Construction of synthetic nucleoli and what it tells us about propagation of sub-nuclear domains through cell division

Abstract

The cell nucleus is functionally compartmentalized into numerous membraneless and dynamic, yet defined, bodies. The cell cycle inheritance of these nuclear bodies (NBs) is poorly understood at the molecular level. In higher eukaryotes, their propagation is challenged by cell division through an “open” mitosis, where the nuclear envelope disassembles along with most NBs. A deeper understanding of the mechanisms involved can be achieved using the engineering principles of synthetic biology to construct artificial NBs. Successful biogenesis of such synthetic NBs demonstrates knowledge of the basic mechanisms involved. Application of this approach to the nucleolus, a paradigm of nuclear organization, has highlighted a key role for mitotic bookmarking in the cell cycle propagation of NBs.

Abbreviations

1°

primary

2°

secondary

CBs

Cajal bodies

CDK

cyclin-dependent kinase

DFC

dense fibrillar component

DJ

distal junction

FCs

fibrillar centers

GC

granular component

HLBs

histone locus bodies

IGS

intergenic spacers

HMG

high mobility group

NBs

nuclear bodies

NORs

nucleolar organizer regions

PJ

proximal junction

PML

promyelocytic leukemia

PNBs

pre-nucleolar bodies

pol

RNA polymerase

pre-rRNA

precursor rRNA

rDNA

ribosomal genes

rRNA

ribosomal RNA; RNP, ribonucleoprotein

TFs

transcription factors

t-UTPs

transcription U 3 proteins

UBF

Upstream binding factor

XEn

Xenopus enhancer

At the turn of the 19^th century, botanists Franz Bauer and Robert Brown discovered the nucleus and cytologist Walther Flemming described the disappearance of nuclei during mitosis in higher eukaryotes. While Lorenzo Oken stated that life comes from life (“omne vivum e vivo”) and Rudolf Virchow that cells result from cell division (“omnis cellula e cellula”), Walther Flemming could then state that nuclei derive from nuclei (“omnis nucleus e nucleo”), also noting that chromosome segregation underpins nuclear division. Two centuries later, the age of fluorescence microscopy and live cell imaging have revealed the true complexity of nuclear organization and cell division. The emerging field of synthetic biology is now allowing us to investigate how nuclear organization is propagated through an “open” mitosis.

Functional Compartmentalization of the Nucleus in the Face of Dynamism

The mammalian cell nucleus is highly organized into numerous membrane-free nuclear bodies (NBs) which are morphologically and molecularly distinct domains (for a review see ref.¹) NBs can be classified into 2 groups. Those intimately associated with chromatin, such as nucleoli and histone locus bodies (HLBs), form around specific chromosomal loci usually in a transcription-dependent manner. Nucleoli form around arrays of active ribosomal genes (rDNA)² and HLBs around active histone gene clusters.³ Others, like Cajal bodies (CBs), promyelocytic leukemia (PML) bodies, speckles and paraspeckles, are largely dispersed through the nucleoplasm forming in a chromatin-independent manner,¹ even though they may be subsequently recruited to chromatin.^4,5

These specialized domains compartmentalize essential nuclear processes, including gene expression and biogenesis of ribonucleoprotein (RNP) complexes, such as ribosomes. Proteomic studies have revealed the plurifunctional nature of NBs which may ensure coordination of inter-related nuclear processes. Hence, while the primary function of nucleoli is ribosome biogenesis (i.e. rDNA transcription, pre-rRNA processing and ribosomal subunits assembly), they are also essential for the biogenesis of other RNPs, cell cycle regulation and stress sensing,^6,7 thus coordinating cell growth, cell proliferation and stress sensing. Similarly, multiple related roles have been ascribed to other NBs including CBs and PML bodies.^8,9 It is generally considered that functional compartmentalization within the nucleus, as evidenced by NBs, increases the efficiency of nuclear processes by concentrating the required machinery while excluding unrelated activities.

High-resolution techniques, such as electron microscopy, have revealed that each NB has its own conserved ultrastructural identity. Small NBs are relatively simple, with CBs organized as 0.2-1.5 μm diameter spheres and PML bodies organized as a 0.1-1 μm ring structure.^1,8 The largest sub-nuclear structure, the nucleolus, varies in size according to cell growth and proliferation with a diameter ranging from less than 0.5 μm in differentiated cells to 3–9 μm in proliferating cells.¹⁰ Strikingly, nucleoli are functionally compartmentalized. In mammalian cells, they are sub-divided into 3 distinct components: the fibrillar centers (FCs), the dense fibrillar component (DFC) and the granular component (GC) (Fig. 1).^10-12 Debate has surrounded the relative functions of the FC and DFC, particularly in relation to the site of rDNA transcription by the dedicated RNA polymerase I (pol I) machinery. The balance of evidence now suggests that rDNA transcription occurs at the interface between FCs and DFCs and that early processing of resulting pre-rRNA occurs in the DFC.^10,12 The FCs contain unengaged transcription factors (TFs) and are devoid of processing factors such as fibrillarin. Their role in ribosome biogenesis is uncertain especially since many organisms have bipartite nucleoli, containing only DFC and GC components.¹³ A role for the GC in late processing of pre-rRNA and assembly of ribosome subunits is beyond doubt.^10,12 Despite such complex architecture, there is no evidence for a physical framework underpinning nucleolar structure other than a shell of heterochromatin often observed in human cells.¹⁴ When considering the spectrum of NBs present in the interphase nucleus, nucleoli are the most obvious paradigm for investigating functional compartmentalization.

Figure 1.

Schematic representation of nucleolar compartmentalization. Top left diagram represents a nucleolus within a DAPI-stained nucleus (blue). Nucleoli are subdivided into FC (green), DFC (yellow) and GC (red). FCs constitute the core of nucleoli and are surrounded by DFCs. These compartments are embedded within the GC that is typically surrounded by perinucleolar heterochromatin (dark blue). rDNA transcription by the pol I transcription machinery (UBF, SL1, Rrn3 and pol I) occurs at the interface between FCs and DFCs (white box). The resulting 47 S pre-rRNA transcripts undergo a highly coordinated series of cleavages (green arrows) and modifications to generate the mature 18 S, 5.8 S and 28 S rRNA. Early processing events occur within the DFC (yellow box), while later processing events and assembly of ribosomal subunits occur within the GC (red box). Note that pseudo-NORs support the view that FCs are the interphase counterparts of the mitotic 2° constrictions and that they contain non-transcribed rDNA sequences and unengaged pol I machinery recruited in a UBF-dependent manner (green box).

Challenging the static picture described thus far, live cell imaging has revealed that NBs are highly dynamic structures with highly mobile constituents that diffuse passively through the nucleoplasm.^15,16 The observation that protein diffusion is slowed down within NBs strongly suggested that functional interactions underpin the establishment of nuclear bodies in a stochastic process referred to as “self-organization”.¹⁷ According to this model, NBs are dynamic steady-state hub structures, and sequences within proteins that target them to particular NBs are in fact interaction domains. For example, some nucleolar proteins are recruited to nucleoli through functional interactions with other nucleolar proteins. Nucleolar localization signals, NOLS, are in fact passive retention signals.

In simple eukaryotes, such as yeast, nuclei remain intact through cell division. In higher eukaryotes, however, evolution has challenged the establishment of nuclear organization with cell division through an “open” mitosis. “Open” mitosis is the most extreme mechanism evolved to solve the problem of nuclear division, as the nuclear envelope and most NBs are disassembled, and high levels of chromatin compaction are required for chromosome segregation. Some of those NBs that form in a chromatin-independent manner appear to remain as mitotic bodies. Paraspeckles, CBs and PML bodies, for example, disperse through the cytoplasm of mitotic cells before gradually exchanging their components with re-forming NBs in post-mitotic nuclei.^1,8,18,19 Conversely, NBs forming in close association with transcriptionally active chromatin loci disappear in prophase as nuclear transcription is shut down. Hence, nucleoli break down during “open” mitosis to re-form in telophase around rDNA arrays that were active in the previous cycle. Components of the pol I transcription machinery remain associated with rDNA,^20,21 while the pre-rRNA processing machineries localize to chromosome periphery.^2,22 This localization of the processing machinery has recently been shown to depend on the nucleolar protein Ki-67 that has a similar distribution on metaphase chromosomes.²³ From work in plants, it has been known for some time that nucleoli re-form around prominent chromosomal features, historically termed secondary (2°) constrictions. These occur at a chromosomal loci commonly referred to as nucleolar organizer regions or NORs.^24,25 We now know that NORs contain rDNA repeats, but what constitutes an NOR as a functioning unit, and the relationship between rDNA and 2° constrictions, remained unclear. As cells exit mitosis, CDK1 inhibition leads to the resumption of rDNA transcription,²⁶ while the processing machinery assembles into pre-nucleolar bodies (PNBs). Then, early processing DFC factors are recruited from PNBs to newly synthesized pre-rRNA, and in a subsequent G1 CDK-dependent manner, late processing GC factors are recruited from PNBs to maturing nucleoli.²⁷ Thus, nucleolar reassembly appears to be controlled by an automatic timer, the cell cycle itself, coordinated by inactivation of mitotic CDKs and activation of G1 CDKs.² While cell biologists have described the chain of events involved in an “open” mitosis, including the disappearance and reappearance of nucleoli, and biochemists have identified the switches involved, the molecular mechanism underlying the propagation of NBs through the cell cycle has remained largely undetermined.

Synthetic Biology, a Tool for Exploring Nuclear Organization

Synthetic biology uses genetic engineering to construct artificial or “synthetic” biological circuitry in living cells.²⁸ Although most commonly used to introduce novel biological functionalities to host cells, the engineering principles of synthetic biology can be used to explore intrinsic cellular functions. Indeed, one way to understand the propagation of dynamic yet defined NBs is to attempt to construct them de novo and, in so doing, understand the basic mechanisms involved. As paradigms of nuclear organization, nucleoli are prime candidates for such an approach. Although the most complex of NBs, nucleoli offer many advantages. Firstly, their primary function, ribosome biogenesis, is well defined and can be easily assessed. Secondly, ribosome biogenesis is performed by a dedicated and specific machinery. Finally, nucleoli form around specific chromosomal loci that are easily identified throughout the cell cycle. A study of nucleolar propagation through “open” mitosis can be conveniently broken down into 2 specific goals. The first goal is to determine how competency is conferred on an NOR. This involves an understanding of the relationship between rDNA chromatin and the morphology of competent NORs. The specific aim of this first question is to produce a novel 2° constriction that models mitotic competent NORs. The second goal is to establish how competent NORs are converted into nucleoli. This addresses the DNA sequence requirements, the importance of chromosomal context and the requirement for a preexisting physical framework for nucleolar formation. In this case, the specific aim is to construct a nucleolus de novo that can produce functional ribosomes.

Pseudo-NORs Model Mitotic Competent NORs and Nucleolar FCs

Human NORs contain rDNA arrays positioned on the short arms of the 5 acrocentric chromosomes.²⁹ However, the actual nature of an NOR is still a matter of debate. Indeed, an evolutionary conserved feature of rDNA arrays is their positioning within heterochromatic regions.^29,30 Establishing the sequence of these regions has revealed that they are shared among human acrocentric chromosomes.³¹ Interestingly, although these rDNA-flanking regions were previously thought to be entirely heterochromatic, their analysis has identified a complex genomic architecture.³¹ Moreover, sequences distal to the rDNA array (DJ sequences) are unique to acrocentric short arms, localized to the nucleolar periphery and are transcriptionally active, suggesting an involvement in nucleolar biology³¹ (Fig. 2). Hence, it is still unclear whether NORs are purely an rDNA array or whether non-rDNA sequences are essential for nucleolar formation. Additionally, while mitotic NORs competent for nucleolar formation appear as 2° constrictions, not all NORs are competent. Non-competent or silent NORs are entirely heterochromatic and not associated with nucleoli in interphase.³² The chromatin of 2° constrictions appears to be 10 times less condensed than their surrounding chromatin,³³ resulting in reduced dye binding and appearance of a gap in stained metaphase chromosomes. Thus, in contrast to the 1° constriction of all chromosomes (the centromere), 2° constrictions are under- rather than over-condensed. Components of the pol I machinery remain associated with competent NORs throughout the cell cycle,²⁰ and are likely participants in establishing their specialized chromatin structure. UBF (Upstream Binding Factor), originally identified as a promoter binding pol I transcription factor, is of particular relevance.^34,35 It contains 4-6 HMG-box DNA binding motifs, first identified in the high mobility group chromatin binding protein HMGB1. Like HMGB1, UBF has a highly acidic C-terminus, but, in contrast, it has an N-terminal dimerization domain. Like HMGB1, UBF appears to compete with the nucleosomal binding of the H1 linker histone,^36,37 but its binding is restricted to rDNA by mechanisms that are not yet fully understood. Finally, UBF is often described as an architectural DNA binding protein due to its ability to bend and loop DNA.^38-40 The realization that UBF binds not only to rDNA promoters, but also extensively throughout the rDNA repeat,⁴¹ provided further impetus to the argument that it might be a key player in organizing the chromatin state of competent NORs.

Figure 2.

UBF depletion results in loss of NOR competency. The top cartoon indicates the features of human NORs, including an array of rDNA repeats (red), surrounded by a conserved distal junction (DJ, green) and proximal junction (PJ, white). Each rDNA repeat is comprised of a transcription unit and an intergenic spacer (IGS). The DJ and PJ were named in regard to their positions relative to the centromere. Competent NORs appear as 2° constrictions, gaps in DAPI staining (blue), on metaphase chromosome (middle left) and participate in nucleolar formation in interphase (bottom left). Interestingly, DJ sequences appear as foci localized within perinucleolar heterochromatin. DJ foci identify individual NORs and indicate that mature nucleoli are comprised of multiple NORs. Note, the association of neo-NORs with endogenous human NORs has revealed the existence of NOR territories, indicated by a dashed white line (bottom left), within mature nucleoli. Upon UBF depletion, a subset of competent NORs loses UBF and become silenced. Consequently, they fail to form 2° constrictions during mitosis (middle right) and fail to participate in nucleolar formation in interphase (bottom right).

The construction of pseudo-NORs that model mitotic competent NORs confirmed UBF's involvement in establishing the specialized morphology of 2° constrictions. Pseudo-NORs are ectopic 0.1-2 Mb arrays composed entirely of XEn (Xenopus Enhancer) elements integrated into chromosomes of human HT1080 fibroblast cells.⁴² These XEn elements are derived from the Xenopus laevis rDNA repeat and consists of blocks of 10 60 or 81 bp repeats.⁴³ In their natural context these elements function as transcriptional enhancers absolutely dependent on UBF binding.^44,45 XEn elements are among the best-characterized UBF-binding sites and, despite a lack of sequence homology with human rDNA, human UBF strongly binds to these elements.⁴⁵ UBF binding to XEn arrays integrated into non-NOR-bearing metacentric chromosomes induced the formation of a novel 2° constriction during mitosis⁴² that is lost upon UBF depletion.⁴⁶ The remaining components of the pol I machinery, but not pol I itself, were also recruited to mitotic XEn arrays in an entirely UBF-dependent manner.^42,47 During interphase, XEn arrays additionally recruit pol I and factors, like t-UTPs (transcription U 3 proteins), that couple rDNA transcription with pre-rRNA processing.⁴⁷ Nevertheless, they remain transcriptionally silent due to the lack of promoter sequences, and fail to recruit the DFC/GC processing machineries. It can thus be argued that nucleolar formation involves more than protein/protein and protein/DNA interactions. XEn arrays were called pseudo-NORs, since their UBF-loaded chromatin mimics mitotic competent NORs but fails to generate nucleoli in interphase. Instead, pseudoNORs form novel bodies, equivalent in composition to nucleolar FCs⁴² (Fig. 3). Interestingly, formation of these novel bodies is strictly UBF dependent.⁴⁷

Figure 3.

Synthetic NORs’ contribution to the understanding of the nucleolar cycle. The left panel summarizes the findings that resulted from the construction of synthetic NORs. Pseudo-NORs, UBF-binding site arrays, demonstrate that UBF seeds the formation of 2° constrictions during mitosis and FCs in interphase (top). Neo-NORs, arrays of UBF-binding sites interspersed with rDNA transcription units, indicate that pre-rRNAs are the only architectural requirement for DFC and GC formation (bottom). While UBF depletion has revealed that mitotic bookmarking is necessary for nucleolar formation, pseudo-NORs establish that it is not sufficient. Thus, the cell cycle inheritance of nucleoli is a staged process (represented in the right panel). Endogenous NORs that were active in the previous interphase are bookmarked by UBF, forming mitotic 2° constrictions that retain most of the FC components. This UBF-dependent bookmarking ensures the reactivation of rDNA transcription in late telophase, with pre-rRNAs seeding the recruitment of DFC and GC factors. Hence, staging establishes a temporal order to nucleolar compartmentalization and ensures the rapid re-formation of nucleoli in early G1. For diagrammatic simplicity, individual chromosomes are merged.

Pseudo-NORs provide experimental support for the view that FCs at the core of nucleoli are the interphase counterparts of the mitotic 2° constrictions.²³ We propose that FCs are composed of non-transcribed rDNA sequences, including the IGS of active repeats, loaded with UBF and associated unengaged pol I TFs (Fig. 1). This presumably accounts for much of the approximately 95% of pol I that is associated with nucleoli, but is not engaged in transcription.⁴⁸ It was previously proposed that the presence of FCs correlated with the large intergenic spacers (IGS) of amniotes.¹³ However, the amphibian Xenopus has a short IGS and nucleoli with clearly observable FCs. Interestingly, the XEn sequences employed to construct pseudo-NORs were derived from the Xenopus IGS. Thus, we think it likely that the presence of FCs correlates more closely with the presence of UBF. We believe that pseudo-NORs provide a model for the early steps in nucleolar formation of higher eukaryotes and suggest that UBF binding is required for NOR competency. UBF levels can be reduced 3-5 fold in human cells without severely impacting proliferation at least in the short term. Analysis of such cells indicates that a subset of NORs becomes silenced, loses associated UBF and 2° constrictions, and fails to participate in nucleolar formation⁴⁶ (Fig. 2).

Neo-NORs are Competent NORs Driving the Formation of Neonucleoli

To confirm that UBF-loaded chromatin, modeled by pseudo-NORs, represents an initial stage in nucleolar formation and to study the full nucleolar cycle, competent synthetic NORs called neo-NORs were constructed. Neo-NORs are 1–4 Mb ectopic arrays of the 20.4 kb neo-NOR cassette, comprised of 6.4 kb XEn UBF-binding sites interspersed with a 14 kb rDNA transcription unit that were integrated into HT1080 chromosomes.⁴⁶ The rDNA transcription unit is composed of a human rDNA promoter (nucleotides −253/+37) fused to mouse pre-rRNA coding sequences and terminator (nucleotides +120/+13,698). Comparison of neo-NORs integrated into metacentric and acrocentric chromosomes provides a test for the importance of chromosomal context and sequence requirement for nucleolar formation.

During mitosis, neo-NORs, integrated into metacentric chromosomes, reproduce the appearance of competent NORs, i.e., the 2° constriction. In interphase, they are transcriptionally active. Their transcription units are composed of human pol I promoters driving mouse pre-rRNA coding sequences. This is to overcome pol I species specificity and facilitate monitoring of neo-NOR products. Neo-NOR-derived mouse transcripts induce the recruitment of human nucleolar DFC and GC components and the formation of synthetic compartmentalized nucleoli that we named neonucleoli.⁴⁶ Neonucleoli are able to utilize the human rRNA maturation apparatus to process mouse pre-rRNA into mature rRNA. It appears that the interactions of the processing machinery with pre-rRNAs have been more conserved throughout evolution than those of pol I TFs with rDNA. Neo-NOR-derived mature rRNAs were assembled into polysome-associated ribosomes. Thus, neo-NORs organize the formation of functionally compartmentalized nucleoli and provide the first definitive evidence that rDNA sequences, specifically UBF binding sites coupled with rDNA transcription units, are sufficient to form an NOR competent for nucleolar biogenesis in human cells (Fig. 3).

The ability of neo-NORs integrated into metacentric chromosomes to form neonucleoli indicates that the conserved rDNA-flanking sequences and perinucleolar heterochromatin shell are not essential for nucleolar biogenesis in human cells. The simplest conclusion is that nucleolar formation can occur independently of chromosomal context and architectural framework. However, we believe that the chromosomal location of rDNA arrays and the complex genomic architectures surrounding them play critical roles in other aspects of nucleolar biology. Indeed, perinucleolar heterochromatin has been implicated in maintaining the genomic stability of rDNA arrays in Drosophila melanogaster⁴⁹ and excluding interfering pol II activities from the nucleoli of human cells.⁵⁰ Interestingly, we noted higher levels of neo-NOR genomic rearrangements and lower levels of neo-NOR transcription than expected. Hence, we believe that the chromosomal context of NORs is essential for their stability and efficiency.

Upon UBF depletion, synthetic neo-NORs like endogenous NORs become condensed, silenced and fail to form neonucleoli.⁴⁶ This provides the final piece of evidence that UBF-loading confers competency on NORs. Thus, the construction and analysis of pseudo-NORs, neo-NORs and neonucleoli has provided compelling evidence that nucleolar biogenesis and propagation through ‘open’ mitosis is a staged process where UBF-dependent mitotic bookmarking precedes pre-rRNA-dependent nucleolar assembly⁴⁶ (Fig. 3). While neo-NORs recapitulate the full nucleolar cycle, pseudo-NORs recapitulate only the bookmarking stage.

Mitotic Bookmarking and Propagation of NBs Through Cell Division

Mitotic bookmarking has been defined as the epigenetic mechanism that retains memory of active genes through mitosis to ensure an early G1 reactivation of essential genes and/or the maintenance of cell linage phenotypes.⁵¹ The term was coined in 1997 when it was realized that genes active in the previous interphase retain chromatin marks and under-condensed promoters during mitosis.⁵² It is now thought that like UBF, pol II TFs could play the role of molecular bookmarks.^53,54 Indeed, it has become apparent that a subset of pol II TFs is also retained during mitosis.⁵⁵ Thus, even though it was not understood at the molecular level, the 2° constrictions reported by Emil Heitz and Barbara McClintock should be considered as the first known and most prominent example of mitotic bookmarking. While bookmarking is only starting to be understood, our synthetic biology experiments have allowed us to determine the nature of 2° constrictions and demonstrate their role in facilitating nucleolar formation. Indeed, this UBF-seeded bookmark results in NORs’ under-condensation and their retention of TFs, ensuring preparedness of rDNA for transcription resumption as early as the end of telophase, prior to global chromosome decondensation.

The realization that UBF is not restricted to vertebrates but present across animal phyla⁵⁶ argues that NOR bookmarking has been conserved through evolution. The lack of UBF in plants, where 2° constrictions were first described, suggests that a related HMG-box protein acts as a substitute for UBF, so influencing the epigenetic state and nuclear position of NORs. Interestingly, yeast HMG-box protein Hmo1 has been reported to organize rDNA chromatin.⁵⁷ However, while Hmo1 and UBF play similar roles in pol I transcription, UBF has evolutionary acquired additional roles that cannot be complemented by Hmo1.⁵⁸ We believe that NOR bookmarking is not required in yeast, as nucleoli remain intact throughout their “closed” mitosis. Thus, mitotic bookmarking appears to be a process acquired with ‘‘open’’ mitosis.

While tethering various CB components seemed to be sufficient to nucleate ectopic CBs,⁵⁹ the generation of ectopic HLBs, speckles, paraspeckles and nuclear stress bodies absolutely required the synthesis or tethering of an RNA constituent.^60-62 Thus, de novo construction of other NBs has highlighted the essential architectural nature of RNAs in the formation of both chromatin-associated and chromatin-independent NBs, suggesting a “seed and grow” model.^63,64 This model postulates that the biogenesis of NBs is triggered by a non-random initial seeding event resulting from a biological process such as gene transcription. The resulting RNA seed then recruits the remaining NB components in a self-organized manner.⁶⁴ An RNA seed provides a high level of adaptability to nuclear organization. However, it raises the issue of cell cycle propagation through “open” mitosis, where transcription is shut down and most sequence specific binding factors are thought to be displaced from mitotic chromosomes. Thus, we expect that some form of mitotic bookmarking will also turn out to be essential to the biogenesis of these NBs and other aspects of nuclear functional compartmentalization.

In conclusion, our synthetic biology approach to studying nucleolar formation has provided a conceptual framework for understanding the rapid re-establishment of nucleoli in post-mitotic cells. It has revealed an essential role in higher eukaryotes for mitotic bookmarking in the cell cycle inheritance of the chromatin-associated nucleolus. This mitotic bookmarking appears to be driven by protein-DNA and protein-protein interactions. We believe that it ensures production of an architectural pre-rRNA in early G1 that seeds nucleolar re-formation in a self-organizational manner, in this case driven by the inclusion of protein-RNA interactions. It will be fascinating in the future to see how this framework can be extended to the reformation of other nuclear bodies.

Funding Statement

BM acknowledges Science Foundation Ireland (PI grant 07/IN.1/B924) for funding work in his laboratory. AG was the recipient of an Empower postdoctoral fellowship from IRCSET.

Acknowledgments

We would like to thank Carol Duffy for reading the manuscript and reviewers for their constructive comments.