Molecular Characterization of a Novel, Cadmium-inducible Gene from the Nematode Caenorhabditis elegans

Cadmium is an environmental contaminant that is both a human toxicant and carcinogen. To inhibit cadmium-induced damage, cells respond by increasing the expression of genes that encode stress-response proteins. We previously reported the identification of 48 cadmium-inducible mRNAs in the nematode Caenorhabditis elegans. Here we describe a new cadmium-responsive gene, designated cdr-1, whose rate and level of inducible expression parallel those of the C. elegansmetallothioneins. The CDR-1 mRNA contains an open reading frame of 831 bp and encodes a predicted 32-kDa, integral membrane protein. Following cadmium exposure, cdr-1 is transcribed exclusively in intestinal cells of post-embryonic C. elegans. In vivo, the CDR-1 protein is targeted specifically to the intestinal cell lysosomes. cdr-1transcription is significantly induced by cadmium but not by other tested stressors. These results indicate that cdr-1expression is regulated by cadmium and in a cell-specific fashion. Inhibition of CDR-1 expression renders C. eleganssusceptible to cadmium toxicity. In conclusion, cdr-1defines a new class of cadmium-inducible genes and encodes an integral membrane, lysosomal protein. This protein functions to protect against cadmium toxicity.

The transition metal cadmium is considered a serious occupational and environmental health threat. It is continuously introduced into the atmosphere through the smelting of ores and the burning of fossil fuels and is commonly found in "Superfund" clean-up sites (1)(2)(3). Humans are exposed to cadmium primarily via inhalation and the ingestion of cadmium-containing foods (4). Toxicological responses of cadmium exposure include kidney damage, respiratory diseases, neurological disorders, and lung, kidney, prostate, and testicular cancers (4).
Cadmium induces intracellular damage via the (a) nonspecific inactivation/denaturation of proteins, by binding to free sulfhydryl residues; (b) displacement of zinc co-factors from a variety of proteins, including transcription factors; and (c) generation of reactive oxygen species, which ultimately oxidize DNA, proteins, and lipids. Although cadmium itself is not redox active in vivo and cannot directly catalyze the reduction of molecular oxygen, it has been suggested that the production of reactive oxygen species is a consequence of a cadmium-induced depletion of cellular glutathione or the inactivation of copper/ zinc-superoxide dismutase (5,6).
To attenuate the toxic effects of cadmium, cells respond by increasing the steady-state levels of a variety of proteins. The functions of these proteins can be broadly defined, because those that repair intracellular damage or those that remove the toxicants (e.g. cadmium, reactive oxygen species). To remove toxicants, cells activate the transcription of genes that encode proteins that are involved in scavenging reactive oxygen species, including heme oxygenase, ␥-glutamylcysteine synthetase, superoxide dismutase, catalase, glutathione peroxidase, and glucose-6-phosphate dehydrogenase (7)(8)(9)(10)(11). Cadmium is essentially removed from the cell through chelation by MT 1 or exported by means of metal ion pumps. These pumps can transport the metal from the cytoplasm into lysosomes or out of the cell (12)(13)(14)(15).
To identify new cadmium-responsive genes that are involved in intracellular defense, the reverse transcriptase-PCR protocol of differential display was used (16). Forty-eight differentially expressed mRNAs, which correspond to the products of thirty-two independent genes, were identified in the non-parasitic nematode Caenorhabditis elegans (17).
C. elegans provides an excellent model system for obtaining an integrated picture of cellular, developmental, and molecular aspects of transition metal toxicity. The adult hermaphrodite is composed of 959 somatic cells. The developmental and cellular biology of C. elegans is thoroughly understood, and the nematode contains highly differentiated muscular, nervous, digestive, and reproductive systems (18). Furthermore, high levels of evolutionary conservation between C. elegans and higher organisms are observed in many of the proteins that are induced as part of a metal-activated stress response. These include MT (19), superoxide dismutase (20,21), ubiquitin (22), heat shock protein 70 (23), glutathione S-transferase (24), catalase (25), and multidrug resistance-associated proteins (14). C. elegans also contains homologues to many of the regulatory proteins that have been implicated in modulating the molecular response to metal exposure (26 -29).
In the collection of cadmium-responsive C. elegans ESTs identified by differential display, four were derived from an identical mRNA (17). Here we describe the isolation and characterization of the full-length mRNA and the gene, designated cdr-1 (cadmium-responsive gene family), from which the ESTs were derived. CDR-1 encodes a novel protein, which is transcribed exclusively in the intestinal cells of post-embryonic C. elegans and is targeted to lysosomes. Transcription of cdr-1 is activated in response to cadmium exposure, whereas several common environmental stressors are not effective inducers. This is unique among metal-inducible genes, which are typically activated by multiple stressors (7,30). In contrast, the level of responsiveness and sensitivity of cdr-1 to cadmium exposure are comparable to those previously reported for the two C. elegans MT genes, mtl-1 and mtl-2 (19). When CDR-1 expression is inhibited, and nematodes are exposed to cadmium in vivo, the pseudo-ceolomic space accumulates fluid. In addition, reproduction and development are inhibited. The results presented in the current report support the hypothesis that cdr-1 represents a new class of metal-responsive genes that encode novel defense/repair proteins.

EXPERIMENTAL PROCEDURES
Growth of C. elegans-The Bristol N2 strain of C. elegans was grown on NGM agar plates at 20°C as previously described (17). To obtain large quantities of C. elegans, nematodes were grown in liquid S medium using Escherichia coli OP50 as a food source. In experiments where nematodes were exposed to cadmium, the S medium was supplemented with 100 M CdCl 2 (19).
Isolation of the CDR-1 cDNA-The four ESTs identified by differential display that are derived from CDR-1 (DDRT2, DDRT7, DDRT16, and DDRT26; accession numbers: AF071362, AF071398, AF0771356, and AF071379, respectively) are between 217 and 240 bp in length and come from the 3Ј-end of the mRNA (17). The full-length sequence of CDR-1 was obtained using the 5Ј-rapid amplification of complementary DNA ends (5Ј-RACE) protocol as previously described (19). Briefly, cDNA was synthesized from 2 g of poly(A) ϩ RNA that was isolated from cadmium-treated C. elegans, and the cDNA was then tailed with ϳ20 dA residues. Subsequently, the tailed-cDNA was mixed with 10 pmol of 5Ј anchor oligo(dT) primer (5Ј-CGTCTAGAGATGCATG-CATC(T) 19 -3Ј), 25 pmol of 5Ј anchor primer (5Ј-CGTCTAGAGATGCAT-GCATC-3Ј) and 25 pmol of the 3Ј gene-specific primer VHCL3 (5Ј-CACATAGTAAAGTCCTTTGG-3Ј). This primer is the inverse complement of nt 829 -848 (relative to the transcription start site) in the cDNA encoding CDR-1. The CDR-1 cDNA was amplified by 30 cycles of the PCR using the following conditions: denaturation at 94°C for 60 s, annealing at 50°C for 90 s, and extension at 72°C for 60 s. The products of this reaction were then subjected to a second round of the PCR using the conditions described above with the following modifications: the 5Ј anchor oligo(dT) primer was omitted, the annealing temperature was raised to 60°C, and the magnesium concentration was increased from 1.5 to 3 mM. In addition, the 3Ј gene-specific primer VHCL3 was replaced with an alternative gene-specific primer VHCL4 (5Ј-TATCTAGAAGTCCTTTTCGAGAACATCC-3Ј). The sequence of VHCL4 is the inverse complement of nt 762-783 in CDR-1. The amplified DNA was subsequently purified, inserted into the pGEM-T vector (Promega), and sequenced.
Sequence analysis showed that this protocol did not produce a fulllength CDR-1 cDNA (i.e. the fragment did not include an initiator ATG codon). To obtain cDNAs that include the 5Ј-end of the CDR-1 mRNA, a third 5Ј-RACE reaction was performed. The cDNA products obtained in the reaction containing the 5Ј anchor primer and VHCL4 were subjected to an additional round of the PCR using the conditions described above except VHCL4 was replaced with the primer VHCL10 (5Ј-TTGTAGAATTCGTCGACAGC-3Ј). VHCL10 is the inverse complement of nt 463-482 and includes a unique EcoRI restriction site, which is present in the CDR-1 cDNA (underlined). Products of this reaction were purified, inserted into the pGEM-T vector, and sequenced. Sequence analysis confirmed that the 5Ј-end of CDR-1 was isolated using this protocol.
A cDNA clone that contained the full-length CDR-1 sequence was subsequently created by first digesting a plasmid that contained the 3Ј-end of CDR-1, pGEM-VHCL3, with NcoI and EcoRI. Following agarose gel electrophoresis, a 3377-bp fragment was isolated. This fragment contained the pGEM-T vector and 377 bp of the 3Ј-end of the CDR-1 cDNA, which includes the translation stop codon. A cDNA fragment that contained the 5Ј-end of the CDR-1 was obtained following the digestion of the plasmid pGEM-VHCL10 with NcoI and EcoRI. A 488-bp fragment, which includes the translation start site, 5Ј-untranslated region, and 15 bp of the pGEM multiple cloning site, was purified, inserted into the 3377-bp plasmid, and sequenced. Sequence analysis confirmed the presence of a full-length CDR-1 cDNA clone. This plasmid was designated pGEM5-CDR-1.
Computer Analysis-Analysis of nucleotide and amino acid sequence data, including assembly of continuous cDNA sequences, sequence comparisons, and data base searches were performed using PCGENE-IntelliGenetics software (IntelliGenetics, Mountainview, CA) and BLAST programs (31,32). Genomic sequence data were obtained from the C. elegans Genome Project (available through the Wellcome Trust Sanger Institute). The location within the C. elegans genome of cdr-1 was determined using ACeDb (A C. elegans Database) (33).
Isolation of the cdr-1 Gene-BLAST nucleotide sequence analysis identified several cosmids that include the entire cdr-1 coding and potential 5Ј-regulatory regions. One of these cosmids, F35E8 (accession number: Z81529), was obtained from the C. elegans Genome Project.
To isolate a genomic fragment that contains cdr-1, F35E8 was digested with XbaI. Based on the sequence data, F35E8 contains two XbaI sites at 15,882 and 21,707, which flank the 5Ј-and 3Ј-ends of cdr-1. Following XbaI digestion, a 5,825-bp fragment was purified and then inserted into pGEM3zf(Ϫ), which was cut with XbaI.
Preparation of RNA and Northern Blot Analysis-Total C. elegans RNA was prepared from washed nematodes as previously described (19). For some experiments poly(A) ϩ RNA was purified using the Poly (A) Tract System, following the manufacturer's instructions (Promega).
Northern blot analysis was performed using 32 P-labeled EST clone DDRT16 or full-length CDR-1 cDNA as probes, as previously described (19). PhosphorImager analysis and ImageQuant software (Molecular Dynamic System) were used to determine the amount of probe that hybridized to the RNA. After the amount of CDR-1 was determined, the bound probe was removed by incubating the membrane at 95°C for 1 h in 0.1% SDS. The RNA was then re-hybridized to a 32 P-labeled C. elegans myosin light-chain probe, which served as a loading control. The steady-state level of CDR-1 mRNA was normalized to that of the constitutively expressed myosin light-chain mRNAs (ϳ1150 and 680 nt) (34).
Kinetics of CDR-1 mRNA Accumulation-The rate of CDR-1 mRNA accumulation following cadmium exposure was determined by exposing C. elegans to 100 M cadmium for different times (19). When cadmium was added, the culture contained nematodes at all stages of development and an adequate supply of food. After exposing C. elegans to cadmium for 1, 2, 4, 8, 16, and 24 h, the nematodes were isolated, total RNA was then prepared and Northern blot analysis performed. The steady-state level of CDR-1 mRNA at each time point was normalized to the levels of the two myosin light-chain mRNAs.
Preparation and Analysis of Transgenic C. elegans-A ϳ3.5-kbp fragment of genomic DNA that is immediately upstream from the initiator ATG in cdr-1 was prepared from the cosmid F35E8 by employing the PCR. The cdr-1 promoter/enhancer fragment was amplified by using Pfu polymerase and the primers VHCL11 and VHCL12. The 5Ј primer VHCL12 (5Ј-CGCGGATCCCTAAGCCAGGATCATGCACC-3Ј) precedes a BamHI site in F35E8 and corresponds to nt 13,809 -13,828 in the cosmid. The 3Ј primer VHCL11 (5Ј-TCCCCCGGGACTTGAGACAG-TAGTACATA-3Ј) is the reverse complement of nt 17,377-17,358 in F35E8. VHCL11 incorporates a SmaI site (underlined) at the 3Ј-end of the amplified product. The amplification conditions were as follows: denaturation at 94°C for 45 s, annealing at 55°C for 60 s, and extension at 72°C for 9 min. Following amplification, the genomic fragment was purified and then digested with BamHI and SmaI. The fragment was then inserted into the C. elegans ␤-galactosidase expression vector pPD95.10, which was cut with the identical enzymes. The cdr-1/lacZ C. elegans vector expresses a ␤-galactosidase reporter gene that is preceded by the SV40 large T antigen nuclear targeting sequence. Thus, the 5Ј-promoter/enhancer region of cdr-1 will control the expression of a ␤-galactosidase fusion enzyme that accumulates in nuclei (35).
C. elegans were transformed by microinjecting a mixture of the recombinant reporter plasmid DNA (100 g/ml) and a plasmid containing the dominant selectable marker gene rol-6 into the gonadal syncytium of young adult C. elegans as described previously (19,35). Transgenic C. elegans were selected and maintained as indicated in previous studies (19).
These lines of transgenic nematodes contain the reporter transgene as a heritable extrachromosomal array, which can be lost after several generations (35). To permanently integrate the array into the C. elegans genome, transgenic nematodes were exposed to 300 J/m 2 of ultraviolet radiation. After several generations, C. elegans that contain the integrated transgene were isolated by identifying individual nematodes of which 100% of the progeny express the rol-6 phenotype. Integrated C. elegans were out-crossed four times with a wild type strain to remove any mutations that may have occurred because of the radiation treatment. The transgenic strain used in the subsequent studies is designated JF9 (mtIs7, cdr-1/lacZ).
Cell-specific, developmentally regulated patterns of cdr-1 transcription were determined in cadmium-exposed and non-exposed C. elegans as previously described (19). Transgenic C. elegans were fixed and stained for ␤-galactosidase activity as described in Freedman et al. (19). Cells that actively transcribe cdr-1 were identified from the level of ␤-galactosidase activity and visualized by microscopy.
Effects of Stressors on cdr-1 Transcription-C. elegans JF9 (mtIs7, cdr-1/lacZ) were used to investigate the effects of exposure to heat shock, transition metals, and oxidative stress on cdr-1 transcription.
The effect of heat shock on cdr-1 transcription was determined using conditions previously described (19,36). Briefly, transgenic C. elegans were incubated at 33°C for 1.5 h. The nematodes were allowed to recover by incubating at 20°C for 16 h. After the recovery period, nematodes were collected and the level of ␤-galactosidase activity was quantified from the amount of blue chromogen produced as previously described (37).
To examine the effects of metals and oxidative stress on cdr-1 transcription, nematodes were collected and then suspended in K medium (38). C. elegans were exposed to CdCl 2 ⅐2. scription in C. elegans (24,36). 2 The effects of these stressors on cdr-1 transcription were then determined from the levels of ␤-galactosidase activity.
In Situ Hybridization-A population of wild type C. elegans at all developmental stages were treated with 100 M cadmium for 24 h. Nematodes were isolated and fixed and the external collagenous cuticle was permeabilized and prehybridized as previously described (40). Digoxigenin-labeled probes were then added to a suspension of fixed, permeabilized nematodes in 40% (v/v) formamide, 5ϫ SSC, 0.1 mg/ml sonicated salmon sperm DNA, 50 g/ml heparin, and 0.1% Tween 20. Hybridization was carried out at 48°C for 16 h.
To synthesize the digoxigenin-labeled probes, anchored PCR was employed using pGEM5-CDR-1 as a template. An antisense CDR-1 DNA probe was generated using plasmid DNA, which was linearized following NcoI digestion, and a SP6 primer. A CDR-1 sense probe was prepared from linearized pGEM5-CDR-1, which was digested with NotI, and a T7 primer. Binding sites for both T7 and SP6 primers are in pGEM5zf(ϩ) and flank the CDR-1 cDNA insert.
In Vivo Localization of CDR-1-To determine the intracellular location of CDR-1 in vivo, a C. elegans expression vector was constructed in which the expression of a CDR-1-eGFP fusion protein is regulated by the cdr-1 promoter. The eGFP contains the Ser-65 to Cys-65 substitution (41). To generate the CDR-1-eGFP fusion cDNA, full-length CDR-1 cDNA was prepared by the PCR using the primers VHCL14 (5Ј-CCG-GTACCTTAATTGACATAGTAAAGTCC-3Ј) and VHCL15 (5Ј-TC-CCCCGGGTCTCAAGTATGTTGG-3Ј). VHCL14 is the inverse complement of nt 834 -854 in the CDR-1 cDNA. This primer introduces a KpnI site (underlined) at the 5Ј-end of the amplified product and converts the stop codon (boldface) to a serine residue (TGA 3 TCA). VHCL15 is identical to nt 6 -22, and it adds a SmaI site to the 5Ј-end of the amplified product (underlined). Following amplification, the DNA was digested with SmaI and KpnI and then inserted into the C. elegans eGFP expression vector pPD117.01, which was digested with the same enzymes. The ϳ3.5-kbp cdr-1 promoter/enhancer fragment, described above, was then inserted into BamHI and SmaI sites that are upstream of the CDR-1-eGFP cDNA in the expression vector. To remove the mec-7 promoter fragment that is present in the pPD117.01 vector, the cdr-1/ CDR-1-eGFP expression plasmid was digested with SacII and BamHI. The ϳ8.1-kbp plasmid was then purified; the ends were subsequently made blunt and then joined. Transgenic C. elegans were generated by microinjection as described above. This strain was designated JF13 (mtEx10, cdr-1/CDR-1-GFP).
Prior to establishing the intracellular location of CDR-1, intestinal cell lysosomes of JF13 (mtEx10, cdr-1/CDR-1-GFP) C. elegans were rhodamine-labeled. This was accomplished by feeding transgenic nematodes RITC-dextran. It has been previously demonstrated that C. elegans accumulate RITC-dextran into intestinal cell lysosomes (42). To label the lysosomes, C. elegans were suspended in BU buffer (70 mM NaCl, 72 mM potassium phosphate, pH 7.0) that contained 10 mg/ml of RITC-dextran (Sigma Chemical Co.) (42). After incubating at 20°C for 8 h, nematodes were washed three times with BU buffer and collected by centrifugation. To remove the unincorporated probe from the intestinal lumen, C. elegans were then placed on NGM plates with food and allowed to feed for 4 h. The nematodes were collected from the plates, suspended in M9 buffer and exposed to 100 M cadmium at 20°C for 16 h. As a control, an identical population of JF13 (mtEx10, cdr-1/CDR-1-GFP) C. elegans was grown in the absence of cadmium. The intracellular locations of CDR-1-eGFP and lysosomes were determined by fluorescence microscopy using eGFP-and rhodamine-specific filter sets, respectively.
RNA interference experiments were performed by growing nematodes on NGM plates under one of four different conditions: (a) 1 mM IPTG; (b) 1 mM IPTG and 50 M cadmium; (c) 1 mM IPTG and 50 g/ml ampicillin; and (d) 1 mM IPTG, 50 g/ml ampicillin and 50 M cadmium. The IPTG is used to induce the expression of T7 RNA polymerase, which is under the control of the lac promoter (45). These plates were seeded with the pPD129.36-CDR-1-transformed BL21(DE3) bacteria, or as a control, non-transformed bacteria (46). L4-stage hermaphrodite C. elegans were placed onto the plates and incubated at 22°C, as described previously (46). The phenotype was investigated in adult C. elegans following 24-, 48-, and 72-h incubations. In addition, the phenotype was examined in the F1 progeny grown at 22°C for 48 h.

RESULTS
Cloning and Sequence Analysis of the CDR-1 cDNA-The full-length CDR-1 cDNA was derived from the EST DDRT16 using 5Ј-RACE. The full-length of CDR-1 was determined from the (a) size of the mRNA transcript, estimated by Northern blot analysis; (b) longest open reading frame; and (c) available C. elegans genomic sequence data.
A Northern blot that contained size-fractioned C. elegans poly(A) ϩ RNA was hybridized to a 32 P-labled DDRT16 cDNA probe. A single mRNA that hybridizes to DDRT16 is present in cadmium-exposed C. elegans. This mRNA has a size of ϳ900 nt and is not present at a significant level in mRNA prepared from non-exposed nematodes (results not shown).
To obtain the full-length sequence of CDR-1, twenty independent cDNA clones that contain the 5Ј-end of CDR-1 were analyzed. The length of the CDR-1 mRNA, determined by assembling overlapping sequences from DDRT16 and 5Ј-RACE products is 894 nt, including 3Ј-and 5Ј-untranslated regions (Fig. 1). This value is in good agreement with the size of the mRNA obtained by Northern blotting.
CDR-1 mRNA contains a single open reading frame of 831 nt. An initiator ATG codon (nt 16 -18) lies within the context of a consensus C. elegans translation start site, (A/G)NNATGT (Fig. 1). The stop codon TGA (nt 847-849) is followed by a 46-nt, 3Ј-untranslated region, with a typical polyadenylation signal (AATAAA) that is 13-18 nt 5Ј of the poly(A) tail.
The 5Ј-ends of many C. elegans mRNAs are shortened and covalently modified by trans-splicing reactions (47). The sequences of the 5Ј-ends of the DNA fragments obtained during the 5Ј-RACE analysis were 100% identical to the genomic sequence in the C. elegans cosmid F35E8 (see below). Thus, the cdr-1 transcript is not trans-spliced. Comparison of mRNA sequences, ascertained by RACE, to genomic sequences allows the transcription start site to be determined. cdr-1 transcription begins at a unique site (A, ϩ1, Fig. 1) and contains a 15-nt 5Ј-untranslated region, which is 24-bp downstream from a consensus TATA box sequence (Table I).
The full-length CDR-1 nucleotide sequence was compared with sequences in several of the GenBank TM data bases using the BLASTN program. CDR-1 is 100% identical to the predicted C. elegans cDNA F35E8.11, which was identified from the F35E8 cosmid sequence data using the Genfinder program (48). The CDR-1 mRNA does not have any significant levels of homology to any non-Caenorhabditis nucleotide sequences currently in the data bases.
Structure and Organization of cdr-1-cdr-1 is located near the center of chromosome V in the C. elegans genome. The size of cdr-1, including the structural gene and the upstream regulatory region, is predicted to be 2475 bp. The predicted gene includes the region of F35E8 between (a) the end of the 3Ј-untranslated region of the CDR-1 mRNA, and (b) the translation start codon of the predicted mRNA F35E8.10, which is located 1016 bp upstream of the cdr-1 transcription start site. F35E8.10 is transcribed from the complementary DNA strand, relative to cdr-1. F35E8.10 is predicted to encode a novel protein; however, homologous C. elegans ESTs have been identified, suggesting that the mRNA is expressed in vivo.
Comparison between the CDR-1 mRNA and genomic sequences enabled the elucidation of the intron/exon organization of the cdr-1 structural gene. The primary CDR-1 transcript is derived from six exons that are contained within ϳ1.5 kbp (Fig.  2). All of the introns in this gene are small, ranging in size from 52 to 347 bp. The occurrence of small introns is a common characteristic of the C. elegans genome. The nucleotide consensus sequences adjacent to the intron/exon junctions in cdr-1 are similar to those described for other C. elegans genes (26,49). Potential UREs were identified by comparing the consensus sequences of UREs that are known to control the transcription of other eukaryotic metal/stress-responsive genes to the contiguous ϳ1-kbp DNA, which flanks the 5Ј-end of cdr-1 (Table I). cdr-1 contains sequences that are 100% identical to the consensus sequences for a TATA box, a single metal response element, two antioxidant response elements, and a heat shock element (50 -52). In addition, cdr-1 contains two potential GATA-transcription factor-binding sites (53,54). A homologous cAMP-response element was also identified (55).
Analysis of the CDR-1 Protein Sequence-The CDR-1 mRNA is predicted to encode a protein of 277 amino acids with a molecular mass of 32,126 and an estimated isoelectric point of 7.60. Hydropathy analysis of the deduced amino acid sequence indicates that CDR-1 is highly hydrophobic, with greater than 80% hydrophobicity over the entire protein (Fig. 3) (56). CDR-1 is predicted to be an integral membrane protein, with two transmembrane spanning domains; one is located at the N terminus (residues 2-23), and the second is near the middle of the protein (residues 157-175) (Fig. 3). The protein may be targeted to one or more membranes of either the endoplasmic reticulum or secretory vesicles (57). CDR-1, however, lacks any of the common protein cleavage signal sequences. CDR-1 contains several potential phosphorylation sites. Prosite analysis reveals consensus motifs for two protein kinase C phosphorylation sites (residues 28 -30 and 84 -86), four casein kinase II phosphorylation sites (residues 123-126, 157-160, 198 -201, and 227-230), and one tyrosine kinase phosphorylation site (residues 148 -155). It also contains an myc-type, helix-loop-helix dimerization domain (residues 87-95). CDR-1 does not contain consensus motifs for glycosylation or myristoylation sites (58).
The CDR-1 amino acid sequence was compared with sequences in GenBank TM data bases using the BLASTP and TBLASTN programs. CDR-1 has the highest levels of homology (27% identity, 46% similarity) with the failed axon connection protein (FAX) from Drosophila melanogaster (accession number: S58776). The Drosophila FAX protein is involved in neuronal development, and it was discovered as a genetic enhancer of a mutation in the tyrosine kinase abl (59). The alignment of their amino acid sequences is presented in Fig. 4. It should be noted that the regions that are adjacent to the predicted transmembrane-spanning domains have the highest levels of homology.
Induction of CDR-1 mRNA by Cadmium-CDR-1 mRNA is not evident in a Northern blot of total RNA prepared from C. elegans grown in the absence of added cadmium. Following a 1-h exposure to 100 M cadmium, CDR-1 mRNA is detected. The steady-state level of CDR-1 mRNA increases 5-fold to obtain a maximal steady-state concentration within 10 -15 h (Fig. 5). The rate of CDR-1 mRNA accumulation (t1 ⁄2 ϳ 4.5 h) is identical to those values reported for mtl-1 and mtl-2 (19).
Cell-specific and Developmental Expression of cdr-1-The ability of cadmium to affect the cell-specific, developmentally regulated pattern of cdr-1 transcription was investigated by exposing several independent lines of transgenic C. elegans, which contain the cdr-1/lacZ reporter transgene, to 100 M cadmium for 24 h. Following cadmium exposure, cdr-1 promoter activity is evident exclusively in the intestinal cells of C. elegans (Fig. 6). Transcription was observed in all post-embryonic stages of development, and not in developing embryos. In the absence of cadmium, reporter transgene activity was not observed.
Whole mount in situ hybridization analysis was used to monitor the level and cellular distribution of CDR-1 mRNA in larval and adult nematodes. When cadmium-exposed C. elegans were hybridized to an antisense CDR-1 cDNA probe, CDR-1 mRNA was observed throughout the intestine in all post-embryonic developmental stages (Fig. 6). In contrast, CDR-1 mRNA was not detected at a significant level in nonexposed nematodes (results not shown). As a negative control, cadmium-treated C. elegans were hybridized to a digoxigeninlabeled CDR-1 sense probe. When the sense probe was used, CDR-1 mRNA was not detected (results not shown).
The results obtained in the in situ hybridization experiments corroborate those found with the cdr-1/lacZ-transgenic C. elegans. Both confirm that cdr-1 transcription is induced by cadmium and is limited to the intestinal cells.
Dose Dependence of cdr-1 Transcription in Transgenic C. elegans-Transgenic C. elegans that contain an integrated cdr-1/lacZ transgene were used to monitor the effects of various concentrations of cadmium on cdr-1 transcription. C. elegans  6. Cell-specific expression of cdr-1. A and B, C. elegans, strain JF9 (mtIs7, cdr-1/lacZ) was exposed to 100 M CdCl 2 for 24 h and then stained for ␤-galactosidase activity as previously described (19). A, reporter gene activity in the intestinal cells of a young adult and an L3 larva. B, cdr-1 promoter activity in the intestinal cells of an L1 larva. C, in situ hybridization of cadmium-treated L1 C. elegans larvae. CDR-1 mRNA was visualized by exposing wild type nematodes to a digoxigenin-dUTP-labeled CDR-1 antisense probe as described under "Experimental Procedures." C shows that the CDR-1 transcript is expressed throughout the intestine of L1. "ph " marks the location of the pharynx in all panels. Nematodes were photographed using Nomarski optics as described previously (19). strain JF9 (mtIs7, cdr-1/lacZ) was exposed to cadmium for 24 h and the level of cdr-1 transcription determined from the yield of blue chromogen (37). The inducible, cell-specific pattern of cdr-1 transcription was identical at all cadmium concentrations (i.e. only occurred in intestinal cells). However, the magnitude of the response was concentration dependent (Fig. 7).
cdr-1 transcription was induced at a minimal cadmium concentration of 1 M. A maximal level of expression was observed following exposure to 25 M cadmium. Using the data presented in Fig. 7, an EC 50 value ϳ2 M was derived. The EC 50 is defined as the concentration of cadmium added to the growth medium that will induce cdr-1 transcription to a level that is 50% of the maximal level detected. The EC 50 value for cdr-1 transcription is similar to that reported for mtl-2, and it is two to three orders of magnitude lower than the 24-h exposure LC 50 for cadmium (37)(38)(39).
Effects of Various Stressors on cdr-1 Transcription-C. elegans strain JF9 (mtIs7, cdr-1/lacZ) was used to examine the effects of environmental stressors on cdr-1 transcription. These stressors included heat shock, oxidative stress, and other transition metals. The exposure conditions (i.e. toxicant concentrations and exposure times) used in these studies were identical to those previously described to induce the transcription of other stress response genes (Table II). The only stressor that was able to induce cdr-1 transcription was cadmium. This characteristic is unique to cdr-1. Typically, the transcription of stress-response genes can be induced by several of the stressors examined.
CDR-1 Protein Localization in Vivo-To determine the intracellular location of CDR-1, transgenic C. elegans were generated that contain CDR-1-eGFP fusion protein whose expression is regulated by the cdr-1 promoter. C. elegans strain JF13 (mtEx10, cdr-1/CDR-1-GFP) was exposed to 100 M cadmium for 24 h. Initial observations of cadmium-treated JF13 (mtEx10, cdr-1/CDR-1-GFP) nematodes clearly demonstrated that the CDR-1-eGFP fusion protein was concentrated in small punctate structures in the intestinal cells (Fig. 8). The size and distribution of these structures were similar to those of lysosomes (42). In contrast, when a non-fusion form of eGFP is expressed in intestinal cells, it accumulates in the cytoplasm but not in vesicles. 3 To confirm that CDR-1 was targeted to the lysosomes, JF13 (mtEx10, cdr-1/CDR-1-GFP) C. elegans were fed RITC-labeled dextran, prior to cadmium exposure. In the double-label experiment, the eGFP and rhodamine signals colocalized to the same structures (Fig. 8). Superimposition of the CDR-1-eGFP and rhodamine signals on the light micrograph confirmed that CDR-1 is targeted to lysosomes (Fig. 8). This observation confirmed that CDR-1 is targeted to intestinal cell lysosomes.
RNA Interference-To determine the role of CDR-1 in the resistance to cadmium toxicity, wild type C. elegans were fed E. coli that were transformed with a plasmid that expresses CDR-1 dsRNA. This technique has been shown to phenocopy null mutations and to be as efficient as injecting dsRNA (46).
The phenotype of C. elegans that were fed CDR-1 dsRNA, in  7. Dose dependence of cdr-1 promoter activity in transgenic C. elegans. A mixed-stage population of JF9 (mtIs7, cdr-1/lacZ) C. elegans were exposed to various concentrations of CdCl 2 for 24 h. The nematodes were then fixed, permeabilized, and stained for ␤-galactosidase activity for 2.5 h. The blue chlorbromindigo precipitate was extracted with dimethylformamide and quantified as previously described (37).  8. Intracellular location of the CDR-1 protein in vivo. Transgenic C. elegans, which contain a cdr-1/CDR-1-eGFP expression vector (JF13 (mtEx10, cdr-1/CDR-1-GFP)), were fed RITC-dextran and then exposed to cadmium as described under "Experimental Procedures." Upper panels, JF13 (mtEx10, cdr-1/CDR-1-GFP) C. elegans L2 larva. "L " and "Ph " indicate the location of the intestinal lumen and the nematode pharynx, respectively. Left panel, the location of CDR-1 visualized by fluorescent microscopy. Right panel, RITC-labeled lysosomes visualized using a rhodamine filter set. Lower panel, composite, high magnification (1000ϫ) view of an adult JF13 (mtEx10, cdr-1/CDR-1-GFP) C. elegans. The nematode is oriented with the gonad on the left and the intestine on the right. Fluorescence images obtained using eGFP and rhodamine filter sets were superimposed on the light microphotograph. Locations where the eGFP and rhodamine signals are coincident can be identified by the blue color. Black arrows indicate eGFP and rhodamine-containing lysosomes. White arrows indicate nonlysosomal gut granules. the absence of cadmium, was similar to control nematodes. These animals reproduced and developed similar to C. elegans grown in the absence of dsRNA and cadmium. In several of the C. elegans, small vesicles were observed in the body cavity ( Fig.  9). In contrast, C. elegans that were fed CDR-1 dsRNA and grown in the presence of 50 M cadmium were small, sterile, and failed to develop into adulthood. Growth and reproduction were slightly inhibited in nematodes grown in cadmium-containing medium, in the absence of dsRNA. However, these nematodes developed into adulthood and successfully reproduced. Microscopic examination of CDR-1 null nematodes exposed to cadmium showed that they accumulated fluid in the pseudo-ceolomic space, which eventually filled the entire space (Fig. 9). In addition, when gravid adult C. elegans, which were maintained in the absence of cadmium and dsRNA, were transferred to plates containing cadmium and dsRNA-containing bacteria, fluid accumulated in the body cavity. JF13 (mtEx10, cdr-1/CDR-1-GFP) C. elegans that were grown in the presence of cadmium and CDR-1 dsRNA exhibited a decrease in the level of CDR-1-eGFP, compared with JF13 (mtEx10, cdr-1/CDR-1-GFP) nematodes grown in cadmium alone (results not shown). The RNA interference results confirmed that CDR-1 is essential for defense against cadmium toxicity.

DISCUSSION
The reverse transcriptase-PCR protocol of differential display was used to identify new cadmium-responsive genes in C. elegans (17). Here we report on the cloning and analysis of a novel cadmium-inducible gene, designated cdr-1. CDR-1 encodes a predicted 277-amino acid, 32-kDa protein (Fig. 1). Amino acid sequence analysis identified two transmembrane-spanning domains, and it predicts that CDR-1 will be an integral membrane protein, which is targeted to the secretory pathway. The later prediction is supported by in vivo studies where it was clearly demonstrated that CDR-1 co-localizes with lysosomes (Fig. 8).
BLAST amino acid sequence analysis of CDR-1 showed that it is a novel protein. However, it does have regions that are homologous to the Drosophila FAX protein. FAX is a 47-kDa protein that is expressed in the Drosophila embryonic mesoderm and axons of the central nervous system. It has been shown that fax is constitutively transcribed during embryogenesis and that the protein is targeted to the plasma membrane (59). There are no reports that address the ability of environmental toxicants to induce fax transcription. In contrast, cdr-1 is transcribed only in intestinal cells in response to cadmium exposure. The function of FAX has not been fully defined. However, genetic data indicate that the protein is involved in signal transduction cascades, which affect neuronal development. The involvement of CDR-1 in signaling pathways is currently unknown, although the protein contains several potential phosphorylation sites. Because FAX and CDR-1 show (a) different patterns of cell-specific expression, (b) intracellular targeting, and (c) transcriptional regulation, their functions may be dissimilar. Because both proteins are targeted to cellular membranes, the shared amino acid homology may reflect intracellular targeting and not a common function.
cdr-1 transcription is induced in response to cadmium exposure (Figs. 5 and 7). In addition, the steady-state level of CDR-1 mRNA rapidly increases in the presence of this toxicant (Fig.  5). These observations confirm that cadmium is a potent activator of cdr-1 transcription. Cadmium-inducible cdr-1 transcription may be controlled by one of several potential UREs, which include AREs, an HSE, and an MRE. Both HSEs and AREs mediate stress-inducible transcription of a variety of eukaryotic genes. Cadmium has also been shown to activate transcription via these UREs (60,61). However, heat shock or oxidative stress did not induce cdr-1 transcription in vivo (Table II), which suggests that these elements may not regulate cdr-1 transcription. The MRE is an obvious candidate as a regulator for cdr-1 transcription. It has been shown that MREs play essential roles in controlling cadmium-activated MT transcription in both vertebrates and invertebrates (62,63). In contrast, there is evidence to suggest that MREs do not function in C. elegans. First, the mtl-1 promoter does not contain a consensus MRE sequence (19). Second, the MRE sequence that is present in mtl-2 does not control metal inducibility. 3 Finally, analysis of the C. elegans genome and predicted protein data bases has not identified a homologue of the MRE-binding transcription factor MTF-1 (64). The contribution of MRE, HSE, and ARE in regulating cdr-1 transcription remains to be confirmed by further analysis.
Transgenic C. elegans and in situ hybridization studies confirm that cdr-1 is transcribed only in intestinal cells. Intestinal cell-specific expression of many C. elegans genes, including those encoding the vitellogenins, gut carboxylesterase, metalinducible mtl-1 and mtl-2, P-glycoprotein, GATA-binding transcription factor-2, and aspartic acid and cysteine proteases, is dependent on the presence of one or more copies of UREs designated GATA elements (65)(66)(67)(68)(69)(70)(71). These elements are binding sites for members of the GATA family of transcription factors (69,72). In C. elegans, GATA elements that are responsible for controlling intestinal cell-specific transcription have the consensus sequence (A/T)GATA(A/G). Consistent with these observations, two consensus GATA elements are present in the promoter/enhancer region of cdr-1 (Table I).
GATA elements are also found in the promoters of the two FIG. 9. Morphological change associated with RNA interference of CDR-1. Wild type C. elegans were grown on NGM plates containing 50 M cadmium and CDR-1 dsRNA (upper), 50 M cadmium alone (middle), or CDR-1 dsRNA alone (lower). The arrowheads in the lower panels indicate the location of vesicles in the body cavity. The images in the upper and middle panels are at 1000ϫ magnification, whereas the images in the lower panels are at 400ϫ. All of the images were collected using a Zeiss Axioscope, equipped with Nomarski optics and a charge-coupled device camera.
C. elegans MT genes. This raises the possibility that GATA elements can regulate metal-inducible transcription. However, with the exception of cdr-1, mtl-1, and mtl-2, metal-activated transcription of genes that contain GATA elements has not been reported. In addition, it has been shown that GATA elements are not responsible for controlling metal-activated transcription of mtl-1 or mtl-2 (67).
The magnitude and rate of mRNA accumulation following cadmium exposure, and the sensitivities (i.e. the minimal concentration of metal that can induce transcription) of cdr-1 are similar to those of mtl-1 and mtl-2 (19,37,73). In addition, the EC 50 for cadmium-inducible cdr-1 transcription is significantly below the 24-h LC 50 for cadmium (2 M versus ϳ8000 M) (38). These observations suggest that cadmium is a specific activator of cdr-1 transcription and that cdr-1 expression is not induced as a secondary response to metal exposure (i.e. activated in response to cell death). In addition, the parallel responses of cdr-1, mtl-1, and mtl-2 to cadmium exposure suggest that they may share common regulatory pathways.
A unique property of cdr-1 is that only cadmium has been found to be able to significantly induce its transcription (Table  II). Typically, the transcription of cadmium-inducible genes is activated by a variety of stressors. For example, the transcription of MT, heme oxygenase, p58, fos, jun, myc, superoxide dismutase, heat shock proteins, multidrug resistance protein, and progesterone receptor increases following exposure to a variety of toxicants, such as other transition metals, heat, oxidative stress, organic chemicals, or endogenous ligands. The mechanism by which cdr-1 transcription is constrained remains to be determined.
To determine the contribution of CDR-1 in the resistance to cadmium toxicity in C. elegans, dsRNA interference was employed. In the absence of cadmium, wild type nematodes in which CDR-1 expression was inhibited, developed and reproduced in a manner similar to control C. elegans. The only phenotypic change observed was the accumulation of a small number of fluid droplets in the pseudo-ceolomic space. In contrast, CDR-1 dsRNA nematodes that were exposed to cadmium were smaller than the controls, failed to reproduce or develop, and had a shorter life span. This phenotype is similar to that observed in C. elegans exposed to cadmium where phytochelatin synthase was inhibited (74). This suggests that the inhibition of development and sterility may be a consequence of cadmium toxicity and not a unique phenotype associated with cdr-1 or pcs-1.
The precise mechanism by which CDR-1 protects against cadmium toxicity and its function in the absence of metal remain to be determined. Several observations, however, support the hypothesis that CDR-1 may function in osmoregulation, possibly to pump cadmium or other ions into lysosomes. First, CDR-1 is predicted to be an integral membrane protein, and it is targeted to lysosomes. Lysosomes function in the intracellular storage of transition metals in species ranging from yeast to humans. Exposing mussels, snails, or rats to cadmium results in the lysosomal accumulation of this metal (75)(76)(77). Second, several transition metal ion pumps that transport cadmium into lysosomes have been identified. In yeast, the cadmium factor gene (YCF1) encodes an ATP-binding cassette protein that transports cadmium-glutathione complexes into vesicles (78). Havelaar et al. (13) identified a lysosomal P-type transition metal transporter in rats. Many of the lysosomal metal-ion pumps function in metal detoxification (79,80). Finally, the accumulation of fluid in the body cavity of CDR-1 dsRNA-treated C. elegans that were exposed to cadmium is reminiscent of the phenotype described when the cells of the secretory-excretory system were ablated (81). Destruction of the pore, duct, or excretory cell resulted in the accumulation of fluid in the body cavity. It has been suggested that these cells may function in osmoregulation in C. elegans.
The amino acid sequence of CDR-1 was compared with members of the three classes of eukaryotic metal transport proteins: P-type ATPases, ABC-type transporters, and the CDF family of transporters (15,82,83). FASTA (84) comparisons between CDR-1 and representative proteins from each class of transporter failed to identify large regions of homology. The CDF family of transporters, which includes two C. elegans proteins, is defined by the consensus sequence (15) as follows: S-X-(ASG)-(LIVMT) 2 -(SAT)-(DA)-(SGAL)-(LIVFYA)-(HDN)-X 3 -D-X 2 -(SA). A region that is partially homologous to the CDF consensus sequence is present in CDR-1 at amino acid residues 129 -143, SVAL-SKFADHHLFFV (residues shown in boldface are identical to the CDF consensus sequence). This suggests that CDR-1 may be a member of this class of transporter. However, CDR-1 is missing the evolutionarily conserved aspartic acid residue (double underlined). In addition, CDF proteins typically contain six transmembrane-spanning domains, whereas only two spanning domains are predicted for CDR-1.
In conclusion, cdr-1 is transcribed in intestinal cells in response to cadmium exposure. The gene encodes a novel protein that is targeted to lysosomes and is required for resistance to cadmium toxicity. The relation between CDR-1 expression, cadmium detoxification, and osmoregulation in C. elegans remains to be resolved.