key: cord-1029583-co20fw3j
authors: Mashimo, T.; Simon-Chazottes, D.; Guénet, J. -L.
title: Innate Resistance to Flavivirus Infections and the Functions of 2′-5′ Oligoadenylate Synthetases
date: 2008
journal: Immunology, Phenotype First: How Mutations Have Established New Principles and Pathways in Immunology
DOI: 10.1007/978-3-540-75203-5_4
sha: be3c5419d94f3aa20abe46f4228fec784f770ebe
doc_id: 1029583
cord_uid: co20fw3j

Mouse susceptibility to experimental infections with flaviviruses is significantly influenced by a cluster of genes on chromosome 5 encoding a family of proteins with enzymatic properties, the 2′–5′ oligoadenylate synthetases (OAS). Positional cloning of the locus in question has revealed that susceptibility of laboratory inbred strains to this class of virus is associated with a nonsense mutation in the gene encoding the OAS1B isoform. Analysis of the molecular structure of the cluster in different mammalian species including human indicates that the cluster is extremely polymorphic with a highly variable number of genes and pseudogenes whose functions are not yet completely established. Although still preliminary, a few recent observations also substantiate a possible role for OAS1 in human susceptibility to viral infections (West Nile virus, SARS, etc.) and its possible involvement in some other diseases such as type 1 diabetes and multiple sclerosis. Finally, convergent observations indicate that the molecules encoded by the 2′–5′ OAS cluster might be involved in other fundamental cellular functions such as cell growth and differentiation, gene regulation, and apoptosis.

The severity of the clinical manifestations of most infectious diseases is greatly influenced by environmental factors and by the physiological status and the immune competence of the infected organism. It is also strongly influenced by genetic factors controlling the virulence of the pathogen and the susceptibility of the host. For this reason, co-evolution of infectious organisms with their hosts has been often compared to the "battle of two genomes" leading in general to infections with less deleterious consequences, which is after all the best way (if not the only way?) to ensure the survival of both organisms in the long term (Lengeling et al. 2001) .

Investigations made with the aim of better understanding the genetic mechanisms that operate during the initial steps of infection are of major interest because they can provide information that may, in turn, help in the development of better strategies for fighting infectious diseases. Experiments of that kind, however, are not easy to perform because many parameters interfere with the experimental protocols, often making difficult the unambiguous delineation of "resistant" and "susceptible" phenotypes in a population of experimentally infected animals. In most cases, resistance or susceptibility to a pathogen depends on the complex interactions of multiple genes that control the host response. In a few cases, however, the situation is greatly simplified by the observation of clear-cut phenotypic differences between various inbred strains of laboratory mice after experimental infection. In this review we describe the historical case of genetic resistance to flaviviruses, the experiments that led to its elucidation at the molecular level, and the consequences of these discoveries.

Flaviviruses are positive-sense, single-stranded, RNA viruses that are generally transmitted to warm-blooded animals through mosquito or tick bites. Many individuals exhibit flavivirus-specific antibodies, suggesting that infections by these viruses are mild or even unapparent, and revealing that some degree of adaptation has occurred between the virus and its host. In some other cases, however, flaviviruses can cause epidemic outbreaks in humans, and infected patients may exhibit a wide range of symptoms ranging from transient febrile illness to life-threatening hemorrhagic fevers (dengue and yellow fever) and meningo-encephalitis syndromes [Japanese encephalitis and West Nile (WN) fever]. The reasons why some flaviviruses cause severe clinical manifestations only in a small percentage of infected individuals are probably numerous and accordingly they have not yet been completely elucidated, but recurrent epidemiological observations and recent scientific data indicate that host-dependent genetic factors might be important.

Variations in innate flavivirus susceptibility in mice were reported for the first time in the early 1930s, and investigations performed during the following decades have fully confirmed these differences. By and large, one can consider that all laboratory inbred strains of mice that have been tested so far, with the exception of strain PL/J, are susceptible to experimental infections while most wild mice or mice from inbred strains recently derived from wild progenitors are resistant. To mention just one example of this difference in susceptibility, we reported that a single intra-peritoneal inoculation, equivalent to 100 times the median lethal dose (LD 50 ) of the WN virus (strain IS-98-ST1), administered to adult mice of the classical laboratory inbred strains BALB/c or C57BL/6 was lethal for all the animals 9.5±1.5 days after inoculation, while mice from unrelated inbred strains recently derived from wild ancestors of either the Mus musculus domesticus (WMP/Pas), Mus m. musculus (MAI/Pas, MBT/Pas, PWK/Pas), or Mus spretus (SEG/Pas, STF/Pas) species were totally resistant to the same treatment (Mashimo et al. 2002) . During this experiment, infectious particles of WN virus could be detected in the brain of all infected mice after 5 days of infection, and the amounts of virus peaked at 10 9 focus forming units (FFU)/g of brain tissue by day 7 (Mashimo et al. 2002) . High levels of anti-WN antibody could also be detected in surviving animals, indicating that the virus replicated in resistant strains. This experiment was just one of many experiments of the same kind performed over the past 40 years with a variety of flaviviruses, using several routes of inoculation and several doses and strains of virus. All these experiments yielded similar results, confirming that the phenotype of resistance/susceptibility is not WNspecific but, on the contrary, extends to other types of flaviviruses as well.

The resistance to flaviviruses was also demonstrated to be controlled by a major locus on chromosome (Chr) 5, designated flavivirus resistance (symbol Flv ), with basically two alleles: Flv r , which is dominant and induces resistance, and Flv s , which is recessive and correlates with susceptibility. Most classical laboratory inbred strains are homozygous for the Flv s allele. A third allele leading to "minor resistance" ( Flv mr ) has also been found segregating in wild mice of the M. m. molossinus subspecies [for historical details on the discovery and genetics of the Flv locus, readers may refer to Brinton and Perelygin (2003) ]. Congenic "resistant" strains have been produced by back-crossing for several generations the successive resistant offspring of an initial cross between resistant wild mice (any species) with a "susceptible" laboratory inbred strain. This classical breeding strategy allowed, at the same time, refinement of the genetic localization of the Flv locus on mouse Chr 5 and allowed the production of unlimited populations of "resistant" and "susceptible" mice with an otherwise similar genetic constitution, a very helpful material for experimentation.

Considering the relatively simple (monofactorial) genetic basis of WN resistance in the mouse and the "genomic" tools that became available after the genome sequencing effort in this species, we decided to embark on the positional cloning of the Flv locus. Readers who may be interested in reading our publication about the positional cloning of Flv (Mashimo et al. 2002) , must know that, since we had no evidence that the gene we were cloning was identical to Flv itself, we provisionally gave it another name ( Wnv for WN virus-with two alleles Wnv r and Wnv s ) even though we had little doubt that the two genes were presumably one-and-the-same entity. We now have molecular proof of this identity. . A first difficulty in this project arose when we found that, among the offspring of an intersubspecific backcross of the type (BALB/ c×MBT/Pas)F1×BALB/c, which we expected to be a mixed population with 50% of the individuals being Flv r / Flv s and the other 50% being Flv s / Flv s , all mice heterozygous for Flv r survived while not all mice with a Flv s / Flv s genetic constitution died as we would have expected. This is a good illustration of a major pitfall in this kind of experiment, where it is always risky or even impossible to trust in a "dead-or-alive" phenotype after an experimental infection, even if the latter is performed in the same highly standardized conditions. To bypass this difficulty and be able to achieve a high-resolution genetic mapping of the Flv locus, an absolutely necessary step in the positional cloning process, we derived a set of subcongenic mice by selecting, with the help of microsatellite markers, those offspring where a crossover event occurred that reduced the critical genetic interval containing the Flv locus. Offspring from these mice (all of the same genetic constitution) were challenged with a standardized dose of virus and finally classified as "resistant" or "susceptible." This rather tedious procedure allowed us to localize, with a very high degree of confidence, the Flv locus within an interval flanked by markers D5Mit408 and D5Mit242 , which is roughly equivalent to 300 kb of DNA (Fig. 1 ).

Over 350 offspring from two intersubspecific backcrosses [(BALB/c×MBT)F1×BALB/c or (C57BL/6×MAI)F1×C57BL/6] were raised and genotyped for the region of Chr 5 flanking the locus for flavivirus resistance and four mice were found with a recombinant haplotype in the critical region (Nos. 27, 117, 222, and 244) . These recombinant mice were then mated with either BALB/c or C57BL/6 susceptible partners, and around 30/50 progenies of these crosses were challenged with West Nile virus and classified as "susceptible" or "resistant." This protocol, because it involves a rather large sample of animals with exactly the same genotype, is highly reliable for the purpose phenotyping based on a dead-or-alive phenotype This genetic interval contains about 30 genes whose sequence and expression pattern are well known (Fig. 2 ). Among these genes, the cluster encoding the interferon inducible oligoadenylate synthetases ( Oas1 , Oas2, Oas3 etc.) appeared top-ranked in the list of candidates for reasons that will be explained later. We then decided to have a more careful look at the sequence of these genes in both "resistant" and "susceptible" mice. We observed several single nucleotide polymorphisms (SNPs) among the different strains or species studied, a finding that was not surprising considering the polyphyletic origins of the laboratory inbred strains of mice (Wade et al. 2002) . We found it remarkable, however, that in one of the elements of the cluster, namely the Oas1b gene, all susceptible mice had a T → C transition in the fourth exon of this gene, replacing an arginine residue with a premature stop codon. The perfect and absolute correlation between susceptibility to viral infection and the occurrence of a stop codon was observed independently in two laboratories (Mashimo et al. 2002; Perelygin et al. 2002) and supported the hypothesis that a truncated, and presumably inactive form of 2´-5´OAS is indeed causative of the innate susceptibility to flavivirus infection. The presence of a stop codon is also compatible with susceptibility behaving as a fully recessive trait and fits perfectly with one of the known functions of the interferon inducible enzyme 2´-5´OAS. In addition to this absolute phenotype/genotype correlation, it has also been reported that a flavivirus-resistant phenotype could be restored in a susceptible mouse strain by replacing the 3´ portion of the susceptible Oas1b sequence in 129/SvJ/RW4 ES cells, by homologous recombination with a 129/SvJ/ RW4 DNA sequence containing four substitutions characteristic of the Oas1b resistance allele, in particular a reversion from TGA to CGA (Scherbik et al. 2007 ).

2´-5´OAS are a relatively homogeneous family of enzymatic proteins with a remote evolutionary origin, since molecules with a similar structure have been identified in a wide range of species including most mammals and birds, and even the marine sponge Geodia cydonium (Cayley et al. 1982; Wiens et al. 1999; Yamamoto et al. 1998 ). In the species where the molecular organization has been studied in detail, the genes encoding 2´-5´OAS have been found to be clustered, with variations in gene copy numbers (orthologs and paralogs) among the different species, indicating that rapid evolutionary changes occurred in these regions. In this section, we will summarize the most recent findings with reference to the corresponding publications.

The human cluster is the simplest with only three genes OAS1 , OAS2 , and OAS3 within a 130-kb stretch of Chr 12 (12q24.13) (Hovnanian et al. 1998 ). These three genes share the same order of transcription and are arranged on the chromosome in the following order: centromere 5´-OAS1 -OAS3 -OAS2 -3´ (Hovanessian and Justesen 2007) . The size of these genes is relatively short (~12 kb for OAS1 ; ~36 kb for both OAS2 and OAS3 ) and analysis of their sequence reveals the presence of a conserved domain of five exons (the first 346-amino acid residues of OAS1 ), designated , Oas3 , and the eight genes in the Oas1 cluster are obvious. From the knowledge we have of their function, they appear as top-ranked candidates in the determinism of susceptibility to flaviviruses the 2´-5´OAS unit, with one copy in OAS1 , two copies in OAS2 , and three copies in OAS3 . This organization and sequence homology suggest that the OAS1 gene is probably the ancestral gene, the other two genes being derived after duplication or triplication of this ancestral gene. The promoter region of the three genes contains an interferon-stimulated response element (ISRE), which is consistent with the fact that most of the 2´-5´OAS proteins are interferon (IFN)-inducible enzymes. OAS1 is transcribed in four isoforms of 42, 44, 46, and 48 kDa respectively, depending on alternative splicing of exons 5 and 6. These isoforms, which have identical aminotermini but different carboxyl-termini, may have different functions. Formation of a human OAS1 tetramer is essential for the catalytic activity of the protein (Torshin 2005) . OAS2 is also spliced in two isoforms of 69 and 71 kDa that share a common amino-terminus of 683 residues, with extensions of 4 and 44 amino acids, respectively. OAS3 encodes a unique protein of 100 kDa (Hovanessian and Justesen 2007; Justesen et al. 2000; Rebouillat and Hovanessian 1999; Fig. 3 ) .

Aside from OAS1 , OAS2 , and OAS3 , another gene-identified by screening a cDNA expression library with anti-OAS3 polyclonal antibodies ) and by screening an EST library (Hartmann et al. 1998 )-with a sequence similar to the one of OAS1 , although somewhat bigger, was discovered. This gene encodes a 56-kDa protein that differs from the other OAS proteins by an approx. 160-amino acid extension at its C-terminus, which has homology to the interferon-inducible protein ISG15 and also has a tandem repeat of two ubiquitin-like domains (Hartmann et al. 1998 ). This protein is devoid of 2´-5´OAS activity and accordingly was named OASL for "OAS like" protein. The human OASL gene exhibits a high degree of homology with the chicken OAS gene, but the latter encodes a highly active 2´-5´OAS (Torshin 2005) . [Sarker et al. 1998 , Rebouillat et al. 1999 , Justesen et al. 2000 The mouse Oas cluster is the second most well known after human. It contains one Oas2 and one Oas3 gene, and ten Oas1 genes (or pseudogenes?) designated Oas1a to Oas1j (Eskildsen et al. 2002; Kakuta et al. 2002; Mashimo et al. 2003; Perelygin et al. 2006; Figs. 2 and 3) . These 12 genes are arranged in tandem within a stretch of DNA that spans approximately 230 kb of mouse Chr 5. The segment harboring the Oas cluster (and a few other flanking genes) is inverted when compared to the human configuration . Mouse Oas2/Oas3 genes have a genomic structure very similar to the one of human, and a very similar pattern of transcription even if the transcription products are somewhat bigger in the mouse (725 and 742 amino acids compared to 687 and 727 for OAS2; 1,138 amino acids compared to 1,087 for OAS3). Here again, based on their sequences and orientation, one can guess that these 12 genes result from the duplication, in several steps, of an ancestral copy after divergence of the primates and rodents phyla. RT-PCR performed with a set of specific primers on RNA extracted from several adult and fetal tissues, either before or after induction with double-stranded RNA (dsRNA), indicated that nine genes of the Oas1 family ( Oas1a to Oas1i ) are transcribed and some variations can be observed in the different patterns of expression. According to Mashimo and coworkers (2003) each of the nine Oas1 genes exhibits a unique battery of transcriptional regulatory elements, suggesting that each of these units has the potential to be differentially regulated. So far only Oas1a has been found to be alternatively spliced yielding two transcripts each including different parts of exon 6 ( Kakuta et al. 2002; Mashimo et al. 2003 ; Fig. 3 ). Two genes, homologous to the human OASL , have also been identified in the mouse: Oasl1 and Oasl2 . Analysis of their structure indicates that they are probably nonfunctional pseudogenes.

Twelve 2´-5´OAS ( Oas ) genes were identified in the rat genome (Chr 12) , including eight Oas1 genes (the orthologs of mouse Oas1a and Oas1e are missing in this species, while two additional isoforms are present: Oas1k and Oas1l ). Two Oas1 pseudogenes, a single Oas2 and a single Oas3 , and two Oas -like genes, Oasl1 and Oasl2 , are also present. The structure and organization of the rat cluster is very similar to the one of the mouse, which is no surprise.

The structure of the 2´-5´OAS cluster has also been investigated in detail in four other mammalian species (pig, dog, cow, and horse) and in chicken (Perelygin et al. 2005 (Perelygin et al. , 2006 . Four OAS genes ( OAS1X , OAS1Y , OAS1Z , and OAS2 ) plus a single copy of OASL1 were detected in the cow genome. The cluster in the pig genome is similar to the OAS cluster of cattle although OAS1Z is absent. Remarkably, the orthologous copy of mouse/human OAS3 is not found in either the pig or the cow genomes. The dog and horse clusters are also similar and are more like the human cluster than either the mouse or rat clusters. Dog and horse have three copies of the Fig. 3 b (continued) Alignment of the predicted amino acid sequences for the proteins encoded by the eight Oas1 genes indicates that Oas1c , Oas1d , Oas1e , Oas1f , and Oas1h are structurally similar and lack several functional domains. These observations suggest that these isoforms may actually be inactive pseudogenes rather than real isoforms. In contrast, Oas1g and Oas1a encode proteins that could be functional in the 2´,5´OAS/RNase L cascade. Note that for both a and b the scale for exons is five times larger than for introns human/mouse OAS3 ortholog ( OAS3C , OAS3M , and OAS3N ) , two copies of OAS2 (OAS2N and OAS2C ) and a single copy of the human OAS1 ortholog. However, while the dog has two copies of the OASL orthologous gene ( OASL1 and OASL2 ) , the horse has only one. Two tandemly duplicated OAS-like ( OASL ) genes were identified in the dog genome but only a single OASL ortholog was found in both the cattle and the pig genomes. The bovine and porcine OASL genes contain premature stop codons and encode truncated proteins, which lack the typical C-terminal double ubiquitin-like domains. Evidence of concerted evolution of all these paralogous 2´-5´OAS genes was obtained in rodents ( Rodentia ) and even-toed ungulates (Artiodactyla ) (Perelygin et al. 2006 ).

The best known molecules among those that are encoded by the 2´-5´OAS cluster are described as interferon-induced enzymes that polymerize ATP into 2´-5´ oligomers of adenosine with the general formula pppA (2´p5´A) n . These enzymes are activated by binding to double-stranded RNA, and their products, the 2´-5´ oligoadenylates, activate the latent endoribonuclease RNase L that finally degrades viral or cellular RNA molecules (Hovanessian and Justesen 2007) . This is how the antiviral activity of these enzymes is generally explained. Several experiments have confirmed this crucial function in the innate, antiflavivirus mechanism of defense in the mouse (Kajaste-Rudnitski et al. 2006; Lucas et al. 2003) although some experiments suggest that RNase L activation is not a major component of the OAS1B-mediated flavivirus resistance phenotype .

Sequence alignment of human OAS1, OAS2, and OAS3 reveals the presence of highly conserved stretches of 7-14 amino acids among which the pentapeptide D-F-L-K 199 -Q has been reported to represent a part of the ATP binding site, while K 199 inside this pentapeptide seems to be essential for catalytic activity (Justesen et al. 2000; Rebouillat and Hovanessian 1999) . In the mouse, interferon is also an inducer of the different OAS molecules, and alignment of the predicted amino acid sequences for the proteins encoded by Oas1c , Oas1d , Oas1e , Oas1f , Oas1h , Oas1i , and Oas1j -although structurally very similar-lack some essential functional domains, such as the LXXXPA motif (Ghosh et al. 1997) , the highly conserved aspartic acid residues in exon 2 (Sarkar et al. 1999) , and the CFK motif (Ghosh et al. 1997) . These observations suggest that, although these isoforms have retained their binding activity to dsRNA, they have lost their Mg 2+ -dependent catalytic activity and accordingly are most probably inactive pseudogenes rather than genes encoding a protein with 2´-5´OAS activity (Sarkar et al. 1999; Shibata et al. 2001 ). Another possibility that should be kept in mind would be that these isoforms, encoded in the above-mentioned genes, have acquired other functions. The three genes, Oas1g , Oas1a , and of course Oas1b , encode proteins that have been proved (or are likely) to be functional in the 2´-5´OAS/RNase L cascade.

The suspected role of the OAS1B isoform in the innate mechanisms of defense in the mouse, hypothesized after experimental infection of mice with flaviviruses and positional cloning of the Flv locus, has been confirmed by the production of a knockin, as already mentioned. It has also been confirmed by experiments performed in vitro in which stable neuroblastoma cell clones overexpressing either the mutant or wild-type OAS1B were infected with WN virus. These experiments indicated that viral replication is less efficient in cells that produce the normal copy of OAS1B than in those expressing the mutant form of the protein . The experiments have been confirmed and reinforced by other experiments performed on genetically engineered fibroblasts that could upregulate OAS1B protein expression under the control of the Tet-Off expression system (Kajaste-Rudnitski et al. 2006) .

The role played by the OAS1B isoform of 2´-5´OAS in the innate mechanisms of defense of the mouse against flavivirus infection now seems firmly established. In human and other mammalian species, however, the role of 2´-5´OAS in viral pathogenesis is much less clear. According to Perelygin and colleagues, the flavivirusspecific activity of the mouse OAS1B isoform on flaviviral replication might be correlated with a 4-amino acid deletion in the P-loop motif that is unique to this isoform and does not appear to exist in human (Perelygin et al. 2002) . This 4-amino acid deletion (12 bp) might be of special importance for the OAS1B protein to specifically interact with the ATP substrate if one considers recent data from the crystalline structure of the porcine OAS1 enzyme (Hartmann et al. 2003) . Even if this hypothesis is supported by other experiments in vitro (Urosevic et al. 1999) , an alternative explanation for the specific activity of OAS1B on flavivirus replication might be found in its promoter sequence, where several binding sites [for NF-κB, GAS, and interferon (IFN)-stimulated specific response element (ISRE)] exhibit a unique organization. In particular, it is noteworthy that OAS1B is the only gene where the NF-κB and ISRE binding sites are closely associated in tandem, producing a genomic structure that has previously been reported as capable of triggering gene expression upon viral induction (Cheng et al. 1998) . Sequencing the promoter regions of the OAS1B isoform in remotely related mouse species did not provide evidence that some particular structural changes in this promoter might be associated with the phenotype of resistance or susceptibility after flavivirus infection. This supports the hypothesis that the stop codon found in the OAS1B coding sequence of most laboratory strains, which is the only obvious structural difference between susceptible and resistant genotypes, indeed is directly related to this phenotype. The 4amino acid deletion in the P-loop motif has not been found in the orthologous region of the rat OAS1B isoform but does occur in rat OAS1F. The rat, like wild mice, seems to be naturally resistant when naturally infected (Eldadah et al. 1967) .

Two recent observations arguing in favor of genetic control of human susceptibility to flavivirus infections have been published recently. The first, by Bonnevie-Nielsen and colleagues, reports a significant correlation between the basal activity of OAS1 and an A/G SNP at the exon 7 splice-acceptor site (AG or AA) of the OAS1 gene . According to these authors, in a cohort of 83 families each containing two parents and two children, allele G had a higher frequency in people with high enzyme activity than in those with low enzyme activity, with the activity being related to this polymorphism in a dose-dependent manner across the GG, GA, and AA genotypes. Allele G generates the p46 enzyme isoform, whereas allele A ablates the splice site and generates a dual-function antiviral/ pro-apoptotic p48 isoform and a novel p52 isoform. The discovery of this genetic polymorphism and of its influence on host susceptibility to flavivirus infections clearly underlines the likely importance of OAS1 in the innate mechanisms of defense.

The second observation was made in a survey performed on 33 individuals hospitalized with WN virus infection. The survey was designed to assess whether a structural change could be detected in the OAS genes of patients with a clinically severe form of the disease. Sequence comparisons between case patients and control subjects identified 23 SNPs, including a synonymous SNP in OASL exon 2 in which the reference allele occurred at a higher frequency in case patients ( p <.004). According to the authors, the RNA transcripts generated from this allele may undergo increased splicing, resulting in a dominant-negative OASL isozyme similar to the nonsense/truncation mutant form of Oas1b in mice (Yakub et al. 2005) . These two reports, although preliminary, are indicative of a possible role for the OAS1 molecules in human mechanisms of defense.

Although the role of the OAS1 molecules in innate immunity against flavivirus infections is now established, at least in the mouse, several experiments indicate that this resistance, unlike resistance to myxovirus associated with the Mx locus (Haller et al. 1998) , does not require induction by interferon (Brinton and Perelygin 2003) . This observation is totally consistent with the observation that plants transgenic for 2´-5´OAS family and for the gene encoding RNase L, were found to be resistant to experimental infections with a number of viruses such as tobacco mosaic virus, cucumber mosaic virus, and potato virus Y (Honda et al. 2003; Mitra et al. 1996; Ogawa et al. 1996) .

Finally, and again concerning the Oas1b gene of the mouse, a likely hypothesis to account for the presence of the same stop codon in virtually all laboratory strains is that all these strains inherited the same segment of Chr 5 from a common ancestor. Such a situation is not uncommon among mouse laboratory strains and was also observed by Staeheli and colleagues when they investigated the genetic basis of susceptibility to orthomyxovirus infection (Staeheli et al. 1986 ). However, whether this occurred by chance only or under some sort of selective pressure is an open question. Nonetheless, it is also clear that the use of inbred strains derived from wild specimens of different species might be a rich source of information for investigating the genetic basis of resistance/susceptibility to infectious diseases.

Aside from their well-established role in the pathology generated by flavivirus infections, 2´-5´OAS molecules (and more specifically those encoded by OAS1 ) may be involved in the outcome of diseases generated by coronaviruses or hepaciviruses. Two independent surveys indicated that SNPs in the OAS1 gene (more precisely in exons 3, 6 or in the 3´-UTR region) were associated with severe acute respiratory syndrome (SARS) susceptibility in Vietnamese or Chinese Han populations (Hamano et al. 2005; He et al. 2006) . Another study suggested that a polymorphism in the 3´-UTR of the OAS1 gene was significantly associated with a higher frequency of self-limiting infection in patients with hepatitis C (Knapp et al. 2003) . It is likely that with time, more associations between OAS1 polymorphisms and resistance to viral infections will be discovered.

Situations where different mammalian genomes harbor orthologous genes with a variable number of copies are not uncommon and it was suggested that such variations are the result of different selective environmental pressures experienced by the ancestors of modern rodents and primates. While infectious agents in natural environments certainly play an important role in natural selection (the "battle of two genomes"), however, the number and range of pathogens is not very different among the different mammalian species. It therefore makes sense to guess that the different OAS molecules have cellular functions other than the one made obvious by the accidental discovery of differential flavivirus resistance in the mouse species. In humans, for example, OAS1 , OAS2 , and OAS3 appear to be differentially induced by interferon, induced in different types of cells, and for some of them expressed even in healthy individuals, which is indicative of an eventual role under physiological conditions and not only after infections. They are also characterized by different subcellular locations. Some OAS proteins might have as-yet-unidentified catalytic activities, suggesting that they may have distinct roles in the cell. In fact, 2´-5´OAS molecules have now been demonstrated to be involved in other cellular processes such as cell growth and differentiation, gene regulation, and apoptosis (Hovanessian and Justesen 2007) .

Some polymorphisms at the OAS1 locus have been reported to be associated with a variety of human pathologies. This is the case, for example, for a SNP generating an A/G splice-site in OAS1 -exon 7, which was found to be associated with a protective effect against type 1 diabetes . This observation was later disputed (Smyth et al. 2006 ) but another polymorphism in the same gene, generating a serine/glycine substitution resulting in a functional variant, was reported as a more likely cause for the observed association with type 1 diabetes (Tessier et al. 2006) . Similarly, SNPs detected in exons 3 and 7 of OAS1 demonstrated an association with risk for multiple sclerosis in 333 patients and 424 healthy controls, suggesting that OAS1 activity is involved in the etiology of this disease (Fedetz et al. 2006 ).

In the mouse species, with the unlimited possibilities of genetic engineering in embryonic stem (ES) cells in vitro, a comprehensive survey of the different functions of the OAS molecules should be undertaken in the forthcoming years, for example by knocking out each and every gene of the cluster. Although of importance, making alterations in the coding sequences of these genes in order to assess their function(s) would not necessarily require that the stop codon in Oas1b exon 4 be "repaired" in advance. Yan and colleagues, for example, demonstrated that mutant mice lacking OAS1D ( Oas1d −/− ) displayed reduced fertility due to defects in ovarian follicle development, decreased efficiency of ovulation, and arrest at the one-cell stage of fertilized eggs (Yan et al. 2005) . This was indeed a totally unexpected function for a protein exhibiting a very high degree of similarity with OAS1B.

As we already noted, the 2´-5´OAS family of genes exhibits both an evolutionarily ancient origin and wide variations in the number of copies between species. Experimental data collected after experiments on mouse flavivirus susceptibility and preliminary observations made in humans suggest that the cluster in question is important for the maintenance of cellular homeostasis since evolutionary (environmental) forces contribute to its "shaping" (Godfrey et al. 2004 ). Since no sequences related to 2´-5´OAS genes could be identified in either Caenorhabditis elegans or in Drosophila melanogaster , however, it seems that the OAS cluster is either not absolutely fundamental for cell physiology or that it is replaced by another structure with similar functions in other developed organisms.

Variation in antiviral 2´,5´-oligoadenylate synthetase (2´5´AS) enzyme activity is controlled by a single-nucleotide polymorphism at a splice-acceptor site in the OAS1 gene

Genetic resistance to flaviviruses

Distribution of the ppp(A2´p)nAbinding protein and interferon-related enzymes in animals, plants, and lower organisms

IP-10 gene transcription by virus in astrocytes requires cooperation of ISRE with adjacent kappaB site but not IRF-1 or viral transcription

Pathogenesis of West Nile virus encephalitis in mice and rats. 1. Influence of age and species on mortality and infection

Gene structure of the murine 2´-5´-oligoadenylate synthetase family

OAS1 gene haplotype confers susceptibility to multiple sclerosis

OAS1 splice site polymorphism controlling antiviral enzyme activity influences susceptibility to type 1 diabetes

Enzymatic activity of 2´-5´-oligoadenylate synthetase is impaired by specific mutations that affect oligomerization of the protein

The mouse olfactory receptor gene family

Mx proteins: mediators of innate resistance to RNA viruses

Polymorphisms of interferon-inducible genes OAS-1 and MxA associated with SARS in the Vietnamese population

p59OASL, a 2´-5´ oligoadenylate synthetase like protein: a novel human gene related to the 2´-5´ oligoadenylate synthetase family

Crystal structure of the 2´-specific and double-stranded RNA-activated interferon-induced antiviral protein 2´-5´-oligoadenylate synthetase

Association of SARS susceptibility with single nucleic acid polymorphisms of OAS1 and MxA genes: a case-control study

Activation of defense-related gene expression and systemic acquired resistance in cucumber mosaic virusinfected tobacco plants expressing the mammalian 2´5´oligoadenylate system

The human 2´-5´oligoadenylate synthetase family: unique interferon-inducible enzymes catalyzing 2´-5´ instead of 3´-5´ phosphodiester bond formation

The human 2´,5´-oligoadenylate synthetase locus is composed of three distinct genes clustered on chromosome 12q24.2 encoding the 100-, 69-, and 40-kDa forms

Gene structure and function of the 2´-5´-oligoadenylate synthetase family

The 2´,5´-oligoadenylate synthetase 1b is a potent inhibitor of West Nile virus replication inside infected cells

Genomic structure of the mouse 2´,5´-oligoadenylate synthetase gene family

Polymorphisms in interferon-induced genes and the outcome of hepatitis C virus infection: roles of MxA, OAS-1 and PKR

The battle of two genomes: genetics of bacterial host/ pathogen interactions in mice

Infection of mouse neurones by West Nile virus is modulated by the interferon-inducible 2´-5´ oligoadenylate synthetase 1b protein

A nonsense mutation in the gene encoding 2´-5´-oligoadenylate synthetase/L1 isoform is associated with West Nile virus susceptibility in laboratory mice

Structural and functional genomics and evolutionary relationships in the cluster of genes encoding murine 2´,5´-oligoadenylate synthetases

A mammalian 2-5A system functions as an antiviral pathway in transgenic plants

Virus-induced cell death in plants expressing the mammalian 2´,5´ oligoadenylate system

Positional cloning of the murine flavivirus resistance gene

Structure of equine 2´-5´oligoade-nylate synthetase (OAS) gene family and FISH mapping of OAS genes to ECA8p15→p14 and BTA17q24→q25

The mammalian 2´-5´ oligoadenylate synthetase gene family: evidence for concerted evolution of paralogous Oas1 genes in Rodentia and Artiodactyla

The human 2´,5´-oligoadenylate synthetase family: interferon-induced proteins with unique enzymatic properties

Molecular cloning and characterization of two related and interferon-induced 56-kDa and 30-kDa proteins highly similar to 2´-5´ oligoadenylate synthetase

The nature of the catalytic domain of 2´-5´-oligoadenylate synthetases

RNase L plays a role in the antiviral response to West Nile virus

Knock-in of the Oas1br allele into a flavivirus-induced disease susceptible mouse generates the resistant phenotype

Cloning of a novel 2´,5´-oligoadenylate synthetase-like molecule, Oasl5 in mice

No evidence for association of OAS1 with type 1 diabetes in unaffected siblings or type 1 diabetic cases

Transcriptional activation of the mouse Mx gene by type I interferon

Type 1 diabetes and the OAS gene cluster: association with splicing polymorphism or haplotype

Three-dimensional models of human 2´-5´ oligoadenylate synthetases: a new computational method for reconstructing an enzyme assembly

Development and characterization of new flavivirus-resistant mouse strains bearing Flv(r)-like and Flv(mr) alleles from wild or wild-derived mice

The mosaic structure of variation in the laboratory mouse genome

Origin of the interferon-inducible (2´-5´)oligoadenylate synthetases: cloning of the (2´-5´)oligoadenylate synthetase from the marine sponge Geodia cydonium

Single nucleotide polymorphisms in genes for 2´-5´-oligoadenylate synthetase and RNase L inpatients hospitalized with West Nile virus infection

Two types of chicken 2´,5´-oligoadenylate synthetase mRNA derived from alleles at a single locus

Mice deficient in oocyte-specific oligoadenylate synthetase-like protein OAS1D display reduced fertility