymatically synthesized (6S)-tetrahydrofolate was used as starting material, the product supported the growth of folatedepleted L1210 cells at one-half the concentration required for authentic (6R,S)-leucovorin. This reaction has been used to produce high specific activity (44 Ci/mmol) [³H](6S)-5-formyltetrahydrofolate in high yield. Experiments with [¹⁴C]formic acid indicate that 1 mol of formate reacted per mol of tetrahydrofolate but that no reaction occurred with a variety of other folate compounds. (6S)-5-Formyltetrahydrofolate, labeled in the formyl group with ¹⁴C, has also been synthesized using this reaction. These easily produced, labeled folates should allow close examination of the transport and utillization of leucovorin and of the mechanism of reversal of methotrexate toxicity by reduced folate cofactors.     

Cation-Dependent Binding of Substrate to the Folate Transport Protein of Lactobacillus casei

Lactobacillus casei cells grown in the presence of limiting folate contained large amounts of a membrane-associated binding protein which mediates folate transport. Binding to this protein at 4°C was time and concentration dependent and at low levels (1 to 10 nM) of folate required 60 min to reach a steady state. The apparent dissociation constant (K_d) for folate was 1.2 nM at pH 7.5 in 100 mM K-phosphate buffer, and it varied by less than twofold when measured over a range of pH values (5.5 to 7.5) or in buffered salt solutions of differing ionic compositions. Conversely, removal of ions and their replacement with isotonic sucrose (pH 7.5) led to a 200-fold reduction in binding affinity for folate. Restoration of the high-affinity state of the binding protein could be achieved by the readdition of various cations to the sucrose medium. K_d measurements over a range of cation concentrations revealed that a half-maximal restoration of binding affinity was obtained with relatively low levels (10 to 50 μM) of divalent cations (e.g., Ca²⁺, Mg²⁺, and ethylenediammonium²⁺ ions). Monovalent cations (e.g., Na⁺, K⁺, and Tris⁺) were also effective, but only at concentrations in the millimolar range. The K_d for folate reached a minimum of 0.6 nM at pH 7.5 in the presence of excess CaCl₂. In cells suspended in sucrose, the affinity of the binding protein for folate increased 20-fold by decreasing the pH from 7.5 to 4.5, indicating that protons can partially fulfill the cation requirement. These results suggest that the folate transport protein of L. casei may contain both a substrate- and cation-binding site and that folate binds with a high affinity only after the cation-binding site has been occupied. The presence of these binding sites would support the hypothesis that folate is transported across the cell membrane via a cation-folate symport mechanism.     

High-performance liquid chromatographic determination of the distribution of naturally occurring folic acid derivatives in rat liver

Procedures which allow extraction and quantitation of labile, reduced folic acid derivatives in rat liver have been developed. These procedures entail extraction of hepatic folates at 100°C in 2% ($\text{w}\text{v}$) sodium ascorbate, 0.2 m 2-mercaptoethanol, pH 7.85. The extract was treated with conjugase to hydrolyze folate polyglutamates and reverse-phase, ion-pair high-performance liquid chromatography was used to separate the resulting monoglutamates which were measured by microbiological assay using Lactobacillus casei. Experiments with HPLC-purified standard derivatives, so treated, showed excellent stability of tetrahydropteroylglutamic acid (H₄PteGlu), 10-formyl-H₄PteGlu, 5-formyl-H₄PteGlu, 5-methyl-H₄PteGlu, and pteroylglutamic acid (PteGlu). Under these conditions, approximately 56% of H₂PteGlu was recovered unchanged while about 27% was converted to PteGlu; 5,10-methylene-H₄PteGlu was quantitatively recovered as H₄PteGlu. These procedures were applied to the task of measuring the distribution of naturally occurring folate cofactors in rat liver. These results indicated that rat liver folates have the following compositions: 5-methyl-H₄PteGlu, 37.2%; H₄PteGlu, 32.7%; 10-formyl-H₄PteGlu, 22.6%; and 5-formyl-H₄PteGlu, 7.7%. Experiments with [³H]PteGlu injection showed that all hepatic folates had the same specific radioactivity as determined by radioassay and L. casei assay, indicating that L. casei exhibited the same growth response to all the folates detected in rat liver.     

Irreversible inhibition of folate transport in Lactobacillus casei by covalent modification of the binding protein with carbodiimide-activated folate

Activated folate formed by reaction of folic acid and 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide irreversibly inhibits the folate transport system of Lactobacillus casei. Complete inhibition of both folate binding to the carrier protein and folate transport was achieved by pretreatment of the cells at low temperature (4 °C) and at neutral pH with 200 nm activated folate. Fifty percent inhibition of binding and transport occurred at 35 and 40 nm activated folate, respectively. Specificity was demonstrated by the fact that excess nonactivated folate added during the pretreatment step afforded complete protection of the binding protein against inhibition, and that activated folate had no effect on the binding or transport of thiamine. Rapid measurements at 4 °C were employed to show that, prior to the appearance of irreversible inhibition, activated folate (K_i = 15 nM) interacted reversibly with the binding site for folate (K_d = 0.8 nM). Cells treated with activated [³H]folate incorporated 1 mol of folate per mole of binding protein. Purification of the labeled protein followed by digestion with Pronase led to the isolation of a compound identified as ?-N-folyl lysine. The ?-amino group of a lysyl residue thus appears to be the nucleophilic group at the binding site that reacts with activated folate.     

Effects of Cultivation Conditions on Folate Production by Lactic Acid Bacteria

A variety of lactic acid bacteria were screened for their ability to produce folate intracellularly and/or extracellularly. Lactococcus lactis, Streptococcus thermophilus, and Leuconostoc spp. all produced folate, while most Lactobacillus spp., with the exception of Lactobacillus plantarum, were not able to produce folate. Folate production was further investigated in L. lactis as a model organism for metabolic engineering and in S. thermophilus for direct translation to (dairy) applications. For both these two lactic acid bacteria, an inverse relationship was observed between growth rate and folate production. When cultures were grown at inhibitory concentrations of antibiotics or salt or when the bacteria were subjected to low growth rates in chemostat cultures, folate levels in the cultures were increased relative to cell mass and (lactic) acid production. S. thermophilus excreted more folate than L. lactis, presumably as a result of differences in the number of glutamyl residues of the folate produced. In S. thermophilus 5,10-methenyl and 5-formyl tetrahydrofolate were detected as the major folate derivatives, both containing three glutamyl residues, while in L. lactis 5,10-methenyl and 10-formyl tetrahydrofolate were found, both with either four, five, or six glutamyl residues. Excretion of folate was stimulated at lower pH in S. thermophilus, but pH had no effect on folate excretion by L. lactis. Finally, several environmental parameters that influence folate production in these lactic acid bacteria were observed; high external pH increased folate production and the addition of p-aminobenzoic acid stimulated folate production, while high tyrosine concentrations led to decreased folate biosynthesis.     

Binding properties of the 5-methyltetrahydrofolate/methotrexate transport system in L1210 cells

A binding component with a high affinity for 5-methyltetrahydrofolate (K_D = 0.11μm) is present on the external surface of L1210 cells. The amount of binder (1 pmol/mg protein) corresponds to 8 × 10⁴ sites per cell. The participation of this component in the high-affinity 5-methyltetrahydrofolate/methotrexate transport system is supported by similarities in the K_D values for 5-methyltetrahydrofolate and methotrexate binding and the K_t values of these compounds for transport. Relative affinities for other folate substrates (aminopterin, 5-formyltetrahydrofolate, and folate) and various competitive inhibitors (thiamine pyrophosphate, ADP, AMP, arsenate, and phosphate) are also similar for both the binding component and the transport system. The measured binding activity does not represent low-temperature transport of substrate into cells, since it is readily saturable with time and is eliminated by either washing the cells with buffer or by the addition of excess unlabeled substrate.