The Essential Vertebrate ABCE1 Protein Interacts with Eukaryotic Initiation Factors*

The ABCE1 gene is a member of the ATP-binding cassette (ABC) multigene family and is composed of two nucleotide binding domains and an N-terminal Fe-S binding domain. The ABCE1 gene encodes a protein originally identified for its inhibition of ribonuclease L, a nuclease induced by interferon in mammalian cells. The protein is also required for the assembly of the HIV and SIV gag polypeptides. However, ABCE1 is one of the most highly conserved proteins and is found in one or two copies in all characterized eukaryotes and archaea. Yeast ABCE1/RLI1 is essential to cell division and interacts with translation initiation factors in the assembly of the pre-initiation complex. We show here that the human ABCE1 protein is essential for in vitro and in vivo translation of mRNA and that it binds to eIF2α and eIF5. Inhibition of the Xenopus ABCE1 arrests growth at the gastrula stage of development, consistent with a block in translation. The human ABCE1 gene contains 16 introns, and the extremely high degree of amino acid identity allows the evolution of its introns to be examined throughout eukaryotes. The demonstration that ABCE1 plays a role in vertebrate translation initiation extends the known functions of this highly conserved protein. Translation is a highly regulated process important to development and pathologies such as cancer, making ABCE1 a potential target for therapeutics. The evolutionary analysis supports a model in which an ancestral eukaryote had large number of introns and that many of these introns were lost in non-vertebrate lineages.

The induction of ribobuclease L (RNase L) 2 represents an important viral defense mechanism of mammalian cells against RNA viruses (1,2). RNase L is present in the cell in an inactive form and can be activated by interferon. Interferon causes the activation of oligoadenylate synthases producing 2Ј-5Ј-oligoadenylate. Bisbal et al. (3) described the isolation of a 68-kDa protein that binds to and inhibits RNase L and cloning of the gene. Although originally termed RNase L inhibitor (RLI) the gene is part of the ABC multigene family and its gene symbol is ABCE1. ABCE1 is induced during infection of cells with HIV-1 and an antisense con-struct directed against ABCE1 resulted in a reduction of viral load (4). In cell-free extracts HIV-1 gag protein can assemble into viral capsids (5). This process is ATP-dependent and was shown to require a 68-kDa protein (HP68) identified as ABCE1/RLI. This same protein also functions in the assembly of HIV-2 and SIVmac (6).
Most of the ABC family genes encode large transport proteins that contain 6 -17 transmembrane domains (7). ABCE1 is one of four human ABC genes that contain only nucleotide binding domains and are therefore not likely to be transporters. Of the 48 human ABC proteins, ABCE1 is the most conserved with a single copy of the gene in every characterized eukaryote, except for Arabidopsis, which has two ABCE1-like genes. In addition, there is an ABCE1-related gene in all characterized archae but not in prokaryotes, demonstrating that this is one of the most conserved genes and is likely the ancestral ABC gene. Because the interferon system is not found outside of mammals, ABCE1 must have another function. Dong et al. (7) documented in Saccharomyces cerevisiae that ABCE1 (known as RLI1) is essential, and we found that yeast spores deficient in RLI1 do not undergo a single cell division. 3 RLI1 binds to the eukaryotic initiation factors eIF2 and eIF5 and forms part of the pre-initiation complex required for the translation of mRNA in yeast. In addition, depletion of RLI1 causes a loss of polysome formation in yeast (7). The N terminus of the ABCE1 protein contains a Fe-S binding domain, and the protein binds Fe-S clusters (8). RLI1 was also found to be required for the processing of ribosomal subunits (9). Yeast RLI1 was able to rescue the lethal phenotype in RLI1 gene deletion in yeast spores, whereas the human ortholog was not able to do so. 3 Therefore, the question remained as to whether the mammalian ABCE1 plays a role in protein synthesis. We present in this report that human anti-ABCE1 antibody specifically inhibits mRNA translation, and ABCE1 protein binds to eIF2␣ and eIF5 initiation factors. Suppressing ABCE1 expression in a human cell line changes the polysome profile, with a decrease in large polysomes and an increase in 80 S ribosomal subunits. We also show that ABCE1 is essential in embryo development of Xenopus laevis, another vertebrate model.
Frogs, Embryos, Antisense, and Injections-X. laevis were purchased from Nasco (Fort Atkinson, WI). The embryo preparation, injection, and culture were performed as described previously (10,11). Antisense ABCE1 morpholino oligonucleotides (MOs) were designed and synthesized by Gene Tools. The sequences are as follows: TGGTAAGTTT-GTCTGCCATGTTGTC (oligonucleotide targets sequence in positions Ϫ6 through ϩ19 relative to the initiation codon) and AAGATTGCACCTATTCAGGAAAGGA (oligonucleotide targets a splice acceptor site) of Xenopus locus BC046573.
Immunoblotting-HeLa cells in 75-ml flasks were lysed when nearly confluent using 1 ml of mammalian protein extraction reagent (Pierce) with mammalian protease inhibitor mixture (Sigma) as described in the vendor's protocol. The lysate was centrifuged at 12,000 rpm for 15 min at 4°C. The supernatant was reacted with primary antibodies. The protein-antibody complex was isolated by protein G/A-agarose beads (Oncogene Inc.). Protein-antibody-protein G/A-agarose complexes were washed with radioimmune precipitation assay buffer (13). The complexes in gel loading buffer (Invitrogen) were heated at 70 -75°C for 10 min. Proteins were fractionated on SDS-PAGE (Invitrogen; NuPAGE, 4 -12% BisTris gel) and transferred electrophoretically to Immobilon-P (Millipore) membranes. The membrane was detected using primary and secondary antibodies and an ECL chemiluminescent kit (Amersham Biosciences).
In Vitro Translation-Rabbit reticulocyte lysate systems, luciferase mRNA, and luciferase assay reagents were purchased from Promega. The reaction conditions and luciferase assay were performed according to the protocol of Promega. The only modification was a reduction in the reaction scale to 0.25 volume of the original volume. In the poly(U) (Sigma) assay, L-[2,3,4,5,6-3 H]phenylalanine (9.25 MBq) was purchased from Amersham Biosciences. The assays were performed basically as described by Monnier et al. (12).
Antibodies-Anti-ABCE1 and anti-eIF2␣ polyclonal antibodies (0.2 and 0.5 mg/ml, respectively) and pre-immune rabbit serum (a control for ABCE1 antibody) were purchased from Novus Biologicals Inc. Anti-eIF5 antibody and eIF5 recombinant protein were kindly provided by Dr. Umadas Maitra (Albert Einstein College of Medicine). The antibody reactions were performed according to manufacturer's instruction and the recommendation of Dr. Maitra. Horseradish peroxidase-conjugated secondary antibodies were purchased from Amersham Biosciences.
siRNA Assay-ABCE1 siRNAs and negative (Silencer TM Negative Control #1 siRNA) and positive (glyceraldehyde-3-phosphate dehydrogenase) control siRNA were designed and purchased from Ambion Inc. HiPerFect transfection reagent was obtained from Qiagen Inc. The HEK 293 cell line was used for siRNA transfection. A reverse transfection procedure was used in all of experiments according to vendor's instructions in 24-well plates. RNA isolation was from cell lysates by TRIzol (Invitrogen), according to Invitrogen's protocol. The real-time PCR method was used for quantitative analysis of mRNA. TaqMan probes (6-carboxytetramethylrhodamine-labeled) and primers and assay reagents were designed and provided by Applied Biosystems Inc. Realtime PCR was performed in an ABI PRISM 7700 sequence detector. The data analysis was according to the instruction of Applied Biosystems Inc.
Analysis of Protein Synthesis-HEK 293 cells were transfected by siRNA 203097 and control siRNA at 10 nM. After 3 days the cells were detached by trypsin and counted, and equal cell numbers for each sample were placed in 96-wells plates and the culture continued for 2 days at 37°C/CO 2 . The cells were labeled with [ 35 S]methionine/cysteine (PerkinElmer Life Science, L-[ 35 S]methionine specific activity, 1175.0 Ci/mmol, 7.9 mCi/ml. Total 35 S activity, 2.3 mCi/181 l) in DMEM lacking methionine/cysteine as described by Gorchakov et al. (13). After 30-min labeling, the cells were lysed by mammalian protein extraction buffer as described above. The aliquots lysate was loaded onto Whatman No. 3MM paper and counted according to the method of Monnier et al. (12).
Polysome Analysis-Cells (6.7 ml, 1 ϫ 10 5 cell/ml) and the complex formed by 3.4 l of siRNA (20 M) and 30 l of HiPerFect in 200 l of DMEM were transfected in a 10-cm plate. After 48 h of transfection, the medium was changed to normal growth medium. On day 3 control cells including control and HiPerFect-treated cells (ϳ60 -70% confluence) were treated with cycloheximide (Sigma) at 100 g/ml concentration for 3 min at 37°C as described (14). Then the cells were detached by 0.25% trypsin containing 100 g/ml cycloheximide, neutralized by 4°C growth medium, and washed twice at 4°C with 1ϫ phosphate-buffered saline. The growth medium and phosphate-buffered saline contained cycloheximide at 50 g/ml. On day 4 the cells treated by siRNA were harvested as described above. Three to four plates of cells were pooled and lysed with 100 l of buffer (15 mM Tris-HCl, pH 7.5, 10 mM MgCl 2, 60 mM NaCl, 1% Triton X-100, 100 mg/ml cycloheximide, 1 mg/ml heparin, and 400 units/ml RNase out) on ice. The lysates were centrifuged at 12,000 rpm for 15 min at 4°C. Lysates (100 -150 l) were layered into 10-ml, 7-47% sucrose gradients for sedimentation at 39,000 rpm for 2 h and 30 min (8). The fraction analysis was performed as described by Dong et al. (7).
Cell and Embryo Photographs-Embryos and cell cultures were photographed on an Olympus 1X70 microscope.
Bioinformatics-For intron analysis the genomic sequence of each gene was identified from GenBank TM and aligned with the cDNA sequence. The position and phase of each intron was manually determined. Phylogenetic analysis was performed by first aligning the amino acid sequence with ClustalX (15) and performing neighbor joining analysis in MEGA version 3.0 (16).

RESULTS
To understand whether the human ABCE1 protein plays a role in the translation of mammalian mRNAs, we incubated rabbit reticulocyte lysates with an antibody to human ABCE1. To confirm that the human ABCE1 antibody reacts with the rabbit ABCE1 protein, we performed an immunoprecipitation with rabbit reticulocyte lysate. The human ABCE1 antibody precipitates a single protein of the expected size ( Fig.  S1). Luciferase mRNA was added to the lysate as a template along with different quantities of ABCE1 antibody. Fig. 1A shows that, at 200 ng of antibody, luciferase activity was inhibited by 90% compared with the no-antibody control. Rabbit IgG was not effective in inhibiting translation of luciferase mRNA even at 10-fold higher concentrations than the ABCE1 antibody. A human eIF5 antibody also suppressed luciferase mRNA translation initiation.
To demonstrate that the effect of ABCE1 on luciferase translation is specific for mRNA, we tested the effect of ABCE1 antibody on the translation of poly(U). Translation of poly(U) is accomplished independent of all initiation factors. The addition of ABCE1 antibody, at levels that severely inhibit translation of luciferase mRNA, had no significant effect on the translation of poly(U) (Fig. 1B). Therefore, the ABCE1 protein is an essential factor in vertebrate translation initiation.
To determine whether the human ABCE1 protein is associated with the pre-initiation complex components eIF5 and eIF2, as in yeast (7), we performed immunoprecipitation with ABCE1, eIF2␣, and eIF5 antibodies in HeLa cell lysates. Precipitated proteins were separated on SDS-PAGE gels and detected with specific antisera. Fig. 1C shows that the eIF5 co-precipitated with ABCE1. In the reverse experiment, ABCE1 co-immunoprecipitated with eIF5 and to a lesser extent with eIF2␣ ( Fig.  1C). Using an eIF2␣ antibody for detection we were also able to determine that eIF2␣ protein interacts with both eIF5 and ABCE1 (data not shown). Therefore, as in S. cerevisiae, ABCE1 associates with both eIF5 and eIF2.
To further study the role of ABCE1 on mammalian cell growth, we employed siRNA to reduce the expression of the ABCE1 gene in a human tumor cell line (HEK 293). The sequences and position in the mRNA of six pairs of siRNA are shown in Table 1. We titrated the capability of these siRNAs to suppress ABCE1 gene expression and to avoid off-target effects, we used low concentrations of siRNA. Transfection with four of siRNAs (at 5-10 nM) for 49 h suppressed ABCE1 mRNA expression more than 70%. At 72 h after transfection we carried out protein analysis of ABCE1. Immunoblotting showed that siRNAs 203097 and 203098, at 10 nM, substantially reduced ABCE1 protein expression (Fig. S2).
To evaluate the effect of suppression of ABCE1 expression on protein synthesis, the polysome profile was analyzed in cells treated with 10 nM siRNA 203097. In these cells, the large polysomes were significantly reduced, with a commensurate increase in free 80 S ribosomes (Fig. 2C). This shift from polysomes to monosomes is consistent with a severe impairment of translation initiation. The inhibition of the rate of translation initiation would be predicted to have profound effects on cell viability. Fig. 4A shows that after 7 days of transfection with siRNA 203097 the proliferation of the cells was dramatically decreased. From day 3 to 7, the cell number of siRNA transfected cells was only increased 18%. However, the cell number of control samples was increased 320 to 350% and nearly reached confluence in a 24-well plate. Morphologically the transfected cells were clearly less dense, and there were many rounded, dying cells (data not shown). Therefore, suppression of ABCE1 expression inhibits cell proliferation in a human tumor cell line. Total incorporation of 35 S into methionine and cysteine was also measured and found to be reduced by over 90% (Fig. 2B).
We have established a conserved function for ABCE1 in the protein translation process and provided evidence that it is critical for the proliferation of human cells in culture. Thus, we would expect an essential role for this protein in vertebrate development. To test this concept, two antisense MOs were designed and injected into X. laevis embryos. These two MOs were injected at various concentrations into two-cell stage embryos and allowed to develop in culture. One MO spanned the ATG start codon sequence, thus inhibiting mRNA translation. The other MO traversed the splice acceptor site and effectively suppressed proper  (24). Radioactivity incorporated was measured by liquid scintillation spectroscopy. C, immunoprecipitation analysis with anti-ABCE1, -eIF2␣, and -eIF5 antibodies. HeLa cells in 75-cm 2 flasks were lysed with 1 ml of protein extraction buffer, and 0.5-ml supernatants were reacted with 1 l of anti-ABCE1 antibody, 1 l of anti-eIF2␣ antibody, 1 l of anti-eIF5 antibody, and 1 l of pre-immune rabbit serum, respectively, for immunoprecipitation. The immunoprecipitation products were subjected to Western blot analysis. Left panel, anti-eIF5 antibody (2000ϫ dilution) reacted with proteins immunoprecipitated by anti-eIF2␣, -eIF5, and -ABCE1 antibodies, respectively. Lane 1 contains 100 ng of purified eIF5 protein. Right panel, anti-ABCE1 antibody (1000ϫ dilution) reacted with proteins immunoprecipitated by anti-eIF5, -eIF2␣, and -ABCE1 antibodies, and pre-immune rabbit serum, respectively. All of second antibody reactions were diluted 10,000 times. The blots were detected by the ECL kit.
processing of ABCE1 mRNA (data not shown). Both MOs consistently arrested embryonic development at the gastrulation stage (Fig. 3). As expected, a control MO had no effect on development. These data demonstrate that ABCE1 is essential for vertebrate development as it is for yeast and Caenorhabditis elegans. While this work was in progress a recessive lethal mutation in the Danio rerio (zebrafish) genome was identified (17). In addition, mutations in CG1703 the Drosophila ortholog have also been described that are lethal (18). Therefore ABCE1 is essential to all examined eukaryotes.
The ABCE1 gene is highly conserved, not only in eukaryotes, but a copy of the gene is found in all archaea. To determine the extent of amino acid conservation of the ABCE1 gene, sequences were extracted from data bases for 21 eukaryotic genes and 24 archaeal genes. The sequences were aligned and used to produce an unrooted neighbor joining phylogenetic tree. All of the eukaryotic genes cluster as expected and provide a relationship of species consistent with other analyses. The only exceptions are the nematode species (C. elegans, Caenorhabditis briggsiae) whose ABCE1 sequence clusters with the single celled eukaryotes (Fig. S3). This indicates that the ABCE1 gene in these species has evolved more rapidly. Within archaea, all of the ABCE1 genes from the Crenarchaea and Euryarchaea cluster together, and the one sequence from a Nanoarchaea sequence appears to be the most divergent.
Because of the very high level of conservation of the ABCE1 gene, and the fact that is present as a single copy, it is one of the few genes that can be used to unambiguously study the evolution of intervening sequences. As expected none of the archae ABCE1 genes have introns, whereas all of the eukaryotic sequences have at least one intron, except for S. cer-  2. A, inhibition of HEK 293 cell proliferation by siRNA. Cells (1 ϫ 10 5 /ml, 0.5 ml) were transfected with siRNA 203097 (at 10 nM final concentration) and control #1 siRNA with 3 l of HiPerfect transfection reagent in 24-well plates, using the reverse transfection procedure. The assay was performed in triplicate. After 48-h transfection, the medium was changed to normal growth medium for an additional 24 h. On the 3rd day the cells were detached using 250 l of trypsin and neutralized by 750 l of normal medium (10% FCS). The cells were counted, and 7.4 ϫ 10 4 cells were subcultured in 0.5 ml of medium in 24-well plates. All of assays were done in triplicate. The cell number was counted at the indicated time. B, 2 ϫ 10 3 treated and control cells in 100 l of growth medium were seeded in different wells of 96-wells plate. After 48-h incubation at 37°C/Co 2 in normal growth medium, the cells were washed with DMEM lacking methionine/cysteine with 2% (v/v) dialyzed FCS once. Then 80 l of washing medium containing 12 Ci of [ 35 S]methionine/cysteine was added to each well. After labeling, the cells in each well were lysed in 80 l of protein lysate buffer. The 20-l lysate was used for the assay of counts/min measurement. All of the assays were performed in triplicate. Blank control was no cells. The counts/ min of all samples represents counts/min of sample minus counts/min of blank control. C, polysome profile. Top, Cytoplasmic extracts (12.1 A 260 optical density) from control cells were sedimented in a sucrose gradient. Middle, control #1 siRNA-treated cells. Bottom, 10 nM siRNA 203097treated cells. MARCH 17, 2006 • VOLUME 281 • NUMBER 11 evisiae (Fig. 4). Each intron in every eukaryotic gene characterized was coded as to position in the protein and phase of the intron. Phase 0 introns splice between codons, phase 1 between the 1st and 2nd base of the codon, and phase 2 between the 2nd and 3rd bases. As can be seen in Fig. S3, all of the vertebrate ABC genes contain a conserved set of 16 introns, with the fish species, Fugu and Tetraodon possessing an additional intron. Therefore the organization of the ABCE1 gene has remained virtually unchanged in vertebrate evolution. The first 5 introns in the Arabidopsis ABCE1 gene are identical to the vertebrate genes, but in the 3Ј-half of the gene this plant species has 5 introns not present in vertebrates. The insects and nematodes have only 2-6 introns in their ABCE1 genes, but nearly all are in identical positions and phases as the vertebrate introns. This is consistent with a model in which an ancestral eukaryote had a large number of introns and many of these have been lost in certain species.

DISCUSSION
The ABCE1 gene is the most conserved member of the ABC gene family and is one of the most conserved genes in vertebrate and archaeal genomes (19). This fact alone suggests that the gene plays an essential role in biology that is common between archae and eukaryotes. In addition, null mutations in the gene are homozygous lethal in every organism that has been examined. In this paper we found that ABCE1 is essential in Xenopus and that suppression of translation or splicing with morpholino oligonucleotides results in the cessation of growth of the embryo during gastrulation, a period when the germ layers of the embryo are formed and the body plan of the mature organism is established. ABCE1 mRNA was detectable by reverse transcription-PCR in oocytes, and growth cessation presumably occurs at the point at which most of the maternal protein has degraded.
The ABCE1 protein was originally identified due to an interaction with and inhibition of RNase L, a nuclease induced by interferon (3). However, RNase L is not found outside of vertebrates, indicating that ABCE1 has alternate functions. The identification of the role of the RLI/ABCE1 protein in ribosome biogenesis and in assembly of the preinitiation complex of the ribosome in S. cerevisiae provides a function that is both essential and universal to eukaryotes (7). Therefore this is likely to be the original role of ABCE1 and the protein has adapted interaction with RNase L as a secondary function. However, we cannot rule out additional roles for ABCE1 in the cell.
The mammalian ribosome is substantially different from the yeast ribosome. For example the initiation complex compnent eIF3 has 14 subunits in mammals and only 6 in yeast. Therefore the importance of ABCE1 in mammalian protein initiation required experimental evidence. The important role of ABCE1 in protein synthesis is extended by the data presented here showing that the protein is essential in in vitro and in vivo translation of mammalian proteins. Antisera to ABCE1 block in vitro translation of mRNA in rabbit reticulocyte lysates but not of poly(U) molecules that can be translated independent of initiation factors. As in yeast, ABCE1 interacts with the eukaryotic initiation factors eIF5 and eIF2 components of the pre-initiation complex. Inhibition of ABCE1 in human cells results in dramatic inhibition of growth, reduction in the amount of large polysomes, and incorporation of labeled amino acids into newly synthesized protein. This is consistent with the results in yeast and supports a critical role for ABCE1 in the initiation of translation.
A single copy of the ABCE1 gene is found in all eukaryotes, except for Arabidopsis, which has two. This provides an excellent situation for the analysis of the phylogenetic relationships between species. From our analysis of a large number of eukaryotic and archae species ABCE1 reproduces the known relationships between species in both kingdoms. The only exception is the nematodes C. elegans and C. briggsiae in which the gene seems to be evolving more rapidly. The ABCE1 gene in C. elegans (Y39E4B.1) is essential and may have acquired other interacting partners that have driven its evolution (20). The conservation of ABCE1 also allows for an analysis of the evolution of intervening sequences in eukaryotes. All vertebrate ABCE1 genes contain 16 introns, except for some fish that have an additional 17th intron. The introns in the 5Ј-half of the gene are largely conserved in the plant (Arabidopsis) ABCE1 gene, but the introns in the 3Ј-half of the gene are in different locations in Arabidopsis as compared with vertebrates. All of the insect and nematode ABCE1 genes contain vastly reduced numbers of introns (2-6), and the introns they do contain are rarely shared, except for closely related species. For example the honey bee, mosquito, and fruit fly share FIGURE 3. Antisense ABCE1 morpholino oligonucleotides antisense to the initiation (ATG) codon inhibits X. laevis embryonic development. A, antisense ABCE1 morpholino oligonucleotide injection. Fertilized eggs were injected by 80, 40, and 20 ng of oligonucleotide and 80 ng of control morpholino oligonucleotide, respectively. Each group contained 11 embryos. After 12-h incubation at 23°C, all 11 embryos injected with 80 or 40 ng of oligonucleotide injection were unable to form a yolk plug, and 9 of 11 embryos injected with 20 ng of oligonucleotide were unable to form a yolk plug. All 11 embryos given control morpholino injection and uninjected controls formed clear yolk plugs. B, two representative control embryos from development stage 39 are also shown.
only one intron. However, all of the 11 introns found in the insect and nematode genes are in identical positions (and frames) as one of the vertebrate introns. This strongly supports the model derived from other genes that an ancestral eukaryote contained a large number of introns, many of which have been lost in certain lineages (21).
The inhibition of ABCE1 could have therapeutic applications. Because the protein is essential to most or all eukaryotes, specific inhibitors could be used in the treatment of pathogens. For example, inhibitors specific to plasmodia, fungi, and/or protozoan parasites could be used to inhibit such organisms as they infect human or other animals. We have shown here that ABCE1 inhibitors efficiently suppress the growth of human tumor cells. It is known that tumor cells have a high capacity for protein translation, and proteins involved in translation such as S6 kinase, mTOR, and 4E-BP1 are molecular targets for cancer therapy (22,23). It is possible that cancer cells are more sensitive to inhibition of protein translation through ABCE1 than are normal cells. Last, ABCE1 is required for the assembly of HIV-1 and other lentiviruses (5). Drugs that interfere with the HIV/ABCE1 interaction could be used as antiretroviral agents.
The ABCE1 protein is unusual in containing a Fe-S cluster binding site. It has long been known that Fe-S clusters are assembled in the mitochondria, and this process is essential to the cell. ABCE1 clearly represents one essential Fe-S containing protein. Whether this is the only essential protein in this class remains to be determined.
In summary we have demonstrated that the ABCE1 protein plays a role in the initiation of translation of proteins, similar to its role in yeast. We also show that this highly conserved gene is essential to Xenopus development and the growth of human cells, just as it is to yeast, C. elegans, Drosophila, and zebrafish. The ABCE1 gene is one of the few genes conserved between archae and eukaryotes and can therefore be used for phylogenetic analysis of species and the evolution of introns.