Open Access

LINEs of evidence: noncanonical DNA replication as an epigenetic determinant

Biology Direct20138:22

DOI: 10.1186/1745-6150-8-22

Received: 10 May 2013

Accepted: 6 September 2013

Published: 13 September 2013

Abstract

LINE-1 (L1) retrotransposons are repetitive elements in mammalian genomes. They arecapable of synthesizing DNA on their own RNA templates by harnessing reversetranscriptase (RT) that they encode. Abundantly expressed full-length L1s and theirRT are found to globally influence gene expression profiles, differentiation state,and proliferation capacity of early embryos and many types of cancer, albeit by yetunknown mechanisms. They are essential for the progression of early development andthe establishment of a cancer-related undifferentiated state. This raises importantquestions regarding the functional significance of L1 RT in these cell systems.Massive nuclear L1-linked reverse transcription has been shown to occur in mousezygotes and two-cell embryos, and this phenomenon is purported to be DNA replicationindependent. This review argues against this claim with the goal of understanding thenature of this phenomenon and the role of L1 RT in early embryos and cancers.Available L1 data are revisited and integrated with relevant findings accumulated inthe fields of replication timing, chromatin organization, and epigenetics, bringingtogether evidence that strongly supports two new concepts. First, noncanonicalreplication of a portion of genomic full-length L1s by means of L1 RNP-driven reversetranscription is proposed to co-exist with DNA polymerase-dependent replication ofthe rest of the genome during the same round of DNA replication in embryonic andcancer cell systems. Second, the role of this mechanism is thought to be epigenetic;it might promote transcriptional competence of neighboring genes linked toundifferentiated states through the prevention of tethering of involved L1s to thenuclear periphery. From the standpoint of these concepts, several hithertoinexplicable phenomena can be explained. Testing methods for the model areproposed.

Reviewers

This article was reviewed by Dr. Philip Zegerman (nominated by Dr. Orly Alter),Dr. I. King Jordan, and Dr. Panayiotis (Takis) Benos. For the complete reviews,see the Reviewers’ Reports section.

Keywords

LINE-1 L1 retrotransposon DNA replication Replication timing Epigenetics Pluripotency Cancer Embryonic stem cells Chromatin domains Origins of replication

Review

Introduction

L1 elements have propagated in mammalian genomes by means of autonomousretrotransposition. Retrotransposition of an L1 element occurs through reversetranscription of its RNA intermediate and subsequent insertion of an L1 cDNA copy ata new location in the genome [1]. As a result of such propagation, L1s comprise ~17%, ~19%, and ~23% of thehuman, mouse, and rat genome, respectively [24]. Among the 516,000 L1 sequences identified in the draft human genome,the majority of the elements are truncated (usually at the 5′ end) L1 copies [2]. Only 7046 L1 sequences in the reference human genome are full-lengthL1 (FL-L1) elements [5], 1000 of which have been classified as potentially active [2] in terms of retrotransposition. Although only ~80–100 active FL-L1sbelonging to the L1Hs subfamily are thought to be present in the reference humangenome [6], active FL-L1s seem to be more abundant in individual genomes [7]. Human FL-L1s are similar in length (~6 kb) but heterogenous insequence composition [5]. This heterogeneity results in a spectrum of functional capabilities ofFL-L1s, ranging from the inability to translate the encoded proteins to highly activeforms in terms of retrotransposition [8]. However, it remains unexplored whether any retrotransposition inactiveFL-L1s are capable of reverse transcription in vivo.

A human FL-L1 element contains a 5′ untranslated region (UTR), two open readingframes (ORF1 and ORF2), and a 3′ UTR followed by an A-rich tail [9]. The L1 5′ UTR houses the sense (the first 100 bp) [10] and antisense (positions 400–600) [11] promoters. Transcription from the antisense promoter is one of the knownmechanisms involved in L1 silencing, and is thought to promote the downregulation oftranscription from the L1 sense promoter because the resultant bidirectionaltranscripts are processed into small interfering RNAs (siRNAs) [12]. L1s with intact ORFs encode two proteins: ORF1p, a nucleic acidchaperone, and ORF2p, which possesses endonuclease (EN) and RT activities [reviewedin [8]]. Both proteins tend to associate with their encoding RNA [13], forming an L1 ribonucleoprotein (RNP) complex that acts as a molecularmachinery of retrotransposition [reviewed in [1]].

It has long been thought that a substantially increased retrotransposition rate islinked to a noticeable synthesis of FL-L1 transcripts and, therefore, occurs inpreimplantation embryos [14], several transformed cell lines [1517], and early meiotic spermatocytes [18]. However, recent evidence shows that retrotransposition occurs mainly inearly embryonic and cancerous cells, not in the germline [1921]. This suggests that the production of FL-L1 RNA per se is notsufficient for retrotransposition, and the factors that allow for retrotranspositionin embryos but not in the germ cell line remain unknown.

Since the acknowledgement of Barbara McClintock’s discovery of mobile geneticelements [22], the transposition and retrotransposition of these elements have been amajor research focus in this field. L1s have successfully propagated in the course ofco-evolution with their hosts’ genomes, whereas diverse mechanisms have evolvedat the genome level to repress the activity of L1s [[8] and references therein]. Given that L1s constitute one fifth of thegenome, it is logical to surmise that their co-evolution with the hosts’genomes has led not only to the evolvement of an effective defence system againstretrotransposition but also to harnessing of L1s for genome functioning. In thisregard, the mechanisms by which L1s contribute to genome functioning remain largelyunexplored. It is also not known whether the ongoing insertional mutagenesis islinked to some programmed L1-dependent processes in the nucleus.

Some efforts have been made to understand the biological significance of theabundance of L1s in the genome in the context of functionally meaningful elements andthe abundance of L1 transcripts in particular cell types. LINEs constitute asubstantial portion of scaffold/matrix attachment regions (S/MARs) in the humangenome [23]. S/MARs play an essential role in the organization of chromatin asfunctional loop domains and thus in the regulation of transcription and DNAreplication [24, 25]. This suggests that numerous L1s may regulate transcription and DNAreplication through their involvement in the establishment of the three-dimensional(3D) structure of chromatin. On the other hand, abundantly expressed FL-L1s are knownto globally influence gene expression profiles, differentiation state, andproliferation capacity of early embryos and many types of cancer, although bymechanisms which remain unclear [26]. Thus far, the S/MAR-related function of L1s remains unexplored inconjunction with their expression status. The global nature of cellular processescontrolled by abundantly expressed FL-L1s suggests that an integrative approach isrequired to study the functional role of upregulated FL-L1s. Specifically, the roleof upregulated FL-L1s should be investigated in a broad context of spatio-temporalorganization and functioning of the genome and chromatin.

An important point in this regard is that the involvement of FL-L1 transcripts in theglobal regulation of early development and carcinogenesis seems to be mediated by L1RT [26]. This raises the question as to whether substantial L1-related reversetranscription exists in early embryonic and cancer cell systems and, if so, what roleit plays. A massive nuclear L1-linked reverse transcription of unknown functionalsignificance has been reported in the mouse zygote and two-cell embryo, which isbelieved to be DNA replication independent [27]. However, this review will argue that the available data do not allow fordefinite conclusions regarding whether or not this L1-linked DNA synthesis by reversetranscription is part of the genomic DNA replication/duplication program. Therefore,it is very important to address this question experimentally.

In this review, an attempt is made to fathom how upregulated FL-L1s and their RTglobally influence the differentiation state and proliferation capacity of earlyembryos and many types of cancer. In this context, the most intriguing phenomenon tobe explored is the massive nuclear L1-linked reverse transcription found at the onsetof embryogenesis. It is difficult, if not impossible, to explain the globalepigenetic role of L1 RT and the nature of massive L1-linked reverse transcriptionwithin the framework of current concepts. Therefore, conceptual advance is the mainchallenge. Herein, available L1 data are revisited and examined in concert withrelevant findings from the fields of replication timing, chromatin organization, DNAtopology, and epigenetics. The broad picture that emerges from this integrativeapproach favors two novel fundamental concepts. First, noncanonical replication of aportion of genomic FL-L1s by means of L1 RNP-driven reverse transcription is likelyto co-exist with DNA polymerase-dependent origin-based replication of the rest of thegenome during the same round of DNA replication in embryonic and cancer cell systems.Second, the role of this mechanism is likely epigenetic. Moreover, endogenousretrotransposition may be associated, to a great extent, with failure of thisnoncanonical DNA replication of an L1 unit. An exploration of this hypothesis showsthat the mechanism of DNA replication is worthy of being retested for specificgenomic locations (distinct FL-L1 sequences) in mammalian early embryonic and cancercell systems. This is important to advance understanding of DNA replication, thebiology of L1s, and mechanisms of pluripotency and carcinogenesis.

L1 RNA and RT are essential for early embryogenesis and carcinogenesis

L1 RNAs and RT, abundantly expressed in preimplantation embryos and some cancer celllines, have been targeted in numerous experiments to investigate their potentialroles. These experiments have brought about very important but overlooked findings.Specifically, they demonstrate that the functional knockdown of L1 expression viaL1-specific RNA interference (RNAi) and the inhibition of RT both independentlyresult in the same biological outcomes [26, 28]. This suggests that both transcription and reverse transcription of L1sare links in the same chain in these cell systems.

The expression of L1s has been shown to be involved in the establishment of anundifferentiated state and a high proliferation rate upon malignant transformation ofcells. For example, the knockdown of L1 expression by L1 ORF2-specific antisenseoligonucleotides drastically inhibited 3H-thymidine incorporation in adose-dependent manner in human transformed hepatoma (Hep3B) cells [28]. In the human A-375 melanoma cell line, both transient and stablesilencing of L1s by ORF1-specific RNAi caused a 50–70% decrease inproliferation rate and promoted differentiation, as was evident from morphologicalchanges and the expression of specific markers [29, 30]. The transcription of the proliferation markers CCND1 andMYC was downregulated in A-375 derivative cells upon L1 silencing [30]. Moreover, both transient and stable downregulation of L1 expression inA-375 cells strongly reduced their tumorigenicity when the cells were inoculated inathymic nude mice [26, 30]. Notably, the targeting of L1 ORF1 by RNAi in melanoma cells wasconcomitant with the drastic reduction of translated ORF2p and RT activity in thesecells [29, 30]. Therefore, it is logical to assume that the observed phenomena are linkedto the transcription and subsequent translation of FL-L1s.

The studies performed in early mouse embryos have shown that L1 transcripts areindispensable for the onset of embryogenesis [31]. When antisense oligonucleotides targeting the 5′ UTR and ORF1 ofthe TF subfamily of FL-L1s were microinjected into the male pronucleus18–20 h after fertilization, a complete and irreversible arrest ofdevelopment occurred at the two- or, to a lesser extent, four-cell stage [31]. Despite the arrested development, the microinjected embryos remainedviable and morphologically normal for several days. However, microinjection of anORF2-specific oligonucleotide neither arrested embryonic development nor decreasedthe RT activity, probably due to a depletion of injected oligonucleotides through thetargeting of 5′-truncated L1 transcripts [31]. In contrast, continuous exposure of Hep3B cells to the oligonucleotidepresent in the culture media [28] could be an effective means to target L1 RNA by ORF2-specific RNAi.Despite the ineffectiveness of the ORF2-specific oligonucleotide at arrestingdevelopment, the fact that the effect caused by the other two types ofoligonucleotides coincided with a significant decrease of the endogenous RT activity [31] suggests that FL-L1 transcripts are essential for the onset ofembryogenesis.

An important question is to whether the role of FL-L1s in early embryos andtransformed cell lines is due to their transcription per se or also due tothe involvement of L1-encoded RT. However, the lack of an L1 RT-specific inhibitor,the questionable effectiveness of available anti-RT drugs, and the abundance of RTexpressed from endogenous retroviruses (ERVs) in embryonic and cancer cells [3234] make this task methodologically challenging. For this reason, the effectsof downregulated expression of L1s versus ERVs have been compared [26].

Nevirapine, a non-nucleoside RT inhibitor that inhibits endogenous RT, affects earlyembryos and cancer cell lines in a manner similar to the L1-specific RNAi [26, 29, 3537]. The exposure of mouse late zygotes and two- and four-cell stage embryosto nevirapine caused developmental arrest at the preimplantation stages [35]. The effect of nevirapine was dose-dependent, and the arrested blastomeresmaintained normal morphology after several days in culture [35]. However, nevirapine did not cause developmental arrest being added toearly zygotes (the first 5 hr after fertilization) and later embryos (from theeight-cell stage onwards) [35]. Exposure of a variety of human and murine tumor cell lines to nevirapinequickly reprogrammed them to differentiating derivatives: the cells exhibiteddrastically decreased proliferation rates, globally changed expression profiles ofseveral hundred genes, and downregulated expression of CCND1 andMYC [[26] with a reference to unpublished data, [29, 30, 38]]. Additionally, nevirapine induced the expression of cell-type-specificdifferentiation markers in many transformed cell lines, including the geneticallyabnormal acute myeloid leukemia (AML) cell lines with t(15;17)PML/RARA and t(8;21) AML1/ETO and primaryblasts from AML patients [29, 37, 38]. Interestingly, the effect of nevirapine was irreversible in early embryos [36] but reversible in tumor cells [29, 37, 38].

The inhibition of telomerase RT is reasoned to be an unlikely cause of thesephenomena [29, 35]. However, the interpretation of the nevirapine-caused effects as L1RT-dependent was questioned because nevirapine was an ineffective inhibitor of L1 RTin cell-based retrotransposition assays [39, 40]. Nevirapine was ineffective when tested on an FL-L1 element [39] at much lower concentrations than were effective in the reprogramming oftransformed cells [38]. The ineffectiveness of nevirapine at inhibiting the synthetic L1 RT [40] could be attributed to conformational changes of the inhibitor binding“pocket”, which could arise in this protein made of the L1 RT domain anda non-L1 segment and post-translationally modified in non-mammalian cells. Despitebeing ineffective at inhibiting retrotransposition in these assays, nevirapinenevertheless completely blocked RT activity when tested on lysates of F9 mouseteratocarcinoma cells [38], which are known to actively express FL-L1s [16]. Efavirenz, another non-nucleoside RT inhibitor, decreased theproliferation rate and promoted the differentiation of cancer cell lines in a mannerakin to nevirapine [29, 37]. It also was found to be an effective L1 RT inhibitor in in vitroretrotransposition assays when used at similar concentrations [40]. Taken together, these data suggest that although nevirapine seems to be aless potent L1 RT inhibitor than efavirenz, it can inhibit endogenous L1 RT when usedat high concentrations.

If nevirapine does inhibit L1 RT in vivo, the unresponsiveness of earlyzygotes, known to have L1 RT carried over by the spermatozoid [27], and late pre-implantation embryos, which also actively express FL-L1s [14], requires further investigation. It can be hypothesized that the presenceof a noticeable lag period between the onset of the exposure of two- and four-cellembryos to nevirapine and the developmental arrest [35] is because the presynthesized L1 RT was incapable of binding this drug.The unresponsiveness of blastocysts to nevirapine could also be because theconcentration of nevirapine reaching cells of the inner cell mass (ICM) was too lowto cause noticeable effects.

Actively transcribed and reverse transcribed L1s, rather than ERVs, are thought to bea driving force of tumorigenic reprogramming and early development progression [26]. L1-interfered A-375 cells exhibited a downregulated expression of HERV-K,the biologically most active family of human ERVs, whereas a functional knockdown ofHERV-K did not affect the level of L1 expression, proliferation rate, or phenotype [30]. Consistent with this observation, downregulated expression of murineendogenous retrovirus-like element (MuERV-L) in mouse zygotes caused only mild andtransient suppression of development [41]. Similarly, stable knockdown of expression of ERVs in early clonedtransgenic pig embryos did not interfere with normal embryonic and post-nataldevelopment [42].

The phenomenon of massive nuclear reverse transcription coinciding with a two-foldincrease of L1 DNA copy number in the mouse zygote and the two-cell embryo as well asthe transient nature of this increase (it diminishes in blastocysts) [27], strongly suggests that L1 RNA is actively and transiently reversetranscribed in preimplantation embryos. This nuclear reverse transcription could bedue to L1 RT rather than ERV RT because, based on current knowledge, ERVs are reversetranscribed in the cytoplasm [43].

Attempts to explain how L1 transcription and reverse transcription can be implicatedin fundamental biological processes in early embryos and cancers have not yet broughtabout any concrete and plausible model. Dr. Spadafora and colleagues hypothesize thatboth L1-dependent transcriptional interference and non-random retrotranspositionevents that are followed, at least in embryos, by the excision of a portion of newlyinserted L1 copies might have a role in these cell systems [27, 36, 38]. However, the term “transcriptional interference”, defined asthe activity of one transcriptional unit modified by the activity of another [44], does not specify the molecular mechanism. Furthermore, the fact that theaddition of anti-RT drugs to cancer cell lines quickly reprograms them to“normal” phenotypes, and their withdrawal abolishes this effect, does notfavor the hypothesis of genetic changes. It is unlikely that L1 RT regulatesfundamental cellular processes by massive retrotransposition in embryos and byanother means in cancers. Spadafora [36] also hypothesized that L1 RT could be implicated in the substantialrepositioning of chromatin in the nuclei and, therefore, in the modulation ofexpression of other genes. This assumption was made based on unpublished data,obtained in his laboratory, that suggest the nuclei of nevirapine-exposed F9teratocarcinoma cells undergo a reorganization of their functional compartments.However, no molecular mechanism has been proposed to explain how L1 RT can beinvolved in chromatin reorganization.

A model to explain how L1 RNAs and RT are implicated in the fundamental processes inearly embryos and certain cancers must address several issues. Specifically itshould: (i) demonstrate the utility of expressed L1 RNAs, ORF1p, and ORF2p, takinginto consideration that ORF2p acts as an RT, synthesizing cDNA; (ii) explain whyearly embryos stop dividing, but transformed cells do not show a complete lack ofproliferation in response to L1-specific RNAi or RT inhibition; (iii) explain why theinhibition of RT (most likely L1 RT as discussed above) causes irreversible arrest ofearly embryonic development versus the reversible effects in cancer cell lines; and(iv) describe how downregulation of L1s can reprogram dedifferentiated cancer cellsto their original cell types but not to other cell types. An attempt to address theseissues is made below.

L1 elements and replication timing programs in pluripotent and cancerous cells

The results of the insightful studies by Dr. Gilbert’s laboratory on changes inreplication timing and chromatin organization linked to the loss of pluripotency indifferentiating embryonic stem cells (ESCs) [45, 46] might shed light on the role of upregulated L1s in establishing anundifferentiated state in a cell.

The replication-timing program is the order in which different chromosomal domainsare replicated during S phase [47]. Genome-wide profiling of replication timing in numerous cell types inmouse and human have indicated that chromosomes consist of alternating early and lateS replicating domains [45, 4850]. Multi-megabase replication domains are prevalent in differentiated cells,whereas alternating small (400–800 kb) early and late S replicationdomains are well represented in mouse and human pluripotent ESCs and in mouse inducedpluripotent stem cells (iPSCs) [45, 49]. Importantly, the replication timing profile of the genome is a dynamic,developmentally regulated feature that is coordinated with the reprogramming of geneexpression and repositioning of chromosome domains within the nucleus [45, 46, 51]. Differentiation of mouse pluripotent ESCs to neural precursor cells(NPCs) is associated with replication timing changes that affect approximately 20% ofthe genome [45].

There has been an attempt to determine whether pluripotency is associated withdistinct features of a replication timing profile in a genomic context [45]. Two features of a replication timing profile were originally consideredto be characteristic of pluripotent cells [45]. One was the presence of small domains that change replication timing fromearly in ESCs to late in NPCs (EtoL) and, vice versa, from late to early (LtoE).These changes result in the merging of small domains into larger, coordinatelyreplicating domains with a consequent 40% reduction in number. The interruption oflate replicating L1-rich AT isohores by small early replicating (EtoL) domains andearly replicating L1-poor GC isohores by late replicating (LtoE) domains was alsothought to be a feature associated with pluripotency [45]. However, it has become evident that the consolidation of replicationdomains and their alignment to AT and GC isochores were more specific to theformation of ectoderm than mesoderm and endoderm [46]. Moreover, the improvement of the correlation of replication timing toGC/L1 content was weaker in differentiating human versus mouse ESCs [49]. In terms of replication timing features, the most notable“fingerprint” or “indicator” of pluripotency in mice wasfound to be the presence of early S replicating domains that reside in a subset ofL1-rich (~27.5%)/AT-rich (~59.7%) isochores with an unusually high (for AT isochores)density of genes [45, 51]. The large EtoL replication-timing switches of these domains are stronglyassociated with loss of pluripotency [45, 51].

A study of replication timing and transcription profiles of a variety of independentcell lines representing different stages of early mouse embryogenesis [46] has revealed that (i) loss of pluripotency is associated with a number ofEtoL replication-timing changes, which are lineage-independent and completed by thelate post-implantation epiblast stage prior to germ layer specification and arestably maintained in all downstream lineages; (ii) these EtoL changes precede thedownregulation of key pluripotency transcription factors [POU5F1 (also known asOCT4)/NANOG/SOX2]; (iii) these EtoL replication-timing changes tend to be accompaniedby a repositioning of these domains toward the nuclear periphery and a downregulationof genes residing in these segments, especially those with low CpG density promoters;(iv) the completion of lineage-independent EtoL changes coincides with a transitionof these EtoL domains to a stable silent epigenetic state, which is very difficult toreprogram back to the pluripotent state in terms of replication timing and theexpression of genes with low CpG density promoters; (v) DNA methylation of genes withlow CpG density promoters within these EtoL domains and activity of several chromatinmodifying enzymes are not a main cause of the established irreversibility; (vi) theacquired stable silencing of lineage-independent EtoL domains on autosomes isreminiscent of the irreversible heterochromatinization of the inactive X chromosome(Xi) in female mammals and occurs within the same time frame in development; (vii)the subnuclear repositioning of EtoL domains occurs in parallel with a dramaticswitch to chromatin compaction along the nuclear envelope; and (viii) theselineage-independent EtoL domains represent 6.1% or 155 Mb of the genome.Interestingly, lineage-dependent EtoL and LtoE changes, occurring after the lateepiblast stage, are easier to reprogram back than lineage-independent EtoL switches.An important conclusion from this study is that loss of pluripotency is associatedwith establishing a very stable epigenetic barrier in the absence of large-scaletranscription changes, and that these epigenetic changes are mapped tolineage-independent L1-rich/gene-rich EtoL domains [46].

It is largely unknown what mechanism drives replication timing changes during loss ofpluripotency and exactly what forces the pluripotency “indicator” domainsto replicate early in ESCs. Rif1 protein has been recently identified as a keydeterminant that establishes the replication timing program and the size ofreplication domains in mouse embryonic fibroblasts and in human transformed HeLacells [52, 53]. Rif1 is thought to perform this role by attaching certain chromatinsegments to the nuclear matrix and establishing restricted access to the Rif1-boundsegments for replication factors in early S phase [52, 53]. Rif1 expression is developmentally regulated [54]; however, the functional significance of the expression patterns and acorrelation with pluripotency are not understood. Although Rif1 is highly expressedin totipotent and many pluripotent cell types (zygotes, cleaving embryos, ESC linesmaintained in vitro, primordial germ cells), it is downregulated in the ICMof the blastocyst [54]. Rif1 becomes downregulated by the downregulation of OCT4 and NANOG [55]. Knockdown of Rif1 leads to differentiation of ESCs [55], which suggests that Rif1 is implicated, at least to some extent, in themaintenance of a pluripotency-specific replication timing profile. However, the lackof a strong correlation between Rif1 expression and pluripotency [54], the fact that Rif1 mainly regulates mid-S replication domains, and itsrole as a preventer and not a promoter of early-S replication [52, 53] suggest that this protein is unlikely to provide early-S replication ofthe EtoL pluripotency “indicator” domains.

A number of observations suggest that late replication is the default state of EtoLdevelopmentally regulated domains, and that an additional as yet unknown propertymust be imposed upon these domains in order to switch them to the early replicationstate [50]. It is worth mentioning that no one has sought to discover whether theactive transcription of L1s, found in both human and mouse ESCs [45, 56, 57], plays a role in the early replication of L1-rich EtoL domains. In thisregard, I propose that specific subsets of FL-L1 transcripts, if present, allow forthe early replication and euchromatinization of the EtoL domains to which they map.The downregulation of this transcription may trigger EtoL replication timing switchesand cause the heterochromatinization of the corresponding domains, thus contributingto loss of pluripotency. The downregulation of transcription of a different subset ofL1s might be involved in loss of totipotency. This idea is supported by the fact thatloss of either totipotency or pluripotency coincides with a wave of chromatincompaction near the nuclear periphery in the absence of large-scale changes oftranscription profiles [46, 58]. Uniformly dispersed chromatin fibers of the pronuclei undergo dramaticreorganization in two- and four-cell stage embryos when heterochromatin blocks emergenear the nuclear envelope, nucleolar precursor body, and in the nuclear interior [58, 59]. The first wave of heterochromatinization is associated with the loss oftotipotency that occurs by the eight-cell stage [60]. It is followed by a conversion of chromatin to a highly dispersedconformation in pluripotent cells but not in the lineage-restricted trophectoderm andprimitive endoderm of the blastocyst [58]. The second wave of chromatin compaction near the nuclear periphery islinked to the loss of pluripotency [46]. It is tempting to speculate that, in both cases, similar epigeneticbarriers would be established through different cohorts of EtoL changes accompaniedby the downregulation of L1 transcription from these EtoL domains in the genome.

A surprising finding might be relevant to the putative link between upregulated L1sand replication timing features: the replication timing profile of human ESCs(hESCs), derived from preimplantation blastocysts, resembles the profile of moremature mouse EpiSCs, derived from the epiblast of post-implantation embryos, but notof mouse ESCs (mESCs) [49]. Mouse EpiSCs can be characterized as cells in which many EtoL domainchanges are completed, and compact chromatin is accumulated near the nuclear envelope [46]. Therefore, a larger portion of the genome is likely to be represented byeuchromatin in mESCs than in hESCs. This can be explained by the fact that FL-L1s areten-fold more abundant in the mouse compared to the human genome [61]. It is reasonable to speculate that the number of upregulated FL-L1 unitsper genome might also be larger in mESCs than in hESCs. This could result in theabundance of early S replicating domains in mESCs, but not in hESCs, and lead to theeuchromatinization of a larger portion of the genome in mESCs when compared withhESCs.

An aberrant execution of the developmental program is thought to be an importantconstituent of carcinogenesis [62]. The characteristic features of replication timing profiles of cancerouscells support this view. Findings in malignant cells from patients with acutelymphoblastic leukemia show that (i) replication-timing changes occur in units of thesame size range (400–800 kb) as normal developmentally regulatedreplication domains; (ii) more than half of these changes align with the boundariesof developmentally regulated replication domains; and (iii) distinct replicationtiming changes can be considered a “pan-leukemic fingerprint”, whichslightly overlaps with a “pluripotent fingerprint” [63]. An overlap of the replication timing profiles of another type ofmalignant cells, teratocarcinoma cells, and pluripotent embryonic cells can be evenmore profound. Teratocarcinoma cells that resemble embryonal carcinoma cells as wellas cells of the ICM [64] are known to develop into normal tissues and germ line cells aftertransplantation to the blastocyst [65]. This suggests that the transplanted teratocarcinoma cells establish thesame “pluripotent” replication timing and gene expression profiles as therecipient cells possess at the blastocyst stage. It is tempting to speculate thatthis can be achieved, at least in part, due to a similarity between single-strandedFL-L1 transcription profiles of teratocarcinoma and the ICM cells. In fact, the L1Hs,L1PA1, and L1PA2 subfamilies equally contribute to L1 transcript profiles of humanembryonal carcinoma and ESCs, whereas older subfamilies are differentiallyrepresented in these cells [57].

Epigenetic repertoire of full-length L1 transcripts

Several recent studies have shown that FL-L1 transcripts and L1 RT are implicated inepigenetic regulation of numerous genes in normal embryonic development and also intumorigenesis [26, 57, 66]. However, the nature of this epigenetic regulation and the involvedmolecular mechanism(s) are largely unexplored and invite numerous futureinvestigations. First, little is known about whether the expressed subsets of FL-L1s,and the putative epigenetic role(s) they might have, change during development.Second, it is not clear whether the active expression of single-stranded FL-L1 RNAsregulates the state of chromatin, and, if so, whether it promotes euchromatinizationor heterochromatinization. Finally, it is not known whether transcription of an FL-L1element, FL-L1 RNA, or reverse transcription of this RNA regulates or modifies thechromatin state.

Although sequence profiles of transcribed FL-L1s and their changes during developmentare largely unknown, some data demonstrate that the transcription of distinct subsetsof L1s is likely developmentally regulated and stage-specific. Different patterns ofexpression of FL-L1s have been found on the X chromosomes during early (day0–4) compared to late (day 8–10) stages of differentiation of femalemESCs [66]. The precise sequence composition of L1s transcribed from the active Xchromosome (Xa) and the Xi, their localization on the chromosome map, and theepigenetic role they might play during early ESC differentiation remain unknown.During the late stages of differentiation, when transcription of L1s in the nucleusand from the Xa is globally reduced, transcription of L1s from the Xi is stilldetectable [66]. This transcription is thought to be bidirectional and play a role in theproduction of siRNAs that promote heterochromatinization in cis and thusdownregulate neighboring genes that escaped Xist-based silencing [66]. Importantly, sense transcription of FL-L1s seems to prevail over thebidirectional transcription in ESCs, which then appears to largely shift tobidirectional transcription of L1s as the cells differentiate. This notion issupported by two findings. First, the frequency of small RNAs derived from L1elements of TF subfamily is two-fold higher on day 5 of mESCdifferentiation than on day 0 [66]. Second, the activity of the L1 sense promoter is markedly more prevalentthan the activity of the antisense promoter in hESCs, which expresses 10 to 15 timesmore sense L1 RNA than in differentiated cells [57]. Together, these data favor the hypothesis that the L1 RNA profiles aredevelopmentally regulated.

Unidirectional (sense) and bidirectional transcription of FL-L1s can coexist in acell, and they likely play opposite epigenetic roles. Both types of transcription ofL1s have been found in ESCs [56, 57, 66]. Bidirectional transcription from the L1 5′ UTR may contribute tosilencing of a portion of the chromatin domains through siRNA-based mechanism inESCs. At the same time, unidirectional transcription of another subset of FL-L1smight promote the euchromatinization in cis of a different cohort ofdomains.

Although no direct evidence demonstrates that sense transcription of FL-L1s isimplicated in euchromatinization in ESCs, this type of transcription of FL-L1s isassociated with euchromatinization in cancer cells. This is supported by the factthat RNAi-based downregulation of the expression of FL-L1s as well as the inhibitionof RT in transformed cells causes the reprogramming of chromatin segments to a morecompact state in their derivates [30, 38]. Because unidirectional sense transcription of FL-L1s appears to shift tobidirectional transcription upon differentiation of ESCs, it is tempting tohypothesize that the epigenetic role of FL-L1 transcripts might change indevelopment.

How FL-L1 RNAs direct or mediate changes of chromatin conformation andtranscriptional activity of neighboring genes is largely unknown. There are at leasttwo potential types of FL-L1 transcripts in the nucleus — assembled andunassembled with ORF1p/ORF2p — that might have different epigenetic roles andunderlying mechanisms. Thus, the sense transcription of an FL-L1 element and/or thetranscripts, incorporated in cis into the chromatin, are essential for theformation and function of a neocentromere and the selective repression of geneswithin or adjacent to this domain [67]. It remains to be determined whether these transcripts are assembled withORF1p/ORF2p or not and whether the sense transcription of FL-L1s inhibits theactivity of neighboring genes in other genomic locations. FL-L1s, which form L1 RNPcomplexes with ORF1p and ORF2p in the cytoplasm [6870], are found in ESCs and many cancer cell lines (discussed below). Uponentering the nucleus, such FL-L1 RNPs might drive a reverse-transcription-basedmechanism linked to the establishment of a totipotent/pluripotent state in embryosand an undifferentiated state in many cancers. This idea is supported by findingsfrom the FL-L1 knockdown and RT inhibition experiments discussed above. The resultsof these experiments also favor the idea that both transcription and reversetranscription of FL-L1s are integral steps of an unknown epigenetic mechanism.L1-encoded proteins preferentially associate with and act on L1 RNA, from which theyare translated (a phenomenon termed cis-preference) [13, 71, 72]. Therefore, it is unlikely that RNAi-based knockdown of transcription ofFL-L1s and the inhibition of L1-encoded RT could target separate epigeneticmechanisms and result in the same outcome. In this context, the question arises as towhat part L1 reverse transcription plays in this mechanism.

Massive L1-linked reverse transcription found in mouse zygotic pronuclei and nucleiof the two-cell embryo is believed to be DNA replication independent for two reasons:(i) the exposure of the zygotes to aphidicolin, an inhibitor of DNA polymerase,4 h after fertilization did not block DNA synthesis as evidenced by asignificant incorporation of 5-bromodeoxyuridine (BrdU), the analogue of thymidine;however, when aphidicolin was used in conjunction with abacavir, a nucleosideinhibitor of reverse transcription, the incorporation of BrdU was strongly inhibited;and (ii) this aphidicolin-resistant abacavir-sensitive synthesis of DNA is observed4–8 h after fertilization, whereas, according to older publications, DNAreplication is thought to start 8–12 h post-fertilization [27]. The first point, namely the interpretation of aphidicolin-resistantsynthesis of DNA as unrelated to genomic DNA replication, is based on the currentconcept of DNA replication that implies genomic DNA is replicated (with the exceptionto telomeres) solely by DNA-directed DNA polymerases [reviewed in [73]]. However, it is worth mentioning that the current concept of DNAreplication, being well-established through numerous experiments and entrenched inthe minds of the scientific community, has not been tested in all genome locations inall cell systems in all organisms at all possible conditions. Potentially unexploredexceptions to the well-known mechanism may exist in distinct genome locations andcell systems. Telomerase is a notable example of a reverse transcriptase carrying itsown RNA molecule, which is used as a template to elongate chromosome ends [74]. L1 RNP could be another example of an enzyme-RNA molecular machinerydriving genome-wide replication of L1 sequences. As research has progressed, it hasbecome apparent that L1 RT and telomerase have remarkable similarities [[75] and references therein]. The second point with respect to DNA replicationstarting in the zygote 8–12 h after fertilization could be fallacious. Thereferences provided by Vitullo and co-authors [27], when traced back to original publications, lead to results obtained bymicrodensitometry of Feulgen stained pronuclei [76], a low sensitivity methodology. The provided references also lead topublications in which dating of post-fertilization events was inferred, probablyincorrectly, from time passed after the injection of human chorionic gonadotropin(HCG) [77, 78]. More accurate estimations of the timing of pronuclear DNA synthesis innaturally ovulated and fertilized mouse eggs of six different genotypes, performed bycytofluorometric measurement of ethidium bromide-stained DNA, have indicated that theS phase starts at ~4 (3.8–4.6) h post-conception and lasts between 6.4 and11.1 h in various genotypes [79]. Accordingly, it is reasonable to assume that the onset of3H-thymidine incorporation in the pronuclei at 21 h post-HCG, whichis thought to correspond to 7–9 h post-fertilization [77], and the onset of labeling with BrdU at 4 hr after fertilization [27] can be attributed to the same event: reverse transcription. The similarityof the early labeling patterns by 3H-thymidine and BrdU in male and femalepronuclei [27, 77] supports this notion. It is also worth noting that the incorporation ofeither 3H-thymidine or BrdU can only be interpreted as DNA synthesis butnot as a particular mechanism thereof.

The DNA synthesis by reverse transcription found at the onset of mouse embryogenesisis thought to be L1-linked [27]. Data obtained by quantitative PCR (qPCR) analyses with primers designedto amplify FL-L1s of the TF subfamily of L1s demonstrate an approximatetwo-fold increase of the L1 DNA copy number per haploid genome in the mouse zygote,two-cell embryo, and morula [27]; however, the time window and the phase of the cell cycle in which theqPCR analyses were performed were not indicated. Consequently, the design of theabove-mentioned experiments [27] has led to results that are inconclusive in terms of whether the L1-linkedDNA synthesis by reverse transcription is DNA replication dependent orindependent.

In this regard, it is important to compare L1-related qPCR data obtained at twopoints of the zygotic cell cycle. The first point should be during the phase of thecell cycle when reverse transcription occurs but DNA polymerase-dependent DNAreplication has not yet started. The second point should be when DNA replication iscomplete (i.e., in G2/mitosis). Although the results of such an experiment cannotprovide evidence of the nature of the observed L1-linked DNA synthesis by reversetranscription (a potential synthesis of extragenomic L1 DNA copies cannot be ruledout), this approach could be a good starting point to test whether this reversetranscription is DNA replication dependent or independent. Therefore, the dataavailable at this time are not convincing evidence of the massive nuclear reversetranscription occurring in early embryos being DNA replication independent.

Hypothesis and rationale: two modes of L1 DNA replication as an epigeneticswitch

In this review, I would like to propose that the L1-linked reversetranscription-based DNA synthesis found in early embryos and also likely to be foundin undifferentiated cancer cells is part of the DNA replication program in thesetypes of cells. This implies that two different mechanisms of DNA replication,canonical and noncanonical, can co-exist to replicate the genome during the sameround of DNA replication in early embryos, ESCs, and many cancers. In these cellsystems, a portion of genomic FL-L1 sequences is proposed to replicate by thenoncanonical mechanism (i.e., L1 RNP-driven reverse transcription starting on an L1RNA template bound to a complementary “parental” genomic L1 DNA sequence)(Figure 1). The noncanonical mechanism is proposed totrigger when FL-L1 RNAs are actively transcribed and translated and when full-size L1RNPs are assembled. Full-size L1 RNP is herein defined as consisting of FL-L1 RNA, L1ORF2p, and multiple trimers of L1 ORF1p (discussed below). Therefore, there can betwo modes of DNA replication of FL-L1 sequences in the genome: canonical andnoncanonical. The noncanonical mode of replication of FL-L1s is proposed to be L1RT-driven, origin-independent DNA replication as a part of normal early development.The canonical (DNA polymerase-driven, origin-dependent) mode of L1 DNA replication islikely to replace the noncanonical replication in differentiating cells when thesynthesis of full-size L1 RNPs is downregulated. The noncanonical mode of FL-L1replication can be recapitulated in cancer cells.
https://static-content.springer.com/image/art%3A10.1186%2F1745-6150-8-22/MediaObjects/13062_2013_Article_386_Fig1_HTML.jpg
Figure 1

A hypothetical mechanism of noncanonical L1 DNA replication. A.Formation of an L1 DNA:RNA duplex. Heterogeneous FL-L1 RNAs, being assembledwith L1 ORF1p and ORF2p, find their “parental” complementarysequences in the genome and form L1 DNA:RNA hybrids. The chaperone activity ofORF1p, which includes the melting of mismatched duplexes, is deemedindispensable for pairing of the L1 RNA with the fully complementary L1 DNA.The displaced DNA strand of an L1 unit is likely stabilized by auxiliaryfactors. B. First-strand cDNA synthesis. ORF2p bound to the 3′ endof the FL-L1 RNA nicks the bottom DNA strand and synthesizes the first cDNAstrand from the liberated 3′-hydroxyl. C. Second nick formation.When ORF2p reaches the 5′ end of the L1 RNA, it nicks the top DNA strandat the 5′ end of the L1 element. ORF2p then switches templates from theRNA to the cDNA. The L1 RNA likely dissociates at this point. D.Second-strand cDNA synthesis on the first cDNA template. E. Nicking atthe genomic DNA-cDNA junctions and the ligation of the segments of the“parental” DNA at the sites of the first and second nicks byauxiliary factors. F. Unpairing of the new L1 cDNA strands and theirpairing with the “parental” strands by auxiliary factors. The endsof the new cDNA strands are joined with the new strands synthesized by thecanonical mechanism on the adjacent segments of the “parental”strands. Each cell division produces two cells with equal amounts of old andnew DNA synthesized by a combination of two different mechanisms.

The next logical question is to why replication of FL-L1 sequences by either thecanonical or the noncanonical mechanism is important for a cell. The answer could bethat the switch from the noncanonical to canonical mode might be a fail-safe means tokeep a large set of embryo-specific genes stably silent when the noncanonicalmechanism of DNA replication is “off”. Specifically, the noncanonicalmechanism of L1 DNA replication may serve as a noncanonical epigenetic determinantthat regulates the transcriptional competence of a large cohort of neighboring genes.This regulation could be implemented through the prevention of a set of L1-rich EtoLdomains from being tethered to the inner nuclear membrane (INM) and from beingpackaged into late-replicating facultative heterochromatin (Figure 2A). It follows then that when FL-L1 sequences are not replicatedby the noncanonical mechanism, they would tend to be silenced due to their sequencecomposition. Sequence features of L1s might favor anchoring to the nuclear matrix andbinding of the origin recognition complex (ORC) – two potential mechanisms thatmay contribute to silencing of L1s and adjacent sequences. The ORC might facilitateheterochromatin assembly and tethering of L1s to the nuclear periphery (discussedbelow). Therefore, origin-based replication of a distinct set of L1s might also beconsidered an epigenetic mechanism, which contributes to the default silencing of theinvolved domains (Figure 2B). The noncanonical replicationof FL-L1s might exert rather specific, albeit different, effects on gene expressionprofiles depending on the subset of polymorphic FL-L1s involved in noncanonical DNAreplication. This implies that a cell type-specific subset of noncanonicallyreplicated FL-L1s determines the cohort of L1-rich EtoL domains that aretranscriptionally competent in this particular type of cells.
https://static-content.springer.com/image/art%3A10.1186%2F1745-6150-8-22/MediaObjects/13062_2013_Article_386_Fig2_HTML.jpg
Figure 2

A hypothetical model of two modes of replication of full-length L1s as anepigenetic switch. A: Undifferentiated (early embryonic and cancer)cells. A subset of small L1-rich domains of the genome (EtoL domains as perHiratani and co-authors [45]) replicates DNA in early S (red arrow). Transcriptional competenceis imposed on these domains during early S phase of DNA replication. TheseL1-rich EtoL domains do not tether to the nuclear lamina and have a looseconformation of chromatin loops. Genes residing in these domains are linked toundifferentiated states of a cell. Noncanonical replication of FL-L1s residingin these domains might prevent them from binding to the nuclear matrix and fromrecruiting the ORCs (panel A, right) and, therefore, from being silencedthrough the ORC/HP1-mediated pathway of heterochromatin assembly. B:Differentiated and differentiating cells. Global programmed downregulation ofFL-L1s upon differentiation of pluripotent embryonic cells results in switching“off” the noncanonical mechanism of L1 DNA replication. In theabsence of L1 RNA paired with a complementary “parental” L1 DNA fornoncanonical L1 DNA replication, L1s might attach to the nuclear matrix andform intrastrand G4 structures. The majority of replication origins areassociated with G4s [80]. The L1-bound ORC might bind HP1 in a distinct chromatin environmentand, thereby, play a crucial role in the establishment of a silent state on theL1-rich/gene-rich EtoL domains (panel B, right). These domains switch to theirdefault state characterized by late replication, dense conformation, andtethering to the nuclear periphery [50] (blue arrow). Basically, the same switch from noncanonical tocanonical replication of L1s residing within EtoL domains might occur upondifferentiation of poorly differentiated cancer cells caused by the knockdownof the expression of L1s.

A notable insight into the initiation of DNA replication in eukaryotic systems hasbrought about the concept of a “relaxed replicator” as a“context-dependent element”, which includes a DNA sequence in conjunctionwith DNA topology, DNA methylation, chromatin-bound proteins, transcriptionalactivity, and short-/long-distance chromatin effects [81]. This concept implies that the binding of the ORC to chromatin is guidedby distinct combinations of sequences, chromatin contexts, and components of nuclearstructure [81, 82]. Accordingly, the replicator-initiator interactions are thought to have anadditional function (or functions) beyond their role in DNA duplication [81]. In this context, it is logical to surmise that numerous ORCs, whichremain bound to DNA by ORC2-5 subunits throughout the cell cycle [83], influence the formation of a certain chromatin environment through therecruitment of chromatin proteins and binding to the nuclear matrix. Indeed, agrowing body of evidence indicates that the ORC is essential for the formation ofheterochromatin in eukaryotes [8487]. In mammals, the ORC recruits heterochromatin protein HP1 [86, 87]. Factors that facilitate this process have begun to be revealed, one ofwhich is an H3K9me3 environment [87].

In the context of nuclear structure, a significant portion of LINEs seem to beORC-binding sites and function as MARs. This is suggested by the fact that originscolocalize with MARs [83, 88, 89] and that human LINEs are overrepresented among S/MARs, comprising 40% ofthe sequences [23]. The high overrepresentation of LINEs among S/MARs could be because S/MARs [90] and L1 sequences (discussed below) share a particular feature: partialunpairing of DNA strands. S/MARs are functionally heterogeneous; SARs are mainlytranscription-linked, and MARs are replication origin/silent gene-associated [25, 91]. Taking into consideration the functional heterogeneity of S/MARs and thetendency of L1-rich domains to be silent and replicate late at the nuclear peripheryin differentiated cells [45, 92], L1s can be even more over-represented among the origin-associated MARsthan “bulk” S/MARs.

Several other facts also support the notion that the sequence composition of L1smakes them prone to bind ORCs. For example, poly(dA:dT) elements (5 mers or longertracts), known to be present within L1s, disfavor nucleosome occupancy not only overthemselves but also over adjacent regions [93]. Low nucleosome occupancy is thought to be a necessary, but notsufficient, requirement for the assembly of ORCs and pre-replication complexes nearthese regions [94]. Another feature of L1 sequences that might be favorable for ORC bindingis a guanine-rich tract known to form an intrastrand tetraplex (G-quadruplex or G4)in the L1 3′ UTR [95]. This feature is present in all L1s with intact 3′ UTRs [95] and conserved throughout mammalian evolution [96]. About 90% of human origins are represented by G4-forming motifs [80], and these structures are known to be nucleosome-free regions [97]. Taken together, these data suggest that ORCs are highly likely to bind toG4 structures of those L1s that tether to the nuclear matrix.

If the sequence features of L1s (G4 structures, the tendency for partial unwinding,and nucleosome disfavoring) do promote ORC binding, the L1-bound ORCs may beessential for establishing a very stable silent state on L1-rich segments of thegenome. One potential mechanism of the ORC-dependent silencing of L1s could be therecruitment of HP1 to the L1-bound ORCs. HP1γ, one of the isotypes of HP1associated with the foci of facultative heterochromatin [98], is known to contribute to the silencing of FL-L1s [99]. Knockdown of the Cbx3 gene that encodes HP1γ activates repressed L1s [99]. The strong binding activity of HP1γ with lamin B receptor, anintegral protein of the INM [100, 101], could also be involved in the sequestration of L1s to the nuclearperiphery. The recruitment of HP1 to the ORC is guided by H3K9me3 [87]. Although H3K9me3 is weakly represented on L1 sequences regardless ofwhether L1s are active or silent [102104], H3K9me3 is overrepresented within L1-rich dark (Q or G) bands [105]. Therefore, it would be timely to gain insight into the putative linkbetween the ORC, HP1, and H3K9me3 with regard to L1 silencing.

Another potential mediator of the ORC-dependent silencing of L1s might be anORC-binding factor ORCA (ORC-associated protein). ORCA associates with the ORC in thepresence of repressive histone marks and methylated DNA and functions as afacilitator of heterochromatin formation [87]. Thus, although it is not completely understood how a very stable silentstate is imposed on L1-rich domains, G4-forming motifs within L1 sequences might be‘landing pads’ for the ORC, the important player in heterochromatinassembly.

A genome-wide origin mapping study in hESCs and embryonic fibroblasts [80] has contributed a very important finding by demonstrating that EtoLdevelopmentally regulated replication domains acquire some additional origins whenthey switch their replication timing from early to late S phase. Despite a generalpositive correlation of early replication with the high density and frequency of theusage of origins, the EtoL replication domains had even slightly lower origin scoreswhen they were early replicating than when late replicating [80]. The exact localization of origins on the sequences of EtoL domains couldclarify whether the additional origins accuired upon the EtoL transition duringdifferentiation of pluripotent hESCs are L1-associated.

Together, these facts favor the hypothesis that L1s within developmentally regulatedEtoL domains can be points of a strong attachment to the INM and peripheral nuclearmatrix, thus keeping these domains in the default silent state. Importantly, such arole may be linked to the binding of the ORC by G4 within the L1 3′ UTR and,therefore, to canonical origin-based replication. L1 RNP, the molecular machinery ofthe proposed noncanonical L1 DNA replication, could be a more successful competitorfor L1 sequences than the nuclear matrix and the ORC, which would preclude the L1silencing scenario. Undoubtedly, L1-MAR and L1-ORC relationships need to beinvestigated in differentiated and non-differentiated cell systems and viewed in thecontext of developmentally regulated replication domains.

Relevant to this discussion, are three important points. First, experimentaltethering of a number of loci to the INM causes their downregulation and therepression of neighboring genes and genes that are located far from the loci.However, experimental untethering by using a competitor compound that binds thetarget site induces the repositioning of the locus and adjoining segments away fromthe nuclear periphery and re-establishes transcriptional competence [106]. Second, transcriptional competence is established at the time ofreplication [107]. Early S replicating sequences are assembled into nucleosomes enrichedwith acetylated histones H3 and H4, the marks of open chromatin, as opposed to late Sreplicating DNA, which is packaged mainly into silent chromatin marked bydeacetylated forms of these same histones [107, 108]. Third, the nuclear periphery, which is essentially a repressiveenvironment, has early S replicating and transcriptionally active subcompartments [59, 109111] that appear to be more prominent in early embryonic and transformed cellsthan in differentiated cells.

Taking all of this into account, it can be speculated that L1s, being untethered fromthe INM and repositioned into early S replication compartments, could then assemblewith acetylated H3/H4. Indeed, activation of L1s in HeLa cells by a carcinogen,benzo(a)pyrene, increases the H3K9ac mark at the L1 5′ UTR [104]. As proposed above, L1 RNP bound to complementary L1 DNA and/ornoncanonical L1 DNA replication might favor the untethering of the implicatedchromatin domains from the INM. These liberated segments can relocate to the nuclearinterior, the location of dominating early S replication [112] and transcriptional competence. Alternatively, these liberated domains canbecome early S replicating and transcriptionally competent without noticeablerepositioning towards the nuclear interior. This idea is consistent with theobservation that the nuclear periphery can be almost entirely (mouse zygote) orpartially (mESCs, many types of cancer cells) represented by euchromatin [58, 59, 113], which appears to replicate in early S phase, at least in the zygote [59]. In ESCs, the small size of alternating early- and late-replicatingdomains, together with the anchorage of late S-replicating segments to the INM [45], suggest that many small L1-rich early S-replicating pluripotency“indicator” domains are restrained in the nuclear periphery. Thelocalization of L1-encoded proteins within the nucleus can be a cue to where L1 RNPsmay act with regard to the nuclear periphery. In A-375 melanoma cells, L1ORF2p-specific fluorescent signals appear as a dense rim in the nuclear periphery andpatches of sparse speckles that protrude into the nuclear interior [30]. However, in the colon cancer cell line H1299, ORF1p-specific signals formmultiple foci across the entire space of the nucleus [114]. This suggests that L1 RNPs may act in the nuclear periphery and in thenuclear interior in a cell type-specific manner.

A replication-timing program, which governs the transcriptional competence ofchromosome domains, is established during early G1 phase, a short window ofopportunity termed the timing decision point (TDP) [115]. Post-mitotic re-establishment of 3D chromatin architecture occurs at theTDP, and developmental cues that change a replication-timing program are likely toact during this short time window [115]. If the proposed noncanonical replication of L1s does occur and functionas a regulator of replication timing and spatial positioning of the involved domains,L1 RNA-L1 DNA interactions for DNA replication should be established no later thanthe TDP. This means that the sites of noncanonical DNA replication are likely to belicensed from early G1 onward, and their licensing could serve as an epigeneticdeterminant. Alternatively, this epigenetic role could be performed by noncanonicalreplication of L1s if it starts at the TDP. The latter could be the case during thefirst round of DNA replication in the embryo. Noticeable DNA synthesis by reversetranscription, which precedes DNA polymerase-dependent DNA replication in mousepronuclei [27], could be the first phase of DNA replication and serve as the epigeneticmechanism implicated in the establishment of the initial replication timing programand chromatin architecture. From this viewpoint, it is not surprising that the DNAsynthesis by reverse transcription is more prominent in the male than the femalepronucleus [27] because the hypercondensed paternal chromatin requires more extensivereorganization than the maternal chromatin. The organization of sperm chromatinfavors the early onset of L1-related reverse transcription in the male pronucleus.Specifically, a small portion of the genome is undermethylated and packaged withhistones into active nuclease-hypersensitive chromatin; these segments of the genomeare highly enriched with L1s [116, 117]. These L1 sequences are found at the periphery of the sperm nucleus [116], the same location where pronuclear reverse transcription occurs [27].

Biological significance of L1 RNP: a step beyond retrotransposition

Two L1-encoded proteins, ORF1p and ORF2p, are translated in unequal amounts from abicistronic FL-L1 transcript [1, 118] and bind to the RNA from which they are translated [1, 13, 72]. This suggests that L1 RNP functions as a molecular machinery invivo. ORF1p forms trimers that polymerize under the very conditions thatsupport high-affinity nucleic acid binding [119]. Polymerized trimers of ORF1p bind to L1 RNA, and one or two molecules ofORF2p attach at or near the L1 RNA poly(A) tail [1, 13, 72, 119]. ORF1p possesses a nucleic acid chaperone activity on oligonucleotidesubstrates in vitro; specifically, it promotes accelerated and stringentannealing of complementary nucleic acid sequences by facilitating the melting ofimperfect duplexes, strand exchange, and the stabilization of perfect duplexes [120, 121]. However, the biological significance of the ORF1p chaperone function ispoorly understood.

First, it is unclear what type(s) of duplexes ORF1p promotes the formation of invivo. On one hand, the chaperone function of ORF1p has been demonstrated onDNA oligonucleotides in in vitro assays [120]; on the other hand, ORF1p preferentially binds to L1 RNA in vivoand in vitro[69, 122]. Considering the complementarity of the poly(A) tail of L1 RNA to thepoly(T) segment of a typical 5′ Tn/An 3′ cleavagesite of L1 EN [123], formation of a short DNA:RNA duplex is proposed to occur to prime reversetranscription during retrotransposition [120]. ORF1p is also speculated to promote the exchange of a DNA:DNA duplex toan RNA:DNA hybrid at the target site [120]. However, the enormous mass of ORF1p trimers that bind to L1 RNA [121] seems excessive to merely promote the formation of a short RNA:DNA duplexto prime cDNA synthesis in vivo. Moreover, because the liberation of3′-OH at the nick site is sufficient to prime reverse transcription on an L1RNA template in vitro[124], it remains uncertain whether such short RNA:DNA duplexes are indeedformed to initiate reverse transcription in vivo.

Second, it is unclear what processes require ORF1p as a chaperone in vivo.Its implication in retrotransposition might not be the only role it plays. EndogenousL1 RNAs, which form L1 RNPs in hESCs, belong not only to retrotranspositionallyactive (L1Hs) but also to retrotranspositionally inactive L1 subfamilies (L1PA2,L1PA3, L1PA4, L1PA6, and L1PA7) [56]. It is unlikely that hESCs synthesize retrotransposition inactive L1 RNPshaving no function. Therefore, ORF1p as a part of retrotranspositionally inactive L1RNP might play a yet unknown role.

ORF1p is deemed essential for the retrotransposition of L1s expressed from L1constructs in transfected cells [121, 125, 126]. This is evidenced by the fact that mutant ORF1 proteins with impairedchaperone function but unaffected RNA-binding activity abolished or reducedretrotransposition in comparison with the wild-type (wt) ORF1p in cell-based assays [125]. Although ORF1p is non-essential for retrotransposition in a cell-freein vitro assay [124], its availability increases the quantity and length of nascent cDNAs andpromotes the initiation of cDNA synthesis at more typical retrotransposition startsites [72]. The role of a non-mutant ORF1p in the retrotransposition of a“synthetic” L1 element in cell-based assays might be the same as in acell-free system. Specifically, it could promote the synthesis of a longer cDNAstrand, including a reporter cassette upstream of the L1 3′ UTR, so that aretrotransposition event is detectable.

While the integration of a “synthetic” L1 element into the genome israndom [127], the integration of endogenous L1s seems to be non-random and biased to asimilar sequence environment. Although post-insertional selection and recombinationinfluence the genomic distribution of L1s, the non-random integration of endogenousL1s appears to be an important factor in the biased localization of L1s inGC-poor/AT-rich regions of the genome [128132]. Analyses of the distribution of L1s in mammalian genomes have led to theconclusion that L1s tend to cluster [130, 133]. However, there is no current consensus on whether clustering is a generalfeature of L1s [130] or more pronounced among old L1 elements [133]. The 100 kb flanking sequences of human L1s of a currently activesubfamily Ta-1 (also known as L1Hs-Ta1) and older L1s (L1PA2 and L1PA5) are enrichedin L1 DNA [130]. Interestingly, the sex chromosomes, which are enriched in ancestral L1s,are much less hospitable for Ta-1 insertions than chromosome 4, which is enriched inTa-1 elements [130]. Although L1s are estimated to insert in pre-existing L1s only 13% of thetime [134], the portion of L1-derived sequences that harbor new L1 insertions can belarger. Remains of the 3′ polyA tails of previous L1 insertions that bear L1 ENrecognition motifs are thought to be common target sites for L1 retrotransposition [123].

Despite the incompleteness of our knowledge regarding the incidence, degree, andlength of sequence similarity between L1 insertions and surrounding regions,available data fit the concept of sectorial mutagenesis introduced by Jurka andKapitonov [128]. This concept implies that new insertions of transposable elements tend tooccur in specific chromosomal regions. Importantly, the density of LINEs correlatesmore strongly with specific orthologous segments of the human and mouse genomes thanwith the local GC content [3].

The factors that determine the non-random integration of endogenous L1s and randominsertions of “synthetic” L1 elements remain unexplored. It has beenhypothesized that the higher frequency of target sites and the open state ofchromatin could contribute to the insertional bias of endogenous L1s [132]. The fact that “synthetic” and endogenous L1s target the sameconsensus sequence [127, 134], but demonstrate different patterns of retrotransposition, does not favorthe notion that the frequency of the target sites could be a key factor of non-randomretrotransposition of endogenous L1s. The open state of chromatin established oncertain chromosomal domains might be a favorable condition rather than adeterminative factor for non-random retrotransposition of L1s.

Not excluding other factors that can contribute to the insertional bias of L1s, Ihypothesize that retrotransposition of endogenous L1s might be linked, at least tosome extent, to noncanonical DNA replication. This may cause non-randomretrotransposition of endogenous L1s if this process fails. More randomretrotransposition of “synthetic” L1s might be caused by the inability ofa reporter cassette bearing L1 RNA to pair with a complementary sequence in thegenome to perform noncanonical DNA replication. Moreover, the chaperone activity ofORF1p might be essential for the recognition of complementary genome sequences by L1RNAs and their pairing. ORF1p that promotes the melting of imperfect duplexes maycontribute to random retrotransposition of “synthetic” and non-randomretrotransposition of endogenous L1s. In addition to the known biased insertions ofL1s, other findings discussed below favor this hypothesis.

An unequal potency of retrotransposition among endogenous FL-L1s capable of producingfunctional proteins is thought to be, at least in part, due to differences in somemeasures of the chaperone activities of ORF1p variants [121]. Importantly, a reference point on the scale of retrotranspositionpotency, also often termed as wt, can, paradoxically, be a measure of the failure ofdistinct ORF1 proteins to perform other biologically essential functions. An exampleof such an L1 element in mice could be a retrotransposition-efficient variant ofL1spa that encodes ORF1p with an aspartic acid codon at residue 159(D159) [121]. In contrast to the D159 variant, another variant of L1spa thatencodes ORF1p with a histidine codon (H159) at this position is known as aretrotransposition-inefficient element [121]. Interestingly, the less active variant, H159 ORF1p, is much moresuccessful at melting a mispaired DNA duplex than the more active D159 ORF1p, whichis not able to fully melt an imperfect duplex in the absence of strand exchange [121, 126]. If L1 RNP does perform an important function on genomic DNA that requiresperfect pairing of L1 RNA and complementary DNA, the efficient melting of mismatchedduplexes by ORF1p could be essential for displacing L1 RNA from a mispaired DNA:L1RNA hybrid and, therefore, for promoting the formation of completely paired L1DNA:RNA hybrids. Consequently, L1 RNA that encodes ORF1p capable of efficient meltingof mismatched duplexes might be less prone to retrotransposition invivo.

Sequence composition of L1s favors the formation of L1 DNA:RNA hybrids reminiscent oflong R loops. An R loop is an unwound DNA segment, one strand of which associateswith the complementary RNA, whereas the second DNA strand appears as a displaced loop [135]. A/T richness and paired stretches of polypurines:polypyrimidines, thecharacteristic features of L1s, are required for dsDNA to be prone to the formationof an R loop [135]. The formation of R loops spanning several kb is possible; however,auxiliary factors are required for the unwinding and stabilization of long ssDNAsegments [135].

If the noncanonical mechanism of DNA replication does exist and new integrationevents of L1s are indeed linked to their noncanonical replication in early embryosand certain types of cancer, then L1 retrotransposition can be expected to occur inthese cell systems rather than in all types of cells where L1s are activelytranscribed. It has become evident that retrotransposition of genomic L1 elementsoccurs mainly in early embryonic cells but not in germline cells, as previouslythought [19, 136], and in certain types of cancer cells but not in normal tissuecounterparts [20, 21]. Despite the fact that L1 RNA is available in female germ cells andtremendously abundant in spermatogenic cell fractions, retrotransposition events arerare in the germlines; this is in contrast to much more frequent integration eventsin preimplantation embryos [19]. Interestingly, L1 RNA that is retrotransposition inactive in thegermlines is carried over into the embryo where it remains stable and then becomesretrotranspositionally active in the cleaving embryo [19]. L1 RNA transcribed in the embryo causes even more retrotranspositionevents than does carried-over L1 RNA [19]. Direct evidence of endogenous L1 retrotransposition associated with L1activation in cancer cells has recently been reported [20, 21]. In transgenic mice carrying the human L1 element, retrotranspositionevents have been found to occur in chemically induced skin tumors but not in theadjacent normal skin tissue [20]. As shown by two high-throughput L1-targeted resequencing methods,retrotransposition of L1Hs occurs in certain human colorectal tumors but not in thesurrounding normal colon tissues [21]. Importantly, the number of new L1 insertions in human colorectal tumorswas not correlated with the degree of hypomethylation of L1 promoters [21]. These findings suggest that the activation of L1 expression as a resultof L1 demethylation is a necessary but not sufficient condition to cause a highretrotransposition rate.

Further investigation of some identified hotspots of L1 insertions is required todetermine conditions and molecular processes that might favor L1 retrotranspositionon the genome scale. Such hotspots have been found in the vicinity of certain genesexpressed in gonads and during embryogenesis [132]. If retrotransposition of L1s is linked to their noncanonical replication,L1 insertions are anticipated to be biased to certain sets of L1-rich EtoL domains,the early replication and transcriptional competence of which is characteristic ofeither embryonic or cancer cells. In this context, it would be interesting to studytwo potential links regarding the L1 integration hotspots: (i) the link between theL1 sequences integrated within the hotspots and FL-L1 RNA species carried over intothe zygote and expressed during development, and (ii) the link between these hotspotsand L1-rich EtoL developmentally regulated domains.

The functional features of L1 ORF2p could potentially make it capable of providingthe putative noncanonical replication of FL-L1 loci. In in vitro assays, L1RT has demonstrated a high processivity on both RNA and ssDNA templates and theability to switch templates from RNA to cDNA in order to synthesize the second strandcDNA [124, 137]. This is consistent with the capability of L1s to generate full-lengthinsertions in vivo. L1 EN generates single strand nicks in dsDNA with apreference for TA dinucleotides within 5′ TTTT/AA 3′ target tracts;additionally, L1 EN is able to efficiently nick other sets of dinucleotides within aloose consensus sequence [123, 138]. From the perspective of the proposed model, this nicking flexibilitymight be essential to generate two nicks in order to prime first and second strandcDNA synthesis. The first nick might occur in the bottom strand complementary to anL1 A-rich tail, which is known to consist of the AATAAA polyA signal followed byAn interrupted by short GT- or T-rich motifs [139]. A putative location of the second nick could be in the top strand at thebeginning of the L1 5′ UTR. An interesting nuance is that L1 EN activityincreases dramatically on an unwound DNA helix [123].

Together, these findings favor the hypothesis that L1 RNP functions as the molecularmachinery of noncanonical replication of L1 units in concert with other cellularfactors that are likely to be available when this mechanism is active. BothL1-encoded proteins appear to be indispensable for the proposed mechanism, and thisimplies that only those FL-L1 transcripts that are assembled with both proteins canfunction in terms of noncanonical replication. The strong preferential binding ofORF1p and ORF2p with their encoding L1 RNA and the chaperone activity of ORF1p canprovide a high level of specificity in recognition of “parental” L1 DNAunits subjected to noncanonical replication and, therefore, in epigenetic targetingon a genomic scale.

L1 RNA and proteins: what, where, and when?

To better understand the epigenetic role(s) of the activated FL-L1s, it is importantto determine the patterns of L1 transcription and synthesis of L1 proteins indifferent types of cells and potential links between these patterns and cellphenotypes. It has long been accepted that the production of FL-L1 RNAs occursnotably in the germline, early embryos, and many types of cancer cells, whereas it ismainly shut down in the majority of normal unstressed somatic tissues. However, ahighly complex picture of L1-involving pathways in a tissues-specific context hasstarted to emerge. First, recent research shows that cells from a broad range ofnormal organs actively synthesize FL-L1 RNAs [140]; however, the majority of these transcripts undergo splicing and/orpremature polyadenylation [140142]. Unfortunately, the scant amount of data on L1 RNA sequences and proteinsin some organs, e.g., in the placenta and esophagus [140, 143, 144], does not allow for a definite conclusion on these patterns. Second, thereis some uncertainty regarding the interpretation of data obtained by methodologicalapproaches appropriate for a less complex system. (For a discussion of these issues,please see Additional file 1). Consequently, the tissuespecificity of the synthesis of FL-L1 RNAs and full-size L1 RNPs and correlationswith distinct phenotypic features can be discussed only with a limited degree ofconfidence. Third, there appears to be different patterns of expression of FL-L1 RNAsthat might not necessarily result in the production of full-size L1 RNPs. Therefore,the relationship between the synthesis of FL-L1s and phenotypic properties might notbe straightforward. Finally, the assembling/functioning of full-size L1 RNPs seems tobe suppressed in gametes, but becomes activated after fertilization.

Currently, there is no convincing evidence that noticeable amounts of full-size L1RNPs are synthesized in either male (adult and prepubertal) or female germline cells.Available data suggest that either FL-L1 RNA and ORF1p are synthesized, but ORF2p ismissing (as observed in early meiotic spermatocytes) or present at a very low level(female gametes), or L1 proteins are produced from shortened L1 transcripts (asobserved in secondary spermatocytes and spermatids) (Figure 3). Therefore, execution of an RT-mediated program might be blocked ingermline cells. Specifically, in adult human testes, L1-related poly(A) RNAs areextremely abundant; however, no FL-L1 RNA has been found by Northern blotting becausethe majority of FL-L1 RNAs undergo premature polyadenylation combined with splicing [140142]. These processed L1 RNAs can potentially translate into either ORF1p orORF2p or their truncated forms. Indeed, both ORF1p and ORF2p (or their truncatedforms, as discussed in Additional file 1) have beendetected by immunostaining in somatic testicular cells, secondary spermatocytes, andimmature spermatids in adult human testes [143]. Similarly, no FL-L1 RNA has been detected in adult mouse testes byNorthern blotting, whereas short L1 transcripts of variable lengths were abundant inboth germ and somatic cells [18]. In adult mouse testes, ORF1p-related immunostaining has been detected insomatic cells and spermatids, but no ORF2p-specific immunostaining has been revealed [18].
https://static-content.springer.com/image/art%3A10.1186%2F1745-6150-8-22/MediaObjects/13062_2013_Article_386_Fig3_HTML.jpg
Figure 3

Patterns of expression of L1 RNA, ORF1p, and ORF2p duringspermatogenesis. Available data suggest that the synthesis of full-sizeL1 RNP consisting of FL-L1 RNA, ORF1p, and ORF2p is repressed duringspermatogenesis. No L1-related products have been found in spermatogonia.Transient expression of FL-L1 RNA and ORF1p (but not ORF2p) occurs inspermatocyte I at the onset of meiosis (leptotene through mid-pachytene stageof prophase I). These upregulated FL-L1s are implicated in chromosome pairing [145]. Transcription of L1s resumes in spermatocyte II and continues inspermatids. However, L1 RNAs are mostly short, spliced, and prematurelypolyadenylated species that translate into either ORF1p or ORF2p. Theirfunctional significance is not known. Transcription of L1s, ORF1p, and ORF2p isdownregulated by the spermatozoa stage. A small amount of FL-L1 RNA (notdetectable by Northern blotting) is available in spermatozoa. ORF2p, not foundin testicular spermatozoa [18], is detectable in the sub-acrosomal space in mature sperm [27].

Although processed L1 transcripts prevail in adult testes, FL-L1 RNAs, which areundetectable by Northern blotting, might be present in early meiotic (leptotene andzygotene) spermatocytes. This cell fraction is rare in adult mouse testes but is muchbetter represented in prepubertal testes where it accounts for the abundant~7 kb sense-strand L1 transcripts [18]. The transient expression of L1s and ORF1p coupled with L1 DNAdemethylation is intrinsic to the onset of normal meiosis (leptotene throughmid-pachytene stage) in every round of spermatogenesis [145, 146]. This type of L1 expression, downregulated in late meiotic prophase I, isunrelated to the production of processed L1 transcripts triggered later inspermatogenesis [18, 145]. The transient expression of FL-L1s and ORF1p is proposed to be aprogrammed, though not understood, event associated with chromosome pairing andassembly of the synaptonemal complexes in male meiosis [145]. Because L1 retrotransposition is highly repressed in the germlinecompared with early embryogenesis [19], and ORF2p appears to be unavailable in early spermatocytes, it can bespeculated that the L1 RNPs, implicated in early male meiosis, are not full-size L1RNPs. Similar to the onset of male meiosis, ORF1p is transiently expressed in femalegerm cells entering meiotic prophase I in the mouse embryonic ovary [144], suggesting the same role of L1 expression in chromosome pairing.

Another category of germ cells likely to accumulate small amounts of FL-L1 RNA aremale and female gametes. This is supported by the fact that FL-L1 RNA carried overinto the zygote by both gametes causes detectable retrotransposition events in theembryo [19]. Because the carried-over FL-L1 RNA remains stable and capable ofretrotransposition during early embryogenesis [19], it might be implicated in the L1-linked RT-dependent synthesis of DNA notonly in the zygote but also in the cleaving embryo. While small amounts of FL-L1 RNAseem to be present in both gametes, it remains unexplored whether this RNA isassembled into RNP with one or both L1-encoded proteins. The synthesis of ORF2p isdownregulated in testicular sperm cells [18] but appears to resume in the epididymal spermatozoa because ORF2p is foundin the sub-acrosomal space of these cells [27]. Therefore, the synthesis of ORF2p seems to restart at the terminal stageof spermiogenesis when the synthesis of ORF1p is downregulated. As shown byimmunostaining, ORF1p and ORF2p are barely detectable at the terminal stages of mouseoogenesis [27, 144]. In a full-size L1 RNP, ORF1p is typically present in great excesscompared to ORF2p [1]; therefore, weak ORF1p-specific immunostaining can reflect thedownregulated synthesis of ORF1p in oocytes. Because of the paucity of ORF2p presentin L1 RNP, [1], weak ORF2p-specific immunostaining dispersed within the cytoplasm of theoocyte [27] may not suggest the lack of ORF2p if compared with the amount of ORF2p inthe epididymal spermatozoid. Together, these findings favor the assumption that smallamounts of FL-L1 transcripts can be stored in both male and female gametes, but theformation of FL-L1 RNA/ORF1p/ORF2p complexes might be blocked due to thedownregulated synthesis of ORF1p. ORF2p, which is synthesized at the very terminalstage of sperm maturation and also seems to be present in the oocyte, could bedestined to initiate the synthesis of L1 DNA by means of reverse transcription inboth zygotic pronuclei.

Preimplantation embryos likely synthesize full-size L1 RNPs; however, systematicstudies of L1 RNAs/ORF1p/ORF2p are required for definite conclusions. Stronglyupregulated expression of L1s [147], noticeable RT-dependent DNA synthesis, and the significant increase of L1copy number in two-cell mouse embryos [27] suggest that L1 RNPs are likely present and function during this stage.The abundance of sense-strand FL-L1 transcripts in mouse blastocysts [14] and the presence of FL-L1 RNAs and ORF1p assembled into RNPs in hESCs andiPSCs [56, 148] favor the idea that full-size L1 RNPs can be present at least inpluripotent cells of the blastocyst.

The exact developmental window when such RNPs are formed remains to be determined.Although genome-wide intense upregulation of L1s occurs and plays an important rolein preimplantation embryos, the less apparent production of sense-strand FL-L1 RNAsand proteins can still be present or transiently reinstated in distinct lineages orcell types later in development. The possibility of L1 expression andretrotransposition in human neural progenitor cells is suggested by the increasedcopy number of endogenous L1s in adult brains when compared with heart and liversamples obtained from the same individuals [149]. Moreover, mouse myogenic precursors, the differentiation of which ispromoted by nevirapine [38], could also be a cell type that synthesizes some amount of full-size L1RNPs.

Several types of cancer cells also seem to synthesize full-size L1 RNPs. With regardto L1-related products, the most studied cancer cells are cell lines derived fromgerm-cell tumors, mostly testicular, that are embryonal carcinomas andteratocarcinomas (teratomas with an embryonal carcinoma component) [64]. Embryonal carcinoma cells are highly malignant counterparts of the ICM:they express pluripotency markers and can be maintained as undifferentiated cells orinduced to differentiate by morphogens [64, 150]. Mouse F9 and C44 embryonal carcinoma and human NTera2D1 teratocarcinomacell lines are known to actively synthesize sense-strand FL-L1 RNAs [15, 16]. These transcripts form RNPs with ORF1p in F9 and C44 cells [16, 68]. The presence of RT activity associated with L1 RNPs in NTera2D1 cells [151] suggests that full-size L1 RNPs may be synthesized in these cells. Thefact that the malignant pluripotent cells originate from germ cells but not othercell types could be explained by the proposition that L1 RNAs, synthesized duringgametogenesis and carried over into the zygote, have a pluripotency-linked functionin the early embryo.

FL-L1s and ORF1p are also upregulated in a range of tumors and transformed celllines, and this upregulation correlates with a transition to undifferentiatedphenotypes, higher tumor grade, and poorer prognosis [17, 114, 140, 152154]. Despite the lack of parallel analyses of ORF2p in many studied cancers,the results discussed in the second section of this review suggest that ORF2p is alsopresent in numerous poorly differentiated tumors. Consequently, the synthesis of L1RNPs is likely a characteristic feature of many cancers.

In addition to many types of cancers, the activation of L1s might be intrinsicallylinked to cell dedifferentiation in certain regenerating cell systems. For example,L1-like retrotransposon that encodes ORF1 and ORF2 is dramatically upregulated in theblastema during axolotl (Ambystoma mexicanum) limb regeneration [155]. This activation of L1s slightly precedes the upregulation of a limbregeneration marker [155]. Interestingly, the completion of the regeneration of the amputated limbwas accompanied by a 16% increase in L1 DNA copy number [155]. Surprisingly, the second wave of regeneration after re-amputation of thesame limb resulted in a 70% increase in L1 DNA copy number [155]. Although the nature of this enormous increase in L1 copy number is notknown, the authors interpret their data as retrotransposition. It is tempting tospeculate that the herein proposed noncanonical mechanism of L1 DNA replication mightbe recapitulated in blastema to allow cell dedifferentiation. Moreover, the increasein L1 DNA copy number after the completion of the regenerative process could be dueto the accumulation of extrachromosomal L1 DNA copies. The synthesis of episomal L1DNA copies (discussed below) and their stockpiling might be part of a cell“memory” mechanism aimed to accelerate noncanonical L1 DNA replicationand dedifferentiation in response to a repetitive severe injury. It has been reportedthat repeated amputation of the axolotl limb results in accelerated regeneration [156], although the underlying mechanism is not understood.

Together, the analysis of L1 expression in a cell type-specific context shows that acorrelation between a noticeable production of FL-L1 RNAs and cell phenotypicproperties is not straightforward. Importantly, the production of FL-L1s might notnecessarily always lead to the synthesis of both L1-encoded proteins and theformation of full-size RNPs. This may occur at the onset of meiosis and during theterminal stages of gametogenesis. The synthesis of full-size L1 RNPs in mitoticallydividing cells appears to be strongly implicated in establishing gene expressionprofiles characteristic for totipotent/pluripotent and poorly differentiatedcells.

A shift in the current L1 paradigm: has the time come?

Barbara McClintock’s theoretical postulates on transposable genetic elements [22] were met with enduring reluctance, but this reluctance eventually evolvedinto acknowledgement of her discovery and revolutionary concept. Paradoxically, thisnow widely accepted concept seems to have become a barrier that impedes conceptualadvances in L1 research.

The current L1 paradigm can be described as retrotransposition-centered: (i)retrotransposition is the only RT-dependent function of L1s considered so far; (ii)the drastic upregulation of L1s in early embryos and cancers is often deemed anon-specific response to general demethylation of the genome because it cannot beintended for retrotransposition, and other possible functions are usually notconsidered; (iii) the upregulation of endogenous L1s is usually thought to be asufficient condition for retrotransposition despite the lack of a correlation betweenthe abundantly expressed L1s and retrotransposition in the male germ line [19]; (iv) while retrotranspositionally active L1s are under scrutiny,retrotranspositionally inactive FL-L1s are neglected as elements that might bereverse transcribed and play an essential role in a cell; and (v) the attributedfunction of premature polyadenylation and splicing of L1 transcripts known to occurin many tissues is to defend against retrotransposition [140142]; however, it is unlikely that L1 RNA is synthesized and processed merelyto be non-functional.

The adherence to this retrotransposition-centered paradigm is reflected in thescarcity of research exploring other potential L1 RT-driven mechanisms. The adherenceto the current paradigm is also evident in the interpretation of data demonstratingsignificant increases in L1 DNA copy number in the mouse zygote and cleaving embryos [27] as well as in regenerated axolotl limbs [155] as a result of numerous retrotransposition events. Although the reportedincrease in L1 DNA copy number may be partially caused by retrotransposition events,it is unlikely that retrotransposition is the sole L1 RT-dependent process in thesecell systems. The activation of L1s in colorectal tumors is accompanied by 0 to17 newinsertions per tumor sample [21]. Even if the degree of L1 activation in the early mouse embryo andregenerated axolotl limb is higher than in tumors, L1 retrotransposition rates inthese cell systems are unlikely to be many times higher than in cancers.Consequently, other possible L1 RT-driven mechanisms are worth exploring.

One such mechanism could be the synthesis of extrachromosomal L1 DNA or L1DNA-containing sequences. Abundant extrachromosomal circular L1 DNA-containingproducts have been found in yeast [157] and certain types of cancer cell lines [158]; however, the biological significance of these products remains unknown.The extrachromosomal L1 DNA copies might be the cause of significantly increased L1DNA copy numbers in the regenerated axolotl limbs. The extrachromosomal L1 DNA copiesmay also be temporarily synthesized during early embryogenesis, thereby causing theamplification of L1 DNA copy number.

The second potential L1 RT-driven mechanism is the noncanonical L1 DNA replicationproposed in this review. In early embryos, this mechanism could account for theqPCR-detectable amplification of L1 DNA copy number in time windows when only L1 DNAis replicated (by the noncanonical mechanism) in all or some embryo cells.

These two mechanisms may co-exist, interplay with each other, and be important forthe establishment of an undifferentiated state of a cell. The noncanonical L1 DNAreplication mechanism could serve as an important epigenetic mark that determinesearly replication of L1-rich developmentally regulated EtoL domains, whereas theformation of extrachromosomal L1 DNA copies could be an auxiliary molecular tool insupport of it.

The proposed model implies that the noncanonical L1 DNA replication mechanism isnormally executed in the totipotent and pluripotent cells of early embryos. Itsinitiation and primary specificity of the involved genomic domains is thought to bedetermined by a subset of L1 RNAs carried over into the zygote. The upregulation ofthe expression of FL-L1s at the two-cell stage and the gradual changes of L1expression profiles during preimplantation development are deemed essential for theestablishment of stage-specific gene-expression profiles. Noncanonical replicationcan potentially be triggered in differentiated somatic cells causing celldedifferentiation and transformation, but not pluripotency because, the embryo- andcancer-specific profiles of FL-L1 RNAs are established under the influence ofdifferent factors.

From the standpoint of the proposed model, the unsolved L1-related issues mentionedin the second section of this review can be explained. Specifically, theco-expression of FL-L1 RNA and RT as well as DNA synthesis by reverse transcription,coinciding with the two-fold increase of the L1 DNA copy number in the early embryos,can be biologically explained. The model also explains the different responses ofearly-cleaving embryos and transformed cells to L1 knockdown and RT inhibition,specifically the complete cessation of divisions versus the continued proliferationat a lower rate. In the zygote, and to some extent the cleaving embryo, thespecificity of a set of domains affected by noncanonical L1 DNA replication likelydepends on L1 RNA delivered by the gametes. The degradation of L1 RNA by theL1-specific RNAi at the onset of embryogenesis does not allow the properreprogramming of the genome. The same situation applies to the effect of RTinhibitors at this embryonic stage. The inability of a cell to proceed with properspatial genome repositioning rather than the failure to complete a DNA replicationround can be a consequence of L1 targeting. Those L1 RNPs that are bound to thegenome for DNA replication may be less likely targets than cytoplasmic molecules.This notion is supported by the fact that the targeting of either L1 RNA or RT intransformed cells does not arrest the cells at a distinct point of the cell cycle. Inpoorly differentiated cancer cells synthesizing L1 RNPs, the experimental impedimentsto the putative noncanonical replication of L1s might switch the involved domainsinto a silent state. As a consequence, the gene expression profiles of transformedcells may change to those reminiscent of their normal counterparts. This may or maynot cause a steady transition to normal cell functioning, depending on the“strength” of counteracting transforming factors and what point of thenoncanonical replication mechanism has been targeted. Both of these aspects couldexplain the reinstated transformed phenotypes in a number of RT inhibitor-treatedcancer cell lines after the withdrawal of the inhibitor. A clue as to whydedifferentiated transformed cells reprogram to their normal counterparts but not toother cell types upon the downregulation of L1s comes from the finding oflineage-dependent EtoL domains that are silenced during the specification of lineages [46]. These domains can more easily be reprogrammed back than pluripotency“indicator” EtoL domains. The changes of the replication timing of aportion of lineage-dependent EtoL domains might also be driven by the switch fromnoncanonical to canonical replication of the resident L1s. The majority ofpluripotency “indicator” domains are likely to remain silent in mostcancers, except for embryonal carcinomas and teratocarcinomas, whereaslineage-dependent EtoL domains might be commonly implicated in malignantdedifferentiation. Therefore, their silencing could favor the reprogramming oftransformed cells into the pathway of their original lineage-specificdifferentiation. The proposed model can also explain why the epigenetic barrierestablished on the L1-rich EtoL pluripotency “indicator” domains is verystable. If L1 transcripts carried over by gametes into the zygote and synthesized inthe early embryo under their direct influence do establish early replication of EtoLpluripotency “indicator” domains, the lack of such transcripts can imposea very stable silencing on these domains.

Some additional findings may or may not contradict the proposed model. First, it isnot clear whether the results of cloning experiments fit the model. The model impliesthat the carried-over FL-L1 transcripts delivered by gametes and ORF2p areindispensable to set up the initial 3D genome architecture and replication timingprogram through noncanonical L1 DNA replication. Because the metaphase II oocyte, thecommon recipient used for somatic cell nuclear transfer [159], contains nuclear factors in its cytoplasm, FL-L1 transcripts might beavailable in the ooplasm if not bound to chromatin. ORF2p is present in theepididymal spermatozoa [27]. However, it is not clear whether ORF2p is lacking in the oocyte or asmall amount of ORF2p is dispersed within the cytoplasm and is therefore barelydetectable. The ability of the ooplasm to support reprogramming of transplantednuclei of somatic cells to the totipotent state challenges the significance of L1 RTdelivered by spermatozoa for genome reprogramming at the onset of embryogenesis.

Second, it is unclear whether the density of FL-L1s within the pluripotency“indicator” and certain lineage-dependent EtoL domains is high enough tocontrol tethering and untethering of these domains with regard to the nuclear lamina.The average length of lineage-specific L1s peaks at regions with a GC content of39–40% in the human and mouse genomes [3] suggesting that FL-L1s might accumulate in the pluripotency“indicator” domains, which have exactly the same GC content [45, 51]. In contrast to the mouse genome, which has ~3000 potentially activeFL-L1s [160], the human genome harbors only ~85 retrotranspositionally active copies of~7000 FL-L1s [134]. Nevertheless, some retrotranspositionally inactive FL-L1s might becapable of reverse transcription in vivo. Because the number of FL-L1scapable of reverse transcription remains unclear, it is perplexing whether the subsetof reverse-transcribed FL-L1s is large enough to establish transcriptional competencefor a large cohort of genes.

The hypothesis proposed in this review is testable. The simplest experimental modelto test whether noncanonical L1 DNA replication occurs would be one-cell mouseembryos. Two factors favor this experimental model: the RT-dependent phase of DNAsynthesis in zygotic pronuclei precedes the DNA polymerase-dependent DNA replication,and the time frames of these events have been defined [27]. Two sequential labelings of synthesizing DNA with halogenated nucleotides(e.g., IdU and CldU) during these two phases, and the subsequent visualization oftheir incorporation by fluorescently labeled antibodies on stretched DNA fiberscombined with parallel L1 DNA-specific fluorescence in situ hybridization(FISH), is expected to be informative. The labeling protocol introduced for thesingle-molecule analysis of replicated DNA [161, 162] can be coupled with proper modification of the method of microfluidicextraction and the stretching of DNA from single nuclei [163]. A modification of the method of microfluidic stretching of DNA isrequired to provide better resolved DNA fibers. The lack of data regarding whetherthe RT-dependent phase of DNA synthesis exists in ESCs and certain transformed celllines, whether it overlaps with or precedes the DNA polymerase-dependent phase, andwhether the cells would be able to resume DNA synthesis after the withdrawal ofaphidicolin makes the suitability of the same approach suggested for one-cell embryosuncertain. Additionally, ChIp-seq of either BrdU-labeled nascent DNAs or nascentDNA-ORF2p complexes obtained from aphidicolin-treated ESCs and transformed cell linescould be considered. The knockdown of specific subsets of FL-L1s, and the inhibitionof L1 RT in ESCs and transformed cell lines, followed by analyses of replicationtiming, gene expression, and S/MAR profiles at the genomic scale could clarifywhether activated FL-L1s regulate gene expression through the establishment ofreplication timing and S/MAR profiles.

Prompted by anti-tumor effects of RT inhibitors in experimental models, an attemptwas made to employ nevirapine for the treatment of non-HIV cancer patients in a smallclinical trial [164]. This clinical trial was also based on positive outcomes of RTinhibitor-based treatment regimes for HIV-related tumors, which could partially beattributed to a direct anti-cancer activity of the drugs [165]. However, this approach did not lead to the anticipated result becausenevirapine appeared to be toxic to some non-HIV-infected cancer patients [164] and was perhaps a suboptimal inhibitor of L1 RT. From the standpoint ofthe model proposed here, targeting the L1 RNP-driven process at the RT level might bean ineffective means to obtain the irreversible differentiation of cancer cells evenif highly specific anti-L1 RT drugs are used. Preventing the licensing of sites ofnoncanonical replication might be a more fruitful approach to obtain sustaineddifferentiation of cancer cells. Uncovering the biological significance and themechanism of L1 RT-dependent DNA synthesis would inform the development of highlytargeted anti-cancer therapies and new approaches to control the reprogramming ofdifferentiated cells into iPSCs. In addition, more detail on the sequences of theFL-L1 RNAs forming the full-size L1 RNPs in cancers would open a new avenue in thefield of cancer biomarkers.

Conclusions

Available data demonstrate that several L1-related phenomena cannot be explained withinthe framework of the current retrotransposition-centered L1 paradigm. A novel concept isrequired to explain the nature of massive L1-linked reverse transcription at the onsetof embryogenesis and how abundantly expressed FL-L1 RNA and RT can globally control theepigenetic state of a cell. A revised L1 paradigm should put into focus the possibilityof L1 RT-driven biologically significant processes other than retrotransposition.

A new concept of noncanonical L1 DNA replication that could exist in early embryos,ESCs, and certain types of cancer has been introduced in this article. This proposedmodel links undifferentiated states of a cell, such as totipotency, pluripotency, andregeneration-/cancer-related dedifferentiation to this mechanism. The hithertounexplained phenomena that demonstrate crucial though different outcomes of thedownregulation of L1s and RT in early embryos and cancers can also be explained. First,the proposed model assigns a biological function to upregulated FL-L1s, L1-encodedproteins, and L1-linked reverse transcription. Second, it suggests how the L1 RNP-drivenprocess could potentially result in transcriptional competence of specific domains ofthe genome that harbor genes associated with undifferentiated states. Moreover, themodel demonstrates how the L1 RNP-driven process could integrate with other fundamentalprocesses in the nucleus. Finally, the model shows how the whole system might beregulated in development and dysregulated in cancer.

An important aspect of this novel concept is that it links retrotransposition ofendogenously expressed L1s to the putative noncanonical L1 DNA replication. Evidencesupporting this claim is provided. Endogenous L1 retrotransposition is clearlynon-random, but seems biased to a similar sequence environment. In addition, L1retrotransposition mainly occurs in proliferating undifferentiated embryonic and cancercells, but not in all types of cells where L1s and FL-L1s are abundantly expressed.

Although the current model of DNA replication seems robust, it should be retested inspecific genome locations (distinct FL-L1 sequences) in early embryonic and cancer cellsystems. This is suggested by the failure of the prevailing L1 paradigm to explainseveral important L1-related phenomena and the plausibility of the proposed model ofnoncanonical L1 DNA replication.

Reviewers’ reports

Reviewer 1: Dr. Philip Zegerman, Wellcome Trust/Cancer Research UK GurdonInstitute, University of Cambridge, Cambridge, UK (nominated by Dr. Orly Alter,University of Utah, Salt Lake City, USA)

Understanding the physiological roles of transposable elements is an importantbiological question. This review aims to link the transcription and duplication of L1elements to other cellular processes including replication timing and changes in thechromatin state.

This review would have benefitted from a clearer and more precise analysis of keyexperiments in a defined manner. Instead sweeping conclusions are made from somesparse data e.g. “the data available at this time show no evidence that themassive nuclear reverse transcription occurring in early embryos is DNA replicationindependent.” p.24, yet the aphidicolin experiment in ref 27 clearlydemonstrates the opposite.

Response: Indeed, data used in this review are often insufficient for definiteconclusions. This is not surprising because the issues discussed and questionsraised in this paper have never been addressed experimentally. However, whensparse data accumulate to the necessary threshold, I think it is timely to drawthe attention of the research community to interpretational or conceptual issues.Some findings have been reported by authors as minor details, but they have acertain value when viewed in a new context or linked with other data and,therefore, are worth being included in this review.

The requirement for well-supported conclusions to be based on strong evidence isappropriate for a paper that employs the deductive approach. This review, on thecontrary, is an inductive paper. I recognize the original text contained somegeneralizations that could sound as sweeping conclusions, and thus I havecritically reassessed the text and changed wording in some instances.

I do not agree with the latter comment regarding the text on p. 24. Theconcluding sentence of the section that is cited is taken out of the context. Itsummarized the there main points of the preceding discussion: 1) theinterpretation of aphidicolin-resistant abacavir-sensitive synthesis of DNA byreverse transcription in the zygote as DNA replication independent was based onthe current concept of DNA replication, which may not be comprehensive; 2) therewere some overlooked timing issues related to initiation of DNA synthesis in thezygote, which question the conclusion made in ref. 27; and 3) there were somedrawbacks in the design of the experiments described in ref. 27, which made theexperiments inconclusive in terms of whether the DNA synthesis by reversetranscription in the zygote was DNA replication dependent or independent.

I would like to emphasize that the endurance of a particular scientifichypothesis does not make it an ultimate truth. It is reasonable to interpret newresults on the basis of a particular hypothesis until some data that support newtestable predictions are obtained. I suggest that this is the case with thecurrent concept of DNA replication. To this end, I have strengthened this point inthe paper. I have also made small changes to clarify the point that experiments inref. 27 were inconclusive with respect to their claim that the DNA synthesis byreverse transcription in the zygote was DNA replication independent.

Another example would be the statement “these data suggest that ORCs are highlylikely to bind to G4 structures of L1s”. p.28.

Response: I have clarified this statement by including an additional point fromthe preceding discussion: “these data suggest that ORCs are highly likely tobind to G4 structures of those L1s that tether to the nuclear matrix.” Thelogic underlying this statement is below. About 225,000 active origins (90% of allactive origins) are associated with G4s[80]; however, the number of inactive G4-bound ORCs is not known. L1s are asubstantial source of G4-forming sequences. Given that the human genome contains~516,000 L1s, most of which are truncated at the 5′ (not at the 3′)end[2], and all L1s with intact 3′ UTRs contain a G-forming tract[95], the number of G4-forming sequences can be significantly greater thancurrent estimates of 375,000 [Todd AK et al., Nucleic Acids Res, 2005,33:2901–2907]. Despite the growing interest in the G4-ORC link, no one hasattempted to estimate what portion of G4s associated with ORCs is represented byL1-derived G4s. With so many unknowns, a landmark for future investigations couldbe what we can see in the nucleus. The abundance of L1s among MARs andpreferential colocalization of origins and MARs suggest that chromatin isorganized in such a way that many L1s likely serve as MARs and ORC binding sitesat the same time.

I would urge the author to reassess the review and re-balance the description andinterpretation of the experimentation. This should be married with a considerablereduction (50%) in the length of the review to allow it to be accessible to as wide acommunity of scientists as possible.

Response: I appreciate the concerns with respect to making the manuscript moreaccessible to as wide a community as possible; however, I believe the paper willlose its value to specialists as well as the integral view if the experimental orinterpretative components are so drastically pruned.

This is a multidisciplinary work that integrates experimental data from a numberof fields. Therefore, some introductory information regarding L1 biology,replication timing, etc., are worth inclusion so the paper is accessible to amultidisciplinary readership. Moreover, retaining the experimental data that mightbe considered as non-key facts is important. From the perspective of theintroduced concepts, the whole picture that emerges from the integration of thekey and non-key facts is a more convincing piece of information than severalfindings standing alone. This is important because the concepts and interpretationof certain experimental data are provocative.

This review is not in a narrative style. As mentioned above, this is an inductivepaper that purports a considerable interpretative component. The interpretativeportion of the manuscript is as important as the experimental with respect tointegrating the experimental material, introducing alternative explanations,pointing out issues pertinent to the current L1 paradigm, proposing conceptualchanges, and examining how the available data fit the model. The discussion ofpotential links between the phenomena that have never been thought linked opensnew avenues for research. I believe there is some value in this intellectualcontribution.

In recognition of the length issue, I have deleted a few details such as thenames of genes that changed their expression levels in response to thedownregulation of L1s in A-375 melanoma cells and the concentrations of the RTinhibitors used to reprogram cancer cell lines and to assess their effects onretrotransposition of L1s. Some redrafting has also been done to make someparagraphs more concise.

Quality of written English: Acceptable

Reviewer 2: Dr. I King Jordan, Georgia Institute of Technology, Atlanta, USA

The manuscript on the functional significance of (potentially) non-canonical L1expression and replication by Ekaterina Belan is a provocative mix of a reviewarticle and a hypothesis paper. The author extensively reviews current experimentalevidence on the role of L1 reverse transcription in early embryos and cancer in lightof recent findings on genome regulation, organization and epigenetics. A key tounderstanding the authors approach is the desire to explore novel functional rolesfor L1s that do not fit within the current paradigm of L1 biology, which focusesmainly on retrotransposition dynamics and host genome mechanisms for the repressionof transposition.

The search for a functional role of L1s rests on the author’s notion that sincethe main role of L1s is not the introduction of genomic variation “it islogical to assume that an important function (or functions) of L1s remains to bediscovered.” While this kind of teleological thinking is tempting, one does notneed to invoke a direct function of L1s to explain their existence and abundance inthe genome (or their regulatory anomalies for that matter). As is held by the selfishDNA theory, the existence of such elements can be explained solely by their abilityto out-replicate the genomes in which they reside.

Response: This is a very good point. I have revised the paragraph to includeconsideration of the evolutionary aspect.

Having said that, once having established themselves in their hosts’ genomes,it is almost certainly the case that elements of this kind can have a profound effecton genome function. Accordingly, what the author refers to as the current‘retrotransposition-centered paradigm’ of L1 biology may indeed lead tointerpretations of experimental evidence that are markedly different from thoseoffered in this manuscript. As such, the alternative hypotheses and views proposedhere do seem to cover new ground, are thought provoking, lead to testable predictions(to some extent), and are thus worthy of publication in Biology Direct.

Some of the interpretations of L1 experimental data presented here are likely to becontroversial, particularly to the extent that they differ from interpretationsoffered by the authors of the studies that generated the data. Thus, the paper hasthe potential to generate a substantive response and a potentially interestingdiscussion in the field and/or the literature. To her credit, the author does providespecific experimental tests of her models as they relate to the occurrence ofnon-canonical L1 DNA replication and the role of full-length L1 expression in genomeregulation.

Finally, it is worth noting that the topics covered in this review, and in particularthe experimental tests proposed, could have biomedical relevance with respect to thelink between L1 reverse transcriptase-dependent DNA synthesis and cancer and/or stemcells. A better understanding of this phenomenon could hold promise for thedevelopment of L1 related anti-cancer therapies and/or novel methods for thereprogramming of differentiated cells to pluripotent stem cells.

Quality of written English: Acceptable

Reviewer 3: Dr. Panayiotis (Takis) Benos, University of Pittsburgh, Pittsburgh,USA

This reviewer provided no comments for publication.

Abbreviations

LINE: 

Long interspersed nuclear element

LINE-1(L1): 

Long interspersed nuclearelement-1

FL-L1: 

Full-length L1

RT: 

Reverse transcriptase

EN: 

Endonuclease

RNP: 

Ribonucleoprotein

cDNA: 

Complementary DNA

L1Hs: 

Subfamily of human-specific (fromHomo sapiens) L1 elements (also known as L1PA1)

L1PA2 L1PA3: 

L1PA4, L1PA6,L1PA7: Subfamilies of primate-specific L1 elements

Ta-1: 

Transpositionally activesubfamily of human L1 elements (also known as L1Hs- Ta1)

UTR: 

Untranslated region

ORF: 

Open reading frame

ORF1: 

Open reading frame 1

ORF2: 

Open reading frame 2

ORF1p: 

Openreading frame 1 protein

ORF2p: 

Open reading frame 2 protein

bp: 

Base pair(s)

kb: 

Kilobase pairs

Mb: 

Megabase pairs

siRNA: 

Small interfering RNA

RNAi: 

RNAinterference

S/MAR: 

Scaffold/matrix attachment region

SAR: 

Scaffold attachment region

MAR: 

Matrix attachment region

ERV: 

Endogenous retrovirus

HERV-K: 

Human endogenousretrovirus family

MuERV-L: 

Murine endogenous retrovirus-like element

AML: 

Acutemyeloid leukemia

ICM: 

Inner cell mass

ESC: 

Embryonic stem cell

hESC: 

Human embryonicstem cell

mESC: 

Mouse embryonic stem cell

iPSC: 

Induced pluripotent stem cell

EpiSC: 

Stem cell derived from the epiblast

NPC: 

Neural precursor cell

EtoL: 

Replicationtiming change from early to late S

LtoE: 

Replication timing change from late to earlyS

Xa: 

Active X chromosome

Xi: 

Inactive X chromosome

BrdU: 

5-Bromodeoxyuridine

IdU: 

Iododeoxyuridine

CldU: 

Chlorodeoxyuridine

HCG: 

Human chorionic gonadotropin

qPCR: 

Quantitative real-time polymerase chain reaction

INM: 

Inner nuclear membrane

ORC: 

Origin recognition complex

ORCA: 

ORC-associated protein

HP1: 

Heterochromatin protein1

G4: 

G-quadruplex

H3K9me3: 

Histone H3 trimethylated at lysine 9

H3K9ac: 

Histone H3acetylated at lysine 9

TDP: 

Timing decision point

3D: 

Three-dimensional

wt: 

Wildtype

R loop: 

RNA•DNA displacement loop

ssDNA: 

Single-stranded DNA

dsDNA: 

Double-stranded DNA

FISH: 

Fluorescence in situ hybridization

ChIp-seq: 

Chromatin immunoprecipitation followed by high-throughput DNA sequencing

HIV: 

Humanimmunodeficiency virus.

Declarations

Grant information

This work was not supported by any funds.

Acknowledgments

I thank Prof. Mark Wilkinson and Rena Okrainetz for valuable comments on themanuscript, and Alexander Gogin for his help with the figures.

Authors’ Affiliations

(1)
Genetics Laboratory, Royal University Hospital

References

  1. Babushok DV, Kazazian HH: Progress in understanding the biology of the human mutagen LINE-1. Hum Mutat. 2007, 28: 527-539. 10.1002/humu.20486.PubMed
  2. Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, Baldwin J, Devon K, Dewar K, Doyle M, FitzHugh W, Funke R, Gage D, Harris K, Heaford A, Howland J, Kann L, Lehoczky J, LeVine R, McEwan P, McKernan K, Meldrim J, Mesirov JP, Miranda C, Morris W, Naylor J, Raymond C, Rosetti M, Santos R, Sheridan A, Sougnez C, et al: Initial sequencing and analysis of the human genome. Nature. 2001, 409: 860-921. 10.1038/35057062.PubMed
  3. Waterston RH, Lindblad-Toh K, Birney E, Rogers J, Abril JF, Agarwal P, Agarwala R, Ainscough R, Alexandersson M, An P, Antonarakis SE, Attwood J, Baertsch R, Bailey J, Barlow K, Beck S, Berry E, Birren B, Bloom T, Bork P, Botcherby M, Bray N, Brent MR, Brown DG, Brown SD, Bult C, Burton J, Butler J, Campbell RD, Carninci P, et al: Initial sequencing and comparative analysis of the mouse genome. Nature. 2002, 420: 520-562. 10.1038/nature01262.PubMed
  4. Gibbs RA, Weinstock GM, Metzker ML, Muzny DM, Sodergren EJ, Scherer S, Scott G, Steffen D, Worley KC, Burch PE, Okwuonu G, Hines S, Lewis L, DeRamo C, Delgado O, Dugan-Rocha S, Miner G, Morgan M, Hawes A, Gill R, Celera , Holt RA, Adams MD, Amanatides PG, Baden-Tillson H, Barnstead M, Chin S, Evans CA, Ferriera S, Fosler C, et al: Genome sequence of the Brown Norway rat yields insights into mammalianevolution. Nature. 2004, 428: 493-521.PubMed
  5. Khan H, Smit A, Boissinot S: Molecular evolution and tempo of amplification of human LINE-1 retrotransposonssince the origin of primates. Genome Res. 2006, 16: 78-87.PubMedPubMed Central
  6. Brouha B, Schustak J, Badge RM, Lutz-Prigge S, Farley AH, Moran JV, Kazazian HH: Hot L1s account for the bulk of retrotransposition in the human population. Proc Natl Acad Sci USA. 2003, 100: 5280-5285. 10.1073/pnas.0831042100.PubMedPubMed Central
  7. Beck CR, Collier P, Macfarlane C, Malig M, Kidd JM, Eichler EE, Badge RM, Moran JV: LINE-1 retrotransposition activity in human genomes. Cell. 2010, 141: 1159-1170. 10.1016/j.cell.2010.05.021.PubMedPubMed Central
  8. Beck CR, Garcia-Perez JL, Badge RM, Moran JV: LINE-1 elements in structural variation and disease. Annu Rev Genomics Hum Genet. 2011, 12: 187-215. 10.1146/annurev-genom-082509-141802.PubMedPubMed Central
  9. Scott AF, Schmeckpeper BJ, Abdelrazik M, Comey CT, O’Hara B, Rossiter JP, Cooley T, Heath P, Smith KD, Margolet L: Origin of the human L1 elements: proposed progenitor genes deduced from aconsensus DNA sequence. Genomics. 1987, 1: 113-125. 10.1016/0888-7543(87)90003-6.PubMed
  10. Swergold GD: Identification, characterization, and cell specificity of a human LINE-1promoter. Mol Cell Biol. 1990, 10: 6718-6729.PubMedPubMed Central
  11. Speek M: Antisense promoter of human L1 retrotransposon drives transcription of adjacentcellular genes. Mol Cell Biol. 2001, 21: 1973-1985. 10.1128/MCB.21.6.1973-1985.2001.PubMedPubMed Central
  12. Yang N, Kazazian HH: L1 retrotransposition is suppressed by endogenously encoded small interfering RNAsin human cultured cells. Nat Struct Mol Biol. 2006, 13: 763-771. 10.1038/nsmb1141.PubMed
  13. Wei W, Gilbert N, Ooi SL, Lawler JF, Ostertag EM, Kazazian HH, Boeke JD, Moran JV: Human L1 retrotransposition: cis preference versus transcomplementation. Mol Cell Biol. 2001, 21: 1429-1439. 10.1128/MCB.21.4.1429-1439.2001.PubMedPubMed Central
  14. Packer AI, Manova K, Bachvarova RF: A discrete LINE-1 transcript in mouse blastocysts. Dev Biol. 1993, 157: 281-283. 10.1006/dbio.1993.1133.PubMed
  15. Skowronski J, Fanning TG, Singer MF: Unit-length line-1 transcripts in human teratocarcinoma cells. Mol Cell Biol. 1998, 8: 1385-1397.
  16. Martin SL, Branciforte D: Synchronous expression of LINE-1 RNA and protein in mouse embryonal carcinomacells. Mol Cell Biol. 1993, 13: 5383-5392.PubMedPubMed Central
  17. Kirilyuk A, Tolstonog GV, Damert A, Held U, Hahn S, Lower R, Buschmann C, Horn AV, Traub P, Schumann GG: Functional endogenous LINE-1 retrotransposons are expressed and mobilized in ratchloroleukemia cells. Nucleic Acids Res. 2008, 36: 648-665.PubMedPubMed Central
  18. Branciforte D, Martin SL: Developmental and cell type specificity of LINE-1 expression in mouse testis:implications for transposition. Mol Cell Biol. 1994, 14: 2584-2592. 10.1128/MCB.14.4.2584.PubMedPubMed Central
  19. Kano H, Godoy I, Courtney C, Vetter MR, Gerton GL, Ostertag EM, Kazazian HH: L1 retrotransposition occurs mainly in embryogenesis and creates somaticmosaicism. Genes Dev. 2009, 23: 1303-1312. 10.1101/gad.1803909.PubMedPubMed Central
  20. Okudaira N, Goto M, Yanobu-Takanashi R, Tamura M, An A, Abe Y, Kano S, Hagiwara S, Ishizaka Y, Okamura T: Involvement of retrotransposition of long interspersed nucleotide element-1 inskin tumorigenesis induced by 7,12-dimethylbenz[a]anthracene and12-O-tetradecanoylphorbol-13-acetate. Cancer Sci. 2011, 102: 2000-2006. 10.1111/j.1349-7006.2011.02060.x.PubMed
  21. Solyom S, Ewing AD, Rahrmann EP, Doucet T, Nelson HH, Burns MB, Harris RS, Sigmon DF, Casella A, Erlanger B, Wheelan S, Upton KR, Shukla R, Faulkner GJ, Largaespada DA, Kazazian HH: Extensive somatic L1 retrotransposition in colorectal tumors. Genome Res. 2012, 22: 2328-2338. 10.1101/gr.145235.112.PubMedPubMed Central
  22. McClintock: Discovery and characterization of transposable elements: the collected papers ofBarbara McClintock. 1987, New York: Garland Publishing, Inc,
  23. Jordan IK, Rogozin IB, Glazko GV, Koonin EV: Origin of a substantial fraction of human regulatory sequences from transposableelements. Trends Genet. 2003, 19: 68-72. 10.1016/S0168-9525(02)00006-9.PubMed
  24. Ottaviani D, Lever E, Takousis P, Sheer D: Anchoring the genome. Genome Biol. 2008, 9: 201-10.1186/gb-2008-9-1-201.PubMedPubMed Central
  25. Linnemann AK, Platts AE, Krawetz SA: Differential nuclear scaffold/matrix attachment marks expressed genes. Hum Mol Genet. 2009, 18: 645-654.PubMedPubMed Central
  26. Spadafora C: A reverse transcriptase-dependent mechanism plays central roles in fundamentalbiological processes. Syst Biol Reprod Med. 2008, 54: 11-21. 10.1080/19396360701876815.PubMed
  27. Vitullo P, Sciamanna I, Baiocchi M, Sinibaldi-Vallebona P, Spadafora C: LINE-1 retrotransposon copies are amplified during murine early embryodevelopment. Mol Reprod Dev. 2012, 79: 118-127. 10.1002/mrd.22003.PubMed
  28. Kuo KW, Sheu HM, Huang YS, Leung WC: Expression of transposon LINE-1 is relatively human-specific and function of thetranscripts may be proliferation-essential. Biochem Biophys Res Commun. 1998, 253: 566-570. 10.1006/bbrc.1998.9811.PubMed
  29. Sciamanna I, Landriscina M, Pittoggi C, Quirino M, Mearelli C, Beraldi R, Mattei E, Serafino A, Cassano A, Sinibaldi-Vallebona P, Garaci E, Barone C, Spadafora C: Inhibition of endogenous reverse transcriptase antagonizes human tumor growth. Oncogene. 2005, 24: 3923-3931. 10.1038/sj.onc.1208562.PubMed
  30. Oricchio E, Sciamanna I, Beraldi R, Tolstonog GV, Schumann GG, Spadafora C: Distinct roles for LINE-1 and HERV-K retroelements in cell proliferation,differentiation and tumor progression. Oncogene. 2007, 26: 4226-4233. 10.1038/sj.onc.1210214.PubMed
  31. Beraldi R, Pittoggi C, Sciamanna I, Mattei E, Spadafora C: Expression of LINE-1 retroposons is essential for mutine preimplantationdevelopment. Mol Reprod Dev. 2006, 73: 279-287. 10.1002/mrd.20423.PubMed
  32. Nelson PN, Carnegie PR, Martin J, Davari Ejtehadi H, Hooley P, Roden D, Rowland-Jones S, Warren P, Astley J, Murray PG: Demystified. Human endogenous retroviruses. Mol Pathol. 2003, 56: 11-18. 10.1136/mp.56.1.11.PubMedPubMed Central
  33. Evsikov AV, de Vries WN, Peaston AE, Radford EE, Fancher KS, Chen FH, Blake JA, Bult CJ, Latham KE, Solter D, Knowles BB: Systems biology of the 2-cell mouse embryo. Cytogenet Genome Res. 2004, 105: 240-250. 10.1159/000078195.PubMed
  34. Peaston AE, Knowles BB, Hutchison KW: Genome plasticity in the mouse oocyte and early embryo. Biochem Soc Trans. 2007, 35: 618-622. 10.1042/BST0350618.PubMed
  35. Pittoggi C, Sciamanna I, Mattei E, Beraldi R, Lobascio AM, Mai A, Quaglia MG, Lorenzini R, Spadafora C: Role of endogenous reverse transcriptase in murine early embryo development. Mol Reprod Dev. 2003, 66: 225-236. 10.1002/mrd.10349.PubMed
  36. Spadafora C: Endogenous reverse transcriptase: a mediator of cell proliferation anddifferentiation. Cytogenet Genome Res. 2004, 105: 346-350. 10.1159/000078207.PubMed
  37. Landriscina M, Fabiano A, Altamura S, Bagala C, Piscazzi A, Cassano A, Spadafora C, Giorgino F, Barone C, Cignarelli M: Reverse transcriptase inhibitors down-regulate cell proliferation invitro and in vivo and restore thyrotropin signaling and iodineuptake in human thyroid anaplastic carcinoma. J Clin Endocrinol Metab. 2005, 90: 5663-5671. 10.1210/jc.2005-0367.PubMed
  38. Mangiacasale R, Pittoggi C, Sciamanna I, Careddu A, Mattei E, Lorenzini R, Travaglini L, Landriscina M, Barone C, Nervi C, Lavia P, Spadafora C: Exposure of normal and transformed cells to nevirapine, a reverse transcriptaseinhibitor, reduces cell growth and promotes differentiation. Oncogene. 2003, 22: 2750-2761. 10.1038/sj.onc.1206354.PubMed
  39. Jones RB, Garrison KE, Wong JC, Duan EH, Nixon DF, Ostrowski MA: Nucleoside analogue reverse transcriptase inhibitors differentially inhibit humanLINE-1 retrotransposition. PLoS One. 2008, 3: e1547-10.1371/journal.pone.0001547.PubMedPubMed Central
  40. Dai L, Huang Q, Boeke JD: Effect of reverse transcriptase inhibitors on LINE-1 and Ty1 reverse transcriptaseactivities and on LINE-1 retrotransposition. BMC Biochem. 2011, 12: 18-10.1186/1471-2091-12-18.PubMedPubMed Central
  41. Kigami D, Minami N, Takayama H, Imai H: MuERV-L is one of the earliest transcribed genes in mouse one-cell embryos. Biol Reprod. 2003, 68: 651-654.PubMed
  42. Ramsoondar J, Vaught T, Ball S, Mendicino M, Monahan J, Jobst P, Vance A, Duncan J, Wells K, Ayares D: Production of transgenic pigs that express porcine endogenous retrovirus smallinterfering RNAs. Xenotransplantation. 2009, 16: 164-180. 10.1111/j.1399-3089.2009.00525.x.PubMed
  43. Prak ET, Kazazian HH: Mobile elements in the human genome. Nat Rev Genet. 2000, 1: 134-144. 10.1038/35038572.PubMed
  44. Whitelaw E, Martin DIK: Retrotransposons as epigenetic mediators of phenotypic variation in mammals. Nat Genet. 2001, 27: 361-365. 10.1038/86850.PubMed
  45. Hiratani I, Ryba T, Itoh M, Yokochi T, Schwaiger M, Chang CW, Lyou Y, Townes TM, Schübeler D, Gilbert DM: Global reorganization of replication domains during embryonic stem celldifferentiation. PLoS Biol. 2008, 6: e245-10.1371/journal.pbio.0060245.PubMedPubMed Central
  46. Hiratani I, Ryba T, Itoh M, Rathjen J, Kulik M, Papp B, Fussner E, Bazett-Jones DP, Plath K, Dalton S, Rathjen PD, Gilbert DM: Genome-wide dynamics of replication timing revealed by in vitro models ofmouse embryogenesis. Genome Res. 2010, 20: 155-169. 10.1101/gr.099796.109.PubMedPubMed Central
  47. Hand R: Eucaryotic DNA: organization of the genome for replication. Cell. 1978, 15: 317-325. 10.1016/0092-8674(78)90001-6.PubMed
  48. Farkash-Amar S, Lipson D, Polten A, Goren A, Helmstetter C, Yakhini Z, Simon I: Global organization of replication time zones of the mouse genome. Genome Res. 2008, 18: 1562-1570. 10.1101/gr.079566.108.PubMedPubMed Central
  49. Ryba T, Hiratani I, Lu J, Itoh M, Kulik M, Zhang J, Schulz TC, Robins AJ, Dalton S, Gilbert DM: Evolutionarily conserved replication timing profiles predict long-range chromatininteractions and distinguish closely related cell types. Genome Res. 2010, 20: 761-770. 10.1101/gr.099655.109.PubMedPubMed Central
  50. Takebayashi SI, Ryba T, Gilbert DM: Developmental control of replication timing defines a new breed of chromosomaldomains with a novel mechanism of chromatin unfolding. Nucleus. 2012, 3: 1-8. 10.4161/nucl.19362.
  51. Hiratani I, Takebayashi S, Lu J, Gilbert DM: Replication timing and transcriptional control: beyond cause and effect –part II. Curr Opin Genet Dev. 2009, 19: 142-149. 10.1016/j.gde.2009.02.002.PubMedPubMed Central
  52. Cornacchia D, Dileep V, Quivy JP, Foti R, Tili F, Santarella-Mellwig R, Antony C, Almouzni G, Gilbert DM, Buonomo SB: Mouse Rif1 is a key regulator of the replication-timing programme in mammaliancells. EMBO J. 2012, 31: 3678-3690. 10.1038/emboj.2012.214.PubMedPubMed Central
  53. Yamazaki S, Ishii A, Kanoh Y, Oda M, Nishito Y, Masai H: Rif1 regulates the replication timing domains on the human genome. EMBO J. 2012, 31: 3667-3677. 10.1038/emboj.2012.180.PubMedPubMed Central
  54. Adams IR, McLaren A: Identification and characterisation of mRif1: a mouse telomere-associated proteinhighly expressed in germ cells and embryo-derived pluripotent stem cells. Dev Dyn. 2004, 229: 733-744. 10.1002/dvdy.10471.PubMed
  55. Loh YH, Wu Q, Chew JL, Vega VB, Zhang W, Chen X, Bourque G, George J, Leong B, Liu J, Wong KY, Sung KW, Lee CW, Zhao XD, Chiu KP, Lipovich L, Kuznetsov VA, Robson P, Stanton LW, Wei CL, Ruan Y, Lim B, Ng HH: The Oct4 and Nanog transcription network regulates pluripotency in mouse embryonicstem cells. Nat Genet. 2006, 38: 431-440. 10.1038/ng1760.PubMed
  56. Garcia-Perez JL, Marchetto MCN, Muotri AR, Coufal NG, Gage FH, O’Shea KS, Moran JV: LINE-1 retrotransposition in human embryonic stem cells. Hum Mol Genet. 2007, 16: 1569-1577. 10.1093/hmg/ddm105.PubMed
  57. Macia A, Muñoz-Lopez M, Cortes JL, Hastings RK, Morell S, Lucena-Aguilar G, Marchal JA, Badge RM, Garcia-Perez JS: Epigenetic control of retrotransposon expression in human embryonic stem cells. Mol Cell Biol. 2011, 31: 300-316. 10.1128/MCB.00561-10.PubMedPubMed Central
  58. Ahmed K, Dehghani H, Rugg-Gunn P, Fussner E, Rossant J, Bazett-Jones DP: Global chromatin architecture reflects pluripotency and lineage commitment inearly mouse embryo. PLoS One. 2010, 5: e10531-10.1371/journal.pone.0010531.PubMedPubMed Central
  59. Ferreira J, Carmo-Fonseca M: Genome replication in early mouse embryos follows a defined temporal and spatialorder. J Cell Sci. 1997, 110: 889-897.PubMed
  60. Mitalipov S, Wolf D: Totipotency, pluripotency and nuclear reprogramming. Adv Biochem Eng Biotechnol. 2009, 114: 185-199.PubMedPubMed Central
  61. Penzkofer T, Dandekar T, Zemojtel T: L1Base: from functional annotation to prediction of active LINE-1 elements. Nucleic Acids Res. 2005, 33: 498-500.
  62. da Costa LF: Return of de-differentiation: why cancer is a developmental disease. Curr Opin Oncol. 2001, 13: 58-62. 10.1097/00001622-200101000-00012.PubMed
  63. Ryba T, Battaglia D, Chang BH, Shirley JW, Buckley Q, Pope BD, Devidas M, Druker BJ, Gilbert DM: Abnormal developmental control of replication timing domains in pediatric acutelymphoblastic leukemia. Genome Res. 2012, 22: 1833-1844. 10.1101/gr.138511.112.PubMedPubMed Central
  64. Andrews PW: From teratocarcinomas to embryonic stem cells. Philos Trans R Soc Lond B Biol Sci. 2002, 357: 405-417. 10.1098/rstb.2002.1058.PubMedPubMed Central
  65. Mintz B, Illmensee K: Normal genetically mosaic mice produced from malignant teratocarcinoma cells. Proc Nat Acad Sci USA. 1975, 72: 3585-3589. 10.1073/pnas.72.9.3585.PubMedPubMed Central
  66. Chow JC, Ciaudo C, Fazzari MJ, Mise N, Servant N, Glass JL, Attreed M, Avner P, Wutz A, Barillot E, Greally JM, Voinnet O, Heard E: LINE-1 activity in facultative heterochromatin formation during X chromosomeinactivation. Cell. 2010, 141: 956-969. 10.1016/j.cell.2010.04.042.PubMed
  67. Chueh AC, Northrop EL, Brettingham-Moore KH, Choo KH, Wong LH: LINE retrotransposon RNA is an essential structural and functional epigeneticcomponent of a core neocentromeric chromatin. PLoS Genet. 2009, 5: e1000354-10.1371/journal.pgen.1000354.PubMedPubMed Central
  68. Martin SL: Ribonucleoprotein particles with LINE-1 RNA in mouse embryonal carcinoma cells. Mol Cell Biol. 1991, 11: 4804-4807.PubMedPubMed Central
  69. Hohjoh H, Singer MF: Cytoplasmic ribonucleoprotein complexes containing human LINE-1 protein andRNA. EMBO J. 1996, 15: 630-639.PubMedPubMed Central
  70. Doucet AJ, Hulme AE, Sahinovic E, Kulpa DA, Moldovan JB, Kopera HC, Athanikar JN, Hasnaoui M, Bucheton A, Moran JV, Gilbert N: Characterization of LINE-1 ribonucleoprotein particles. PLoS Genet. 2010, 6:
  71. Kulpa DA, Moran JV: Ribonucleoprotein particle formation is necessary but not sufficient for LINE-1retrotransposition. Hum Mol Genet. 2005, 14: 3237-3248. 10.1093/hmg/ddi354.PubMed
  72. Kulpa DA, Moran JV: Cis-preferential LINE-1 reverse transcriptase activity in ribonucleoproteinparticles. Nat Struct Mol Biol. 2006, 13: 655-660. 10.1038/nsmb1107.PubMed
  73. Pavlov YI, Shcherbakova PV, Rogozin IB: Roles of DNA polymerases in replication, repair, and recombination ineukaryotes. Int Rev Cytol. 2006, 255: 41-132.PubMed
  74. Blackburn EH, Greider CW, Szostak JW: Telomeres and telomerase: the path from maize, Tetrahymena and yeast to humancancer and aging. Nat Med. 2006, 12: 1133-1138. 10.1038/nm1006-1133.PubMed
  75. Kopera HC, Moldovan JB, Morrish TA, Garcia-Perez JL, Moran VJ: Similarities between long interspersed element-1 (LINE-1) reverse transcriptaseand telomerase. Proc Natl Acad Sci U S A. 2011, 108: 20345-20350. 10.1073/pnas.1100275108.PubMedPubMed Central
  76. Howlett SK, Bolton VN: Sequence and regulation of morphological and molecular events during the firstcell cycle of mouse embryogenesis. J Embryol Exp Morphol. 1985, 87: 175-206.PubMed
  77. Luthardt FW, Donahue RP: Pronuclear DNA synthesis in mouse eggs. An autoradiographic study. Exp Cell Res. 1973, 82: 143-151. 10.1016/0014-4827(73)90256-5.PubMed
  78. Bouniol-Baly C, Nguyen E, Besombes D, Debey P: Dynamic organization of DNA replication in one-cell mouse embryos: relationship totranscriptional activation. Exp Cell Res. 1997, 236: 201-211. 10.1006/excr.1997.3708.PubMed
  79. Schabronath J, Gärtner K: Paternal influence on timing of pronuclear DNA synthesis in naturally ovulated andfertilized mouse eggs. Biol Reprod. 1988, 38: 744-749. 10.1095/biolreprod38.4.744.PubMed
  80. Besnard E, Babled A, Lapasset L, Milhavet O, Parrinello H, Dantec C, Marin JM, Lemaitre JM: Unraveling cell type-specific and reprogrammable human replication originsignatures associated with G-quadruplex consensus motifs. Nat Struct Mol Biol. 2012, 19: 837-844. 10.1038/nsmb.2339.PubMed
  81. Gilbert DM: In search of the holy replicator. Nat Rev Mol Cell Biol. 2004, 5: 848-855. 10.1038/nrm1495.PubMedPubMed Central
  82. Gilbert DM, Miyazawa H, DePamphilis ML: Site-specific initiation of DNA replication in Xenopus egg requires nuclearstructure. Mol Cell Biol. 1995, 15: 2942-2954.PubMedPubMed Central
  83. Ohta S, Tatsumi Y, Fujita M, Tsurimoto T, Obuse C: The ORC1 cycle in human cells: II. Dynamic changes in the human ORC complex duringthe cell cycle. J Biol Chem. 2003, 278: 41535-41540. 10.1074/jbc.M307535200.PubMed
  84. Bell SP: The origin recognition complex: from simple origins to complex functions. Genes Dev. 2002, 16: 659-672. 10.1101/gad.969602.PubMed
  85. Sasaki T, Gilbert DM: The many faces of the origin recognition complex. Curr Opin Cell Biol. 2007, 19: 337-343. 10.1016/j.ceb.2007.04.007.PubMed
  86. Prasanth SG, Shen Z, Prasanth KV, Stillman B: Human origin recognition complex is essential for HP1 binding to chromatin andheterochromatin organization. Proc Natl Acad Sci U S A. 2010, 107: 15093-15098. 10.1073/pnas.1009945107.PubMedPubMed Central
  87. Chakraborty A, Shen Z, Prasanth SG: “ORCanization” on heterochromatin: linking DNA replication initiationto chromatin organization. Epigenetics. 2011, 6: 665-670. 10.4161/epi.6.6.16179.PubMed
  88. Boulikas T: Common structural features of replication origins in all life forms. J Cell Biochem. 1996, 60: 297-316. 10.1002/(SICI)1097-4644(19960301)60:3<297::AID-JCB2>3.0.CO;2-R.PubMed
  89. Courbet S, Gay S, Arnoult N, Wronka G, Anglana M, Brison O, Debatisse M: Replication fork movement sets chromatin loop size and origin choice in mammaliancells. Nature. 2008, 455: 557-560. 10.1038/nature07233.PubMed
  90. Bode J, Winkelmann S, Götze S, Spiker S, Tsutsui K, Bi CAKP, Benham C: Correlations between scaffold/matrix attachment region (S/MAR) binding activityand DNA duplex destabilization energy. J Mol Biol. 2006, 358: 597-613. 10.1016/j.jmb.2005.11.073.PubMed
  91. Linnemann AK, Krawetz SA: Silencing by nuclear matrix attachment distinguishes cell-type specificity:association with increased proliferation capacity. Nucleic Acids Res. 2009, 37: 2779-2788. 10.1093/nar/gkp135.PubMedPubMed Central
  92. Berezney R, Dubey DD, Huberman JA: Heterogeneity of eukaryotic replicons, replicon clusters, and replication foci. Chromosoma. 2000, 108: 471-484. 10.1007/s004120050399.PubMed
  93. Segal E, Widom J: What controls nucleosome positions?. Trends Genet. 2009, 25: 335-343. 10.1016/j.tig.2009.06.002.PubMedPubMed Central
  94. Lubelsky Y, Sasaki T, Kuipers MA, Lucas I, Le Beau MM, Carignon S, Debatisse M, Prinz JA, Dennis JH, Gilbert DM: Pre-replication complex proteins assemble at regions of low nucleosome occupancywithin the Chinese hamster dihydrofolate reductase initiation zone. Nucleic Acids Res. 2011, 39: 3141-3155. 10.1093/nar/gkq1276.PubMedPubMed Central
  95. Howell R, Usdin K: The ability to form intrastrand tetraplexes is an evolutionarily conserved featureof the 3′ end of L1 retrotransposons. Mol Biol Evol. 1997, 14: 144-155. 10.1093/oxfordjournals.molbev.a025747.PubMed
  96. Smit AF, Tóth G, Riggs AD, Jurka J: Ancestral, mammalian-wide subfamilies of LINE-1 repetitive sequences. J Mol Biol. 1995, 246: 401-417. 10.1006/jmbi.1994.0095.PubMed
  97. Halder K, Halder R, Chowdhury S: Genome-wide analysis predicts DNA structural motifs as nucleosome exclusionsignals. Mol Biosyst. 2009, 5: 1703-1712. 10.1039/b905132e.PubMed
  98. Minc E, Courvalin JC, Buendia B: HP1gamma associates with euchromatin and heterochromatin in mammalian nuclei andchromosomes. Cytogenet Cell Genet. 2000, 90: 279-284. 10.1159/000056789.PubMed
  99. Brown JP, Bullwinkel J, Baron-Lühr B, Billur M, Schneider P, Winking H, Singh PB: HP1gamma function is required for male germ cell survival and spermatogenesis. Epigenetics Chromatin. 2010, 3: 9-10.1186/1756-8935-3-9.PubMedPubMed Central
  100. Ye Q, Worman HJ: Interaction between an integral protein of the nuclear envelope inner membrane andhuman chromodomain proteins homologous to Drosophila HP1. J Biol Chem. 1996, 271: 14653-14656. 10.1074/jbc.271.25.14653.PubMed
  101. Kourmouli N, Theodoropoulos PA, Dialynas G, Bakou A, Politou AS, Cowell IG, Singh PB, Georgatos SD: Dynamic association of heterochromatin protein 1 with the nuclear envelope. EMBO J. 2000, 19: 6558-6568. 10.1093/emboj/19.23.6558.PubMedPubMed Central
  102. Martens JH, O’Sullivan RJ, Braunschweig U, Opravil S, Radolf M, Steinlein P, Jenuwein T: The profile of repeat-associated histone lysine methylation states in the mouseepigenome. EMBO J. 2005, 24: 800-812. 10.1038/sj.emboj.7600545.PubMedPubMed Central
  103. Mikkelsen TS, Ku M, Jaffe DB, Issac B, Lieberman E, Giannoukos G, Alvarez P, Brockman W, Kim TK, Koche RP, Lee W, Mendenhall E, O’Donovan A, Presser A, Russ C, Xie X, Meissner A, Wernig M, Jaenisch R, Nusbaum C, Lander ES, Bernstein BE: Genome-wide maps of chromatin state in pluripotent and lineage-committed cells. Nature. 2007, 448: 553-560. 10.1038/nature06008.PubMedPubMed Central
  104. Teneng I, Montoya-Durango DE, Quertermous JL, Lacy ME, Ramos KS: Reactivation of L1 retrotransposon by benzo(a)pyrene involves complex genetic andepigenetic regulation. Epigenetics. 2011, 6: 355-367. 10.4161/epi.6.3.14282.PubMedPubMed Central
  105. Barski A, Cuddapah S, Cui K, Roh T-Y, Schones DE, Wang Z, Wei G, Chepelev I, Zhao K: High-resolution profiling of histone methylations in the human genome. Cell. 2007, 129: 823-837. 10.1016/j.cell.2007.05.009.PubMed
  106. Ruault M, Dubarry M, Taddei A: Re-positioning genes to the nuclear envelope in mammalian cells: impact ontranscription. Trends Genet. 2008, 24: 574-581. 10.1016/j.tig.2008.08.008.PubMed
  107. Zhang J, Xu F, Hashimshony T, Keshet I, Cedar H: Establishment of transcriptional competence in early and late S phase. Nature. 2002, 420: 198-202. 10.1038/nature01150.PubMed
  108. Lande-Diner L, Zhang J, Cedar H: Shifts in replication timing actively affect histone acetylation during nucleosomereassembly. Mol Cell. 2009, 34: 767-774. 10.1016/j.molcel.2009.05.027.PubMedPubMed Central
  109. McNairn AJ, Gilbert DM: Epigenomic replication: linking epigenetics to DNA replication. Bioessays. 2003, 25: 647-656. 10.1002/bies.10305.PubMed
  110. Luo L, Gassman KL, Petell LM, Wilson CL, Bewersdorf J, Shopland LS: The nuclear periphery of embryonic stem cells is a transcriptionally permissiveand repressive compartment. J Cell Sci. 2009, 122: 3729-3737. 10.1242/jcs.052555.PubMedPubMed Central
  111. Easwaran HP, Baylin SB: Role of nuclear architecture in epigenetic alterations in cancer. Cold Spring Harb Symp Quant Biol. 2010, 75: 507-515. 10.1101/sqb.2010.75.031.PubMedPubMed Central
  112. O’Keefe RT, Henderson SC, Spector DL: Dynamic organization of DNA replication in mammalian cell nuclei: spatially andtemporally defined replication of chromosome-specific alpha-satellite DNAsequences. J Cell Biol. 1992, 116: 1095-1110. 10.1083/jcb.116.5.1095.PubMed
  113. Zink D, Fischer AH, Nickerson JA: Nuclear structure in cancer cells. Nat Rev Cancer. 2004, 4: 677-687. 10.1038/nrc1430.PubMed
  114. Harris CR, Normart R, Yang Q, Stevenson E, Haffty BG, Ganesan S, Cordon-Cardo C, Levine AJ, Tang LH: Association of nuclear localization of a long interspersed nuclear element-1protein in breast tumors with poor prognostic outcomes. Genes Cancer. 2010, 1: 115-124. 10.1177/1947601909360812.PubMedPubMed Central
  115. Gilbert DM: Cell fate transitions and the replication timing decision point. J Cell Biol. 2010, 191: 899-903. 10.1083/jcb.201007125.PubMedPubMed Central
  116. Pittoggi C, Renzi L, Zaccagnini G, Cimini D, Degrassi F, Giordano R, Magnano AR, Lorenzini R, Lavia P, Spadafora C: A fraction of mouse sperm chromatin is organized in nucleosomal hypersensitivedomains enriched in retroposon DNA. J Cell Sci. 1999, 112: 3537-3548.PubMed
  117. Pittoggi C, Zaccagnini G, Giordano R, Magnano AR, Baccetti B, Lorenzini R, Spadafora C: Nucleosomal domains of mouse spermatozoa chromatin as potential sites forretroposition and foreign DNA integration. Mol Reprod Dev. 2000, 56: 248-251. 10.1002/(SICI)1098-2795(200006)56:2+<248::AID-MRD7>3.0.CO;2-V.PubMed
  118. Alisch RS, Garcia-Perez JL, Muotri AR, Gage FH, Moran JV: Unconventional translation of mammalian LINE-1 retrotransposons. Genes Dev. 2006, 20: 210-224. 10.1101/gad.1380406.PubMedPubMed Central
  119. Callahan KE, Hickman AB, Jones CE, Ghirlando R, Furano AV: Polymerization and nucleic acid-binding properties of human L1 ORF1 protein. Nucleic Acids Res. 2012, 40: 813-827. 10.1093/nar/gkr728.PubMedPubMed Central
  120. Martin SL, Bushman FD: Nucleic acid chaperone activity of the ORF1 protein from the mouse LINE-1retrotransposon. Mol Cell Biol. 2001, 21: 467-475. 10.1128/MCB.21.2.467-475.2001.PubMedPubMed Central
  121. Martin SL: Nucleic acid chaperone properties of ORF1p from the non-LTR retrotransposon,LINE-1. RNA Biol. 2010, 7: 706-711. 10.4161/rna.7.6.13766.PubMedPubMed Central
  122. Hohjoh H, Singer MF: Sequence-specific single-strand RNA binding protein encoded by the human LINE-1retrotransposon. EMBO J. 1997, 16: 6034-6043. 10.1093/emboj/16.19.6034.PubMedPubMed Central
  123. Cost GJ, Boeke JD: Targeting of human retrotransposon integration is directed by the specificity ofthe L1 endonuclease for regions of unusual DNA structure. Biochemistry. 1998, 37: 18081-18093. 10.1021/bi981858s.PubMed
  124. Cost GJ, Feng Q, Jacquier A, Boeke JD: Human L1 element target-primed reverse transcription in vitro. EMBO J. 2002, 21: 5899-5910. 10.1093/emboj/cdf592.PubMedPubMed Central
  125. Martin SL, Cruceanu M, Branciforte D, Wai-Iun Li P, Kwok SC, Hodges RS, Williams MC: LINE-1 retrotransposition requires the nucleic acid chaperone activity of the ORF1protein. J Mol Biol. 2005, 348: 549-561. 10.1016/j.jmb.2005.03.003.PubMed
  126. Martin SL, Bushman D, Wang F, Li PW, Walker A, Cummiskey J, Brancifirte D, Williams MC: A single amino acid substitution in ORF1 dramatically decreases L1retrotransposition and provides insight into nucleic acid chaperone activity. Nucleic Acids Res. 2008, 36: 5845-5854. 10.1093/nar/gkn554.PubMedPubMed Central
  127. An W, Han JS, Wheelan SJ, Davis ES, Coombes CE, Ye P, Triplett C, Boeke JD: Active retrotransposition by a synthetic L1 element in mice. Proc Natl Acad Sci USA. 2006, 103: 18662-18667. 10.1073/pnas.0605300103.PubMedPubMed Central
  128. Jurka J, Kapitonov VV: Sectorial mutagenesis by transposable elements. Genetica. 1999, 107: 239-248. 10.1023/A:1003989620068.PubMed
  129. Pavlícek A, Jabbari K, Paces J, Paces V, Hejnar JV, Bernardi G: Similar integration but different stability of Alus and LINEs in the humangenome. Gene. 2001, 276: 39-45. 10.1016/S0378-1119(01)00645-X.PubMed
  130. Boissinot S, Entezam A, Young L, Munson PJ, Furano AV: The insertional history of an active family of L1 retrotransposons in humans. Genome Res. 2004, 14: 1221-1231. 10.1101/gr.2326704.PubMedPubMed Central
  131. Abrusán G, Krambeck HJ: The distribution of L1 and Alu retroelements in relation to GC content on humansex chromosomes is consistent with the ectopic recombination model. J Mol Evol. 2006, 63: 484-492. 10.1007/s00239-005-0275-0.PubMed
  132. Graham T, Boissinot S: The genomic distribution of L1 elements: the role of insertion bias and naturalselection. J Biomed Biotechnol. 2006, 2006: 75327-PubMedPubMed Central
  133. Jurka J, Kohany O, Pavlicek A, Kapitonov VV, Jurka MV: Duplication, coclustering, and selection of human Alu retrotransposons. Proc Natl Acad Sci U S A. 2004, 101: 1268-1272. 10.1073/pnas.0308084100.PubMedPubMed Central
  134. Szak ST, Pickeral OK, Makalowski W, Boguski MS, Landsman D, Boeke JD: Molecular archeology of L1 insertions in the human genome. Genome Biol. 2002, 3: research0052
  135. Yagil G: Paranemic structures of DNA and their role in DNA unwinding. Crit Rev Biochem Mol Biol. 1991, 26: 475-559. 10.3109/10409239109086791.PubMed
  136. van den Hurk JAJM, Meij IC, Seleme MC, Kano H, Nikopoulos K, Hoefsloot LH, Sistermans EA, de Wijs IJ, Mukhopadhyay A, Plomp AS, de Jong PT, Kazazian HH, Cremers FP: L1 retrotransposition can occur early in human embryonic development. Hum Mol Genet. 2007, 16: 1587-1592. 10.1093/hmg/ddm108.PubMed
  137. Piskareva O, Schmatchenko V: DNA polymerization by the reverse transcriptase of the human L1 retrotransposon onits own template in vitro. FEBS Lett. 2006, 580: 661-668. 10.1016/j.febslet.2005.12.077.PubMed
  138. Jurka J: Sequence patterns indicate an enzymatic involvement in integration of mammalianretroposons. Proc Natl Acad Sci USA. 1997, 94: 1872-1877. 10.1073/pnas.94.5.1872.PubMedPubMed Central
  139. Belancio VP, Whelton M, Deininger P: Requirements for polyadenilation at the 3′ end of LINE-1 elements. Gene. 2007, 390: 98-107. 10.1016/j.gene.2006.07.029.PubMed
  140. Belancio VP, Roy-Engel AM, Pochampally RR, Deininger P: Somatic expression of LINE-1 elements in human tissues. Nucleic Acids Res. 2010, 38: 3909-3922. 10.1093/nar/gkq132.PubMedPubMed Central
  141. Perepelitsa-Belancio V, Deininger P: RNA truncation by premature polyadenylation attenuates human mobile elementactivity. Nat Genet. 2003, 35: 363-366. 10.1038/ng1269.PubMed
  142. Belancio VP, Hedges DJ, Deininger P: LINE-1 RNA splicing and influences on mammalian gene expression. Nucleic Asids Res. 2006, 34: 1512-1521. 10.1093/nar/gkl027.
  143. Ergün S, Buschmann C, Heukeshoven J, Dammann K, Schnieders F, Lauke H, Chalajour F, Kilic N, Strätling WH, Schumann GG: Cell type-specific expression of LINE-1 open reading frames 1 and 2 in fetal andadult human tissues. J Biol Chem. 2004, 279: 27753-27763. 10.1074/jbc.M312985200.PubMed
  144. Trelogan SA, Martin SL: Tightly regulated, developmentally specific expression of the first open readingframe from LINE-1 during mouse embryogenesis. Proc Natl Acad Sci USA. 1995, 92: 1520-1524. 10.1073/pnas.92.5.1520.PubMedPubMed Central
  145. van der Heijden GW, Bortvin A: Transient relaxation of transposon silencing at the onset of mammalian meiosis. Epigenetics. 2009, 4: 76-79. 10.4161/epi.4.2.7783.PubMed
  146. Soper SFC, van der Heijden GW, Hardiman TC, Goodheart M, Martin SL, de Boer P, Bortvin A: Mouse maelstrom, a component of nuage, is essential for spermatogenesis andtransposon repression in meiosis. Dev Cell. 2008, 15: 285-297. 10.1016/j.devcel.2008.05.015.PubMedPubMed Central
  147. Peaston AE, Evsikov AV, Graber JH, de Vries WN, Holbrook AE, Solter D, Knowles BB: Retrotransposons regulate host genes in mouse oocytes and preimplantationembryos. Dev Cell. 2004, 7: 597-606. 10.1016/j.devcel.2004.09.004.PubMed
  148. Wissing S, Muñoz-Lopez M, Macia A, Yang Z, Montano M, Collins W, Garcia-Perez JL, Moran JV, Greene WC: Reprogramming somatic cells into iPS cells activates LINE-1 retroelementmobility. Hum Mol Genet. 2012, 21: 208-218. 10.1093/hmg/ddr455.PubMedPubMed Central
  149. Coufal NG, Garcia-Perez JL, Peng GE, Yeo GW, Mu Y, Lovci MT, Morell M, O’Shea KS, Moran JV, Gage FH: L1 retrotransposition in human neural progenitor cells. Nature. 2009, 460: 1127-1131. 10.1038/nature08248.PubMedPubMed Central
  150. Skotheim RI, Lind GE, Monni O, Nesland JM, Abeler VM, Fosså SD, Duale N, Brunborg G, Kallioniemi O, Andrews PW, Lothe RA: Differentiation of human embryonal carcinomas in vitro and invivo reveals expression profiles relevant to normal development. Cancer Res. 2005, 65: 5588-5598. 10.1158/0008-5472.CAN-05-0153.PubMed
  151. Deragon JM, Sinnett D, Labuda D: Reverse transcriptase activity from human embryonal carcinoma cells NTera2D1. EMBO J. 1990, 9: 3363-3368.PubMedPubMed Central
  152. Asch HL, Eliacin E, Fanning TG, Connolly JL, Bratthauer G, Asch BB: Comparative expression of the LINE-1 p40 protein in human breast carcinomas andnormal breast tissues. Oncol Res. 1996, 8: 239-247.PubMed
  153. Florl AR, Löwer R, Schmitz-Dräger BJ, Schulz WA: DNA methylation and expression of LINE-1 and HERV-K provirus sequences inurothelial and renal cell carcinomas. Br J Cancer. 1999, 80: 1312-1321. 10.1038/sj.bjc.6690524.PubMedPubMed Central
  154. Roman-Gomez J, Jimenez-Velasco A, Agirre X, Cervantes F, Sanchez J, Garate L, Barrios M, Castillejo JA, Navarro G, Colomer D, Prosper F, Heiniger A, Torres A: Promoter hypomethylation of the LINE-1 transposable elements activatessense/antisense transcription and marks the progression of chronic myeloidleukemia. Oncogene. 2005, 24: 7213-7223. 10.1038/sj.onc.1208866.PubMed
  155. Zhu W, Kuo D, Nathanson J, Satoh A, Pao GM, Yeo GW, Bryant SV, Voss SR, Gardiner DM, Hunter T: Retrotransposon long interspersed nucleotide element-1 (LINE-1) is activatedduring salamander limb regeneration. Dev Growth Differ. 2012, 54: 673-685. 10.1111/j.1440-169X.2012.01368.x.PubMedPubMed Central
  156. Bliakher LI: Proceedings of the Institute of Experimental Morphogenesis: vol 5. Investigation of repeated regeneration in amphibians. 1936, Moscow, in Russian
  157. Han JS, Shao S: Circular retrotransposition products generated by a LINE retrotransposon. Nucleic Acids Res. 2012, 40: 10866-10877. 10.1093/nar/gks859.PubMedPubMed Central
  158. Schmidt H, Taubert H, Lange H, Kriese K, Schmitt WD, Hoffmann S, Bartel F, Hauptmann S: Small polydispersed circular DNA contains strains of mobile genetic elements andoccurs more frequently in permanent cell lines of malignant tumors than in normallymphocytes. Oncol Rep. 2009, 22: 393-400.PubMed
  159. Lorthongpanich C, Solter D, Lim CY: Nuclear reprogramming in zygotes. Int J Dev Biol. 2010, 54: 1631-1640. 10.1387/ijdb.103201cl.PubMed
  160. Goodier JL, Ostertag EM, Du K, Kazazian HH: A novel active L1 retrotransposon subfamily in the mouse. Genome Res. 2001, 11: 1677-1685. 10.1101/gr.198301.PubMedPubMed Central
  161. Norio P, Schildkraut CL: Plasticity of DNA replication initiation in Epstein-Barr virus episomes. PLoS Biol. 2004, 2: e152-10.1371/journal.pbio.0020152.PubMedPubMed Central
  162. Norio P, Kosiyatrakul S, Yang Q, Guan Z, Brown NM, Thomas S, Riblet R, Schildkraut CL: Progressive activation of DNA replication initiation in large domains of theimmunoglobulin heavy chain locus during B cell development. Mol Cell. 2005, 20: 575-587. 10.1016/j.molcel.2005.10.029.PubMed
  163. Wang X, Takebayashi S, Bernardin E, Gilbert DM, Chella R, Guan J: Microfluidic extraction and stretching of chromosomal DNA from single cell nucleifor DNA fluorescence in situ hybridization. Biomed Microdevices. 2012, 14: 443-451. 10.1007/s10544-011-9621-8.PubMedPubMed Central
  164. Landriscina M, Fabiano A, Lombardi V, Santodirocco M, Piscazzi A, Fersini A, De Vis K, Barone C, Cignarelli M: Nevirapine toxicity in non-HIV cancer patients. Chemotherapy. 2008, 54: 475-478. 10.1159/000159273.PubMed
  165. Landriscina M, Spadafora C, Cignarelli M, Barone C: Anti-tumor activity of non-nucleosidic reverse transcriptase inhibitors. Curr Pharm Des. 2007, 13: 737-747. 10.2174/138161207780249191.PubMed

Copyright

© Belan; licensee BioMed Central Ltd. 2013

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative CommonsAttribution License (http://​creativecommons.​org/​licenses/​by/​2.​0), whichpermits unrestricted use, distribution, and reproduction in any medium, provided theoriginal work is properly cited.

Advertisement