|Home | About | Journals | Submit | Contact Us | Français|
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
All organisms are confronted with external variations in trace element abundance. To elucidate the mechanisms that maintain metal homeostasis and protect against heavy metal stress, we have determined the transcriptome responses in Drosophila to sublethal doses of cadmium, zinc, copper, as well as to copper depletion. Furthermore, we analyzed the transcriptome of a metal-responsive transcription factor (MTF-1) null mutant. The gene family encoding metallothioneins, and the ABC transporter CG10505 that encodes a homolog of ‘yeast cadmium factor’ were induced by all three metals. Zinc and cadmium responses have similar features: genes upregulated by both metals include those for glutathione S-transferases GstD2 and GstD5, and for zinc transporter-like proteins designated ZnT35C and ZnT63C. Several of the metal-induced genes that emerged in our study are regulated by the transcription factor MTF-1. mRNA studies in MTF-1 overexpressing or null mutant flies and in silico search for metal response elements (binding sites for MTF-1) confirmed novel MTF-1 regulated genes such as ferritins, the ABC transporter CG10505 and the zinc transporter ZnT35C. The latter was analyzed in most detail; biochemical and genetic approaches, including targeted mutation, indicate that ZnT35C is involved in cellular and organismal zinc efflux and plays a major role in zinc detoxification.
Metal ions are vital for many biological processes, such as transcription, respiration and growth. However, overaccumulation of essential metals such as copper and zinc or of non-essential toxic metals like cadmium and mercury is detrimental. Copper, for instance, can catalyze the generation of aggressive reactive oxygen species (ROS) via the so-called Fenton reaction (1). Zinc is not redox-active under physiological concentrations and is less toxic than copper, but at high concentrations zinc may bind to inappropriate sites in proteins or cofactors and interfere with their functions. Cadmium does not undergo redox changes, nonetheless, it depletes antioxidant components, notably glutathione, which results in enhanced ROS production (2). Genetic disorders or gross environmental fluctuations can distort metal homeostasis and may lead to a number of deficiency/overload diseases. Copper imbalance is the cause of serious diseases, such as Menkes syndrome and Wilson disease, and has also been implicated in common neurologic conditions, like Alzheimer's disease, Parkinson's disease and the prion-type diseases (3,4). Zinc deficiency can lead to severe syndromes, notably acrodermatitis enteropathica, which is characterized by a decreased absorption of ingested zinc, due to a defective zinc importer (5). Zinc toxicity is associated with reduced iron absorption, impaired immune function (6) and neuronal death (7,8). Cadmium, unlike copper or zinc, is a non-essential metal and is toxic even at very low concentrations. A classic example of industrial cadmium pollution is Itai–Itai disease which occurred in Japan several decades ago (9).
As misregulation of copper and zinc homeostasis can lead to disease and cadmium poses a significant health risk to humans and animals, there is considerable interest in understanding how the organism can defend itself against an excess of a particular metal, yet provide all tissues with appropriate amounts of metals under limiting conditions. The fruit fly Drosophila is a convenient system to address these questions since many aspects of metal homeostasis are conserved between flies and humans. In recent years numerous studies led to a better understanding of uptake, distribution, detoxification and elimination of metal ions. For copper, the Ctr transporter family mediates cellular uptake and three high-affinity copper importers, Ctr1A, Ctr1B and Ctr1C, are known in Drosophila (10). Membrane-bound P-type ATPase transporters are responsible for cellular copper efflux. In mammals, two of them, ATP7A and ATP7B, are implicated in Menkes and Wilson diseases and DmATP7, their common ortholog in Drosophila, was recently characterized (11–16). Zinc transport is mediated via two families of solute linked carrier (SLC) proteins: SLC39A, also referred to as ZIPs, that function in the uptake of zinc into the cytoplasm, and SLC30A, also called ZnTs, that reduce cytoplasmic zinc concentrations by promoting zinc efflux to the exterior or into intracellular vesicles (17). More than 10 zinc transporter genes are annotated in Drosophila melanogaster based on sequence similarities to vertebrate zinc transporters. The ZIP family gene foi (fear of intimacy) was characterized in Drosophila and shown to be a zinc importer (18,19). Along with metal transport, an important aspect of heavy metal homeostasis is the sequestration of metals by metallothioneins (MTs). MTs are small, cysteine-rich proteins with outstanding metal binding capacity. The fly genome encodes four MTs termed MtnA to -D (20–22).
MT gene expression is dependent on the metal-responsive transcription factor-1 (MTF-1), a key regulator of heavy metal homeostasis and detoxification in higher eukaryotes from insects to mammals (23–25). For transcriptional activation, MTF-1 binds through its zinc fingers to metal response elements (MREs) in the promoter/enhancer regions of the target genes. MRE motifs have a core consensus TGCRCNC (where R stands for A or G and N for any of the bases). In mammals, MTF-1 is essential for liver development as mice lacking MTF-1 die in utero due to liver degeneration (26). In contrast, MTF-1 knockout flies (MTF-1140-1R, hereafter termed MTF-1 KO) are viable under laboratory conditions but very sensitive to elevated concentrations of heavy metals, and also to copper scarcity (22). Besides MTs, mammalian MTF-1 contributes to the expression of ZnT1, tear albumin/lipocalin, placenta growth factor, selenoprotein W, N-myc downstream regulated gene 1 (ndrg1), cysteine and glycine-rich protein 1 (csrp1), cEBP alpha, Kruppel-like factor 4 (klf4) and hepatitis A virus cellular receptor 1 (27–31). In Drosophila, the best characterized MTF-1 target genes are metallothioneins. Recently, the copper importer Ctr1B was shown to be induced in an MTF-1-dependent manner upon copper starvation (32). However the knockout of the Drosophila MT family or of Ctr1B did not reproduce the whole spectrum of the MTF-1 mutant phenotype (10,33). Notably, neither absence of metallothioneins nor Ctr1B absence has an appreciable effect on the fly's zinc sensitivity, suggesting the existence of other MTF-1 target genes at least for zinc defense.
Here we present the results of a genome-wide search for Drosophila MTF-1 target genes using in silico and microarray-based approaches. Furthermore, we present a detailed analysis of transcriptional responses to cadmium, copper and zinc as well as to copper starvation in Drosophila larvae. Our study unveiled a great number of genes that respond to environmental metal fluctuations. It also revealed novel MTF-1 regulated genes, such as the ferritin heavy and light chain homologs, the ABC transporter CG10505 that encodes a homolog of ‘yeast cadmium factor’ (YCF1) and a zinc transporter which we named ZnT35C. In this report, we characterize ZnT35C, a member of SLC30A family in Drosophila and show that it is required for zinc tolerance in the fly.
Animals were raised on standard cornmeal molasses-based food. In the third larval instar (fourth day of the development) the animals were transferred from normal to supplemented food containing either 0.05 mM CdCl2, 0.5 mM CuSO4, 5 mM ZnCl2 or 5 mM FeCl3. After 6 h of feeding on the supplemented food, RNA was extracted. For the treatment with the copper chelator BCS (bathocuproine disulfonic acid), animals were kept continuously in the chelator-containing food, since copper stores of a normally fed animal are not exhausted within a few hours. In BCS larval development is delayed, so the RNA was extracted on the fifth day after egg laying. To control for the handling of the larvae during transfer to supplemented food, the normal food controls and the larvae grown in BCS were also removed at third instar and transferred for the last 6 h to normal or BCS containing food, respectively. Total RNA was extracted using TRIzol reagent (Life Technologies).
Nuclease S1 protection assay was performed as described (34). The dried gels were exposed to storage phosphor screens and analyzed using a PhosphorImager (Molecular Dynamics).
For ZnT35Ca-GFP fusion protein analysis, flies were allowed to deposit eggs in the food and raised until third instar larvae. Larvae were dissected and analyzed under a Leica DRB fluorescence stereomicroscope (Figure 7A) and a Leica TCS SP spectral confocal microscope (Figure 7B).
Groups of ~15 adult male flies (2–3 days old, raised on the indicated metal concentrations from egg), were analyzed by inductively coupled plasma mass spectrometry (ICP-MS), as described previously (35). Flies were dissolved in 65% HNO3 in a microwave oven and then diluted to a 6.5% HNO3 solution for analysis. ICP-MS was performed using a HP4500 Series 300 ShieldTorch System instrument (Agilent, Waldbronn, Germany) in peak-hopping mode with spacing at 0.05 a.m.u., 3 points/peak, 5 scans/replicate, 2–3 replicates/sample and an integration time of 400 ms/point. The rate of plasma flow was 15 l/min with an auxiliary flow of 1.0 l/min. The RF power was 1.2 kW. The sample was introduced using a cross-flow nebulizer at a flow rate of 1.06 l/min. The apparatus was calibrated using a 6.5% HNO3 solution containing Cu at 5, 10, 25, 50 and 100 p.p.b. with 103Rh, the internal standard for all isotopes of Cu.
Microarray experiments were done in triplicates using a pool of RNA from at least 30 animals for each assay. cDNA was synthesized from larval total RNA with SuperScript reverse transcriptase (Invitrogen/Life Technologies cDNA Synthesis Kit) using the T7-(T)24 primer 5′-GGCCAGTGAATTGTAATACGACTCACTATAGGGAGGCGGTTTTTTTTTTTTTTTTTTTTTTTT-3′. The resulting cDNA was purified with Phase Lock Gels and concentrated by ethanol precipitation. Synthetic double-stranded cDNA was in vitro transcribed into biotin-labeled cRNA with biotinylated 11-CTP and 16-UTP (Ambion MEGAscript T7 kit). Biotin-labeled cRNA was then isolated with an RNeasy Mini Kit (Qiagen). Adjusted cRNA (15 μg) was taken for fragmentation. Fragmented cRNA (50–200 nt fragments) (11.5 μg) was used to probe the Drosophila Genome Array (Affymetrix), which contains more than 13500 mRNA transcripts from known Drosophila genes and predicted ORFs. Probe arrays were treated with streptavidin, anti-streptavidin goat antibody, biotinylated goat IgG antibody, and stained with streptavidin phycoerythrin. Arrays were scanned twice with an Agilent G2500 Genome Array Scanner.
Raw data processing was performed using the Affymetrix Microarray Suite Ver. 5.0 (MAS5) Software. After hybridization and scanning, probe cell intensities were calculated and summarized for the respective probe sets by means of the MAS5 algorithm (36). In order to compare the expression values of the genes from chip to chip, global scaling was performed resulting in the normalization of the trimmed mean of each chip to a target intensity (TGT value) of 500 as described in the Statistical Algorithms Description Document (Affymetrix, 2002). Quality control measures were considered before performing the statistical analysis. These included adequate scaling factors (between 10 and 17 for all samples) and appropriate total number of ‘Present Calls’ per chip (26–30%) calculated by application of a signed-rank call algorithm (37). Furthermore, the optimal 3′/5′ hybridization ratios (~1) for the housekeeping genes (GAPDH, actin) as well as for the spike controls (BIOB, BIOC, CREX and BIODN), added as hybridization controls into the hybridization cocktail were taken into consideration. After filtering of genes with unreliable expression, using the Cross-Gene Error Model implemented in the Gene Spring software 5.1. (Silicongenetics, 2003), an unequal variance t-test was applied to detect significantly differentially expressed genes. In general, a significance level of 0.05 was chosen. Furthermore, the signal-rank call algorithm from the MAS5 software (37) was applied as an additional filter. Within one comparison of two conditions, each gene was taken into account for further analysis if the algorithm attributed ‘Present Calls’ to at least 50% of the values.
For fly stocks, gene nomenclature, RT–PCR, EMSA (electrophoretic mobility shift assay), cloning and generation of ZnT35C mutant alleles: see Supplementary Data.
Drosophila larvae have a remarkable ability to cope with a wide range of heavy metal concentrations in the ingested food. Wild-type (wt) flies survive up to 3 mM copper, 300 μM cadmium or 14 mM zinc, concentrations that are much higher than what is found in standard laboratory food. Elevated metal concentrations in the food also result in increase of copper, cadmium or zinc content of the fly body (Supplementary Figure 1) (38). When the copper concentration in the food rises from 5 μM Cu in normal food to 500 μM copper, total body copper increases from 5 to 80 ng/mg. Total body zinc content rises ~7-fold when comparing flies on normal food (0.2 mM zinc) with those raised on 4 mM zinc containing food. Thus total body copper and zinc do not increase proportionally to the metal concentration in the food, owing to the existence of homeostatic mechanisms regulating metal uptake and efflux. To elucidate the cellular response to heavy metals, we conducted microarray experiments with RNA from Drosophila third instar feeding larvae kept on normal, non-supplemented food (hereafter NF) or food containing 50 μM CdCl2, 500 μM CuSO4 or 5 mM ZnCl2 for 6 h. To gain more insights into metal acquisition and homeostasis, we also performed microarrays with RNA from larvae raised on food containing a copper chelator, BCS. To assess the transcriptome response of the whole organism we have chosen to use RNA from whole larvae. One caveat of this approach is that any differential expression of genes with a restricted tissue specificity could be underestimated. After applying statistical filters (Materials and Methods) and using a 1.5-fold difference as a threshold change we observed 37 genes upregulated and 11 downregulated in response to 50 μM Cd (Supplementary Table 1), 11 genes induced, and 71 downregulated in 500 μM Cu (Supplementary Table 2), and 299 upregulated and 82 downregulated genes in response to 5 mM Zn (Supplementary Table 3). Strikingly, for all three metals, metallothioneins were found at the top of the list of induced genes. Only eight genes were induced by all three metals, the four Drosophila metallothioneins being among them (Figure 1). In addition, CG30152 (CG10404) was also induced by the three metals. This gene of unknown function is conserved from bacteria to humans. Using RNA from three independent experiments we confirmed the metal induction in S1 nuclease protection assays (Figure 2A). CG10505, an ABC transporter related to the YCF1 was found to be induced by zinc and cadmium, and transcript mapping experiments revealed an induction also by copper (Figure 2B). Eleven genes were in common between cadmium and zinc but not with copper (Figure 1). Remarkably, two genes of the insect Delta class glutathione S-transferases (Gst), GstD2 and GstD5, were highly induced both by cadmium and zinc. We checked the expression of GstD2 and GstD5 by semi-quantitative RT–PCR. The larvae fed with metals had clearly elevated expression of both Gst genes (data not shown). GSTs comprise a large family of detoxification enzymes conserved from yeast to mammals. They are induced by heat shock, ROS, fungal and bacterial infections (39), and by a wide range of chemical agents including heavy metals (40–43). Two putative zinc transporter genes of the SLC30A family, CG17723, hereafter termed ZnT63C (for details of the nomenclature see Supplementary Data), and also CG3994, hereafter ZnT35C, were upregulated by elevated concentrations of zinc or cadmium; however, ZnT35C appeared only when less-stringent statistical screening criteria were applied (P-value 0.06). A subsequent quantification of ZnT35C transcripts in the S1 assay clearly showed induction by cadmium and zinc but not by copper treatments (Figure 2C). The genes encoding the ferritin light and heavy chain homologs (Fer2LCH and Fer1HCH) were 2- to 3-fold induced by zinc. These genes were also almost 2-fold induced by cadmium treatment, an induction that appeared however to be statistically less significant. Transcript quantification revealed a moderate yet significant increase in ferritin mRNAs by cadmium and copper treatments, and more than 2-fold induction by zinc or iron (Figure 2D). Cadmium also elevated the expression of the Hsp22 gene in Drosophila larvae. This is consistent with the observed heat shock response induced by cadmium in yeast (44,45), mice (46), and various human cell lines (47,48) presumably due to protein misfolding caused by ectopic incorporation of cadmium.
In copper, following the metallothioneins, the transcripts coding for proteins involved in immune defense against microbial infection were induced significantly. The high-affinity copper importer Ctr1B was strongly downregulated in response to copper treatment [Supplementary Table 2, see also Refs (10,32)]. The other copper transporters Ctr1A and Ctr1C showed no changes in their mRNA levels. The transcriptome response to the 5-day-exposure to copper chelator BCS yielded a great number of differentially expressed genes (Supplementary Table 4). Under conditions of copper deprivation, transcripts of the copper importer Ctr1B were increased several fold. Also, five cytochrome P450 genes were induced, while another four were downregulated between 5.3- and 1.6-fold by BCS treatment. It is known that P450 enzymes have a wide spectrum of enzymatic activities; they are able to metabolize many xenobiotic as well as cellular compounds such that they serve as direct GST substrates (49). As reduced copper levels likely affect copper containing proteins, scarcity of copper might result in oxidative stress due to compromised Cu-Zn SOD (superoxide dismutase) activity. In line with this, we observed an ~7-fold upregulation in GstD5 expression. Furthermore, metallothioneins were significantly downregulated by Cu depletion, in agreement with their scavenger function upon metal load (Supplementary Table 4 and Figure 2E).
Since the MRE-binding transcription factor MTF-1 is a major player in the transcriptional response to heavy metals in mammals, we reasoned that at least some differentially expressed genes as determined by our microarray experiments are regulated by the Drosophila homolog of MTF-1. Moreover, several genes that emerged in our study do harbor MREs in their promoter regions, indicating a possible role for MTF-1 in their regulation, as was recently shown for the copper transporter Ctr1B (32). To find the genes regulated by MTF-1 we have conducted microarray experiments with RNA from MTF-1 KO and wt larvae kept in NF or transferred to 50 μM cadmium or 500 μM copper containing food for 6 h during the third instar. MTF-1 target genes are expected to lose their metal regulation in the MTF-1 mutants or also to be downregulated in the mutants compared to the wt in the non-supplemented food.
In MTF-1 KO Drosophila, 43 genes were downregulated >2-fold (P-value cutoff 0.05) (Table 1 and Supplementary Table 5). As expected, the transcripts of metallothioneins were absent in the knockout animal and remained undetectable after metal treatment. CG4716, a gene of predicted methylene-tetrahydrofolate dehydrogenase activity, was 8.7-fold downregulated in the MTF-1 mutant. This gene is also zinc-inducible (Table 1) and was found to be upregulated with increasing age in Drosophila (50). It harbors a single MRE 63 bp upstream of the transcription start. Both the copper importer Ctr1B and zinc exporter-like ZnT35C were strongly downregulated in the MTF-1 KO. Two triacylglycerol lipases, CG5966 and CG6283, were downregulated at least 4-fold in MTF-1 mutants, and another lipase-like gene, CG17191, was more than 2-fold downregulated (Supplementary Table 5). Interestingly, the transcripts of the heat shock proteins 23 and 67Bb were also reduced in the KO, in line with the role of MTF-1 in cell stress response. Furthermore, mRNA levels of several structural proteins, TFs and carrier proteins were downregulated, however only 2- to 3-fold. Not unexpectedly, several genes, such as metallothioneins and the zinc transporter-like ZnT35C that are affected by the absence of MTF-1, are also induced by metals in MTF-1 wt larvae (Table 1). In line with MTF-1 being the main regulator of these genes, their metal induction was abolished in the MTF-1 KO (Figure 2C and E and data not shown).
The expression of the other metal inducible genes such as ferritins, ABC transporter CG10505, zinc transporter-like ZnT63C, CG30152 or Hsp22 was not significantly changed in the MTF-1 mutant larvae compared to the wt when both genotypes were raised on NF. However, in MTF-1 KO larvae we did not detect the metal-induced expression of ferritins, CG30152, the ABC and zinc transporters as we observed in wt larvae. Thus MTF-1 is required in the presence of excess metal to induce the expression of these genes but does not seem to be involved in their basal expression. However, unlike ferritins and the ABC transporter, CG30152 and ZnT63C do not contain classical MREs in their putative gene-regulatory regions, suggesting rather an indirect regulation by MTF-1.
A peculiar group of genes comprise glutathione S-transferases D2 and D5. As noted before, these genes are strongly induced by cadmium and zinc. Curiously, when MTF-1 KO larvae were challenged with cadmium, the induction of both genes became more pronounced. Notably, in MTF-1 mutant animals, GstD5 was 5-fold induced already in the normal food (Supplementary Table 5) suggesting a compensatory role of this gene for the loss of MTF-1 and an important function in anti-cadmium (and zinc) defense [see also Ref. (30)].
As mentioned earlier, MTF-1 activates transcription via binding to the MREs invariably present in the promoter regions of metallothionein genes. A closer look at the 18 MREs from the genomic region of Drosophila metallothioneins revealed a striking conservation of thymidine nucleotides 5′ to the consensus (Supplementary Figure 2A). Also, the MRE core sequences are frequently followed by a guanine and position 6 in the consensus is predominantly adenine. This ‘refined’ Drosophila MRE sequence is very similar to the mammalian extended MRE sequence (51). We attempted to find MTF-1 target genes by searching through the Drosophila genome for clusters of extended MREs using the ‘Fly Enhancer’ search engine developed by Markstein et al. (52). Twelve regions in the genome contained two or more extended MRE sites within a span of 300 bp or less. Three of them are associated with MtnB, MtnC and MtnD. One cluster is in the upstream region of Ctr1B, a target gene of MTF-1 under conditions of copper scarcity. Another cluster containing four MREs is in the locus of divergently transcribed ferritin genes, namely in the first intron of the Fer1HCH and 0.7 kb upstream of the Fer2LCH transcription start (Supplementary Figure 2B). A closer look at the ferritin region revealed two more clusters of MRE core sequences located in the 5′-untranslated region (5′-UTR) and in the coding region of Fer2LCH gene (Supplementary Figure 2B). The comparison of these sequences between seven Drosophila species revealed 100% conservation of the three Fer1HCH intronic MREs (Supplementary Figure 2C). The MREs in the two exonic clusters are conserved mainly between phylogenetically closer species (D.melanogaster, Drosophila erecta, Drosophila yakuba and Drosophila pseudoobscura). Our genome-wide search located another MRE cluster in the fifth exon of glutamate receptor gene Glu-RI, a gene involved in synaptic transmission and cation transport. The other six clusters of MREs are either at least 1.5 kb away from a start of any transcription unit or in the intergenic regions of unknown genes. Since ferritin genes are induced by metals, we investigated the potential role of MTF-1 in their inducibility.
Ferritin is well characterized as an iron binding protein. There are also reports that it binds to a variety of other metals such as beryllium, copper, zinc, cadmium, lead and aluminium (53). Transcripts and protein levels for both Drosophila ferritin subunits have been shown to strongly increase when larvae are fed a diet rich in iron (54,55). We confirmed the iron inducibility of Fer1HCH and Fer2LCH transcript levels (Figure 2D). Furthermore, both genes were considerably upregulated by zinc and to a less extent by cadmium and copper. We tested the possible role of MTF-1 in the basal and metal-induced expression of Drosophila ferritins. Indeed, cadmium, copper and zinc induction was abolished in the MTF-1 KO, whereas the iron induction remained (Figure 2D). Moreover, there is also a significant upregulation of both ferritin transcripts in larvae overexpressing Drosophila MTF-1 compared to the ones with only endogenous MTF-1 (Figure 2F). These data imply that ferritins are MTF-1 target genes under metal, notably zinc (but not iron) stress. A direct transcriptional regulation by MTF-1 is suggested by the presence of evolutionarily conserved MRE motifs, which strongly bind MTF-1 in EMSA (Supplementary Figure 2D).
CG10505 is a member of the ABC transporter family. Members include YCF1 found in Saccharomyces cerevisiae, and multidrug resistance associated proteins (MRP/CFTR family) found in humans, that confer resistance to cadmium or to various cytotoxic drugs (56,57). The promoter/enhancer region of CG10505 contains one MRE 0.8 kb upstream of the annotated translation start. We tested the expression of CG10505 by the S1 nuclease protection assay. Here also we saw a loss of metal induction in the MTF-1 KO (Figure 2B). Moreover, overexpression of Drosophila MTF-1 increased transcript levels of CG10505 (Figure 2F). For the indicated concentrations of metals used zinc is the best inducer of the gene. To test the function of this transporter in vivo we expressed the CG10505 coding sequence under the control of synthetic, Gal4-inducible promoter. The cross of transgenic flies with a fly strain that expresses Gal4 transactivator under the actin promoter (actin-Gal4) did not produce any obvious phenotype in normal food or when challenged with different concentrations of zinc, cadmium or copper. We then tested if the overexpression of CG10505 could rescue the metal sensitivity phenotype of MTF-1 mutants. MTF-1 mutants do not survive in food containing either 5 mM zinc, 100 μM cadmium or 500 μM copper (22). In a competitive situation where both MTF-1 KO and wt flies are raised in the same tube, the KO genotype suffers even more and fails to survive in 2 mM zinc or 40 μM cadmium. To test for rescue of the MTF-1 KO by CG10505 we brought UAS-CG10505 and actin-Gal4 driver transgenes into the MTF-1 mutant background and let the offspring of crosses grow up in different concentrations of metals. Under these conditions, we observe a partial rescue of the zinc and copper sensitivity phenotype (Figure 3A). However, somewhat unexpectedly, considering the relation of this ABC transporter to YCF1, the MTF-1 KO did not benefit from CG10505 overexpression in cadmium containing food. Based on the transcript quantification and genetic rescue results, we propose that CG10505 is a target gene of MTF-1 with a role in zinc and copper homeostasis.
Next, we examined ZnT35C, a gene that was >6-fold downregulated in the MTF-1 KO larvae. In agreement with the role of MTF-1, ZnT35C transcripts remained at background level in metal-treated MTF-1 KO larvae (Figure 2C). Furthermore, expressing MTF-1 under the tubulin promoter in addition to the wt MTF-1, caused the transcripts of ZnT35C to markedly increase (Figure 2F). In D.melanogaster, two alternative transcripts of the gene, ZnT35Ca and ZnT35Cb, are annotated. They differ in the first few exons and share the 5th to 10th exons (Figure 4A). The TMpred software, that predicts membrane spanning regions of proteins, identified six strong transmembrane helices in both variants of the ZnT35C protein. A search for putative MTF-1 binding site revealed two MREs in the promoter/enhancer region of ZnT35Ca variant and a third one in the 5′-UTR of ZnT35Ca. Two more MREs are found further downstream before the Znt35Cb transcription start.
The ZnT35Ca variant predominates over the b form at all developmental stages, with a particularly high expression in adult females (Figure 4B). Neither variant could be detected in cultured S2 cells. In larvae, the expression of the gene was very low; it is tempting to speculate that at this stage of rapid growth zinc is probably of great need, and that the low zinc exporter expression would help to maintain zinc supply. Next, we checked the metal induction of each transcript variant in the S1 nuclease protection assay. Here also, the ZnT35Ca was induced upon cadmium and zinc challenge, whereas we could hardly detect any mRNA from ZnT35Cb (Figure 4C). To test the functions of the gene in vivo, we made Gal4-inducible expression constructs by fusing the coding sequences of both isoforms to the UAS promoter. The overexpression of ZnT35C yielded zinc-super-resistant flies; at least 3-fold more ZnT-overexpressing flies eclosed at 14 mM Zn compared to wt flies (Figure 4D). However, in cadmium and not unexpectedly also in copper, ZnT35C overexpression did not confer any advantage.
Keeping in mind the MTF-1-dependent expression of ZnT35C on the one hand, and the sensitivity of MTF-1 KO to metal load on the other, we attempted to rescue the metal sensitivity of MTF-1 KO flies by overexpressing ZnT35C. The actin-driven overexpression of ZnT35Cb isoform did not show any difference in viability; however, ZnT35Ca overexpression gave an advantage to MTF-1 KO Drosophila in zinc, but not in cadmium food (Figure 3B). Rescue was not complete, presumably because of unspecific general expression.
Next, we generated a mutant of ZnT35C by gene targeting based on homologous recombination (22,58). The targeting in the ZnT35C locus resulted in a truncation of the gene that eliminated 60% of the C-terminal coding sequence including the part coding for three transmembrane domains. The S1 probe designed to hybridize to the exons that are removed in our truncation mutant did not reveal any signal in the homozygous mutant flies, confirming successful targeting in the ZnT35C locus (Figure 5A). The mutant of ZnT35C (ZnT35Cmut) is viable with no apparent phenotype under normal laboratory conditions. However, the mutant animals are extremely sensitive to zinc load in that no mutant larvae develop in 2 mM zinc (Figure 5B) whereas wt larvae develop to adulthood in as high as 14 mM zinc concentrations. Even adult ZnT35Cmut flies die within a few days when transferred to 3 mM zinc food while wt flies are not affected by this condition. ZnT35Cmut animals did not suffer from high cadmium or copper concentrations, and were similarly affected as wt by starvation in 5-fold diluted standard food, thus excluding a generally reduced viability (Figure 5B and C). We also generated a genomic deletion of the ZnT35C locus by imprecise excision of an adjacent P-element that removes the transcription start of the ZnT35Ca isoform (allele 11-9). The flies that carry 11-9 over a deficiency of the ZnT35C region (Def rd9) are almost as sensitive to zinc load as ZnT35Cmut (Figure 5D). Interestingly, as revealed by an S1 nuclease protection assay, the ZnT35C promoter is hyperactivated in ZnT35Cmut, suggesting that the absence of functional protein compromises zinc efflux, leading to metal accumulation and induction of metal defense mechanisms (Figure 6A). In agreement with such a scenario, metallothionein (MtnA) expression was noticeably elevated in ZnT35Cmut (Figure 6B). Furthermore, flies lacking ZnT35C have more than 2-fold elevated total body zinc, and the animals heterozygous for ZnT35Cmut contain at least 50% more zinc (Figure 6C). Manganese or copper amounts remained unchanged in all tested genotypes.
Finally, we also checked the expression pattern and cellular localization of ZnT35C, using a transgene where the ZnT35C promoter drives a fusion construct of ZnT35Ca-GFP. The malpighian tubules (the insect analog of a vertebrate kidney) of third instar larvae grown on normal food displayed a faint green fluorescence, which was strongly induced by zinc treatment (Figure 7A). In tubular cells, expression was primarily confined to the apical membrane (Figure 7).
The imbalance of metals due to genetic disorders or environmental pollution is a worldwide problem associated with adverse health effects. Our study unveils a number of genes and regulatory mechanisms involved in maintaining heavy metal homeostasis. The metals tested in our study, cadmium, copper and zinc, induce a wide range of responses, including induction of genes coding for metallothioneins, transporters, components of the glutathione-mediated detoxification pathway, antimicrobial peptides, ubiquitin-conjugating enzymes, heat shock proteins and cytochrome P450 enzymes (Supplementary Tables 1–4). The universal induction of metallothioneins by all three metals and the downregulation by copper depletion is in line with their metal scavenger function and protective role against metal toxicity. However, there are differences in the strength of the response of individual metallothioneins: MtnA induction is more prominent in copper (Figure 2E and Supplementary Tables 1–3), whereas MtnB and MtnC are more strongly induced by zinc and cadmium (Supplementary Tables 1–3). Indeed, a recent study suggests that individual metallothioneins preferentially defend against specific metals, with MtnA playing a major role in handling copper excess, MtnB during cadmium and zinc excess, while MtnC and MtnD play only minor roles in defending against these three metals (59). In absolute terms, expression of MtnA is higher than that of other metallothioneins, with an impressive basal level and consequently a lesser fold-induction, suggesting MtnA to be the major metal scavenger at Drosophila larval stage.
Another essential system for metal ion homeostasis is metal export and import. Although transcriptional control is only one of several regulatory processes, we could follow numerous changes in characterized, as well as putative, Drosophila metal transporters in the microarray experiment. To provide adequate intracellular copper concentrations, the transcripts of the high-affinity copper importer Ctr1B are strongly induced by copper depletion; conversely, they are downregulated at high Cu concentration in order to reduce copper uptake. Two zinc transporter genes (ZnT35C and ZnT63C), both encoding ZnT family members involved in zinc efflux, as well as two ATPase transporters (CG6263 and CG18419) presumably involved in cation and lipid transport, were upregulated by Zn and Cd treatments. Of note, several ubiquitin-conjugating enzymes (Ubc) were induced by zinc. In this context it is worth mentioning that the yeast zinc importer Zrt1p is ubiquitinated and subsequently degraded upon zinc load (60). The upregulation of Ubc might also help to degrade misfolded proteins which are probably formed upon metal load.
Our results indicate that the larval transcriptome responses to zinc and cadmium share several features. Even though they exert different biological effects, the chemical properties of Zn and Cd are similar, thus part of the regulatory mechanisms to cope with excess zinc are also functional in cadmium stress. Another example of the similarity between the zinc and cadmium responses is the specific and strong induction of two members of the glutathione S-transferase Delta class genes. There are more than 40 GST genes in Drosophila grouped into six classes (61). Three GSTs, CG17524, CG6776 and CG1681 with homology to epsilon, omega and theta class GSTs, respectively, are preferentially upregulated by zinc. GSTs catalyze the conjugation of glutathione (GSH) to a variety of harmful compounds. The resulting glutathione conjugates (GS-X) can then be sequestered by the vacuole or excreted from the cell by ATP-dependent GS-X pumps (62). The details of this process vary with different compounds and in different species. In S.cerevisiae, e.g. cadmium ions are conjugated to GSH and the Cd·GS2 conjugate is transported to the vacuole by YCF1, an ATP-binding cassette transporter (56). It is plausible that the GSTs induced in our studies catalyze the conjugation of Cd and/or Zn to GSH, which eventually will be removed from the cytoplasm. Alternatively, or in addition to conjugating metals, they might conjugate ROS which can occur as byproducts of heavy metal metabolism. Interestingly, a specific ATP-binding cassette transporter CG10505, that shares 26% identity to YCF1, is strongly upregulated by all tested metals. It is tempting to speculate that CG10505 acts in a similar fashion to its yeast counterpart in exporting the metal-GSH conjugates. Nevertheless, the metal specificity of Drosophila CG10505 differs from that of YCF1 because the former contributes to elevated resistance to zinc and copper, rather than to cadmium. Another enzyme involved in GSH conjugate metabolism and excretion, gamma-glutamyl transferase, which is an integral part of the gamma-glutamyl cycle involving the degradation and neo-synthesis of GSH (63) was also induced by zinc in Drosophila. Taken together, these data suggest that GSH conjugation plays a substantial role in the response of Drosophila to metal stress. Support for such a scenario comes from a recent study in the mouse showing the existence of two branches of cellular anti-cadmium defense, one via MTF-1 and its target genes, notably metallothioneins, and the other via glutathione (30). In agreement with this concept our data indicate that when the MTF-1-dependent response is abrogated, the glutathione-associated pathway is upregulated to compensate for this compromised genetic background.
The knockout of the Drosophila MTF-1 gene altered the expression of more than 50 genes in our transcriptome assay. However, there is no obvious phenotype unless the animal is subjected to metal stress: copper, zinc, cadmium and mercury load, or copper starvation (22,64). The copper importer Ctr1B and metallothionein genes are well-established targets of Drosophila MTF-1. The Ctr1B mutant is sensitive to changes in copper concentration and the KO of the Drosophila metallothionein family is sensitive to cadmium and copper load; however, these mutants hardly show any phenotype upon zinc challenge (33). Here we suggest a major role in zinc detoxification for several other genes.
Next to the aforementioned ABC transporter CG10505, zinc exporter Zn35C contributes to zinc homeostasis in that the truncation mutant ZnT35Cmut is extremely sensitive even to a mild zinc load of 2 mM, while wt tolerates up to 14 mM zinc excess. Also, an overexpression of this transporter confers resistance to zinc and can partially rescue the zinc sensitivity of MTF-1 knockout flies. Given the strong zinc sensitivity phenotype of the mutant, we propose that ZnT35C is the major transporter of cellular zinc, similar to the mammalian zinc exporter ZnT1 (65). It remains to be seen whether the other zinc exporter-like genes annotated in the Drosophila genome transport zinc from the cytoplasm to vesicular compartments, as do several of the mouse zinc transporters, such as ZnT2, 3, 5 and 7 (66–69). Drosophila ZnT35C localizes to the plasma membrane both at normal and elevated zinc concentrations; besides, mutant flies have a high zinc content, in support of a role in zinc excretion. We note that the mouse zinc exporter ZnT1 was also found to be induced both by cadmium and zinc treatment in an MTF-1-dependent fashion similar to ZnT35C (27). Thus not only the regulation of metallothioneins but also that of zinc transporters is conserved between insects and mammals. In other cases, metal responses of Drosophila and mammals diverge: the fly orthologs of the mouse cEBP alpha, ndrg1 or klf4 were unchanged in our microarray experiments.
An unexpected result of the present study is the induction of ferritin heavy and light chain homologue genes by heavy metals. Ferritin is well known to detoxify, store and transport iron. It was also shown to bind other metal ions and has been speculated to function as a general metal detoxicant in mammals (53). Our results suggest an induction of Drosophila ferritins by metals other than iron, notably zinc. The latter induction depends on MTF-1, based on the following findings: (i) the presence of multiple MREs, (ii) the loss of zinc induction in the MTF-1 mutant, and (iii) elevated ferritin transcripts upon MTF-1 overexpression. In contrast, iron inducibility must be regulated independently since it is not altered by the absence of MTF-1. In line with different signaling pathways responding to iron load versus excess of zinc, copper or cadmium, we note that the classical MTF-1 dependent MtnA gene is not inducible by iron (Figure 2E). Yet these pathways apparently overlap in the ferritin genes.
Supplementary Data are available at NAR Online.
We are grateful to Dr Ulrich Wagner for assistance in microarray data analysis, to Andrea Patrignani, Bruno Schmid and Dr Antonia Manova for technical support, Dr Andreas Simons (FU Berlin) for metal measurements, Viola Günther for providing nuclear protein extracts, Fritz Ochsenbein for the preparation of figures, Dr Michael Fetchko for critical reading of the manuscript and Alla Vardanyan for helpful discussions. We thank the Functional Genomics Center Zurich (FGCZ) for advice and for financial support. This work was supported by Kanton Zürich and by the Swiss National Science Foundation. Funding to pay the Open Access publication charges for this article was provided by the Kanton Zürich.
Conflict of interest statement. None declared.