Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

1 Reciprocal monoallelic expression of ASAR lncRNA genes

2 controls replication timing of human chromosome 6.

3

4 Michael Heskett1, Leslie G. Smith2, Paul Spellman1, and Mathew J. Thayer2*

5

6 1Department of Chemical Physiology and Biochemistry

7 Oregon Health & Science University

8 3181 S. W. Sam Jackson Park Road

9 Portland, Oregon 97239, USA

10

11 2Department of Chemical Physiology and Biochemistry

12 Oregon Health & Science University

13 3181 S. W. Sam Jackson Park Road

14 Portland, Oregon 97239, USA

15

16 *Corresponding author.

17 ORCID: 0000-0001-6483-1661

18 TEL: (503) 494-2447

19 FAX: (503) 494-7368

20 Email:[email protected]

21

22 Running Title: ASAR lincRNA genes on chromosome 6

23 Key Words: Replication timing, cis-acting element, non-coding RNA

1 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

24 Abstract

25 DNA replication occurs on mammalian chromosomes in a cell-type distinctive temporal

26 order known as the replication timing program. We previously found that disruption of the

27 noncanonical lncRNA genes ASAR6 and ASAR15 results in delayed replication timing

28 and delayed mitotic chromosome condensation of human chromosomes 6 and 15,

29 respectively. ASAR6 and ASAR15 display random monoallelic expression, and display

30 asynchronous replication between alleles that is coordinated with other random

31 monoallelic genes on their respective chromosomes. Disruption of the expressed allele,

32 but not the silent allele, of ASAR6 leads to delayed replication, activation of the previously

33 silent alleles of linked monoallelic genes, and structural instability of human chromosome

34 6. In this report, we describe a second lncRNA gene (ASAR6-141) on human

35 chromosome 6 that when disrupted results in delayed replication timing in cis. ASAR6-

36 141 is subject to random monoallelic expression and asynchronous replication, and is

37 expressed from the opposite chromosome 6 homolog as ASAR6. ASAR6-141 RNA, like

38 ASAR6 and ASAR15 RNAs, contains a high L1 content and remains associated with the

39 chromosome territory where it is transcribed. Three classes of cis-acting elements control

40 proper chromosome function in mammals: origins of replication, centromeres; and

41 telomeres, which are responsible for replication, segregation and stability of all

42 chromosomes. Our work supports a fourth type of essential chromosomal element, the

43 “Inactivation/Stability Center”, which expresses ASAR lncRNAs responsible for proper

44 replication timing, monoallelic expression, and structural stability of each chromosome.

45

2 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

46 Introduction

47 Numerous reports over the past 50+ years have described an abnormal DNA replication

48 phenotype affecting individual chromosomes in mitotic preparations from mammalian

49 cells [reviewed in (Thayer 2012)]. For example, we found that certain tumor derived

50 chromosome translocations display a delay in replication timing (DRT) that is

51 characterized by a >3 hour delay in the initiation and completion of DNA synthesis along

52 the entire length of individual chromosomes (Smith et al. 2001). Chromosomes with DRT

53 also display a delay in mitotic chromosome condensation (DMC), which is characterized

54 by an under-condensed appearance during mitosis and a concomitant delay in the mitotic

55 phosphorylation of histone H3 (Smith et al. 2001; Chang et al. 2007). We also found that

56 ~5% of chromosome translocations induced by exposing human or mouse cells to

57 ionizing radiation (IR) display DRT/DMC (Breger et al. 2004). To characterize the

58 DRT/DMC phenotype further, we developed a Cre/loxP system that allowed us to create

59 chromosome translocations in a precise and controllable manner (Breger et al. 2004;

60 Breger et al. 2005). Using this Cre/loxP system, we carried out a screen in human cells

61 designed to identify loxP integration sites that generate translocated chromosomes with

62 DRT/DMC (Breger et al. 2004; Breger et al. 2005; Stoffregen et al. 2011; Donley et al.

63 2015). We found that ~5% of Cre/loxP induced translocations display DRT/DMC (Breger

64 et al. 2005). Therefore, ~5% of translocations induced by two different mechanisms (IR

65 or Cre/loxP) result in DRT/DMC.

66 Our Cre/loxP screen identified five cell lines that generate balanced translocations,

67 affecting eight different autosomes, all displaying DRT/DMC (Breger et al. 2005).

68 Characterization of two of these translocations identified discrete cis-acting loci that when

3 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

69 disrupted result in DRT/DMC on human chromosomes 6 or 15 (Stoffregen et al. 2011;

70 Donley et al. 2015). Molecular examination of the disrupted loci identified two lncRNA

71 genes, which we named ASynchronous replication and Autosomal RNA on chromosome

72 6 (ASAR6) and on chromosome 15 (ASAR15) (Stoffregen et al. 2011; Donley et al. 2015).

73 These studies defined the first cis-acting loci that control replication timing, monoallelic

74 gene expression, and structural stability of individual human autosomes (Stoffregen et al.

75 2011; Donley et al. 2015).

76 The vast majority of genes on mammalian autosomes are expressed from both alleles.

77 However, some autosomal genes are expressed preferentially from only one allele,

78 achieving a state of “autosome pair non-equivalence” (Singh et al. 2003; Ensminger and

79 Chess 2004). The most extreme form of differential allelic expression is often referred to

80 as monoallelic expression, where a single allele is expressed exclusively [reviewed in

81 (Gendrel et al. 2016)]. Differential allelic expression can arise from distinct mechanisms.

82 For example, variation in gene expression can be heritable and has been mapped to the

83 genomes of humans and model organisms as expression quantitative trait loci [eQTL;

84 (Petretto et al. 2006)]. eQTL are genetic loci where sequence variation is associated with

85 differential expression of one or more genes. The consequences of eQTL can be

86 observed either in cis or in trans (Bryois et al. 2014; Signor and Nuzhdin 2018). In

87 contrast, differential allelic expression can also occur in the absence of DNA sequence

88 polymorphisms and is connected to situations where there is a “programmed”

89 requirement to regulate gene dosage or to provide exquisite specificity [reviewed in

90 (Alexander et al. 2007; Li et al. 2012; Lin et al. 2012; Gendrel et al. 2014)]. One well

91 established form of programmed monoallelic expression occurs in a parent of origin

4 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

92 specific manner, and is known as genomic imprinting [reviewed in (Bartolomei 2009)]. In

93 addition, monoallelic expression occurring in a random manner was observed from as

94 many as 8% of autosomal genes (Gimelbrant et al. 2007; Chess 2012). One unusual

95 characteristic of all programmed monoallelic genes is asynchronous replication between

96 alleles (Mostoslavsky et al. 2001; Singh et al. 2003; Ensminger and Chess 2004;

97 Schlesinger et al. 2009). This asynchronous replication is present in tissues where the

98 genes are not transcribed, indicating that asynchrony is not dependent on transcription

99 (Singh et al. 2003; Ensminger and Chess 2004; Donley et al. 2013; Donley et al. 2015).

100 Furthermore, asynchronous replication of random monoallelic genes is coordinated with

101 other random monoallelic genes on the same chromosome, indicating that there is a

102 chromosome-wide system that coordinates replication asynchrony of programmed

103 random monoallelic genes (Singh et al. 2003; Ensminger and Chess 2004; Donley et al.

104 2013; Donley et al. 2015). We use the following criteria to classify genes as being subject

105 to Programed Random Monoallelic Expression (PRME): 1) differential allelic expression

106 is detected in multiple unrelated individuals, which rules out rare DNA polymorphisms in

107 promoters or enhancers; 2) differential expression of either allele is detected in single

108 cell-derived subclones from the same individual, which rules out genomic imprinting and

109 eQTL; and 3) asynchronous replication between alleles is present and coordinated with

110 other random monoallelic genes on the same chromosome, indicating that the monoallelic

111 gene is regulated by a chromosome-wide system that coordinates asynchronous

112 replication along the chromosome pair. Using these criteria, we previously found that

113 ASAR6 and ASAR15 are subject to PRME (Stoffregen et al. 2011; Donley et al. 2013;

114 Donley et al. 2015).

5 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

115 Recent reports have described very long intergenic non-coding (vlinc) RNAs expressed

116 in numerous human tissues (Kapranov et al. 2010; St Laurent et al. 2012; St Laurent et

117 al. 2016). The vlincRNAs are RNA Pol II products that are nuclear, non-spliced, non-

118 polyadenylated transcripts of >50 kb of contiguously expressed sequence that are not

119 associated with protein coding genes. The initial reports annotated 2,147 human

120 vlincRNAs from 833 samples in the FANTOM5 dataset (St Laurent et al. 2013; St Laurent

121 et al. 2016). A more recent study identified an additional 574 vlincRNAs expressed in

122 childhood acute lymphoblastic leukemia (Caron et al. 2018). Therefore, there are

123 currently >2,700 annotated vlincRNAs that are encoded by >10% of the human genome

124 (St Laurent et al. 2013; St Laurent et al. 2016; Caron et al. 2018). ASAR6 and ASAR15

125 RNAs share several characteristics with the vlincRNAs, including: RNA Pol II products,

126 long contiguous transcripts (>50 kb) that are non-spliced, non-polyadenylated, and are

127 retained in the nucleus (Stoffregen et al. 2011; Donley et al. 2013; Donley et al. 2015).

128 Therefore, given these shared characteristics between ASAR6, ASAR15 and vlincRNAs,

129 we consider the vlincRNAs as potential ASAR candidates.

130 In addition to monoallelic expression, ASAR6 and ASAR15 RNAs share additional

131 characteristics, including: the RNAs are retained within the chromosome territories where

132 they are transcribed, and they contain a high long interspersed element 1 (LINE1 or L1)

133 content (Stoffregen et al. 2011; Donley et al. 2013; Donley et al. 2015). In this report, we

134 used these “ASAR” characteristics to identify a second lncRNA gene that controls

135 replication timing of human chromosome 6, which we designate as ASAR6-141. The

136 ASAR6-141 gene is located at ~141 mb of human chromosome 6, is subject to random

137 monoallelic expression and asynchronous replication, and disruption of the expressed

6 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

138 allele, but not the silent allele, leads to delayed replication of human chromosome 6 in

139 cis. ASAR6-141 RNA, which was previously annotated as vlinc273 (St Laurent et al.

140 2016), is ~185 kb in length, contains ~29% L1 sequences, remains associated with the

141 chromosome 6 territory where it is transcribed, and is expressed in trans to the expressed

142 allele of ASAR6. These observations support a model that includes reciprocal monoallelic

143 expression of different ASAR lincRNA genes that control replication timing of homologous

144 chromosome pairs.

145

7 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

146 Results

147 Reciprocal random monoallelic expression of ASAR lncRNAs on chromosome 6:

148 With the goal of identifying nuclear RNAs with “ASAR” characteristics expressed from

149 human chromosome 6, we carried out RNA-seq on nuclear RNA isolated from HTD114

150 cells. HTD114 cells are a human fibrosarcoma cell line, where we previously carried out

151 the Cre/loxP screen that led to the identification and functional characterization of ASAR6

152 and ASAR15 (Breger et al. 2005; Stoffregen et al. 2011; Donley et al. 2015). Figure 1

153 shows the UCSC Genome Browser view of chromosome 6 showing the RNA-seq reads

154 from a region, between 140.3 mb and 141.3 mb, previously annotated as expressing 6

155 different vlincRNAs (St Laurent et al. 2016). We found that vlinc273 RNA is expressed in

156 HTD114 cells, but vlinc271, vlinc1010, vlinc1011, vlinc1012, and vlinc272, showed little

157 or no expression. Consistent with the characterization of vlincRNAs as being

158 predominantly nuclear (St Laurent et al. 2016), we found that vlinc273 RNA is enriched

159 (>100X) in the nuclear fraction in the human cell lines GM12878, HepG2, and K562 (Table

160 1). Figure 1 also shows the location of Fosmids used in the RNA-DNA FISH analyses

161 (see below), the Long RNA-seq track (showing contiguous transcripts from the human

162 cell lines, GM12878, HepG2, and K562) from ENCODE/Cold Spring Harbor, and the

163 Repeat Masker Track showing the location of repetitive elements (see Table S1). In

164 addition, vlinc273 RNA was found to have significantly more L1 elements than intronic

165 regions of similar length. For this analysis, we created a beta distribution of the fraction

166 of L1 sequences within introns by repeated random sampling of intronic regions of length

167 equal to vlinc273 that was then used to generate an empirical p-value (L1 fraction within

168 vlinc273 is 0.29, p=0.013).

8 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

169 One prominent characteristic of both ASAR6 and ASAR15 is monoallelic expression

170 (Stoffregen et al. 2011; Donley et al. 2013; Donley et al. 2015). Therefore, to determine if

171 the vlinc273 transcripts detected in HTD114 cells also show monoallelic expression, we

172 used reverse transcribed RNA as input for PCR, followed by sequencing at heterozygous

173 SNPs. Figure 2A shows sequencing traces from two different SNPs that are heterozygous

174 in genomic DNA, but a single allele was detected in RNA isolated from HTD114 cells,

175 indicating that these transcripts are monoallelic (see Fig. S1A and Table S2). In addition,

176 we previously generated two chromosome 6 mono-chromosomal hybrids to aid in

177 mapping heterozygous SNPs onto the HTD114 chromosome 6 homologs (Stoffregen et

178 al. 2011). These two hybrid cell lines are mouse L cell clones, each containing one of the

179 two chromosome 6s from HTD114, which we arbitrarily name as CHR6A and CHR6B.

180 Using these mono-chromosomal hybrids, we previously found that ASAR6 is expressed

181 from CHR6A (Stoffregen et al. 2011). Sequence traces generated from genomic DNA

182 isolated from these two mono-chromosomal hybrids indicated that the vlinc273 transcripts

183 are derived from CHR6B (Fig. 2A). A similar analysis of a third heterozygous SNP within

184 vlinc273 also indicated that the RNA is monoallelic and expressed from CHR6B (Fig. S1B;

185 and Table S2). These results indicate that vlinc273 is expressed from the opposite

186 chromosome 6, or in trans, to ASAR6 in HTD114 cells.

187 Another prominent characteristic shared between ASAR6 and ASAR15 is that their

188 RNAs remain associated with the chromosome territories where they are transcribed

189 (Stoffregen et al. 2011; Donley et al. 2013; Donley et al. 2015). Therefore, we next

190 assayed expression of vlinc273 using RNA-DNA FISH in HTD114 cells. For this analysis

191 we used two different Fosmid probes to detect vlinc273 RNA (see Fig. 1), plus a

9 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

192 chromosome 6 centromeric probe to detect chromosome 6 DNA. As expected from the

193 RNA-seq analysis, we detected expression of vlinc273 RNA, and the RNA FISH signal

194 remains associated with one of the chromosome 6 homologs. Figure 2B-F show

195 examples of this analysis and indicates that the relatively large clouds of RNA are

196 adjacent to, or overlapping with, one of the chromosome 6 centromeric DNA signals. We

197 detected single sites of vlinc273 RNA expression in >80% of HTD114 cells (Fig. S2A). In

198 addition, we used RNA-DNA FISH to detect both ASAR6 and vlinc273 RNA in

199 combination with a chromosome 6 whole chromosome paint as probe to detect the

200 chromosome 6 territory. Figure 2G-J shows examples of this analysis and indicates that

201 vlinc273 and ASAR6 RNAs are detected on opposite chromosome 6 homologs.

202 Quantitation of the RNA FISH signals indicated that >90% of the cells express vlinc273

203 RNA in trans to ASAR6 RNA (Fig. S2B). We also note that the size of the RNA FISH

204 signals detected by the ASAR6 and vlinc273 probes were variable, ranging from large

205 clouds occupying the entire chromosome 6 territory, to relatively small spots of

206 hybridization. We interpret the variability in the size of the RNA clouds to reflect the phase

207 of the cell cycle. Thus, both ASAR6 and vlinc273 RNAs are not detectable in mitotic cells

208 (not shown), which suggests that the RNA clouds must be reformed every cell cycle as

209 cells exit mitosis and re-enter the G1 phase. Thus, cells with small RNA FISH signals are

210 predicted to be in early G1, whereas cells with large RNA FISH signals are predicted to

211 be in S or G2 phases.

212

213 Random monoallelic expression of vlinc273:

10 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

214 The observations described above indicate that vlinc273 and ASAR6 RNAs are

215 expressed from opposite chromosome 6 homologs in the clonal cell line HTD114. This

216 monoallelic expression could be due to either DNA sequence polymorphisms within

217 regulatory elements (i.e. eQTL), genomic imprinting, or to PRME. Therefore, to begin to

218 distinguish between these possibilities, we determined if vlinc273 is monoallelically

219 expressed in EBV transformed lymphoblastoid cells, which have been used extensively

220 in the analysis of autosomal monoallelic expression in humans (Ensminger and Chess

221 2004; Gimelbrant et al. 2007; Stoffregen et al. 2011; Donley et al. 2013). For this analysis,

222 we used RNA-DNA FISH to assay expression of vlinc273 in GM12878 lymphoblastoid

223 cells. We used two different Fosmid probes to detect vlinc273 RNA, plus a chromosome

224 6 centromeric probe to detect DNA. For comparison, we also assayed expression of

225 KCNQ5, a gene on chromosome 6 known to be subject to PRME (Gimelbrant et al. 2007;

226 Donley et al. 2013). Note that the RNA FISH probe for KCNQ5 is derived from the first

227 intron, which is ~400 kb in length, and this probe results in a robust RNA FISH signal. We

228 found that vlinc273 and KCNQ5 show similar levels of monoallelic expression with single

229 sites of RNA hybridization in >84% of cells (Fig. 3A-E and Fig. S2C).

230 Next, to determine if vlinc273 is monoallelic in human primary cells, we assayed

231 expression of vlinc273 using RNA FISH in primary blood lymphocytes (PBLs) isolated

232 from two unrelated individuals. For comparison, we also included a probe for KCNQ5. We

233 found that vlinc273 and KCNQ5 show similar levels of monoallelic expression with single

234 sites of RNA hybridization in >69% and two RNA FISH signals in <13% of cells (Fig. 3F-

235 J and Fig. S2C).

11 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

236 To determine if the monoallelic expression of vlinc273 is random, and not imprinted,

237 we tested whether or not vlinc273 expression was coordinated with a gene on

238 chromosome 6 that is known to be subject to PRME. The premise of this experiment is

239 that PBLs are non-clonal, and individual PBL cells are anticipated to express either the

240 paternal or the maternal allele of any gene subject to PRME at approximately equal

241 frequencies. Thus, if vlinc273 expression is always expressed from the same or opposite

242 chromosome 6 homolog as a known PRME gene then we can conclude that vlinc273

243 expression is also subject to PRME. In contrast, if vlinc273 expression is detected from

244 the same chromosome 6 homolog as a known PRME gene in only ~50% of cells, then

245 we would suspect that vlinc273 expression was subject to imprinting. For this analysis,

246 we tested if expression of vlinc273 RNA was detected on the same or opposite

247 chromosome, i.e. either in cis or in trans, as KCNQ5 RNA, which is known to be subject

248 to PRME (Gimelbrant et al. 2007; Stoffregen et al. 2011; Donley et al. 2013). For this

249 analysis, we used a two-color RNA FISH assay to detect expression of vlinc273 in

250 combination with a probe from the first intron of KCNQ5 on PBLs isolated from two

251 unrelated individuals. Quantification of the number of RNA FISH signals in >100 cells

252 indicated that vlinc273 and KCNQ5 were expressed from the same chromosome 6

253 homolog in >89% of cells from both individuals (see Fig. 3F-I and Fig. S2B). A similar

254 assay on GM12878 lymphoblastoid cells, which are also non-clonal, indicated that

255 vlinc273 and KCNQ5 were expressed from the same chromosome 6 homolog in 94% of

256 cells (Fig. S2B). Therefore, because the PBLs and GM12878 cells are non-clonal, and

257 KCNQ5 expression is subject to PRME (Gimelbrant et al. 2007; Stoffregen et al. 2011;

258 Donley et al. 2013), we conclude that the monoallelic expression of vlinc273 must also

12 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

259 be random and therefore not imprinted. We note that we detected two sites of

260 hybridization for both probes in ~2% of cells (Fig. 3J). Finally, to directly compare the

261 appearance of the RNA FISH signals detected for vlinc273 to XIST RNA expressed from

262 the inactive X chromosome we assayed vlinc273 and XIST RNAs simultaneously in

263 female PBLs. Figure 3K and 3I show the clouds of RNA detected by the vlinc273 probe

264 in relation to the relatively larger clouds of RNA hybridization detected by the XIST probe.

265

266 Asynchronous replication of vlinc273 is coordinated on chromosome 6.

267 All programmed monoallelically expressed genes share the property of asynchronous

268 replication (Goldmit and Bergman 2004). We previously used Replication Timing-Specific

269 Hybridization (ReTiSH) (Schlesinger et al. 2009) to assay asynchronous replication of

270 chromosome 6 loci, including ASAR6. In the ReTiSH assay, cells are labeled with BrdU

271 for different times and then harvested during mitosis (see Fig. 4A). Regions of

272 chromosomes that incorporate BrdU are visualized by a modification of chromosome

273 orientation-fluorescence in situ hybridization (CO-FISH), where the replicated regions

274 (BrdU-labeled) are converted to single stranded DNA and then hybridized directly with

275 specific probes (Schlesinger et al. 2009). Since mitotic chromosomes are analyzed for

276 hybridization signals located on the same chromosome in metaphase spreads, the

277 physical distance between the loci is not a limitation of the ReTiSH assay (Schlesinger et

278 al. 2009). We previously used this assay to show that the asynchronous replication of

279 ASAR6 was coordinated, either in cis or in trans, with other random monoallelic loci on

280 human chromosome 6 (Stoffregen et al. 2011; Donley et al. 2013). For this analysis, we

281 used PBLs and a three-color hybridization scheme to simultaneously detect the vlinc273

13 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

282 locus, ASAR6, and the chromosome 6 centromere. The chromosome 6 centromeric probe

283 was included to unambiguously identify both chromosome 6s. Because centromeric

284 heterochromatin is late replicating, centromeric probes hybridize to both copies of

285 chromosome 6 at the 14 and 5 hour time points (Schlesinger et al. 2009; Donley et al.

286 2013). Using this assay, we found that the vlinc273 alleles were subject to asynchronous

287 replication that is coordinated in cis with ASAR6 (Fig. 4B-D; and Table 2). Therefore,

288 because the asynchronous replication of ASAR6 is coordinated with other random

289 monoallelic loci on chromosome 6 (Stoffregen et al. 2011; Donley et al. 2013), we conclude

290 that the asynchrony at the vlinc273 locus is also part of a chromosome-wide system that

291 coordinates the asynchronous replication of random monoallelic loci on chromosome 6.

292 One shared characteristic of the ASAR6 and ASAR15 genes is that the silent alleles

293 replicate before the expressed alleles on their respective chromosomes (Stoffregen et al.

294 2011; Donley et al. 2013; Donley et al. 2015). Therefore, one unanticipated result from our

295 ReTiSH assay is that asynchronous replication of vlinc273 and ASAR6 is coordinated in

296 cis. Thus, the earlier replicating vlinc273 allele is on the same homolog as the earlier

297 replicating ASAR6 allele. Therefore, to determine if the asynchronous replication of

298 vlinc273 and ASAR6 is also coordinated in cis in HTD114 cells, where they are expressed

299 from opposite homologs (see Fig. 2), we analyzed the asynchronous replication of vlinc273

300 and ASAR6 using the same three color ReTiSH assay describe above (Schlesinger et al.

301 2009). HTD114 cells contain a centromeric polymorphism on chromosome 6, and the

302 chromosome with the larger centromere is linked to the later replicating and expressed

303 allele of ASAR6 (Donley et al. 2013). We found that the asynchronous replication of

304 vlinc273 and ASAR6 is coordinated in cis in HTD114 cells (Fig. 4E-G; and Table 2). These

14 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

305 observations are consistent with our previous finding that ASAR6 is expressed from the

306 later replicating allele [(Donley et al. 2013); i.e. CHR6A], and indicate that vlinc273 is

307 expressed from the earlier replicating allele in HTD114 cells (i.e. CHR6B; see Fig. 2A).

308 Regardless, we found that the vlinc273 locus is subject to random monoallelic expression

309 and asynchronous replication that is coordinated with other random monoallelic loci on

310 chromosome 6 and therefore vlinc273 is subject to PRME.

311

312 Deletion of the expressed allele of vlinc273 results in delayed replication in cis.

313 To determine if the genomic region containing the vlincRNA cluster located on

314 chromosome 6 at 140.3-141.3 mb regulates replication timing, we used CRISPR/Cas9 to

315 delete the entire locus in HTD114 cells (see Figs. 1 and S1A). For this analysis we

316 designed single guide RNAs (sgRNAs) to unique sequences flanking the locus (see Fig

317 S1A; and Table S2). We expressed sgRNA-1, sgRNA2 and sgRNA-3 in all pairwise

318 combinations with Cas9 and screened clones for deletions using PCR primers that flank

319 the sgRNA binding sites (see Fig. S1A and Table S2). Because vlinc273 expression is

320 monoallelic in HTD114 cells, we isolated clones that had heterozygous deletions affecting

321 either CHR6A or CHR6B. We determined which allele was deleted based on retention of

322 the different base pairs of heterozygous SNPs located within the deleted regions (see Figs.

323 S1, S3 and S4; and Table S2). We also determined the deletion junctions by sequencing

324 PCR products that span the deleted regions (see Fig. S1A and Table S3).

325 To confirm that the deletions on CHR6B removed the expressed allele of vlinc273, we

326 carried out RNA-DNA FISH on cells containing deletions of either vlinc273 alone or the

327 entire vlincRNA cluster. Figure S5A shows the quantitation of this analysis and indicates

328 that vlinc273 is no longer detected in >86% of cells containing deletions on CHR6B. A

15 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

329 similar analysis of cells containing deletions of ASAR6 from CHR6A indicated that ASAR6

330 RNA was undetectable (Fig. S5B). In addition, to determine if deletions of the expressed

331 allele of vlinc273 affected expression of ASAR6, and vice versa, we assayed expression

332 of ASAR6 in the vlinc273 deletions on CHR6B, and expression of vlinc273 in the ASAR6

333 deletions on CHR6A. Figure S5A and S5B shows that vlinc273 and ASAR6 expression

334 are not affected by deletions of the expressed alleles of their reciprocal counterparts.

335 We previously found that the most sensitive assay to detect a chromosome-wide delay

336 in replication timing involves a BrdU “terminal label” assay (Smith and Thayer 2012). This

337 assay is distinct from the ReTiSH assay, and does not provide gene-specific information

338 related to asynchrony at individual loci, but does provide a sensitive assay to detect

339 chromosome-band and whole chromosome-scale changes in replication timing (Smith and

340 Thayer 2012). Using this assay, we previously found that prior to any genetic alterations

341 the two chromosome 6 homologs replicate synchronously in HTD114 cells (Breger et al.

342 2005; Stoffregen et al. 2011; Donley et al. 2013; Platt et al. 2017). In addition, we

343 previously took advantage of a centromeric polymorphism in HTD114 cells to

344 unambiguously distinguish between the two chromosome 6 homologs [(Donley et al. 2013;

345 Platt et al. 2017); see Fig. 4E-G]. The chromosome 6 with the larger centromere is linked

346 to the expressed allele of ASAR6 [(Donley et al. 2013); CHR6A], and therefore the

347 expressed allele of vlinc273 is linked to the chromosome 6 with the smaller centromere

348 (CHR6B; see Figs. 5 and S6). For this replication timing assay, cultures were incubated

349 with BrdU for 5.5 hours and mitotic cells harvested, processed for BrdU incorporation and

350 subjected to FISH using a chromosome 6 centromeric probe. As expected, prior to

351 disruption of the vlinc cluster, CHR6A and CHR6B incorporate comparable levels of BrdU

16 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

352 (Fig. 5F). In contrast, cells containing a deletion of the entire vlincRNA cluster on the

353 CHR6B allele contain significantly more BrdU incorporation into CHR6B than in CHR6A

354 (Fig. 5A-E). Quantification of the BrdU incorporation in multiple cells indicated that deletion

355 of the CHR6B allele, which contains the expressed allele of vlinc273, results in significantly

356 more BrdU incorporation into CHR6B (Fig. 5F). This is in contrast to cells containing a

357 deletion of the vlincRNA cluster from the CHR6A, which is silent for all 6 vlincRNAs, where

358 the BrdU incorporation is comparable between CHR6A and CHR6B (Fig. 5F). These

359 results indicate that deletion of the vlincRNA cluster on CHR6B results in delayed

360 replication of the entire chromosome in cis. In addition, replication timing analysis of

361 heterozygous deletions encompassing only the vlinc273 locus (using sgRNA-2 and

362 sgRNA-3) indicated that deletion of the expressed allele (CHR6B), but not the silent allele

363 (CHR6A), resulted in delayed replication of chromosome 6 (Fig. 5F). Finally, deletion of

364 the vlinc271, vlinc1010, vlinc1011, vlinc1012 and vlinc272 loci (using sgRNA-1 and

365 sgRNA-2) on CHR6B, did not result in delayed replication of chromosome 6 (Fig. 5F). For

366 an additional comparison, we included the chromosome 6 replication timing data from

367 HTD114 cells containing heterozygous deletions of ASAR6 on the expressed allele

368 (CHR6A) and on the silent allele (CHR6B) (Fig. 5F; also see Fig. S6). Taken together

369 these results indicate that deletion of the expressed allele of vlinc273 results in delayed

370 replication of chromosome 6 in cis, and because vlinc273 also displays PRME, the

371 vlinc273 locus is an ASAR. Because vlinc273 is the second ASAR identified on human

372 chromosome 6, and is located at ~141 mb, we designate this gene as ASAR6-141.

373

374 Discussion

17 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

375 Chromosome associated lncRNAs have become well established as regulators of

376 chromosome scale replication timing, gene expression and structural stability [reviewed

377 in (Thayer 2012; Galupa and Heard 2018)]. In this report, we identified a second

378 chromosome 6 lncRNA gene, ASAR6-141, that when disrupted results in delayed

379 replication timing of the entire chromosome in cis. ASAR6 and ASAR6-141 are subject to

380 PRME, are expressed from opposite chromosome 6 homologs, and disruption of the

381 expressed alleles, but not the silent alleles, leads to delayed replication timing of human

382 chromosome 6 in cis. ASAR6 and ASAR6-141 RNAs share certain characteristics,

383 including RNA Pol II products that are non-spliced, non-polyadenylated, contain a high

384 L1 content and remain associated with the chromosome territories where they are

385 transcribed. Taken together our results indicate that the replication timing of human

386 chromosome 6 is regulated by the reciprocal monoallelic expression of two different

387 ASAR lncRNA genes (Fig. 6).

388 We previously found that ASAR6 controls the transcriptional activity of nearby

389 monoallelic genes in cis (Stoffregen et al. 2011). Thus, we found that genetic disruption

390 of the expressed allele of ASAR6 leads to the transcriptional activation of the previously

391 silent allele of the protein coding gene FUT9 (Stoffregen et al. 2011). Our work supports

392 a model where the reciprocal monoallelic expression of ASAR genes controls the

393 transcriptional activity of linked protein coding genes in cis. Human chromosome 6

394 contains numerous random monoallelic genes, including KCNQ5, an olfactory receptor

395 cluster at 6p22.1, and the HLA locus (Gimelbrant et al. 2007). Interestingly, loss of

396 function mutations in KCNQ5 have been shown to cause autosomal dominant mental

397 retardation-46 (MRD46) (Lehman et al. 2017). Taken together, these observations raise

18 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

398 the possibility that PRME of protein coding genes can cause an autosomal dominant

399 phenotype in individuals with heterozygous loss of function mutations due to the exclusive

400 expression of the loss of function allele, creating the null phenotype, in 50% of cells.

401 However, whether or not PRME contributes to autosomal dominant phenotypes in

402 humans remains to be established. In addition, chromosome 6 also contains numerous

403 imprinted genes, e.g. IGF2R, PLAGL1, and HYMAI (Luedi et al. 2007). Whether or not

404 ASAR genes interact with imprinted loci is currently not known.

405 One hallmark of genes that are subject to PRME is coordination in the asynchronous

406 replication between alleles (Singh et al. 2003; Ensminger and Chess 2004; Donley et al.

407 2013; Donley et al. 2015). This coordination can be either in cis, i.e. the early replicating

408 alleles of two genes are always on the same homolog; or in trans, i.e. the early replicating

409 alleles are always on opposite homologs (Donley et al. 2013). In this report, we found that

410 the asynchronous replication of ASAR6-141 is coordinated in cis with ASAR6. This

411 observation is consistent with our previous findings that human chromosome 6 contains

412 loci that display random asynchronous replication that is coordinated both in cis and in

413 trans, that some of these asynchronous loci are separated by >100 megabases of

414 genomic DNA, and that the coordinated loci are on either side of the centromere

415 (Stoffregen et al. 2011; Donley et al. 2013). One important unanswered question about

416 the asynchronous replication that occurs at monoallelic genes is whether or not there is

417 a difference in origin usage on the two alleles. Detecting replication initiation events is

418 arguably the biggest obstacle in the human replication field. The reason why we

419 understand so little about human DNA replication is because of the large number of sites

420 that can serve as replication origins and their stochastic usage (Fragkos et al. 2015;

19 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

421 Dileep and Gilbert 2018). The existing ensemble methods used to identify origins suffer

422 from a low signal-to-noise ratio, due to stochastic selection of origins. Single molecule

423 methods are needed to understand such a stochastic process, and these methods are

424 currently too low throughput to be effective. Regardless, it will be interesting to determine

425 if the two alleles of monoallelic loci use the same or different origins, and whether or not

426 chromosomes with delayed replication timing also have altered origin usage.

427 Asynchronous replication of random monoallelic genes is an epigenetic mark that

428 appears before transcription and is thought to underlie the differential expression of the

429 two alleles of identical sequence (Mostoslavsky et al. 2001). Therefore, because the

430 asynchronous replication at PRME genes is coordinated along each chromosome, the

431 expression pattern of PRME genes is also anticipated to be coordinated, i.e. in cis- always

432 expressed from the same homolog; or in trans- always expressed from opposite

433 homologs. We previously found that ASAR6 and ASAR15 are expressed from the later

434 replicating alleles (Donley et al. 2013; Donley et al. 2015). In contrast, the FUT9 protein

435 coding gene, which is closely linked to ASAR6 (see Fig. S7), is expressed from the early

436 replicating allele (Donley et al. 2013; Donley et al. 2015). Therefore, PRME genes can be

437 expressed from either the early or the late replicating alleles. One unanticipated result

438 from our allelic expression and asynchronous replication assays described here is that

439 ASAR6-141 is expressed from the early replicating allele, which is the first example of an

440 ASAR that is expressed from the early allele. Nevertheless, we found that disruption of

441 the expressed allele, but not the silent allele of ASAR6-141 results in delayed replication

442 of chromosome 6, indicating that expression and not asynchronous replication is a critical

443 component of ASAR function. This conclusion is consistent with our previous observation

20 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

444 that ASAR6 RNA mediates the chromosome-wide effects of ASAR6 forced expression

445 (Platt et al. 2017). Therefore, the role of asynchronous replication at ASAR loci may serve

446 as a mechanism to help establish which allele will be transcribed. Thus, the epigenetic

447 mark that establishes early and late replication between the two alleles of PRME genes

448 may function to establish asymmetry between alleles, and then depending on the

449 regulatory promoter/enhancer elements at different PRME genes either the early or late

450 replicating allele will be transcribed.

451 One striking feature of both ASAR6 and ASAR15 is that they contain a high density of

452 L1 retrotransposons, constituting ~40% and ~55% of the expressed sequence,

453 respectively (Stoffregen et al. 2011; Donley et al. 2015). L1s were first implicated in

454 monoallelic expression when Dr. Mary Lyon proposed that L1s represent “booster

455 elements” that function during the spreading of X chromosome inactivation (Lyon 1998;

456 Lyon 2003). In humans, the X chromosome contains ~27% L1 derived sequence while

457 autosomes contain ~13% (Bailey et al. 2000). In addition, L1s are present at a lower

458 concentration in regions of the X chromosome that escape inactivation, supporting the

459 hypothesis that L1s serve as signals to propagate inactivation along the X chromosome

460 (Bailey et al. 2000). Further support for a role of L1s in monoallelic expression came from

461 the observation that L1s are present at a relatively high local concentration near both

462 imprinted and random monoallelic genes located on autosomes (Allen et al. 2003). L1s

463 have also been linked to DNA replication timing from the observation that differentiation-

464 induced replication timing changes are restricted to AT rich isochores containing high L1

465 density (Hiratani et al. 2004). Another potential link between L1s and DNA replication is

466 the observation that ~25% of origins in the human genome were mapped to L1 sequences

21 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

467 (Bartholdy et al. 2015). While this observation is suggestive of a relationship between

468 origins and L1s, it is not clear what distinguishes L1s with origin activity from L1s without

469 (Bartholdy et al. 2015).

470 We previously proposed a model in which the antisense L1 sequences function to

471 suppress splicing, and to promote stable association of the RNA with the chromosome

472 territories where they are transcribed (Platt et al. 2017). Consistent with this interpretation

473 is the finding that a de novo L1 insertion, in the antisense orientation, into an exon of the

474 mouse Nr2e3 gene results in inefficient splicing, accumulation of the transcript to high

475 levels, and retention of the transcript at the mutant Nr2e3 locus (Chen et al. 2006). In

476 addition, a more recent study found that the antisense strand of L1 RNA functions as a

477 multivalent “hub” for binding to numerous nuclear matrix and RNA processing proteins,

478 and that the L1 antisense RNA binding proteins repress splicing and 3’ end processing

479 within and around the L1 sequences (Attig et al. 2018). We recently used ectopic

480 integration of transgenes and CRISPR/Cas9-mediated chromosome engineering and

481 found that L1 sequences, oriented in the antisense direction, mediate the chromosome-

482 wide effects of ASAR6 and ASAR15 (Platt et al. 2017). In addition, we found that

483 oligonucleotides targeting the antisense strand of the one full length L1 within ASAR6

484 RNA restored normal replication timing to mouse chromosomes expressing an ASAR6

485 transgene. These results provided the first direct evidence that L1 antisense RNA plays

486 a functional role in replication timing of mammalian chromosomes (Platt et al. 2017).

487 Because ASAR6-141 RNA also contains numerous L1 sequences oriented in the

488 antisense direction (Table S1), we anticipate that the antisense L1 sequences within

489 ASAR6-141 RNA will likely represent the functional elements.

22 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

490 The vlincRNAs were identified as RNA transcripts of >50 kb of contiguous RNA-seq

491 reads that have no overlap with annotated protein coding genes (St Laurent et al. 2016).

492 The vlincRNAs were identified from the FANTOM5 Cap Analysis of Gene Expression

493 (CAGE) dataset, indicating that the vlincRNAs contain 5’ caps and consequently

494 represent RNA Pol II transcripts (St Laurent et al. 2016). We previously found that ASAR6

495 and ASAR15 are also transcribed by RNA Pol II (Stoffregen et al. 2011; Donley et al.

496 2015). ASAR6-141 RNA shares certain characteristics with ASAR6 and ASAR15 RNAs

497 that distinguish them from other canonical RNA Pol II lncRNAs. Thus, even though

498 ASAR6-141, ASAR6 and ASAR15 RNAs are RNA Pol II products they show little or no

499 evidence of splicing or polyadenylation and remain associated with the chromosome

500 territories where they were transcribed [(Stoffregen et al. 2011; Donley et al. 2015; St

501 Laurent et al. 2016); and see Figs. 1, 2 and 3]. Our work supports a model where all

502 mammalian chromosomes express “ASAR” genes that encode chromosome associated

503 lncRNAs that control the replication timing program in cis. In this model, the ASAR

504 lncRNAs function to promote proper chromosome replication timing by controlling the

505 timing of origin firing. In addition, because both ASAR6, and ASAR6-141 are

506 monoallelically expressed, our model includes expression of different ASAR genes from

507 opposite homologs [(Platt et al. 2017); see Fig. 6].

508 We previously found that ~5% of chromosome translocations, induced by two

509 different mechanisms (IR and Cre/loxP) display DRT/DMC (Breger et al. 2004; Breger et

510 al. 2005). Because ~5% of translocations display DRT/DMC and only one of the two

511 translocation products has DRT/DMC (Breger et al. 2005), indicates that ~2.5% of

512 translocation products display DRT/DMC (Breger et al. 2004; Breger et al. 2005;

23 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

513 Stoffregen et al. 2011; Donley et al. 2015). Taken with the observation that the

514 translocations that display DRT/DMC have disrupted ASAR genes (Stoffregen et al. 2011;

515 Donley et al. 2015), these observations suggest that ~2.5% of the genome is occupied

516 by ASARs. There are currently >2,700 annotated human vlincRNAs, and they are

517 expressed in a highly cell type-specific manner (St Laurent et al. 2013; St Laurent et al.

518 2016; Caron et al. 2018). Because many of the vlincRNAs are encoded by regions of the

519 genome that do not overlap with protein coding genes, many of the vlincRNAs contain a

520 high density of repetitive elements, including L1s (see Fig. 1 and Table S1 for examples).

521 In this report, we found that the genomic region annotated as vlinc273 has all of the

522 physical and functional characteristics that are shared between ASAR6 and ASAR15, and

523 therefore vlinc273 is an ASAR (designated here as ASAR6-141). In addition, while

524 ASAR6 RNA was not annotated as a vlincRNA in any previous publication, our RNA-seq

525 data from HTD114 cells indicates that ASAR6 RNA has all of the characteristics of a

526 vlincRNA (see Fig. S7). Furthermore, we note that there are two annotated vlincRNAs

527 (vlinc253 and vlinc254) that map within the ~1.2 mb domain of asynchronous replication

528 that we previously associated with the ASAR6 locus [(Donley et al. 2013); Fig. S7].

529 Therefore, vlinc253 and vlinc254 display asynchronous replication that is coordinated with

530 ASAR6, ASAR6-141 and all other PRME genes on human chromosome 6 [see (Donley

531 et al. 2013)]. Taken together, these observations raise the intriguing possibility that these

532 other vlincRNAs are also ASARs. Finally, the clustering of vlincRNA genes with ASAR

533 characteristics, and their apparent tissue-restricted expression patterns (see Fig. 1 and

534 S7), supports a model in which each autosome contains clustered ASAR genes, and that

535 these ASAR clusters, expressing different ASAR transcripts in different tissues, function

24 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

536 as “Inactivation/Stability Centers” that control replication timing, monoallelic gene

537 expression, and structural stability of each chromosome.

538

25 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

539 Methods:

540 Cell culture

541 HTD114 cells are a human APRT deficient cell line derived from HT1080 cells (Zhu et

542 al. 1994), and were grown in DMEM (Gibco) supplemented with 10% fetal bovine serum

543 (Hyclone). GM12878 cells were obtained from ATCC and were grown in RPMI 1640 (Life

544 Technologies) supplemented with 15% fetal bovine serum (Hyclone). Primary blood

545 lymphocytes were isolated after venipuncture into a Vacutainer CPT (Becton Dickinson,

546 Franklin Lakes, NJ) per the manufacturer’s recommendations and grown in 5 mL RPMI

547 1640 (Life Technologies) supplemented with 10% fetal bovine serum (Hyclone) and 1%

548 phytohemagglutinin (Life Technologies). All cells were grown in a humidified incubator at

549 37C in a 5% carbon dioxide atmosphere.

550

551 RNA-seq:

552 Nuclei were isolated from HTD114 cells following lysis in 0.5% NP40, 140 mM NaCl, 10

553 mM Tris-HCl (pH 7.4), and 1.5 mM MgCl2. Nuclear RNA was isolated using Trizol reagent

554 using the manufacturer’s instructions, followed by DNase treatment to remove possible

555 genomic DNA contamination. RNA-seq was carried out at Novogene. Briefly, ribosomal

556 RNAs were removed using the Ribo-Zero kit (Illumina), RNA was fragmented into 250-

557 300bp fragments, and cDNA libraries were prepared using the Directional RNA Library

558 Prep Kit (NEB). Paired end sequencing was done on a NoaSeq 6000. Triplicate samples

559 were merged and aligned to the human genome (hg19) using the STAR aligner (Dobin et

560 al. 2013) with default settings. Duplicate reads and reads with map quality below 30 were

561 removed with SAMtools (Li et al. 2009).

26 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

562

563 DNA FISH

564 Mitotic chromosome spreads were prepared as described previously (Smith et al. 2001).

565 After RNase (100µg/ml) treatment for 1h at 37ºC, slides were washed in 2XSSC and

566 dehydrated in an ethanol series and allowed to air dry. Chromosomal DNA on the slides

567 was denatured at 75ºC for 3 minutes in 70% formamide/2XSSC, followed by dehydration

568 in an ice cold ethanol series and allowed to air dry. BAC and Fosmid DNAs were labeled

569 using nick translation (Vysis, Abbott Laboratories) with Spectrum Orange-dUTP,

570 Spectrum Aqua-dUTP or Spectrum Green-dUTP (Vysis). Final probe concentrations

571 varied from 40-60 ng/µl. Centromeric probe cocktails (Vysis) and/or whole chromosome

572 paint probes (Metasystems) plus BAC or Fosmid DNAs were denatured at 75ºC for 10

573 minutes and prehybridized at 37ºC for 10 minutes. Probes were applied to denatured

574 slides and incubated overnight at 37ºC. Post-hybridization washes consisted of one 3-

575 minute wash in 50% formamide/2XSSC at 40ºC followed by one 2-minute rinse in PN

576 (0.1M Na2HPO4, pH 8.0/2.5% Nonidet NP-40) buffer at RT. Coverslips were mounted with

577 Prolong Gold antifade plus DAPI (Invitrogen) and viewed under UV fluorescence

578 (Olympus).

579

580 ReTiSH

581 We used the ReTiSH assay essentially as described (Schlesinger et al. 2009). Briefly,

582 unsynchronized, exponentially growing cells were treated with 30μM BrdU (Sigma) for 6

583 or 5 and 14 hours. Colcemid (Sigma) was added to a final concentration of 0.1 μg/mL for

584 1 h at 37°C. Cells were trypsinized, pelleted by centrifugation at 1,000 rpm, and

27 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

585 resuspended in prewarmed hypotonic KCl solution (0.075 M) for 40 min at 37°C. Cells

586 were pelleted by centrifugation and fixed with methanol-glacial acetic acid (3:1). Fixed

587 cells were drop gently onto wet, cold slides and allowed to air-dry. Slides were treated

588 with 100μg/ml RNAse A at 37°C for 10 min. Slides were rinsed briefly in H20 followed by

589 fixation in 4% formaldehyde at room temperature for 10 minutes. Slides were incubated

590 with pepsin (1 mg/mL in 2N HCl) for 10 min at 37°C, and then rinsed again with H20 and

591 stained with 0.5 μg/μL Hoechst 33258 (Sigma) for 15 minutes. Slides were flooded with

592 200μl 2xSSC, coversliped and exposed to 365-nm UV light for 30 min using a UV

593 Stratalinker 2400 transilluminator (Stratagene). Slides were rinsed with H20 and drained.

594 Slides were incubated with 100μl of 3U/μl of ExoIII (Fermentas) in ExoIII buffer for 15 min

595 at 37°C. The slides were then processed directly for DNA FISH as described above,

596 except with the absence of a denaturation step. ASAR6 DNA was detected with BAC

597 RP11-767E7, and ASAR6-141 DNA was detected with BAC RP11-715D3.

598

599 RNA-DNA FISH

600 Cells were plated on glass microscope slides at ~50% confluence and incubated for 4

o 601 hours in complete media in a 37 C humidified CO2 incubator. Slides were rinsed 1X with

602 sterile RNase free PBS. Cell Extraction was carried out using ice cold solutions as follows:

603 Slides were incubated for 30 seconds in CSK buffer (100mM NaCl/300mM sucrose/3mM

604 MgCl2/10mM PIPES, pH 6.8), 10 minutes in CSK buffer/0.1% Triton X-100, followed by

605 30 seconds in CSK buffer. Cells were then fixed in 4% paraformaldehyde in PBS for 10

606 minutes and stored in 70% EtOH at -20°C until use. Just prior to RNA FISH, slides were

607 dehydrated through an EtOH series and allowed to air dry. Denatured probes were

28 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

608 prehybridized at 37°C for 10 min, applied to non-denatured slides and hybridized at 37oC

609 for 14-16 hours. Post-hybridization washes consisted of one 3-minute wash in 50%

610 formamide/2XSSC at 40ºC followed by one 2-minute rinse in 2XSSC/0.1% TX-100 for 1

611 minute at RT. Slides were then fixed in 4% paraformaldehyde in PBS for 5 minutes at RT,

612 and briefly rinsed in 2XSSC/0.1% TX-100 at RT. Coverslips were mounted with Prolong

613 Gold antifade plus DAPI (Invitrogen) and slides were viewed under UV fluorescence

614 (Olympus). Z-stack images were generated using a Cytovision workstation. After

615 capturing RNA FISH signals, the coverslips were removed, the slides were dehydrated in

616 an ethanol series, and then processed for DNA FISH, beginning with the RNase treatment

617 step, as described above.

618

619 Replication timing assay

620 The BrdU replication timing assay was performed as described previously on

621 exponentially dividing cultures and asynchronously growing cells (Smith and Thayer

622 2012). Mitotic chromosome spreads were prepared and DNA FISH was performed as

623 described above. The incorporated BrdU was then detected using a FITC-labeled anti-

624 BrdU antibody (Roche). Coverslips were mounted with Prolong Gold antifade plus DAPI

625 (Invitrogen), and viewed under UV fluorescence. All images were captured with an

626 Olympus BX Fluorescent Microscope using a 100X objective, automatic filter-wheel and

627 Cytovision workstation. Individual chromosomes were identified with either chromosome-

628 specific paints, centromeric probes, BACs or by inverted DAPI staining. Utilizing the

629 Cytovision workstation, each chromosome was isolated from the metaphase spread and

630 a line drawn along the middle of the entire length of the chromosome. The Cytovision

29 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

631 software was used to calculate the pixel area and intensity along each chromosome for

632 each fluorochrome occupied by the DAPI and BrdU (FITC) signals. The total amount of

633 fluorescent signal in each chromosome was calculated by multiplying the average pixel

634 intensity by the area occupied by those pixels. The BrdU incorporation into human

635 chromosome 6 homologs containing CRISPR/Cas9 modifications was calculated by

636 dividing the total incorporation into the chromosome with the smaller chromosome 6

637 centromere (6B) divided by the BrdU incorporation into the chromosome 6 with the larger

638 centromere (6A) within the same cell. Boxplots were generated from data collected from

639 8-12 cells per clone or treatment group. Differences in measurements were tested across

640 categorical groupings by using the Kruskal-Wallis test (Kruskal 1964) and listed as P-

641 values for the corresponding plots.

642

643 CRISPR/Cas9 engineering

644 Using Lipofectamine 2000, according to the manufacturer’s recommendations, we co-

645 transfected HTD114 cells with plasmids encoding GFP, sgRNAs and Cas9 endonuclease

646 (Origene). Each plasmid encoded sgRNAs were designed to bind at the indicated

647 locations (Fig. 1; also see Table S1). 48h after transfection, cells were plated at clonal

648 density and allowed to expand for 2-3 weeks. The presence of deletions in were confirmed

649 by PCR using the primers described in Supplemental Table S1. The single cell colonies

650 that grew were analyzed for heterozygous deletions by PCR. We used retention of a

651 heterozygous SNPs (see Table S1) to identify the disrupted allele (CHR6A vs CHR6B),

652 and homozygosity at this SNP confirmed that cell clones were homogenous.

653 654

30 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

655 References 656 657 Alexander MK, Mlynarczyk-Evans S, Royce-Tolland M, Plocik A, Kalantry S,

658 Magnuson T, Panning B. 2007. Differences between homologous alleles

659 of olfactory receptor genes require the Polycomb Group protein Eed. J

660 Cell Biol 179: 269-276.

661 Allen E, Horvath S, Tong F, Kraft P, Spiteri E, Riggs AD, Marahrens Y. 2003.

662 High concentrations of long interspersed nuclear element sequence

663 distinguish monoallelically expressed genes. Proc Natl Acad Sci U S A

664 100: 9940-9945.

665 Attig J, Agostini F, Gooding C, Chakrabarti AM, Singh A, Haberman N, Zagalak

666 JA, Emmett W, Smith CWJ, Luscombe NM et al. 2018. Heteromeric RNP

667 Assembly at LINEs Controls Lineage-Specific RNA Processing. Cell 174:

668 1067-1081.e1017.

669 Bailey JA, Carrel L, Chakravarti A, Eichler EE. 2000. Molecular evidence for a

670 relationship between LINE-1 elements and X chromosome inactivation:

671 the Lyon repeat hypothesis. Proc Natl Acad Sci U S A 97: 6634-6639.

672 Bartholdy B, Mukhopadhyay R, Lajugie J, Aladjem MI, Bouhassira EE. 2015.

673 Allele-specific analysis of DNA replication origins in mammalian cells. Nat

674 Commun 6: 7051.

675 Bartolomei MS. 2009. Genomic imprinting: employing and avoiding epigenetic

676 processes. Genes Dev 23: 2124-2133.

31 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

677 Breger KS, Smith L, Thayer MJ. 2005. Engineering translocations with delayed

678 replication: evidence for cis control of chromosome replication timing. Hum

679 Mol Genet 14: 2813-2827.

680 Breger KS, Smith L, Turker MS, Thayer MJ. 2004. Ionizing radiation induces

681 frequent translocations with delayed replication and condensation. Cancer

682 Research 64: 8231-8238.

683 Bryois J, Buil A, Evans DM, Kemp JP, Montgomery SB, Conrad DF, Ho KM, Ring

684 S, Hurles M, Deloukas P et al. 2014. Cis and trans effects of human

685 genomic variants on gene expression. PLoS Genet 10: e1004461.

686 Caron M, St-Onge P, Drouin S, Richer C, Sontag T, Busche S, Bourque G,

687 Pastinen T, Sinnett D. 2018. Very long intergenic non-coding RNA

688 transcripts and expression profiles are associated to specific childhood

689 acute lymphoblastic leukemia subtypes. PLoS One 13: e0207250.

690 Chang BH, Smith L, Huang J, Thayer M. 2007. Chromosomes with delayed

691 replication timing lead to checkpoint activation, delayed recruitment of

692 Aurora B and chromosome instability. Oncogene 26: 1852-1861.

693 Chen J, Rattner A, Nathans J. 2006. Effects of L1 retrotransposon insertion on

694 transcript processing, localization and accumulation: lessons from the

695 retinal degeneration 7 mouse and implications for the genomic ecology of

696 L1 elements. Hum Mol Genet 15: 2146-2156.

697 Chess A. 2012. Mechanisms and consequences of widespread random

698 monoallelic expression. Nat Rev Genet 13: 421-428.

32 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

699 Dileep V, Gilbert DM. 2018. Single-cell replication profiling to measure stochastic

700 variation in mammalian replication timing. Nature communications 9: 427.

701 Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, Batut P,

702 Chaisson M, Gingeras TR. 2013. STAR: ultrafast universal RNA-seq

703 aligner. Bioinformatics (Oxford, England) 29: 15-21.

704 Donley N, Smith L, Thayer MJ. 2015. ASAR15, A cis-Acting Locus that Controls

705 Chromosome-Wide Replication Timing and Stability of Human

706 Chromosome 15. PLoS Genet 11: e1004923.

707 Donley N, Stoffregen EP, Smith L, Montagna C, Thayer MJ. 2013. Asynchronous

708 Replication, Mono-Allelic Expression, and Long Range Cis-Effects of

709 ASAR6. PLoS Genet 9: e1003423.

710 Ensminger AW, Chess A. 2004. Coordinated replication timing of monoallelically

711 expressed genes along human autosomes. Hum Mol Genet 13: 651-658.

712 Fragkos M, Ganier O, Coulombe P, Mechali M. 2015. DNA replication origin

713 activation in space and time. Nat Rev Mol Cell Biol 16: 360-374.

714 Galupa R, Heard E. 2018. X-Chromosome Inactivation: A Crossroads Between

715 Chromosome Architecture and Gene Regulation. Annu Rev Genet 52:

716 535-566.

717 Gendrel AV, Attia M, Chen CJ, Diabangouaya P, Servant N, Barillot E, Heard E.

718 2014. Developmental dynamics and disease potential of random

719 monoallelic gene expression. Dev Cell 28: 366-380.

33 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

720 Gendrel AV, Marion-Poll L, Katoh K, Heard E. 2016. Random monoallelic

721 expression of genes on autosomes: Parallels with X-chromosome

722 inactivation. Semin Cell Dev Biol.

723 Gimelbrant A, Hutchinson JN, Thompson BR, Chess A. 2007. Widespread

724 monoallelic expression on human autosomes. Science 318: 1136-1140.

725 Goldmit M, Bergman Y. 2004. Monoallelic gene expression: a repertoire of

726 recurrent themes. Immunol Rev 200: 197-214.

727 Hiratani I, Leskovar A, Gilbert DM. 2004. Differentiation-induced replication-

728 timing changes are restricted to AT-rich/long interspersed nuclear element

729 (LINE)-rich isochores. Proc Natl Acad Sci U S A 101: 16861-16866.

730 Kapranov P, St Laurent G, Raz T, Ozsolak F, Reynolds CP, Sorensen PH,

731 Reaman G, Milos P, Arceci RJ, Thompson JF et al. 2010. The majority of

732 total nuclear-encoded non-ribosomal RNA in a human cell is 'dark matter'

733 un-annotated RNA. BMC biology 8: 149.

734 Kruskal JB. 1964. Multidimensional scaling by optimizing goodness of fit to a

735 nonmetric hypothesis. Psychometrika 29: 1–27.

736 Lehman A, Thouta S, Mancini GMS, Naidu S, van Slegtenhorst M, McWalter K,

737 Person R, Mwenifumbo J, Salvarinova R, Guella I et al. 2017. Loss-of-

738 Function and Gain-of-Function Mutations in KCNQ5 Cause Intellectual

739 Disability or Epileptic Encephalopathy. Am J Hum Genet 101: 65-74.

740 Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis

741 G, Durbin R. 2009. The Sequence Alignment/Map format and SAMtools.

742 Bioinformatics (Oxford, England) 25: 2078-2079.

34 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

743 Li SM, Valo Z, Wang J, Gao H, Bowers CW, Singer-Sam J. 2012. Transcriptome-

744 wide survey of mouse CNS-derived cells reveals monoallelic expression

745 within novel gene families. PLoS One 7: e31751.

746 Lin M, Hrabovsky A, Pedrosa E, Wang T, Zheng D, Lachman HM. 2012. Allele-

747 biased expression in differentiating human neurons: implications for

748 neuropsychiatric disorders. PLoS One 7: e44017.

749 Luedi PP, Dietrich FS, Weidman JR, Bosko JM, Jirtle RL, Hartemink AJ. 2007.

750 Computational and experimental identification of novel human imprinted

751 genes. Genome Res 17: 1723-1730.

752 Lyon MF. 1998. X-chromosome inactivation: a repeat hypothesis. Cytogenet Cell

753 Genet 80: 133-137.

754 -. 2003. The Lyon and the LINE hypothesis. Semin Cell Dev Biol 14: 313-318.

755 Mostoslavsky R, Singh N, Tenzen T, Goldmit M, Gabay C, Elizur S, Qi P,

756 Reubinoff BE, Chess A, Cedar H et al. 2001. Asynchronous replication

757 and allelic exclusion in the immune system. Nature 414: 221-225.

758 Petretto E, Mangion J, Dickens NJ, Cook SA, Kumaran MK, Lu H, Fischer J,

759 Maatz H, Kren V, Pravenec M et al. 2006. Heritability and tissue specificity

760 of expression quantitative trait loci. PLoS Genet 2: e172.

761 Platt EJ, Smith L, Thayer MJ. 2017. L1 retrotransposon antisense RNA within

762 ASAR lncRNAs controls chromosome-wide replication timing. J Cell Biol.

763 Schlesinger S, Selig S, Bergman Y, Cedar H. 2009. Allelic inactivation of rDNA

764 loci. Genes Dev 23: 2437-2447.

35 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

765 Signor SA, Nuzhdin SV. 2018. The Evolution of Gene Expression in cis and

766 trans. Trends Genet 34: 532-544.

767 Singh N, Ebrahimi FA, Gimelbrant AA, Ensminger AW, Tackett MR, Qi P,

768 Gribnau J, Chess A. 2003. Coordination of the random asynchronous

769 replication of autosomal loci. Nat Genet 33: 339-341.

770 Smith L, Plug A, Thayer M. 2001. Delayed Replication Timing Leads to Delayed

771 Mitotic Chromosome Condensation and Chromosomal Instability of

772 Chromosome Translocations. Proc Natl Acad Sci U S A 98: 13300-13305.

773 Smith L, Thayer M. 2012. Chromosome replicating timing combined with

774 fluorescent in situ hybridization. J Vis Exp 10: e4400.

775 St Laurent G, Savva YA, Kapranov P. 2012. Dark matter RNA: an intelligent

776 scaffold for the dynamic regulation of the nuclear information landscape.

777 Frontiers in genetics 3: 57.

778 St Laurent G, Shtokalo D, Dong B, Tackett MR, Fan X, Lazorthes S, Nicolas E,

779 Sang N, Triche TJ, McCaffrey TA et al. 2013. VlincRNAs controlled by

780 retroviral elements are a hallmark of pluripotency and cancer. Genome

781 Biol 14: R73.

782 St Laurent G, Vyatkin Y, Antonets D, Ri M, Qi Y, Saik O, Shtokalo D, de Hoon

783 MJ, Kawaji H, Itoh M et al. 2016. Functional annotation of the vlinc class of

784 non-coding RNAs using systems biology approach. Nucleic Acids

785 Research 2 44: 3233-3252.

36 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

786 Stoffregen EP, Donley N, Stauffer D, Smith L, Thayer MJ. 2011. An autosomal

787 locus that controls chromosome-wide replication timing and mono-allelic

788 expression. Hum Mol Genet 20: 2366-2378.

789 Thayer MJ. 2012. Mammalian chromosomes contain cis-acting elements that

790 control replication timing, mitotic condensation, and stability of entire

791 chromosomes. Bioessays 34: 760-770.

792 Zhu Y, Bye S, Stambrook PJ, Tischfield JA. 1994. Single-base deletion induced

793 by benzo[a]pyrene diol epoxide at the adenine phosphoribosyltransferase

794 locus in human fibrosarcoma cell lines. Mutat Res 321: 73-79.

795 796

37 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

797 Figure Legends 798

799 Figure 1. UCSC Genome Browser view of the vlinc cluster on chromosome 6 between

800 140.3 and 141.3 mb. The genomic locations of vlinc271, vlinc1010, vlinc1011, vlinc1012,

801 vlinc272 and vlinc273 are illustrated using the UCSC Genome Browser (see Table S4 for

802 genomic locations). RNA-seq data from nuclear RNA isolated from HTD114 is shown

803 (HTD114 RNA-seq). Long RNA-seq data from the ENCODE Project (Cold Spring Harbor

804 Lab) is shown using the Contigs view. Expression from the human cell lines GM12878

805 (red), HepG2 (magenta), and K562 (blue) are shown. RNA from total cellular Poly A+ (cel

806 pA+), total cellular Poly A- (cel pA-), nuclear Poly A+ (nuc pA+), nuclear Poly A- (nuc pA-

807 ), cytoplasmic Poly A+ (cyt pA+), and cytoplasmic Poly A- (cyt pA-) are shown. Also shown

808 are the Repeating Elements using the RepeatMasker track. The location of five RNA FISH

809 probes (Fosmids) that were used to detect expression of vlinc1012 (G248P87615G6),

810 vlinc272 (G248P89033E3 and G248P83206B11) and vlinc273 (G248P85904G6 and

811 G248P81345F10) are shown.

812

813 Figure 2. Monoallelic expression and nuclear retention of vlinc273 in HTD114 cells. (A)

814 DNA sequencing traces from PCR products designed to detect SNPs rs989613623 and

815 rs2328092 (see Fig. S1A and Table S2). PCRs were carried out on genomic DNAs

816 isolated from HTD114, two mono-chromosomal hybrids containing the two different

817 chromosome 6s from HTD114 {L(Hyg)-1 contains chromosome 6A (CHR6A) and

818 expresses ASAR6, and L(Neo)-38 contains chromosome 6B (CHR6B) and is silent for

819 ASAR6 (Stoffregen et al. 2011)}. The top and bottom panels also show the sequencing

38 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

820 traces from HTD114 cDNA (RNA). The asterisks mark the location of the heterozygous

821 SNPs. (B-F) RNA-DNA FISH to detect vlinc273 expression in HTD114 cells. Fosmid

822 G248P81345F10 was used as probe to detect vlinc273 RNA (green), and a chromosome

823 6 centromeric probe was used to detect chromosome 6 DNA (red). The nuclear DNA was

824 stained with DAPI. Bars are 2.5 uM. (G-J) RNA-DNA FISH to detect vlinc273 and ASAR6

825 expression in HTD114 cells. Fosmid G248P81345F10 was used as probe to detect

826 vlinc273 RNA (green), Fosmid G248P86031A6 was used as probe to detect ASAR6 RNA

827 (red), and a chromosome 6 paint was used to detect chromosome 6 DNA (magenta). The

828 nuclear DNA was stained with DAPI. Bars are 2.5 uM. Quantitation of the number of RNA

829 FISH signals is shown in Fig. S2A and S2B.

830

831 Figure 3. Monoallelic expression and nuclear retention of vlinc273 in EBV transformed

832 lymphoblasts and primary blood lymphocytes. (A-E) RNA-DNA FISH to detect vlinc273

833 expression in GM12878 EBV transformed lymphocytes. Fosmid G248P81345F10 was

834 used to detect vlinc273 RNA (green), and a chromosome 6 centromeric probe (CHR6

835 cen) was used to detect chromosome 6 DNA (red). (F-J) RNA FISH to detect coordinated

836 expression of vlinc273 and KCNQ5, a known random monoallelic gene, in primary blood

837 lymphocytes. Fosmid G248P81345F10 was used to detect vlinc273 RNA (green), and

838 Fosmid G248P80791F6 was used to detect expression of the first intron of KCNQ5. (K

839 and I) RNA FISH to detect expression of vlinc273 and XIST RNAs, in female primary

840 blood lymphocytes. The nuclear DNA was stained with DAPI. Bars are 2.5 uM.

841 Quantitation of the number of RNA FISH signals per nucleus and the cis versus trans

842 expression is shown in Fig. S2B-D.

39 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

843

844 Figure 4. Coordinated asynchronous replication timing on chromosome 6. (A) Schematic

845 representation of the ReTiSH assay. Cells were exposed to BrdU during the entire length

846 of S phase (14 hours) or only during late S phase (5 hours). The ReTiSH assay can

847 distinguish between alleles that replicate early (E) and late (L) in S phase. (B-D) Mitotic

848 spreads from human primary blood lymphocytes were processed for ReTiSH and

849 hybridized with three different FISH probes. First, each hybridization included a

850 centromeric probe to chromosome 6 (magenta). Arrows mark the centromeric signals in

851 panels B (5 hours) and C (14 hours). Each assay also included BAC probes representing

852 ASAR6 (RP11-374I15; green) and vlinc273 (RP11-715D3; red). Panel D shows the two

853 chromosome 6s, from both the 5 hour and 14 hour time points, aligned at their

854 centromeres. The ASAR6 BAC and the vlinc273 BAC show hybridization signals on the

855 same chromosome 6 at the 5-hour time point, and as expected hybridized to both

856 chromosome 6s at the 14 hour time point. The chromosomal DNA was stained with DAPI.

857 (E-G) ReTiSH assay on HTD114 cells. Each ReTiSH assay included a centromeric probe

858 to chromosome 6 (magenta). Arrows mark the centromeric signals in panels E (5 hours)

859 and F (14 hours). Each assay also included BAC probes for ASAR6 (RP11-374I15; green)

860 and vlinc273 (RP11-715D3; red). The chromosomal DNA was stained with DAPI. The

861 ASAR6 BAC and the vlinc273 BAC show hybridization signals to one chromosome 6

862 homolog (CHR6A) at the 5-hour time point, and as expected hybridized to both

863 chromosome 6s at the 14-hour time point.

864

40 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

865 Figure 5. Delayed replication of chromosome 6 following disruption of vlinc273. (A

866 and B) A representative mitotic spread from BrdU (green) treated HTD114 cells

867 containing a deletion of the expressed allele of the vlinc273 locus. Mitotic cells were

868 subjected to DNA FISH using a chromosome 6 centromeric probe (red). The larger

869 centromere resides on the chromosome 6 with the expressed ASAR6 allele and the silent

870 vlinc273 allele (6A). (C) The two chromosome 6s were extracted from panels A and B

871 and aligned to show the BrdU incorporation and centromeric signals. (D) Pixel intensity

872 profiles of BrdU incorporation and DAPI staining along the (6A) and (6B) chromosomes

873 from panel C. (E) BrdU quantification along 6A and 6B from panel D. (F) The ratio of DNA

874 synthesis into the two chromosome 6s was calculated by dividing the BrdU incorporation

875 in CHR6B by the incorporation in CHR6A in multiple cells. The dots show the BrdU

876 incorporation ratios for the individual cells assayed. The box plots show the ratio of

877 incorporation before (Intact, dark blue), and in heterozygous deletions of the entire locus

878 (6A dark purple; and 6B light blue), which included vlinc271, vlinc1010, vlinc1011,

879 vlinc1012, vlinc272, and vlinc273, see maps in Figs. 1 and S1A. Heterozygous deletions

880 affecting vlinc273 only from the silent (6A orange) or expressed (6B green) alleles are

881 shown. A heterozygous deletion affecting vlinc271, vlinc1010, vlinc1011, vlinc1012, and

882 vlinc272 on CHR6B (6B) is shown in pink. Also shown are heterozygous deletions

883 affecting ASAR6 from the expressed (6A magenta) or silent (6B yellow) alleles. Error

884 bars are SD. P values of <1 x10-4 are indicated by ***, and P values of >1 x 10-1 are

885 indicated by *, and were calculated using the Kruskal-Wallis test.

886

41 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

887 Figure 6. “ASAR” model of replication timing on chromosome 6. The two homologs

888 of human chromosome 6 are shown (gray) with hypothetical origins of replication depicted

889 as blue bars. Expression of ASAR6 and ASAR6-141 genes is monoallelic, resulting in a

890 reciprocal expression pattern with an expressed or active ASAR (green or red clock) and

891 a silent or inactive ASAR (white clock) on each homolog. The red and green clouds

892 surrounding the chromosomes represent “ASAR” RNA expressed from the different active

893 “ASARs” on each homolog.

894

42 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Table 1. Nuclear enrichment of vlinc273 RNA Fold vlinc273 nuclear Cell line RPKM enrichment GM12878 Nucleus 0.83 172 GM12867 Cytoplasm 0.0048

HEPG2 Nucleus 4.67 212 HEPG2 Cytoplasm 0.022

K562 Nucleus 0.93 103 K562 Cytoplasm 0.009

895 896 897

43 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

898 Table 2. Coordinated Asynchronous Replication Timing by ReTiSH. PBLs Locus 1 Locus 2 cis (%) trans (%) P value ASAR6-141 ASAR6 75 25 <1 x 10-3

HTD114 Locus 1 Locus 2 cis (%) trans (%) P value ASAR6-141 CHR6 cen* 77 23 <1 x 10-4 ASAR6 CHR6 cen* 88 12 <1 x 10-6 ASAR6-141 ASAR6 76 24 <1 x 10-4

CHR6 cen* indicates the large centromere on one of the chromosome 6s in HTD114 cells.

899

900

44 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

901 Supporting Information 902 903 Figure S1. Monoallelic expression and heterozygous deletions of the vlinc cluster at ~141 904 mb of chromosome 6. (A) UCSC Genome Browser view of the vlinc cluster on 905 chromosome 6 between 140.3 and 141.3 mb showing the location of the vlincRNA genes 906 (vlinc271, vlinc1010, vlinc1011, vlinc1012, vlinc272, and vlinc273), Fosmids, PCR 907 primers (red half arrows marked i-viii; see Table S2), sgRNAs (see Table S2), and 908 heterozygous SNPs (see Table S2). Green triangles show the location of the 909 heterozygous SNPs corresponding to: (*) rs2328092, rs989613623, and A/T at 910 140,943,663, (#) rs685606, rs17070386, rs72990126, and (&) rs9399336. (B) DNA 911 sequencing traces from PCR products (using primers iii and iv; see Table S2), which show 912 the A/T SNP at 140,943,663. PCRs were carried out on genomic DNAs isolated from 913 HTD114, two mono-chromosomal hybrids containing the two different chromosome 6s 914 from HTD114 {L(Hyg)-1 contains chromosome 6A (CHR6A) and expresses ASAR6, and 915 L(Neo)-38 contains chromosome 6B (CHR6B) and is silent for ASAR6 (Stoffregen et al. 916 2011)}. The sequencing trace from HTD114 cDNA (RNA) is shown. The bottom two 917 panels, marked 6B(R57-37) and 6A(R57-27), show the sequencing traces from PCR 918 products produced from genomic DNA isolated from heterozygous deletions of the entire 919 vlinc cluster (using sgRNA-1 and sgRNA-3) from CHR6B and CHR6A, respectively. Note 920 that these PCR products are generated from the non-deleted alleles. The arrows mark 921 the location of the heterozygous SNP. 922 923 Figure S2. Quantitation of RNA FISH signals. (A) RNA-DNA FISH to detect vlinc273 and 924 ASAR6 expression in HTD114 cells. Fosmids G248P81345F10 and G248P85904G6 925 were used as probe to detect vlinc273 RNA and Fosmid G248P86031A6 was used as 926 probe to detect ASAR6 RNA. RNA FISH signals were scored in >200 cells and the 927 percent of cells with 1 (green), 2 (red), or No (blue) signals are shown. (B) RNA-DNA 928 FISH was used to detect vlinc273 (G248P81345F10 probe) plus ASAR6 (G248P86031A6 929 probe) RNAs simultaneously in HTD114 cells. A chromosome 6 whole chromosome paint 930 or chromosome 6 centromeric probe was used to detect chromosome 6 DNA. RNA-DNA 931 FISH was used to detect vlinc273 (G248P81345F10 probe) plus KCNQ5

45 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

932 (G248P880791F6) expression in primary blood lymphocytes isolated from 2 unrelated 933 individuals, one female (XX) and one male (XY), and in GM12878 cells. The percentage 934 of nuclei showing hybridization to the same (i.e. in cis) or opposite (i.e. in trans) 935 chromosome 6s. A minimum of 100 cells with both signals were scored for each assay, 936 and the percentage of nuclei showing hybridization to the same (i.e. in cis; green) or 937 opposite (i.e. in trans; red) chromosome 6s is shown. (C) RNA-DNA FISH was used to 938 detect vlinc273 (G248P81345F10 probe) plus KCNQ5 (G248P880791F6 probe) RNAs in 939 GM12878 cells. A chromosome 6 centromeric probe was used to detect chromosome 6 940 DNA. RNA FISH signals were scored in >200 cells and the percent of cells with 1 (green), 941 2 (red), or No (blue) signals are shown. (D) RNA-DNA FISH was used to detect vlinc273 942 (G248P81345F10 probe) plus KCNQ5 (G248P880791F6 probe) RNAs simultaneously in 943 primary blood lymphocytes isolated from 2 unrelated individuals, one female (XX) and 944 one male (XY). A chromosome 6 centromeric probe was used to detect chromosome 6 945 DNA. RNA FISH signals were scored in >200 cells and the percent of cells with 1 (green), 946 2 (red), or No (blue) signals are shown. 947 948 Figure S3. DNA sequencing traces from PCR products generated using either primers 949 vii plus viii (HTD114, CHR6A, and CHR6B), iii plus viii {6A(1-15-7), 6B(4-44-36), 950 6B(4-45-53); see Fig. S1A and Table S2} or i plus viii {6A(R56-1), 6A(R57-27), 951 6B(R57-37)}. All of the sequence traces show the DNA region containing the 952 heterozygous SNP rs9399336 (see Table S2), which is marked by arrows. PCR products 953 were generated using primers vii plus viii on genomic DNA isolated from HTD114 and two 954 mono-chromosomal hybrids containing the two different chromosome 6s from HTD114 955 {L(Hyg)-1 contains chromosome 6A (CHR6A), and L(Neo)-38 contains chromosome 6B 956 (CHR6B) (Stoffregen et al. 2011)}. In addition, sequence traces from 6 independent 957 CRISPR/Cas9-mediated deletions, containing deletions of either vlinc273 alone {6A(1- 958 15-7), 6B(4-44-36), 6B(4-45-53)} or deletions of the entire vlincRNA gene cluster 959 {6A(R56-1), 6A(R57-27), 6B(R57-37). Note that these primer combinations amplify 960 junction fragments generated when only vlinc273 is deleted (primers iii plus viii), or when 961 the entire vlinc cluster is deleted (primers i plus viii). Therefore, these junction PCR 962 products are generated from the deleted chromosome, i.e. either 6A or 6B.

46 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

963 Figure S4. Examples of DNA sequencing traces from PCR products generated using 964 primers v plus vii (see Fig. S1A and Table S2). The sequence traces show the DNA 965 regions containing the heterozygous SNPs rs685606, rs17070386, and rs72990126 (see 966 Table S2) The location of the SNPs are marked by arrows. PCR products were generated 967 from genomic DNA isolated from HTD114 and the two mono-chromosomal hybrids 968 containing the two different chromosome 6s from HTD114 {L(Hyg)-1 contains 969 chromosome 6A (CHR6A), and L(Neo)-38 contains chromosome 6B (CHR6B) 970 (Stoffregen et al. 2011)}. In addition, sequence traces from 4 independent CRISPR/Cas9- 971 mediated deletions, containing deletions of vlinc273 {6A(1-15-7), 6B(4-66-36), 6B(4- 972 66-48), 6B(4-45-53)}. Note that these PCR products are generated from the non-deleted 973 allele. The arrows mark the locations of the heterozygous SNPs. 974 975 Figure S5. Heterozygous deletions affect the expressed alleles of vlinc273 or ASAR6. 976 (A) RNA-DNA FISH was used to detect vlinc273 and ASAR6 expression in cells with 977 heterozygous deletions of the expressed alleles of vlinc273 {6B(vlinc273)} or the entire 978 vlinc RNA cluster {6B(vlinc271, vlinc1010, vlinc1011, vlinc1012, vlinc272, vlinc273)}. 979 Fosmid G248P81345F10 was used as probe to detect vlinc273 RNA and Fosmid 980 G248P86031A6 was used as probe to detect ASAR6 RNA. RNA FISH signals were 981 scored in >200 cells and the percent of cells with 1 (green), 2 (red), or No (blue) signals 982 are shown. (B) RNA-DNA FISH was used to detect vlinc273 and ASAR6 expression in 983 cells with heterozygous deletions of the expressed alleles of ASAR6 {6A(ASAR6-1) and 984 6A(ASAR6-2)}. Fosmid G248P81345F10 was used as probe to detect vlinc273 RNA 985 and Fosmid G248P86031A6 was used as probe to detect ASAR6 RNA. RNA FISH 986 signals were scored in >200 cells and the percent of cells with 1 (green), 2 (red), or No 987 (blue) signals are shown. 988 989 Figure S6. Delayed replication of chromosome 6 following disruption of ASAR6. (A and 990 B) A representative mitotic spread from BrdU (green) treated cells containing a deletion 991 of the expressed allele of ASAR6 (Platt et al. 2017). Mitotic cells were subjected to DNA 992 FISH using a chromosome 6 centromeric probe (red). The larger centromere resides on 993 the chromosome 6 with the expressed ASAR6 allele (CHR6A). Bar is 10 uM. (C) The two

47 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

994 chromosome 6s were extracted from panels A and B and aligned to show the BrdU 995 incorporation and centromeric signals. (D) Pixel intensity profiles of BrdU incorporation 996 and DAPI staining along the (6A) and (6B) chromosomes from panel C. The long (q) and 997 short (p) arms of chromosome 6 are indicated. Bar is 2 uM. (E) BrdU quantification along 998 6A and 6B from panel D. Figure 5F shows the ratio of DNA synthesis into the two 999 chromosome 6s in multiple cells by dividing the BrdU incorporation in 6B by the 1000 incorporation in 6A. 1001 1002 Figure S7. Expression and asynchronous replication of ASAR6. UCSC Genome

1003 Browser view of the ASAR6 RNA-seq data, showing the reads from the plus and minus

1004 strands in separate tracks, from HTD114 nuclear ribo-minus RNA. We previously mapped

1005 the ~1.2 mb asynchronous replication domain associated with ASAR6 as indicated {see

1006 (Donley et al. 2013)}. We note that the RNA-seq reads associated with ASAR6 fulfill all

1007 of the criteria described for vlincRNAs {see (St Laurent et al. 2013; St Laurent et al.

1008 2016)}. Two additional vlincRNAs (253 and 254) also map to the asynchronous replication

1009 domain. Also shown is the Repeat Masker Track.

1010

1011 Table S1. Repetitive elements within vlinc273 (ASAR6-141). The chromosome

1012 position, orientation, size and total bases occupied by LINE, Alu, and other repeats,

1013 from RepeatMasker are shown.

1014

1015 Table S2. DNA oligonucleotides used for sgRNAs and PCR primers. The DNA

1016 sequence of the oligonucleotides and the position on chromosome 6 of the sequences

1017 used for sgRNAs and PCR primers used to screen for deletions. Also shown are the

48 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

1018 heterozygous SNPs within the PCR products used to determine which allele was

1019 expressed and/or deleted following CRISPR/Cas9 expression.

1020

1021 Table S3. Sequence junctions of the CRISPR/Cas9 deletion clones. PCR products

1022 were generated using primers iii and viii on clones 6A(1-15-7), 6B(4-66-48), 6B(4-

1023 44-36), 6B(4-45-53) containing deletions of vlinc273, or i plus viii on clones 6A(R57-

1024 27), 6B(R57-37), 6A(R56-1) containing deletions of the entire vlinc cluster

1025 (vlinc271, vlinc1010, vlinc1011, vlinc1012, vlinc272, vlinc273), or i plus iv on clone

1026 6B(L65-1) containing a deletion of vlinc271, vlinc1010, vlinc1011, vlinc1012, and

1027 vlinc272. PCR products were subjected to Sanger sequencing and the proximal and

1028 distal junctions were determined using BLAST. The deleted alleles were determined

1029 by loss of heterozygosity at 7 different SNPs (for examples see Figs. S1, S3 and S4).

1030

1031 Table S4. Genomic location of vlincRNA genes. The chromosome position,

1032 orientation and size of ASAR6 and the vlincRNA genes used in this study.

49 Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press Downloaded from rnajournal.cshlp.org on October 5, 2021 - Published by Cold Spring Harbor Laboratory Press

Reciprocal monoallelic expression of ASAR lncRNA genes controls replication timing of human chromosome 6.

Michael Heskett, Leslie G Smith, Paul Spellman, et al.

RNA published online March 6, 2020

Supplemental http://rnajournal.cshlp.org/content/suppl/2020/03/06/rna.073114.119.DC1 Material

P

Accepted Peer-reviewed and accepted for publication but not copyedited or typeset; accepted Manuscript manuscript is likely to differ from the final, published version.

Open Access Freely available online through the RNA Open Access option.

Creative This article, published in RNA, is available under a Creative Commons License Commons (Attribution 4.0 International), as described at License http://creativecommons.org/licenses/by/4.0/.

Email Alerting Receive free email alerts when new articles cite this article - sign up in the box at the Service top right corner of the article or click here.

To subscribe to RNA go to: http://rnajournal.cshlp.org/subscriptions

Published by Cold Spring Harbor Laboratory Press for the RNA Society