Towards an understanding of regulating Cajal body activity by protein modification

ABSTRACT

The biogenesis of small nuclear ribonucleoproteins (snRNPs), small Cajal body-specific RNPs (scaRNPs), small nucleolar RNPs (snoRNPs) and the telomerase RNP involves Cajal bodies (CBs). Although many components enriched in the CB contain post-translational modifications (PTMs), little is known about how these modifications impact individual protein function within the CB and, in concert with other modified factors, collectively regulate CB activity. Since all components of the CB also reside in other cellular locations, it is also important that we understand how PTMs affect the subcellular localization of CB components. In this review, we explore the current knowledge of PTMs on the activity of proteins known to enrich in CBs in an effort to highlight current progress as well as illuminate paths for future investigation.

KEYWORDS:

• Cajal body
• coilin
• phosphorylation
• post-translational modification
• SMN
• telomerase
• WRAP53

Introduction

There are different types of ribonucleoproteins (RNPs), which are comprised of non-coding RNA and associated proteins. RNPs take part in fundamental cellular activities such as translation and pre-mRNA splicing. While the nucleolus serves as the initial hub for the generation of RNPs that take part in translation (ribosomes), the subnuclear domain known as the Cajal body (CB) contributes to the biogenesis of small nuclear RNPs (snRNPs), which are involved in pre-mRNA splicing.¹ In addition to snRNPs, CBs also contain another type of RNP: small Cajal body-specific RNPs (scaRNPs). The presence of scaRNPs in the CB serves to modify the snRNA component of snRNPs. The modification of snRNAs allows for the full-functionality of the snRNP. Base paring between the scaRNA of the scaRNP with the snRNA of the snRNP guides the modification of target sites within the snRNA by ribose methylation (for box C/D scaRNPs) or pseudouridylation (for box H/ACA scaRNPs).² CBs also take part in the biogenesis of small nucleolar RNPs (snoRNPs). Like scaRNPs, there are 2 classes of snoRNPs: box H/ACA and box C/D. The target sites for snoRNPs are rRNA and the RNA polymerase III transcribed, nucleolar-trafficked, U6 snRNA.² In addition to snRNPs, scaRNPs and snoRNPs, CBs also contain the RNP responsible for the formation of telomeres, telomerase.³ The RNA component of telomerase, known as TERC or TR (hTR if referring to that in human) is an H/ACA class of small non-coding RNA and is considered to be a scaRNA. Mature telomerase contains hTR in a complex with core proteins and the telomerase reverse transcriptase (hTERT).⁴ The CB is thought to be the assembly point for the incorporation of hTERT into the nascent telomerase complex. Along with snRNPs, scaRNPs and telomerase, the CB is enriched for many proteins. All of these proteins localize to other cellular locales in addition to the CB. The functions of these proteins in the CB is diverse, but collectively are thought to contribute to the RNP biogenesis mission of the CB. Like other cellular processes, RNP biogenesis, and thus CB activity, is regulated but an understanding of this regulation is far from complete. For example, many of the proteins enriched within the CB are post-translationally modified, but the functional consequences of these modifications are, for the most part, unclear. In this review, we will summarize what is known about the regulation of CB activity by exploring the current literature in regards to the post-translational modification of proteins enriched in the CB. Specifically, we will detail the (putative) function of proteins within the CB and describe any evidence as to if post-translational modification impacts these functions. The components that conduct these modifications, if they are known, will also be examined. We will also discuss the signals that govern these modifications in order to understand the local and global controls that influence the RNP biogenesis activity of CBs.

We will begin our analysis with 3 major components of the CB: coilin, SMN and WRAP53. We will then discuss other proteins that accumulate in CBs and contribute to snRNP, telomerase, scaRNP, or snoRNP formation. Table 1 lists the proteins we will examine, along with known modifications of these proteins. The modification information was obtained from UniProt and PhosphoSitePlus websites and includes evidence obtained from large-scale mass spectrometric approaches and directed studies. As stated above, although all of the proteins listed in Table 1 localize to other regions of the cell in addition to CBs, we will focus our discussion on the activity of these proteins in CBs. However, we will also provide information about non-CB functions when appropriate to give an indication as to how post-translational modification influences the function of a given protein.

Table 1. Modified CB proteins^*.

Name	Uniprot ID	Phos	Meth	Acet	Sumo	Ubiq
Coilin	P38432	+	+	+	+	+
CRM1	O14980	+	−	+	−	+
DAXX	Q9UER7	+	−	+	+	+
Dyskerin	O60832	+	−	+	+	+
Fam118B	Q9BPY3	+	−	+	−	+
Fibrillarin	P22087	+	+	+	−	+
GAR1	Q9NY12	+	+	+	−	+
Nopp140	Q14978	+	+	+	+	+
PA28γ	P61289	+	+	+	+	+
PHAX	Q9H814	+	+	+	−	−
SART3	Q15020	+	+	+	−	+
SmD1	P62314	+	+	−	−	+
SMN	Q16637	+	+	+	+	+
TERT	O14746	+	−	+	−	+
TGS1	Q96RS0	+	−	−	−	+
TOE1	Q96GM8	+	−	−	−	+
USPL1	Q5W0Q7	+	−	−	−	−
WRAP53	Q9BUR4	+	−	−	−	+

* Phos = phosphorylation, meth = methylation, acet = acetylation, sumo = sumoylation, ubiq = ubiquitination. Modification information obtained from UniProt and PhosphoSitePlus websites.

Coilin

What does coilin do in the CB?

Human coilin (UniProt # P38432) contains numerous modifications (Table 1) and is considered the CB marker protein.^5,6 Given coilin's interaction with other proteins and RNAs in the CB, it likely plays a structural role in CB formation.^7-13 Reduction or depletion of coilin abolishes canonical CBs.^14,15 Coilin has been shown to have RNA degradation/processing activity.¹⁶ Although the functional consequence of this activity is not fully understood, it is possible that coilin may directly contribute to snRNA, scaRNA and hTR processing and/or incorporation of these RNA into RNPs.

How does modification impact coilin's CB activity?

Phosphorylation

The first report of coilin modification took place in 1993, when Lamond and colleagues observed that coilin is phosphorylated, and this phosphorylation increases during mitosis when CBs disassemble.¹⁷ Work by this same group demonstrated that one potential phosphorylation site within coilin, S202, mis-localizes coilin to the nucleolus when this residue is changed to mimic a constitutively phosphorylated residue (S202D).¹⁸ The Lamond lab also showed that phosphatase inhibition by okadaic acid treatment disrupts CBs and results in the mis-localization of coilin to the nucleolus.¹⁹ This finding clearly indicates that phosphorylation impacts CB formation. Our studies have shown that coilin is a self-interacting protein, and the extent of this self-interaction decreases during mitosis when coilin is hyperphosphorylated and CBs disassemble.⁷ Hyperphosphorylation of coilin, therefore, may trigger CB disassembly. We have also observed that coilin in cell lines with CBs (such as HeLa) is less phosphorylated than that found in some primary cell lines (such as WI-38), which have few CBs.²⁰ In fact, coilin phosphorylation in interphase primary cells is apparently about the same as that found in mitotic HeLa cells. Point mutations of known phosphorylation sites further demonstrate that phosphorylation impacts coilin function, and thus CB activity, as well as proliferation rate.²¹ Specifically, one change in coilin (S184A) results in the mis-localization of coilin to the nucleolus⁷ while another change (S489D) decreases the in vitro RNA processing activity of coilin.²² Besides coilin self-association, phosphorylation also appears to impact coilin association with other proteins and RNA. For example, phosphorylation of the C-terminus of coilin may help disengage the SMN complex from CBs and allow for snRNP accumulation in this subnuclear domain.¹¹ Additionally, the interaction profile of RNAs associated with coilin changes in response to nocodazole, which arrests cells in mitosis and generates hyperphosphorylated coilin.¹² Specifically, nocodazole treatment decreases the level the box C/D scaRNA 2 and 9 in the endogenous coilin immunoprecipitation complex approximately 4-fold compared to that present in untreated cells. Since another study has observed that ectopically expressed coilin-GFP associates with hundreds of non-coding RNAs,¹³ it is likely, therefore, that coilin phosphorylation impacts CB activity by influencing the composition of the CB both in terms of proteins and RNAs. In addition, coilin may play a more direct role in snRNP, scaRNP, snoRNP and telomerase biogenesis than previously appreciated, and this activity is also likely to be regulated by phosphorylation. Recent work has shown that vaccinia-related kinase (VRK1) may play an important role in regulating CB dynamics in a cell cycle dependent manner by phosphorylating coilin (at S184) and protecting coilin from ubiquitination and proteasomal degradation.²³ More studies of this kind will be necessary to understand the role that coilin phosphorylation has on CB integrity and activity.

Methylation

Coilin has been shown to contain symmetrical dimethyl arginine (sDMA) in a region of the protein known as the RG (arginine/glycine) box.^9,24 Coilin interaction with SMN is contingent upon the presence of sDMAs in coilin.^8,9,24 Mutations in the coilin RG-box decrease interaction with SMN and result in coilin foci that do not enrich for SMN. Instead, nuclear SMN in cells expressing coilin with a mutated RG box form foci known as Gems.⁸ In agreement with these results, hypomethylated coilin is correlated with decreased SMN interaction and increased Gem frequency.^9,24,25 Other work has shown that coilin localization is dependent upon its modification: hypomethylated coilin is enriched in the nucleolus while methylated coilin localizes to CBs.²⁶ Interestingly, hypomethylated coilin in the nucleolus is not associated with SMN or snRNPs,²⁶ clearly demonstrating that sDMA modification of coilin can affect its localization and association with other proteins.

Acetylation, sumoylation and ubiquitination

Large-scale mass spectrometric work has shown that coilin is modified by acetylation (at residues K204, K405, K496), sumoylation (K127, K204, K281) and ubiquitination (K476, K496, K555). Unfortunately, the functional consequences of these modifications on coilin activity are unknown. We have previously reported²⁷ that coilin interacts with PIASy, which is a SUMO E3 ligase enriched within another nuclear domain known as the PML body. It is not known if coilin is a substrate for PIASy-mediated sumoylation, but it is known that SUMO-1 can be detected in CBs.²⁸ The interaction between coilin and PIASy may thus provide an avenue for the sumoylation of proteins enriched within the CB.

SMN

What does SMN do in the CB?

Human SMN (UniProt # Q16637) is extensively modified (Table 1) and very well studied, in part because this protein is mutated in most cases of the neurodegenerative disease spinal muscular atrophy (SMA). The best-characterized activity of SMN actually takes place in the cytoplasm, not the CB.^29-34 In the cytoplasm, SMN plays crucial roles in ensuring that Sm proteins assemble properly onto U1, U2, U4 and U5 spliceosomal snRNAs during snRNP biogenesis, and the import of these nascent snRNPs to the nucleus, where they first target to CBs.^32,35-38 The exact function of SMN in the CB is unclear, but likely involves more than being a delivery vehicle for newly assembled, yet not mature, snRNPs to the CB. Before snRNPs are ready for splicing, additional snRNP-specific proteins are assembled onto the RNA. Additionally, the snRNA component of the snRNP contains modifications (ribose methylations and pseudouridylations) that are required for full functionality of the snRNP.³⁹ These snRNA modifications take place in the CB via scaRNPs.² Since scaRNPs are similar to snRNPs in that they are comprised of an RNA (scaRNA) and associated proteins, it is possible the SMN in the CB helps assemble scaRNPs in a similar manner as that found for snRNPs. Furthermore, SMN in the CB may also contribute toward telomerase holoenzyme assembly, which is known to take place in the CB. Another idea as to what SMN may be doing in the CB involves the recycling or regeneration of snRNPs after a splicing reaction takes place.^10,40 Reduction of SMN by RNAi disrupts CBs,¹⁵ clearly indicating that this protein is vital for proper CB formation and composition. Regarding the function of SMN that when disrupted causes SMA, there are some opposing ideas as to if alterations in the snRNP biogenesis-promoting role of SMN is the main culprit.^41,42,43 The disruption of other functions for SMN, such as those involved in muscle formation and neuromuscular junctions, as well as in the afferent nerves, may also contribute to the SMA phenotype.^44,45 Recent work has shown that SMN takes part in the formation of messenger RNPs which are comprised of mRNA and mRNA binding proteins.⁴⁶ Therefore, reduction in the level of functional SMN may cause disruptions in many cellular processes, some of which take place in the CB, but all of which result in SMA.

How does modification impact SMN's CB activity?

Phosphorylation

Somewhat surprisingly, the first report of SMN post-translational modification (phosphorylation) was published in 2004,⁴⁷ 9 y after the identification of SMN, which took place in 1995.⁴⁸ A functional consequence of SMN phosphorylation was soon identified, and centered upon the differential phosphorylation of cytoplasmic and nuclear SMN.⁴⁹ Briefly, SMN was shown to be phosphorylated on serines 28 and 31, and these modifications were more enriched in cytoplasmic SMN compared to nuclear SMN. This group then examined the activity of cytoplasmic SMN versus nuclear SMN using an snRNP assembly assay and observed that cytoplasmic SMN, which is phosphorylated on S28 and S31, is more active than nuclear SMN which is hypophosphorylated relative to cytoplasmic SMN. Hence phosphorylation of cytoplasmic SMN is an activator of the SMN complex.⁴⁹ Other studies have shown that dephosphorylation of SMN, which occurs via the action of the PPM1G phosphatase, is necessary for SMN localization to CBs.⁵⁰ Another phosphatase, PP1γ, likely also acts upon SMN (or other members of the SMN complex).⁵¹ It is not known if PP1γ targets the same residues as PPM1G, but reduction of these 2 phosphatases results in different affects on SMN localization: PPM1G reduction decreases SMN enrichment in CBs⁵⁰ while PP1γ reduction increases SMN CB localization.⁵¹ Other work, however, has shown that, like PPM1G, PP1γ reduction decreases SMN localization in CBs.⁵² Phospho-null mutants of SMN (S28A/S31A) fail to show large cytoplasmic aggregates like those formed by WT SMN when overexpressed in cell lines.⁵³ This finding suggests that the phosphorylation of SMN may promote self-association and SMN complex formation. By forming these complexes, it is hypothesized that phosphorylated SMN is more stable and less subjected to the ubiquitin proteasome system.⁵⁴ Additional insight into SMN phosphorylation has come from work conducted by Gruss and colleagues, who have demonstrated that 7 out of the 9 components of the SMN complex are phosphorylated.⁵⁵ SMN contains serine and threonine phosphorylations, and mutation of these sites disrupts localization to CBs and association with coilin.⁵⁵ Three tyrosine residues have also been shown to be phosphorylated in SMN, and simultaneous mutation of these 3 tyrosine sites to phenylalanine (Y109F, Y127F, Y130F; demarcated in paper as SMN_Y) greatly reduces enrichment in CBs and nearly abolishes interaction with coilin.⁵⁵ Collectively, this work clearly demonstrates that SMN, and the SMN complex, is regulated by phosphorylation and future work clearly elucidating the kinases and phosphatases responsible for this regulation is of great importance. This work has already started in earnest.⁵²

Methylation

SMN is methylated on arginine 204.⁵⁶ The functional consequence of this modification on SMN activity is unknown. SMN has been shown to have preferential interaction with proteins containing symmetrical dimethyl arginine (sDMA). Proteins with sDMAs that have been shown to increase interaction with SMN include coilin,^9,24 SmD1 and SmD3,⁵⁷ SmB/B',^57,58 Lsm4,⁵⁸ Epstein-Barr Virus Nuclear Antigen 2 (EBNA2)⁵⁹; FGF-2²³⁶⁰ and the C-terminal domain of RNA polymerase II.⁶¹

Acetylation, sumoylation and ubiquitination

SMN is acetylated at 3 residues (Table 1), but nothing is known about how this modification impacts its activity. Recent work has shown that SMN is modified by SUMO1 at residue K119.⁶² Sumoylated SMN is not enriched in the CB, which contains the desumoylation enzyme USPL1. This suggests that desumoylation of SMN takes place in the CB. This work has also identified a SUMO-interacting motif (SIM)-like motif in close proximity to the sumoylated K119 residue, both of which reside within the Tudor domain. The authors' propose that SMN sumoylation and interaction with proteins via the SIM-like motif are important for CB formation and/or composition.⁶² Regarding ubiquitination, mass spectrometric approaches have shown that SMN is ubiquitinated on residues K41, K51, K184, K186 and K209. The first report of ubiquitinated SMN showed that SMN protein levels could be increased in SMA patient cells upon treatment with the proteasome inhibitor MG132.⁶³ As mentioned above, the impact of ubiquitination and phosphorylation on SMN stability and complex formation was also examined.⁵⁴ The protein UCHL1 (ubiquitin C-terminal hydrolase L1) was found to be increased in SMA patient cells, and overexpression of UCHL1 decreased SMN levels.⁶⁴ Moreover, UCHL1 can ubiquitinate SMN, likely facilitating SMN degradation. Inhibition of UCHL1 activity increased SMN levels, suggesting that UCHL1 could be a therapeutic target for the treatment of SMA. However, another report has found that inhibition of UCHL1 actually exacerbates SMA in a mouse model.⁶⁵ The effect of monoubiquitination of SMN has also been explored.⁶⁶ The ubiquitin ligase Itch interacts with SMN and monoubiquitinates it at residues K179 and K209. Monoubiquitination of SMN affects its cellular localization. Ubiquitination-incompetent SMN mutant is capable of forming SMN oligomers and the SMN complex, but this mutant accumulates primarily in many nuclear foci with some nucleolar accumulations. In contrast, a SMN mutant fused with ubiquitin to mimic monoubiquitination primarily localizes to the cytoplasm. Nuclear accumulations with this mutant are present but appear to be in nucleoli. Hence the monoubiquitination of SMN may be a signal for nuclear export or inhibition of nuclear import. Both SMN mutants (ubiquitination-incompetent and monoubiquitination mimic) disrupt canonical CBs.⁶⁶ More broadly speaking, the disruption of ubiquitin homeostasis, via SMN interaction with the ubiquitin-like modifier activating enzyme 1 (UBA1), may, along with alterations in the β-catenin signaling, promote SMA.^65,67

WRAP53

What does WRAP53 do in the CB?

Unlike coilin and SMN, Human WRAP53 (UniProt # Q9BUR4) is not extensively modified, containing only phosphorylated and ubiquitinated residues (Table 1). WRAP53 (also known as TCAB1 or WDR79) was first identified in 2009 as the protein that binds the CAB motif present in box H/ACA scaRNAs and telomerase RNA, and targets these RNAs to the CB.^68,69 Like SMN, WRAP53 localizes to the cytoplasm and CB. Reduction of WRAP53 by RNAi disrupts CBs and mis-localizes coilin to the nucleolus,⁷⁰ demonstrating that this protein plays an important role in CB integrity and nuclear organization. Mutations in WRAP53 are found in patients with dyskeratosis congenita, a disease caused by insufficient levels of telomerase.⁷¹ These findings show that WRAP53 is a crucial player in the trafficking and formation of the telomerase RNP. Given its interaction with the CAB motif present in box H/ACA scaRNAs, WRAP53 is also an important factor in that class of scaRNP biogenesis. It is important to note that human box C/D scaRNAs lack a CAB motif and WRAP53 interaction with these RNAs is far lower that that observed for box H/ACA scaRNAs.⁷² This finding brings into question how box C/D scaRNAs are targeted to the CB for incorporation into a scaRNP. For some box C/D scaRNAs (those that are intron encoded), it appears that a motif known as the G.U/U.G wobble stem associates with WRAP53, allowing for the accumulation of these scaRNAs in the CB.⁷² We have provided evidence that coilin, which highly associates with the box C/D scaRNAs 2, 9 and 17, may also play a role in the targeting of these scaRNAs to the CB.¹² Apart from the role of WRAP53 in CB-centered telomerase and scaRNP biogenesis, other work has shown that this protein is an essential player in the DNA double strand break repair pathway.⁷³

How does modification impact WRAP53s CB activity?

Phosphorylation and ubiquitination

No studies have specifically investigated how post-translational modification of WRAP53 influences its activities. Clearly, therefore, this is an open area of investigation ripe for the picking. WRAP53 contains 20 residues that are phosphorylated (Table 1), 15 of which reside in the N-terminal 150 amino acids of the protein. It is possible that WRAP53 interactions through its WD40 domain are regulated by phosphorylation of the N-terminus. Additionally, phosphorylation of WRAP53 may regulate its association with scaRNAs and telomerase RNA as well as its trafficking and localization to CBs. Regarding ubiquitination, WRAP53 residues K176 and K388 are modified in this manner. Like phosphorylation, however, it is not known how these modifications impact WRAP53 activity, localization or stability.

CRM1

What does CRM1 do in the CB?

CRM1 (UniProt # O14980) localizes to the nucleoplasm and CB.^74,75 CRM1 is part of the export complex that takes newly transcribed snRNAs from the nucleus to the cytoplasm for further maturation steps.⁷⁶ Before the journey to the cytoplasm, the export complex likely transits through the CB for remodeling steps which include the removal of the integrator complex responsible for the 3´-end processing of snRNA.^1,77-79 CRM1 also plays a role in the targeting of snoRNPs to the nucleolus, which, like snRNPs, also appear to transit through the CB during their biogenesis.^75,80-82 Inhibition of CRM1, and hence export of nascent spliceosomal snRNAs to the cytoplasm, disrupts CB composition and dynamics,^75,83 thereby reinforcing the idea that the CB is a dynamic nuclear structure responsive to the RNP demands of the cell. Moreover, selective inhibitors of nuclear export (SINE) that target CRM1 are current being developed to treat triple negative breast cancer,⁸⁴ implicating CRM1 RNP biogenesis activity as a potential target to treat cancer.

How does modification impact CRM1s CB activity?

Phosphorylation, acetylation and ubiquitination

CRM1 is modified by phosphorylation, acetylation and ubiquitination (Table 1). Given its important role in nuclear export, it is not surprising that many reports regarding the impact of PTMs do not necessarily directly involve modifications of CRM1. Rather, most of these reports discuss how the modification of cargo proteins influences their export by the CRM1 complex. One exception is the finding that CRM1 S391 phosphorylation by CDK1/cyclinB during mitosis controls CRM1 interactions, localization and mitotic functions.⁸⁵ Considering that CRM1 is phosphorylated on 20 residues, including S391, it will be of great interest to decipher how these PTMs regulate the RNP biogenesis-promoting activities of the CRM1 complex.

DAXX

What does DAXX do in the CB?

DAXX (UniProt # Q9UER7) is a well-studied transcriptional co-repressor protein that is involved in many pathways and interacts with many factors.⁸⁶ DAXX localizes to the cytoplasm and the nucleus, and in the nucleus DAXX is sequestered into PML bodies in order to regulate its activity.⁸⁶ Recent work has shown that DAXX can localize to CB in a cell cycle-dependent manner, with maximal enrichment in CBs taking place in early to mid S phase.⁸⁷ Since telomerase reverse transcriptase (TERT) accumulation with CBs also peaks during this phase of the cell cycle, coinciding with the assembly of the telomerase holoenzyme,^88-90 it is possible that the presence of DAXX in CBs promotes telomerase assembly (via interactions with telomerase subunits) and targeting of telomerase to telomeres.⁸⁷

How does modification impact DAXX's CB activity?

Phosphorylation, acetylation, sumoylation and ubiquitination

DAXX is extensively modified (Table 1), and the impact of these modifications on non-CB activities of DAXX has been well investigated.⁹¹ Since the report about DAXX's involvement in telomerase biogenesis is relatively recent (2015), it is not surprising that the impact of post-translational modifications on this facet of DAXX function has not been examined. However, given that DAXX is a protein of 740 amino acids that contains 50 phosphorylated residues, it is likely that the telomerase-promoting activity of DAXX is regulated by phosphorylation.

Dyskerin

What does dyskerin do in the CB?

Dyskerin (UniProt # O60832) localizes to the nucleolus and CB.⁹² Mutations in dyskerin cause X-linked dyskeratosis congenita,^93,94 demonstrating the importance of this protein in telomerase biogenesis. The co-transcriptional association of dyskerin and the proteins NHP2 and NOP10 with hTR is an early step in telomerase biogenesis.³ This early complex also contains NAF1, which is replaced by GAR1 in the mature complex. Dyskerin is also the enzymatic component of box H/ACA small nucleolar RNPs (snoRNPs) and scaRNPs, which are responsible for the pseudouridylation of rRNA and snRNA, respectively.⁹⁵ Because hTR lacks any complementary of its hairpin pockets to any known target sequence, telomerase lacks the dyskerin-mediated pseudouridylase activity found in other H/ACA RNPs.³ We have recently identified novel interactions between coilin and SMN with the telomerase components dyskerin and NAF1.⁹⁶ We have also found that alterations in the level of coilin and SMN impact the composition of the dyskerin complex and alter the association of dyskerin with hTR.⁹⁶ Hence dyskerin incorporation into both telomerase and box H/ACA scaRNPs/snoRNPs may be regulated by associations with CB proteins.

How does modification impact dyskerin's CB activity?

Phosphorylation and acetylation

Although dyskerin contains 17 phosphorylated sites (Table 1), there is no published information about how this modification impacts telomerase, scaRNP or snoRNP biogenesis. Nor is there information about if phosphorylation regulates the pseudouridylase activity of dyskerin. In mouse dyskerin, S121 is phosphorylated (Table 1, PhosphoSitePlus). Interestingly, this site is conserved in human dyskerin and a mutation of this residue (S121G) is associated with dyskeratosis congenita (DC).⁹⁷ It is possible, therefore, that some cases of DC are the result of altered dyskerin activity due to insufficient phosphorylation. In terms of acetylation, the influence of this modification on the activity of dyskerin activity is unknown.

Sumoylation and ubiquitination

Many mutations in dyskerin associated with DC are clustered in 2 regions: amino acid 31-72 and amino acid 314-420.⁹⁸ The 31-72 region contains a sumoylated residue (K39), that is mutated (K39E) in DC.⁹⁹ The 31-72 region also contains a hydrophobic cluster sumoylation motif that encompasses K39, and each residue of this motif is mutated in DC.⁹⁹ This group also found that mutation of K39 that prevents its sumoylation (K39R) results in decreased hTR levels as well as reduced telomerase activity and telomere maintenance.⁹⁹ The sumoylation of K39 was also shown to increase dyskerin protein stability, while the K39R mutation did not show this phenotype.⁹⁹ Since K39 is also likely to be ubiquitinated (Table 1),^100,101 modification of this residue by sumoylation or ubiquitination most likely represents a crucial regulatory control that impacts overall dyskerin levels in the cell.

Fam118B

What does Fam118B do in the CB?

Fam118B (UniProt # Q9BPY3) was identified in a screen for coilin interacting proteins.¹⁰² Alteration of Fam118B expression levels (either by reduction or overexpression) disrupts canonical CB structure and composition. Depletion of Fam118B was also shown to decrease splicing capacity, inhibit cell proliferation and reduce the amount of symmetrical dimethyl arginine present in SmD1.¹⁰² Consequently, Fam118B is an important player not only in the formation of CBs, but also in the interaction of SMN with Sm proteins which is required for snRNP biogenesis.

How does modification impact Fam118B's CB activity?

Phosphorylation, acetylation, ubiquitination

Fam118B contains 5 phosphorylation sites, 2 of which take place on tyrosines (Table 1). Currently, there is no information about the role of phosphorylation on Fam118B activity. This protein is also acetylated and ubiquitinated, but, like phosphorylation, the role that these PTMs have on Fam118B activity, and by extension CB activity, is unknown at present.

Fibrillarin

What does fibrillarin do in the CB?

Fibrillarin (UniProt # P22087) localizes to the nucleolus and CB^103,104 and is an essential component of box C/D scaRNPs and snoRNPs which are responsible for the modification of snRNA and rRNA, respectively, by methylation of ribose. Fibrillarin is not only a core component of box C/D scaRNPs and snoRNPs, but is also an RNA binding protein that interacts with scaRNAs and snoRNAs, and is the methyltransferase which modifies target sites within snRNA and rRNA.^105-107 Fibrillarin has also been shown to interact with SMN via a glycine/arginine rich (GAR) domain present within fibrillarin,^108,109 suggesting that SMN may contribute to the formation of box C/D scaRNPs and snoRNPs. In Drosophila, fibrillarin recruitment to CBs has been shown to be contingent upon poly(ADP-ribose) polymerase (PARP1), a well-studied modifier of nuclear proteins by ADP-ribosylation.¹¹⁰

How does modification impact fibrillarin's CB activity?

Phosphorylation, methylation, acetylation and ubiquitination

Nothing is known about how acetylation and ubiquitination regulate fibrillarin incorporation into scaRNPs/snoRNPs, interaction with scaRNAs/snoRNAs, methyltransferase activity, or localization to CBs. Likewise, although it is known that fibrillarin is phosphorylated on 8 residues, the impact of this PTM on fibrillarin function and localization is unknown. The most well-studied PTM of fibrillarin is methylation. Arginines within the GAR domain of fibrillarin are dimethylated,^103,111-114 indicating that this modification may serve to regulate the interaction between fibrillarin and SMN. It should be pointed out that the type of methylation present in fibrillarin, asymmetric dimethylarginine (aDMA),¹¹⁵ is more common than the symmetric dimethylarginine (sDMA) modification present within some Sm proteins and coilin, which strengthens their interaction with SMN.^9,24,58,116 Unlike sDMA, the aDMA modification present within the GAR domain of fibrillarin does not influence its interaction with SMN.¹¹⁷ It is not clear if aDMA modifications in fibrillarin impact other aspects of its function, including localization, assembly into box C/D scaRNPs/snoRNPs, RNA binding and interaction with other proteins.

GAR1

What does GAR1 do in the CB?

Like fibrillarin, GAR1 (UniProt # Q9NY12) localizes to the nucleolus and CB.¹¹⁸ GAR1 replaces NAF1 in telomerase biogenesis, and is present in the mature telomerase RNP.³ GAR1 is also a core component of box H/ACA scaRNPs and snoRNPs, which are involved in the pseudouridylation of target sites within snRNA and rRNA, respectively.¹¹⁹ GAR1 is so named because it contains 2 glycine/arginine rich (GAR) domains,¹¹⁹ which contain aDMA modifications.^117,120 GAR1 interacts with SMN via either of its GAR domains, one of which is located at the N-terminus and the other is found at the C-terminus.^108,117 The exchange of NAF1 for GAR1 in both telomerase RNP and box H/ACA snoRNP (and presumably H/ACA scaRNPs) is thought to take place in the CB.¹²¹

How does modification impact GAR1s CB activity?

Phosphorylation, methylation, acetylation and ubiquitination

GAR1 contains PTMs (Table 1), but little is known about how/if these modifications regulate GAR1 activity. GAR1 contains 5 phosphorylated residues, 4 of which are tyrosines. None of the phosphorylation sites lie within the GAR domains present at the N- and C-terminus. The GAR domain does contain aDMA,^117,120 but this modification does not alter GAR1 interaction with SMN.¹¹⁷ It is possible that the aDMA modifications found within fibrillarin and GAR1 do in fact alter interaction with SMN in cells, but the experimental techniques used to assess the SMN/fibrillarin, SMN/GAR1 interaction may lack the resolution to detect such aDMA-mediated alterations. As with fibrillarin, the impact of aDMA on the activity of GAR1 as a core component of telomerase and box H/ACA snoRNPs and scaRNPs is unknown. It is possible that this modification regulates the exchange with NAF1 in the formation of the mature RNPs that takes place in the CB. We have recently observed that NAF1 is more highly recovered in SMN immunoprecipitation (IP) complexes than that found in coilin IP complexes.⁹⁶ Hence the interplay between SMN, coilin, GAR1, NAF1 and fibrillarin, which is likely regulated by PTMs, may contribute toward the formation of mature telomerase and H/ACA RNPs.

Nopp140

What does Nopp140 do in the CB?

Nopp140 (UniProt # Q14978) is enriched in the CB and nucleolus^92,122 and plays a crucial role in the formation of the ribosome.¹²³ It forms a complex with dyskerin, which also localizes to the CB and nucleolus.⁹² Additionally, Nopp140 has been shown to interact with coilin¹²⁴ and both classes (box H/ACA and box C/D) of snoRNPs,¹¹⁸ giving rise to the hypothesis that Nopp140 serves as a snoRNP chaperone and the link between the CB and nucleolus.¹¹⁸ Given these characteristics, it is possible that Nopp140 also contributes to scaRNP biogenesis in the CB. Intriguingly, studies using SMA cell lines have revealed a link between SMN and Nopp140 in the CB: the loss of Nopp140 from CBs is correlated with the severity of SMA.¹²⁵ In other words, lines derived from SMA type I (the most severe) patients had the lowest proportion of CBs with Nopp140 compared to lines derived from SMA type II or type III patients.¹²⁵ These findings indicate that SMN may contribute to Nopp140 functions in the CB, possibly related to H/ACA snoRNP or scaRNP biogenesis.¹²⁵

How does modification impact Nopp140s CB activity?

Phosphorylation, methylation, acetylation and ubiquitination

The role of methylation, acetylation and ubiquitination on Nopp140 activity is largely unexplored. In contrast, Nopp140 phosphorylation has been investigated. Nopp140 is one of the most highly phosphorylated proteins in the cell,^126,127 and is a substrate for casein kinase II (CK2)¹²⁷ and protein kinase A.¹²⁸ Phosphatases acting upon Nopp140 in the cell are at present unknown. Using an in vitro assay, the phosphorylation of Nopp140 was found to be required for the association of Nopp140 with snoRNPs.¹²⁹ Therefore, kinases and phosphatases likely regulate the snoRNP chaperone activity of Nopp140. Nopp140 has been shown to interact with CK2,¹³⁰ and it is likely that this interaction inhibits the activity of CK2.¹³¹ In particular, the phosphorylation of Nopp140 S574 greatly increases its affinity for CK2, providing a molecular switch that inactivates CK2. CK2 repression by Nopp140 is antagonized by inositol hexakisphosphate (IP₆), which reactivates CK2 by competing with Nopp140 for binding at the CK2 substrate binding site.¹³¹

PA28γ

What does PA28γ do in the CB?

PA28γ (UniProt # P61289) is a very well studied proteasome activator.¹³² In conditions of stress, such as that induced by UV-C exposure, CBs are disrupted and PA28γ co-localization with coilin is increased.¹³³ It is important to note that PA28γ is not present in CBs in untreated cells, but rather is dispersed uniformly in the nucleoplasm.¹³³ Interestingly, reduction of PA28γ by RNAi attenuates the impact of UV-C-mediated disruption on CBs, while overexpression of PA28γ promotes CB disassembly, verifying that PA28γ contributes to CB integrity.¹³³ In support of the idea that stress results in the increased association of coilin with PA28γ, Lafarga and colleagues have observed that motor neurons from patients with type I SMA have increased co-localization of coilin and PA28γ both in CBs and nucleoli compared to control cells.¹³⁴ These findings strongly indicate that PA28γ is responsive to various stresses and contributes to the reorganization of CB components.

How does modification impact PA28γ's CB activity?

Phosphorylation, methylation, acetylation, sumoylation and ubiquitination

PA28γ contains several PTMs (Table 1). Specific information about how these modifications impact the function of PA28γ in CB integrity is lacking. Regarding the UV-C-induced disruption of CBs, it was observed that PA28γ levels do not change significantly after this treatment. Consequently, one idea as to how the association of PA28γ with coilin is regulated is by phosphorylation of coilin or PA28γ. The kinase MEKK3 may participate in this process by modifying PA28γ¹³³, but this has not been proven as of yet.

PHAX

What does PHAX do in the CB?

Like CRM1, PHAX (UniProt # Q9H814) takes part in the export of spliceosomal snRNA^1,135 and is localized to the CB and nucleoplasm. Along with the cap binding complex (CBC) and ARS2, PHAX interacts with the 7-methylguanosine cap of nascent snRNA to form an export complex that also contains CRM1.^1,135,136 After remodeling steps that take place in the CB, the export complex transits to the cytoplasm.^1,78,82 The directionality of the nuclear export transport is conferred by CRM1-associated RAN bound to GTP. Once in the cytoplasm, GTP is hydrolyzed resulting in RAN-GDP and disassembly of the export complex. As detailed below, phosphorylation of PHAX also regulates the formation of the export complex and influences its directionality. Reduction of PHAX by RNAi disrupts CBs,¹⁵ demonstrating that snRNP biogenesis is required for CB formation.

How does modification impact PHAX's CB activity?

Phosphorylation, methylation and acetylation

PHAX contains several modifications (Table 1) but the PTM of PHAX that has been most investigated is phosphorylation. Indeed, PHAX stands for phosphorylated adapter for RNA export.¹³⁵ Nuclear PHAX is phosphorylated, and this modification is essential to form the export complex.¹³⁵ After cargo is delivered to the cytoplasm, PHAX is dephosphorylated and this, along with GTP hydrolysis by CRM1-associated RAN, disassembles the complex allowing for the recycling of PHAX and the other export components back into the nucleus.¹³⁵ Since, like CRM1, PHAX is also involved in the intranuclear transport of U3, U8 and U13 snoRNAs as well as telomerase RNA,^82,137 it is possible that phosphorylation also regulates this aspect of PHAX activity. In follow up work, Ohno and colleagues identified the sites within mouse PHAX that are phosphorylated: S14, S16, S18, S56, S57, S60 and S63.¹³⁸ Of these sites, S56, S57, S60 and S63 are the most important for export complex formation. The kinase responsible for PHAX phosphorylation is the nucleus-enriched casein kinase 2 and the cytoplasmic phosphatase that acts upon PHAX is the protein phosphatase 2A (PP2A).¹³⁸ It will be of interest in future studies to investigate how phosphorylation of these sites within PHAX regulates the intranuclear transport of U3, U8 and U13 snoRNAs, as well as telomerase RNA and scaRNAs.

SART3

What does SART3 do in the CB?

SART3 (UniProt # Q15020) is an snRNP assembly factor that interacts with U6 snRNA and accumulates in CBs.¹³⁹ It is suspected that SART3 association with the U6 snRNP promotes annealing with the U4 snRNP, forming the U4/U6 di-snRNP, and this activity takes place in the CB.¹³⁹ This annealing of U4 and U6 snRNPs is hypothesized to take place upon the synthesis of new snRNPs as well as after a splicing reaction takes place.¹³⁹ Building upon this work, Stanek and colleagues observed that inhibition of snRNP assembly induces CBs in primary cell lines (WI-38) that normally have very few CBs.¹⁴⁰ SART3, which is a coilin interacting protein, is required for the induction of CBs and the accumulation of incomplete snRNPs with coilin in CBs. These findings suggest that the CB is formed in response to an increased demand for snRNPs that outpaces the capacity of the nucleoplasm to accurately form these snRNPs. Hence the CB is involved in the quality control of snRNP assembly.¹⁴⁰

How does modification impact SART3s CB activity?

Phosphorylation, methylation, acetylation and ubiquitination

SART3 is extensively modified, containing ubiquitinated, acetylated, methylated and phosphorylated residues (Table 1). Somewhat surprisingly, there is no published information regarding the impact of these PTMs on the snRNP assembly activity of CBs. Nor is there any information about if the PTMs in SART3 affect CB induction in primary cell lines upon inhibition of snRNP assembly. Since SART3 is phosphorylated on 20 residues, it will be very interesting to ascertain if the modification of these residues is part of the regulatory system that governs CB formation, number and size.

SmD1

What does SmD1 do in the CB?

SmD1 (UniProt # P62314) is a core component of snRNPs. We are using SmD1 as representative of all Sm proteins, although not all Sm proteins share the same PTMs as found on SmD1. Along with SmB/B´, SmD2, SmD3, SmE, SmF and SmG, SmD1 is loaded onto the Sm motif cis element present within spliceosomal snRNAs.^1,2 This Sm ring assembly step of snRNP formation takes place in the cytoplasm, under the control of the SMN complex.^1,2 After additional cytoplasmic modifications of the snRNA, including 3´-end processing and trimethylguanosine cap formation,^1,2 the nascent snRNP is imported into the nucleus with the help of the SMN complex and other nuclear import factors,^1,141 where it first localizes to CB.³⁶ In addition to Sm proteins, the newly imported snRNP acquires additional snRNP-specific proteins, and it possible that this takes place in the CB.²

The CB is also the site wherein the snRNA component of the snRNP is modified by pseudouridylation (by box H/ACA scaRNPs) or methylation (by box C/D scaRNPs).⁸¹ These modifications are guided by complementary base pairing of the scaRNA component of the scaRNP with target sites within snRNA of the snRNP.^106,142-145 After all these biogenesis steps, snRNPs localize to active sites of transcription where splicing takes place or are stored in speckles.^1,2 Coilin interaction with Sm proteins and SMN at the CB may disengage snRNPs from the SMN complex.^10,11

How does modification impact SmD1s CB activity?

Phosphorylation, methylation and ubiquitination

Although SmD1 is phosphorylated and ubiquitinated (Table 1), the most extensively studied PTM of SmD1 is methylation. Specifically, SmD1 contains sDMAs, which increases its interaction with SMN.¹¹⁶ The sDMA modification within SmD1 facilitates interaction with the SMN complex in the cytoplasm so that Sm ring assembly takes place efficiently and on the proper site of the snRNA, thereby preventing Sm binding to other RNAs and forming a deleterious complex. It is not known if methylation or phosphorylation of SmD1, or PTMs found on other Sm proteins, alters their localization in CBs or interactions with CB components such as coilin. We have found, however, that one Sm protein (SmB´) preferentially interacts with a phosphomimetic coilin fragment but SMN prefers binding to a phospho-null coilin fragment.^11,146 This finding clearly shows that PTMs likely regulate CB protein dynamics and influence RNP biogenesis within this subnuclear organelle. The major methyltransferase that generates sDMAs on Sm proteins and Lsm proteins (which are part of the U6 snRNP), as well as on coilin is PRMT5.^{24,57,58,111,116,147} Recent work has shown a functional interaction between PRMT5 and huntingtin (Htt), and observed that sDMA modification of known PRMT5 substrates, such as coilin, is reduced in Huntington disease.¹⁴⁸ Thus, disrupted modification of PRMT5 target proteins may contribute to Huntington disease pathology. In addition to PRMT5, another methyltransferase, PRMT7, also contributes to sDMA modifications found in Sm proteins and, like PRMT5, is required for snRNP assembly.¹⁴⁹

TERT

What does TERT do in the CB?

TERT (UniProt # O14746) is the telomerase reverse transcriptase. This subunit of telomerase is responsible for adding the repeat sequence TTAGGG to the ends of chromosomes using telomerase RNA as the template.³ Telomerase holoenzyme assembly is suspected to take place in the CB,¹⁵⁰ which may then deliver telomerase to telomeres.¹⁵¹

How does modification impact TERT's CB activity?

Phosphorylation, acetylation and ubiquitination

TERT contains several PTMs (Table 1). Most notably, TERT contains 11 phosphorylated residues. There has been some effort at understanding how PTMs regulate TERT function. Specifically, TERT has been shown to be a substrate for the E3 ubiquitin ligase Hdm2, thereby negatively regulating telomerase activity.¹⁵² Also, several studies have shown that phosphorylation affects TERT: the Akt kinase-mediated phosphorylation of TERT increases telomerase activity,¹⁵³ but the c-Abl tyrosine kinase, which is activated by double stranded DNA breaks, inhibits TERT activity upon phosphorylation.¹⁵⁴ Protein phosphatase 2A (PP2A) has been shown to inhibit telomerase activity.¹⁵⁵ Furthermore, phosphorylation of TERT, especially by Akt at S227, has been shown to be important for the nuclear localization of TERT.^156,157 Despite these efforts, no studies have investigated how TERT PTMs regulate its association with the CB, or the delivery of telomerase by CBs to telomeres.

TGS1

What does TGS1 do in the CB?

The localization of TGS1, trimethylguanosine synthase I (UniProt # Q96RS0), is similar to SMN in that it is found in the cytoplasm and CBs.¹⁵⁸ Additionally, TGS1 directly interacts with SMN and forms the 5′ cap of spliceosomal snRNAs in the cytoplasm.¹⁵⁸ Trimethylguanosine cap formation of snoRNAs, which takes place in the CB, is catalyzed by a shorter TGS1 isoform compared to the full-length isoform which acts upon snRNAs in the cytoplasm.¹⁵⁹

How does modification impact TGS1s CB activity?

Phosphorylation and ubiquitination

Proteasomal processing of TGS1, resulting in the generation of a shorter nuclear isoform, may very well be regulated by the phosphorylation status of TGS1, which contains 30 phosphorylated residues. Unfortunately, no studies regarding the impact of phosphorylation on the trimethylguanosine synthase activity of TGS1, its interaction with the SMN complex, its localization, or its processing have been reported.

TOE1

What does TOE1 do in the CB?

TOE1 is a target of the immediate early transcription factor Egr1, and plays a role as an inhibitor of cell growth by influencing the level of the cyclin-dependent kinase inhibitor p21.¹⁶⁰ Curiously, TOE1 (also known as hCaf1z) localizes to the nucleus and CB.¹⁶¹ Additionally, TOE1 forms a complex with hCcr4d (which is also found in the CB) and has deadenylation activity.¹⁶¹ Both coilin and SMN interact with TOE1, and reduction of TOE1 disrupts CBs, decreases pre-mRNA splicing and inhibits cellular proliferation.¹⁶² A possible function for TOE1 (and the TOE1/hCcr4d complex) in the CB is not entirely obvious considering that RNA with polyA tails do not significantly accumulate in the CB. Indeed, the RNAs that accumulate in CBs (i.e. snRNAs, snoRNAs, scaRNAs) are not polyadenylated. However, fairly recent work has identified a polyadenylation pathway for the maturation of hTR.^163,164 These authors posit that much of this process occurs in the nucleolus, but leave open the possibility that final trimming of hTR may occur in CBs. If so, it stands to reason that TOE1, a 3′ to 5′ exonuclease, may contribute to hTR processing. TOE1 may also contribute to the 3′ processing of other RNAs that accumulate in CBs.¹⁶¹

How does modification impact TOE1s CB activity?

Phosphorylation

Human TOE1 is phosphorylated on 7 residues, 6 serines and 1 threonine. At present, there is no information regarding the kinases and phosphatases that modify TOE1. Also lacking are any functional studies that indicate how phosphorylation of TOE1 impacts its function, interactions, or localization to CBs.

USPL1

What does USPL1 do in the CB?

USPL1, ubiquitin-specific protease-like 1 (UniProt # Q5W0Q7), was recently identified as a component of CBs that is essential for proliferation and canonical CB formation.¹⁶⁵ USPL1 is the third type of SUMO protease reported in the literature, but this activity does not underlie the alterations in CB formation and impaired proliferation observed upon reduction of USPL1.¹⁶⁵ Additional work on USPL1 has shown that snRNA transcription is reduced upon depletion of USPL1 by RNAi.¹⁶⁶ Moreover, snRNP assembly and pre-mRNA splicing are negatively impacted upon reduction of USPL1.¹⁶⁶ USPL1 also associates with U1 and U2 snRNA gene loci and interacts with components of the little elongation complex, which is involved in the regulation of snRNA gene transcription.¹⁶⁶ Together, these data support the idea that USPL1 plays an important role in the transcription of snRNA genes by RNA polymerase II.¹⁶⁶

How does modification impact USPL1s CB activity?

Phosphorylation

Not surprisingly considering that USPL1 was only recently identified, no studies regarding the regulatory role of PTMs on USPL1 activities in the CB have been reported. Considering that USPL1 has 12 phosphorylated residues, it is likely that this post-translational modification contributes to USPL1 function and, by extension, CB formation and activity.

What signals promote modification of CB proteins?

A variety of conditions have been shown to impact CBs.^146,167 In regards to CB formation, a parallel can be drawn between the biogenesis of nucleoli at sites of rDNA transcription and the nucleation of CBs at sites of snRNA gene transcription (such as RNU1 and RNU2).^146,167 Astonishingly, recent work has found that CB association with certain gene loci contributes to genome organization, especially of chromosome 1, and CBs impact the expression of many genes and RNA splicing.¹⁶⁸ Inhibition of transcription disrupts canonical CBs,^146,167 clearly indicating that the CB is a dynamic subnuclear structure that is responsive to the overall snRNA levels in the cell. It is likely that cellular cues regulating snRNA transcription levels reach the CB via signal transduction cascades that culminate in the post-translational modification of proteins enriched within the CB, but a clear elucidation of such pathways awaits further investigation. Interestingly, reduction of CB components, such as WRAP53, resulting in CB disruption appears to induce a feedback control mechanism designed to increase important CB proteins, such as coilin, by stimulating the transcription of genes, such as COIL, that encode these proteins.¹⁶⁸ Since coilin is extensively modified, this putative compensatory mechanism must also include the regulation of enzymes that properly configure coilin.

Although not extensively discussed here, it is also highly likely that the activity of the SNAP_c, little elongation and integrator complexes, all of which contribute to snRNA gene transcription, are regulated by PTMs. One example, ICE1, which is part of the little elongation complex, contains (according to PhosPhoSitePlus) 72 phosphorylated residues. Unfortunately, a PubMed search for ICE1 and phosphorylation yields zero results. Additionally, INTS1, a component of the integrator complex, is phosphorylated on 37 residues but no functional information regarding these modifications is present. Regarding the SNAP_c complex, a bit more information regarding PTMs and the function of this complex is known compared to the little elongation or integrator complexes. For example, it has been shown that CK2 phosphorylates SNAP190, which in turn regulates SNAPc complex binding and transcription of the U6 snRNA gene by RNA polymerase III.¹⁶⁹ Clearly, therefore, the post-translational regulation of these 3 transcription complexes is largely unchartered territory and fertile ground for investigation.

Additional, but unknown at present, signaling pathways likely also impact the modification of CB proteins during the cell cycle and development. For example, CBs are disassembled during mitosis^170,171 and the composition and number of CBs changes during development.^172,173 Additionally, immortalization and transformation both increase the frequency and size of CBs,¹⁷⁴ while viral infection by adenovirus and herpes virus disrupts CBs and alters the localization of some components, such as coilin.^175-177 Finally, serum starvation, DNA damage and inhibition of translation, nuclear export and phosphatases all disrupt canonical CBs.¹⁴⁶ In regards to serum starvation, 2 proteins that respond to mitogenic stimuli and impact CBs are FGF-2²³ and ZPR1. FGF-2²³, as mentioned above, interacts with SMN and contains sDMAs.⁶⁰ Moreover, FGF-2²³ is part of the INFS (integrative nuclear FGR1 signaling) pathway and thus may transmit mitogenic signals to SMN and CBs.^178,179 A connection between the epidermal growth factor receptor (EGFR) and CBs is made via ZPR1. ZPR1 binds the cytoplasmic domain of EGFR and, upon EGF binding to EGFR, ZPR1 is released and can then interact with elongation factor 1α and SMN. The interaction between ZPR1 and SMN increases the localization of SMN in CBs.¹⁸⁰ Together with TOE1, therefore, ZPR1 and FGF-2²³ may be mediators of proliferative signals to CBs, possibly resulting in the modification of CB-enriched proteins. Another signaling protein found in CBs is peroxiredoxin V, which is responsive to reactive oxygen species.¹⁸¹ It is possible that peroxiredoxin V accumulation in CBs is an indicator of a stress response pathway that may alter the composition of protein modification within the CB.

Although the conditions listed above (cell cycle, development, etc.) most likely trigger signal transduction cascades that result in the modification of CB proteins, the identification of the modifying enzymes and the functional consequence of such PTMs on CB proteins and CB function is largely unknown. Known modification factors that act upon the CB proteins listed in Table 1 are shown in Table 2. As an example of how we need more clarification regarding PTMs and the CB, it is well established that CBs disassemble during mitosis, which is when coilin is hyperphosphorylated.¹⁷ Since the self-association of coilin is reduced upon hyperphosphorylation,⁷ a model can be developed in which the phosphorylation state of coilin impacts CB formation. In order for this model to be complete, however, one needs to identify the kinases and phosphatases that modify coilin in a cell cycle-dependent manner. Recent work has shown that VRK1 may be one such kinase.²³

Table 2. Modifiers of CB proteins.

Substrate	Modification	Modifier	Reference
Coilin	Phos	Cdk2/cyclin E	^182
	Phos	VRK1	^23,183
	Phos	VRK2	^183
	Phos	CK2	^7
	Dephos	PPM1G	^20
	Meth (sDMA)	PRMT5	^9,24
CRM1	Phos	Cdk1/cyclin B	^85
DAXX#	Phos	ATM	^91
Dyskerin		Not known
Fam118B		Not known
Fibrillarin	Meth (aDMA)	PRMT1	^115
GAR1	Meth (aDMA)	PRMT1	^117,120
Nopp140	Phos	CK2	^127,130,131
	Phos	Protein kinase A	^128
PA28γ#	Phos	Chk2	^184
PHAX	Phos	CK2	^138
	Dephos	PP2A	^138
SART3		Not known
SmD1	Meth (sDMA)	PRMT5	^57,147
	Meth (sDMA)	PRMT7	^149
SMN	Phos	PKA	^185
	Dephos	PPP4	^186
	Dephos	PP1γ	^51,52
	Dephos	PPM1G	^49,50
	Dephos	HD-PTP/PTPN23	^52
	Ubiq	Itch	^66
	Ubiq	UHCL1	^64
	Ubiq	Mind bomb 1 E3	^187
	Deubiq	Usp9X	^188
TERT#	Phos	Akt	^153,157
	Phos	c-Abl	^154
	Dephos	PP2A	^155
	Ubiq	Hdm2	^152
TGS1		Not known
TOE1		Not known
USPL1		Not known
WRAP53		Not known

# For simplicity, we only list one modifier of DAXX and PA28γ, and a few modifiers of TERT.

Several figures have been provided which depict the current knowledge about the PTM of CB-enriched proteins. In Fig. 1, we show the impact of PTMs on snRNP biogenesis, with specific emphasis on SMN modifications as well as the snRNA export pathway. Fig. 2 focuses on coilin modifications and modifiers. Fig. 3 specifically examines the impact of PTMs on the interactions and localization of SMN and coilin.

Figure 1. PTMs and snRNP biogenesis. A model showing known modifications of proteins involved in snRNP biogenesis, including the snRNA export pathway. Enzymes that modify these proteins are shown. Phosphorylated, methylated, ubiquitinated and sumoylated proteins are indicated. Processes marked by question marks indicate that there is no information on how PTMs affect localization or interaction; or where the PTM is known, but the enzyme(s) responsible for the modification are not known. Proteins that are modified, such as TGS1, but the functional consequence of this modification and the modifying enzymes are unknown, are indicated in the figure but lack denoted modification. Also not shown in the model are proteins with combinations of PTMs. (P = Phosphorylation; Me = Methylation; So = Sumoylation; Ub = Ubiquitination).

Figure 2. Modifications and modifiers of coilin, the CB marker protein. Enzymes known to modify coilin are shown. Not shown are coilin proteins with multiple different PTMs. During mitosis, coilin is hyperphosphorylated (coilin with 6 phosphorylations), which correlates with decreased coilin self-association and disassembled CBs. Hypomethylated coilin, which is enriched within the nucleolus, is indicated. Also shown in the nucleolus is a coilin phosphomutant (S184A). More details in text.

Figure 3. The impact of PTMs on the interactions and localizations of SMN and coilin. It is hypothesized that coilin in the nucleoplasm is hyperphosphorylated compared to coilin in CBs. During mitosis, coilin is hyperphosphorylated and CBs disassemble. CBs reform early to mid G1 (indicated in model). Mitotic coilin is reduced in its self-association compared to interphase coilin (paired coilin in the CB). Coilin interacts with SMN, Sm proteins, sRNPs and WRAP53. Methylation of coilin (and certain SM proteins) increases it interaction with SMN. Phosphorylation of coilin differentially affects its interaction with SMN and Sm proteins, with SMN preferentially binding to coilin that is less phosphorylated than that preferred by Sm proteins. The differential interaction of coilin with SMN and Sm proteins may disengage the nascent snRNP from the SMN complex in the CB, allowing for subsequent snRNP biogenesis steps, including snRNA modification and snRNP-specific protein association. Many non-coding RNAs, such as scaRNAs, snRNAs and snoRNAs are enriched in the CB and coilin complex (curved lines). Coilin interaction with RNAs decreases when coilin is hyperphosphorylated.¹²

Future directions

Moving forward, a more complete understanding as to how PTMs impact the function of proteins enriched within the CB will be crucial in order to decipher how this subnuclear domain is formed and its activity regulated. At present, there is little information that details protein modification within the CB vs the modification status of these same proteins in other cellular locales. This type of information is known for SMN in that cytoplasmic SMN is more likely to be phosphorylated at serine 28 and 31 compared to nuclear SMN⁴⁹ and coilin since it has been shown that hypomethylated coilin localizes to the nucleolus.²⁶ Ideally, the modification of all CB proteins shown in Table 1 could be determined for these proteins in all the cellular localizations they accumulate, be it the CB, nucleolus, nucleoplasm or cytoplasm. In so doing, the critical modifications that impact the localization and function of these proteins could be more easily ascertained. This information may also help identify the cellular location of the enzymes that modify these CB-enriched proteins. Along these lines, the cellular cascades that trigger CB formation and govern their numbers also need to be more fully elucidated. Strikingly, all 18 of the proteins we evaluated here are phosphorylated, but astonishingly little has been done to evaluate how this modification impacts their activity, either in the CB or other cellular locations. It is our hope that the presentation of these gaps in our knowledge will spur the research community to make a concerted effort culminating in an understanding of regulated Cajal body activity.

Since the CB plays an important role in the biogenesis of snRNP, scaRNP, snoRNP and telomerase, all of which are required for efficient and rapid cell proliferation, identification of signaling pathways that could be inhibited should be of therapeutic value in the treatment of cancer. In support of this hypothesis, reduction of many different inhabitants of the CB, as detailed above, results in decreased proliferation of cancer cell lines. Additionally, SMA and DC, both of which are diseases that have alterations in RNPs, may in fact be the result of more widespread disruptions in CB functions related to the biogenesis of other RNPs besides snRNPs for SMA and telomerase for DC. Studies that seek to clarify how PTMs regulate the formation of these RNPs, therefore, may shed light on the SMA and DC disease states.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.