DNA STRUCTURE
ERVWE1 is a 10.2-kb long full-length provirus integrated on chromosome 7q21.2. Like the proviral form of simple exogenous retroviruses, ERVWE1 is structurally composed of two long terminal repeats (LTRs), flanking the internal sequence containing gag, pol and env genes. Each LTR is composed of three regions, i.e. from 5 to 3U3, R and U5. As in other proviruses, the U3 region of ERVWE1 5LTR serves as proviral promoter and the R region of the 3LTR acts as a polyadenylation signal. Both gag and pol genes, normally coding respectively for the matrix, capsid and nucleocapsid proteins, and for the viral enzymatic machinery, are disrupted by stop codons. Only the full-length env gene that codes for the envelope glycoprotein, Syncytin-1, is functionally preserved. In addition ERVWE1 contains a 2-kb intron, at the 5LTR/gag junction, with no trivial homology or known function.
CLASSIFICATION
The current classification and nomenclature of ERVs is complex and varies between and within species. Retroviral classification is initially based on virion morphology during maturation and assembly of particles at the cell membrane. Accordingly, retroviruses are designated A-, B-, C- and D-type. ERVs are also classified on the basis of the similarity of their pol region to those of exogenous retroviruses. This point is illustrated in class I and II which group the pol MuLV-like (Murine Leukemia Virus) and the pol MMTV-like (Mouse Mammary Tumor Virus) ERVs, respectively. The International Committee on Taxonomy of Viruses (I.C.T.V., http://www.ncbi.nlm.nih.gov/ICTVdb/index.htm) has established seven genera of Retroviridae, Alpha-, Beta-, Gamma-, Delta-, Epsilon-retrovirus, Lentivirus and Spumavirus. The ERV nomenclature is heterogeneous and complex due to the difficulty to associate HERVs with physiopathological functions. It is mainly based on the Primer Binding Site (PBS) sequence, which is recognized by a specific tRNA whose one letter code then becomes the ERV suffix. As the PBS located 4 bp downstream from the U5 subdomain of the 5 ERVWE1 LTR showed extensive homology with the avian retroviruses PBS used by tRNATrp (single letter code: W) for minus-strand DNA synthesis, this family was tentatively named HERV-W. Phylogenetic trees within the pol region showed that the HERV-W family is related to ERV-9 and RTVL-H families and thus belongs to the class I endogenous retroviruses. The homologies within the pol and env genes with the murine type C and simian type D retroviruses, respectively, suggest a chimeric genome structure as described for baboon endogenous virus. Based on the size criteria, such a chimerism seemed to exist within the LTR: the 247-nt U3 and the 79- to 81-nt R elements were comparable to avian or type D retrovirus U3 and mammalian type C R elements, respectively, although the 410- to 455-nt U5 element remained unclassified as unusually long. According to the new classification, HERV-W elements belong to the genus of gammaretroviruses. No replication-competent elements could be found within the HERV-W family and no corresponding exogenous retrovirus have been characterized.
EVOLUTION

HERV-W family: ERV-W elements are detectable in the genome of all Catarrhini, i.e. Hominidae (the human, chimpanzees, gorillas, orangutans), Hylobatidae (gibbons) and Cercopithecidae (Old World monkeys), but not in Platyrrhini (New World monkeys) nor in other more distant primates. Thus, the HERV-W family entered and began to spread in the genome of primates after the divergence of the Catarrhini and Platyrrhini, i.e. less than 40 million years ago (MYA). It has been estimated, based on the divergence of sequences, that the main period of spreading of the HERV-W family within the genome lasted for about 5 M years, between 15.5 to 28.6 MYA, before the Catarrhini divergence in Hominoidae and Cercopithecidae. The HERV-W family thus remained transcriptionally active over a short period of primate evolution, in contrast to other HERV families. This spreading gave rise in the human haploid genome to at least 1401 integration events, that can be classified into 3 subfamily, the subfamily 1 being the oldest. Many of them have recombined into solitary LTRs (343 to 1100 solo LTRs), and a large portion have been generated by Line1 retrotranspositional machinery (176 pseudogenes). Today only 29 to 39 proviruses are still full length and none has remained competent for replication due to frameshifts and stop codons within the ORFs for gag, pol and/or env. Only 30 env HERV-W related regions can be found in the genome, as compared to 70 gag and 100 pro regions, indicating a preferential loss of the envelope. In addition, only 5 ORFs longer than 1000 bp have been preserved: one gag, one pro, one pol, one full-length env (Syncytin-1) and one truncated env. Among these ORFs, only two proteins seem to be produced: a Gag protein encoded on a HERV-W pseudogene, and Syncytin-1, encoded by ERWE1, the unique full-length envelope glycoprotein.

ERVWE1 integration: ERVWE1 integration occurred in the germ line of a Catarrhini ancestor before Hominoidae and Cercopithecidae divergence more than 19-25 MYA, as indicated by the identification of a deleted ERVWE1 orthologous locus in Old World Monkeys. ERVWE1 integrated on chromosome 7q21.2, into an old MaLR-e1 solo LTR. The presence in several species of 4-6 bp direct repeats, although degenerated, at both ends of the provirus attestes that ERVWE1 was generated by a retrovirus-like integration event, i.e. by using a functional ERV-W reverse transcriptase. However, whether it resulted from a re-infection, cis- or trans-retrotranspostion event remains unknown. The presence of a 12bp deletion in the env gene of ERVWE1 provirus in Hominoids and Cercopithecs suggests that this deletion, crucial for the Env fusogenic activity (see below, Protein), occured originally in a primary Catarrhini ancestor possibly soon after integration, in the youth of the ERV-W family. As ERVWE1 is part of the subfamily 3 of HERV-W elements, which is probably the youngest, this support the hypothesis that the HERV-W family remained active during a short period during primates evolution. This 12-bp deletion constitutes a signature, specific to ERVWE1. As it was shown to be unique among all ERV-W copies in human and chimpanzee genomes, this suggests that ERVWE1 did not retrotransposed after its integration and furthermore was not expressed in the hominoid germ line, as opposed to many other HERV-W loci which retrotransposed as cDNA structures (pseudogenes) using Line-1 reverse transcriptase machinery.

ERVWE1 conservation and selection in Hominoidae:
- Conservation: After integration within the Catarrhini genome, ERVWE1 genomic structure followed two divergent evolutionary pathways, a genetic drift in Cercopithecus sp. versus a domestication, in Hominoidae sp. (Hominidae and Hylobatidae). More precisely in Cercopithecus sp. an about 9kb region was deleted comprising the 4.3kb LTR-gag-pol 5 fraction of the ERVWE1 provirus, and the env ORF became further inactivated by accumulation of stop codons and frameshift mutations. In Hominoidae (Hominidae and Hylobatidae), on the contrary, ERVWE1 structure has been preserved, as well as the surrounding genomic structure within a 30kb area. gag and pol regions contain numerous stop codons and frameshifts in all Hominoidae, while Syncytin-1 ORF has been conserved.
- Selection: Diverse facts emphasize the selective pressure on ERVWE1 locus in the Hominoidae lineage and the recruitement of its envelope gene to become a bona fide gene involved in placental morphogenesis. Notably comparative analysis of the ERVWE1 locus for 24 individuals showed a variable pattern of sequence variation along the proviral locus , that is compatible with a positive pressure on the elements critical for env activity maintenance.
- Enhancer
The MALR-e1 LTR portion located upstream of the ERVWE1 provirus has been shown to act as a trophoblast specific enhancer (TSE) that co-opted with the ERVWE1 5LTR promoter, conferring on Syncytin-1 a specific and high activity in the placenta. This sequence is particularly conserved in humans as no polymorphism was observed in 48 sequences analyzed and is also strictly identical in all Hominidae sp. analyzed. In contrast, the portion located downstream of the provirus is different for each Hominidae species. The MaLR co-optation however seems to be Hominidae-specific, as in the gibbon (Hylobatidae), the MaLR is deficient in enhancer activity. In contrast the gibbon 5LTR presents higher promoter activity.
- Promoter
The 5LTR exhibits an unusually low polymorphism (one variable site in 18.0 kb) as compared to the variability described for noncoding sequences (one every 0.47 kb) and repeated sequences (one every 0.31 kb), suggesting that there has been a selective sweep of this region. Conversely, the variability of the 3LTR (one in 0.5 kb) is typical of repeated sequences. On line with this, the functional analysis of all U3 elements revealed that the human and other apes ERVWE1 5 LTRs were always more active in BeWo cells than the ERVWE1 3 LTRs.
- Transcriptional termination and postranscriptional signals
In all Hominoidae including the gibbon, the poly-A signal within the 3LTR and the post-transcriptional regulation elements (env ATG context, 5 and 3 UTRs and splice sites including those for the env mRNA processing) are strictly identical.
- Envelope
ERVWE1 env, Syncytin-1, was shown to be the most conserved env ORF of the 16 human proviruses (from 9 HERV families) still containing a env gene, even though it is the fourth oldest. The observed variability of the env ORF (one variable site every 2.2 kb) fell within the same range as the variability described for human coding sequences (one every 1.08-2.00 kb), highlighting that the behavior of this gene of retroviral origin is similar to any essential cellular gene, as opposed to infectious retroviruses or more generally RNA-based organism. The critical domains essential for classical retroviral envelope expression and function are highly conserved and clearly under functional constraint in the entire Hominoidae lineage. Most of the amino-acid changes in Syncytin-1 evolution are located in positions that are variable across env proteins (surface domain involved in receptor recognition and binding, intracytoplasmic tail involved in fusogenic activity regulation), which could represent gradual adjustment to its cellular function. Interestingly, based on sequences comparison and according to the most parsimonious scenario, one of the nonsense mutations found in Cercopithecus lineage, which eliminates the last 30 amino-acids of the env protein, occured in the Catarrhini ancestor but reverted in Hominoidae re-establishing the full-length env ORF. Furthermore, the ERVWE1 signature, which consists of four amino-acid (12-bp) deletions in the intracytoplasmic tail of the glycoprotein, were shown to be crucial for the envelope fusogenicity and all tested Hominoidae Syncytin-1 proteins present similar fusogenic activity in heterologous cell fusion assays.

Convergent evolution of endogenous retroviral envelopes: Another fusogenic endogenous retroviral envelope, Syncytin-2, has been functionally preserved in the human genome. Like Syncytin-1, Syncytin-2 is highly expressed in the placenta and thought to be involved in placental morphogenesis. Although not fusogene, a third endogenous reroviral envelope is believed to be involved in placenta development: it is the envelope gene of ERV3 provirus, whose function is associated with cell differentiation/proliferation. ERVWE1/Syncytin-1 and ERVFRDE1/Syncytin-2 are specific to primates and thus do not exist in other placentae. However, this apparent endogenous retrovirus hijacking for placentation use is not restricted to the primates. Indeed two unique endogenous envelope genes of retroviral origin have been found in the mouse, i.e Syncytin-A and -B. They both display fusogenic activity and specific expression in the syncytiotrophoblast-containing labyrinth. They are found in all Muridae tested and show striking conservation of their env coding status. In addition, the envelope of a particular class of sheep ERV, endogenous Jaagsiekte sheep retroviruses (enJSRVs), regulates trophectoderm growth and differentiation in the periimplantation conceptus and when blocked leads to pregnancy loss. Altogether the data strongly argue for convergent evolution of endogenous retroviral envelopes to serve for placentation in mammals.
REGULATION

Note: ERVWE1 transcriptional activity was shown to be regulated by several transcription factors (cAMP/PKA pathway, Oct-1, GCMa, AP-2, Sp-1), hormones (steroid hormones: oestrogen and progesterone), cytokines (TNF-alpha, INF-gamma, IFN-beta, IL-6, IL-1), environmental conditions (hypoxia), exogenous virus infections and epigenetic methylation processes.

Promoter region description: The ERVWE1 promoter is a bipartite element consisting of a retroviral promoter (the 5LTR) and a "cellular" regulatory region, called the URE, within the 436pb sequence of the directly upstream ERVWE1 integration site.
- URE: The 436bp URE is composed of three sub-domains: a cellular positive regulatory region from -436 to -128, a negative regulatory region from -128 to -67, and a trophoblast specific enhancer (TSE) from -67 to -35, all contained in the MaLR-e1 LTR. In the distal positive URE (-436 to -128), computational analyses indicated putative binding sites for transcription factors previously found to be involved in placental promoter regulation (Ap-2, Sp-1, PPAR-gamma, GATA, c-Myb, and GCMA), as well as for NF-kappaB (-214/-204) and Ap-1 (-146/-136). The p65/p50 NF-kappaB heterodimer binds to the distal URE NF-kappaB site. This binding site and the Ap-1 binding site are crucial for the positive regulation of Syncytin-1 expression by TNF-alpha, IFN-gamma, IL-beta, IL-6 and PMA as well as for its negative regulation by IFN-beta in astrocytes. However, activation by the two interleukins (IL) does not require binding of the p65 /p50 heterodimer to the NF-kappaB site. The central URE negative regulatory region contains another Ap-1 binding site as well as binding sites for glucocorticoid and progesterone receptors. However, these sites remain putative. The distal positive URE (-436 to -128) and the central negative URE (-128 to -67) have compensatory effects in the BeWo placenta cell line. Putative binding sites for ubiquitous Ap-2, Sp-1 and placenta-specific GCMa are essential constituents of the TSE. Sp-1 and GCMa effectively bind to these Sp-1 and GCMa binding sites respectively. The binding of GCMa greatly enhances promoter activity specifically in placenta cells. The most proximal region to the 5LTR (-35 to +1) contains putative binding sites for c-Myc and two sites for placenta specific GATA transcription factors. GATA 2 and 3, but not GATA 1 and 4, are able to bind in this region, and the integrity of both GATA sites seems to be needed for GATA 2 and GATA 3 binding and placenta-specific positive regulation of ERVWE1 transcription.
- 5LTR: The retroviral promoter is formed by the U3 region, like other simple retroviruses. A little inhibitor activity has been found in the U5 region. The U3 region is 247bp long and can be sub-divided into two regions: a region responsible for basal placental activity from +1 to +125, and a core promoter from +125 to +310. Indeed, progressive deletions into the first 125 bp of the 5 region induce a corresponding progressive decrease in the promoter activity of the LTR, in different cell types, but activity always remains higher in placenta cells. This region contains several putative binding sites for ubiquitous and placenta specific transcription factors (GATA, Pit-1a, two Sp-1, Ap-2, Oct-1, PPAR-RXR and another GATA). The Sp-1 and Ap-2 binding sites spanning the region +55 to +74 have been found to be essential for LTR activity, but remain putative. Interestingly both of them are localised in an oestrogen response element (nt +51 to +87). DNA binding assays showed that purified ERalpha bound specifically to this ERE. Oestrogen (E2, 2-OH-E2, 4-OH-E2) as well as progesterone induces Syncytin-1 transcription up to 20 fold (after three days of growth and treatment) and also induces the other ERVWE1 mRNA species. The distal part of the U3 region (+125 to +247) is defined as the minimal promoter region. Indeed, this core promoter is active in all cell types. It contains an Oct-1 binding site and conventional CAAT and TATA boxes located at 43 and 26 bp upstream of the R CAP site (+248), and separated from each other by one Ap-2 and two Sp-1 putative binding sites. Oct-1, but not Oct-2, Oct-4 or Oct-6, binds to the Oct-1 binding site, which is also essential to core promoter activation. The functional roles of CAAT and TATA boxes have been confirmed by mutant analyses, and this region was found to be induced by the cAMP/Protein Kinase A pathways. Furthermore cAMP/PKA stimulates GCMa association with CBP, and GCMa acetylation by CBP. This leads to the stimulation of ERVWE1 promoter activity. Note that GCMa also binds on a very distal binding site (about 2535) and is also involved in the high placenta transcriptional activity of ERVWE1.

Exogene stimuli:
- Environment: Hypoxia reduces the Syncytin-1 transcription level as demonstrated in cytotrophoblast BeWo cells, primary trophoblasts cultures, ex vivo perfused placental cotyledons under hypoxic conditions and observed in placental diseases associated with hypoxia, and overexpression of SOD-1 (superoxide dismutase-1), a regulator of oxygen species, also reduced Syncytin-1 transcription. Gene repressive effects of oxygen deficiency can be compensated by induction of the cAMP/PKA pathway. in vitro serum deprivation gives rise to an increase in the transcription of HERV-W elements, including ERVWE1.
- Virus transactivation: Herpes virus simplex type I (HSV-1) and Influenza virus A/WSN/33 infections can transactivate ERVWE1 5LTR as observed in vitro in neuronal and brain endothelial cells and in tumoral cell types. HSV immediate early proteins IE1 and IE3 both trigger ERVWE1 5LTR activity, and when co-expressed act synergically to stimulate the ERVWE1 5LTR promoter. HSV-1 IEs seem to mediate enhanced binding of Oct-1 to its cognate binding site (nt 144-168). The mechanisms by which Influenza A/WSN/33 activates the ERVWE1 promoter are unknown. However ERVWE1 transactivation by both viruses correlated with a corresponding elevation in IFN-beta transcription in SK-N-MC neuroepithelioma cells. Conversely, the observation that IFN-beta decreases ERVWE1 URE-LTR promoter activity in U-87MG astrocytic glial cells suggests complex regulatory mechanisms.

Epigenetic: Imprint hypothesis: it has been suggested that the Syncytin-1 locus, ERVWE1, could be regulated by an imprinting mechanism, and more particularly by a maternal imprint. The hypothesis is based on the biological function of Syncytin-1 in placental, and thus embryonic, development, and on the proximity of the locus to two other maternally imprinted genes that are temporally regulated in the same manner as Syncytin-1 during pregnancy, one of these two imprinting genes being another retroelement (PEG10). However, this hypothesis has not yet been confirmed. The influence of methylation on the ERVWE1 U3 retroviral promoter has been investigated and was shown to suppress its activity in HeLa and BeWo cells. Correlating with methylation control of ERVWE1 transcriptional capacity, the U3 region was found to be highly methylated in several cell types and lines (skin fibroblasts, PBMC, one breast carcinoma sample, HeLa) where Syncytin-1 is not expressed, or not systematically expressed in the case of breast cancers. It was found likewise to be highly hypomethylated in full-term placenta samples and completely unmethylated in placental BeWo cells. Note that a global hypomethylation of HERV-W LTRs elements was found in ovarian malignant tissues.

RNA: SPLICING STRATEGY
ERVWE1 full-length transcript, which would include the 2-kb intron, has never been detected. Though, ERVWE1 provirus produces 3 major monospliced transcripts. The first one is 7.4kb long. It corresponds to a splice of the 2kb intronic sequence (SD1/SA1), and thus contains the gag, pro/pol and env frames. The second one is 3.1kb long and results from the splice of the 2-kb intron, gag and pro/pol sequence (SD1/SA2). It thus contains only the env gene and is responsible for Syncytin-1 translation. The third produced transcript is a 1.3kb long fully spliced transcript (SD1/SA3). Other splice variants may theoretically exist, as at least another splice donor located at the env 5UTR/ORF junction (SD2) was found to be used in association with the SA3 splice acceptor near the end of the ORF, eliminating the full env region in one placenta cDNA. Physiological transcription of the ERVWE1 locus has been detected in several tissues. Transcription levels however are mainly low as demonstrated by the need for sensitive detection techniques such as RT-PCR or EST analyses. Besides, placental and, to a lesser extent, testicular tissues have high ERVWE1 transcriptional activity as indicated by Northern blotting detection. In addition, ERVWE1/Syncytin-1 mRNAs have been detected in multiple sclerosis and tumoral tissues as well as cancerous cell lines. Table 1 shows all tissues where ERVWE1 and/or HERV-W env type transcripts have been reported. The finding of ERVWE1-specific transcripts are indicated, however the list may be not exhaustive. Moreover, these results must be treated with caution as (i) the biological significance of low expression levels could be questioned, and (ii) the Syncytin-1 sequence is present within both the 7.4 kb transcript (e.g. in the testis and placenta), and the Syncytin-1 producing 3.1 kb mRNA (to date observed exclusively in the placenta).
TRANSCRIPTION
Placental Syncytin-1 3.1kb mRNA expression occurs specifically in trophoblast cells (extravillous cytotrophoblasts, villous cytotrophoblasts, and the syncytiotrophoblast layer), but not in placenta parenchymal cells such as fibroblasts. The variation in Syncytin-1 mRNA levels during pregnancy has been the focus of several studies but results are to some extent controversial. Thus during the first trimester ERVWE1/Syncytin-1 expression is relatively high but stable. Conflicting results were obtained concerning the latter trimesters: (i) the level of expression is reduced during the second trimester and increases again in the third trimester to reach its highest level; (ii) this level increases progressively from the second to the third trimester and falls suddenly at term, (iii) at term, the expression level remains higher than in the first trimester or becomes lower. These discrepancies may be due to the amplification method (region amplified gag-pol versus env, ERVWE1 specificity, mRNA species specificity i.e. 7.4-, 3- kb or both) or the physiological sample. Indeed the observed loss of transcription during the second trimester is considered by the authors to be potentially an artefact linked to the medical -unknown- reason that lead to interruption of pregnancy in the second trimester.

ERVWE1 encodes a 538 amino-acid, 73-kDa glycosylated (53 kDa unglycosylated) envelope protein, Syncytin-1. Structurally, Syncytin-1 protein consists of a 20AA leader peptide at the amino end, a surface subunit (SU) (AA21-317) and a transmembrane subunit (TM) (AA318-538) at the carboxy end. Syncytin-1 is synthesized as a glycosylated gPr73 precursor that associates as a homotrimeric structure. Each precursor undergoes cleavage into two mature proteins: a gp50 surface unit (SU), and a gp24 transmembrane unit (TM). The cleavage occurs at a furin cleavage site (RNKR) located at the SU/TM junction. SU and TM are further covalently linked through a disulphid bond between CWIC and CX6CC motifs of the SU and TM respectively and reach the cellular membrane. SU is responsible for recognizing and binding to specific receptors on the host cell. TM presents a hydrophobic fusion peptide (AA320-340), and a fusion core made of N- and C-terminal heptad repeats (AA352-392 and AA407-440 respectively). Heptad repeats are also involved in the homotrimerization of the above-mentioned precursors. In addition TM contains an immunosuppressive region inside the C-terminal heptad repeat (AA377-396), a carboxy-transmembrane domain (AA444-469) for protein anchoring in the membrane and ends in a cytoplasmic tail.

The conservation of ERVWE1 provirus genomic localisation and envelope open reading frame have been screened in 155 individuals. They are conserved in all individuals tested so far. Moreover, sequencing of critical elements of ERVWE1, including the env ORF but also LTR elements involved in transcriptional regulation and the splice sites necessary to generate subgenomic env mRNA, showed striking conservation among the 24 individuals (48 alleles) analysed. All the polymorphic variants of Syncytin-1 are fusogenic.
Envelope allelic variants: Five mutations have been found within Syncytin-1 ORF. One is a synonymous mutation, while the four others are non-synonymous. These non-synonymous mutations are dispersed among five Syncytin-1 protein variants: V129-R138-S307-S477 (67% of the sequenced population), VRnS (25%), VqnS (4%), aRSS (2%), VqSf (2%). Each of the 24 analyzed individuals had at least one of the two major genotypes, i.e. VRSS and VRnS. Altogether, amino-acid variants (frequencies) are: AA129 V(0.979), a(0.021); AA138 R(0.9375), q(0.0625); AA307 S(0.7083), n(0.2917). All variants are functional and display the same fusogenicity.

None yet described

None yet described