Microsporidia are amitochondriate eukaryotic microbes with fungal affinities and a common status of obligate intracellular parasites. A set of 13 potential genes encoding ATP-binding cassette (ABC) systems was identified in the fully sequenced genome of Encephalitozoon cuniculi. Our analyses of multiple alignments, phylogenetic trees and conserved motifs support a distribution of E. cuniculi ABC systems within only four subfamilies. Six half transporters are homologous to the yeast ATM1 mitochondrial protein, a finding which is in agreement with the hypothesis of a cryptic mitochondrion-derived compartment playing a role in the synthesis and transport of Fe–S clusters. Five half transporters are similar to the human ABCG1 and ABCG2 proteins, involved in regulation of lipid trafficking and anthracyclin resistance respectively. Two proteins with duplicated ABC domains are clearly candidate to non-transport ABC systems: the first is homologous to mammalian RNase L inhibitor and the second to the yeast translation initiation regulator GCN20. An unusual feature of ABC systems in E. cuniculi is the lack of homologs of P-glycoprotein and other ABC transporters which are involved in multiple drug resistance in a large number of eukaryotic microorganisms.
In the microbial world, microsporidia occupy a special place as both amitochondriate unicellular eukaryotes and obligate intracellular parasites. In the last decade, their phylogenetic placement has been debated and they are currently considered as fungi-like organisms. These microparasites are widespread among animal hosts and humans. Several species, including three of the Encephalitozoon genus (E. cuniculi, E. hellem and E. intestinalis), are known to cause opportunistic diseases in immunocompromised persons. The set of known anti-microsporidial drugs is restricted to albendazole, fumagillin and some derivatives. However, the susceptibility to these agents may depend on the considered microsporidian genus or species. For example, rather limited responses to albendazole treatment were observed in AIDS patients infected with Enterocytozoon bieneusi. As judged by in vitro growth-inhibition assays, three benzimidazoles were found to be more effective against E. cuniculi and E. hellem than against E. intestinalis. Such differences in susceptibility to microtubule disrupters may be hypothesized to result from divergences in tubulin sequences as well as from variations in rate of penetration of the drugs and/or in detoxification capacity.
Some major protozoan parasites such as Plasmodium have been demonstrated to overexpress a homolog of P-glycoprotein (P-gp), an ATP-binding cassette (ABC) transporter conferring multidrug resistance in human cells. In the case of microsporidia, an immunofluorescence study supported the association of P-gp epitopes with the cell surface of Encephalitozoon developing stages. Verapamil and cyclosporin A were shown to act as chemosensitizing agents for albendazole, suggesting that the effectiveness of this drug may be increased through the inhibition of P-gp pump activity. The completion of sequencing of the highly reduced nuclear genome (2.9 Mbp) of E. cuniculi led us to the identification of a relatively limited set of genes encoding ABC systems. Classification of E. cuniculi ABC proteins into families may help to elucidate their functions. With this primary purpose, it is necessary to re-evaluate the diversity of the potential E. cuniculi ABC proteins.
ABC proteins constitute one of the largest families of paralogous sequences, and they are involved in many fundamental biological processes. They share a highly conserved ATPase domain, the ABC, which has been demonstrated to bind and hydrolyze ATP, thereby providing energy for a large number of biological processes. The amino acid sequence of this cassette displays three major conserved motifs, the Walker motifs A and B commonly found in ATPases together with a specific signature LSGGQ motif also known as the linker peptide. Most ABC systems are transporters which are composed of four structural domains: two very hydrophobic membrane spanning domains (called thereafter IM domains) and two hydrophilic cytoplasmic domains containing the ABC (called thereafter ABC domains). ABC transporters are involved in the import or the export of a wide variety of substances. Importers constitute mainly the prokaryotic transporters dependent upon a substrate-binding protein whose function is to provide bacteria with essential nutrients. Usually, they carry IM and ABC domains on separate polypeptide chains. Exporters are found in both prokaryotes and eukaryotes and are involved in the extrusion of noxious substances and drugs, the secretion of extracellular toxins and the targeting of membrane components. Generally, their constitutive IM or ABC domains are fused in various combinations. ‘Half-size’ transporters are composed of a single IM domain fused to an ABC domain, a structural organization that could be symbolized as ‘IM–ABC’ or ‘ABC–IM’ depending on the N- or C-terminal location of the transmembrane domain. ‘Full-size’ transporters are generated by duplication and fusion of ‘half-size’ transporters. An increasing number of ABC proteins are apparently not involved in transport but rather in cellular processes such as DNA repair, translation or regulation of gene expression. These systems, symbolized as ABC2, do not have IM domains and are composed of two ABC domains fused together.
The present inventory, based on sequence similarities with ABC systems in other living organisms, indicates that E. cuniculi exhibits 13 putative genes corresponding to only four families of ABC systems and lacks representatives of the major full-size ABC transporters found in eukaryotes, such as the P-gp family involved in multidrug resistance. Rather surprisingly, 11 of these are homologous to ‘half-size’ ABC transporters belonging to two subfamilies and for which a role in multidrug resistance has not been clearly demonstrated.
Materials and methods
Identification and classification of ABC proteins
The open reading frames (ORFs) identified in the E. cuniculi genome were assigned to families and subfamilies of ABC systems in two steps. First, the total E. cuniculi ORF content was compared to the custom annotated sequences of ABCISSE (http://www.pasteur.fr/recherche/unites/pmtg/abc/database.html) by using the BLASTP program. This specific annotation mentions the family and the subfamily of about 2000 ABC proteins. The statistical significance of the similarity scores was assessed by using the E-value parameter of the BLAST package. When an E. cuniculi ORF was homologous (E-value <20) to at least 10 proteins of the same family, it was provisionally assigned to this family. Second, the E. cuniculi ORFs selected by the procedure described above were compared to the whole set of ABCISSE sequences of the different families by using the ClustalW program. The co-clustering of E. cuniculi sequences with members of specific subfamilies allows a definitive assignation to a given subfamily.
Phylogenetic analysis of ABC proteins
The sequences of ABC proteins identified in E. cuniculi were compared to homologous sequences in ABCISSE by using the ClustalW program. Phylogenetic trees were computed by the NJ method implemented in the ClustalW package and were viewed by the Macintosh application TreeView. The list of the HMT (mitochondrial transporters) subfamily proteins analyzed in this study could be retrieved from the ABCISSE database Internet link http://www.pasteur.fr/cgi-bin/pmtg/requete.plform html&cmd FS&attri_1 HMT. WHI (white transporters homologs) subfamily proteins could be retrieved at (http://www.pasteur.fr/cgi-bin/pmtg/requete.plform html&cmd FS&attri_1 WHITE).
Computer prediction of primary and secondary structure elements
Online resources were used to analyze the sequences of the 13 ABC systems found in the genome of E. cuniculi. The same type of analysis was also performed on experimentally characterized ABC proteins (ATM1_YEAST, ABC7_HUMAN, ABG2_HUMAN in the SwissProt database) and the two E. intestinalis ABC transporters EiABC1 and EiABC2 (protein ID in the GenPept database (http://www.ncbi.nlm.nih.gov: AAK96046 and AAK56394, respectively). Candida albicans protein sequences were extracted from the genome sequence by using ACeDB, a database-managing environment developed originally by J. Thierry-Mieg and R. Durbin (http://www.acedb.org). The genome of C. albicans was downloaded from the University of Stanford (ftp://cycle.stanford.edu/pub/projects/candida). The identification of conserved motifs was performed by using Pfam through its user interface at Sanger Institute (http://www.sanger.ac.uk/Pfam). The TMpred interface at EMBnet was used to identify transmembrane segments (http://www.ch.embnet.org/software/TMPRED_form.html). Identification of subcellular sorting signals was performed by using the programs Psort II (http://psort.nibb.ac.jp/form2.html), SignalP (http://www.cbs.dtu.dk/services/SignalP-2.0), TargetP (http://www.cbs.dtu.dk/services/TargetP), Mitoprot (http://mips.gsf.de/cgi-bin/proj/medgen/mitofilter) and Predotar (http://www.inra.fr/Internet/Produits/Predotar).
The analysis of the genome sequence of E. cuniculi revealed that 13 ORFs, including the two perfectly duplicated ORFs 01_0200 and 01_1410 carried by the two highly conserved subterminal regions of chromosome 1, could be assigned to the ABC superfamily. We noticed that representatives of the ABC superfamily are lacking on chromosomes 2, 6 and 9. As summarized in Table 1, we distinguished four clusters corresponding to four families or subfamilies of ABC systems. The first cluster (six ORFs) belongs to the HMT subfamily of the DPL (export of peptides, lipids and drugs; ABCB) family whereas the second cluster (five ORFs) is assigned to the WHI subfamily of the EPD (eye pigment and drugs; ABCG) family. The two remaining ORFs are assigned to the REG (regulation of translation; ABCF) subfamily of the ART (antibiotic resistance and translation regulation) family and to the RLI (RNase-L inhibitor; ABCE) family. The phylogenetic distribution of E. cuniculi ABC systems is shown in Fig. 1 and the data are summarized in Table 1. No representative of the multidrug-resistance full transporters P-gp (P-gp subfamily) and the conjugate-drug-resistance transporters (MRP subfamily), which are characterized by two tandemly repeated copies of each IM and ABC domains and symbolized as ‘(IM–ABC)2’, was identified in the E. cuniculi genome. In the following sections, we will describe the predicted properties of E. cuniculi putative ABC systems according to their clustering within ABC families.
|ABC systems||Topology||HCNG||Assigned Encephalitozoon ORFs|
|1||DPL||HMT||IM–ABC █•||ABCB||01_0200, 01_1410, 03_0240, 04_0480, 10_1230, 11_1200 EiABC1a, EiABC2a|
|1||EPD||WHI||ABC–IM •█||ABCG||03_0390, 07_0560, 08_0110, 10_1520, 11_1340|
|ABC systems||Topology||HCNG||Assigned Encephalitozoon ORFs|
|1||DPL||HMT||IM–ABC █•||ABCB||01_0200, 01_1410, 03_0240, 04_0480, 10_1230, 11_1200 EiABC1a, EiABC2a|
|1||EPD||WHI||ABC–IM •█||ABCG||03_0390, 07_0560, 08_0110, 10_1520, 11_1340|
Class 1 systems are known as exporters with fused transmembrane and ATPase domains. Involved in non-transport cellular processes, class 2 systems have no transmembrane domain. The identification number for each E. cuniculi ORF indicates the corresponding chromosome followed by the ORF position on the chromosome. Family and subfamily names are reminiscent to their predicted functions. Topology: Organization of structural domains from N- to C-terminus. IM (rectangle); ABC (circle). HCNG: Alternative nomenclature of ABC families adopted by the Human Gene Nomenclature Committee.
aPreviously identified E. intestinalis ORFs
E. cuniculi DPL (ABCB) family, HMT subfamily transporters
The six members assigned to the DPL family correspond to ‘half transporters’ having a structural organization symbolized as ‘IM–ABC’. It is expected that these proteins would homo- or heterodimerize to form a functional transporter. All E. cuniculi DPL family putative transporters belong to the HMT subfamily. The prototype of this family is the ATM1 protein of Saccharomyces cerevisiae, which is located in the mitochondrial inner membrane and is essential for the generation of cytosolic Fe/S proteins, by mediating export of Fe/S cluster precursors out of mitochondria. The second well-characterized member is the Schizosaccharomyces pombe HMT1 protein, involved in heavy metal tolerance by a vacuole-sequestration mechanism of phytochelatins also know as heavy metal-binding proteins. The phylogenetic tree computed from multiple alignments reveals that the E. cuniculi genes cluster independently of other systems of the family (Fig. 2). Two ABC transporter genes (EiABC1 and EiABC2) have been recently sequenced in the related microsporidian species E. intestinalis. These genes cluster with E. cuniculi HMT subfamily ORFs on phylogenetic trees indicating that they belong to the same subfamily (Fig. 1). EiABC1 is closely related and probably orthologous to the E. cuniculi ORF 03_0240 while EiABC2 is clearly divergent from other E. cuniculi ORFs (Fig. 2).
The organization of transmembrane regions (ABC_membrane Pfam motif) is quite diverse, as judged by TMpred predictions (Fig. 3). Putative mitochondrial pre-sequences were identified for ORFs 01_0200/01_1410, 03_0240 and 11_1200. The hydrophobic portion of the C-terminal part of the putatively excised peptide was found to include the first transmembrane segment. In particular, ORF 11_1200 appears to be the closest homolog of the S. cerevisiae and C. albicans ATM1 proteins with respect to secondary structure predictions (Fig. 3). By contrast, ORF 10_1230 displays a putative signal peptide rather than a mitochondrial pre-sequence, suggesting that it might be secreted via the endoplasmic reticulum–Golgi pathway.
E. cuniculi EPD family (ABCG), WHI subfamily transporters
The second cluster of sequences (03_0390, 07_0560, 08_0110, 10_1520 and 11_1340) was assigned to the WHI subfamily (Table 1). WHI subfamily systems are also half transporters but with respect to HMT subfamily transporters, their topological disposition is inverted with an N-terminal ATPase domain fused to the C-terminal transmembrane domain leading to a structural organization symbolized as ‘ABC–IM’. The most well-characterized members of this subfamily are the white, brown and scarlet gene products of Drosophila melanogaster involved in the transport of eye pigment precursors. It is thought that the white and brown proteins form a heterodimeric complex involved in guanine transport, while the white and scarlet proteins form a tryptophan transporter. The phylogenetic tree computed from multiple alignments reveals that these genes cluster independently of other systems of the family (Fig. 4). No ORF in the E. cuniculi genome was found to correspond to the related fungal PDR subfamily involved in pleiotropic multiple drug resistance, which is composed of proteins containing two tandemly repeated copies of each ABC and IM domains and whose structural organization is symbolized as ‘(ABC–IM)2’.
Two different types of organization in transmembrane domains can be distinguished among E. cuniculi WHI subfamily ORFs, which are consistent with the dendrogram derived from the phylogenetic analysis of their sequences (Fig. 5A). The first is represented by ORFs 03_0390 and 07_0560. The transmembrane domain of 03_0390 contains the B44 and B775 PfamB motifs which are specific to fungal WHI subfamily proteins, particularly the ADP1 proteins of S. cerevisiae and C. albicans (Fig. 5B). However, these yeast proteins have a long N-terminal extension (half of the size of the protein) beginning with a signal peptide, which is lacking in other WHI subfamily proteins. In fact, ORF 03_0390 is more closely related to the human homolog ABG2. Pairwise comparisons also confirmed that E. cuniculi ORFs are more similar to human and worm EPD family transporters than to fungal ADP1 proteins. ORF 07_0560 displays only the B775 PfamB motif. The second type of organization is represented by ORFs 08_0110, 10_1520 and 11_1340 that display none of the Pfam motifs mentioned above. These ORFs have a larger C-terminal region including an additional transmembrane segment. A mitochondrial pre-sequence was predicted for ORF 08_0110.
The E. cuniculi ORFs putatively involved in cellular processes other than transport
The two remaining ORFs (05_1190 and 11_1580) do not contain transmembrane domains and are comprised of two ABC domains fused together. Experimentally investigated homologs of these proteins are not involved in transport. ORF 11_1580 was assigned to the RLI (ABCE) family of proteins which contain a ferredoxin 4 Fe/4S motif and whose most well-characterized member is the human RNase L inhibitor (ABCE1). The 05_1190 protein was assigned to the REG (ABCF) subfamily of the ART family, the prototype of which is the yeast protein GCN20. The latter was shown to interact with the protein GCN1 to stimulate the activity of a kinase (GCN2) that phosphorylates the eukaryotic translation initiation factor eIF2. This leads to an increased translation of the transcriptional activator GCN4 in amino acid-starved cells.
One of the most remarkable features of this inventory of the microsporidian E. cuniculi ABC systems is that all transporters fall into two discrete subfamilies of ABC systems, the HMT and WHI subfamilies. Only 13 ORFs in this genome were identified as ABC systems, the smallest number reported in eukaryotes. That six ORFs are attributed to the HMT subfamily including principally mitochondrial transporters is intriguing because there is no morphological evidence for mitochondria in E. cuniculi and in other microsporidia. However, a mitochondrial-type HSP70 has been identified in three microsporidian species, suggesting that they originate from a mitochondrion-bearing ancestor [24,,26]. Thus, microsporidia have probably lost their ATP-producing organelle by regressive evolution. An alternative hypothesis consisting in the preservation of a simplified organelle, the ‘mitosome’, has been recently proposed on the basis that several E. cuniculi ORFs display significant similarities with known mitochondrial proteins, especially those involved in the generation of Fe–S clusters. The product of ORF 11_1200 appears to be the best candidate to function in the export of these clusters from a cryptic mitosome towards the cytosol, as judged by its high similarity to the yeast ATM1 protein and a very significant probability for the presence of an N-terminal extension resembling a mitochondrial-type pre-sequence. Indeed, a putative Fe/S-binding site was identified in the RNAse-L inhibitor homolog 11_1580.
The fact that six E. cuniculi ORFs were homologous to proteins of the HMT subfamily does not mean that they perform exactly the same function. At least four human genes were shown to belong to this family. Two of these, ABCB6 and ABCB7, are probably performing the same function as yeast ATM1 protein in Fe–S cluster export since they complement a yeast strain mutated in the gene encoding this protein. ABCB7 is associated with the X-linked inherited disease sideroblastic anemia and ataxia. The function of the remaining two genes is presently unknown. Direct experimental investigation is therefore needed to establish the physiological function of HMT subfamily transporters in microsporidia. Recently, two ABC transporters (EiABC1 and EiABC2) were identified in E. intestinalis. EiABC1 was considered as an ATM1 homolog whereas EiABC2 was claimed to be closer to P-gp. We found that EiABC1 and EiABC2 are half-size transporters that belong to the HMT subfamily. EiABC1 is a possible ortholog of E. cuniculi ORF 03_0240. EiABC2 seems quite different from all E. cuniculi ABC proteins which is suggestive of a rather rapid intrageneric evolution. It might be a specific type of ABC transporter not represented in all microsporidian species and more specifically in E. cuniculi.
No orthology relationship has been found for the EPD family by comparing the E. cuniculi ABC ORFs to those of S. cerevisiae (by using BLASTP) and to the C. albicans genome DNA sequence (by using TBLASTN). Among the proteins assigned to the WHI subfamily, the 03_0390 sequence is the best representative, sharing homology to the C-terminal region of yeast ADP1 and the N-terminal region of human ABCG2. There are indications that some WHI subfamily systems are involved in the regulation of lipid trafficking and/or in drug resistance. The mammalian white gene homolog ABCG1 (ABC7) is indeed highly induced in lipid-loaded macrophages, suggesting a role in cholesterol and phospholipid trafficking, reviewed in. Recently, it was found that phytositosterolemia (elevation of plasma levels of plant sterols due to enhanced intestinal absorption and reduced removal) was caused by mutations in the human ABCG5 and ABCG8 transporters, reviewed in. Moreover, ABCG2 (also called MXR or BCRP) was shown to confer anthracyclin drug resistance when overexpressed in certain breast cancer cell lines, reviewed in. Thus, it could be speculated that at least ORF 03_0390 plays a role in a drug-resistance phenomenon.
The RLI system could play a role in the regulation of RNA stability in mammalian cells. Inhibiting protein synthesis by cleaving mRNAs, RNase L is activated by adenylate oligomers synthesized by the 2–5A synthase, an enzyme induced by interferons. Its activity is regulated by RLI through inhibition of 2–5A binding. Thus, ORF 11_1580 might be involved in the regulation of transcription. The need of a GCN20 homolog (ORF 05_1190) for microsporidian translation regulation through a possible interaction with elongating ribosomes, as in other eukaryotes and in Plasmodium falciparum, may indeed be considered.
In contrast with other parasitic microorganisms, the genome of E. cuniculi does not contain sequences homologous to full-size ABC transporters involved in multidrug resistance in other eukaryotic pathogens. The lack of such sequences is also supported by the analysis of conserved motifs (data not shown). However, immunofluorescence images of Encephalitozoon-infected mammalian cells obtained after reaction with anti-P-gp antibodies showed significant labeling of merogonial parasite stages adjacent to the parasitophorous vacuole, where Encephalitozoon cells are known to proliferate. The question whether that labeling was due to host or parasite P-gp is still an open question. Whether an E. cuniculi ORF might account for a protein capable of reacting with anti-mammalian P-gp antibodies remains to be elucidated. No ORF similar to the MRP subfamily, whose members are involved in drug-conjugate export are found in E. cuniculi. This is in agreement with the observation that monoclonal antibodies against MRP do not react with E. intestinalis-infected cells.
Among the future lines of research in E. cuniculi ABC systems, it will be important to ascertain that all corresponding genes are expressed and, if this is the case, to determine the subcellular localization of the different transporters. Whether predicted mitochondrial pre-sequences represent signals for targeting HMT transporters to a mitochondrion derived-compartment, or not, may deserve specific investigations for testing their functionality in a convenient mitochondrion-bearing eukaryotic model system. Alternatively, electron microscopy and immunocytochemistry with antibodies raised against specific epitopes of potential E. cuniculi ABC transporters might help to solve this issue. The role of ABC transporters in the development of resistance to albendazole would be determined.
- atp-binding cassette transporters
- drug resistance, multiple
- encephalitozoon cuniculi
- membrane transport proteins
- mitochondrial proteins
- open reading frames
- ribonuclease, pancreatic
- trees (plant)
- play behavior