The interferon-inducible p47 (IRG) GTPases in vertebrates: loss of the cell autonomous resistance mechanism in the human lineage
© Bekpen et al.; licensee BioMed Central Ltd. 2005
Received: 4 June 2005
Accepted: 7 October 2005
Published: 31 October 2005
Members of the p47 (immunity-related GTPases (IRG) family) GTPases are essential, interferon-inducible resistance factors in mice that are active against a broad spectrum of important intracellular pathogens. Surprisingly, there are no reports of p47 function in humans.
Here we show that the p47 GTPases are represented by 23 genes in the mouse, whereas humans have only a single full-length p47 GTPase and an expressed, truncated presumed pseudo-gene. The human full-length gene is orthologous to an isolated mouse p47 GTPase that carries no interferon-inducible elements in the promoter of either species and is expressed constitutively in the mature testis of both species. Thus, there is no evidence for a p47 GTPase-based resistance system in humans. Dogs have several interferon-inducible p47s, and so the primate lineage that led to humans appears to have lost an ancient function. Multiple p47 GTPases are also present in the zebrafish, but there is only a tandem p47 gene pair in pufferfish.
Mice and humans must deploy their immune resources against vacuolar pathogens in radically different ways. This carries significant implications for the use of the mouse as a model of human infectious disease. The absence of the p47 resistance system in humans suggests that possession of this resistance system carries significant costs that, in the primate lineage that led to humans, are not outweighed by the benefits. The origin of the vertebrate p47 system is obscure.
It is generally assumed that the immune system of the mouse is a good experimental model for that in humans. However, several studies suggest that immune mechanisms have been evolving rather differently in the human and mouse lineages (for review, see Mestas and Hughes ). The p47 (immunity-related GTPases (IRG) family; see Nomenclature, below) GTPases present a uniquely striking example of this divergence.
In mice the interferon-γ-inducible p47 GTPases constitute one of the most powerful resistance systems against several important intracellular pathogens [2–4]. The proteins localize on intracellular membrane systems in interferon-induced cells, some (IGTP, IIGP1) favoring the endoplasmic reticulum [5, 6] and others (LRG-47, GTPI) the Golgi membranes [6, 7] (for names of individual IRG GTPases see Additional data file 1). Infection or phagocytosis, however, initiates redistribution of the p47 GTPases to the phagocytic vacuole [6–8]. The p47 GTPases probably act specifically against vacuolar pathogens. Thus, Gram-positive and Gram-negative bacteria, mycobacteria, and protozoal pathogens are all resisted by the p47 GTPases, whereas no viral target has yet been confirmed.
The p47 GTPase IIGP1 is a low-affinity nucleotide binding protein with a slow GTP turnover . At high protein concentrations and in the presence of GTP, IIGP1 oligomerizes and increases GTP turnover by up to 20-fold. These properties are distinct from those of the classical signaling GTPases and are reminiscent of the dynamins and p65 (GBP-1) GTPases [10, 11]. The crystal structure of IIGP1 exhibits a H-Ras-1-like nucleotide-binding domain flanked by amino-terminal and carboxyl-terminal helical domains that are unknown in other GTPases . This basic structure is probably common to the whole family. However, the divergent sequences of published p47 GTPases  and the patterns of susceptibility in knockout strains (for reviews, see Taylor  and MacMicking [3, 4]) show that the proteins are highly diversified. Thus, a subgroup of three proteins (the GMS GTPases) have a radical substitution (the substitution of Methinine (M) for Lysine (K)) in the conserved P-loop G1 motif of the nucleotide binding site (Walker A motif) and correlated sequence variation elsewhere in the G-domain , implying a distinct catalytic mechanism for GTP hydrolysis. In the case of IIGP1 and LRG-47, the cell biology of the two proteins is distinct; IIGP1 associates with the endoplasmic reticulum membrane primarily through an amino-terminal myristoylation sequence, whereas LRG-47 associates with Golgi membrane via an amphipathic helix in the subterminal domain . We recently showed that IIGP1 participates in a novel effector mechanism in Toxoplasma gondii infected astrocytes involving vesiculation and ultimately destruction of the parasitophorous vacuole membrane . In contrast, there is evidence that LRG-47 is involved in accelerated acidification of the phagocytic vacuole containing Mycobacterium tuberculosis .
The p47 GTPases are thus a functionally diverse resistance system with many signs of adaptive divergent evolution. Surprisingly, there are no reports of p47 GTPase function in humans. To address this imbalance, we analyzed the p47 GTPase gene family in depth. We conclude that although the mouse has 23 p47 GTPases, of which up to 20 may be functional in resistance, the resistance system is entirely absent from humans. This finding carries important implications for our understanding of human and mouse immunity to vacuolar pathogens.
Genomic organization of the p47 GTPase (Irg) genes of the C57BL/6 mouse
The complex block of 13 genes on chromosome 11 contains the most divergent sequences (Figure 1c,d; Additional data file 2), including all three GMS (Irgm) GTPases , suggesting that this cluster is relatively ancient. In contrast, the eight Irga genes clustered on chromosome 18 are also clustered phylogenetically, suggesting more recent divergence, probably from a translocated member of the Irgb (TGTP) cluster on chromosome 11. The isolated Irg gene on chromosome 7, Irgc, is an ancient root with no obvious systematic relationship to the other subfamilies. Within the chromosomal clusters, more recent duplication events are apparent. The sibling pair Irgb3 and Irgb4 differ by only nine nucleotides in the open reading frame. The genes Irgb1, Irgb3, Irgb4, and Irgb8 appear to have been duplicated in tandem with Irgb2, Irgb5, and Irgb9, respectively. The pattern of divergence in the mouse p47 tree suggests an old gene family that has undergone a succession of duplication-divergence cycles over time - a pattern of evolution that is still actively continuing in several of the subfamilies.
The structure of p47 GTPase genes and their splicing patterns
Identification of interferon-stimulatable elements in putative promoters of Irggenes
The isolated p47 gene, Irgc, on chromosome 7 is a clear exception. No clustered or isolated ISRE or GAS elements could be identified up to 10 kilobases (kb) 5' of the putative transcription start of this transcribed gene, and Irgc was not induced in interferon-stimulated fibroblasts (Figure 3b, panel i left). A weak Sox-related element was detected in the proximal promoter region. In view of the close homology of Irgc to the interferon-inducible Irg genes, we considered whether Irgc is induced in tissues of mice 24 hours after infection with Listeria monocytogenes [13, 16]. No induction of Irgc was detected in liver, lung, or spleen after 50 cycles of amplification, whereas Irga2, used as a positive control, was induced in all three tissues (Figure 3b; panel i right). However, Irgc, unlike Irga2, was constitutively expressed in the mature mouse testis (Figure 3b; unpublished data). We conclude that mouse Irgc is expressed in a tissue-specific manner and is not induced by infection.
The coding sequences of the p47 GTPases
The p47 GTPase genes of the human genome
At the protein level the shortest isoform of IRGM is shorter than a canonical G-domain, being truncated in the middle of β-strand 6 just before the G5 sequence motif, which interacts with the guanine base of the bound nucleotide (Figures 6 and 8b; also see Ghosh and coworkers ). The longer isoforms are terminated by short sequence extensions that are unrelated to known GTPase domains. A rabbit antiserum raised against recombinant human IRGM produced in Escherichia coli failed to detect signal by immunofluorescence or Western blot in human cell lines (data not shown).
IRGgenes of the dog
Is the mouse (order Rodentia) or the human (order Primata) the exception? We looked for IRG genes in a third order of mammals, the Carnivora. We recovered a total of eight IRG genes from the public genome database of the dog Canis familiaris (Figures 5 and 6) as well as a partial sequence of a 9th gene (not shown). Of these, one (not shown) is a pseudo-gene by a number of criteria, another is clearly dog IRGC, whereas the partial sequence is novel but most closely related to IRGC. The remainder assort into segments of the phylogeny already established for the interferon-inducible mouse IRG genes (Figure 5). Both GMS and GKS genes are represented and are inducible by interferon in dog MDCK epithelial cells (Additional data file 4). The three dog GMS genes appear to have diversified independently from the mouse GMS genes (Figure 5). As in humans and mouse, dog IRGC was not induced by interferon-γ (Additional data file 4). Overall, the IRG gene status of the dog clearly resembles that of mouse rather than that of humans.
IRGgenes in fish genomes
IRG GTPases are at least as old as the vertebrates. We have identified at least two distinct irg genes in the freshwater pufferfish Tetraodon nigriviridis, a closely linked pair of irg genes in the saltwater pufferfish Fugu rubripes, and at least 11 partially clustered irg genes in the zebrafish Danio rerio (Figures 5 and 6, and Additional data file 5). The fish irg genes fall into separate clades from the mammalian genes (Figure 5). A specific IRGC homolog is not immediately apparent. GMS subfamily IRGM genes are absent from fish. The pufferfish and zebrafish irgf genes have one intron identically positioned at the end of helix 4 of the G-domain (indicated on Figure 6; also see Additional data file 5). This intron is 81 bp long in both pufferfish species but is substantially longer in the zebrafish genes. The distinct irge subfamily of the Danio irg genes are intronless in the open reading frame, like mammalian IRG genes.
IRGhomologs with divergent nucleotide-binding regions: the quasi-GTPases
The mouse, human and zebrafish genomes encode proteins that are homologous to the IRG GTPases but are radically modified in the GTP-binding site. The mammalian protein FKSG27 (IRGQ), a protein of unknown function that is 70% conserved between man and mouse, is extended amino-terminally relative to a p47 GTPase by about 100 residues encoded on three short exons. The remaining 420 residues, encoded on a single long exon, are clearly homologous to and colinear with the IRG proteins (Figure 6 and Additional data file 6), especially in the amino- and carboxyl-terminal parts of the exon. The region of lowest similarity is in the G-domain, and conserved GTP-binding motifs are lacking (Figure 6, and Additional data files 6 and 7). Thus, FKSG27 (IRGQ) is not a GTPase despite its phylogenetic relationship to the IRG proteins. FKSG27 (IRGQ) is closely linked to IRGC in humans and mouse (Figure 7a).
The zebrafish genome contains three IRG homologs with more or less modified GTP-binding motifs (irgq1-irgq3; Figures 5 and 6, and Additional data file 7). Their homology to IRG genes is stronger than that of FKSG27 (IRGQ), but as with FKSG27 (IRGQ) their function as GTPases is doubtful. The irgq1 gene is clustered on a single BAC clone with four apparently normal irge genes and immediately downstream of a truncated p47 gene, irgg, with which irgq1 is transcribed as the carboxyl-terminal half of a tandem transcript. Thus, the hypothetical protein product would be a carboxyl-terminally truncated p47 GTPase, linked at its carboxyl-terminus to a similarly truncated p47 homolog probably without GTPase function.
We propose to term the modified IRG proteins without GTPase function 'quasi IRG' proteins, hence IRGQ. IRGQ sequences reveal their phylogenetic relationship to the IRG proteins, but they are nevertheless more or less radically modified, primarily in the nucleotide binding site. In view of the substantial divergence between the IRGQ genes and functional p47 GTPases, it was unexpected not to find close homologs of the Danio irgq sequences in either the Fugu or Tetraodon genomes. The evolution and diversity of the Danio irgq genes is apparently linked to the evolution and diversity of the GTPase-competent IRG sequences.
IRG homologs outside the vertebrates
No unambiguous IRG homologs have been found outside the vertebrates. However, two possibly related sequence were recovered from the Caenorhabditis elegans genome, and several groups of putative GTPases of unknown function exist in the bacteria that have sequence features reminiscent of IRG GTPases. Perhaps the most striking of these are found in the Cyanobacteria (see Additional data file 1 for accession numbers for these sequences). Among other features, all of these sequences have in common with the IRG GTPases the presence of a large hydrophobic residue in place of the familiar catalytic Q61 of H-Ras-1, but this feature is far from diagnostic for the IRG GTPases . Despite several suggestive characteristics of these invertebrate and bacterial GTPase sequences, it is not possible on the basis of sequence criteria alone to establish their phylogenetic relationship with vertebrate IRG proteins.
The p47 GTPases (IRG proteins) are an essential resistance system in the mouse for immunity against pathogens that enter the cell via a vacuole. In this study we reached several unexpected conclusions about the evolution of the system. First, the IRG resistance system, despite its importance for the mouse, is absent from humans because it has been lost during the divergent evolution of the primates. Second, the IRG resistance system is at least as old as the bony fish but missing in the invertebrates. Finally, the IRG proteins appear to be accompanied phylogenetically by homologous proteins, here named IRGQ proteins, that probably lack nucleotide binding or hydrolysis function, and that may form regulatory heterodimers with functional IRG proteins. We consider these points in order.
The argument for the absence of the IRG resistance system in humans relies on several findings. The system is reduced from 23 genes in mouse to one full-length gene and a transcribed G-domain in humans, and the residual genes lack the character of functional resistance genes. Thus, IRGC is highly conserved in humans, dog and mouse, is not interferon or infection inducible, and is expressed constitively in mature testis. IRGM, although clearly derived from a typical GMS subfamily resistance gene, is transcribed constitutively from an endogenous retroviral LTR, is unresponsive to interferon, and appears to be structurally damaged in several ways.
We argue that the IRG resistance system has been lost from primates (the situation in chimpanzee is identical to that in humans; unpublished data) rather than gained by the murine rodents (including rat; unpublished data) on the following grounds. First, like the mouse, the dog genome has several complete, interferon-inducible IRG genes in addition to IRGC. Second, humans and chimpanzees possess a degraded member of the GMS subfamily of IRG proteins, confirming that this distinctive subfamily, present and functional as resistance genes in dog and mouse, was widely distributed at the origin of the mammalian radiation. Finally, the IRG system is present in bony fish, representing ancient vertebrates. Rapid expansion and contraction of multigene families associated with pathogen resistance has frequently been documented in both animals and plants [23–28]. In all of these cases, however, the resistance mechanism itself has been retained as its protein mediators have evolved or even, in the natural killer receptor case, been replaced by a different molecular species . The IRG case may be different because here the resistance mechanism itself has apparently been lost during primate evolution.
It will be of interest and of considerable importance to analyze the different strategies by which humans, dog and mouse deploy resistance mechanisms effective against vacuolar pathogens. None of the known mechanisms active in humans against vacuolar pathogens, namely nitric oxide and oxygen radicals [30–32], tryptophan depletion [33, 34], accelerated acidification by Rab5a , cation depletion [36–38] or autophagy [39, 40], is missing from the mouse. Nevertheless, it remains possible that the distinctive resistance actions of the p47 GTPases [7, 8] are performed by an unrelated and thus far unidentified molecular machine in primates.
The loss of a highly evolved and complex resistance system that is active against vacuolar pathogens needs an adaptive explanation. The evolution of a successful avoidance strategy by the pathogens is unlikely, because many different pathogens and pathogen classes are controlled by IRG proteins in the mouse [2, 4]. Nevertheless very recent evidence suggests that Chlamydia spp. divergence between humans and mouse may indeed be partially driven by differences in the deployment of p47 GTPases, in this case Irga6 (IIGP1) . Human Chlamydia trachomatis is controlled by IIGP1 in interferon-treated mouse oviduct epithelial cells, whereas control of the extremely closely related mouse C. muridarum is independent of interferon. However, a more plausible general model is the evolution in the primate lineage of improvements to the battery of parallel mechanisms, rendering the IRG system redundant. Interestingly, in this context it was noted in the Chlamydia study quoted above that interferon-stimulated mouse oviduct epithelial cells did not express the important resistance factor indoleamine deoxygenase, which is responsible for tryptophan depletion, whereas this is well expressed in interferon-stimulated human HeLa cells . In general, pathogen resistance mechanisms also carry costs, for example autoimmunity and allergy, arising from the adaptive immune system and many others arising from innate immunity [42–46]. Indeed, the interferon-inducible dynamin-like GTPases, the Mx proteins, which confer on mice strong resistance to certain RNA viruses and have been considered functionally related to the p47 GTPases [4, 47–49], exist in a balanced polymorphism in the wild over null alleles [50, 51], and have been lost spontaneously from all except two laboratory mouse strains . It is not yet obvious what specific costs might be associated with possession of the Mx or IRG resistance systems.
The IRG proteins are well represented in the bony fish but, although they are abundant and diverse in the zebrafish, in Fugu there are only two very closely linked and similar genes that are, indeed, annotated as a single tandem gene in ENSEMBL (although we judge this not to be the case). Thus, the available annotated fish genomes seem to mirror the IRG situation in the mammals, with Fugu and Tetraodon reflecting the reduced human case and Danio the complex dog and mouse. However, it has not yet been reported whether any fish IRG genes respond to infection . Both Danio and the pufferfish are Actinopterygian fish, in which where there is increasing agreement that the genome has been amplified by three rounds of duplication [54, 55]. It is plausible that the complex IRG representation in Danio may be attributed to preservation of these genes on more than one of the potential eight paralogons, whereas only a single copy carries IRG genes in the pufferfish. Further clarification of this issue awaits the completion of the genomes.
The phylogenetic origin of the IRG proteins is obscure. Because the family is conserved at least down to the bony fishes with little structural modification, the IRG genes are not strictly fast-evolving and their basic conservatism makes them easy to identify. Thus, their apparent absence from most known invertebrate genomes is probably real. Although many components of the adaptive immune system appear to have evolved close to the chordate-vertebrate boundary , this is not generally the case for innate immune mechanisms [57, 58] There seems to be no reason in principle why the IRG system should not work in invertebrates because it is cell-autonomous. However, the putative GTPase sequences that we have recovered from C. elegans and from Cyanobacteria are too distantly related outside the G-domain for a clear phylogenetic relationship to IRG proteins to be established from sequence similarity alone, whereas the similarities within the G-domains, although occasionally striking, are to some extent forced by the maintenance of a highly conserved function, namely regulated GTP hydrolysis. A stronger case for a meaningful phylogenetic relationship between these proteins and the vertebrate IRG proteins would follow from structural evidence that they display the distinctive IRG fold exemplified by mouse Irga6 (IIGP1) and from a detailed analysis of their catalytic mechanism.
The basic unit of IRG protein function may be a dimer because several genes we have identified occur in pairs in a head-to-tail arrangement, are expressed as tandem transcripts, and are presumably expressed as dimeric proteins. This conclusion is consistent with the dimer of IIGP1 (Irga6) observed in the crystal, shown by site-directed mutagenesis of the dimer interface to be essential for GTP-dependent oligomerization and cooperative hydrolysis . However, a second dimer interface is also required for oligomerization (unpublished data), and which of the two dimer structures the constitutive IRG dimers represent is of considerable interest. Unlike the observed homodimer of IIGP1, the products of the two putative tandem genes in the mouse Irgb2/b1 and Irgb5/b4 are heterodimers, implying that the two IRG subunits serve distinct functions in the protein. The same may be true for the tandem pair of irg genes of Fugu, which are annotated as a single tandem gene in ENSEMBL. These latter genes have diverged very recently, because with three exceptions they are identical at the nucleotide level over the first 290 amino acids. However, they have diverged substantially in the carboxyl-terminal region (Figure 6), suggesting a recent selective force. If the two tandem IRG domains are indeed expressed as a tandem protein (as favored by the ENSEMBL annotation), then it will be as a heterodimer with significant sequence variation at the carboxyl-terminus. The extreme case of heterodimer differentiation in IRG tandem genes may occur in Danio, in which gene irgg, a canonical (although truncated) IRG gene, is apparently expressed in tandem with the adjacent downstream gene irgq1, which is a modified (and also truncated) quasi-GTPase gene that is unlikely to function as a GTPase (Additional data file 7). In this case the role of the irgq1 domain may be regulatory for the amino-terminal irgg domain. Other IRGQ proteins may also be regulators of IRG proteins, interacting with the functional IRG proteins with a symmetry resembling one or other of the two dimer structures. Thus IRGQ proteins would coevolve with IRG proteins. This would explain why the three irgq genes of the zebrafish have no homologs in the pufferfishes, with their single tandem pair of irg proteins, and recalls the recent observation that the GAP protein of the small GTPase, Rap1, is itself probably derived from a GTPase ancestor, retaining the G-domain structure but not the sequence to reveal its origin .
Better understanding of the mechanism of action and regulation of the p47 GTPases is needed before their complex evolutionary history can be put in context. From the evidence we present here, however, it is already clear that effective resistance to vacuolar pathogens in humans and mouse must be organized on radically different principles.
We introduce here a general nomenclature on phylogenetic principles for the p47 GTPases, based on the stem name IRG (immunity-related GTPases). This stem was favored over other possibilities because the name IRG-47 has priority in the literature as the first description of a p47 GTPase . The phylogenetic basis of the nomenclature is apparent from Figure 5; each deep monophyletic clade is identified by a single-letter suffix as IRGM or IRGC. The nomenclature proposed here in Figure 5 and in Additional data file 1 has been accepted by the gene nomenclature committees of human and mouse, and by the zebrafish sequencing project. We have tried throughout to use the different forms of gene name accepted by the nomenclature committees for mouse (Irg) human (IRG), dog (IRG) and zebrafish (irg). The nomenclature of the IRGQ genes departs from the phylogenetic principle. The IRGQ nomenclature simultaneously recognizes the affinity of these sequences to IRG genes and stresses their anomalous GTP-binding domain features. It is, however, highly unlikely that the IRGQ sequences of humans, mouse, and fish represent a monophyletic group. It is more likely that the IRGQ genes of each taxonomic group derive from IRG genes of other clades of that group. This pattern is already apparent by inspection of the irgq protein sequences of the zebrafish in Figure 6, but it cannot be discerned from the G-domain-based phylogeny shown in Figure 5 because of the specific divergence of the G-domains of the irgq proteins.
Materials and methods
Use of database resources
All available public databases were extensively screened by BLAST and related searches for sequences belonging to the IRG family. In the case of the mouse, transcript sequences derived from the C57BL/6 strain were given preference over sequences of other and undefined strain origin, and compared in all cases with genomic sequence available via the ENSEMBL (v28.33d.1, February 2005) array of websites . A systematic study of polymorphism has not yet been completed, but it is already clear that nearly all IRG sequences derived from the CZECHII cDNA libraries (Mus musculus musculus) differ from C57BL/6 sequences. These differences make allocation of many CZECHII sequences to individual clade members of the C57BL/6 mouse problematic. Identification of certain Irg sequences with recognized gene symbols was achieved through the Mouse Genome Initiative web resources . Where ambiguities persist in the mouse genomic map, especially on chromosome 18 in the region of IrgA6-IrgA8 (Mb60.878-60.958), and on chromosome 11 in the region from PA28βψ to IrgB7ψ (Mb57.570-57.700), we used primary BAC and cosmid sequences to reach a consensus view.
Human and dog IRG sequences were identified from the available public databases (ENSEMBL, National Center for Biotechnology Information) and confirmed wherever possible by multiple sequence comparisons at transcriptional and genomic levels. Fugu material was obtained and analyzed through [63–65]. Tetraodon sequence was initially assembled from the GSS sequence database at National Center for Biotechnology Information and subsequently from the University of California at Santa Cruz compiled genome database  via the BLAST server. Zebrafish sequence was obtained from zebrafish genome resources at the Sanger Centre  and analyzed in an Acedb database using the Spandit annotation tool.
Chromosomal locations and synteny analysis of mouse and human chromosomes was initiated through ENSEMBL . Further details were obtained through the Sanger Centre . Nucleotide sequences and translated open reading frames of IRG family members used in this paper are given in Additional data files 9 and 10, and can also be accessed at the IRG family database at our laboratory .
Phylogeny and alignment protocols
Routine sequence analysis and local sequence database management was handled using DNA-Strider 1.3f12, Vector-Nti, and MacVector 7.2. The identity and similarity matrix on protein and nucleotide sequences (Additional data file 2) are based on GeneDoc (version number 2.6.002). Phylogenetic analysis was conducted using the neighbor-joining method , as implemented in the MEGA2 program . We used p-distances for constructing the phylogenetic trees. Reliability of the neighbor-joining trees was examined using the bootstrap test .
Alignments were performed via the BCM multiple alignment programme suite  and EBI-ClustalW  using the default options and manipulated according to the crystal structure of IIGP1 . Shading of alignments was performed using Boxshade  and additional sequences were shaded manually according to the default options of Boxshade.
Identification of transcription factor binding sites
Promoter regions (2 kb upstream of putative transcription start point) were screened for putative transcription factor binding sites with the Transcription Element Search System [76, 77], and the results were further analysed and confirmed manually. Additional promoter analysis of Irgc (mouse Cinema) and IRGC (human CINEMA) was performed with ConSite  based on phylogenetic footprinting .
RT-PCR on cells and tissues
C57BL/6J mice were obtained from the animal house at the Institute for Genetics, University of Cologne. Listeria monocytogenes infection was performed as described previously . Twenty-four hours after infection, the mice were killed, and liver, lung and spleen were removed and snap frozen in liquid nitrogen. Mouse L929 fibroblasts were stimulated for 24 hours with 200 U/ml interferon-γ or 200 U/ml interferon-β (R&D System GmBH, Weisbaden-Nordenstadt, Germany and Calbiochem-Novabiochem Corparation La Jolla, CA, respectively). Human cell lines (Hela, GS293 (GeneSwitch™ -293, Invitrogen GmbH Karlsruhe, Germany), HepG2, T2, THP1, MCF-7, SW-480, Primary foreskin fibroblast-HS27) were stimulated for 24 hours with 2,000 U/ml interferon-β or 200 U/ml interferon-γ (PBL Biomedical Laboratories, NY, USA and Peprotech/Cell concepts GmbH UmKirsh, Germany, respectively). Total RNA was extracted from tissues and cells using the RNAeasy mini kit (QIAGEN, Hilden, Germany), except for testis, for which the RNAeasy Lipid Tissue Kit (QIAGEN) was used. Poly(A) RNA was isolated from total RNA using the Oligotex mRNA kit (QIAGEN). Total RNA from human tissues was purchased from Biochain (Hayward, CA, USA). cDNA was generated from mRNA and total RNA using the Super Script First-Strand Synthesis System for RT-PCR (Invitrogen, Carlsbad, CA, USA). The generated cDNAs were screened for the presence of p47 (IRG) GTPase transcripts by PCR. A list of the primers used is given in Additional data file 8. The amplified fragments were confirmed by sequencing.
Additional data files
The following additional data are included with the online version of this paper: A list of all IRG gene family members described in this paper (gives names, synonyms, accession numbers and further information for each IRG gene; Additional data file 1); nucleotide and amino acid identities based on G-domain of mouse Irg family (gives percentage of identity on both protein and nucleotide level within the mouse Irg family; Additional data file 2); ISRE and GAS elements of mouse IRG family genes (contains the positions and exact sequences of all ISRE and GAS elements found in putative promoters of mouse IRG genes; Additional data file 3); inducibility of Dog p47 (IRG) GTPases (shows interferon inducibility of members of the p47 (IRG) GTPases present in the dog; Additional data file 4); genomic organization of Danio rerio p47 (irg) GTPases (illustrates the genomic organization of all p47 (irg) GTPases found in zebrafish to date; Additional data file 5); protein similarity matrix of Irgc and Irgq (contains comparison between the mouse p47 GTPase Irgc and the long coding exon of the closely linked quasi-GTPase Irgq (FKSG27); Additional data file 6); divergent nucleotide-binding motifs in quasi-GTPases (compares the nucleotide binding motifs of quasi-GTPases to those of the classical mouse p47 GTPases; Additional data file 7); a list of the primers used (contains the sequences of all primers used in this study; Additional data file 8); nucleotide sequences of all IRG family members (Additional data file 9); protein sequences of all IRG family members (Additional data file 10).
We are greatly indebted to Lois Maltais of the Mouse Genome Database at The Jackson Laboratory; Ruth Lovering, Gene Nomenclature Advisor, HUGO Gene Nomenclature Committee (HGNC); and Yvonne Edwards of the Fugu Genomics Project at the UK Human Genome Mapping Project (HGMP) Resource Centre for their time and effort in developing a useful nomenclature for the p47 GTPases. We are grateful to Kerstin Jekosch, Informatics & Systems Groups, Sanger Centre, for help with analyzing and annotating the zebrafish genes; to Cornelia Stein, Institute for Genetics, Cologne for communicating unpublished zebrafish material; and to Natasa Papic, Institute for Genetics, Cologne for assistance in editing the long sequence alignments. This study was supported by the Centre for Molecular Medicine, Cologne, and DFG grants SPP1110, SFB243 and SFB635. Iana Parvanova was supported by the DFG Graduate College 'Genetics of Cellular Systems'; and Cemalettin Bekpen and Julia Hunn were supported by the Cologne Graduate School in Genetics and Functional Genomics. We are particularly grateful to the anonymous referee who drew our attention to the candidate p47 GTPase sequence C46E1.3 in C. elegans.
- Mestas J, Hughes C: Of mice and not men: differences between mouse and human immunology. J Immunol. 2004, 172: 2731-2738.PubMedView ArticleGoogle Scholar
- Taylor GA, Feng CG, Sher A: p47 GTPases: regulators of immunity to intracellular pathogens. Nat Rev Immunol. 2004, 4: 100-109. 10.1038/nri1270.PubMedView ArticleGoogle Scholar
- MacMicking JD: Immune control of phagosomal bacteria by p47 GTPases. Curr Opin Microbiol. 2005, 8: 74-82. 10.1016/j.mib.2004.12.012.PubMedView ArticleGoogle Scholar
- MacMicking JD: IFN-inducible GTPases and immunity to intracellular pathogens. Trends Immunol. 2004, 25: 601-609. 10.1016/j.it.2004.08.010.PubMedView ArticleGoogle Scholar
- Taylor GA, Stauber R, Rulong S, Hudson E, Pei V, Pavlakis GN, Resau JH, Vande Woude GF: The inducibly expressed GTPase localizes to the endoplasmic reticulum, independently of GTP binding. J Biol Chem. 1997, 272: 10639-10645. 10.1074/jbc.272.16.10639.PubMedView ArticleGoogle Scholar
- Martens S, Sabel K, Lange R, Uthaiah R, Wolf E, Howard JC: Mechanisms regulating the positioning of mouse p47 resistance GTPases LRG-47 and IIGP1 on cellular membranes: retargeting to plasma membrane induced by phagocytosis. J Immunol. 2004, 173: 2594-2606.PubMedView ArticleGoogle Scholar
- Martens S, Parvanova I, Zerrahn J, Griffiths G, Schell G, Reichman G, Howard JC: Disruption of Toxoplasma gondiiparasitophorous vacuoles by the mouse p47 resistance GTPases. PLoS Pathogens. 2005, Google Scholar
- MacMicking J, Taylor GA, McKinney J: Immune control of tuberculosis by IFN-gamma-inducible LRG-47. Science. 2003, 302: 654-659. 10.1126/science.1088063.PubMedView ArticleGoogle Scholar
- Uthaiah R, Praefcke GJM, Howard JC, Herrmann C: IIGP-1, a interferon-γ inducible 47 kDa GTPase of the mouse, is a slow GTPase showing co-operative enzymatic activity and GTP-dependent multimerisation. J Biol Chem. 2003, 278: 29336-29343. 10.1074/jbc.M211973200.PubMedView ArticleGoogle Scholar
- Tuma PL, Collins CA: Activation of dynamin GTPase is a result of positive cooperativity. J Biol Chem. 1994, 269: 30842-30847.PubMedGoogle Scholar
- Prakash B, Renault L, Praefcke GJ, Herrmann C, Wittinghofer A: Triphosphate structure of guanylate-binding protein 1 and implications for nucleotide binding and GTPase mechanism. EMBO J. 2000, 19: 4555-4564. 10.1093/emboj/19.17.4555.PubMedPubMed CentralView ArticleGoogle Scholar
- Ghosh A, Uthaiah R, Howard JC, Herrmann C, Wolf E: Crystal structure of IIGP1: a paradigm for interferon-inducible p47 resistance GTPases. Mol Cell. 2004, 15: 727-739. 10.1016/j.molcel.2004.07.017.PubMedView ArticleGoogle Scholar
- Boehm U, Guethlein L, Klamp T, Ozbek K, Schaub A, Fütterer A, Pfeffer K, Howard JC: Two families of GTPases dominate the complex cellular response to interferon-γ. J Immunol. 1998, 161: 6715-6723.PubMedGoogle Scholar
- p47 (IRG) GTPase Database. [http://db.aghoward.uni-koeln.de/public/database2/global/]
- Gilly M, Damore MA, Wall R: A promoter ISRE and dual 5' YY1 motifs control IFN-gamma induction of the IRG-47 G-protein gene. Gene. 1996, 179: 237-244. 10.1016/S0378-1119(96)00366-6.PubMedView ArticleGoogle Scholar
- Zerrahn J, Schaible UE, Brinkmann V, Guhlich U, Kaufmann SH: The IFN-inducible Golgi- and endoplasmic reticulum-associated 47-kDa GTPase IIGP is transiently expressed during listeriosis. J Immunol. 2002, 168: 3428-3436.PubMedView ArticleGoogle Scholar
- Pai EF, Krengel U, Petsko GA, Goody RS, Kabsch W, Wittinghofer A: Refined crystal structure of the triphosphate conformation of H-ras p21 at 1.35 A resolution: implications for the mechanism of GTP hydrolysis. EMBO J. 1990, 9: 2351-2359.PubMedPubMed CentralGoogle Scholar
- Resh MD: Fatty acylation of proteins: new insights into membrane targeting of myristoylated and palmitoylated proteins. Biochim Biophys Acta. 1999, 1451: 1-16. 10.1016/S0167-4889(99)00075-0.PubMedView ArticleGoogle Scholar
- Maurer-Stroh S, Gouda M, Novatchkova M, Schleiffer A, Schneider G, Sirota FL, Wildpaner M, Hayashi N, Eisenhaber F: MYRbase: analysis of genome-wide glycine myristoylation enlarges the functional spectrum of eukaryotic myristoylated proteins. Genome Biol. 2004, 5: R21-10.1186/gb-2004-5-3-r21.PubMedPubMed CentralView ArticleGoogle Scholar
- Ling J, Pi W, Bollag R, Zeng S, Keskintepe M, Saliman H, Krantz S, Whitney B, Tuan D: The solitary long terminal repeats of ERV-9 endogenous retrovirus are conserved during primate evolution and possess enhancer activities in embryonic and hematopoietic cells. J Virol. 2002, 76: 2410-2423. 10.1128/jvi.76.5.2410-2423.2002.PubMedPubMed CentralView ArticleGoogle Scholar
- Singh G, Lykke-Andersen J: New insights into the formation of active nonsense-mediated decay complexes. Trends Biochem Sci. 2003, 28: 464-466. 10.1016/S0968-0004(03)00176-2.PubMedView ArticleGoogle Scholar
- Mishra R, Gara SK, Mishra S, Prakash B: Analysis of GTPases carrying hydrophobic amino acid substitutions in lieu of the catalytic glutamine: implications for GTP hydrolysis. Proteins. 2005, 59: 332-338. 10.1002/prot.20413.PubMedView ArticleGoogle Scholar
- Delarbre C, Jaulin C, Kourilsky P, Gachelin G: Evolution of the major histocompatibility complex: a hundred-fold amplification of MHC class I genes in the African pigmy mouse Nannomys setulosus. Immunogenetics. 1992, 37: 29-38. 10.1007/BF00223542.PubMedView ArticleGoogle Scholar
- Mashimo T, Glaser P, Lucas M, Simon-Chazottes D, Ceccaldi PE, Montagutelli X, Despres P, Guenet JL: Structural and functional genomics and evolutionary relationships in the cluster of genes encoding murine 2',5'-oligoadenylate synthetases. Genomics. 2003, 82: 537-552. 10.1016/S0888-7543(03)00176-9.PubMedView ArticleGoogle Scholar
- Trowsdale J, Barten R, Haude A, Stewart CA, Beck S, Wilson MJ: The genomic context of natural killer receptor extended gene families. Immunol Rev. 2001, 181: 20-38. 10.1034/j.1600-065X.2001.1810102.x.PubMedView ArticleGoogle Scholar
- Noel L, Moores TL, van der Biezen EA, Parniske M, Daniels MJ, Parker JE, Jones JDG: Pronounced intraspecific haplotype divergence at the RPP5 complex disease resistance locus in Arabidopsis. Plant Cell. 1999, 11: 2099-2111. 10.1105/tpc.11.11.2099.PubMedPubMed CentralView ArticleGoogle Scholar
- Angata T, Margulies EH, Green ED, Varki A: Large-scale sequencing of the CD33-related Siglec gene cluster in five mammalian species reveals rapid evolution by multiple mechanisms. Proc Natl Acad Sci USA. 2004, 101: 13251-13256. 10.1073/pnas.0404833101.PubMedPubMed CentralView ArticleGoogle Scholar
- Leister D: Tandem and segmental gene duplication and recombination in the evolution of plant disease resistance gene. Trends Genet. 2004, 20: 116-122. 10.1016/j.tig.2004.01.007.PubMedView ArticleGoogle Scholar
- Barten R, Torkar M, Haude A, Trowsdale J, Wilson MJ: Divergent and convergent evolution of NK-cell receptors. Trends Immunol. 2001, 22: 52-57. 10.1016/S1471-4906(00)01802-0.PubMedView ArticleGoogle Scholar
- Nathan C, Shiloh MU: Reactive oxygen and nitrogen intermediates in the relationship between mammalian hosts and microbial pathogens. Proc Natl Acad Sci USA. 2000, 97: 8841-8848. 10.1073/pnas.97.16.8841.PubMedPubMed CentralView ArticleGoogle Scholar
- Murray HW, Nathan CF: Macrophage microbicidal mechanisms in vivo: reactive nitrogen versus oxygen intermediates in the killing of intracellular visceral Leishmania donovani. J Exp Med. 1999, 189: 741-746. 10.1084/jem.189.4.741.PubMedPubMed CentralView ArticleGoogle Scholar
- Fang FC: Antimicrobial reactive oxygen and nitrogen species: concepts and controversies. Nat Rev Microbiol. 2004, 2: 820-832. 10.1038/nrmicro1004.PubMedView ArticleGoogle Scholar
- Adams O, Besken K, Oberdorfer C, MacKenzie CR, Takikawa O, Daubener W: Role of indoleamine-2,3-dioxygenase in alpha/beta and gamma interferon-mediated antiviral effects against herpes simplex virus infections. J Virol. 2004, 78: 2632-2636. 10.1128/JVI.78.5.2632-2636.2004.PubMedPubMed CentralView ArticleGoogle Scholar
- Daubener W, MacKenzie CR: IFN-gamma activated indoleamine 2,3-dioxygenase activity in human cells is an antiparasitic and an antibacterial effector mechanism. Adv Exp Med Biol. 1999, 467: 517-524.PubMedView ArticleGoogle Scholar
- Prada-Delgado A, Carrasco-Marin E, Pena-Macarro C, Del Cerro-Vadillo E, Fresno-Escudero M, Leyva-Cobian F, Alvarez-Dominguez C: Inhibition of Rab5a exchange activity is a key step for Listeria monocytogenes survival. Traffic. 2005, 6: 252-265. 10.1111/j.1600-0854.2005.00265.x.PubMedView ArticleGoogle Scholar
- Forbes JR, Gros P: Divalent-metal transport by NRAMP proteins at the interface of host-pathogen interactions. Trends Microbiol. 2001, 9: 397-403. 10.1016/S0966-842X(01)02098-4.PubMedView ArticleGoogle Scholar
- Flo TH, Smith KD, Sato S, Rodriguez DJ, Holmes MA, Strong RK, Akira S, Aderem A: Lipocalin 2 mediates an innate immune response to bacterial infection by sequestrating iron. Nature. 2004, 432: 917-921. 10.1038/nature03104.PubMedView ArticleGoogle Scholar
- Schaible UE, Kaufmann SH: Iron and microbial infection. Nat Rev Microbiol. 2004, 2: 946-953. 10.1038/nrmicro1046.PubMedView ArticleGoogle Scholar
- Ogawa M, Yoshimori T, Suzuki T, Sagara H, Mizushima N, Sasakawa C: Escape of intracellular Shigella from autophagy. Science. 2005, 307: 727-731. 10.1126/science.1106036.PubMedView ArticleGoogle Scholar
- Gutierrez MG, Master SS, Singh SB, Taylor GA, Colombo MI, Deretic V: Autophagy is a defense mechanism inhibiting BCG and Mycobacterium tuberculosis survival in infected macrophages. Cell. 2004, 119: 753-766. 10.1016/j.cell.2004.11.038.PubMedView ArticleGoogle Scholar
- Nelson DE, Virok DP, Wood H, Roshick C, Johnson RM, Whitmire WM, Crane DD, Steele-Mortimer O, Kari L, McClarty G, Caldwell HD: Chlamydial IFN-γ immune evasion is linked to host infection tropism. Proc Natl Acad Sci USA. 2005, 102: 10658-10663. 10.1073/pnas.0504198102.PubMedPubMed CentralView ArticleGoogle Scholar
- Modlin RL: Activation of toll-like receptors by microbial lipoproteins: role in host defense. J Allergy Clin Immunol. 2001, 108 (4 Suppl): S104-S106. 10.1067/mai.2001.118299.PubMedView ArticleGoogle Scholar
- Tian D, Traw MB, Chen JQ, Kreitman M, Bergelson J: Fitness costs of R-gene-mediated resistance in Arabidopsis thaliana. Nature. 2003, 423: 74-77. 10.1038/nature01588.PubMedView ArticleGoogle Scholar
- Burdon JJ, Thrall PH: The fitness costs to plants of resistance to pathogens. Genome Biol. 2003, 4: 227-10.1186/gb-2003-4-9-227.PubMedPubMed CentralView ArticleGoogle Scholar
- Weatherall DJ, Miller LH, Baruch DI, Marsh K, Doumbo OK, Casals-Pascual C, Roberts DJ: Malaria and the red cell. Hematology (Am Soc Hematol Educ Program). 2002, 35-57.Google Scholar
- Schmid-Hempel P: Variation in immune defence as a question of evolutionary ecology. Proc R Soc Lond B Biol Sci. 2003, 270: 357-366. 10.1098/rspb.2002.2265.View ArticleGoogle Scholar
- Haller O, Kochs G: Interferon-induced MX proteins: dynamin-like GTPases with antiviral activity. Traffic. 2002, 3: 710-717. 10.1034/j.1600-0854.2002.31003.x.PubMedView ArticleGoogle Scholar
- Praefcke GJK, McMahon HT: The dynamin superfamily: universal membrane tubulation and fission molecules?. Nat Rev Mol Cell Biol. 2004, 5: 133-147. 10.1038/nrm1313.PubMedView ArticleGoogle Scholar
- Klamp T, Boehm U, Schenk D, Pfeffer K, Howard JC: A giant GTPase, very large inucible GTPase-1, is inducible by IFNs. J Immunol. 2003, 171: 1255-1265.PubMedView ArticleGoogle Scholar
- Haller O, Acklin M, Staeheli P: Influenza virus resistance of wild mice: wild-type and mutant Mx alleles occur at comparable frequencies. J Interferon Res. 1987, 7: 647-656.PubMedView ArticleGoogle Scholar
- Jin H, Yamashita T, Ochiai K, Haller O, Watanabe T: Characterization and expression of the Mx1 gene in wild mouse species. Biochem Genet. 1998, 36: 311-322. 10.1023/A:1018741312058.PubMedView ArticleGoogle Scholar
- Staeheli P, Grob R, Meier E, Sutcliffe J, Haller O: Influenza virus-susceptible mice carry Mx genes with a large deletion or a nonsense mutation. Mol Cell Biol. 1988, 8: 4518-4523.PubMedPubMed CentralView ArticleGoogle Scholar
- Trede NS, Langenau DM, Traver D, Look AT, Zon LI: The use of zebrafish to understand immunity. Immunity. 2004, 20: 367-379. 10.1016/S1074-7613(04)00084-6.PubMedView ArticleGoogle Scholar
- Christoffels A, Koh EG, Chia JM, Brenner S, Aparicio S, Venkatesh B: Fugu genome analysis provides evidence for a whole-genome duplication early during the evolution of ray-finned fishes. Mol Biol Evol. 2004, 21: 1146-1151. 10.1093/molbev/msh114.PubMedView ArticleGoogle Scholar
- Meyer A, Van de Peer Y: From 2R to 3R: evidence for a fish-specific genome duplication (FSGD). Bioessays. 2005, 27: 937-945. 10.1002/bies.20293.PubMedView ArticleGoogle Scholar
- Flajnik MF, Du Pasquier L: Evolution of innate and adaptive immunity: can we draw a line?. Trends Immunol. 2004, 25: 640-644. 10.1016/j.it.2004.10.001.PubMedView ArticleGoogle Scholar
- Kimbrell DA, Beutler B: The evolution and genetics of innate immunity. Nat Rev Genet. 2001, 2: 256-267. 10.1038/35066006.PubMedView ArticleGoogle Scholar
- Fujita T, Matsushita M, Endo Y: The lectin-complement pathway: its role in innate immunity and evolution. Immunol Rev. 2004, 198: 185-202. 10.1111/j.0105-2896.2004.0123.x.PubMedView ArticleGoogle Scholar
- Daumke O, Weyand M, Chakrabarti PP, Vetter IR, Wittinghofer A: The GTPase-activating protein Rap1GAP uses a catalytic asparagine. Nature. 2004, 429: 197-201. 10.1038/nature02505.PubMedView ArticleGoogle Scholar
- Gilly M, Wall R: The IRG-47 gene is IFN-gamma induced in B cells and encodes a protein with GTP-binding motifs. J Immunol. 1992, 148: 3275-3281.PubMedGoogle Scholar
- EnsembleArchive. [http://feb2005.archive.ensembl.org/index.html]
- The Mouse Genome Informatics. [http://www.informatics.jax.org/]
- MRC RFCGR: The Fugu Genomics Project. [http://fugu.biology.qmul.ac.uk/]
- ENSEMBL: Fugu Genome Browser. [http://www.ensembl.org/Fugu_rubripes/]
- IMCB Fugu Genome Project. [http://www.fugu-sg.org/]
- Human Genome Browser Gateway. [http://genome.ucsc.edu/cgi-bin/hgGateway]
- The Danio rerio Sequencing Project (Sanger). [http://www.sanger.ac.uk/Projects/D_rerio]
- Mouse-human synteny alignments (Sanger). [http://www.sanger.ac.uk/Projects/M_musculus/publications/fpcmap-2002/mouse-s.shtml]
- Ensembl Genome Browser. [http://www.ensembl.org/index.html]
- Saitou N, Nei M: The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol. 1987, 4: 406-425.PubMedGoogle Scholar
- Kumar S, Tamura K, Jakobsen IB, Nei M: MEGA2: molecular evolutionary genetics analysis software. Bioinformatics. 2001, 17: 1244-1245. 10.1093/bioinformatics/17.12.1244.PubMedView ArticleGoogle Scholar
- Felsenstein J: Confidence limits on phylogenies: an approach using the bootstrap. Evolution. 1985, 39: 783-791.View ArticleGoogle Scholar
- BCM search Launcher. [http://searchlauncher.bcm.tmc.edu/multi-align/multi-align.html]
- EMBL-EBI clustal-w. [http://www.ebi.ac.uk/clustalw/]
- BOXSHADE 3.21. [http://www.ch.embnet.org/software/BOX_form.html]
- Schug J, Overton GC: TESS: Transcription Element Search Software on the WWW: Technical Report CBIL-TR-1997-1001-V.0.0. 1997, Pennsylvania: Computational Biology and Informatics Laboratory, School of Medicine, University of PennsylvaniaGoogle Scholar
- Transcription Element Search System. [http://www.cbil.upenn.edu/tess]
- Lenhard B, Sandelin A, Mendoza L, Engstrom P, Jareborg N, Wasserman WW: Identification of conserved regulatory elements by comparative genome analysis. J Biol. 2003, 2: 13-10.1186/1475-4924-2-13.PubMedPubMed CentralView ArticleGoogle Scholar
- Tools for Phylogenetic Footprinting Purposes. [http://www.phylofoot.org]
- Li Y, Chambers J, Pang J, Ngo K, Peterson PA, Leung WP, Yang Y: Characterization of the mouse proteasome regulator PA28b gene. Immunogenetics. 1999, 49: 149-157. 10.1007/s002510050476.PubMedView ArticleGoogle Scholar
- Lafuse WP, Brown D, Castle L, Zwilling BS: Cloning and characterization of a novel cDNA that is IFN-gamma-induced in mouse peritoneal macrophages and encodes a putative GTP-binding protein. J Leukoc Biol. 1995, 57: 477-483.PubMedGoogle Scholar
- Carlow D, Marth J, Clark-Lewis I, Teh H-S: Isolation of a gene encoding a developmentally regulated T cell specific protein with a guanine nucleotide triphosphate-binding motif. J Immunol. 1995, 154: 1724-1734.PubMedGoogle Scholar
- Sorace JM, Johnson RJ, Howard DL, Drysdale BE: Identification of an endotoxin and IFN-inducible cDNA: possible identification of a novel protein family. J Leukoc Biol. 1995, 58: 477-484.PubMedGoogle Scholar
- Taylor GA, Jeffers M, Largaespada DA, Jenkins NA, Copeland NG, Vande Woude GF: Identification of a novel GTPase, the inducible expressed GTPase, that accumulates in responses to interferon γ. J Biol Chem. 1996, 271: 20399-20405. 10.1074/jbc.271.34.20399.PubMedView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.