Regulatory RNPs: A novel class of ribonucleoproteins that contribute to ribosome heterogeneity

Ribonucleoproteins (RNPs), which are comprised of non-coding RNA and associated proteins, are involved in essential cellular processes such as translation and pre-mRNA splicing. One class of RNP is the small Cajal body-specific RNP (scaRNP), which contributes to the biogenesis of small nuclear RNPs (snRNPs) that are central components of the spliceosome. Interestingly, three scaRNAs are internally processed, generating stable nucleolus-enriched RNAs of unknown function. Here we provide evidence that these RNAs become part of novel RNPs we term regulatory RNPs (regRNPs). We postulate that regRNPs can impact rRNA modifications via interactions with the guide RNA component of small nucleolar RNPs (snoRNPs). Most modifications within rRNA (predominantly pseudouridylation and ribose 2’-O-methylation) are conducted by snoRNPs, and we hypothesize that the activity of at least some of these snoRNPs is under the control of regRNPs. Ribosome heterogeneity leading to specialized ribosomes is an exciting emerging concept. Because modifications within rRNA can vary in different physiological or pathological situations, rRNA modifications are thought to be the major source of ribosome heterogeneity. Our identification of regRNPs thus provides important and timely insight into how ribosome heterogeneity may be accomplished. This work also provides additional functional connections between the Cajal body and the nucleolus. Summary Statement Processed scaRNAs give rise to a novel regulatory RNP which regulates the modification of ribosomal RNA. These findings provide insight into the mechanisms governing ribosome heterogeneity.


Introduction
Since the processed fragments derived from scaRNA2, 9 and 17 are enriched in the nucleolus (Tycowski et al., 159 2004), we next examined the localization of scaRNA9 ∆ leader compared to WT scaRNA9. For this work, cells 160 were transfected with DNA encoding WT or ∆ leader scaRNA9, and RNA FISH, using a probe indicated by D in 161 Fig. 1, followed by immunofluorescence was conducted (Fig. 2C). Based on the binding location of probe D, 162 full-length scaRNA9 as well as the processed mgU2-30 domain will be detected by this probe. As expected, 163 WT scaRNA9 is enriched in CBs, nucleoli and nucleoplasm (Fig. 2C), consistent with the full-length scaRNA9 164 being localized in the nucleoplasm/CB and the processed mgU2-30 domain being nucleolar. CBs were 165 detected by anti-coilin staining. The scaRNA9 ∆ leader RNA was also detected in CBs and nucleoli, but the 166 intensity of the nucleolar staining relative to that found in the nucleoplasm was significantly increased for the 167 scaRNA9 ∆ leader compared to the WT scaRNA9 signal (histogram). These localization data support the 168 Northern results and strongly indicate that the 28 nt 5' leader sequence of scaRNA9 is important for the 169 stability of full-length scaRNA9. In the absence of this leader sequence, most of the scaRNA9 transcribed from 170 the plasmid DNA is processed into the mgU2-30 domain, which results in greater levels of nucleolar signal 171 compared to that observed with WT scaRNA9. 172 173 Coilp1 positively influences the processing of scaRNA 9 and 17 174 We have previously shown that COILP1, a pseudogene, encodes a 203 amino acid protein which contains 175 sequence homology with coilin's RNA binding domain, but has an unique 77 amino acid N-terminal sequence 176 (Poole et al., 2016). We further showed that endogenous coilp1 complexes and bacterially purified coilp1 bind 177 in vitro transcribed scaRNA2, scaRNA9, and hTR (Poole et al., 2016). Moreover, we found that GFP and myc 178 tagged coilp1 localizes to the nucleolus. Since scaRNA 2, 9, and 17 give rise to stable processed fragments 179 that localize to the nucleolus (Tycowski et al., 2004), we were interested in determining if coilp1 overexpression 180 or knockdown would, like coilin and SMN (Enwerem et al., 2015), affect the biogenesis and/or stability of these 181 processed fragments. For this work, coilp1 was reduced by RNAi, followed by transfection with DNA encoding 182 scaRNA 2, 9 or 17. To examine the effect of coilp1 overexpression on scaRNA 2, 9 and 17 processing, cells 183 were co-transfected with GFP or GFP-coilp1 and DNA encoding scaRNA 2, 9 or 17. Isolated RNA was then 184 The nucleolus enriched processed fragments of scaRNA 2, 9 and 17: A new class of RNP? 197 Although it has long been known that the fragments derived from scaRNA 2, 9 and 17 are enriched in the 198 nucleolus, the function of these processed RNAs in this cellular locale is unknown. It was previously thought 199 that these RNAs would guide modification of snRNAs that traffic through the nucleolus (Tycowski et al., 2004). 200 Of the U1, U2, U4, U5 and U6 snRNAs in mammalian cells, however, only the pol III-derived U6 has a clear 201 nucleolar pathway (Kiss, 2004). Since the guide RNAs derived from scaRNA 2, 9 and 17 (mgU2-61, mgU2-19, 202 mgU2-30 and mgU4-8) are thought to act upon U2 and U4, not U6, this leaves open the question as to what 203 these RNAs are doing in the nucleolus. It seems highly unlikely that these RNAs are non-functional by-204 products given their stability, which strongly indicates the RNAs are part of an RNP. We initially postulated that 205 scaRNA 2, 9 and 17-derived RNAs become snoRNPs, and aid in the modification of rRNA or the nucleolar-206 trafficked U6 snRNA. Since the processed fragments are box C/D RNAs (and thus guide methylation 207 modifications), we queried several websites, such as snoSCAN (Schattner et al., 2005), that predict target sites 208 for scaRNA 2, 9 and 17 derived RNAs on rRNA. Although these websites returned alignments and methylation 209 target sites on rRNA, these locations did not correspond to any sites with experimentally verified methylation. 210 This lead us to question if these processed RNAs do, in fact, directly modify rRNA. In the process of doing 211 BLAST searches of the scaRNA17-derived mgU4-8 using a snoRNA/scaRNA database (Lestrade et al., 2006), 212 we observed that the 3′ loop region of mgU4-8 can base pair with the snord16 (U16) snoRNA (Fig. 4, bottom  213 right). The snord16 snoRNP guides the methylation of the A484 site of 18S rRNA. We hypothesize that the 214 interaction of mgU4-8 with snord16 might interfere with the association of the snord16 snoRNP with 18S rRNA, 215 thereby inhibiting the methylation A484. In the nucleolus, therefore, mgU4-8 may be part of a new class of 216 RNPs, which we term regulatory RNPs. The key feature of these putative regulatory RNPs is that they are not 217 directly involved in the modification of target sites within the nucleolus, but instead regulate the activity of 218 snoRNPs by interacting with the RNA component of the snoRNP, or the target RNA. Additionally, we further 219 hypothesize that both methylation and pseudouridylation can be regulated by the actions of regulatory RNPs. 220 As shown in Fig. 4, each of the 4 RNAs derived from scaRNA 2, 9 and 17 likely become part of a regulatory 221 RNP that may influence the modification of four sites within 28S rRNA, two sites within 18S rRNA, and two 222 sites within U6 snRNA (Table 1). Regarding U6 snRNA, the 5′ loop region of mgU4-8, intriguingly, contains 8 223 bases that are found in U6 snRNA, and this loop can base pair with U94 and HBII-166. U94 and HBII-166 224 guide the methylation of U6 snRNA at C62 and C60, respectively. 225 226

Reduction of scaRNA17 increases the level of A484 methylation within 18S rRNA 227
To garner experimental support for the existence of regulatory RNPs, cells were treated with control or 228 scaRNA17 siRNAs. In a typical experiment, RNAi reduced scaRNA17 levels by 60% as assessed by qRT-229 PCR. RNA isolated from these cells was subjected to methylation analysis using a primer extension technique 230 that takes advantage of the fact that reverse transcriptase pauses at sites of methylation when dNTP levels are 231 low (Maden et al., 1995). Primers were designed to interrogate the methylation status of the A484 site of 18S 8 rRNA. The A484 site of 18S rRNA is known to be modified, presumably due to association with the snord16 233 box C/D snoRNP. Since the mgU4-8 fragment derived from scaRNA17 is enriched in the nucleolus and has 234 extensive base pairing with snord16 snoRNA (Fig. 4), and this interaction may inhibit snord16 interaction with 235 18S rRNA, we predicted that scaRNA17 reduction (which would lead to the reduction of mgU4-8) would 236 increase the methylation level of A484 within 18S rRNA. This is exactly what we have observed (Fig. 5). RNA 237 obtained from cells in which scaRNA17 was reduced had consistently more (1.8-fold) A484 methylation 238 compared to that observed in RNA from control siRNA treated cells (n = 3, P<0.05). In contrast, 2′-O-239 methylation levels were essentially the same at the other two known 2′-O-methylation sites (Am436 and 240 Am468), which are not guided by snord16 or targeted by mgU4-8 (derived from scaRNA17). We also observed 241 an increase in the pause site (denoting increased methylation) of the A484 site in scaRNA17-reduced RNA 242 using a primer labeled with digoxigenin (Fig. 5B). These results thus argue that the processed fragment 243 derived from scaRNA17 can form a regulatory RNP that downregulates the methylation of the A484 site in 18S 244 rRNA. 245 246

Disruption of snord16/18S rRNA interaction by a fragment of scaRNA17 247
Previous work has shown that methylation of rRNA by some box C/D snoRNPs is facilitated by "extra base 248 pairings" between the snoRNA and target rRNA (van Nues et al., 2011). Most of these extra base pairings are 249 the result of loops within the snoRNA, allowing for additional snoRNA:rRNA interactions. The interaction 250 between snord16 snoRNA with 18S rRNA is an example of an association that contains extra base pairings 251 ( Fig. 6A). An additional 10 base pairings between snord16 snoRNA and 18S rRNA is made possible by a loop 252 of snord16, between nucleotide 24 and 48. In so doing, it is expected that the methylation of A484 of 18S rRNA 253 is increased as a consequence of these extra base pairings compared to the level of methylation if only 254 nucleotides (nt) 14-24 of snord16 base paired with 18S rRNA. Very interestingly, the scaRNA17-derived 255 nucleolar fragment mgU4-8 interacts with snord16 via nucleotides present in the snord16 loop region (Fig. 6B). 256 It is possible, therefore, that the association of mgU4-8 with the looped region of snord16 disrupts the 257 interaction between snord16 with 18S rRNA, resulting in a decrease in the level of 18S rRNA A484 258 methylation. Given this regulatory activity, we propose that the mgU4-8 domain generated by processing of 259 scaRNA17 be renamed regulatory RNP17 (regRNP17). 260

261
To begin an analysis into the mechanism by which regRNP17 imparts a regulatory effect upon the modification 262 of rRNA, we conducted in vitro RNase protection assays using fragments of 18S rRNA, U16 snoRNA and 263 scaRNA17. The 18S rRNA fragment is 31 nt in length, contains the A484 site and is 3' end labeled with 264 digoxigenin (Dig). The snord16 fragment is 45 nt in length and encompasses nucleotides 14-57 (Fig. 6). The 265 regRNP17 fragment is 33 nt in length and contains the bases that interact with snord16. The basis for RNase 266 protection is that interacting RNAs will not be subjected to degradation by RNase A/T1, which preferentially 267 cleaves single stranded RNA. Consequently, we expect that the interaction of the snord16 fragment with the 268 18S fragment will result in a protected fragment approximately 22 nt in length. The addition of the regRNP17 269 fragment to the 18S/snord16 mixture, however, is predicted to disrupt the interaction between snord16 and 270 18S, decreasing the amount of Dig-labeled 18S fragment that is protected from RNase A/T1 degradation. This 271 is what we have observed (Fig. 6C). In a reaction containing just the Dig-labeled 18S rRNA fragment incubated 272 without RNase A/T1 (lanes 1 and 6), a 31 nt band is detected. This band is digested upon incubation with 273 RNase A/T1 (lane 3). The addition of snord16 to the reaction with the 18S rRNA results in a protected fragment 274 (lane 4), indicating that the interaction between snord16 and 18S precludes RNase A/T1 from fully digesting 275 the 18S RNA. In a reaction containing all three fragments (18S, snord16 and regRNP17), however, the amount 276 of 18S protected fragment is decreased (lane 5), suggesting that the regRNP17 disrupts the interaction 277 between snord16 and 18S rRNA. A reaction containing regRNP17 and 18S does not result in a protected 278 fragment (lane 2), demonstrating that snord16 is required to generate the protected fragment. These findings 279 strongly suggest that regRNP17 influences A484 methylation levels by altering the interaction between 280 snord16 and 18S rRNA. 281 282

A novel RNP: the regulatory RNP 284
In 2004 it was found that scaRNA 2, 9 and 17 are internally processed, resulting in the generation of stable 285 nucleolus-enriched fragments (Tycowski et al., 2004). Base pairing between the full-length scaRNA 2, 9 and 17 286 with U2, U4 and U12 snRNAs strongly suggest that scaRNA 2, 9 and 17 guide the 2'-O-methylation of specific 287 sites within these snRNAs. Like other scaRNAs, therefore, full-length scaRNA 2, 9 and 17 contribute towards 288 the biogenesis of snRNPs, which are crucial parts of the spliceosome necessary for pre-mRNA splicing. The 289 function of the stable, nucleolus-enriched fragments derived from these three scaRNAs, however, remains 290 murky. Specifically, the function of mgU2-61 (derived from scaRNA2), mgU2-19 and mgU2-30 (both derived 291 from scaRNA9) and mgU4-8 (derived from scaRNA17) (Fig. 1) is not evident considering that their putative 292 targets (U2 and U4 snRNA) do not have a clear nucleolar pathway (Kiss, 2004). By conducting BLAST 293 searches of the unpaired loops of these processed fragments against a snoRNA database (Lestrade et al., 294 2006), we found that each of the four fragments derived from scaRNA 2, 9 and 17 can base pair with snoRNA 295 (Fig. 4). This led us to speculate that the processed fragments do not form RNPs which directly methylate 296 targets, but instead may regulate the activity of snoRNPs responsible for 2'-O-methylation and 297 pseudouridylation (Fig. 4, Table 1). Experimental evidence in support of this hypothesis is shown in Figs. 5 and 298 6, and argues for the possibility that the nucleolar-enriched processed fragments of scaRNA 2, 9 and 17 form 299 novel RNP complexes that we term regulatory RNPs (regRNPs). We propose that regRNPs affect the 300 modification of rRNA and U6 snRNA (which traffics though the nucleolus) by interactions with the snoRNA 301 component of snoRNPs. It is highly unlikely that there are only four regRNPs present (one each from scaRNA 302

RegRNPs and ribosome heterogeneity 307
An exciting emerging concept is that of ribosome heterogeneity, leading to specialized ribosomes. This 308 concept is based on the realization that ribosomes are not all the same, all the time, but can vary in response 309 to different physiological or pathological situations (Lafontaine, 2015). One major contributor to ribosome 310 heterogeneity is rRNA modification. Since each ribosome in human contains around 100 each of 311 pseudouridines and ribose methylations, there is a vast potential for ribosome specialization in regulating the 312 level of these modifications in rRNA (Lafontaine, 2015). Very significantly, three methylation sites within rRNA 313 that we predict are subject to regRNP 2, 9 and 17 control (G3923 in 28S, U1804 and A484 in 18S) have been 314 shown to be differentially modified in endogenous ribosomes (Krogh et al., 2016) (Incarnato et al., 2017. In 315 fact, all three of these sites show upwards of 25% variability (Krogh et al., 2016) (Incarnato et al., 2017. 316 Interestingly, comparison of the U1804 and G3923 methylation levels in HeLa and HCT116 cells shows 317 significant differences between these lines (Krogh et al., 2016). These differences were not correlated with 318 altered levels of snoRNPs, indicating that another factor besides snoRNP availability is responsible for the 319 differential modification of rRNA (Krogh et al., 2016). Our data presented in Figs. 5 and 6 showing that 320 regRNP17 impacts the level of A484 methylation in 18S rRNA supports the hypothesis that ribosomes may be 321 optimized for a given cell type or physiological/pathological situation by controlling the level of rRNA 322 modifications using regRNPs. 323

18S rRNA/snord16/regRNP17 interactions in context with other snoRNAs 325
Given that rRNA has extensive secondary structure as well as numerous modifications, it is not surprising that 326 the accessibility and regulation of snoRNP activity must be subject to some type of control. An example of 327 where one would expect some type of ordered snoRNP activity is shown in Fig. 7A, which displays the 328 snoRNAs that interact with 18S rRNA in the region of A484. In particular, snord11, snord56 and snord70 have 329 overlapping binding sites on the 18S rRNA. Moreover, a 9 nt region of 18S rRNA starting at T514 (underlined 330 in Fig. 7A) is found exactly in snord16. In fact, snord16 nt 76-86 can complementary base pair with snord56 at 331 nt 29-19. In so doing, interactions between snord16 and snord56 may contribute towards the regulation of 332 C517 methylation. Since regRNP17 interacts with the looped region of snord16, and this association appears 333 to disrupt the interaction of snord16 with 18S rRNA (Fig. 6), it is further possible that regRNP17 may also be 334 involved in the regulation of snord56-mediated C517 methylation. Snord16 has many RNA interactions that 335 may regulate its functions, and, conversely, allow snord16 to regulate the activities of the RNAs it interacts with 336 ( Fig. 7B). In fact, 65% of snord16 can base pair with other RNAs (66 nt out of 101 nt). A key regulator of these 337 snord16 interactions may be regRNP17, which, by binding the looped region of snord16, could greatly 338 influence snord16 associations with 18S rRNA and other snoRNAs. 339 repeat regions in scaRNA 2 and 9 ( Fig. 1) impacts their processing (Enwerem et al., 2015;Poole et al., 2016), 344 and here identify another cis element: the 28 nt leader sequence of scaRNA9 (Fig. 2). We have shown that the 345 28 nt leader sequence of scaRNA9 is critical for its stability and influences its protein interactions. Since 346 scaRNA9 is processed at its 5' and 3' end by exonucleases, any nucleic acid sequence not protected by a 347 protein complex should be processed. This indicates that there are protein components which bind to the 348 scaRNA9 leader sequence that block it from being processed. We show here that coilp1 could be one of these 349 interactors, as coilp1's interaction with scaRNA9 ∆ leader is drastically reduced compared to WT scaRNA9 350 ( Fig. 2A). Furthermore, the increased recovery of coilin (2.2-fold) using the scaRNA9 ∆ leader bait in the RNA 351 pulldown compared to the amount of coilin obtained with WT scaRNA9 may indicate that coilp1 negatively 352 regulates the amount of coilin in the binding complex. Not surprisingly, only full-length scaRNA9 and the 5'-353 most processed fragment (mgU2-19) levels are deleteriously impacted by deletion of the leader sequence (Fig.  354   2B). An interesting future direction involves deciphering the mechanism which governs the amount of scaRNA 355 2, 9 and 17 that is processed into regRNPs. 356 357

Coilp1 contributes to the availability of regRNPs 358
The data presented here further characterizes the recently discovered pseudogene encoded protein coilp1, 359 and implicates it in the biogenesis of the newly described regRNPs. At this point, the exact function of coilp1 in 360 this process is not clear. ScaRNA 9 and 17 showed an increase in processing when coilp1 was overexpressed, 361 and a decrease in processing when coilp1 expression is reduced (Fig. 3). However, we did not observe any 362 change in processing for scaRNA2. ScaRNAs 2 and 17 are both independently transcribed by RNA 363 polymerase II (Gerard et al., 2010;Tycowski et al., 2004). Possibly due to non-canonical box C and D 364 sequences, mgU2-25 (scaRNA2) and mgU12-22 (scaRNA17) are inherently unstable, and do not accumulate 365 within the cell once the scaRNAs are internally processed. ScaRNA9 is contained within a host gene, and, as 366 such, must be spliced and processed at its 5' and 3' ends. ScaRNA9's 5' domain, mgU2-19, is capable of 367 accumulating, albeit a lower levels than its 3' domain, mgU2-30. Indeed, work previously done by our lab has 368 shown that scaRNA2 is less efficiently processed by coilin than scaRNA9 (Enwerem et al., 2015). Given that 369 coilp1 shares sequence homology with coilin, coilp1's differential contribution to the availability of regRNPs processing. We further predict that coilin negatively regulates the activity of coilp1 but SMN and WRAP53 374 promote this activity (Fig. 8). Another possibility is that coilp1 may chaperone regRNPs to the nucleolus. Given 375 that Nopp140 is a snoRNP chaperone, interacts with coilin and localizes to the CB and nucleolus (Isaac et al., 376 1998), it is also well positioned to serve as an important factor in the formation of regulatory RNPs. Indeed, 377 deletion of Nopp140 in fly reduces rRNA methylation (He et al., 2015), supporting a role for this protein in the In conclusion, our work has provided a possible function for the processed fragments derived from scaRNA 2, 381 9 and 17 as regRNPs. Although not an exact parallel to what we are proposing, the Stamm group has shown 382 that box C/D snoRNAs can have dual functions in rRNA modification and alternative pre-mRNA splicing 383 (Falaleeva et al., 2016). Thus there is some degree of precedent for a guide RNA having additional functions. 384 Our concept of a regulatory RNP, however, is novel. These regRNPs are predicted to contribute to the 385 heterogeneity of RNA, leading to specialized ribosomes. These findings further add to the complexity of 386 ribosome biogenesis. When ribosome biogenesis is negatively affected, it results in the ribosomopathy disease 387 state, which is characterized by any dysfunction in any of the hundreds of components which facilitate 388 ribosome formation (Wang et al., 2015;Zhou et al., 2015). Interestingly, recent work in human cells has shown 389 that p53 down regulates fibrillarin levels, and in cancer cells lacking functional p53 the level of rRNA 390 methylation is increased (Marcel et al., 2013). This increase in rRNA methylation results in ribosomes with a 391 lower fidelity (i.e. stop codons are bypassed) and a greater likelihood to initiate translation through internal 392 ribosome entry sequences (IRESs) (Belin et al., 2009;Marcel et al., 2015;Marcel et al., 2013). As a 393 consequence of these changes in rRNA methylation, the translation of messages with IRESs is increased; 394 such as those whose products promote tumor development (IGF-1R, c-myc, VEGF-A and FGF1) (Marcel et al., 395 2015). Clearly, therefore, the regulation of modifications within rRNA is of great importance. With our 396 identification of regRNPs, we demonstrate that one non-coding RNA can, based on its localization and 397 processing, influence both the splicing and translation machinery. Our current efforts seek to experimentally 398 verify additional regRNPs. 399 400

IF and RNA FISH 480
Cells were fixed in 4% PFA for 10 minutes and permeabilized in 0.5% Triton for 5 minutes. Slides were then 481 rinsed 3 times in 1x PBS. Slides were blocked using 10% normal goat serum (NGS) for 30 minutes at 37°C. 482 Slides were then probed with 10% NGS containing 1:200 coilin antibody for 30 minutes at 37°C. Slides were 483 then washed 3 times for 5 minutes, and incubated with 1:600 AlexaFlour-488 Goat anti-rabbit secondary 484 antibody in 10% NGS for 30 minutes at 37°C. Slides were then washed 5 times for 5 minutes in 1x PBS. 485 Following immunostaining, cells were post-fixed in 4% PFA for 10 minutes. Slides were then washed twice in 486 2x SSC for 5 minutes followed by two washes in 70% ethanol. Cells were then dehydrated by a series of 487 ethanol washes (80%, 95%, and 100% ethanol) for 3 minutes each, and slides were allowed to air dry for 5 488 minutes. 100 μl of probe solution (10% dextran sulfate, 2 mM VRC, 0.02% BSA, 40 μg E. coli tRNA, 2x SSC, 489 50% formamide, and 30 ng probe) was added to the slides and incubated overnight at 37°C. Probe used for 490 detection of scaRNA9 (Integrated DNA Technologies, Coralville, Iowa, USA): 5'-491 /5Alex594N/TCATAGTTACAAAGATCAGTAGTAAAACCTTTTCATCATTG-3'. Slides were then washed 3 492 times for 5 minutes each in 2x SSC, and DAPI stained for 5 minutes. Slides were destained in 2x SSC for 5 493 minutes prior to mounting. 494

BLAST searches 496
Using a publicly available database (Lestrade and Weber, 2006), BLAST were carried out by using sequences 497 from scaRNAs 2, 9, and 17 that were previously reported to be contained within loop structures (Tycowski et 498 al., 2004). Antisense hits were then compared to reported structures of predicted targets, where applicable, to 499 validate the accessibility of the complementary region.   with snoRNA. Also shown is the potential for the 3′ loop of mgU2-30 (top right) to base pair with 28S rRNA, 675 possibly blocking ACA3 binding to this site, as well as serve as a binding site for U20 snoRNA. The 5′ loop of 676 scaRNA17-derived mgU4-8 (bottom right), which contains a sequence found in U6 snRNA, might serve as a 677 binding sink for U94 and HBII-166 (not shown) snoRNAs, thereby altering C60 and C62 methylation of U6 678 snRNA. We propose that these fragments derived from scaRNA 2, 9 and 17 form regulatory RNPs (regRNPs) 679 that influence the modification of sites within 18S rRNA, 28S rRNA and the nucleolar trafficked U6 snRNA 680 (Table 1) Primer extension was conducted using only low (2.5 uM) dNTP and the primer is 5' Dig labeled. Quantification 687 demonstrates that the A484 pause signal is 1.8-fold more abundant in the scaRNA17 knockdown condition 688 compared to control knockdown (n = 3 experimental repeats, p < 0.05).   a n t i -Co i l i n a n t i -F i b r i l l a r i n a n t i -S MN a n t i -T u b u l i n L a n e : 2 I n p u t B e a d s Un t r a n s f e c t e d B e GF P : F D E C 2 L a n e : 1 2 L a n e