J Biol Inorg Chem (2015) 20:1229–1238 DOI 10.1007/s00775-015-1302-2
ORIGINAL PAPER
Mitochondrial iron supply is required for the developmental pulse of ecdysone biosynthesis that initiates metamorphosis in Drosophila melanogaster Jose V. Llorens1 · Christoph Metzendorf2 · Fanis Missirlis3 · Maria I. Lind1
Received: 13 July 2015 / Accepted: 5 October 2015 / Published online: 14 October 2015 © SBIC 2015
Abstract Synthesis of ecdysone, the key hormone that signals the termination of larval growth and the initiation of metamorphosis in insects, is carried out in the prothoracic gland by an array of iron-containing cytochrome P450s, encoded by the halloween genes. Interference, either with iron-sulfur cluster biogenesis in the prothoracic gland or with the ferredoxins that supply electrons for steroidogenesis, causes a block in ecdysone synthesis and developmental arrest in the third instar larval stage. Here we show that mutants in Drosophila mitoferrin (dmfrn), the gene encoding a mitochondrial carrier protein implicated in mitochondrial iron import, fail to grow and initiate metamorphosis under dietary iron depletion or when ferritin function is partially compromised. In mutant dmfrn larvae reared under iron replete conditions, the expression of halloween genes is increased and 20-hydroxyecdysone (20E), the active form of ecdysone, is synthesized. In contrast, addition of an iron chelator to the diet of mutant dmfrn larvae disrupts 20E synthesis. Dietary addition of 20E has little effect on the growth defects, but enables approximately one-third of the iron-deprived dmfrn larvae to successfully turn into pupae and, in a smaller percentage, into adults. This partial rescue is not observed with dietary supply of * Fanis Missirlis
[email protected] * Maria I. Lind
[email protected]
ecdysone’s precursor 7-dehydrocholesterol, a precursor in the ecdysone biosynthetic pathway. The findings reported here support the notion that a physiological supply of mitochondrial iron for the synthesis of iron-sulfur clusters and heme is required in the prothoracic glands of insect larvae for steroidogenesis. Furthermore, mitochondrial iron is also essential for normal larval growth. Keywords Development · Insect · Mitochondria · Mitoferrin · Cholesterol Abbreviations 7DHC 7-dehydrocholesterol 20E 20-hydroxyecdysone BPS Bathophenanthroline disulfonate dfh Drosophila frataxin dib disembodied dmfrn Drosophila mitoferrin E74A Ecdysone-induced protein 74EF FAC Ferric ammonium citrate Fer1HCH Ferritin 1 heavy chain homolog GFP Green fluorescent protein Gp93 Glycoprotein 93 Hsc20 Heat shock protein cognate 2 MRS3/4 Yeast mitoferrins RNAi RNA interference Rp49 Ribosomal protein L32 sad shadow Tb Tubby
1
Department of Comparative Physiology, Uppsala University, Norbyvägen 18A, Uppsala, Sweden
2
Heidelberg University Biochemistry Center (BZH), University of Heidelberg, Im Neuenheimer Feld 328, Heidelberg, Germany
Introduction
Departamento de Fisiología, Biofísica y Neurociencias, Centro de Investigación y de Estudios Avanzados, Av. IPN 2508, Mexico City, Mexico
Iron is an essential micronutrient for all eukaryotic organisms [1]. Stable in multiple oxidation states, iron serves as
3
13
1230
an indispensible protein cofactor. Protein-bound iron is in the form of heme or as an iron-sulfur cluster or as a single ion or in the form of a di-iron center [2]. Essential steps in the biosynthesis of heme and iron-sulfur clusters take place in mitochondria, placing these organelles at the heart of cellular iron distribution and metabolism [3, 4]. Thus, iron transport into the mitochondrial matrix is required for biosynthesis of these cofactors. The only known mitochondrial iron transporter in eukaryotes is mitoferrin [5]. Yeast [6–9], zebrafish [10] and mouse [11] mutants in mitoferrin homolog genes display defects in mitochondrial iron import and, as a consequence, impairment in heme and iron-sulfur cluster biosynthesis. Similar results have been obtained by RNA interference (RNAi) in the nematode Caenorhabditis elegans [12] and in mammalian cell cultures [13]. Mitoferrin is a member of the mitochondrial carrier protein family [14]. So far, direct assays attributing iron transport activity to mitoferrins have only been performed with the yeast mitoferrin homologs, MRS3/4 [5, 15]. Nevertheless, known mitoferrin homologs from other species conserve the histidine residues implicated in iron transport and can rescue yeast MRS3/4deficient strains [5, 10, 16]. The genome of Drosophila melanogaster includes a single mitoferrin gene (dmfrn), which affects cellular iron homeostasis [16] and is essential for male fertility [17]. The sterility of hypomorph dmfrn mutants depends on dietary iron availability, suggesting that the mitochondrial iron transport function of mitoferrin is required for spermatogenesis [17]. More recently, dmfrn RNAi was successfully used to reduce pathologic mitochondrial iron accumulation associated with a fly model of Friedreich’s ataxia [18]. This result again supports the notion that dmfrn is involved in the trafficking of iron into the mitochondrial matrix. In addition, dietary iron restriction in combination with dmfrn deletion mutants, or in combination with dmfrn depletion by RNAi, compromised the development of larvae to adulthood [17, 18]. The physiological demand for iron during the larval stages (even after normalizing for growth) increases prior to pupariation [19]. Both RNAi of frataxin (dfh) [20–22] and Heat shock protein cognate 20 (Hsc20) mutants [23] lead to developmental arrest in the third instar larva. Hsc20 and dfh are involved in the biosynthesis of iron-sulfur clusters. Furthermore, RNAi of two nuclear-encoded mitochondrial ferredoxin genes, specifically in the prothoracic glands of Drosophila, resulted in failure to synthesize ecdysone [20]. Ecdysone, produced by the prothoracic gland, is the key steroid hormone that triggers larvae to transition to the pupal stage to undergo metamorphosis and it acts on target cells through its nuclear receptors [24–26]. A phenotypic rescue was possible by dietary administration
13
J Biol Inorg Chem (2015) 20:1229–1238
of 20-hydroxyecdysone (20E), the functional form of ecdysone [20]. These results suggest that iron-sulfur cluster biosynthesis within the prothoracic gland is required for steroidogenesis. The main stimulator for ecdysone synthesis and release is prothoracicotropic hormone, although inputs from juvenile hormone [27, 28], insulin [29–32] and monoamines [33] are part of the more complex regulatory system of the ecdysone production [34, 35]. The transcription factors responding to these biochemical signals induce ecdysone biosynthetic genes, collectively referred to as the halloween genes [36–38]. A subset of these genes encode cytochrome P450s, heme-containing enzymes that typically carry out oxidation of their substrates [39]. To do so, cytochrome P450s require electron donors; in mitochondria, these electrons are delivered from iron-sulfur containing ferredoxins. Another example of a biosynthetic step in 20E biosynthesis requiring an iron-sulfur containing enzyme is the conversion of cholesterol into 7-dehydrocholesterol (7DHC) catalyzed by the evolutionarily conserved Rieske-domain oxygenase Neverland [40, 41]. Even mutations affecting mitochondrial fusion dynamics were found to block the formation of 20E in prothoracic glands and thereby leading to an arrest in larval development, underscoring the importance of mitochondria in steroidogenesis [42]. Collectively, these observations suggested that mitochondrial iron-sulfur cluster biogenesis and heme synthesis are both required for ecdysone biosynthesis (for detailed reviews of Drosophila melanogaster metamorphosis see [43–49]). Here we investigate iron-dependent developmental defects of dmfrn mutant flies in more detail and show that the development of dmfrn mutants is sensitive not only to low availability of dietary iron but also to a genetically modified form of ferritin, in which the green fluorescent protein (GFP) is spliced into a subset of ferritin 1 heavy chain homolog (Fer1HCH) subunits resulting in the Fer1HCHG188 allele [50]. Homozygous Fer1HCHG188 mutants fail iron-loading into ferritin [50] and impair ferritin secretion leading to embryonic lethality [51], whereas heterozygous Fer1HCHG188/+ flies assemble a GFP-tagged ferritin, which incorporates iron less efficiently, resulting in an overall reduction of total body iron and a likely reduction in bioavailable iron [52]. Importantly, we show that the developmental arrest of dmfrn mutants can be partially rescued by dietary supplementation with 20E, but not by its precursor 7DHC. Thus, mitochondrial iron transport by mitoferrin in the prothoracic glands is required for sufficient biosynthesis of ecdysone, whose release in a major pulse ends the larval stage and initiates metamorphosis. A more ubiquitous function of mitochondrial iron in larval growth was also observed.
J Biol Inorg Chem (2015) 20:1229–1238
Materials and methods Drosophila stocks The 693 bp-deletion allele dmfrnDf13, which removes a large part of the first dmfrn exon, including the first 38 amino acids of the encoded protein, was generated previously by P-element excision [17]. The approximately 11 kbp deletion Df(3R)ED6277, which removes dmfrn and CG5514 completely and parts of Mes-4 and Gp93, was generated by recombination as part of the DrosDel project [53] and was obtained from the Szeged Drosophila Stock Centre. The P-element insertion mutant dmfrnEY01302 was obtained from the Bloomington Drosophila Stock Center at Indiana University and has also been characterized previously [17]. The rescue transgene dmfrnvenusB32 contains the dmfrn gene region with a C-terminal venus fluorescent protein tag [17]. The Fer1HCHG188 allele was generated through a protein-trap screen [54] and has been extensively characterized [50–52]. The ferritin allele Fer1HCHG188 was introduced into the mitoferrin mutant chromosome of Df(3)ED6277 by recombination to produce Df(3)ED6277, Fer1HCHG188. Lines Df(3R)ED6277, dmfrnEY01302 and dmfrnvenusB32 have been outcrossed to w1118 for more than ten generations to remove off-target mutations. Culture media and conditions Drosophila stocks were maintained at 25 °C on standard potato sucrose medium in a 12 h/12 h light/dark cycle. Low iron conditions were achieved as before [17, 52, 55, 56] by adding to the normal diet 100, 133 or 166 μM of the iron chelator bathophenanthroline disulfonate (BPS, Sigma). High dietary iron conditions were achieved by addition of 0.8 mg/ml ferric ammonium citrate (FAC, Sigma), which corresponds to an increase in food iron by 2.5 mM. 20E was purchased from Abcam (Cambridge, UK) and 7DHC from Sigma; these were dissolved in 100 % ethanol obtaining 7 mM 20E and 40 mM 7DHC stock solutions, respectively. 20E and 7DHC were added to the fly medium at final concentrations of 0.33 mg/ml (0.7 mM) and 1.5 mg/ ml (4 mM), respectively. These concentrations were chosen based on previous work [57]. Higher 20E concentrations resulted in lethality of larvae. Phenotype scoring Heterozygous dmfrnDf13/+ flies, used as controls, were generated by crossing dmfrnDf13/TM6c, Tb, Sb to Vno/TM6c, Tb, Sb. To obtain dmfrn transheterozygous deficiency flies, crosses were carried out between dmfrnDf13/TM6c, Tb, Sb
1231
and dmfrnDf(3R)ED6277/TM6c, Tb, Sb. The TM6c balancer carries a mutation in the Tubby (Tb) gene, which serves as a marker as it causes larvae to be visibly shorter in length, but slightly wider (hence the name for the gene). This phenotype allowed us to distinguish dmfrn heterozygous larvae from transheterozygous dmfrn deletions. The same considerations were applicable to the rescue experiments using dmfrnvenusB32 and to the interaction study of dmfrn with Fer1HCHG188. All crosses were carried out in normal food tubes and after 48 h larvae were placed into plates containing food of the indicated treatment. As the combination TM6c/TM6c is embryonic lethal, the expectation from these crosses is that 66 % of larvae carry a dmfrn deficiency over a balancer (33 % dmfrnDf13/ TM6c and 33 % dmfrnDf(3R)ED6277/TM6c or, in the control cross, Vno/TM6c). Furthermore, another 33 % are transheterozygous for the deficiencies dmfrnDf13/dmfrnDF(3R) ED6277 or in the control cross dmfrnDf13/Vno. The ratio between Tb− and Tb+ reflects the larvae carrying the TM6c balancer (Tb) to the larvae of the genotype of interest (Tb+), respectively. As 66 % of the larvae are expected to be Tb− and 33 % are expected to be Tb+ the ratio Tb/Tb+ should be 2 if all genotypes show the same survival until the time of analysis. The ratio was calculated at the pupal stage and all crosses were done in triplicate and at different times. Values higher than 2 represent pre-pupal lethality or developmental delays of dmfrnDf13/Df(3R)ED6277 larvae (see Table 1). To quantify growth differences at least 10 larvae of the indicated genotypes and dietary conditions were collected on the 5th day after embryo oviposition, washed in phosphate saline buffer and imaged using a USB microscope (Plexgear®). MicroCapture software was used to measure length (L) and width (diameter—d). Larval volume was determined as previously [58], using the formula 4/3π × (L/2)2 × (d/2). Finally, the percentage of Tb+ larvae that had initiated metamorphosis was plotted for 12 h intervals. From these graphs we scored the time needed by larvae to reach the pupal stage by determining in a given population when 50 % of the larvae initiated metamorphosis. Fertility assay For the male fertility assay, dmfrnEY01302/TM6c females were crossed to either dmfrnDf(3R)ED6277/TM6c or dmfrnDf(3R)ED6277, Fer1HCHG188/TM6c males. Single 3- to 4-days-old Tb+ male flies, collected from flies reared on indicated food sources, were mated with 2–3 virgin w1118 females reared on normal food, respectively. After 5 days of mating, the fraction of fertile flies was determined by the presence of larvae in the vials.
13
1232
J Biol Inorg Chem (2015) 20:1229–1238
Table 1 Developmental effects of dmfrn deletion mutants raised in different diets Genotype or cross
Diet
Tb+ larval volume (mm3) Tb+ time to pupa (days) Tb/Tb+ ratio Total N pupae
w1118
Normal
16.6 ± 1.1
5.6
–
150
133 µM BPS 19.9 ± 1.4 2.5 mM FAC 16.3 ± 0.9
5.4 5.5
– –
150 150
5.6
2.7
187
5.6 5.2 5.5 5.7 6.2 5.6 6.3
1.5 2.3 1.9 1.5 1.6 2.0 2.4
91 112 174 61 152 60 400
– 6.4 6.6 6.7 1 pupa 10.5 10 No pupae 5.6 – 5.7 –
1.6 2.0 4.8 1.6 398 32 58 ∞ 1.5 1.8 1.8 1.8
1105 148 157 108 399 152 59 93 434 972 66 832
5.6 5.8 – 5.6 5.6 – 5.7 9.0
1.7 1.7 5.1 5.5 1.7 1.8 1.6 41
992 82 239 67 801 738 71 958
9.0 No pupae 6.4
75 ∞ 2.2
76 90 1128
6.6
2.0
45
dmfrndf13/TM6c, Tb, Sb was crossed to: Vno/TM6c, Tb Sb
Normal
dmfrnDf(3R)ED6277/TM6c, Tb Sb
16.9 ± 1.0 100 µM BPS 18.4 ± 1.4 133 µM BPS 14.2 ± 0.9 15.1 ± 0.9 166 µM BPS 16.1 ± 1.1 2.5 mM FAC 16.4 ± 2.0 Normal –
100 µM BPS 133 µM BPS
166 µM BPS 2.5 mM FAC
B32; dmfrnDf(3R)ED6277/TM6c, Tb Sb
Normal
13.9 ± 1.5
– 9.8 ± 0.6 6.2 ± 0.7 9.6 ± 1.0 – 4.4 ± 0.4 2.0 ± 0.3 2.9 ± 0.9 – – 9.4 ± 1.0 –
– 10.9 ± 0.3 133 µM BPS – 7.4 ± 1.0 2.5 mM FAC – – 15.7 ± 0.4 – dmfrnDf(3R)ED6277, Fer1HCHG188/TM6c, Tb Sb Normal 1.0 ± 0.2 133 µM BPS 0.26 ± 0.06 2.5 mM FAC – 3.8 ± 0.4
Genetic interaction with Fer1HCHG188 and effects of the rescue transgene B32 are also shown. The description of the experiments can be found in the main text; each line indicates a different experimental run
Real‑time polymerase chain reaction (RT‑PCR) Total RNA was isolated from 30 third instar larvae (5 days old) reared on normal food using RNeasy Minikit (Qiagen), using QIAshredder columns to homogenize the sample and treated with DNAse I (Ambion). The RNA pellets were resuspended in nuclease-free water. Total RNA (2.5 μg), primed by oligo(dT)20-primers, was reverse transcribed
13
with ThermoScript (Invitrogen). Quantitative PCR was carried out using QuantiTect SYBR green (Qiagen) in a MX3000P (Stratagene) qPCR system. Primer performance was assayed with a dilution series of cDNA. The method ‘comparative quantification’ was used to quantify relative transcript levels. Relative transcript levels of the housekeeping gene Rp49 were used to normalize data obtained for genes of interest. The primers used were:
J Biol Inorg Chem (2015) 20:1229–1238
Rp49 forward 5′-CCG CTT CAA GGG ACA GTA TCT G-3′, reverse 5′-CAC GTT GTG CAC CAG GAA CTT-3′, E74A forward 5′-TCC GAG AGC AAC TTC GAG AT-3′, reverse 5′-TTG ATC AAA TCG CCA CAG AG-3′, dib forward 5′-TGCCCTCAATCCCTATCTGGTC-3′ reverse 5′-ACAGGGTCTTCACACCCATCTC-3′, sad forward 5′-CCGCATTCAGCAGTCAGTGG-3′ reverse 5′-ACCTGCCGTGTACAAGGAGAG-3′
20E quantification Quantification of 20E was performed using the competitive Enzyme Immunoassay (EIA; Cayman Chemicals, Inc, USA) with the protocol kindly provided by Véronique Monnier [20]. The standard curve was obtained with 20E purchased from Abcam (Cambridge, UK). For sample preparation, 15 larvae of dmfrnDf13/+ and 45 larvae of dmfrnDf13/Df(3R)ED6277 were weighed, homogenized in 600 μl of methanol and centrifuged (20 min at 18,000×g). The supernatant was removed and this step was repeated by adding another 600 μl of methanol to the pellet material. The supernatants were combined, dried and resuspended in 50 μl of EIA Buffer prior the EIA assay. Ellmann reagent (Cayman Chemicals, Inc, USA) was used for the chromogenic reaction and absorbance was read at 412 nm on a Victor 3 photometer (Perkin Elmer). Three biological replicates were performed. Statistical analysis All statistical analyses were carried out with GraphPad Prism 5.03 software. For comparison of means, we performed one-way analysis of variance at followed by Tukey’s post hoc test. Group differences (at p
