Swarthmore College Works

Biology Faculty Works Biology

3-1-2021

Ectopic Expression Of Pericentric HSATII RNA Results In Nuclear RNA Accumulation, MeCP2 Recruitment, And Division Defects

C. C. Landers

C. A. Rabeler , '20

E. K. Ferrari

Lia R. D'Alessandro , '21

D. D. Kang

See next page for additional authors Follow this and additional works at: https://works.swarthmore.edu/fac-biology

Part of the Biology Commons, and the Genetics Commons Let us know how access to these works benefits ouy

Recommended Citation C. C. Landers; C. A. Rabeler , '20; E. K. Ferrari; Lia R. D'Alessandro , '21; D. D. Kang; J. Malisa; Safia M. Bashir , '20; and Dawn M. Carone. (2021). "Ectopic Expression Of Pericentric HSATII RNA Results In Nuclear RNA Accumulation, MeCP2 Recruitment, And Defects". Chromosoma. Volume 130, Issue 1. 75-90. DOI: 10.1007/s00412-021-00753-0 https://works.swarthmore.edu/fac-biology/611

This work is licensed under a Creative Commons Attribution 4.0 License. This work is brought to you for free by Swarthmore College Libraries' Works. It has been accepted for inclusion in Biology Faculty Works by an authorized administrator of Works. For more information, please contact [email protected]. Authors C. C. Landers; C. A. Rabeler , '20; E. K. Ferrari; Lia R. D'Alessandro , '21; D. D. Kang; J. Malisa; Safia M. Bashir , '20; and Dawn M. Carone

This article is available at Works: https://works.swarthmore.edu/fac-biology/611 Chromosoma (2021) 130:75–90 https://doi.org/10.1007/s00412-021-00753-0

ORIGINAL ARTICLE

Ectopic expression of pericentric HSATII RNA results in nuclear RNA accumulation, MeCP2 recruitment, and cell division defects

Catherine C. Landers1 & Christina A. Rabeler 2 & Emily K. Ferrari3 & Lia R. D’Alessandro2 & Diana D. Kang4 & Jessica Malisa5 & Safia M. Bashir2 & Dawn M. Carone2

Received: 11 May 2020 /Revised: 16 January 2021 /Accepted: 19 January 2021 / Published online: 13 February 2021 # The Author(s) 2021

Abstract Within the pericentric regions of reside large arrays of tandemly repeated sequences. Expression of the human pericentric satellite HSATII is prevented by extensive heterochromatin silencing in normal cells, yet in many cancer cells, HSATII RNA is aberrantly expressed and accumulates in large nuclear foci in cis. Expression and aggregation of HSATII RNA in cancer cells is concomitant with recruitment of key chromatin regulatory proteins including methyl-CpG binding protein 2 (MeCP2). While HSATII expression has been observed in a wide variety of lines and tissues, the effect of its expression is unknown. We tested the effect of stable expression of HSATII RNA within cells that do not normally express HSATII. Ectopic HSATII expression in HeLa and primary fibroblast cells leads to focal accumulation of HSATII RNA in cis and triggers the accumulation of MeCP2 onto nuclear HSATII RNA bodies. Further, long-term expression of HSATII RNA leads to cell division defects including lagging chromosomes, chromatin bridges, and other chromatin defects. Thus, expression of HSATII RNA in normal cells phenocopies its nuclear accumulation in cancer cells and allows for the characterization of the cellular events triggered by aberrant expression of pericentric satellite RNA.

Keywords Noncoding RNA . Pericentric heterochromatin . Chromatin instability . Cancer biology . Human

Introduction uniquely localized within the . While LINE and SINE elements are heterogeneously interspersed genome-wide, Nearly 50% of the human genome consists of repetitive DNA tandemly repeated elements are more positionally defined. sequence elements. These include -based Satellite DNAs, defined by tandemly repeating units of repeats such as long and short interspersed nucleotide ele- DNA, localize primarily to centric or pericentric regions and ments (LINES and SINES) and simple sequence repeats (tan- to of all chromosomes, and are classified based on dem DNA) such as ribosomal DNAs and satellite sequences their sequence and length of repeating unit. The major classes (Richard et al., 2008). Distinct classes of repeat elements are of human satellite DNA are the following: (1) alphoid (alpha satellite; α-Sat) DNA resident at all centromeres; (2) telomeric repeats; (3) beta satellite DNA (β-Sat) on acrocentric chromo- Catherine C. Landers and Christina A. Rabeler contributed equally to this work. somes; and (4) satellites HSATI, HSATII, and HSATIII found at the pericentric regions of a subset of chromosomes. * Dawn M. Carone Large blocks of HSATII DNA are located on 11 human [email protected] chromosomes, with the largest arrays residing within the pericentromeres of human chromosomes 1 and 16. Satellite 1 Department of Nutritional Sciences, University of Connecticut , DNA blocks may span several megabases and consist of Storrs, CT, USA tandemly repeated elements ranging from five to several hun- 2 Department of Biology, Swarthmore College, Swarthmore, PA, USA dred nucleotides (Richard et al., 2008). Length and sequence 3 Children’s Hospital of Philadelphia, Philadelphia, PA, USA composition of satellites, including HSATII, varies in the hu- 4 Division of Pharmaceutics and Pharmacology College of Pharmacy, man population and may be used to study human genetic Ohio State University, Columbus, OH, USA variation (Miga, 2019); however, the full extent and diversity 5 Stanford University School of Medicine, Stanford, CA, USA of HSATII sequences is currently unknown due to its location 76 Chromosoma (2021) 130:75–90 within large sequence assembly gaps in the human genome 2005). MeCP2 is recruited to HSATII RNA accumulations (Altemose et al., 2014; Eichler et al., 2004). HSATII is rough- and may be sequestered in CAST bodies in cancer cells, but ly defined as a ~26 bp repeat, but exhibits significant hetero- the dynamics, consequences, and direct role of HSATII RNA geneity in sequence and has been poorly studied, despite its in this recruitment are currently unknown. prominence in the human (Tagarro et al., 1994)and Inappropriate expression of satellite RNAs can be an indi- abundance in the human genome. A computational effort to cator of heterochromatic instability, which may be increasing- generate a satellite reference database (Altemose et al., 2014) ly common in cancers, and has wide-ranging implications successfully demonstrates the ability to cluster HSAT se- (Carone and Lawrence, 2013). Given that maintenance of quences into subfamilies, which are likely to reside on indi- centric/pericentric heterochromatin is key to normal chromo- vidual chromosomes, but a comprehensive HSATII map has some segregation, structural changes to the underlying chro- yet to be completed. In contrast, significant progress has been matin may result in functional changes, including satellite made in mapping alpha satellite arrays, consisting of 171-bp expression and aberrant cell division. Supporting this, prior monomer repeats, particularly on the X (Miga et al., 2019)and studies analyzing the effect of satellite expression have dem- Y (Jain et al., 2018) chromosomes, to generate full - onstrated that forced expression of α-sat transcripts leads to to-telomere linear sequence. cell division defects, chromosomal instability, and aneuploidy Due to their localization near centromeres, satellite se- in normal cells (Chan et al., 2017;Ichidaetal.,2018;Zhu quences are likely to be subject to strict regulation to maintain et al., 2018). Paradoxically, some α-sat expression is observed both genetic and epigenetic stability. While expression of in normal cells (Hall et al., 2017) and low levels of α-sat pericentric satellites has been observed in many species, includ- expression are thought to be necessary for centromere func- ing yeast, , and mammalian cells, it is clear that when tion and constitutive heterochromatin maintenance (Johnson satellites are expressed, their expression is highly regulated et al., 2017; McNulty et al., 2017). Further, some studies have (reviewedin(Halletal.,2012; Perea-Resa and Blower, 2018; found no phenotype following forced expression of satellite Smurova and De Wulf, 2018)). The increased presence of sat- sequences (Ideue et al., 2014;Rošić et al., 2014). Thus, the ellite RNA has recently emerged as an indicator of instability, functional role of satellites has been an enigma, despite their as demonstrated by satellite overexpression in cancer cells and prevalence and structural conservation within pericentric re- tumor tissues (Hall et al., 2017;Tingetal.,2011;Zhuetal., gions. This conserved, tandemly repetitive structure of centric 2011). Whereas α-sat is expressed at low levels in normal cells and pericentric satellites is in direct juxtaposition to the dem- (Halletal.,2017; McNulty et al., 2017) and increases expres- onstrated lack of sequence conservation between species sion in cancer, HSATII expression is restricted to cancer cells, (Henikoff et al., 2001) and the discovery of functional and thus, the presence of HSATII RNA is a potential biomarker neocentromeres lacking satellite sequences (Voullaire et al., of cancer (Hall et al., 2017;Tingetal.,2011). Overexpression 1993). Pericentric satellites, in particular, have little known of chromosome-specific pericentric HSATII loci (e.g., Chr 7) function despite their abundance within the pericentric se- occurs within nuclei in which other HSATII chromosomal lo- quence landscape. It has become clear that the mechanisms cations (e.g., Chr 1) are not expressed, due to their accumula- governing transcriptional silencing of pericentric satellites are tion of repressive polycomb proteins, thus indicating that complex, with many histone modifications contributing to HSATII expression is regulated in a locus-specific manner maintaining repression. Thus, a ubiquitous role for pericentric (Halletal.,2017). Accumulation of HSATII RNA occurs in repeats, the mechanisms governing their regulation, and the cis, with the RNA accumulating in cancer-associated satellite satellite sequences that comprise them, has remained poorly transcript (CAST) bodies, adjacent to the HSATII locus from understood. Given that pericentric satellite expression is which it is transcribed. Within a given tumor or cancer cell line, misregulated in a wide range of cancer cell lines and tissues, expression of HSATII maintains this locus-specific expression and the chromatin structure of pericentric satellites is compro- pattern in most cells (Hall et al., 2017). mised in both cancer and senescent cells (Brückmann et al., HSATII RNA within CAST bodies recruits and directly 2018; Hédouin et al., 2017; Slee et al., 2012; Swanson et al., binds MeCP2 and its protein-binding partner, Sin3A (Hall 2013;Tassellietal.,2016), the direct effect of aberrant et al., 2017). MeCP2 is canonically classified as a DNA pericentric satellite expression warrants investigation. methyl-binding protein (Hite et al., 2009); however, more re- Here, we established cell lines stably expressing HSATII cent evidence indicates that MeCP2 also contains an intrinsi- RNA and α-sat RNA in order to test the long-term effects of cally disordered domain, which has the capacity to bind RNA expression of these distinct satellite transcripts. We demon- (Castello et al., 2016), and can function as a transcriptional strate that ectopically expressed HSATII RNA accumulates in activator (Hite et al., 2009). MeCP2 is mutated in Rett syn- nuclear bodies reminiscent of CAST bodies in both HeLa and drome, where it is known to bind transcripts in the brain and primary human fibroblast cells. The focal accumulation of affect alternative splicing, thus suggesting MeCP2 has multi- HSATII transcripts is unique to HSATII, as stable ectopic functional roles in regulation and splicing (Young et al., expression of α-sat RNA does not result in focal RNA Chromosoma (2021) 130:75–90 77 accumulation in the nucleus. Further, HSATII RNA accumu- to less than 5% of α-sat and empty vector control transfected lates in cis, immediately adjacent to the genomic integration cells (Fig. 1b, d). Nucleoplasmic and cytoplasmic diffuse ex- site, and recruits MeCP2 to these nuclear foci. Expression of pression was also observed at this early timepoint, likely due both α-sat and HSATII transcripts leads to cell division de- to high levels of expression driven by the CMV promoter. fects including chromatin bridges, blebbing, and micronuclei Cells transfected with α-sat displayed a similar level of ex- formation, thus indicating that expression of both centromeric pression, with roughly 23% of cells displaying α-sat RNA by and pericentric satellite sequences has the potential to generate RNA FISH (Fig. 1c, e). However, a striking difference was chromosomal instability and aberrant cell division. The study observed in the distribution of HSATII and α-sat RNA in the of the effect of HSATII expression represents a unique oppor- nucleus. Distinct focal accumulations of HSATII RNA (2–3 tunity to shed light on the function of pericentric satellite se- per nucleus on average) were observed (Fig. 1b), while α-sat quences more generally, in addition to testing the functional RNA appeared as a diffuse, primarily nuclear RNA signal consequences of expression of pericentric satellite sequences (Fig. 1c). Expression of HSATII or α-sat was dependent on in cancer. transfection with the respective insert-containing vector, thus demonstrating construct delivery specificity, which was ob- served upon three independent transient transfections. Further, Results the percentage of cells expressing the desired sequence insert was significantly different from controls (empty vector) (Fig. Establishment of a system to study the 1f). RT-qPCR also confirmed high levels of HSATII expres- effect of HSATII RNA expression sion in HSATII-transfected cell lines compared to alpha-sat transfected and controls (Fig. 1g). Since RT-qPCR was per- Expression of HSATII RNA in cancer cells has previously formed from total cellular RNA, results here cannot distin- been observed in both cancer cells grown in culture and from guish between nuclear RNA accumulations and diffuse tissue biopsies (Hall et al., 2017; Ting et al., 2011). While there RNA (nuclear or cytoplasmic), thus the greater than eightfold are distinct loci that harbor HSATII DNA sequences within the increase in HSATII expression shown for one transfection pericentric regions of roughly 11 human chromosomes, only a (Fig. 1g), likely illustrates the total amount of HSATII over- subset of these HSATII sequences are expressed, and the expression compared to α-sat and control cells. HSATII RNA transcripts expressed from these loci remain in While HeLa represents an easily transfected cell line that does cis in CAST bodies. HSATII sequences resident on human not express HSATII RNA, they are cancer cells that have been chromosomes 7 and 10 are among those sequences displaying maintained in culture since the , thus are not representative a preference for expression in a variety of different cancer cell of “normal” cells. Following successful demonstration of tran- lines and tissues (Hall et al., 2017). In order to study the direct sient HSATII and α-sat expression in HeLa cells, stably effect of expression of HSATII RNA, we developed a cell transfected cell lines were generated for both HeLa and a primary culture model to stably express HSATII sequence derived (non-transformed) human fibroblast cell line to ensure that the from Chr 7 in cell lines that do not endogenously express results observed were not simply an effect of expressing satellite HSATII. To further examine the effect of HSATII expression RNA in cancer cells, in which the nuclear and chromatin envi- irrespective of its location of expression, stable cell lines were ronment can be drastically different from normal cells (Carone created in which the Chr 7 HSATII-expression construct had and Lawrence, 2013; Zink et al., 2019). Tig-1 cells are female been randomly integrated into the genome. primary fibroblasts maintained at low passage number that retain An HSATII cDNA sequence derived from Chr 7 was an inactive X chromosome and a stable karyotype (2n= 46) cloned into a designed for mammalian expression (Ohashi et al., 1980), do not express HSATII (Hall et al., and stable integration, containing a CMV promoter and a neo- 2017), and can be transfected using lipid-mediated transfection. mycin selectable marker (Fig. 1a). HeLa cells, while a cancer A lipid-mediated transfection protocol was optimized for suc- cell line, do not endogenously express HSATII RNA (Hall cessful transfection of Tig-1 primary fibroblasts, to ensure high et al., 2017); thus, initial transfection experiments were con- transfection efficiency in addition to high cell viability (Fig. S1). ducted in HeLa cells due to their ease of transfection by lipid- Following 2 weeks of neomycin selection, mock mediated transfection. To test for HSATII-specific effects, transfected and control HeLa and Tig-1 cultures displayed control vectors containing an α-sat sequence derived from complete cell death, while a subset of viable cells (ranging Chr 4 and no insert (empty vector) were also used concurrent- from 10-15%) harboring HSATII, α-sat, and empty vector ly to transfect HeLa cells (Fig. S2). Cells were assayed for integrants were selected and further expanded for a period of transient satellite expression 24 h after transfection by RNA several weeks. Cells from three independent stably transfected FISH and RT-qPCR. Twenty-four hours after transfection, cell lines were fixed for DNA/RNA FISH and pelleted and approximately 20% of HSATII-transfected HeLa cell nuclei harvested for RNA extraction at weekly intervals to assay displayed nuclear accumulations of HSATII RNA compared satellite expression (Fig. 1a). To test for genomic integration 78 Chromosoma (2021) 130:75–90

Expression a construct

Lipid-DNA complex

Cell membrane

Nucleus

Add to plated cells

2 week neomycin selection Cytoplasm (stable lines)

b HeLac HeLa

DAPI DAPI HSATII RNAA HSATII RNA αSAT RNAA αSAT RNA

d 25 e 25 HSATII Alpha Sat 20 nuclear expression nuclear expression 20 (24 hours) (24 hours) 15 15

10 10

5 5 Percent of cells scored Percent of cells scored 0 0 y a y Alph Empt HSATII Alpha Empt HSATII Transfection Vector Transfection Vector

25 f *** g 10 *** 20 8 HSATII/ HSATII nuclear expression BACT 15 Alpha nuclear expression 6 10 4 5 2 0 Relative RT-qPCR enrichment 0

Percent cells with nuclear expression a HSATII Alpha Empty Alph Empty Transfection Vector HSATII Control Transfection Vector Fig. 1 Transient transfection of satellite expression results in nuclear FISH. Percent of cells (out of 500 nuclei) with (d) HSATII RNA nuclear satellite RNA accumulation. (a) Transfection scheme for transient and expression and (e) α-sat nuclear expression. (f) Nuclear RNA signal stable integration expression. A plasmid harboring HSATII, α-sat detected by RNA FISH is dependent on the sequence harbored within (αsat), or no insert (empty vector) is introduced to cultured HeLa or the transfected construct. Asterisks denote significant differences from Tig-1 primary fibroblast cells via lipid-mediated transfection and cells empty vector transfected cells by Chi-square test, p<.0001. (g) qRT- are then fixed on coverslips or harvested for RNA isolation. Stable cell PCR of HSATII RNA 24 h after transfection. Expression values shown lines are further selected with neomycin (G418) for 2 weeks prior to relative to housekeeping gene expression (ß-) for each transfection fixation or harvesting. Twenty-four hours after transfection, nuclei are condition. Scale bar, 5μm scored for expression of (b)HSATIIand(c) α-sat RNA signal by RNA Chromosoma (2021) 130:75–90 79 of the , DNA hybridization was performed integration, 15 of those also had focal HSATII RNA signal, using a biotin-labeled probe complementary to the plasmid with 8 of those demonstrating an adjacent localization via (pTargeT) backbone, which indicated that the majority of cells sequential RNA-DNA FISH experiments. An example of ac- with detectable signal had one or two sites of integration (Fig. cumulated HSATII RNA without adjacent localization to the S3). Analysis of chromosome spreads harvested from pTargeT backbone is shown in Figure S5.Itislikelythatthere transfected cell lines confirmed random integration of the are genomic locations into which the expression vector has HSATII expression vector, with the majority of spreads with integrated that are not permissible to , that will hybridization signal harboring one or two integration sites, not tolerate accumulation of the RNA, or for which the CMV which were randomly distributed on chromosomes (Fig. S4). promoter may no longer be active that may account for sites Three weeks following transfection, HeLa cells retained the that have no detectable HSATII RNA signal. It is also possible same pattern of satellite expression, with α-sat RNA being that accumulation of HSATII RNA occurs only when proteins diffuse and nuclear (Fig. 2a). In contrast, ~5% of nuclei in are recruited to the RNA to form CAST bodies (Hall et al., HSATII-transfected HeLa cells displayed nuclear accumula- 2017), as it is currently unknown whether the transcripts alone tions of HSATII RNA. This pattern was similarly observed in accumulate. We observed that focal accumulation of ectopi- Tig-1 primary fibroblasts, with independent transfections cally expressed HSATII RNA is reminiscent of the focal ac- demonstrating 7–28% of HSATII transfected cells harboring cumulations in cancer cells (CAST bodies) in both their ap- nuclear accumulations of HSATII RNA (Fig. 2e, i)andα-sat pearance and in the pattern in which they form near their site RNA displaying a diffuse nuclear signal by RNA FISH (Fig. of transcription. Therefore, we sought to examine whether 2d). RT-qPCR further confirmed higher levels of HSATII HSATII RNA foci recruited MeCP2, one of the protein com- RNA in HSATII-transfected HeLa (Fig. 2h) and Tig-1 (Fig. ponents within HSATII CAST bodies in cancer cells (Hall 2j) cells. Of note, while transiently transfected cell lines et al., 2017), using co-/RNA FISH. In displayed some cytoplasmic satellite RNA, stably transfected HeLa cells, all HSATII RNA focal accumulations had an cell lines had very little cytoplasmic satellite RNA (α-sat or overlapping focal signal with MeCP2 (100% of those nuclei HSATII), suggesting that the vast majority of satellite RNA scored) (Fig. 3c–d). Colocalization analysis of HSATII RNA transcribed ectopically in these cell lines is retained in the and MeCP2 in HeLa cells indicated a Pearson correlation nucleus. Taken together, these results indicate that stably ex- coefficient of r~1, indicating that these signals are nearly per- pressing HeLa and primary fibroblast cell lines can be used to fectly overlapping, despite being detected in different fluores- examine the effect of ectopic satellite RNA (HSATII and α- cent channels (Fig. 3b). This was in contrast to a lack of sat) expression in cells that do not normally express HSATII. colocalization of HSATII RNA and pTargeT backbone Further, results from transient and stably transfected cell lines DNA in HeLa cells (−0.5< r <0.7) (Fig. 3b), as was expected indicate that ectopic HSATII RNA accumulates in nuclear based on their adjacency (Fig. 3a). In Tig-1 cells, some foci, while ectopic α-sat RNA is primarily diffuse in the nu- colocalization with MeCP2 was observed for a subset of cleus, suggesting that when these two satellite RNAs are HSATII RNA accumulations (Fig. 3g) (8% colocalization expressed in both of these cell types, their transcripts may for one transfected cell line scored). Since not all HSATII behave in very different ways within the nuclear environment, RNA accumulations in primary fibroblasts recruited MeCP2, irrespective of their site of expression within the genome. this may suggest that the cellular context may influence the potential for recruitment of MeCP2 into CAST bodies (Fig. HSATII RNA accumulates adjacent to the location S5b). It might also be possible that additional proteins are from which it is expressed and recruits MeCP2 recruited to CAST bodies independently of MeCP2, or that HSATII RNA, alone, has the ability to condense into focal In cancer cells that express HSATII RNA, the RNA accumu- accumulations in cis. Additionally, there were many focal lates in cis immediately adjacent to their site of transcription accumulations of MeCP2 in Tig-1 cells that did not overlap (Hall et al., 2017). Therefore, we asked whether the accumu- HSATII RNA accumulations (Fig. 3g). These MeCP2 foci lated HSATII RNA foci in stably transfected cell lines also were also observed in control cells (Fig. S5c) and likely rep- remain in cis. Sequential RNA FISH followed by DNA FISH resent normal nuclear accumulations of MeCP2 that do not confirmed HSATII RNA accumulated adjacent to the site into overlap HSATII transcripts. These data suggest that ectopical- which the expression construct (pTargeT backbone) was inte- ly expressed HSATII RNA accumulates in cis and remains in grated into the genome in both HeLa (Fig. 3a–b) and Tig-1 nuclear foci near, though not overlapping, their site of geno- cells (Fig. 3e–f). While not all sites of integration had a focal mic integration. Further, these HSATII RNA foci recruit accumulation of RNA, when HSATII RNA was detected, it MeCP2, a protein known to be in CAST bodies in cancer was adjacent to the location in which the vector had integrated cells. Differential MeCP2 recruitment in HeLa and primary in 53% of nuclei scored. For example, out of 53 Tig-1 HSATII fibroblast cells to HSATII RNA accumulations suggests stably transfected nuclei with observed HSATII DNA MeCP2 colocalization may be dependent on both the cellular 80 Chromosoma (2021) 130:75–90

HeLa Tig-1 a HeLa-αSAT d Tig1-αSAT

DAPI DAPI αSAT RNAA αSAT RNAA DAPI αSAT RNAA αSAT RNAA e Tig1-HSATII b HeLa- HSATII

DAPI HSATII RNA HSATII RNA HSATII RNA HSATII RNA f Tig1-Empty c HeLa-Empty

DAPI HSATII RNA HSATII RNA HSATII RNA HSATII RNA

0.020 gh6 40 ij8 HSATII/ HSATII/ 30 BACT 6 0.015 BACT 4

20 4 0.010

2 10 2 0.005

0 0 0.000 Relative Expression (RT-qPCR) 0

y Relative Expression (RT-qPCR) a a I I h

Percent cells with nuclear foci (Week 3) Percent cells with nuclear foci (Week Alpha Empty

Percent cells with nuclear foci (Week 3) Percent cells with nuclear foci (Week Alph Empty Alpha Empt HSATII Alp Empty HSATII HSATI HSATI Transfection Vector Transfection Vector Transfection Vector Transfection Vector Fig. 2 Stable expression of satellite RNA results in focal nuclear empty vector have no satellite RNA expression detectable by RNA FISH accumulation of HSATII RNA in a subset of cells. (a) Following (c, f). (g, i) Quantification of cells with HSATII nuclear expression 3 integration and neomycin selection for the expression construct weeks following transfection in HeLa (g)andTig-1(i). (h, j)RT-qPCR (HSATII, αSAT, empty), a subset of HeLa (a–c) and Tig-1 (d–f) nuclei of HSATII RNA in transfected cell lines 3 weeks following transfection. exhibit satellite RNA expression by RNA FISH. While α-sat RNA (a, d) Data is shown normalized to ß-actin (housekeeping gene). Error bars is primarily diffuse in the nucleus, HSATII RNA accumulates in distinct show 95% confidence interval. Scale bar, 5μm nuclear foci (b, e) in both HeLa and Tig-1 cells. Cells transfected with an context and/or chromatin environment throughout the accumulates in a similar manner to HSATII RNA that is nucleus. endogenously expressed from pericentric regions. It has been demonstrated that aberrant α-satellite expression Satellite RNA expression induces cell division defects can lead to cell division defects and chromosomal instabil- and instability ity (Chan et al., 2017; Ichida et al., 2018;Zhuetal.,2018; Zhu et al., 2011). Therefore, we examined whether cell Analysis of cell lines established to stably express HSATII division defects could be observed in cells that were also RNA suggests that ectopically expressed HSATII RNA forced to express pericentric HSATII, in direct comparison Chromosoma (2021) 130:75–90 81

a HeLa-HSATII HeLa-HSATII HeLa-HSATII b 1.5

1.0

0.5

0.0 Pearson’s Coefficient Pearson’s DAPI -0.5 HSATII RNA HeLa HSATII RNA/ HeLa HSATII RNA/ pTargeT DNA HSATII RNA pTargeT DNA pTargeT MeCP2 c HeLa-HSATII HeLa-HSATII HeLa-HSATII d HeLa MeCP2 HSATII RNA linescan 50000 MeCP2 40000 HSATII RNA

30000

20000

Signal Intensity 10000 DAPI HSATII RNA 0 MeCP2 01234 HSATII RNA MeCP2 Distance[ m] e Tig-1-HSATII Tig-1-HSATII Tig-1-HSATII f Tig-1 pTargeT DNA HSATII RNA linescan

4000 pTargeT DNA HSATII RNA 3000

2000

Signal Intensity 1000 DAPI HSATII RNA 0 pTargeT DNA HSATII RNA pTargeT DNA 0123 Distance[ m]

Tig-1 MeCP2 HSATII RNA linescan g Tig-1-HSATII Tig-1-HSATII Tig-1-HSATII h 6000 MeCP2 5000 HSATII RNA

4000

3000

2000 Signal Intensity

DAPI 1000 HSATII RNA MeCP2 HSATII RNA MeCP2 0 012 Distance[ m] Fig. 3 HSATII RNA accumulates adjacent to site of integration and demonstrated by Pearson correlation coefficients plotted in (b). (d)Dual recruits MeCP2. (a) RNA hyb for HSATII RNA (red) followed by se- channel linescan of inset shown in (c) demonstrating colocalization of quential DNA hyb using a probe to detect the pTargeT backbone (green) MeCP2 (red) and HSATII RNA (green). (e) RNA hyb for HSATII RNA demonstrates HSATII RNA accumulates primarily adjacent to sites of (red) followed by sequential DNA hyb using a probe to detect the pTargeT genome integration in HeLa cells stably expressing HSATII 3 pTargeT backbone (green) demonstrates HSATII RNA accumulates ad- weeks after transfection. All boxes with white border indicate regions of jacent to, and not overlapping, sites of genome integration in Tig-1 cells interest with zoomed image at bottom right. Individual channels images stably expressing HSATII 3 weeks after transfection. (f) Dual channel shown in black and white. (b) Scatter plot of Pearson correlation coeffi- linescan of inset in (e) demonstrating the HSATII RNA (red) and cients for individual nuclei (n=10) hybridized for sequential HSATII pTargeT DNA (green) signals are adjacent, and not overlapping. (g) RNA/pTargeT DNA and HSATII RNA with MeCP2 co-IF Tig-1 fibroblasts stably expressing HSATII RNA for 3 weeks demon- (HSATII RNA/MeCP2). Black lines indicate mean; grey error bars indi- strate overlapping signals for HSATII RNA and MeCP2. Box with white cate 95% confidence interval. (c) Co-HSATII RNA hyb and IF for border indicates region of interest with zoomed image at bottom right. (h) MeCP2 indicate a near perfect colocalization of HSATII RNA and Dual channel linescan of inset shown in (g) demonstrating colocalization MeCP2 signals within HeLa nuclei 3 weeks following transfection, as of MeCP2 (red) and HSATII RNA (green). Scale bar, 5μm

to cells expressing ectopic α-sat RNA. Stable cell lines controlcells.Incontrast,at21–28 days following transfec- expressing HSATII and α-sat RNA were compared to cells tion,anincreaseinchromatinand cell division defects was with stable integration of the pTargeT backbone alone observed, including an increase in the presence of chroma- (empty vector) and control (untransfected) cells at identical tin bridges (CB), nuclear blebbing/micronuclei (MN), lag- passage numbers. Two weeks following transfection, some ging chromosomes (LC), and abnormally shaped nuclei cell division defects were observed in a small percentage of (Fig. 4a) in HeLa cells. While a significant increase in cells; however, this was not significantly different from chromatin bridges and overall chromatin defects (sum of 82 Chromosoma (2021) 130:75–90 all defects) was observed in both HSATII and α-sat ex- Discussion pressing cell lines compared to empty vector transfected cells, an increase in blebbing alone was significant (Chi- In order to determine the consequences of the presence of square, p<.05) only in HSATII-expressing HeLa cells. As HSATII RNA within cells, we developed a cell culture system expected for karyotypically normal fibroblast cells, the fre- to test the effect of HSATII expression by establishing inde- quency of total cell division defects in Tig-1 was much pendent cell lines that stably express either HSATII or α-sat lowerthaninHeLa(Fig.4b). Observed cell division de- satellite sequences from randomly integrated sites in the ge- fects were significantly different for all HSATII or α-sat nome. We found that centromeric α-sat and pericentromeric expressing Tig-1 cell lines 24 days after transfection com- HSATII transcripts behave differently when ectopically pared to empty vector transfected cells, including increases expressed, yet the presence of HSATII and α-sat transcripts in chromatin bridges and nuclear blebbing/micronuclei, both trigger cell division defects. HSATII RNA accumulates which are only very rarely seen (1 cell per 500 cells) in in nuclear bodies in cis, immediately adjacent to the integra- this primary fibroblast cell line when normally maintained tion site (Fig. 5), while α-sat RNA is overall more diffuse and at low passage numbers. Similar levels of cell division does not accumulate appreciably in nuclear bodies, suggesting defects were observed in α-sat and HSATII stably express- that these two distinct tandemly repeated RNAs have different ing cells established from independently transfected and dynamics within the nuclear environment, and supporting that maintained HeLa and Tig-1 cell lines (Fig. S6). Primary they may have distinct functions. Nuclear accumulations of fibroblast cells exhibited increased cell division defects HSATII RNA recruit MeCP2 into these nuclear bodies, which irrespective of which satellite RNA (HSATII or α-sat) are reminiscent of CAST bodies in cancer cells (Hall et al., was expressed (Fig. 4b and Fig. S6d–f), whereas HeLa 2017)(Fig.5). Although the transcripts localize in very dif- cells only exhibited a significant increase in blebbing/MN ferent ways, expression of both α-sat and HSATII induce cell andabnormallyshapednucleiinHSATII-expressingcells, division defects including chromatin bridges, blebbing, and suggesting that there is an HSATII-expression-dependent micronuclei formation, thus have the potential to impact cell effect in HeLa cells that is not present in primary fibro- division and cause further instability. blasts (Fig. 4). This is intriguing in light of previous evi- Previous research has focused on the effect of expression of dence that HeLa cells exhibit high levels of α-sat expres- alpha satellite RNA, which is normally expressed at low levels sion when normally maintained in culture (Hall et al., in all cell types and thought to be involved in centromere 2017). Thus, it is possible that HeLa cells respond more protein recruitment (Hall et al., 2017; Johnson et al., 2017; marginally to increased α-sat expression, whereas the pres- McNulty et al., 2017;Wongetal.,2007). Though a low level ence of HSATII RNA induces a more drastic range of of α-sat expression is normal and likely necessary for centro- phenotypes. In contrast, untransfected Tig-1 cells exhibit mere protein recruitment, several groups have described that less α-sat expression (Hall et al., 2017), thus may be more when α-sat is overexpressed (Chan et al., 2017;Ichidaetal., sensitive to expression of alpha satellite in the nucleus. 2018; Zhu et al., 2018) or knocked down (Ideue et al., 2014; Supporting this, karyotypic analysis of stably expressing Rošić et al., 2014), this has detrimental effects on cell division, Tig-1 cells maintained in culture for 60 days after transfec- suggesting that the levels of α-sat expression are critical to tion revealed that while the majority of empty vector maintaining proper cell division. In contrast, HSATII RNA is transfected and HSATII expressing cells had a normal kar- not expressed in normal cells, but is highly overexpressed in yotype (2n= 46), α-sat expressing cell lines had an in- cancer cells and tissues (Hall et al., 2017; Ting et al., 2011). creased number of chromosome spreads exhibiting aneu- HSATII RNA accumulates in large nuclear foci within these ploidy (Fig. S7). Forced alpha satellite expression has cells, where it recruits nuclear regulatory proteins including been previously shown to induce cell division defects MeCP2. Thus, previous evidence suggested that α-sat and and generate chromosomal instability (Chan et al., HSATII satellite sequences are likely to have unique regula- 2017;Ichidaetal.,2018; Zhu et al., 2018; Zhu et al., tory mechanisms governing their expression. Our results here 2011); therefore, similar mechanisms may be induced suggest that in addition to distinct transcriptional regulatory upon aberrant α-sat expression in primary fibroblasts. mechanisms, the transcripts themselves behave in unique Further investigation will be required to determine the ways within the nuclear environment. The accumulation of mechanism by which satellite expression impacts cell HSATII RNA and recruitment of MeCP2 is reminiscent of division and results in the formation of chromatin brid- “toxic repeat RNAs,” which function to sequester nuclear reg- ges and micronuclei, but the observed increase in these ulatory proteins in diseases such as frontotemporal dementia/ cell division aberrations in cells stably expressing α-sat ALS and myotonic dystrophy (Swinnen et al., 2020). Further or HSATII suggests that expression of either pericentric work will be required to understand the molecular interactions or centromeric satellite sequences induces cell division between HSATII RNA and MeCP2 and the cellular condi- defects. tions required for MeCP2 recruitment to HSATII RNA foci, Chromosoma (2021) 130:75–90 83

a

Bleb MN CB

CB DAPI HeLa-HSATII

Bleb 4 Chromatin bridges

*** *** CB 3

2

1 DAPI HeLa-HSATII DAPI HeLa-HSATII 20 0 Blebbing/MN 25 *** Total chromatin defects

Percent cells with chromatin bridges 15 *** 20 HSATII Alpha Empty Control Transfection Vector 15 10 *** 10 5 5

Percent cells with blebbing 0 0

TII TII A HS Empty HSA Alpha Empty Control Alpha Control Percent cells with chromatin defects Transfection Vector Transfection Vector b

Bleb

Bleb/MN LC MN

DAPI Tig-HSATII

1.5 Chromatin bridges ***

1.0 *** DAPI Tig-HSATII DAPI Tig-HSATII 0.5

1.0 8 ** ** Blebbing/MN Total chromatin defects *** 0.0 0.8 6

Percent cells with chromatin bridges ATII 0.6 ** HS Empty Alpha Control 4 0.4 Transfection Vector 0.2 2

Percent cells with blebbing 0.0 0

TII

Percent cells with chromatin defects A HSATII Alpha Empty Control HS Alpha Empty Control Transfection Vector Transfection Vector 84 Chromosoma (2021) 130:75–90

ƒFig. 4 Ectopic expression of satellite RNA leads to cell division defects. HSATIII RNA, which is transcribed from Chr 9q12 during (a) Representative images of HeLa cells transfected and stably expressing stress, assembles into nuclear stress bodies (nSBs), recruiting HSATII RNA for 3 weeks exhibiting cell division defects including chromatin bridges (CB), blebbing (Bleb), and the formation of HSF1 and numerous splicing factors to promote intron reten- micronuclei (MN), as demonstrated in the images in the top 3 panels tion and suppress splicing of mRNAs during recovery from (DAPI stained, greyscale). Arrows indicate the type of defect observed. heat shock (Biamonti and Vourc'h, 2010; Jolly et al., 2004; Graphs depict percentages of total cells counted displaying these defects: Ninomiya et al., 2020). Intriguingly, HSATII and HSATIII chromatin bridges (bottom left), blebbing/MN (bottom middle), and total chromatin defects (bottom right) out of 500 total cells for each transfec- both derive from CATTC pentamer repeats, despite their di- tion vector used or control (untransfected) cells. (b) Representative im- vergence and distinct chromosomal localizations. HSATII ages of Tig-1 normal fibroblast cells transfected and stably expressing displays a widespread distribution within the pericentric re- HSATII RNA for 3 weeks displaying cell division defects (lagging chro- gions of ~11 human chromosomes, while the bulk of mosomes (LC), blebbing, MN) as indicated by arrows. Graphs depict percentages of cells displaying these defects: chromatin bridges (bottom HSATIII resides on Chr 9 and Y, with some smaller loci left), blebbing/MN (bottom middle), and total chromatin defects (bottom interspersed within pericentric regions of additional chromo- right) out of 500 total cells for each transfection vector used (X axis). somes (Altemose et al., 2014; Tagarro et al., 1994). HSATII Frequencies of chromatin defects significantly (p<.05) different from and HSATIII RNA appear to recruit different protein-binding empty vector transfected cells by Chi-square test are indicated with two asterisks (**). Frequencies of chromatin defects significantly (p<.01) dif- partners, yet they both display conserved functional aspects in ferent from empty vector transfected cells by Chi-square test are indicated their ability to accumulate in cis and recruit/sequester protein with three asterisks (***). Untransfected control cells included for com- binding partners. It is possible that the tandemly repetitive parison. Scale bar, 5μm nature of pentamers embedded within HSATII/HSATIII re- peats may facilitate this recruitment given that identical se- quences (monomers) are present in high copy number within a linear array. Nuclear stress bodies form on HSATIII non- but our results suggest that the recruitment of MeCP2 is likely coding RNA and recruit proteins into nuclear condensates, a to be a sequence-dependent mechanism, rather than a theme which is emerging as a conserved feature of RNP gran- location-dependent mechanism since HSATII is expressed ules that are likely to exist within liquid-liquid phase separated here from random integration sites in both HeLa and Tig-1 domains in the nucleus. Recent work highlights the role of primary fibroblasts (Figs. S3 and S4).

Fig. 5 Timeline of ectopic Timeline of ectopic HSATII expression HSATII expression. HSATII RNA (green) is expressed tran- siently 24 h after transfection in multiple nuclear accumulations Transfection of HSATII and some diffuse nuclear RNA. expression construct Following antibiotic selection for 2 weeks (14 days), HSATII RNA bodies form, which recruit MeCP2 (orange), and accumulate 24 hours adjacent to the genomic integra- HSATII expressed tion site of the HSATII expression -Nuclear accumulation

vector (purple). At this point, Interphase nucleus KEY HSATII RNA is restricted to HSATII RNA these nuclear bodies and little Antibiotic selection Integration site diffuse nuclear RNA is observed. HSATIIHSATII RRNANA wwithith MMeCP2eCP2 Approximately 3 weeks after Chromatin/nucleus transfection (21 days), significant cell division defects are observed, HSATII body formation including lagging chromosomes ~ 14 days - random integration (LC) between dividing daughter - RNA accumulation in cis cells and the formation of -recruitment of MeCP2 micronuclei (MN) Interphase nucleus

MN HSATII expression-induced chromatin defects MN LC ~ 21 days -lagging chromsomes -blebbing

-micronuclei formation Dividing G1 DaughtersDAPI Chromosoma (2021) 130:75–90 85

RNA in modulating the biophysical properties of liquid drop- et al., 2018; Zhu et al., 2011). Overexpression of α-sat has lets, and the accumulation of misregulated transcripts may also been linked to chromosomal instability (CIN) and segre- lead to mislocalization of RNA-binding proteins containing gation errors resulting in copy number changes in daughter intrinsically disordered domains (IDRs) (Maharana et al., cells (Ichida et al., 2018). Yet, paradoxically, other studies 2018). MeCP2, though classically described as a DNA- found no cell division phenotypes associated with overexpres- binding protein, has been more recently identified in an sion of pericentric satellite sequences (Ideue et al., 2014; RNA-binding protein (RBP) screen (Castello et al., 2016) Rošić et al., 2014). It is important to note that the cell division and binds mRNAs in the brain, where it affects alternative defects we observe here are long-term effects of ectopic splicing (Young et al., 2005). MeCP2 also contains an IDR, HSATII expression resulting 3–4 weeks after the initial trans- and is bound to HSATII RNA within CAST bodies, which we fection; thus, the cell division defects we observe here may note also resemble nuclear condensates. reflect long-term effects of continual satellite overexpression. Since initially found to be expressed in cancer cells, the Furthermore, differences in the effects of satellite overexpres- functional consequence of HSATII transcripts within the nu- sion could be due to the location of the satellite DNA se- cleus has remained elusive; however, induction and injection quences in the genome (i.e., centric/pericentromeric) or the of HSATII RNA were found to result in the formation of localization and function of the satellite RNA (i.e., cis vs cDNA intermediates, which can facilitate the expansion of trans). Results obtained here for HSATII RNA expression pericentromeric HSATII-containing DNA at endogenous sites support previous reports of centromeric satellite expression- in the genome (Bersani et al., 2015). Though the mechanism induced cell division defects (Bouzinba-Segard et al., 2006; facilitating this copy number gain is not well understood, it is Chan et al., 2017;Ichidaetal.,2018;Sleeetal.,2012), and thought to result from the formation of RNA-DNA hybrids extend this phenomenon to expression of pericentromeric following reverse transcription of HSATII RNA transcripts. In HSATII RNA. While we did observe changes in chromosome addition to cancer cells, HSATII RNA has recently been number following long-term α-satellite expression, we did not found to be induced upon herpesvirus infection, where it is observe changes in chromosome number in HSATII- thought to be involved in the viral cycle (Nogalski et al., expressing cells (Fig. S7); thus, our evidence does not suggest 2019) and in human FSHD cell models where it resides in that long-term expression of HSATII RNA leads to CIN. nuclear dsRNA foci that also aggregate proteins (Shadle Despite their abundant presence within pericentromeres et al., 2020). HSATII RNA has also been implicated in the and evidence suggesting that low-level expression of satellite growth of cancer cells (Ting et al., 2011), and its expression is RNA may be integral to both centromere protein recruitment further induced in cells grown under 3D culturing conditions andcelldivision(Johnsonetal.,2017; McNulty et al., 2017), (Bersani et al., 2015). Thus, there may be convergent mecha- the role of tandemly repeated satellite sequences within these nisms by which HSATII RNA can more globally affect both regions has been elusive. Recent work suggests that growth and viral replication. Though these effects have been pericentric satellites within fruit fly and mouse heterologous observed, the specific role that HSATII transcripts play in chromosomes may act to tether chromosomes within nuclei mediating these phenotypes has remained unclear. Our obser- via the formation of chromocenters (Jagannathan et al., 2018). vation that HSATII RNA accumulates and recruits regulatory Perturbation of proteins within chromocenters results in the protein in cis, irrespective of the site from which it is formation of micronuclei due to budding from the nucleus, expressed, may shed some significant insight into the molec- supporting that tethering of pericentric regions to chromocen- ular function of HSATII RNA. The ability of HSATII RNA to ters is key to nuclear organization and overall nuclear integrity bind, and potentially sequester, large amounts of regulatory (Jagannathan et al., 2018). Alteration of epigenetic regulatory proteins is likely to have a large effect on transcription and mechanisms within heterochromatin is also known to cause genome regulation more broadly, and may ultimately lead to chromosomal instability (Slee et al., 2012); thus, impaired widespread heterochromatin instability, which is observed in maintenance of heterochromatin is likely to lead to both sat- cancer and other diseases (Carone and Lawrence, 2013). ellite expression and chromosomal instability. Our results here Supporting a more global effect of HSATII accumulations add to a growing body of evidence suggesting that the pres- within the nucleus, our results implicate HSATII transcripts in ence of aberrant satellite transcripts, themselves, are likely to generating cell division defects. When HSATII RNA is induce defects in chromosome segregation. This observation expressed ectopically in both HeLa and Tig-1 primary fibro- is supported by previous studies of transient α-sat expression, blasts for several weeks, we observed increased frequencies of but the more long-term and specific effect of the presence of chromatin bridges, lagging chromosomes, and micronuclei satellite RNA has been unclear. By randomly integrating and formation (Fig. 4 and Fig. S6). These phenotypes have previ- expressing ectopic HSATII and α-sat RNA, we demonstrate ously been observed upon introduction of satellite transcripts the ability to observe the specific behavior of these satellite into both mouse and human cells, where they may induce transcripts within the nucleus, irrespective of their site of tran- DNA damage via the formation of RNA-DNA hybrids (Zhu scription. Despite displaying strikingly different nuclear 86 Chromosoma (2021) 130:75–90

Materials and methods

Key Resource Table

Reagent type Designation Source or Identifiers Additional information (species) or reference resource Antibody DyLight 488 Streptavidin Vector Vector (1:500) Laboratories Laboratories:SA-5488 Antibody DyLight 594 Streptavidin Vector Vector (1:500) Laboratories Laboratories:SA-5594 Antibody MeCP2 (D4F3) XP® Rabbit mA Cell Signaling Cell Signaling:3456 (1:250) Technology Antibody Goat anti-Mouse IgG (H+L) Invitrogen Thermo Fisher:A-11005 (1:500) or (1:250) Cross-Adsorbed Secondary (Thermo Antibody, Alexa Fluor 594 Fisher Scientific) Cell line HeLa (H. sapiens) Cell line Tig-1 Coriell Coriell:AG06173 (H. sapiens) Chemical Geneticin™ Selective Antibiotic Gibco (Thermo Thermo compound, (G418 Sulfate) Fisher Fisher:10131027 drug Scientific) Commercial StrataClone PCR Cloning Kit Agilent Agilent:240205 assay or kit Commercial pTargeT™ Mammalian Expression Promega Promega:A1410 assay or kit Vector System Commercial iScript™ Reverse Transcription Bio-Rad Bio-Rad:1708841 assay or kit Supermix for RT-qPCR Commercial LipofectamineⓇ LTX with Invitrogen Thermo assay or kit PLUS™ Reagent (Thermo Fisher:15338030 Fisher Scientific) Commercial FuGENE® HD Transfection Promega Promega:E2311 assay or kit Reagent Sequence-based Alpha_sat_Chr4_F Integrated DNA 5′-CTGCACTACCTGAAGAGGAC-3' reagent (used for cloning ASAT) Technologies Sequence-based Alpha_sat_Chr4_R Integrated DNA 5′-GATGGTTCAACACTCTTACA-3' reagent (used for cloning ASAT) Technologies Sequence-based HSATII_F (used for cloning and Integrated DNA 5′-GGAACCGAATGAATCCTCATTGAATG reagent qRT-PCR of HSATII) Technologies -3' Sequence-based HSATII_R (used for cloning and Integrated DNA 5′-ATTCGATTCCATTCGATGATGATTCC-3' reagent qRT-PCR of HSATII) Technologies sequence-based asat_FITC_DNA_probe Integrated DNA 5′-/56-FAM/CCTTTTGATAGAGCAGTTTT reagent Technologies GAAACACTCTTTTTGTAGAATCTG CAAGTGGATATTTGG-3' Sequence-based HSATII_bio_LNA_probe Qiagen GG 5′biotin-ATTCCATTCAGATTCCATTC reagent #LCD0156269-BKH, GATC-3' product 339500 Sequence-based bACT F QRT (used for qRT-PCR) Integrated DNA 5′-CATGTACGTTGCTATCCAGGC-3' reagent Technologies Sequence-based bACT R QRT (used for qRT-PCR) Integrated DNA 5′-CTCCTTAATGTCACGCACGAT-3' reagent Technologies Transfected GFP construct Other Gift from P. Jones laboratory construct (H. sapiens) Transfected pTargeT-ASAT This paper construct (H. sapiens) Transfected pTargeT-HSATII This paper construct (H. sapiens) Chromosoma (2021) 130:75–90 87 localization, the presence of both α-sat and HSATII RNA 312.5 μL LipofectamineⓇ LTX and supplemented with leads to chromatin bridge formation and the formation of 12.5 μLPLUS™ Reagent. Twenty-four hours following micronuclei, suggesting convergent roles for these transcripts transfection, cells on coverslips were fixed for FISH ex- in the perturbation of cell division. periments and cells in flasks were pelleted and frozen for RNA extraction and qRT-PCR analysis. Stably transfected lines of HeLa cells were cultured in T-25 flasks, also using a ratio of 3 μg plasmid DNA:315 μL Construction of DNA LipofectamineⓇ LTX and 12.5 μLPLUS™ Reagent per flask. In parallel with the flasks transfected with the empty The pTargeT™ Mammalian Expression Vector System pTargeT™ vector, pTargeT-αSAT, and pTargeT-HSATII (Promega, Madison, WI), containing a CMV promoter and plasmids, untransfected HeLa cells were cultured in parallel neomycin resistance, was used for transfection and long- as a control. Cells were transfected 24 h after seeding in flasks, term selection of stably transfected lines. Insert DNA was and then maintained in standard growth medium containing derived from total RNA extracted from U2OS cells and then 10% FBS for 3 days before adding G418 sulfate- reverse transcribed using an iScript™ Reverse Transcription supplemented media. Cells were cultured in selective media Supermix (Bio-Rad Laboratories, Hercules, CA). A 140-bp for 2 weeks, during which the media was exchanged every 2 region of alpha satellite sequence from chromosome 4 days. After this selection period, the stably transfected lines (GenBank M38467.1) and a 349-bp region of HSATII from were maintained for an additional several weeks. Cells were chromosome 7 were independently cloned into pTargeT™ then plated on coverslips and fixed for FISH experiments DNA backbone via StrataClone TA cloning (Agilent twice a week, and harvested 3 and 4 weeks following the Technologies, Santa Clara, CA). end of selection, for RNA extraction and qRT-PCR analysis. To optimize lipid-mediated transfection for Tig-1 pri- Cell culture and transfection mary fibroblasts for both expression levels as well as overall health of transfected cells, a GFP expression vec- HeLa human epithelioid cervix carcinoma cells and Tig-1 tor (obtained from P. Jones) was transiently transfected human primary fibroblast cells (Coriell AG06173) were into cells on coverslips using a range of conditions, with cultured in the presence of 5% CO2 prior to transfection 3 μg plasmid DNA transfected per well of a 6-well plate. in Minimal Essential Medium (Gibco, Thermo Fisher LipofectamineⓇ LTX was tested at volumes of 9 μLand Scientific, Waltham, MA) supplemented with 10% 12 μL per well, and FuGENE® HD was tested at a range (HeLa) or 15% (Tig-1) fetal bovine serum (HyClone, of concentrations (3:1, 4:1 and 5:1 per μg DNA). After 24 GE Healthcare Life Sciences, Marlborough, MA; h, media was exchanged and cells maintained in standard Avantor Seradigm, VWR, Radnor, PA), 100 units/mL media for an additional 24 h. Cells were then visualized in penicillin and 100 μg/mL streptomycin (Gibco, Thermo culture prior to fixation and DAPI staining as described Fisher Scientific) and 2 mM L-glutamine (Gibco, Thermo below. Transfection efficiency was evaluated by scoring Fisher Scientific). After transfection, cells were main- GFP expression and cellular health gauged by scoring the tained in the above media, replacing the penicillin/ number of cells per field of view. streptomycin with geneticin (G418 sulfate) (Gibco, Tig-1 cells were stably transfected using FuGENE® Thermo Fisher Scientific) at a concentration of 700 μg/ HD Transfection Reagent (Promega), following the man- mL for HeLa cells and 500 μg/mL for Tig-1 cells for ufacturer’s protocol. Cells in T-25 flasks were grown to selection over a minimum of 14 days. 60% confluency and then transfected using 8 μgplasmid Transient transfections of the empty pTargeT™ vector, DNA and 44 μL FuGENE® HD Transfection Reagent per pTargeT-αSAT, and pTarget-HSATII were performed in flask. Three control flasks were cultured in addition to the parallel in HeLa cells using LipofectamineⓇ LTX with pTargeT lines: a wild-type untransfected flask not treated PLUS™ Reagent (Invitrogen, Thermo Fisher Scientific). with selective media, a wild-type untransfected flask se- Cells were seeded on 22×22mm coverslips or in T-25 lected with G418 sulfate, and a mock flask treated with flasks such that they were 70–90% confluent at the time the transfection reagent and then selected with G418 sul- of transfection. The plasmid DNA-lipid complexes for fate. Media containing transfection reagent was removed transfection were prepared following the manufacturer’s 24 h post-transfection and replaced with selective media. protocol. For cells in 6-well plates, 2.5 μg DNA was Selection was complete after 10–14 days, after which the transfected per well using a 2.5 μg:22.5 μL ratio of plas- stably transfected cells were maintained for several mid DNA:LipofectamineⓇ LTX, supplemented with weeks. During this time, cells were fixed on coverslips 3.5 μLPLUS™ Reagent. Cells in T-25 flasks were weekly and harvested for RNA extraction at multiple time transfected with 3 μg plasmid DNA per flask, with points. 88 Chromosoma (2021) 130:75–90

Preparation of chromosome spreads for hybridization, 50 ng of nick translated probe per coverslip was combined with 12 μg of human Cot-1 DNA (Roche), Chromosome spreads were prepared as previously described 10 μg salmon sperm ssDNA (Sigma-Aldrich), and 20 μg (Howe et al., 2014), with the following modifications. Cells in E. coli tRNA (Sigma-Aldrich), dried with a Speed Vac, and flasks were treated with colcemid for 1–2 h depending on the then resuspended in 10 μLformamide.Coverslipswerefirst cell line and monitored for mitotic cells, at which point the treated with 0.2 N NaOH in 70% ethanol to remove RNA, and cells were harvested, treated with hypotonic solution (0.075M then dehydrated in 100% ethanol and air dried. Interphase KCl), and fixed with Carnoy’s fixative (3:1 ratio of cells or chromosome spreads were denatured in 70% formam- methanol:glacial acetic acid). Fixed cells were then dropped ide in 2× SSC, pH 7.0, for 2 min at 80°C, followed by 5 min in onto clean glass slides under humid conditions to produce cold (4°C) 70% ethanol and 5 min in cold 100% ethanol prior chromosome spreads. Slides were serially dehydrated with to air drying. Probes were diluted in the hybridization buffer ethanol and stored at − 20°C. Thawed slides were dehydrated used for RNA FISH, to a final concentration of 50% formam- in 100% ethanol prior to DNA FISH experiments. ide, before applying to coverslips and incubating overnight at 37°C in a humid chamber. Coverslips were then washed as for Fluorescence in situ hybridization (FISH) and RNA FISH, with the adjustment of the first wash to 50% immunostaining formamide/2× SSC. The probe was detected with anti- digoxigenin-fluorescein (Roche), diluted 1:500 in 4× SSC+ For all FISH and immunostaining on interphase cells, 1% BSA. Secondary incubation, washes, DAPI staining and transfected cell lines were grown on 22×22 mm coverslips mounting were carried out as for the RNA oligo hybridization. and then fixed as follows: 3–5 min in cytoskeletal buffer RNA and DNA co-hybridization was performed by first com- (Byron et al., 2013) plus 0.5% Triton X-100 (Sigma- pleting the RNA hybridization as described, followed by fix- Aldrich, St. Louis, MO) and 10 mM ribonucleoside vanadyl ation in 4% paraformaldehyde for 10 min prior to DNA FISH, complex (New England Biolabs, Ipswich, MA), 10 min in 4% with the removal of the NaOH treatment step. paraformaldehyde (Ted Pella, Redding, CA) in 1× PBS, In HeLa cells, MeCP2 and HSATII RNA co-visualization followed by storage in 70% ethanol at 4°C. was performed by first hybridizing oligo probes to RNA as RNA FISH was performed using a FITC-labeled DNA described, followed by a 10-minute fixation in 4% parafor- oligo probe complementary to alpha satellite (Integrated maldehyde prior to staining for MeCP2. For Tig-1 cells, DNA Technologies, Newark, NJ) or a biotinylated locked MeCP2 was stained first, and then the signal was fixed in (LNA) oligo probe complementary to HSATII 4% paraformaldehyde for 10 min before proceeding with the (QIAGEN, Hilden, Germany) (see Key Resources Table). RNA FISH protocol. The HSATII probe was detected with For each coverslip, 0.25 pmol of oligonucleotide was dena- DyLight 488 streptavidin and the MeCP2 with an Alexa Fluor tured for 10 min at 80°C in 30% formamide, and then diluted 594 goat anti-rabbit secondary. To stain for MeCP2, cover- in hybridization buffer to a final concentration of 15% form- slips were rinsed for 10 min in 1× PBS, then incubated with a amide. Hybridization buffer contained 2 mg/mL bovine serum 1:250 dilution of MeCP2 rabbit antibody (Cell Signaling albumin (BSA) (Roche, Basel, Switzerland), 10% dextran sul- Technology, Danvers, MA) in 1× PBS+1% BSA at 37°C for fate, and 2× SSC, supplemented with RNasin Plus RNase 1–3 h. Coverslips were washed at room temperature: 10 min Inhibitor (Promega). The probe was then applied to coverslips, in 1× PBS, 10 min in 1× PBS+0.1% Triton X-100, and 10 min incubated overnight in a humid chamber at 37°C and then in 1× PBS. The secondary antibody (diluted in 1× PBS+1% washed for 15 min in 15% formamide in 2× SSC at 37°C, BSA, (1:500) for HeLa cells, and (1:250) for Tig-1 cells) was 15 min in 2× SSC at 37°C, 15 min in 1× SSC at RT, and 5 min then applied and incubated as for the primary, with the same in 4× SSC at RT. HSATII RNA was detected with either set of washes. Following both MeCP2 staining and HSATII 1:500 DyLight 488 streptavidin or DyLight 594 streptavidin RNA hybridization, coverslips were DAPI stained and (Vector Laboratories, Burlingame, CA) in 4× SSC+1% BSA. mounted. Secondary-detected coverslips were then washed at room tem- perature for 10 min in 4× SSC, 10 min in 4× SSC+0.1% Triton X-100, and 10 min in 4× SSC. All coverslips were stained Image acquisition and analysis with 2 μg/mL DAPI, mounted with VECTASHIELD® Antifade Mounting Medium (Vector Laboratories), and sealed Slides were imaged using a ZEISS Axio Observer Z1 with fingernail polish. epifluorescent microscope equipped with a Hamamatsu ApTargeT™ vector was labeled by nick with ORCA-Flash 4.0 Digital CMOS camera or ZEISS Axiocam digoxigenin-11-dUTP (Roche) (Byron et al., 2013)todetect 702 mono camera with a ×100 oil objective. Both single plane integration of the construct via DNA FISH in both interphase images and Z-stacks were captured and analyzed using ZEISS cells and mitotic chromosome spreads. To prepare the probe ZEN2 imaging software. Chromosoma (2021) 130:75–90 89

Counts of nuclear body formation or chromatin defects Supplementary Information The online version contains supplementary were scored manually, while colocalization was measured material available at https://doi.org/10.1007/s00412-021-00753-0. using the ZEN2 Colocalization Module, including generation Acknowledgements ’ We would like to thank S. Akkipeddi for assistance of Pearson s correlation coefficient. Linescans were generated with maintenance of stably transfected cell lines and H. Sen for assistance using the Profile tool in ZEN2. DNA integration foci from with chromosome spreads and DNA hybridizations. We are grateful to J. pTargeT DNA hybridizations on interphase cells were scored Rubien, S. Akkipeddi, and B. Carone for helpful comments and critical on a subjective scale for relative intensity as determined by reading of the manuscript. observation through the microscope oculars. This scale cate- Author’s contribution All authors were involved in the design, imple- gorized relative foci brightness into the following five subjec- mentation, data collection, and data interpretation. DMC devised the tive classes: 0 (no foci), + (dim), ++ (easily visible), +++ (very study and wrote the manuscript. All authors read and approved the final bright), or ++++ (extremely bright). DNA integration foci manuscript. from chromosome spreads hybridized with the pTargeT probe were categorized as being on acrocentric or submetacentric Funding This work was supported by funding from the Charles E. Kaufman Foundation of the Pittsburgh Foundation (KA2017-91790) chromosomes, and further scored based on signal location and the National Institutes of Health (R15 GM134495) to DMC. relative to the centromere as well as chromosome size. Chromatin defects were scored by imaging DAPI stained Declarations slides and evaluating nuclei for abnormal phenotypes. This scoring was completed by multiple individuals in a blind man- Competing interests The authors declare no competing interests. ner (with the labels removed from slides) to ensure robustness and reproducibility in data scoring. Categories for these de- fects included the following: chromatin bridges between two References nuclei (including lagging chromosomes), micronuclei which were fully separated from the primary nucleus, abnormally Altemose N, Miga KH, Maggioni M, Willard HF (2014) Genomic char- shaped nuclei (including nuclei in the process of blebbing), acterization of large heterochromatic gaps in the human genome assembly. PLoS Comput Biol 10:e1003628 and other defects (including nuclei with burst chromatin and Bersani F, Lee E, Kharchenko PV, Xu AW, Liu M, Xega K, MacKenzie holes or indentations in the chromatin). All scoring data were OC, Brannigan BW, Wittner BS, Jung H, Ramaswamy S, Park PJ, analyzed and graphed using Prism8 (GraphPad), including Maheswaran S, Ting DT, Haber DA (2015) Pericentromeric satellite statistical analyses using Chi-square test (as shown in Figs. 1 repeat expansions through RNA-derived DNA intermediates in can- – and 4,andFigureS6). cer. Proc Natl Acad Sci 112:15148 15153 Biamonti G, Vourc'h C (2010) Nuclear stress bodies. Cold Spring Harb Perspect Biol 2:a000695 Bouzinba-Segard H, Guais A, Francastel C (2006) Accumulation of small Quantitative RT-PCR analysis murine minor satellite transcripts leads to impaired centromeric ar- chitecture and function. Proc Natl Acad Sci U S A 103:8709–8714 RNA was extracted from pelleted cells using either an RNeasy Brückmann NH, Pedersen CB, Ditzel HJ, Gjerstorff MF (2018) Epigenetic Reprogramming of Pericentromeric Satellite DNA in Mini kit and QIAshredder (QIAGEN) or a Direct-zol RNA Premalignant and Malignant Lesions. Mol Cancer Res 16:417–427 Miniprep Kit (Zymo Research, Irvine, CA), following manu- Byron M, Hall LL, Lawrence JB (2013) A multifaceted FISH approach to facturers’ protocols for each. Samples were then treated with study endogenous RNAs and DNAs in native nuclear and cell struc- TURBO™ DNase (Invitrogen, Thermo Fisher Scientific) fol- tures. Curr Protoc Hum Genet Chapter 4, Unit 4.15 lowing manufacturer’s guidelines. RNA was purified using Carone DM, Lawrence JB (2013) Heterochromatin instability in cancer: from the Barr body to satellites and the nuclear periphery. Semin Agencourt AMPure XP beads (Beckman Coulter, Brea, CA) Cancer Biol 23:99–108 and quantified via NanoDrop™ (Thermo Fisher Scientific). Castello A, Fischer B, Frese CK, Horos R, Alleaume A-M, Foehr S, Curk Reverse transcription was performed using iScript™ T, Krijgsveld J, Hentze MW (2016) Comprehensive identification of – Reverse Transcription Supermix for RT-qPCR (Bio-Rad), in- RNA-binding domains in human cells. Mol Cell 63:696 710 cluding RT-controls for all RNA samples. Quantitative RT- Chan DYL, Moralli D, Khoja S, Monaco ZL (2017) Noncoding centro- meric RNA expression impairs chromosome stability in human and PCR was carried out with PerfeCTa® SYBR® Green murine stem cells. Dis Markers 2017:7506976 FastMix® Reaction Mix (Quantabio, Beverley, MA), using Eichler EE, Clark RA, She X (2004) An assessment of the sequence gaps: primers included in the Key Resources Table. All reactions unfinished business in a finished human genome. Nat Rev Genet 5: – were performed in triplicate, using a CFX96 Real-Time 345 354 Hall LE, Mitchell SE, O'Neill RJ (2012) Pericentric and centromeric tran- C1000 Thermal Cycler (Bio-Rad) or MJ Research PTC-200 scription: a perfect balance required. Chromosom Res 20:535–546 thermocycler equipped with a Chromo4 Real-Time PCR Hall LL, Byron M, Carone DM, Whitfield TW, Pouliot GP, Fischer A, Detector (Bio-Rad). Analysis of relative enrichment was cal- Jones P, Lawrence JB (2017) Demethylated HSATII DNA and culated using the ΔΔC(t) method (Pfaffl, 2001), where re- HSATII RNA foci sequester PRC1 and MeCP2 into cancer- – ported values are normalized to β-actin expression. specific nuclear bodies. Cell Rep 18:2943 2956 90 Chromosoma (2021) 130:75–90

Hédouin S, Grillo G, Ivkovic I, Velasco G, Francastel C (2017) CENP-A Richard GF, Kerrest A, Dujon B (2008) Comparative genomics and mo- chromatin disassembly in stressed and senescent murine cells. Sci lecular dynamics of DNA repeats in . Microbiol Mol Biol Rep 7:1–14 Rev 72:686–727 Henikoff S, Ahmad K, Malik HS (2001) The centromere paradox: stable Rošić S, Köhler F, Erhardt S (2014) Repetitive centromeric satellite RNA inheritance with rapidly evolving DNA. Science 293:1098–1102 is essential for formation and cell division. J Cell Biol Hite KC, Adams VH, Hansen JC (2009) Recent advances in MeCP2 207:335–349 structure and function. Biochem Cell Biol 87:219–227 Shadle SC, Bennett SR, Wong C-J, Karreman NA, Campbell AE, van der Howe B, Umrigar A, Tsien F (2014) Chromosome preparation from Maarel SM, Bass BL, Tapscott SJ (2020) DUX4-induced bidirec- cultured cells. In: Chromosome preparation from cultured cells. In tional HSATII satellite repeat transcripts form intranuclear double- J Vis Exp stranded RNA foci in human cell models of FSHD. Hum Mol Genet Ichida K, Suzuki K, Fukui T, Takayama Y, Kakizawa N, Watanabe F, Slee RB, Steiner CM, Herbert BS, Vance GH, Hickey RJ, Schwarz T, Ishikawa H, Muto Y, Kato T, Saito M, Futsuhara K, Miyakura Y, Christan S, Radovich M, Schneider BP, Schindelhauer D, Grimes Noda H, Ohmori T, Konishi F, Rikiyama T (2018) Overexpression BR (2012) Cancer-associated alteration of pericentromeric hetero- of satellite alpha transcripts leads to chromosomal instability via chromatin may contribute to chromosome instability. Oncogene 31: – segregation errors at specific chromosomes. Int J Oncol 3244 3253 Ideue T, Cho Y, Nishimura K, Tani T (2014) Involvement of satellite I Smurova K, De Wulf P (2018) Centromere and pericentromere transcription: … noncoding RNA in regulation of chromosome segregation. roles and regulation in sickness and in health. Front Genet 9:674 Cells 19:528–538 Swanson EC, Manning B, Zhang H, Lawrence JB (2013) Higher-order unfolding of satellite heterochromatin is a consistent and early event Jagannathan M, Cummings R, Yamashita YM (2018) A conserved func- – tion for pericentromeric satellite DNA in cell . J Cell Biol 203:929 942 Swinnen B, Robberecht W, Van Den Bosch L (2020) RNA toxicity in Jain M, Olsen HE, Turner DJ, Stoddart D, Bulazel KV, Paten B, Haussler D, non-coding repeat expansion disorders. EMBO J 39:e101112 Willard HF, Akeson M, Miga KH (2018) Linear assembly of a human Tagarro I, Fernández-Peralta AM, González-Aguilera JJ (1994) centromere on the Y chromosome. Nat Biotechnol 36:321–323 Chromosomal localization of human satellites 2 and 3 by a FISH Johnson WL, Yewdell WT, Bell JC, McNulty SM, Duda Z, O'Neill RJ, method using oligonucleotides as probes. Hum Genet 93:383–388 Sullivan BA, Straight AF (2017) RNA-dependent stabilization of Tasselli L, Xi Y, Zheng W, Tennen RI, Odrowaz Z, Simeoni F, Li W, SUV39H1 at constitutive heterochromatin. Elife 6 Chua KF (2016) SIRT6 deacetylates H3K18ac at pericentric chro- Jolly C, Metz A, Govin J, Vigneron M, Turner BM, Khochbin S, Vourc'h matin to prevent mitotic errors and cellular senescence. Nat Struct C (2004) Stress-induced transcription of satellite III repeats. J Cell MolBiol23:434–440 – Biol 164:25 33 Ting DT, Lipson D, Paul S, Brannigan BW, Akhavanfard S, Coffman EJ, Maharana S, Wang J, Papadopoulos DK, Richter D, Pozniakovsky A, Contino G, Deshpande V, Iafrate AJ, Letovsky S, Rivera MN, Poser I, Bickle M, Rizk S, Guillen-Boixet J, Franzmann TM et al Bardeesy N, Maheswaran S, Haber DA (2011) Aberrant overexpres- (2018) RNA buffers the phase separation behavior of -like sion of satellite repeats in pancreatic and other epithelial cancers. – RNA binding proteins. Science 360:918 921 Science 331:593–596 McNulty SM, Sullivan LL, Sullivan BA (2017) Human centromeres Voullaire LE, Slater HR, Petrovic V, Choo KH (1993) A functional produce chromosome-specific and array-specific alpha satellite tran- marker centromere with no detectable alpha-satellite, satellite III, scripts that are complexed with CENP-A and CENP-C. Dev Cell 42: or CENP-B protein: activation of a latent centromere? Am J Hum – 226 240.e6 Genet 52:1153–1163 Miga KH (2019) Centromeric satellite DNAs: hidden sequence variation Wong LH, Brettingham-Moore KH, Chan L, Quach JM, Anderson MA, in the human population. In Genes (Basel), vol. 10 Northrop EL, Hannan R, Saffery R, Shaw ML, Williams E, Choo Miga KH, Koren S, Rhie A, Vollger MR, Gershman A, Bzikadze A, KHA (2007) Centromere RNA is a key component for the assembly Brooks S, Howe E, Porubsky D, Logsdon GA et al (2019) of nucleoproteins at the and centromere. Genome Res 17: Telomere-to-telomere assembly of a complete human X 1146–1160 chromosome Young JI, Hong EP, Castle JC, Crespo-Barreto J, Bowman AB, Rose Ninomiya K, Adachi S, Natsume T, Iwakiri J, Terai G, Asai K, Hirose T MF, Kang D, Richman R, Johnson JM, Berget S, Zoghbi HY (2020) LncRNA-dependent nuclear stress bodies promote intron (2005) Regulation of RNA splicing by the methylation-dependent retention through SR protein phosphorylation. EMBO J 39:e102729 transcriptional repressor methyl-CpG binding protein 2. Proc Natl Nogalski MT, Solovyov A, Kulkarni AS, Desai N, Oberstein A, Levine Acad Sci U S A 102:17551–17558 AJ, Ting DT, Shenk T, Greenbaum BD (2019) A tumor-specific Zhu Q, Pao GM, Huynh AM, Suh H, Tonnu N, Nederlof PM, Gage FH, endogenous repetitive element is induced by herpesviruses. Nat Verma IM (2011) BRCA1 tumour suppression occurs via Commun 10:90 heterochromatin-mediated silencing. Nature 477:179–184 Ohashi M, Aizawa S, Ooka H, Ohsawa T, Kaji K, Kondo H, Kobayashi Zhu Q, Hoong N, Aslanian A, Hara T, Benner C, Heinz S, Miga KH, Ke T, Noumura T, Matsuo M, Mitsui Y, Murota S, Yamamoto K, Ito H, E, Verma S, Soroczynski J et al (2018) Heterochromatin-encoded Shimada H, Utakoji T (1980) A new human diploid cell , TIG- satellite RNAs induce breast cancer. Mol Cell 70:842–853.e7 1, for the research on cellular aging. Exp Gerontol 15:121–133 Zink D, Fischer AH, Nickerson JA (2019) Nuclear structure in cancer Perea-Resa C, Blower MD (2018) Centromere Biology: transcription cells. Nat Rev Cancer 4:677–687 goes on stage Pfaffl MW (2001) A new mathematical model for relative quantification Publisher’snote Springer Nature remains neutral with regard to jurisdic- in real-time RT–PCR. Nucleic Acids Res 29:e45 tional claims in published maps and institutional affiliations.