Transcription of highly repetitive tandemly organized DNA in amphibians and birds: A historical overview and modern concepts

ABSTRACT

Tandemly organized highly repetitive DNA sequences are crucial structural and functional elements of eukaryotic genomes. Despite extensive evidence, satellite DNA remains an enigmatic part of the eukaryotic genome, with biological role and significance of tandem repeat transcripts remaining rather obscure. Data on tandem repeats transcription in amphibian and avian model organisms is fragmentary despite their genomes being thoroughly characterized. Review systematically covers historical and modern data on transcription of amphibian and avian satellite DNA in somatic cells and during meiosis when chromosomes acquire special lampbrush form. We highlight how transcription of tandemly repetitive DNA sequences is organized in interphase nucleus and on lampbrush chromosomes. We offer LTR-activation hypotheses of widespread satellite DNA transcription initiation during oogenesis. Recent explanations are provided for the significance of high-yield production of non-coding RNA derived from tandemly organized highly repetitive DNA. In many cases the data on the transcription of satellite DNA can be extrapolated from lampbrush chromosomes to interphase chromosomes. Lampbrush chromosomes with applied novel technical approaches such as superresolution imaging, chromosome microdissection followed by high-throughput sequencing, dynamic observation in life-like conditions provide amazing opportunities for investigation mechanisms of the satellite DNA transcription.

Background

Satellite DNA was first defined as an additional peak of a DNA fraction in a density gradient of CsCl.¹ Satellite DNA is composed of highly repetitive sequences with preferential “head-to-tail” monomer orientation. Although the tandem DNA repeats make up the main component of (peri-)centromeric and (sub-)terminal heterochromatin,² they can still be transcribed. Satellite DNA families and their non-protein-coding transcripts participate in the spatial organization of genome in interphase nuclei, maintenance of the chromosome structure, chromocenter and centromere formation, and cellular stress response.^3-9 Fundamental knowledge about biogenesis and functions of the satellite non-coding transcripts has found application in medicine.^8-10 It is thus clear that satellite DNAs and their transcripts possess essential structural and functional roles in the cell.

Many reviews have addressed the satellite DNA structure, classification, transcription and function in vertebrates and invertebrates, as well as in plants and yeast.^{2,5,7,8,10-17} However, the data on tandemly organized DNA repeats transcription in the amphibian and avian taxa is fragmentary and insufficiently represented in the literature.^18,19 Meanwhile, the amphibians and birds are promising model organisms to investigate the phenomena of satellite DNA transcription. Birds have relatively small genomes with a characteristic distribution of tandem repeats.²⁰ Specifically, sex W chromosome in chicken contains a number of tandemly repetitive families.²¹ Another example is the germ-line restricted chromosome (GRC) in finches, largely consisting of highly repetitive DNA.^22,23 The amphibian genomes, as a rule, have high C-values and large fraction of tandemly repetitive sequences that facilitates investigation of satellite DNA transcription.^18,24-26 Genomes of model organisms such as chicken (Gallus gallus) or clawed frog (Xenopus) are thoroughly characterized, enabling detailed studies of the tandem DNA repeats.^20,26,27

Birds and amphibians are of a special interest due to the widespread transcription of tandem repeats at the diplotene stage of oogenesis when chromosomes acquire specific lampbrush form.^18,19,24 Exceptionally large size of the lampbrush chromosomes and high rates of transcription during the lampbrush stage enables to study tandem repeats transcriptional activity with striking longitudinal resolution.

Here we summarize historical and modern data on tandemly repetitive DNA transcription in representative species of amphibians and birds. The review will concentrate on transcription of tandem DNA repeats in the female gametes and somatic cells. Hypotheses explaining widespread transcription of satellite DNA during the lampbrush stage will be put forward and possible functions of the resulting non-coding RNAs will be discussed.

Lampbrush chromosomes of amphibians and birds – a brief overview

A brief overview of lampbrush chromosomes is given to highlight their advantages in investigation of transcription and nascent RNA processing. Chromosomes at the diplotene stage of the first meiotic division in oogenesis of amphibians and birds are called lampbrush chromosomes due to their unusual conformation, similar to the oil lamp or test tube brushes.²⁴ Lampbrush chromosomes are half-bivalents with a well-developed chromomere-loop organization: the chromosome axes are composed by linearly distributed beads of compact chromatin – chromomeres, from which one or more pairs of lateral loops emerge (Fig. 1A). The lateral loops differ by size, shape and position along the chromosomes. Excellent description of the amphibian and avian lampbrush chromosome biology was given in the monograph and several recent reviews.^{18,19,24,28,29}

Figure 1.

Localization and transcription of highly repetitive tandemly organized DNA on lampbrush chromosomes. (A) Overall view of lampbrush chromosome (half-bivalent) that consists of central axes and lateral loops emerged from chromomeres. Lateral loops containing actively transcribed repetitive DNA sequences are shown in red. Transcription of tandem repeats can be observed at (peri)centromere, (sub-)telomere, intersticial and polymorphic regions of lampbrush chromosomes. (B-C) Lampbrush chromosome of anuran amphibian Pelophylax lessonae. Phase contrast image (B) and DNA/(DNA+RNA) in situ hybridization with the probe specific to centromere RrS1 tandem repeat (red) (C). Chromosomes are counterstained with DAPI. Lateral loops with detected transcripts of RrS1 repeat are shown by arrows. Enlarged fragments (B’, C’) show transcription units of RrS1 repeat. Nu – extrachromosomal nucleoli. Scale bars: 10 μm. Images B, C, B′, C′ were redrawn from Dedukh et al., 2013 [40].

Considerable decondensation of chromatin in lateral loops is caused by extremely active transcription of looped-out loci. On lampbrush chromosomes, attached to deoxynucleoprotein axes of the loop transcripts with associated proteins form cytologicaly detectable ribonucleoprotein matrix (RNP-matrix) thus making the analysis of the transcription morphology possible. The length of lampbrush chromosomes of the newt Triturus viridescens ranges from 0.4 to 0.8 mm, with the length of the lateral loops varying from 10 to 200 µm.³⁰ The size of the avian lampbrush chromosomes ranges from 5 to 200 µm,^31,32 and the length of their lateral loops varies from 15 to 80 µm.^19,33 These features make lampbrush chromosomes a splendid model for cytological visualization of the DNA transcription process. It is important that fundamental aspects of transcription and nascent RNA processing are the same in different model objects. Thus data from lampbrush chromosomes can be extrapolated on vertebrates in general.

Lateral loops of the lampbrush chromosomes contain transcriptional units, where RNA polymerases move along the DNA axis.^18,19,24 Electron microscope studies demonstrate that in each transcription unit, transcription starts close to the “thin” end of the loop with the minor RNP-matrix and terminates at the “thick” end with increasing length of the nascent transcripts (Fig. 1A).²⁴

The lateral loops of the lampbrush chromosomes can be generally divided into 2 major types: “normal”²⁴ or “simple”¹⁸ and “beaded”/“lumpy” loops²⁴ also known as loops with “complex” morphology.¹⁸ Interestingly, it was demonstrated that different types of satellite DNA repeats can form loops with distinct morphology variants (either “simple” or “complex”).

Milestones of investigation of satellite DNA transcription on lampbrush chromosomes are presented on Fig. 2.

Figure 2.

Milestones: Transcription of satellite DNA on lampbrush chromosomes: from discovery to our days. Lampbrush chromosomes (LBCs), tandem repeats (TRs), long-terminal repeat (LTR), immunofluorescence staining with subsequent fluorescent in situ hybridization( imunoFISH), ribonucleoprotein matrix (RNP-matrix), RNA polymerase II (RNA-pol II).

Amphibian lampbrush chromosomes – first cytological evidence of satellite DNA transcription

Biochemical evidence on the satellite DNA transcription³⁴ has required further cytological demonstration. Application of novel DNA/RNA-transcripts in situ hybridization (ISH) technique to the preparations of lampbrush chromosomes enabled the first cytological demonstration of satellite DNA transcription.³⁵

Early ISH experiments with radioactively labeled probes specific to different types of satellite DNA revealed nascent RNA transcripts on the lateral loops of lampbrush chromosomes in a number of amphibian species (Table S1; Fig. 2).^1,25,35-39 Pre-treatment of the lampbrush chromosome preparations with RNase A, RNase T1 and DNase I confirmed probe hybridization to the nascent RNA transcripts.^1,35,37,39

It became evident that satellite DNA is expressed during the lampbrush stage regardless of where they localize on the chromosomes (Table S1). For example, centromeric satellite repeats TkS1 in Triturus cristatus group and X1-741 in Xenopus laevis are transcribed on lampbrush chromosomes.^25,37 RrS1 repeat is also transcribed on lateral loops arising from centromeric regions of Pelophylax ridibundus and P. lessonae lampbrush chromosomes (Fig. 1B and C).⁴⁰

Satellite 1 of the Notophthalmus viridescens with (peri-)centromeric and 2 non-centromeric sites is transcriptionally active on the lampbrush chromosomes.³⁶ One of the 2 satellite DNA sequences (sat G family), called pTvml, with presumably pericentromeric localization is transcribed on the lampbrush chromosomes of T. vulgaris meridionalis.⁴¹

Repeats distributed throughout the genome, such as satellite 2 of the newt N. viridescens,³⁸ X132C in X. laevis²⁵ or satellite 1 in bullfrogs Rana catesbeiana and R. pipiens³⁹ are transcribed during the lampbrush stage. Transcription occurs on different types of repetitive DNA sequences dispersed throughout the P. waltl genome⁴² and satellite DNA restricted on the long arm of chromosome I of the newt T. c. carnifex.^1,35

Transcribed tandemly repetitive sequences X132A in X. laevis were found at the terminal regions of the chromosomes.²⁵ Crucially, transcripts of X132 repetitive family were originally found in the oocyte RNA and polyrRNA fractions of X. laevis by RNA dot blot hybridization.⁴³ RNA ISH results revealed transcripts of telomeric repeat at the termini of lampbrush chromosomes of anuran amphibians P. ridibundus and P. lessonae (Fig. 1B and C).⁴⁰

The direction of satellite DNA transcription on chromosomal lateral loops as well as the transcribed strands can vary in different species. For instance, X1-741 repeat of X. laevis, satellites 1 and 2 of the N. viridescens and satellite DNA of T. cristatus carnifex are transcribed from both strands.^{1,25,35,36,38} An opposite direction of the transcription in arrays of satellite DNA was shown on the lateral loops of T. cristatus carnifex and N. viridescens.^1,35,36 The RNA of satellite 2 of N. viridescens was found on almost all lateral loops emerging from the corresponding large chromomeres.³⁸ Transcripts of both strands of bullfrog satellite 1 form transcription units with opposite direction of RNP-matrix.³⁹

Thus a number of studies have revealed transcription of tandemly organized highly repetitive DNA sequences on lampbrush chromosomes of anuran and tailed amphibians. Works on the lampbrush chromosome preparations in the early 80's were the pioneering studies originally proving expression of satellite DNA. These works facilitated studies of different aspects of satellite DNA transcription in vertebrates.

Chicken lampbrush chromosomes: Example of a conserved pattern of tandemly repetitive DNA transcription among Vertebrata

Data on the transcription of highly repetitive DNA sequences on giant chromosomes from growing oocytes can be extrapolated to somatic cells of the same and even different species. A remarkable example is the telomeric repeat transcription discovered for the first time on the chicken lampbrush chromosomes, and then demonstrated in cell lines of various species of the order Vertebrata.

On the chicken lampbrush chromosomes, transcription of one of the strands of (TTAGGG)_n repeat was shown originally using a cytological approach.⁴⁴ On the terminal loops of the chicken lampbrush chromosomes, telomere repeat transcripts were detected after DNA/RNA ISH with (TAACCC)₅ probe.⁴⁴

A similar strand-specific telomere-repeat containing non-coding RNAs were detected on the lampbrush chromosomes of turkey (Meleagris gallopavo) and domestic pigeon (Columba livia)⁴⁴ and subsequently Japanese quail (Coturnix japonica).⁴⁵ In each case the labeled RNP-matrix was revealed only after hybridization with (TAACCC)₅ probe; no labeling was observed in experiments with (TTAGGG)₅ probe.⁴⁴ Thus, it was demonstrated that C-rich strand serves as a template for transcription of the avian telomere repeat.⁴⁴

Only after 10 years after the discovery of telomeric repeat transcription in chicken oocytes, telomeric repeat containing RNA, or TERRA RNA, was identified in cancer cells and normal mammalian cell lines.^46-48 Recently, telomeric repeat transcripts were revealed on the lampbrush chromosomes of frogs (P. ridibundus and P. lessonae).⁴⁰

Strand-specific pattern of the telomere repeat transcription is identical in birds, amphibians and mammals. In mammalian cells transcription starts from the subtelomeric regions on several chromosomal ends along with the C-rich strand of the repeat.^46-49 Telomere repeat transcripts on the anuran amphibian lampbrush chromosomes form small caps at the ends of half-bivalents and were detected with (TAACCC)₅ FISH probe.⁴⁰

Mammalian TERRA RNA is heterogeneous in length and is transcribed by RNA polymerase II.^46,47 In the somatic cells, TERRA RNA forms a complex with proteins and accumulates into the nuclear foci known as TERF (TElomeric Repeat-binding Factor) foci.⁵⁰ Likewise the telomere repeat transcripts on giant loops-containing termini of chicken and pigeon lampbrush chromosomes can be associated with RNP-rich structures, called GITERA (GIant TErminal RNP Aggregates).⁴⁵ However, transcription of the telomere repeat on its own does not induce GITERA formation. GITERA probably serves as nuclear domains for accumulation or nuclear retention of non-nascent poly(A) RNAs, including telomere repeat transcripts, subtelomere RNAs and specific sets of RNA-binding proteins.⁴⁵

Evolutionary conserved pattern of telomere (TTAGGG)n repeat containing RNA transcription is an illustration of the common denominator of genome expression among Vertebrata.

Transcription of satellite DNA on avian lampbrush chromosomes: More examples

About 60 avian genomes are available today²⁰ opening up additional opportunities for examining satellite DNA structure and transcription. For instance, Z-macrosatellite sequences (Z-amplicons) are transcribed during the lampbrush stage on chicken Z chromosome and on a number of macrochromosomes.²¹ MHM (male hypermethylated region) is another region that is located on the chicken sex chromosome and is formed by more than 200 copies of tandem repeats. Strand-specific transcription of the MHM region during the lampbrush stage has also been shown to occur on the chicken Z chromosome. Antisense transcripts of MHM exist as heterogeneous non-coding RNA of high molecular-mass.⁵¹

Sequence- and locus-conserved tandem PO41 (pattern of 41 bp) repeat is also transcribed during the lampbrush stage. In chicken and Japanese quail the PO41 repeat is mainly located at the subtelomere regions of both arms of macrochromosomes 1 and 2, pseudoautosomal region of sex chromosone and on the pericentromeric regions of acrocentric microchromosomes.⁵² In turkey the PO41 repeats reside on the pseudoautosomal locus, microchromosomes, at subtelomere regions of p- and q- arms of chromosome 1 and p-arm of chromosome 3.⁵² The PO41 repeat is transcribed from both strands on chicken and Japanese quail lampbrush chromosomes, in addition C- and G-rich PO41 RNAs were shown to be present within one transcription unit due to “head-to-head” organization of repeat copies.⁵² It should be noted that the PO41 RNAs were detected on the lampbrush chromosomes inside intact 3D-preserved nuclei (germinal vesicles) manually isolated from the chicken and Japanese quail oocytes, complementing cytological observations obtained by spreading technique.⁵³

Arrays of the conserved PO41 repeat lie adjacent to the other 41 bp repeat families: the species-specific CNM (chicken nuclear-membrane-associated sequence) and the BglII- (BglII-digested) repeats which localize on the chicken and Japanese quail chromosomes correspondingly.⁵² CNM repeat is transcribed from both strands forming loop-specific pattern of G- and C-rich transcripts.⁵⁴ The C-rich transcripts of the CNM repeat were discovered on the lateral loops originating from the centromere chromomeres of microchromosomes and from the subterminal site of macrochromosome 3, while transcripts from opposite strand of the CNM repeat (G-rich transcripts) were detected on q-terminal regions of microchromosomes.^52,54

Chromosome-specific differences involving transcription of the PO41 and BglII- repeats were revealed. Strand-specific transcripts of PO41 and BglII-repeats were detected in one or several transcription units. Up to 9 loop-specific variants of transcription units organization with different directions were determined.⁵² Moreover, at ZW sex bivalent transcription of PO41 repeat proceeds without transcription of BglII-repeat. On a number of Japanese quail microchromosomes failure of transcription of both repeats was observed.⁵²

PR1 is a highly repetitive sequence located in the centromeric regions of the pigeon chromosomes.⁵⁵ In the domestic pigeon (Columba livia), only one site of PR1 repeat transcription was shown to be present close to the centromere of chromosome 2.⁵⁵ Multiple sites of the PR1 transcription were detected on the loops arising from the centromeres on lampbrush chromosomes of the wood-pigeon (C. palumbus).⁵⁵ An array of lumpy loop 2 tandemly organized repeat (LL2R) that forms marker loops on the chicken lampbrush chromosome 2 is also transcribed prior or during the lampbrush stage of oogenesis.⁵⁶

To summarize, on avian lampbrush chromosomes satellite DNA can be actively transcribed on the sex chromosomes, at the (peri-)centromeric, (sub-)telomeric regions, and also throughout the autosomes.

Composition of the RNP-matrix on lateral loops bearing satellite DNA transcripts

The proteins associated with the nascent transcripts, that form RNP-matrix of lateral loops, may reflect certain steps of their processing such as RNA editing, splicing, cleavage and packaging.^18,19,24

Small nuclear ribonucleoproteins (snRNPs) and heterogeneous nuclear ribonucleoproteins (hnRNPs) are involved in co-transcriptional processing and packaging of nascent RNAs.^57,58 A significant number of studies revealed spliceosomal snRNP and hnRNP components on almost all lateral loops of amphibian and avian lampbrush chromosomes.^18,19,24,59 The RNP-matrix of satellite DNA transcription units on giant loops of bivalent II of the newt N. viridescens³⁶ lacks general hnRNPs and splicing snRNPs^57,60, but is enriched with hnRNP L.⁵⁷

The composition of RNPs associated with the nascent satellite DNA transcripts was thoroughly investigated on the avian lampbrush chromosomes with example of short 41 bp repeats. Particularly, the RNP-matrix of lateral loops bearing C-rich transcripts of the PO41, CNM and BglII-repeats on the chicken and Japanese quail lampbrush chromosomes is enriched by hnRNP K/J.⁵² hnRNP K/J preferentially binds to the C-rich RNA.^61,62 The lateral loops bearing transcripts of PO41, CNM and BglII-repeats are also characterized by the lack of splicing snRNPs.⁵² On the other hand, transcripts of the LL2R tandem repeat recruit snRNPs and SC35 splicing factors, likely exhibiting functions of architectural non-coding RNAs.⁵⁶ The RNP-matrix of lateral loops carrying telomere RNA is enriched with splicing factor PSF/SFPQ.⁴⁵ Thus co-transcriptional stages of satellite RNA processing differ from co-transcriptional stages of mRNA processing.

Recently, Kaufmann et al.⁶³ applied superresolution imaging approach – spectral position determination microscopy – for studying the organization of transcription units in lateral loops. Such an approach might be useful for studying the organization of the transcription units of satellite DNA sequences on giant lampbrush chromosomes.

Widespread transcription of the satellite DNA during the lampbrush stage: Possible mechanisms of initiation

Despite multiple attempts at its elucidation, the mechanism underpinning the widespread transcription of the satellite DNA on lampbrush chromosomes was inscrutable in general. It was assumed that transcription begins at upstream promoters of protein coding genes and the RNA-polymerase fails to terminate at the stop codons, resulting in the read-through transcription of the satellite DNAs (Fig. 3A, upper panel).^1,35,36 For example, satellite 1 of the newt N. viridescens is located between clusters of histone genes where transcription probably starts and then proceeds without interruption into the adjacent satellite DNA region.³⁶ However, Bromley and Gall⁶⁴ have shown that transcription can be initiated downstream of histone gene cluster or at a random site which is often located within satellite 1 sequence. These results indicate that the transcription start can be upstream of the histone gene locus or initiated in the enhancer-like sequences of satellite 1.⁶⁴

Figure 3.

Transcription process and possible fate of non-coding RNA derived from satellite DNA. (A) Hypotheses explaining possible mechanisms of initiation of widespread tandem repeats transcription during the lampbrush stage. According to the “read-through hypothesis” transcription starts from upstream protein coding gene promoters and proceeds without interruption into adjacent non-protein-coding sequences including arrays of tandem repeats. The long terminal repeat (LTR) activation hypothesis holds that transcription begins at active LTR promoters of the endogenous retrovirus elements or even solo LTRs. Nascent transcripts contain non-coding sequences in the neighboring regions including arrays of tandem repeats. Protein coding gene is shown as a blue arrow, non-coding sequences (intergenic spacer) as a pink line; retrotransposon as a green arrow. Start of the transcription is shown as black arcuate arrow, transcripts as a black curly line. (B) A hypothetical scheme describing the fate of non-coding transcripts in oogenesis and early development. Both strands of tandem repeat sequences are often transcribed on lateral loops of lampbrush chromosomes. Resulting long single stranded transcripts (ssRNA) can align into long double stranded form (dsRNA) that activate RNase III enzyme. RNase III enzyme then produces short double stranded non-coding RNAs that accumulate in the oocyte cytoplasm and transfer to the early embryo after fertilization. Short dsRNA participate in regulation of the early stages of embryogenesis before activation of the embryo genome. The fate of transcripts is shown on example of CNM tandem repeat, whose transcription from both strands during lampbrush stage was demonstrated.⁵⁴

Figure 2.,3.

Milestones: Transcription of satellite DNA on lampbrush chromosomes: from discovery to our days. Lampbrush chromosomes (LBCs), tandem repeats (TRs), long-terminal repeat (LTR), immunofluorescence staining with subsequent fluorescent in situ hybridization( imunoFISH), ribonucleoprotein matrix (RNP-matrix), RNA polymerase II (RNA-pol II).

The second scenario is in accordance with the explanation offered for initiation of transcription of the 41 bp tandem repeats on avian lampbrush chromosomes. Bioinformatical analysis showed that arrays of chicken CNM and PO41 repeats often co-locate with the long terminal repeat (LTR) of avian endogenous retroviruses (ERVs).^52,56 Therefore, active LTR promoters as a part of full length ERVs or even as solo LTRs might drive the transcription of CNM and PO41 repeats. As an alternative to read-through hypotheses we offer an LTR-activation hypothesis for the widespread satellite DNA transcription on the lampbrush chromosomes (Fig. 3A, lower panel). We propose that transcription on lateral loops often begins at adjacent active LTR promoters. Interestingly, activation of transcription from scattered LTR promoters in oocytes was recently shown in mammals,^65-67 thus implying evolutionary conservation of the transcription initiation mechanism during oogenesis.

Impossibility of transcription activation at an upstream promoter of protein-coding genes has been suggested for telomeric repeat on chicken lampbrush chromosomes.⁴⁴ Neighboring telomere and subtelomere regions were shown to be transcribed by the RNA polymerase II on chicken and Japanese quail lampbrush chromosomes.⁵² However, it was shown that subtelomere tandem DNA repeats are transcribed in a direction opposite to telomere. Thus in lampbrush chromosomes of chicken, transcription of (TTAGGG)_n likely starts from bi-directional promoter located between the subtelomere and telomere repeat arrays. We suggest that the boundary elements between transcription units can be examined at genomic level by microdissection of the individual lateral loops followed by high-throughput sequencing of the dissected fragments.⁶⁸

Alternatively, the inner promoter sequences were found within Bgl II- tandem repeat of the newt T. vulgaris meridionalis and satellite 2 of the N. viridescens.^69,70 These promoter regions represent functional analog of small nuclear RNA (snRNA) genes promoters. Authors found that inner promoters are involved in transcription of Bgl II-repeat and satellite 2 and can even participate in the transcription of true snRNA genes.^69,70

To summarize, the transcription of highly repetitive tandemly organized DNA may be initiated from promoters of protein-coding genes, scattered LTR promoters and even inner promoter sequences.

Transcription of highly repetitive tandem DNA sequences in somatic tissues of amphibians and birds

Transcription of satellite DNA is not restricted to the lampbrush stage of gametogenesis in amphibians

There are only a few examples of protein-coding genes which are specifically expressed in heterochromatic regions (for a review see refs. in 7). At the same time, position effect variegation and “heterochromatin character” of highly repetitive tandemly organized sequences in somatic tissues were associated with transcriptional repression. A new period began when several groups independently demonstrated transcription of the satellite DNA under stress conditions.^71,72 These examples removed any questions regarding satellite DNA complete transcriptional inactivity.

In the 80's it was demonstrated that satellite DNA transcription beginning during the lampbrush stage of oogenesis can be continued in early embryos as well as in adult somatic tissues (Table S1). For instance, transcripts of the satellite 2 of N. viridescens were demonstrated on the lateral loops of lampbrush chromosomes by DNA/RNA ISH and in ovary, spleen, oviduct, intestine and liver by northern blotting. Satellite 2 transcripts were found in cytoplasm and were heterogeneous in size.³⁸ Moreover, satellite 2 transcripts undergo self-cleavage in vitro and in vivo, forming dimers and multimers with hammerhead-like structures.^73,74 Location of satellite 2 throughout the genome and self-cleavage of the transcripts resemble the behavior of infectious plant RNAs. This resemblance is indicative of common evolutionary origins and significance of the satellite 2 for Salamandridae family.⁷³

Transcripts of the A, B and C subfragments of the X132 tandem repeat were demonstrated in the RNA fraction isolated from X. laevis oocyte and polyribosomes at the embryo stage 40.⁴³ Hybridization with the X132 probe showed a weak signal in oocyte RNA and a strong signal in embryonic RNA. Notably, the transcripts of both strands of the X132A repetitive sequences were demonstrated in the polyrRNA fraction of the Xenopus embryo.⁴³ In contrast, only transcripts of the opposite strand of the X132A repeat sequences were detected on Xenopus lampbrush chromosomes.²⁵

In the X. laevis embryo, tissue-, spatial- and stage-specific pattern was demonstrated for satellite 1 transcription,^75,76 also known as the X1-741 satellite DNA [25], or OAX (Oocyte Activation in Xenopus) region.^77,78 To summarize, transcripts of both strands of the X1-741 satellite DNA sequences were identified by ISH on X. laevis lampbrush chromosomes.²⁵ OAX / satellite 1 transcripts are absent until tailbud stage and cannot be revealed in adult liver or oocytes as demonstrated by RNase protection assay, dot blot hybridization and ISH in cryosections of whole-mount samples.^75,76,78 The later stages of development are characterized by tissue-specific OAX / satellite 1 transcription. The non-coding RNA from both strands of the OAX transcripts was detected in the cement gland and somites.⁷⁸

Thus, satellite DNA expression is not only associated with widespread transcription in growing amphibian oocytes, but represents a common phenomenon of tandem repeats activation.

Specific pattern of transcription of tandemly repetitive DNA in chicken embryogenesis

In chicken, long non-coding MHM RNA formed by tandem repeats is transcribed from antisense strand on lampbrush chromosomes⁵¹ and also from both strands during female embryogenesis.⁷⁹ MHM non-coding transcripts were heterogeneous in size, generally polyadenylated and had predominantly nuclear localization.⁵¹

Northern blotting and whole-mount ISH demonstrated more abundant sense MHM RNAs in gonads, limbs, heart, branchial arch and brain of developing chicken embryo. Sense MHM transcripts are localized in the nucleus, while antisense transcripts are localized in the cytoplasm. Antisense MHM RNAs were detected in the developing ovarian cortex and scattered medulla cells.⁷⁹ Interestingly, only antisense MHM RNAs were detected by RNA gel blot hybridization in chicken embryonic fibroblasts derived from 40-day old female embryo, kidney, spleen, liver, heart, lung, adrenal, brain and ovary.⁵¹ Additionally, antisense MHM RNAs was demonstrated in tissues from triploid intersex female (ZZW), but not from triploid (ZZZ) males.⁵¹ In male tissues transcripts of MHM region were detected only after hypomethylation by 5-azacytidine treatment.⁵¹

These results indicate different sex-, temporal-, strand- and tissue-specific pattern of MHM region transcription. MHM transcripts can be accumulated at the transcription site and can participate in sex determination during embryogenesis.⁵¹ Additionally, mis-expression of the sense and antisense MHM RNAs can result in developmental abnormalities in brain and gonads in chicken embryos.⁷⁹

Transcripts of short tandem repeats as potential house-keeping regulatory non-coding RNAs in Galliformes

Keeping in mind that the subtelomeric tandem PO41 repeat is transcribed during the lampbrush stage of oogenesis,⁵² we have recently demonstrated transcription of both strands of the PO41 repeat in somatic tissues. The PO41 repeat transcripts were found in brain, muscles, oviduct, intestine and eye of adult chicken and Japanese quail females, during chicken embryogenesis and also in chicken MDCC-MSB1 malignant cell line.^80,81 In embryonic and adult somatic and malignant cells the PO41 RNAs exhibit identical distribution pattern during interphase: all transcripts appear to be nuclear-retained and form one or 2 major foci with dispersed RNA in euchromatin (Fig. S1).^80,81 Reverse transcription polymerase chain reaction with specific primers confirmed PO41 transcription at different stages of chicken embryogenesis.⁸¹ Pretreatment of fixed MDCC-MSB1 cells with different RNases followed by RNA hybridization demonstrated that PO41 RNAs exist in predominantly single-stranded form with short double-stranded regions, likely due to inverted units in the arrays of the PO41 repeat.⁸⁰

PO41 RNA demonstrates definite distribution during the cell cycle progression in transformed MDCC-MSB1 cells, embryonic cells and normal cultured embryonic fibroblasts. Long-lived transcripts of the PO41 repeat distribute through the mitosis along with the chromosomes. For instance, at metaphase the PO41 repeat transcripts formed small dots around chromosome plates, whereas in anaphase transcripts lined up at mid-body, and at telophase they were concentrated around the chromosome termini.^80,81

Many satellite DNA transcripts have tissue-, spatial- and temporal-specific patterns of transcription.⁷ In contrast, the PO41 RNAs have similar distribution in the embryonic and adult somatic cells, among different somatic tissues and in normal and transformed cultured cells in a number of Galliformes species.^80,81 Other prominent examples of the widespread transcription are the Xist long non-coding RNA and, its antagonist, the Tsix RNA^82,83 as well as TERRA RNA in mammals.⁴⁷

Uniform character of the PO41 repeat transcription resembles the expression pattern of house-keeping protein-coding and non-protein-coding genes.^80,81 Thus, PO41 transcripts likely execute essential role during meiosis and mitosis. Transcription from both strands of the PO41 repeat can potentially cause generation of the double-stranded RNA and activation of the Dicer-depending mechanism of the heterochromatin formation and maintenance at subtelomere regions.^80,81 It was shown that inhibition of the pericentric satellite transcription leads to disruption of chromocenter formation and cell arrest at 2-cell stage in mouse embryo development.⁶ The definitive function of the PO41 house-keeping RNAs is yet to be determined.

Functions of the non-coding tandemly repetitive DNA transcripts in amphibians and birds

Non-coding transcripts of the tandem repeats are vital for gametogenesis

Functions of the non-coding RNAs derived from tandem repeats are one of the cornerstones of modern cell biology. Different models of biogenesis were introduced for transcripts derived from different types of highly repetitive tandemly organized DNA. It ought to be noted that transcripts of the telomere tandem repeat have a conserved pattern of expression throughout the evolution that can be associated with universal function of TERRA RNA in different organisms. In fact, the (TTAGGG)_n transcripts participate in different aspects of telomere biology in normal and cancer cells.^49,50 However, telomere RNA function during meiosis is still underinvestigated.⁴⁹ That is, during prophase I in human oocyte and spermatocyte development TERRA RNA associates with telomerase reverse transcriptase (TERT) at chromosomal ends. This fact supports the idea that telomeric transcripts might participate in telomere stability during gametogenesis that allows cells to divide properly and to form balanced gametes.^84,85 It should be noted that it is quite difficult to obtain mammalian female gametes preparations, especially preparations of oocytes at the first meiotic division, which limited the studies of the biological role of TERRA RNA and TERT in gametogenesis. Transcription of telomere repeat that was demonstrated at the termini of lampbrush chromosomes^40,44,45 makes amphibian and avian oocytes an attractive model for elucidation of TERRA RNA functions and biogenesis.

Repetitive DNA sequences of the Passeriformes GRC represent an additional enigmatic example of tandem repeats, whose non-coding RNAs could be important for gametogenesis.^22,23,86,87 GRC does not exist in somatic cells of adults and embryos of both sexes, and has distinct behavior and chromatin features in oogenesis and spermatogenesis.²² It is expected that the GRC repetitive sequences are transcribed only during oogenesis, with resulting non-coding transcripts being essential for female gametes maturation or tissue-specific dosage compensation.^22,86,87 However, the transcription of the GRC repetitive DNA sequences on lampbrush chromosomes of the finches has not been studied yet.

Transcripts of tandem repeats can be essential for the progression of early embryonic stages

It was recently discovered that long non-coding RNAs can encode small polypeptides with certain biological activity.⁸⁸ In light of this data it is worth to mention that some amphibian satellite DNA transcripts can appear in the polysomal fraction, such as the X132 RNA of X. laevis⁴³, while others demonstrate ribozyme activity, such as satellite 2 transcripts of newts N. viridescens and Triturus, being active in ovarian and somatic tissues.^38,73,74,89 Moreover, it was proposed that satellite 1 (OAX or X741) of the X. laevis has a potential open reading frame (ORF).⁷⁷ However, so far it is unknown if the other transcripts of amphibian and avian tandem repeats have ORFs or ribozyme activity.

It has been suggested that abundant maternal RNA in oocyte are essential for fertilization, zygote formation and early cleavage stages of development before embryo genome activation.^90-98 However, mRNAs represents only a small fraction of RNAs synthesized in amphibian or avian oocytes, with some of the house-keeping genes being inactive at the lampbrush stage of oogenesis.^19,24,95 For example, only about 1.2% and 3% of transcribed sequences on the Xenopus and Triturus lateral loops serve as mRNA.^91,98 In contrast, mRNAs content in the ooplasm of other metazoan representatives ranged from 35% to 65%, with a large part of mRNAs possibly eliminated later.⁹⁴ The majority of RNA species represent transcripts of repetitive sequences, including tandem repeats,⁹⁵ indicating that their non-coding RNAs are not just waste products of read-through transcription.

Transcription of the major part of tandem repeats from both strands merits special attention since they may form double strand RNA (dsRNA) and thus may be involved in the RNA dependent silencing and heterochromatin formation. In fission yeast, transcripts complementary to both strands are cleaved by Dicer into small interfering RNAs (siRNAs). siRNAs in turn recruit proteins required for heterochromatin formation.^7,8 Large fraction of endogenous siRNAs was also revealed in mouse oocytes.⁹⁹

The Dicer-dependent mechanism of heterochromatin formation operates in chicken somatic cells.¹⁰⁰ Additionally, short RNAs accumulate in chicken oocyte cytoplasm.¹⁰¹ These short regulatory RNAs transferred to the embryo with the ooplasm can be potentially involved in transfer of epigenetic information through generations.¹⁰¹ Parts of these short RNAs can originate from long processed transcripts of tandem DNA repeats (Fig. 3B),¹⁰¹ including the PO41 and CNM repeats, whose transcription from both strands occurs on the lampbrush chromosomes of Galliformes (Fig. 3B).^52,54 However, it is not clear yet whether these non-coding RNAs anneal to form dsRNA. We propose a hypothesis that short regulatory RNAs originating from tandemly repetitive sequences participate in heterochromatin formation at the early stages of embryo development.

Conclusions

Representatives of amphibian and bird taxa are model objects of many investigations. Elegant cytological demonstration of satellite DNA transcription in the amphibian and avian lampbrush chromosomes has represented a remarkable discovery supplementing biochemical studies. According to the current-day model, transcription of tandem DNA repeats during oogenesis is initiated predominantly at active LTR-promoters. Novel approaches applied to the giant lampbrush chromosomes such as superresolution imaging, chromosome microdissection followed by high-throughput sequencing, total RNA next generation sequencing and dynamic observation of nascent RNA processing by labeled molecules provide amazing opportunities for investigation into the mechanisms of satellite DNA transcription in rigorous details and in life-like conditions.

In relation to the current knowledge on non-coding RNAs biogenesis, some of the tandem repeat transcripts likely participate in the regulatory mechanisms of embryo development, epigenetic inheritance, heterochromatin formation by co-transcription silencing, nuclear RNA retention or have ribozyme activity.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgments

We are grateful to Herbert Macgregor (University of Exeter) for critical reading of the manuscript and constructive comments. The authors acknowledge resource centers “The Environmental Safety Observatory,” “Molecular and Cell Technologies,” “Biobank,” and “Chromas” (Saint-Petersburg State University) for the access to the experimental equipment and technical assistance.

The research was supported by Russian Science Foundation grant # 14-14-00131.