Identification and characterization of stem-bulge RNAs in Drosophila melanogaster

ABSTRACT

Non-coding Y RNAs and stem-bulge RNAs are homologous small RNAs in vertebrates and nematodes, respectively. They share a conserved function in the replication of chromosomal DNA in these two groups of organisms. However, functional homologues have not been found in insects, despite their common early evolutionary history. Here, we describe the identification and functional characterization of two sbRNAs in Drosophila melanogaster, termed Dm1 and Dm2. The genes coding for these two RNAs were identified by a computational search in the genome of D. melanogaster for conserved sequence motifs present in nematode sbRNAs. The predicted secondary structures of Dm1 and Dm2 partially resemble nematode sbRNAs and show stability in molecular dynamics simulations. Both RNAs are phylogenetically closer related to nematode sbRNAs than to vertebrate Y RNAs. Dm1, but not Dm2 sbRNA is abundantly expressed in D. melanogaster S2 cells and adult flies. Only Dm1, but not Dm2 sbRNA can functionally replace Y RNAs in a human cell-free DNA replication initiation system. Therefore, Dm1 is the first functional sbRNA described in insects, allowing future investigations into the physiological roles of sbRNAs in the genetically tractable model organism D. melanogaster.

Introduction

Non-coding RNAs (ncRNAs) regulate many fundamental pathways in eukaryotic organisms. Two families of ncRNAs play essential functional roles during chromosomal DNA replication: Y RNAs in vertebrates and stem-bulge RNAs (sbRNAs) in nematodes [1,2]. Y RNAs have been shown to be essential for the initiation of chromosomal DNA replication in a human  cell-free system, for DNA replication and cell proliferation in cultured vertebrate cells and for early development and viability of Xenopus laevis and the zebrafish, Danio rerio [3–5]. sbRNAs are able to functionally replace endogenous Y RNAs in a human DNA replication initiation system. They are also essential for DNA replication and cell proliferation in the nematode Caenorhabditis elegans, as well as its development and viability [6].

Both Y and sbRNAs show homology in function and structure [1,2,7,8]. Structurally, they comprise short stem-loop RNAs of around a hundred nucleotides in length. The partially complementary 5ʹ and 3ʹ ends hybridize to form a double-stranded stem structure with a central single-stranded loop. Common to both Y and sbRNAs are conserved nucleotide sequence elements in the stem structure. They comprise a short helix of 7–10 base pairs flanked by G-C base pairs either end, and a highly conserved double-stranded GUG-CAC tri-nucleotide motif near the center of this domain. This domain is located in the upper stem of both ncRNA families, which opens up into the central loop domain. Importantly, these conserved motifs are essential for the function of these RNAs during the initiation of chromosomal DNA replication in a cell-free system because mutations in these elements abrogate their function [3,6,9,10].

Despite the similarities, there are a few differences between Y and sbRNAs. On the one hand, vertebrate Y RNAs contain a second helical motif with a bulged C residue towards the terminus of the stem-loop, which bind orthologues of the Ro60 protein. In fact, Y RNAs were originally described as the RNA component of Ro-ribonucleotide particles (Ro-RNPs), based on their association with Ro60 [11]. However, neither the Ro-binding domain nor the Ro60 protein are essential for Y RNA function during DNA replication [3,9,12], and they are not found in sbRNAs [1,2,6]. On the other hand, sbRNAs contain a highly conserved UUAUC penta-nucleotide motif at the 5ʹ end of the central single-stranded domain, which appears to play an additional stimulatory role for DNA replication [1,2,6].

In evolutionary terms, functional Y RNAs are found in all vertebrates investigated so far, and sbRNAs in several nematodes including C. elegans and Caenorhabditis briggsae [1,6,13,14]. Yet there is a clear lack of information about the presence of these RNAs in the large group of insects, which diverged from a common ancestor after the nematodes. Computational searches have not provided evidence for a wide-scale conservation of either RNA family in insects, including the major model organism Drosophila melanogaster [8]. However, there are isolated descriptions of a candidate Y RNA in Anopheles gambiae [8] and an sbRNA in the silkworm Bombyx mori [15], suggesting that Y or sbRNAs might be conserved to some extent in insects. It is unknown if these isolated examples play a functional role in DNA replication.

In this study, we conducted a computational search for sbRNA genes in D. melanogaster based on homology motifs present in nematode sbRNAs. We present the first two candidate sbRNA genes in D. melanogaster, coding for two sbRNAs named Dm1 and Dm2. Both their predicted secondary structures resemble sbRNAs and show stability in molecular dynamics simulations. Dm1 is more abundantly expressed than Dm2 in D. melanogaster S2 cells and in adult flies. Only Dm1 can functionally replace Y RNAs in a human cell-free DNA replication initiation system. Therefore, Dm1 is the first functional sbRNA described in insects, allowing future investigations into the physiological roles of sbRNAs in the genetically tractable model organism D. melanogaster.

Results

Identification of sbRNA candidate genes in Drosophila melanogaster

To identify candidate sbRNA genes in the genome of fruit fly Drosophila melanogaster, we searched for sequences with predicted transcription start and stop sites containing conserved nucleotide motifs present in nematode sbRNAs (see Materials and Methods). The search resulted in two candidate sbRNA genes, named Dm1 and Dm2. They are located on chromosomes X and 3L, contain putative upstream RNA polymerase III promoter elements and transcribe into RNA molecules of 85 and 89 nucleotides, respectively (Figure S1 and S2).

We then assessed the predicted secondary structures of these two candidate sbRNAs using the mfold algorithm (Figure 1). Dm1 provided a single structure, while Dm2 provided seven alternative structures. The difference of free energy between the predicted structures of Dm2 was approximately 0.1 Kcal, suggesting that this molecule alternates between the seven related conformations. The most stable predicted secondary structures of Dm1 and Dm2 show an overall stem-bulge shape (Figure 1) that is highly similar to vertebrate Y RNAs and nematode sbRNAs [2]. Both Dm1 and Dm2 contain a double-stranded stem, subdivided into small domains by internal loops and bulges, and a central single-stranded loop (Figure 1). They lack a terminal helix with a bulged C forming a binding site for Ro60 protein, which is typically found in vertebrate Y RNAs. Both candidates also contain a conserved double-stranded GUG-CAC motif present in both nematode sbRNAs and vertebrate Y RNAs, which is essential for their function in chromosomal replication [2]. Both Dm1 and Dm2 contain a UUUAC penta-nucleotide motif downstream of the GUG motif, similar to the UUAUC motif found in nematode sbRNAs at this position. However, unlike nematode sbRNAs, Dm1 has a partially base-paired UUUAC penta-nucleotide motif and a non-conserved bulged G next to the trinucleotide motif (Figure 1a). Dm2 has an unusually short double-stranded stem around the trinucleotide motif and lacks flanking G-C base pairs (Figure 1b), which are conserved in both vertebrate Y RNAs and nematode sbRNAs.

Figure 1.

Predicted secondary structures of two putative Drosophila melanogaster sbRNAs. An evolutionarily conserved trinucleotide domain present in stem-bulge and Y RNAs that is functionally essential for DNA replication activity is highlighted in red/blue, a variable region in yellow, and a second conserved sequence present in sbRNAs in orange.

We conclude that we have identified two candidate sbRNA genes in the genome of D. melanogaster, which share several key elements with vertebrate Y RNAs and nematode sbRNAs.

Molecular dynamics

Computational simulations of the molecular dynamics of small sbRNAs and Y RNAs allow predictions about 3D structure stability and rigidity. We therefore performed these computational simulations on the D. melanogaster candidate sbRNAs Dm1 and Dm2 in comparison to small reference vertebrate Y RNAs (Figures 2, 3, S3-S10). The 3D structures of Dm1 and Dm2 reached thermodynamic equilibrium after 15 ns and 30 ns, respectively (Figure 2). However, both Dm1 and Dm2 were much more stable than the reference Y RNAs. The stability is evaluated as a function of the oscillation. The smaller the oscillation, the greater the stability. The oscillation of Dm1 and Dm2 are the ones closest to a straight line after 30ns of simulation (Figure 2b). The radius of gyration of all RNAs was constant over time for all RNAs analyzed, indicating that these structures maintained their folded configuration over the course of the simulation (Figure 2a). Together, these results suggest that the proposed three-dimensional structures of both Dm1 and Dm2 are stable.

Figure 2.

Parameters of computational molecular dynamics simulations. Root mean square deviation (A) and Radius of gyration (B) obtained from the MD simulation from different Y RNAs and sbRNAs. 3D representations of D. melanogaster Dm1 and Dm2 RNAs, Y3 RNA from C. griseus, H. sapiens and M. musculus and Y5 RNA from M. mulatta were used for this analysis (see Figure S5-S10, corresponding .pdb files).

Figure 3.

Parameters of computational molecular dynamics simulations. Root mean square fluctuation (rmsf) of C1ʹ from D. melanogaster Dm1 (A) and Dm2 (B) RNAs. Red boxes highlight the rmsf of the GUG triplet and blue boxes of the CAC triplet.

We next performed an analysis of individual nucleotide dynamics of Dm1 and Dm2, focusing on the conserved GUG-CAC trinucleotide motifs. The analysis of the root mean square fluctuation (rmsf) for C_1ʹ atoms of each nucleotide suggests that the stem region containing the conserved GUG-CAC motif is a region of high mobility in Dm1 (Figure 3a). This region is much more stable in Dm2 (Figure 3b), whose behavior was similar to the reference Y RNAs (Figure S3). We expected the pairing of this functional triplet in Dm1 to have up to eight hydrogen bonds. However, the average number of hydrogen bonds in this region was 2.4 ± 1.41 during the simulation, while Dm2 had 1.7 ± 1.22, less than half of that expected (Figure S4). Although Dm1 presents a higher average number of H-bonds than Dm2, their number of H-bonds equalized at the end of the simulation (equilibrium region). The reduced numbers of hydrogen bonds suggest that this region has high mobility, in contrast to known sbRNAs [15].

The GUG-CAC motif of Dm1 has a greater flexibility compared to other Y RNAs. The most likely explanations for this behavior are: (1) the loop proximity may have promoted instability in the region; or (2) the long period of simulation (50 ns) may have expanded the imperfections in the force field parameters throughout the simulation.

Previous analyses of the conserved upper stem domain of human Y1 RNA by circular dichroism and solution state NMR indicated that this domain assumes an overall A-form RNA helix, but central bases including the GUG-CAC motif are destabilized and may actually dynamically flip out of the helix [10]. Our molecular dynamics simulations suggest that Dm1, but not Dm2, might show a similar feature of localized high flexibility. Importantly, mutations within the GUG-CAC motif in Y1 RNA led to both structural disturbances and loss of function during the initiation step of chromosomal DNA replication in human cell nuclei [3,9,10]. It is therefore possible that Dm1 RNA might be a functional homologue of vertebrate Y RNAs as it contains this essential motif and associated structural features.

The prediction of 3D structures by homology modeling or by ab initio methods is always accompanied by a certain degree of uncertainty. Molecular dynamics simulations have proved to be a very useful tool to give greater reliability to these proposed structures, since, if the conformation is not stable throughout the simulation time and tends to unfold or collapse, the proposed structure is not feasible and should be reviewed. The conformations proposed in this work do not have homologous structures solved by experimental methods of X-ray crystallography or by nuclear magnetic resonance, therefore, it is impossible to compare them with other templates. In this way, the computational simulation of thermodynamic equilibrium was one of the tools found to generate reliability to the proposed structures, since they proved to be stable along the simulation time.

In the next set of experiments, we turned to phylogenetic and functional analyses of the Dm1 and Dm2 candidate sbRNAs in relation to vertebrate Y and nematode sbRNAs.

Phylogenetic analysis

Vertebrate Y RNAs and nematode sbRNAs share structural and functional homologies [1,2]. Therefore, we performed a phylogenetic analysis of Dm1 and Dm2 in comparison to Y and sbRNAs (Figure 4). Vertebrate Y RNAs separated into the four distinct major clades of Y1, Y3, Y4 and Y5 RNAs (and into minor outgroups, one comprising teleost Y1/3 RNAs and one some amphibian Y5 RNAs), consistent with earlier reports [7,8]. Nematode sbRNAs form a distinct clade from the vertebrate Y RNAs (Figure 4). This clade is more closely related to Y5, than to the three other Y RNA clades. Both Dm1 and Dm2 RNAs associate with the nematode sbRNA clade, together with an sbRNA candidate of the silkworm Bombyx mori described earlier [15]. Taken together, these data strongly suggest that these small insect candidate sbRNAs are more closely related to nematode sbRNAs than to vertebrate Y RNAs.

Figure 4.

Global Phylogenetic Tree of Y RNAs and sbRNAs. sbRNA family (Blue). Y5 RNA family (Red). Y4 RNA family (Purple). Y3 RNA family (Orange). Y1 RNA family (Green). Gene species are presented by: Dm, Drosophila melanogaster; Bm, Bombyx mori; Ce, Caenorhabditis elegans; h, Homo sapiens; Ch, Cricetulus griseus; m, Mus musculus; Rn, Rattus norvegicus, Cp, Cavia porcellus; Oc, Oryctolagus cuniculus; Pt, Pan troglodytes; Mm, Macaca mulatta; Cf, Canis familiaris; Bt, Bos taurus; Md, Monodelphis domestica; z, Danio rerio; La, Loxodonta africana; x, Xenopus laevis; Xt, Xenopus tropicalis. This analysis involved 62 nucleotide sequences. There were a total of 149 positions in the final dataset. The bootstrap consensus tree inferred from 10,000 replicates is taken to represent the evolutionary history of the taxa analyzed [32]. Phylogenetic analyses were conducted in MEGA X [25] and the tree was drawn in Figtree v1.4.2.

These phylogenetic data suggest that insect sbRNAs and nematode sbRNAs may have evolved in their respective lineages from a common ancestor and may therefore share some common functionality. From these data, however, it is unclear whether vertebrate Y RNAs have also evolved from a common ancestor  with nematode or insect sbRNAs, or whether they are the product of a convergent evolution into a partially shared functionality [2]. To address this issue, phylogenetic work would need to be conducted with a substantially larger number of functionally and structurally related sbRNA homologues from other insects and arthropods, and from crustacean, mollusk and even more distantly related groups of organisms. As these data are currently unavailable, future work aimed at identifying such additional distant homologues would be an important goal in the field.

Expression of Dm1 and Dm2 sbRNAs

Y RNAs and sbRNAs are abundantly expressed in vertebrates and nematodes, respectively, at levels approximately ten times below those of ribosomal 5S RNAs [3]. However, individual expression levels vary substantially between different types of Y and sbRNAs within a cell type or tissue, and between different cell types and different tissues of the same organism [2,3,6,16,17]. Next, we therefore determined the relative expression levels of Dm1 and Dm2 sbRNAs in D. melanogaster S2 cells and adult flies (Figure 5). In S2 cells, the expression of the Dm1 and Dm2 genes was approximately 250-fold and 1,000-fold lower than the 5S gene, respectively. In adult flies, Dm1 and Dm2 sbRNA expression was 66-fold and 900-fold lower than the 5S gene, respectively.

Figure 5.

Expression of Dm1 and Dm2 sbRNAs, relative to 5S rRNA. Total RNA was prepared from Drosophila melanogaster S2 cells and adult flies, and quantified by qRT-PCR. Percentages of relative expression are shown as mean values ± standard deviations of n = 4 independent experiments.

We conclude that the relative expression level of Dm1 sbRNA in D. melanogaster is similar to sbRNAs in nematodes and Y RNAs in vertebrates, whereas Dm2 sbRNA is expressed at one to two orders of magnitude below that.

Functional homologies to human Y RNA

Vertebrate Y RNAs and nematode sbRNAs have essential functions in chromosomal DNA replication [2]. Depletion of Y RNAs and sbRNAs in vivo by RNA interference or antisense morpholino oligonucleotides leads to an inhibition of DNA replication and cell proliferation, and to death during early development [3,4,6]. In a cell-free system derived from human cells, Y RNAs are required in late G1 phase nuclei specifically for the initiation of chromosomal DNA replication, but not for subsequent elongation synthesis [3,5]. Importantly, non-human vertebrate Y RNAs and several nematode sbRNAs can functionally substitute for human Y RNAs in this system [6,9], indicating that they are functionally redundant and homologous. To address whether Drosophila Dm1 and Dm2 sbRNAs were functionally homologous to these RNAs, we tested whether they could functionally substitute for vertebrate Y RNAs in this system (Figure 6).

Figure 6.

DNA replication initiation function. (A) Synthesis of Dm1 and Dm2 sbRNAs. The indicated RNAs were synthesised by in vitro transcription and visualised after denaturing polyacrylamide gel electrophoresis. Molecular sizes of single-stranded DNA oligonucleotide markers are indicated. (B) RNA-dependent initiation of chromosomal DNA replication in a human cell-free system. Human late G1 phase template nuclei were incubated in the absence or presence of control or Y RNA-depleted human cytosolic extracts supplemented with 100 ng of purified Dm1 or Dm2 sbRNAs, as indicated. Percentages of replicating template nuclei are shown for each reaction as mean values ± standard deviations of n = 5 independent experiments. Brackets indicate results of t-tests (unpaired, two-tailed with unequal variance) of control against the experimental samples (*** p = 0.007; n.s., not significant).

We synthesised Dm1 and Dm2 sbRNAs in vitro from synthetic templates and purified the full-length RNAs (Figure 6a). We then tested these RNAs in the human cell-free DNA replication initiation system (Figure 6b). Incubation of late G1 phase template nuclei in a control extract from proliferating HeLa cells resulted in a significant initiation of chromosomal DNA replication in about 35% of these nuclei. This is above the background of about 5% of contaminating S phase nuclei replicating in the absence of extract, as described previously [3,18]. Degradation of the endogenous human Y RNAs resulted in an inhibition of replication initiation, reducing the percentages of replicating nuclei to near-background values (Figure 6b), as described previously [3]. Significantly, addition of Dm1 restored the initiation function of the Y RNA-depleted extract, whereas addition of Dm2 did not (Figure 6b). We therefore conclude that Dm1 sbRNA, but not Dm2 sbRNA, is a functional homologue of human Y RNAs in this DNA replication initiation system.

Discussion

In this study, we identified by a computational search for conserved sequence motifs present in nematode sbRNAs two short candidate RNAs in the genome of the fruit fly Drosophila melanogaster, which we termed Dm1 and Dm2 sbRNAs. Both RNAs have predicted secondary structures partially resembling nematode sbRNAs and both show stability in molecular dynamics simulations. These candidate RNAs are phylogenetically closer to nematode sbRNAs than to homologous vertebrate Y RNAs. However, we found that only Dm1 sbRNA, but not Dm2, is abundantly expressed and can functionally replace Y RNAs in a human cell-free DNA replication initiation system.

We have tested the functionality of the two candidate sbRNAs in a heterologous DNA replication system derived from human cells. This approach is very powerful to identify homologues of Y and sbRNAs in vitro, yet it requires subsequent in vivo validation in the cognate organism. Using this system, it was shown that exogenous non-human vertebrate Y RNAs and nematode sbRNAs can replace endogenous human Y RNAs to initiate chromosomal DNA replication in human cell nuclei in vitro [6,9]. Independent subsequent studies using microinjection of antisense morpholino oligonucleotides in zebrafish, Xenopus laevis and C. elegans then confirmed a functional role of these RNAs during DNA replication, cell proliferation and organismal viability in the homologous developing organism [4,6]. Future experiments would thus be required to define the physiological functions of Dm1 and Dm2 RNAs in D. melanogaster, for instance by disrupting their functionality by injection of antisense morpholino oligonucleotides into fertilized eggs, or by knocking out their genes using genetic tools.

With this caveat in mind, our results support the conclusion that Dm1 is a genuine stem-bulge RNA expressed in D. melanogaster, which is functionally related to nematode sbRNAs and vertebrate Y RNAs. A different short sbRNA candidate gene has previously been described in another insect, the silkworm Bombyx mori [15], and our phylogenetic analysis reported here suggests that it is homologous to Dm1 sbRNA. However, it is yet unknown whether this BmsbRNA has an activity during DNA replication and would thus be functionally homologous, or orthologous, to Dm1 sbRNA, other nematode sbRNAs or vertebrate Y RNAs.

To our knowledge, Dm1 sbRNA would therefore be the first functional sbRNA described in insects. This discovery should stimulate future work into identifying and characterizing additional functional sbRNA or Y RNA homologues in insects, and to study their physiological roles during development and tissue maintenance in whole organisms.

In contrast, the Dm2 sbRNA gene is only very weakly expressed in D. melanogaster (if at all). Dm2 shares fewer conserved elements with vertebrate Y RNAs and nematode sbRNAs than Dm1. In the phylogenetic tree however, Dm2 still groups with other nematode sbRNAs, but it did not show any significant activity in initiating DNA replication in vitro. It is important to note that not every nematode sbRNA was active in a human cell-free DNA replication initiation system either [6,19,20]. These observations suggest that Dm2 sbRNA may not be a fully functional Y RNA homologue in the context of initiating DNA replication in vitro, but it may have a different physiological role. Interestingly, the overall structure of Dm2 is not inconsistent with the possibility that it could be a miRNA precursor [6,19,20]. We will address this possibility separately in a future communication.

In conclusion, we have shown that small non-coding stem-loop RNAs with conserved structural elements and a homologous function in DNA replication are not only present in nematodes and vertebrates, but also in insects. This finding should stimulate further work into the evolution of these RNAs and into elucidating molecular mechanisms underlying their physiological function using the powerful genetic tools available in D. melanogaster.

Materials and methods

Search for sbRNA gene sequences

The search for sequences similar to sbRNAs was carried out in the genome of Drosophila melanogaster (genome id: 47), as previously described [15], starting from fragments of conserved sequences for sbRNAs containing potential start and stop points of transcription [21,22]. Sequences found were submitted to mfold web server [23], using default parameters, in order to obtain their secondary structure. This secondary structure was compared with the standard model for sbRNAs, described for Caernohabditis elegans [1].

Phylogenetic analysis

The alignment (Figure S11), necessary to construct the phylogenetic tree, was carried out in Clustal Omega [24], using sequences for Y RNAs [8] and sbRNAs related to DNA replication [1,6,7]. The phylogenetic analysis was carried out with MEGA version X [25], using an unrooted Maximum Likelihood method with 10,000 bootstrap steps. Then, the tree was drawn in Figtree v1.4.2 (http://tree.bio.ed.ac.uk/software/figtree).

Molecular dynamics

RNA secondary structures were constructed with Varna Applet [26] and used for graphical representation. The 3D representations were generated by RNA Composer server [27], and used for molecular dynamics simulations (MD). These 3D representations are supplied as .pdb files (supplementary figures S5-S10, Supplementary Material online).

The MD simulations were carried out and analyzed in the programs NAMD2 [28] and VMD [29] using Charmm C35b2/C36a2 force field [30]. For this procedure, the spatial coordinates of the RNAs were virtually immersed in a periodic box containing TIP3 water and sufficient amount of sodium counter ions to neutralize the system charges. The box size was at least 15 Å away from the outer surface of the RNA. The simulations were carried out in steps. In the first one, all atoms of the system were minimized by 20,000 steps of Conjugated Gradient (CG). In the second step, water and ions were equilibrated by 60 ps. In the third step, all atoms of the system were minimized again by another 20,000 steps of CG. In the fourth and final step, all atoms of the system were equilibrated during 50 ns using 1 atm pressure, 300 K temperature, and a constant number of atoms (NPT ensemble). The other simulation parameters were adjusted according to the protocol established by [31]. The simulations took place in 20 nodes of an Intel Xeon E5-2695v2 Ivy Bridge, 2.4 GHZ (480 cores) processors of the St Dumont cluster at LNCC, Brazil.

Relative quantification of Drosophila RNAs by qRT-PCR

Total RNA was extracted from D. melanogaster cell cultures and adult wild-type flies. D. melanogaster S2 cells (provided by Dr. Alain Debec, Institut Jacques Monod, Paris, France), derived from the embryo of D. melanogaster, were maintained in Shields and Sang M3 insect culture media (Sigma-Aldrich), supplemented with 10% heat-inactivated fetal bovine serum (FBS) (Gibco) and antibiotics at a temperature of 26°C. For RNA extraction, the media was discarded, followed by extraction with TRIzol LS (Invitrogen). D. melanogaster flies were acquired from the State University of Maringá, Maringá – Paraná, Brazil. For whole organism RNA extraction, adult flies were macerated in liquid nitrogen, followed by extraction with TRIzol LS (Invitrogen).

Total RNA sample concentrations were quantified by nanodrop spectroscopy using NanoDrop2000 (Thermo-Fisher) and standardized for a total of 2000 ng of RNA for the treatment with DNAse I (Biolabs). The reverse transcriptase kit used to synthesize the first cDNA was iScript (Bio-Rad). The qPCR reactions were carried out by the following kit: Power SYBR Green PCR Master Mix (Thermo Fisher), with 40 cycles and annealing temperature of 58°C, in the LightCycler® 96 System (Roche) equipment. The primer sequences used were:

• Dm1: 5’ GGGGTGGGTGTACCCGGAAA-3' and 5’AAAACGAGGAGGAACTATGAGGG-3’

• Dm2: GGCCATGGTTAGCGACGCG and 5’AAAAAAGACTATGACCCCCGCC-3’

• 5S rRNA: 5’ GCCAACGACCATACCACGC-3’ and 5’AAAAAGTTGTGGACGAGGCCAA-3’

The relative amount of expression (Ar) was calculated (Equation 1). Each sample was carried out in duplicates and calculations were based on 4 independent experiments.

$$Ar=2^{\left(-\mathrm{\Delta }CT\right)}=2^{-\left(Ctofthesample-Ctof5Sgene\right)}$$ (1)

RNA synthesis in vitro

Dm1 and Dm2 sbRNAs were synthesized in vitro by bacteriophage SP6 RNA polymerase from two annealed complementary synthetic DNA oligonucleotides (Sigma-Aldrich Co. LLC), as described previously [10]. Sequences of the four DNA oligonucleotides with added SP6 promoter sequences (underlined) are as follows:

• Dm1 forward 5’ATTTAGGTGACACTATAGGGGGTGGGTGTACCCGGAAAGTGGTGTTTACACGTCCCGTCCATGAGTAATCTGCGCACTTTCCCTCATAGTTCCTCCTCGTTTT-3’

• Dm1 reverse 5’AAAACGAGGAGGAACTATGAGGGAAAGTGCGCAGATTACTCATGGACGGGACGTGTAAACACCACTTTCCGGGTACACCCACCCCCCTATAGTGTCACCTAAAT-3’

• Dm2 forward 5’ATTTAGGTGACACTATAGGGCCATGGTTAGCGACGCGGTGGTGTTTACCCACATGTAGGCATATGCATATGCTGATTTCCACCAAGGCGGGGGTCATAGTCTTTTTT-3’

• Dm2 reverse 5’AAAAAAGACTATGACCCCCGCCTTGGTGGAAATCAGCATATGCATATGCCTACATGTGGGTAAACACCACCGCGTCGCTAACCATGGCCCCTATAGTGTCACCTAAAT-3’

Transcription produced the following synthetic RNAs:

• Dm1 sbRNA 5’GGGGUGGGUGUACCCGGAAAGUGGUGUUUACACGUCCCGUCCAUGAGUAAUCUGCGCACUUUCCCUCAUAGUUCCUCCUCGUUUU-3’

• Dm2 sbRNA 5’GGCCAUGGUUAGCGACGCGGUGGUGUUUACCCACAUGUAGGCAUAUGCAUAUGCUGAUUUCCACCAAGGCGGGGGUCAUAGUCUUUUUU-3’

Following in vitro synthesis, template DNA was degraded by DNAse I treatment and the newly transcribed RNA was purified by phenol/chloroform extraction and ethanol precipitation. The RNA was then visualised by electrophoresis on denaturing polyacrylamide gels containing 8M urea, as described previously [3,4].

DNA replication reactions

Human HeLaS3 and EJ30 cells were grown as proliferating monolayers, and template nuclei were prepared from mimosine-arrested late G1 phase EJ30 cells as described previously [18].

DNA replication initiation reactions were performed as detailed previously [3,18]. These contained late G1 phase template nuclei prepared from mimosine-arrested human EJ30 cells, a cytosolic extract from proliferating HeLaS3 cells, in vitro-synthesised RNAs, and a buffered mix of ribo- and deoxyribonucleoside triphosphates. The mix included digoxigenin-dUTP as a tracer for newly synthesised DNA. Inclusion of creatine phosphate/phosphocreatine kinase provided an energy regenerating system. Endogenous human Y RNAs were depleted from the HeLa cell extract by endogenous RNAseH activity targeted by antisense DNA oligonucleotides, as detailed previously [3].

After a 3-hour reaction time, nuclei were fixed and sedimented on polylysine-coated glass coverslips. Digoxigenin-labelled DNA, the product of the replication reactions, was detected by anti-digoxigenin fluorescein-conjugated F_ab fragments (Roche), and total DNA was stained with propidium iodide, as described previously [3,18]. Confocal fluorescence microscopy was performed on a Leica SP1 microscope using 40x lens magnification; individual channels were recorded simultaneously. The percentages of replicating nuclei were determined from randomly chosen microscopic fields. At least 200 nuclei were scored per reaction.

Funding Statement

This work was supported by the Conselho Nacional de Desenvolvimento Científico e Tecnológico [305960/2015-6]; Fundação Araucária [174/2016]; Fundação Araucária [40/2016].

Acknowledgments

This work was supported by Fundação Araucária (147/14 and 40/16), Coordination for the Improvement of Higher Education Personnel - Brazil (CAPES) and National Council for Scientific and Technological Development – Brazil - CNPq (grant number 305960/2015-6). SLP was supported by the Department of Zoology, University of Cambridge. FFDJ, JRPJ and LGP received graduated CAPES fellowships; PSAB, FSR and ACR received graduated fellowships from CNPq; DC and AFSL received undergraduate fellowships from CNPq and Fundação Araucária, respectively. The authors would like to thank Dr. Alain Debec for providing the Drosophila melanogaster S2 strain cells, LNCC for computational facilities; COMCAP/UEM for equipment facilities; and Dr. Omar Cleo Neves Pereira, Dr. Cláudio Henrique Zawadzki and Milena Pavlickova for critical discussions.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Supplementary data

Supplemental data for this article can be accessed here.