The Diverse Cellular Functions of Inner Nuclear Membrane Proteins

Abstract

The nuclear compartment is delimited by a specialized expanded sheet of the endoplasmic reticulum (ER) known as the nuclear envelope (NE). Compared to the outer nuclear membrane and the contiguous peripheral ER, the inner nuclear membrane (INM) houses a unique set of transmembrane proteins that serve a staggering range of functions. Many of these functions reflect the exceptional position of INM proteins at the membrane–chromatin interface. Recent research revealed that numerous INM proteins perform crucial roles in chromatin organization, regulation of gene expression, genome stability, and mediation of signaling pathways into the nucleus. Other INM proteins establish mechanical links between chromatin and the cytoskeleton, help NE remodeling, or contribute to the surveillance of NE integrity and homeostasis. As INM proteins continue to gain prominence, we review these advancements and give an overview on the functional versatility of the INM proteome.

During evolution, the nuclear envelope (NE) arose as a separating membrane sheath between the genetic material and the cytoplasm. This physical barrier became one of the hallmarks of eukaryotic cells and defined nuclear compartmentalization. The emergence of the NE did not only generate a protective cover of the genome but also spatially separated transcription and translation, allowing for new forms of their regulation. Over the course of a billion years, the NE got equipped with a set of membrane-embedded proteins that endowed it with diverse tasks, yielding a functionally versatile compartment interface.

One of the most prominent features of the NE is its asymmetry. The NE is built by two double lipid bilayers, an outer nuclear membrane (ONM) and an inner nuclear membrane (INM), which are compositionally distinct with respect to their membrane protein assortment although both membranes are interconnected at numerous sites. Compared to the ONM and the connected endoplasmic reticulum (ER), the INM is enriched in a unique collection of membrane proteins that maintain close associations with chromatin and a network of intermediate filament proteins called nuclear lamins.

Clearly, the molecular identity of the INM is reflective of its functional specialization. Early work had identified only a handful of INM proteins (Senior and Gerace 1988; Worman et al. 1988; Foisner and Gerace 1993; Manilal et al. 1996; Lin et al. 2000). Later, the mass spectrometric characterization of the NE proteome from different sources, including rat liver and muscle (Schirmer et al. 2003; Wilkie et al. 2011; Korfali et al. 2012), human leukocytes (Korfali et al. 2010), and mouse neuroblastoma cells (Dreger et al. 2001), identified nearly 1000 putative nuclear envelope transmembrane proteins (NETs). However, as the NE is interconnected with ER, a fraction of the identified putative NETs may not be NE-specific and represent membrane proteins of the peripheral ER. Therefore, protein localization studies are a decisive element of NET definition. Remarkably, many of these NETs showed a high level of tissue specificity, suggesting that the NE proteome is adapted to cell functionality. Only a small fraction of NETs has so far been confirmed to be enriched at the NE, and most of those are also residents of the INM. To date, we know roughly 35 integral membrane proteins that were found to localize at the INM of mammalian cells (Table 1). These include prominent, well-studied INM proteins like the lamin B receptor (LBR), the LAP2-Emerin-Man1 (LEM)-domain family members lamina-associated polypeptide 2 (LAP2), emerin, MAN1, and LEM2, the Sad1p, Unc-84 (SUN)-domain proteins SUN1 and SUN2, as well as LAP1 and NET5 (Table 1; Fig. 1). Various others are less well characterized (Table 1) and the biological significance of their INM localization remains unexplored.

Table 1.

Well-characterized (in bold) and other putative inner nuclear membrane (INM) proteins, listing their predicted molecular weight, the number of transmembrane segments, membrane orientation, and references pertaining to their localization at the INM

Gene name/NET designation	Protein name	Predicted molecular weight (kD)	Predicted transmembrane (TM) segments	Orientation	References for localization at the INM
LBR	Lamin B receptor (LBR)	58 (observed), 70.7 (predicted)	8	Multipass	Worman et al. 1988
SUN1	SUN domain-containing protein 1 (SUN1)	90	1	Single-pass, type II	Dreger et al. 2001
SUN2	SUN domain-containing protein 2 (SUN2)	80.3	1	Single-pass, type II	Hodzic et al. 2004
TMPO	Lamina-associated polypeptide 2 (LAP2)	50.6 (isoform β)	1	Single-pass, type II	Foisner and Gerace 1993
EMD	Emerin (EDMD)/LEMD5	28.9	1	Single-pass, type II	Manilal et al. 1996
LEMD3/MAN1	INM protein MAN1 (MAN1)/LEMD3	99.9	2	Multipass	Lin et al. 2000
LEMD2	LEM-domain-containing protein 2 (LEMD2)	56.9	2	Multipass	Brachner et al. 2005
TMEM201/NET5	TM protein 201/SAMP1	43.3	4	Multipass	Buch et al. 2009
TOR1AIP1/LAP1	Torsin-1A-interacting protein 1 (TOR1AIP1)/Lamina-associated polypeptide 1 (LAP1)	66.2 (isoform B)	1	Single-pass, type II	Senior and Gerace 1988
NRM	Nurim	29.3	6	Multipass	Rolls et al. 1999
LUMA/TMEM43	TM protein 43 (TMEM43)/LUMA	44.8	4	Multipass	Dreger et al. 2001
TMEM120A/NET29	TM protein 120A (TMEM120A)/TMPIT	42.8	5	Multipass	Malik et al. 2010
MOSPD3/NET30	Motile sperm domain containing 3 (MOSPD3)	25.5	2	Multipass	Malik et al. 2010
SCARA5/NET33a	Scavenger receptor class A, member 5 (SCARA5)	43.2	1	Single-pass, type II	Malik et al. 2010
SLC39A14/NET34a	Metal cation symporter ZIP14 (ZIP14)/SLC39A14	52.8	7	Multipass	Malik et al. 2010
MYORG/NET37	Myogenesis-regulating glycosidase (MYORG)/KIAA1161	81	1	Single-pass, type II	Malik et al. 2010
PLPP7/NET39	Inactive phospholipid phosphatase 7 (PLPP7)/PPAPDC3/C9orf67	21.9	4	Multipass	Liu et al. 2009; Malik et al. 2010
TM7SF2/NET47b	Delta (14)-sterol reductase TM7SF2/ANG1/DHCR14A	46.4	7	Multipass	Malik et al. 2010
DHRS7/NET50	Dehydrogenase/reductase (SDR family) member 7 (DHRS7)	38.2	1	N.D.	Malik et al. 2010
ERG28/NET51b	Ergosterol biosynthetic protein 28 (ERG28)/C14orf1	15.8	4	Multipass	Malik et al. 2010
APH1B/NET55b	Anterior pharynx defective 1B (APH1B)	28.4	6	Multipass	Malik et al. 2010
NCLN/NET59b	Nicalin (NCLN)	62.9	2	Multipass	Malik et al. 2010
NCEH1b	Neutral cholesterol ester hydrolase 1 (NCEH1)/AADACL1/KIAA1363	45.8	1	Single-pass, type II	Korfali et al. 2010
MAGT1a	Magnesium transporter protein 1 (MAGT1)/IAG2	38	4	Multipass	Korfali et al. 2010
METTL7Ab	Methyltransferase-like protein 7A (METTL7A)	28.3	1	N.D.	Korfali et al. 2010
TMEM41A	TM protein 41A (TMEM41A)	29.7	5	Multipass	Korfali et al. 2010
STT3Ab	Dolichyl-diphosphooligosaccharide—protein glycosyltransferase subunit STT3A(STT3-A)/ITM1/TMC	80.5	12	Multipass	Korfali et al. 2010
TMEM38Ab	Trimeric intracellular cation channel type A (TRIC-A)/TMEM38A	33.6	7	Multipass	Wilkie et al. 2011
LPCAT3b	Lysophospholipid acyltransferase 5 (LPCAT3)/MBOAT5/OACT5	56	9	Multipass	Wilkie et al. 2011
NEMP1	Nuclear envelope integral membrane protein 1 (NEMP1)/TMEM194/KIAA0286	50.6	5	Multipass	Wilkie et al. 2011
TMEM214c	TM protein 214 (TMEM214)	77	2	Multipass	Wilkie et al. 2011
LRRC59b	Leucine-rich repeat-containing protein 59 (LRRC59)/p34	35	1	Single-pass, type II	Blenski and Kehlenbach 2019

All the putative INM proteins listed here fulfilled two criteria: (1) resistance to prefixation extraction with triton X-100, and (2) localization of a fraction of the total population at the INM observed through superresolution microscopy. Please note that possibly all membrane proteins that are inserted in the endoplasmic reticulum (ER) and possess extraluminal domains <60 kDa can access the INM. However, this does not ensure their enrichment and functional relevance at the INM. The other predominant cellular localizations for the putative INM proteins are annotated in the footnotes. Please note that SUN3, SUN4 (Spag4), and SUN5 (Frohnert et al. 2011; Calvi et al. 2015; Pasch et al. 2015) are specifically expressed in testis where they enrich at the NE. However, their localization to the INM remains to be formally proven.

(N.D.) No data available.

^aProteins also localize at the plasma membrane as observed for SCARA5 (Huang et al. 2010), SLC39A14 (Taylor et al. 2004; Tuschl et al. 2016), and MAGT1 (Zhou and Clapham 2009).

^bProteins primarily localize at the ER as observed for SLC39A14 (Malik et al. 2010), TM7SF2 (Holmer et al. 1998), ERG28 (Malik et al. 2010), APH1B (Malik et al. 2010), NCLN (Dettmer et al. 2010), NCEH1 (Igarashi et al. 2010), MAGT1 (Cherepanova et al. 2014), METTL7A (McKinnon and Mellor 2017), STT3A (Braunger et al. 2018; Zhu et al. 2019), TMEM38A (Yazawa et al. 2007; Shrestha et al. 2020), LPCAT3 (Zhao et al. 2008), and LRRC 59 (Zhen et al. 2012).

^cPredominantly localizes at the ER and the Golgi complex (Li et al. 2013).

Figure 1.

Schematic representation of well-characterized inner nuclear membrane (INM) proteins indicating their membrane topology, domain organization, and known interaction partners. INM proteins bind chromatin, chromatin-associated factors, and nuclear lamins as indicated. Certain INM proteins like lamin B receptor (LBR) and LAP2-Emerin-Man1 (LEM)-domain proteins bind transcription factors and thereby modulate gene expression. Additionally, some INM proteins bind to each other while others, like LBR, self-interact, altogether creating a complex network of interactions. At the INM, Sad1p, Unc-84 (SUN) domain proteins form trimers via their coiled-coil domains and interact with three KASH peptides of outer nuclear membrane (ONM)-resident Nesprins to form the LINC (linker of nucleoskeleton and cytoskeleton) complex.

Most transmembrane proteins of the INM are cotranslationally inserted into the ER membrane and assume either a type II or a polytopic topology, with their extended amino termini facing the cytosol (Fig. 1). From the ER, these membrane proteins distribute to the ONM and INM by diffusion, passing through narrow membrane-proximal openings of nuclear pore complexes (NPCs) (Soullam and Worman 1995; Ohba et al. 2004; Theerthagiri et al. 2010; Zuleger et al. 2011; Boni et al. 2015; Ungricht et al. 2015; Pawar et al. 2017). These peripheral channels prevent the passage of membrane proteins with extraluminal domains larger than 60 kDa. The accumulation of INM-destined proteins at the nuclear face of the NE is determined by their retention on nuclear lamins and/or chromatin (Soullam and Worman 1995; Zuleger et al. 2011; Boni et al. 2015; Ungricht et al. 2015; Pawar et al. 2017). Thus, in principle, ER proteins with small extraluminal domains can also partition to the INM (Soullam and Worman 1995; Zuleger et al. 2011; Boni et al. 2015; Ungricht et al. 2015; Pawar et al. 2017), but in contrast to bona fide INM proteins, they are not enriched at the NE. Notably, in dividing cells undergoing open mitosis, both INM and ER proteins disperse in the mitotic ER network. During mitotic exit, a nearly synchronous enrichment of INM proteins on chromatin drives the process of NE reformation during late anaphase and telophase, rapidly reestablishing the INM membrane territory, representing a second, postmitotic mode of targeting proteins to the INM (Mattaj 2004).

The INM can be viewed as a functionally multifaceted membrane domain (Fig. 2). INM proteins engage in an array of genome-regulatory functions ranging from spatial genome organization (Solovei et al. 2013), epigenetic silencing (Somech et al. 2005), DNA replication (Martins et al. 2003), DNA damage repair (Lei et al. 2012), and transcriptional regulation. They also play important roles in diverse signaling pathways (Berk et al. 2013), in lipid synthesis (Tsai et al. 2016), and mediate mechanobiological tasks, including nuclear anchorage, migration, and mechanotransduction (Lammerding et al. 2005; Rothballer and Kutay 2013; Cho et al. 2017; Lee and Burke 2018). Interestingly, many of these functions converge and are essential for differentiation and development. They are realized through an intricate spectrum of biochemical interactions of INM proteins with chromatin, the nuclear lamina, and regulatory factors (Fig. 2).

Figure 2.

The diverse functions of inner nuclear membrane (INM) proteins. (A) Many INM proteins tether chromatin to the nuclear envelope (NE) and facilitate repression of genomic elements. Some INM proteins assist processes such as splicing, DNA replication, or the DNA damage response. (B) Members of the LINC (linker of nucleoskeleton and cytoskeleton) complex have mechanical and structural roles, interlinking the nucleoskeleton and cytoskeleton and thereby acting as force transmission devices across the NE. LINC complexes perform a wide range of functions, including nuclear anchorage and migration, chromosome movements, and the maintenance of NE membrane spacing. Other INM proteins (e.g., emerin and SAMP1) may functionally collaborate with LINC complexes in mechanotransduction. (C) INM proteins interact with transcription factors to regulate gene expression, thereby contributing to diverse signaling pathways. (D) Certain INM proteins possess enzymatic functions. Lamin B receptors (LBRs) exhibit C14 sterol reductase activity that is used in cholesterol synthesis. Others, like LAP1 and LEM2, directly or indirectly regulate the activities of partner proteins such as Torsins and the ESCRT-III complex, respectively.

An appreciation of the functional complexity of INM proteome is imperative to achieve an integrated view of INM biology. This is of particular importance in light of the numerous links that exist between INM protein dysfunction and human diseases. The idea that INM proteins are physiologically relevant was first realized by the discovery that emerin, mutations of which cause X-linked Emery–Dreifuss muscular dystrophy, is located at the INM (Bione et al. 1994; Manilal et al. 1996). This placed an INM protein at the causative epicenter of a disease, which paved ways for the intensification of research directed toward the cellular functions of INM proteins and their implications in human health. Presently, over 15 syndromes have been linked to the dysfunction of INM proteins (Table 2), and the list of diseases continues to expand. Every new disease link reiterates the significance of the INM proteome and stimulates novel ways to perceive how human physiology is influenced by the cellular functions of INM proteins. In this review, we use the examples of well-characterized, widely expressed INM proteins to portray their broad functional repertoire, giving priority to mammalian representatives.

Table 2.

Human diseases (in bold) related to either mutations in genes encoding for inner nuclear membrane (INM) proteins or a misregulation of INM protein levels

Disease	Pathology/clinical features	Implicated INM protein	Associated mutations/defects	References
Pelger-Huët anomaly	Hyposegmentation and abnormal chromatin organization in granulocyte nuclei	Lamin B receptor (LBR)	P119L, P569R, R377X*, W436X*, L534X*, V11EfsX24, G382DfsX39, S167TfsX176, loss of splicing of exon 3, 11, 14, altered splicing of exon 13	Hoffmann et al. 2002; Best et al. 2003; Clayton et al. 2010; Tsai et al. 2016
Greenberg skeletal dysplasia	Abnormal sterol metabolism, hydrops, ectopic calcification, and moth-eaten (HEM) skeletal dysplasia	LBR	N547D, R583Q, L534X, Y468TfsX475, V11EfsX24	Waterham et al. 2003; Clayton et al. 2010; Tsai et al. 2016; Turner and Schlieker 2016
Pelger-Huët anomaly with mild skeletal anomaly	Bilobed neutrophil nuclei with mild skeletal dysplasia	LBR	I218DfsX19, R586H, R76X, N547S	Borovik et al. 2013; Sobreira et al. 2015
Reynolds syndrome	Primary biliary cirrhosis, cutaneous systemic sclerosis	LBR	R372C	Gaudy-Marqueste et al. 2010
Emery–Dreifuss muscular dystrophy (EDMD) or EDMD-related myopathies	Joint contracture, muscle wasting and weakness, cardiac diseases with conduction defects, and arrhythmia	Emerin	∼100 mutations in emerin have been associated with X-linked EDMD, including M1V, S54F, Q133H, P183H, P183T, S143X, W200X, Q44X, Q86X, S171X, Q219X, Q228X, Y41X, S66X, S120X, W226X, Δ95–99, R207GfsX30, L84fsX7, D9GfsX24, F39SfsX17, R45KfsX16, F190YfsX19, R203AfsX34, R204PfsX7, S52QfsX9, L84PfsX7, Y85LfsX8, E11SfsX2, S49LfsX11, and others	Bione et al. 1994; Mora et al. 1997; Ellis et al. 1999; Morris and Manilal 1999; Astejada et al. 2007; Ben Yaou et al. 2007; Brown et al. 2011; Dai et al. 2019; Zhou et al. 2019
		LUMA	E85K, I91V	Liang et al. 2011
		MAN1	G88V, R230T occur along with mutations in dysferlin, plectin, ryanodine receptor 3, ankyrin2, nesprin1, integrator complex subunit 1, or titin	Meinke et al. 2020
		NET39	M92K, R252P occur along with mutations in plectin, collagen, dysferlin, or integrator complex subunit 1	Meinke et al. 2020
		SUN1	A230V, G76A, and W377C occur along with mutations in either emerin or lamin A/C: G68D, G338S, W377C	Meinke et al. 2014
		SUN2	A56P and V378I occur along with mutations in lamin A/C: R620C, E438D	Meinke et al. 2014
		TMEM201/Samp1	G15A, G18S, and G597S occur along with mutations in lamin A/C, nesprin1, or SUN1	Meinke et al. 2020
		TMEM38A	V247M and D260N occur along with mutations in plectin, collagen, dysferlin, or integrator complex subunit 1	Meinke et al. 2020
		TMEM214	G236S and R179H occur along with mutations in nesprin3 or collagen	Meinke et al. 2020
Buschke–Ollendorff syndrome	Connective tissue nevi combined with osteopoikilosis	MAN1	W621X, Y441X, R625X, L611X, S619X, W855X, R655X, and L638fs mutation at exon 12/intron 12 boundary causing defective splicing	Hellemans et al. 2006; Kobayashi et al. 2007; Zhang et al. 2009; Burger et al. 2010; Korekawa et al. 2012; Brodbeck et al. 2016
Isolated osteopoikilosis	Small and round spots of increased bone density that are mainly located in the epiphyseal regions of the tubular bones	MAN1	E601X, Q153X, C1323A, R537X, L478X, R678X, and R735X	Hellemans et al. 2004, 2006; Couto et al. 2007; Mumm et al. 2007; Baasanjav et al. 2010
Isolated melorheostosis	“Dripping wax” appearance in the cortex of long bones, accompanied by abnormalities of adjacent soft tissues	MAN1	L638X	Hellemans et al. 2006
Hutterite-type cataract (with or without arrhythmic cardiomyopathy)	Clouding of the lens of the eye; some patients have suffered sudden death presumably of arrhythmogenic origin	LEM2	L13R	Boone et al. 2016
Severe dystonia with cerebellar atrophy	Progressive dystonia, severe contractures of the Achilles tendons, and feet deformations	LAP1	E482A	Dorboz et al. 2014
Limb–girdle muscular dystrophy type 2	Proximal and distal weakness and atrophy, rigid spine and contractures of the proximal and distal interphalangeal hand joints, cardiomyopathy, and respiratory involvement	LAP1	L394P, E62DfsX25, P43fsX15	Kayman-Kurekci et al. 2014; Ghaoui et al. 2016
Unnamed multisystemic disease	Psychomotor retardation, cataract, heart malformation, sensorineural deafness	LAP1	R321X	Fichtman et al. 2019
Arrhythmogenic right ventricular cardiomyopathy/dysplasia (ARVC5)	Cardiomyocyte apoptosis, lethal ventricular tachycardia, and fibro-fatty infiltration mainly in the right ventricle	LUMA	S358L	Stroud et al. 2018
Dilated cardiomyopathy	Increase in left ventricular systolic and diastolic diameter and decrease in ejection fraction	LAP2α	R690C	Taylor et al. 2005
Hutchinson–Gilford progeria syndrome	Premature aging disease: individuals exhibit low body weight, decreased joint mobility, scleroderma, and die around the age of 13	SUN1	Up-regulated levels of SUN1 in patients expressing mutant lamin A/C	Chen et al. 2012
Hypertrophic cardiomyopathy	Cardiac arrythmia, supraventricular extrasystoles	SUN2	M50T, V378I occur along with a mutation in myosin binding protein C	Meinke et al. 2014
Mandibuloacral dysplasia type A	Growth retardation, craniofacial anomalies, mottled cutaneous pigmentation, skin rigidity, partial lipodystrophy, insulin resistance	SUN2	Mislocalization of SUN2 due to mutation in LMNA	Camozzi et al. 2012
Cancers	Subtype: cervical, colon, colorectal, gastric, pancreatic cancers, myeloma, lymphoma, medulloblastoma, and others	LAP2α	Up-regulated levels of LAP2α	Reviewed in Brachner and Foisner (2014)
	Subtype: digestive tract cancers	LAP2β	Up-regulated levels of LAP2β	Kim et al. 2012
	Subtype: ovarian cancer	Emerin	Reduced expression of emerin	Capo-chichi et al. 2009
	Subtype: hepatocellular carcinoma and neoplasms	NET33	Down-regulation of NET33	Liu et al. 2018; Ulker et al. 2018
	Subtype: aggressive breast, ovarian and prostate cancers	AADACL1	Up-regulated levels/activity of AADACL1	Chiang et al. 2006; Chang et al. 2011

The table does not cover diseases associated with nuclear lamins and barrier to autointegration factor (BAF).

(X) Termination codon, (fs) frameshift, (*) the amino acids at which premature termination occurs were allocated using Uniprot.

LAMIN B RECEPTOR (LBR)

When radioactive lamins were incubated with erythrocyte NE membranes, B-type lamins bound a 58 kDa protein with high specificity (Worman et al. 1988). This protein, designated as the “lamin B receptor,” was the first integral INM protein to be identified and remains one of the most well-characterized proteins of the INM to date. Structurally, LBR is a polytopic membrane protein with a hydrophilic amino-terminal domain that protrudes into the nucleoplasm and is composed of a Tudor domain, a serine/arginine-rich (RS) hinge region, and a second globular domain. This is followed by a hydrophobic region containing eight membrane-spanning helices, and a short nucleoplasmic carboxy-terminal tail (Fig. 1). Functionally, LBR is a versatile protein. Whereas its extraluminal domain contributes to chromatin organization in the nuclear periphery, the transmembrane segments exhibit sterol reductase activity.

Early biochemical studies had revealed a striking enrichment of heterochromatin marks on chromatin pulled down by LBR (Makatsori et al. 2004). Later, a combination of histone tail peptide arrays, chromatin immunoprecipitation, and direct binding experiments indicated that LBR can directly recognize the heterochromatin mark H4K20me2 via its amino-terminal Tudor domain (Hirano et al. 2012). Besides nucleosomes, LBR engages in a multitude of other interactions with nuclear partners, including lamin B1 (Ye and Worman 1994), dsDNA (Ye and Worman 1994; Duband-Goulet and Courvalin 2000), and RNA (Chen et al. 2016), which all bind to the RS domain of LBR; HA95 (Martins et al. 2000) and the methylated DNA-binding protein MeCP2 (Guarda et al. 2009), as well as the heterochromatin organizer HP1 that associates with LBR's membrane-proximal globular domain (Ye et al. 1997; Lechner et al. 2005). The nucleoplasmic domain of LBR thereby constitutes a multifaceted molecular backbone for heterochromatin tethering to the NE. In addition, homo-oligomerization of LBR may even contribute to the compaction of heterochromatin at the nuclear periphery (Hirano et al. 2012).

The functional importance of LBR's interaction with chromatin is highlighted by studies on chromatin organization during mammalian development and differentiation. The spatial segregation of transcriptionally repressed heterochromatin to the nuclear periphery of differentiated mouse cells was shown to rely on two molecular assemblies, referred to as the A-type and B-type tethers (Solovei et al. 2013). LBR constitutes the molecular pillar of the B-type tether, whereas the A-type tether consists of lamins A/C and (an) unknown protein(s) of the INM. One of the two tethers is sufficient to maintain heterochromatin at the nuclear periphery. While developing tissues primarily rely on LBR for peripheral heterochromatin tethering, differentiated tissues seem to be more dependent on lamin A/C (Solovei et al. 2013). Remarkably, loss of both LBR and lamin A/C can lead to “nuclear inversion,” an atypical collapse of heterochromatin into the nuclear interior, which is physiologically relevant for light perception in the photoreceptor rod cells of nocturnal animals (Solovei et al. 2013).

Beyond tethering and compacting chromatin at the nuclear periphery, LBR also facilitates transcriptional repression. Myoblast transcriptome analyses, for example, revealed that LBR is involved in suppression of muscle-specific gene expression during early stages of differentiation (Solovei et al. 2013). Furthermore, LBR has been suggested to promote X-chromosome inactivation (XCI) during development in mammals. The association of LBR with the long noncoding RNA Xist contributes to the recruitment of the inactivated X chromosome to the nuclear periphery and was also suggested to ensure its transcriptional silencing (McHugh et al. 2015; Chen et al. 2016). However, in a recent X-chromosome-wide analysis of Xist-mediated silencing, loss of LBR had only a minor effect on silencing (Nesterova et al. 2019). Discrepancies in these observations may be linked to the different silencing assays, different times of Xist induction, as well as allelic versus nonallelic analyses of XCI used in these studies. Thus, although the LBR–Xist interaction might be necessary for positioning of the inactive X chromosome and/or to stabilize gene expression, the exact impact of LBR on XCI still remains to be determined.

LBR–chromatin association may also be relevant for cellular senescence. Initial studies established a link between a reduction of lamin B1 and cellular senescence, suggesting that the nuclear lamina undergoes profound changes as cells enter senescence (Shimi et al. 2011; Freund et al. 2012; Dreesen et al. 2013; Shah et al. 2013; Dou et al. 2015). More recent studies have revealed that LBR is also down-regulated during transition into senescence (Ivanov et al. 2013; Lenain et al. 2015; Lukášová et al. 2017; Arai et al. 2019; En et al. 2020). The cellular transition to senescence is associated with an extensive reorganization of chromatin and changes in gene expression. Induction of senescence through γ-irradiation of cancer cells significantly decreased the levels of LBR and caused relocation of centromeric heterochromatin from the nuclear periphery to the nuclear interior (Lukášová et al. 2017). Upon induction of senescence via replicative or oncogene-induced stress, heterochromatin-rich fragments bud off from nuclei associated with a down-regulation of lamin B1 and LBR (Ivanov et al. 2013; Dou et al. 2015). The significant decrease in the levels of both LBR and lamin B1 in all these studies may imply that the regulation of their levels during senescence is interrelated. Importantly, however, down-regulation of LBR and lamin B1 levels alone is insufficient to induce senescence phenotypes (Dreesen et al. 2013; Lukášová et al. 2017), indicating that reduction of lamin B1 and LBR is a hallmark and the consequence of senescence rather than its cause.

Last, LBR–chromatin interactions have also been implicated in postmitotic NE assembly. This is supported by early observations showing that immunodepletion of LBR reduces chromatin recruitment of the NE vesicles in vitro (Pyrpasopoulou et al. 1996) and that deletion of LBR slightly delays NE assembly in cultured mammalian cells (Anderson et al. 2009).

Compared to the plethora of studies on the chromatin-associated functions of the nucleoplasmic domain of LBR, its carboxy-terminal membrane domain, which shares extensive homology to the human C14 sterol reductase TM7SF2 (Holmer et al. 1998) (also called DHCR14 or SR-1), has remained neglected for a long time. Because most enzymes involved in the cholesterol biosynthesis pathway localize to the ER, such a function seemed unexpected for an INM protein (Schuler et al. 1994). However, human LBR complemented yeast C14 reductase mutants, which provided first evidence for the enzymatic activity of LBR (Silve et al. 1998; Prakash et al. 1999). Recent findings have now unraveled that the C14-sterol reductase activity of LBR is even essential for the viability of mammalian cells, which can explain its relevance to human physiology (Tsai et al. 2016). Two congenital diseases are associated with mutations in LBR: Pelger–Huët anomaly, an autosomal dominant disorder resulting in abnormal hypolobulation of granulocyte nuclei, and Greenberg skeletal dysplasia, an autosomal-recessive condition resulting in abnormal bone development, fetal hydrops, and the ultimate nonviability of the fetus (Turner and Schlieker 2016). Most of these mutations map to the enzymatic membrane domain of LBR and strongly perturb LBR's ability to engage in cholesterol synthesis, albeit through two different mechanisms. Whereas some mutations interfere with LBR's ability to bind the enzymatic cofactor NADPH, others lead to LBR degradation. Importantly, all investigated disease-causing mutants fail to complement the cholesterol auxotrophy imposed by LBR deficiency in cultured mammalian cells, thus establishing LBR as a major sterol reductase required for cholesterol synthesis (Tsai et al. 2016). Peculiarly, the sterol reductase domain of LBR also resembles a conserved domain of isoprenylcysteine carboxyl methyltransferases, which methylate the carboxyl group of prenylated cysteine as a last step of CaaX modification. Such a function for LBR has been speculated to aid the final modification steps of substrates such as prelamin A and/or lamin B (Li et al. 2015), but still awaits supporting evidence.

SUN DOMAIN PROTEINS

Members of the conserved family of SUN (Sad1p, Unc-84)-domain proteins localize at the INM and interact with KASH (Klarsicht/ANC-1/Syne-1 homology)-domain proteins of the ONM in the perinuclear space, forming NE-spanning LINC (linker of nucleoskeleton and cytoskeleton) complexes (Crisp et al. 2006). SUN domain proteins oligomerize into homotrimers with the help of their long intraluminal coiled-coil regions that also force the carboxy-terminal SUN domains into a trimeric arrangement (Sosa et al. 2012; Wang et al. 2012). The formation of LINC complexes relies on the tight intercalation of structurally extended KASH peptides at the SUN domain interfaces, building force-resistant coupling devices within the NE. The extraluminal domains of different SUN and KASH proteins, in turn, engage in a variety of interactions with nuclear components and the cytoskeleton, respectively, conferring the LINC complex with a diverse range of functions. These include nuclear anchorage, nuclear migration, insertion of NPCs or spindle pole bodies into the NE, the coupling of centrosomes to the nucleus, NE membrane spacing, and NE remodeling at the onset of mitosis (Malone et al. 1999; Starr et al. 2001; Crisp et al. 2006; Liu et al. 2007; Hiraoka and Dernburg 2009; Friederichs et al. 2011; Talamas and Hetzer 2011; Gundersen and Worman 2013; Cain et al. 2014; Guilluy et al. 2014; Turgay et al. 2014). As these processes have been covered in elaborate reviews (Starr and Fridolfsson 2010; Gundersen and Worman 2013; Rothballer and Kutay 2013; Rothballer et al. 2013; Burke and Stewart 2014; Chang et al. 2015; Burke 2018; Lee and Burke 2018), we will limit ourselves here to some prominent intranuclear functions of LINC complexes.

From yeast to humans, SUN domain proteins establish specialized links between the INM and chromatin. The best-defined role is the anchorage of telomeres. In Saccharomyces cerevisiae, the SUN domain protein Mps3 mediates telomere tethering to the NE by chromatin silencing factors like Sir4 (Bupp et al. 2007; Horigome et al. 2011) or telomerase subunits like Est1 (Antoniacci et al. 2007; Schober et al. 2009). The association between telomeres and the NE is thought to suppress telomeric transcription and protect telomeres from harmful recombination (Gartenberg 2009; Schober et al. 2009; Mekhail and Moazed 2010).

In addition, the association between telomeres and SUN proteins is of particular importance for meiosis. During meiotic prophase I, rapid prophase movements (RPMs) led by telomeres facilitate the alignment and pairing of homologous chromosomes to ensure their faithful partitioning. Clustering of telomeres requires the concerted effort of LINC complexes, the cytoskeleton, and telomere-associated proteins. In mammalian cells, a specialized LINC complex consisting of SUN1 and KASH5 is required to couple telomeres via cytoplasmic dynein to the microtubule system (Morimoto et al. 2012; Horn et al. 2013; Lee et al. 2015). Telomere association of SUN1 relies on a complex of the membrane-associated junction protein (MAJIN1) and the telomere repeat-binding proteins 1 and 2 (TERB1 and 2) (Shibuya et al. 2014, 2015), and is regulated by cyclin-dependent kinase 2 (CDK2) (Viera et al. 2015). Mice deficient in either SUN1 or KASH5 are infertile due to meiotic arrest, highlighting the physiological relevance of RPMs and subsequent meiotic bouquet formation (Ding et al. 2007; Horn et al. 2013).

Similarly, in Schizosaccharomyces pombe, the SUN and KASH domain proteins Sad1 and Kms1 promote the formation of the meiotic chromosome bouquet. They mediate RPMs through an association of Kms1 with cytoplasmic dynein and of Sad1 with the telomere bouquet proteins Bqt1-4 and Rap1 and Taz1, components of the shelterin complex (Chikashige et al. 2006, 2009; Hiraoka and Dernburg 2009). The function of LINC complexes in bouquet formation is also conserved in the budding yeast S. cerevisiae (Trelles-Sticken et al. 2000; Conrad et al. 2008), in which the telomere-led RPMs are actin dependent (Trelles-Sticken et al. 2005; Koszul et al. 2008), and in Caenorhabditis elegans (Penkner et al. 2007, 2009; Sato et al. 2009; Woglar and Jantsch 2014), with the difference that RPMs rely on the association of specialized chromosome pairing centers with SUN1 (MacQueen et al. 2005). The evolutionary conserved meiotic chromosome movements nicely illustrate how cytoskeletal forces are transmitted via LINC complexes across the NE, here to impart changes in intranuclear organization.

SUN domain proteins have also been prominently implicated in NE-associated DNA repair. The yeast SUN domain protein Mps3, for instance, assists the tethering of persistent DNA double-strand breaks (DSBs) to the NE, potentially providing an environment supportive for alternative repair pathways (Oza et al. 2009; Horigome et al. 2014). Also in higher organisms, several lines of evidence suggest a role for LINC complexes in the DNA damage response (DDR) (Aymard et al. 2017; Marnef et al. 2019). DDR is accomplished either through error-prone nonhomologous end joining (NHEJ) or by a more precise repair pathway involving homologous recombination (HR). Initial studies revealed that mouse SUN1 and SUN2 interact with the DNA-dependent protein kinase (DNAPK), which plays a role in NHEJ repair and thereby assists the process (Lei et al. 2012). A later study then demonstrated that human cells use the DNA repair factor 53BP1, SUN1/2, and dynamic microtubules to promote the mobility of both dysfunctional telomeres and DSBs, potentially to facilitate NHEJ of DSBs (Lottersberger et al. 2015). Whether the mobility of DSBs is facilitated via direct and specific interactions between 53BP1-bound DSBs and the LINC complex or is the result of LINC complex-mediated transduction of forces onto chromatin in an untargeted manner remain interesting possibilities to be explored. On the other hand, the C. elegans SUN domain protein, UNC-84 sequesters components of the NHEJ pathway at the NE while promoting the recruitment of Fanconi anemia nuclease (FAN-1) to the sites of DNA interstrand cross-links. Thereby, UNC-84 has been proposed to guide repair pathway choice through inactivation of NHEJ and promotion of FAN-1-mediated HR at chromosomal breaks (Lawrence et al. 2016).

A surge of recent work reaffirmed the significance of LINC complexes in coupling nuclear mechanics and chromatin-associated processes. First, shear stress applied to the mammalian cell surface was found to stretch chromatin regions and led to transcriptional up-regulation of a transgene inserted in a stretched region, perhaps due to stretch-enabled chromatin opening and binding of RNA polymerase (RNAP) II (Tajik et al. 2016). Importantly, both chromatin displacement and transgene expression were affected upon depletion of SUN proteins, supporting the requirement of SUN proteins in coupling mechanical signals to transcription. Second, LINC complexes were suggested to influence a mechanically induced blockade of adipogenesis in mesenchymal stem cells (MSCs) (Uzer et al. 2018). Here, depletion of SUN1 and SUN2 diminished the nuclear localization of β-catenin, which normally counteracts adipocytic commitment by down-regulation of adipogenic transcription factors (Sen et al. 2008). Finally, extreme mechanical stimulation imposed by a high frequency tensile strain in human MSCs causes a rapid phosphorylation and turnover of SUN2. This contributes to a decoupling of the cytoskeleton and the nucleus, thereby imparting protection from strain-induced DNA damage (Gilbert et al. 2019).

Interestingly, through its association with lamin A/C, human SUN2 is also known to contribute to the maintenance of the latent HIV provirus in a transcriptionally repressed chromatin domain at the NE, thereby promoting the latency of HIV infection (Sun et al. 2018). Yet, SUN proteins may not only modulate viral latency but also affect early steps of HIV infection. The mechanistic details are, however, not well understood (Donahue et al. 2016; Lahaye et al. 2016; Schaller et al. 2017; Luo et al. 2018). Likely, even to date, many other exciting roles of LINC complexes remain to be discovered.

LEM-DOMAIN PROTEINS

LEM-domain proteins, named after the founding family members LAP2, emerin, and MAN1, are defined by the presence of a bihelical motif called the LEM domain (Lin et al. 2000). This domain confers interaction with the widespread metazoan DNA-binding protein, the barrier to autointegration factor (BAF), which has been implicated in important processes such as gene regulation, chromatin condensation, and nuclear assembly (Margalit et al. 2007; Samwer et al. 2017). Whereas BAF and LEM domains are restricted to metazoans, some INM proteins in both lower and higher eukaryotes possess a structurally related helix-extension-helix fold (Heh) that confers direct DNA interaction.

Metazoan LEM-domain proteins strongly associate with the nuclear lamina (Sullivan et al. 1999; Östlund et al. 2006; Ulbert et al. 2006) and are organizers of a complex interaction network at the NE–lamina–chromatin interface. In this network, LEM-domain proteins functionally overlap in recruiting chromatin-modifying proteins such as the histone deacetylase HDAC3 (Somech et al. 2005; Demmerle et al. 2012), and transcriptional regulators such as germ cell-less (GCL) and Btf (Holaska et al. 2003; Haraguchi et al. 2004; Mansharamani and Wilson 2005) (see also Fig. 1). Thereby, they coordinate the peripheral immobilization and repression of genomic elements, and the inhibition of transcription (Nili et al. 2001; Brachner and Foisner 2011; Ho et al. 2013; Guilluy et al. 2014; Lee et al. 2017). However, individually, they also perform some specialized functions, which will be discussed in the following sections.

LAP2

LAP2 proteins comprise at least six distinct protein isoforms (α, β, γ, δ, ε, ζ) that are generated by alternative splicing in mammals (Dechat et al. 2000). The most abundant among them are LAP2α and LAP2β, both of which interact with lamins and chromatin. LAP2α diverges significantly from LAP2β and the other LAP2 isoforms, as it is a nucleoplasmic protein, whereas the others are membrane embedded. LAP2β, the longest membrane-bound isoform, possesses an extended nucleoplasmic domain composed of an amino-terminal LEM-like domain through which it interacts with DNA, followed by the LEM domain that interacts with BAF, a low complexity region encompassing a lamina-binding domain, a single transmembrane segment, and a short luminal domain (Fig. 1). Early studies demonstrated a role for LAP2β in postmitotic targeting of membranes to chromatin and postmitotic nuclear growth (Foisner and Gerace 1993; Yang et al. 1997; Gant et al. 1999; Anderson et al. 2009).

In interphase cells, Lap2β plays an important role in fundamental cellular processes such as transcription and DNA replication. First, LAP2β contributes to the formation of a repressive chromatin environment at the INM (Cutler et al. 2019). Both on its own and in complex with the transcriptional repressor GCL, it represses the transcriptional activity of the E2F-DP3 heterodimer (Nili et al. 2001). Transcriptional repression by LAP2β is further regulated by its interaction with epigenetic modifiers like the histone deacetylase 3 (HDAC3) (Somech et al. 2005). Importantly, the LAP2β-HDAC3 complex also binds the transcriptional repressor cKrox that recognizes GAGA-type DNA elements within lamina-associated chromatin domains, for instance those spanning the developmentally regulated gene loci IgH and Cyp3a in human cells, leading to their peripheral localization and transcriptional silencing (Zullo et al. 2012). Second, LAP2β has also been suggested to control DNA replication by direct association with the chromatin factor HA95 that in turn interacts with Cdc6, an essential component of the prereplication complex (preRC). Abolishing the HA95-LAP2β interaction induces proteasome-mediated degradation of Cdc6 and thereby inhibits the initiation of DNA replication (Martins et al. 2003).

In LAP2α, the transmembrane domain is replaced by a unique coiled-coil domain that confers A-type lamin-binding activity (Dechat et al. 1998; Vlcek et al. 1999). Notably, both LAP2α and the nucleoplasmic pool of lamin A/C associate with euchromatin (Gesson et al. 2016). Furthermore, LAP2α is known to stabilize the tumor suppressor retinoblastoma protein pRb by anchoring it to nucleoplasmic lamin A/C (Markiewicz et al. 2002; Dorner et al. 2006). Thereby LAP2α regulates several functions of pRb such as cell cycle control and terminal differentiation of adipose and muscle tissues (Hansen et al. 2004; Huh et al. 2004). Overexpression of LAP2α in pre-adipocytes promotes cell cycle exit and initiation of differentiation to adipocytes in vitro (Dorner et al. 2006). On the other hand, LAP2α loss impairs pRb function causing inefficient cell cycle arrest in mouse fibroblast cultures and hyperproliferation of epidermal and erythroid progenitor cells in vivo (Naetar et al. 2008).

But do the soluble and membrane-bound LAP2 isoforms have completely independent functions? A recent study has demonstrated that LAP2α and LAP2β use their common LEM-like domain to modulate the activity of GLI1 (Mirza et al. 2019), a zinc finger transcription factor that controls the hedgehog pathway during tumorigenesis (Oro et al. 1997). Acetylated GLI1 is anchored at the INM by LAP2β, creating an inactive but dynamic nuclear reserve of GLI1 at the NE. Nucleoplasmic LAP2α does not only compete for binding of GLI1 but also drives GLI1 activation together with its binding partner HDAC1, which converts GLI1 into the active, deacetylated form on chromatin. Thus, LAP2 proteins form an isoform-based nuclear chaperoning system that controls the balance between inactive GLI1 in the nuclear periphery and active GLI1 in the nucleoplasm, promoted by the release of GLI1 from LAP2β by aPKC to ensure maximal GLI activation. As LAP2 proteins were found to interact quite broadly with zinc-finger proteins via their LEM-like domain, this may suggest a novel and general nuclear scaffolding function toward zinc-finger transcription factors (Mirza et al. 2019).

EMERIN

Emerin is one of the best-studied INM proteins, as mutations in the EMD gene causes X-linked Emery–Dreifuss muscular dystrophy (EDMD) (Bione et al. 1994). It is a tail-anchored membrane protein possessing an amino-terminal LEM domain followed by a so-called regulator-binding domain (RBD), a lamin-binding region, and a second RBD overlapping with an adenomatous polyposis coli-like (APC-L) domain (Berk et al. 2013; Koch and Holaska 2014).

Emerin perfectly epitomizes the degree of biochemical complexity at the NE as it can interact with a plethora of partners and is involved in diverse cellular processes. First, it associates with factors involved in genome organization and regulation. These include epigenetic regulators, proteins involved in signaling, transcription, mRNA splicing, and of course BAF (Holaska and Wilson 2007; Wilson and Foisner 2010; Koch and Holaska 2014). Following the paradigm of formation of repressive chromatin at the NE, emerin epigenetically modulates gene expression by interacting with and regulating the enzymatic activity of HDAC3, which forms the catalytic subunit of nuclear corepressor complex (NCoR). Emerin-null fibroblasts exhibit epigenetic changes that strikingly resemble those of HDAC3 knockout (KO) cells. These include a global increase in H4K5 acetylation and decreased H3K27 and H3K9 trimethylation (Demmerle et al. 2012). Besides epigenetic modulation, emerin influences gene expression by scaffolding a variety of gene-regulatory partners at the INM through its RBDs and APC-L domain. Prominent among them are GCL (Holaska et al. 2003), Btf (Haraguchi et al. 2004), Lim domain only (Lmo7) (Holaska et al. 2006), β-catenin (Markiewicz et al. 2006), and the splicing factor YT521-B (Wilkinson et al. 2003). Through these interactions, emerin regulates fundamental processes such as cell cycle progression, apoptosis, myogenic differentiation, and mRNA splicing. Not surprisingly, due to its involvement in several gene regulatory networks, emerin also has an established role in signaling. A large number of genes regulated by the transforming growth factor β (TGF-β), Notch, JNK, MAPK, NF-κB, integrin, and IGF pathways, are misexpressed upon loss of emerin (Muchir et al. 2007; Koch and Holaska 2012; Berk et al. 2013). Interestingly, most of these signaling pathways regulate myogenic differentiation—an intriguing aspect given the involvement of emerin in EDMD (Massague et al. 1986; Bione et al. 1994; Polesskaya et al. 2003).

Second, emerin associates with a variety of structural components such as lamin A/C (Clements et al. 2000), the LINC complex (Mislow et al. 2002; Haque et al. 2010), actin (Holaska et al. 2004), myosin (Holaska and Wilson 2007), and spectrin (Holaska and Wilson 2007). Owing to these interactions, emerin is perfectly positioned to integrate mechanical impetuses at the INM. Emerin-null mouse embryonic fibroblasts display alterations in nuclear morphology, NE plasticity, and response to mechanical stimulation as well as an impaired viability under mechanical strain (Lammerding et al. 2005; Rowat et al. 2006). This suggests that emerin may help cells to adapt to mechanical load. Indeed, upon mechanical stimulation, emerin gets phosphorylated at tyrosine residues 74 and 95. This strengthens the interaction between the LINC complex and lamin A/C, and initiates actin bundle formation at the nuclear periphery, which may provide structural rigidity to the nucleus (Guilluy et al. 2014). Additionally, in response to the mechanically induced increase in actin polymerization, emerin regulates nuclear accumulation and activity of mechanosensitive transcription factors such as MLK1-SRF, and thus stimulates mechanosensitive gene expression (Olson and Nordheim 2010; Ho et al. 2013; Willer and Carroll 2017). Interestingly, the function of emerin as a mechanosensor also impinges on the positioning of chromosome territories. Growing cells on softer matrices leads to phosphorylation of emerin at tyrosine residue 99, a spatial repositioning of chromatin domains to the nuclear interior and transcriptional deregulation of associated genes, albeit to different extents (Pradhan et al. 2018).

Further investigations into how perturbation of signaling pathways and mechanosensing influence muscle regeneration and cardiac conduction may prove beneficial for the development of EDMD treatments.

MAN1

MAN1/LEMD3 is the longest of all LEM family members. It is anchored at the INM by two transmembrane segments encompassing a 130-residues-long protein segment that is located in the perinuclear space. The LEM domain of MAN1 resides at the amino terminus of its first, long nucleoplasmic domain, whereas the shorter carboxy-terminal nucleoplasmic domain comprises a DNA-binding winged-helix domain, also known as the Man1-Src1p carboxy-terminal (MSC) domain, and a carboxy-terminal RNA recognition motif (RRM) (Lin et al. 2000; Caputo et al. 2006).

MAN1 has been shown to interact with lamin A/C, BAF, R-SMADs, and PPM1A (Smad2/3 phosphatase). Much of the cellular function of MAN1 is attributed to the interaction of its RRM with R-SMADs (Osada et al. 2003; Raju et al. 2003; Hellemans et al. 2004; Lin et al. 2005; Pan et al. 2005; Cohen et al. 2007). R-SMAD proteins are key regulators of multiple signaling pathway. Two types of R-SMADs exist in mammals: TGF-β-responsive (SMAD2 and SMAD3) and bone morphogenic protein (BMP)-responsive (SMAD1, SMAD5, and SMAD8) Smads. Interaction of the carboxyl terminus of MAN1 with SMAD2/3 sequesters them at the NE and competes for their binding to transcriptional activator complexes, thereby regulating expression of TGF-β target genes. Indeed, MAN1 has been shown to antagonize TGF-β signaling in mammalian cells (Lin et al. 2005; Pan et al. 2005; Chambers et al. 2018) and mice embryos (Ishimura et al. 2006; Cohen et al. 2007). Inactivation of MAN1 in mouse embryos is associated with increased transcriptional activity of SMAD2/3 and an increased expression of TGF-β, causing perturbations in vascular remodeling and angiogenesis, which leads to embryonic lethality (Ishimura et al. 2006; Cohen et al. 2007). Similarly, MAN1 also negatively regulates the BMP signaling pathway via interactions with SMADs 1, 5, and 8. In Drosophila, MAN1-mediated inhibition of the BMP pathway has been shown to ensure proper synaptic growth and the integrity of neuromuscular junctions, which is required for proper locomotor activity (Wagner et al. 2010; Laugks et al. 2017).

Notably, MAN1 has also been found to activate the promoter of BMAL1, which is one of the small numbers of “clock genes” that are responsible for generating the internal circadian rhythm. Consistently, depletion of MAN1 resulted in a prolonged circadian period, whereas its overexpression led to a reduced period length in mammalian cells and Drosophila (Lin et al. 2014). This study brings to light two interesting findings: first, it presents the possibility that INM proteins can also affect gene transcription in an activating manner. Second, it reveals an unexpected function of an INM protein in determining the circadian rhythm.

Beyond transcriptional control, MAN1 also influences the dynamic organization of the NE during the cell cycle in various organisms. In the fission yeast Schizosaccharomyces japonicus, the NE partially ruptures during late anaphase and reseals following mitotic exit. During anaphase, NPCs redistribute toward the spindle poles where they cosegregate with chromatin prior to NE breakage. Here, Man1 has been implicated in connecting segregating chromatin to NPCs to ensure equal partitioning of the nucleus and NPCs between daughters. Cells lacking Man1 exhibit a failure in anaphase NPC distribution, evident by an irregular clustering of NPCs between segregated chromosomes in daughter nuclei (Yam et al. 2013). In S. pombe, Man1 collaborates with Lem2 to perform some key functions of the missing nuclear lamina, including the maintenance of NE structure and stability as well as the anchorage of telomeres at the nuclear periphery (Gonzalez et al. 2012). Finally, MAN1 functionally cooperates with other LEM-domain proteins also in C. elegans, where Ce-MAN1 and Ce-emerin ensure proper chromosome segregation and cell division (Liu et al. 2003).

LEM2

Similar to MAN1, LEM2 is anchored in the INM by two transmembrane segments and possesses two nucleoplasmic domains (Fig. 1), which comprise an amino-terminal LEM domain and a winged-helix MSC domain, respectively. Both domains have been proposed to mediate interactions with DNA and/or chromatin. The LEM2 homolog of fission yeast is part of a network of silencing factors at the nuclear periphery that collectively ensure perinuclear heterochromatin repression (Banday et al. 2016; Barrales et al. 2016). S. pombe Lem2 mediates centromere tethering via its LEM domain, whereas heterochromatin silencing and the anchorage of telomeres requires its MSC domain. For transcriptional silencing, Lem2 cooperates with other factors, such as the RNAi machinery and the telomere-associated protein Taz1 (Barrales et al. 2016). Additionally, Lem2, via its MSC domain, promotes the binding of the Snf2-like/HDAC repressor complex (SHREC) to chromatin, which in turn suppresses the recruitment of the anti-silencing factor Epe1 (Banday et al. 2016; Barrales et al. 2016). This function seems key to the role of Lem2 in silencing of telomeres, as deletion of epe1⁺ completely suppresses the respective defect of a lem2Δ mutant. Lem2 has also been implicated in the maintenance of genome stability in S. cerevisiae, where the Lem2 homolog Heh1 along with its interacting partner Nur1 forms the “chromosome linkage INM protein” (CLIP) complex that physically links rDNA repeats to the nuclear periphery (Mekhail et al. 2008). Deletion of either heh1 or nur1 causes release of rDNA repeats from the NE and leads to chromosome instability by promoting aberrant recombination events in the rDNA repeats.

A suite of recent studies in both yeast and mammalian cells have placed LEM2 into a new spotlight for its role in the recruitment of the ESCRT-III complex to the NE. Here, ESCRT-III supports the constriction of tubular membrane structures culminating in membrane fission at different occasions, in line with the general role of ESCRT-III in membrane scission events (Henne et al. 2013; McCullough et al. 2018). NE-associated functions of ESCRT-III include the surveillance of NPC assembly (Webster et al. 2014), postmitotic NE closure (Olmos et al. 2015; Vietri et al. 2015), repair of NE ruptures (Denais et al. 2016; Raab et al. 2016), and the remodeling of heterochromatin-INM contacts (Pieper et al. 2020). Recruitment of the ESCRT-III complex to its site of action depends on specific adaptor proteins, and LEM2 has emerged as one player that promotes the recruitment of ESCRT-III to a specific site at the NE.

The role of LEM2/Lem2 in directing the ESCRT-III complex to the NE initially emerged from studies on NPC assembly in yeast (Webster et al. 2014). The inhibition of NPC assembly in S. cerevisiae leads to formation of an NE subdomain named “storage of improperly assembled NPCs” (SINC) compartment, which is enriched in NPCs that may represent improper assembly intermediates. Heh1, the yeast homolog of LEM2, recruits Chm7 (the CHMP7 homolog) to these sites and promotes SINC formation (Webster et al. 2016). Recent work has now even revealed that the Heh1-Chm7 axis does not only support the surveillance of NPC formation but also the sealing of NE ruptures. Heh1 and Chm7 are usually physically separated on opposite sides of the nuclear border. However, damage of the NE leads to the local encounter of Heh1 and Chm7 at the site of perturbation. Heh1 then activates the membrane shaping function of Chm7, leading to rapid repair of the nuclear border (Thaller et al. 2019). However, there seems to be an additional backup mechanism for the repair of NE ruptures in cultured mammalian cells where repair of laser-induced NE ruptures can occur independently of Chmp7 but requires recruitment of LEM-domain proteins including LEM2 to chromatin through BAF (Halfmann et al. 2019).

LEM2 does not only cooperate with ESCRT-III in interphase cells, but also during mitosis. In mammalian cells, postmitotic NE sealing depends on ESCRT-III, which is recruited to sites where spindle microtubules penetrate the reforming NE during anaphase (Olmos et al. 2015; Vietri et al. 2015). ESCRT-III recruitment is mediated by its component CHMP7, which interacts with the MSC domain of LEM2 (Gu et al. 2017). Initially, LEM2 tethers membranes to postmitotic chromatin disks by binding of its LEM domain to BAF. At the same time, a low-complexity domain within LEM2 undergoes liquid–liquid phase separation to coat bundles of spindle microtubules that need to be cleared. Finally, the MSC domain of LEM2 activates CHMP7. These two proteins copolymerize around microtubule bundles to form a molecular O-ring that initiates recruitment of spastin and ESCRT-III to severe the microtubules and promote nuclear compartmentalization, respectively (von Appen et al. 2020).

The contribution of LEM2 in establishing nuclear compartmentalization is also evident in S. japonicus that undergoes semi-open mitosis during which the NE is broken and then resealed at a single site. For NE sealing, the intersecting spindle becomes tightly enwrapped by the nuclear membrane, and both Lem2 and its interacting partner Nur1 are enriched at these sites called NE “tails.” However, the accumulation of LEM2 at these “tails” demands the release of a pool of LEM2 from their interaction with pericentromeric and telomeric heterochromatin. This is mediated by the ESCRT-III-associated AAA⁺-ATPase Vps4, which ensures that the association between LEM2 and peripheral heterochromatin is transient and subject to continuous remodeling. In the absence of Vps4, the interactions between Lem2 and heterochromatin are locked, causing a defect in bulk release of chromosomes from the NE at mitotic entry and a failure in the reestablishment of nuclear compartmentalization during mitotic exit (Yam et al. 2011; Pieper et al. 2020). Taken together, LEM2/Lem2 safeguards the integrity of the cell nucleus in response to numerous insults that are associated with a lack of proper nuclear compartmentalization.

LEM2 might also protect cells from DNA damage by interacting with several components of the nucleotide excision repair (NER) machinery. Indeed, cells depleted of LEM2 show increased sensitivity to UV-induced DNA damage and increased phosphorylated γ-H2AX protein levels, indicating an impaired or delayed DNA damage response (Moser et al. 2020). Furthermore, LEM2 plays a role in regulating MAP kinase and AKT signaling pathway during embryonic development in mice (Tapia et al. 2015) and acts together with emerin to regulate the ERK signaling during myoblast differentiation (Huber et al. 2009). And, last but not least, S. pombe Lem2 was suggested to restrict changes in nuclear size imposed by alterations in membrane synthesis and nucleocytoplasmic transport, thereby contributing to the maintenance of a constant nucleus to cell volume ratio, perhaps by acting as a barrier to membrane flow into and out of the NE (Kume et al. 2019).

NET5 ALIAS SAMP1

SAMP1 is a polytopic membrane protein with four central transmembrane segments that is well conserved from yeast to human. The amino-terminal nucleoplasmic domain of SAMP1 contains two conserved zinc-finger domains that are required for NE localization (Gudise et al. 2011). From a functional perspective, the zinc-fingers might be used for chromatin interaction. In line with this, Ima1, the SAMP1 homolog in S. pombe, attaches centromeric heterochromatin to the nuclear periphery (King et al. 2008). Likewise, human SAMP1 has been suggested to contribute to the organization of peripheral heterochromatin in U2OS cells (Bergqvist et al. 2019) and its overexpression was shown to enhance the peripheral localization of chromosome 5 (Zuleger et al. 2013).

Originally, mammalian SAMP1 had been linked to mitotic functions (Buch et al. 2009; Larsson et al. 2018). During mitosis, it localizes to membranes in the vicinity of the spindle poles. The depletion of SAMP1 was associated with some spindle defects, prolonged mitosis, and chromosome mis-segregation. Because SAMP1 interacts with γ-tubulin and HAUS6, a subunit of the augmin complex involved in microtubule nucleation, it has been suggested that SAMP1 may promote the recruitment of the augmin complex and γ-tubulin to the mitotic spindle (Larsson et al. 2018).

Other interaction partners of SAMP1 include the LINC complex, emerin, and lamin A/C (Gudise et al. 2011). The association of SAMP1 with the LINC complex plays an important role in nuclear movement during fibroblast polarization and centrosome positioning (Gudise et al. 2011; Borrego-Pinto et al. 2012). Similar to the defects observed upon depletion of lamin A and emerin, knockdown of SAMP1 perturbs the differentiation of myoblasts (Jafferali et al. 2017; Le Thanh et al. 2017). Supporting the role of SAMP1 in early differentiation, recent studies have revealed that SAMP1 levels increase in differentiating induced pluripotent stem cells (iPSCs) concomitantly with those of lamin A/C and ectopic expression of SAMP1 induces a rapid differentiation of iPSCs even under pluripotent culturing conditions. Whereas the idea that an INM protein can drive differentiation of iPSCs is interesting, mechanistic insights into how Samp1 may regulate this process are lacking and require further investigation (Bergqvist et al. 2017).

LAP1

LAP1/TOR1AIP1 is a lamina-binding type II membrane protein that is restricted to metazoans. It exists in three isoforms (A, B, C) in rodents (Foisner and Gerace 1993), of which only two, LAP1B and LAP1C, are expressed in human cells (Santos et al. 2014). Whereas LAP1 is a bona fide component of the INM, its paralog LULL1 is distributed in the peripheral ER. Both possess a well-conserved luminal domain that adopts a RecA-like fold and is used for the activation of Torsin AAA⁺ ATPase family members, first and foremost Torsin-1A and Torsin-1B (Goodchild and Dauer 2005; Jungwirth et al. 2010; Kim et al. 2010; Zhao et al. 2013; Brown et al. 2014; Sosa et al. 2014). Torsins reside in the perinuclear space and the ER lumen; however, their precise molecular function remains elusive to date. Torsin activation by LAP1 and LULL1 relies on an active site complementation mechanism in which a critical arginine finger complements the active site of Torsins (Brown et al. 2014; Sosa et al. 2014; Rose et al. 2015). Interestingly, Torsin-1A, the predominant Torsin isoform in the brain, is prominently associated with early onset torsion dystonia (DYT1), a severe movement disorder (Ozelius et al. 1997). Similarly, mutations in the TOR1AIP1 gene have been linked to primary dystonia (Table 2; Dorboz et al. 2014; Rebelo et al. 2015), and the ablation of Lap1 phenocopies both the nuclear blebbing and perinatal lethality observed in Tor1A KO mice (Kim et al. 2010).

In comparison, the nucleoplasmic domain of LAP1, which interacts with lamins A/C and B1 (Foisner and Gerace 1993; Senior and Gerace 1988) and emerin (Shin et al. 2013), has remained less well studied. Functional investigations have suggested that LAP1 collaborates with emerin in the maintenance of skeletal muscle cells, supported by the observation that combined loss of LAP1 and emerin causes a significantly more pronounced myopathy than that observed upon loss of the individual proteins (Shin et al. 2013). Some recent work uncovered that the nucleoplasmic domain of LAP1 also directly interacts with chromatin, similar to other INM proteins. Surprisingly, however, it was observed that LAP1 cannot be released from chromatin during mitosis if Torsin functionality is compromised or LAP1 is overexpressed, resulting in chromosome segregation defects and binucleation (Luithle et al. 2020), revealing an unexpected function of Torsins in the ER lumen in modulating chromatin association of LAP1 in the nucleoplasm.

Future work aimed at characterizing the molecular function of LAP1, LULL1, and the enigmatic Torsin family of proteins will hopefully allow rationalizing the cellular and organismal phenotypes associated with their dysfunction.

CONCLUDING REMARKS

In the past years, a wealth of research has pushed our perception of the NE from a mere shield for the genome to a dynamic biochemical factory that supports a multitude of essential cellular processes. Several converging themes emerge from the functional characterization of INM proteins. Principally, most of the INM proteins can be appreciated for their association with chromatin and for their role in the establishment of a repressive environment for gene expression. Yet, we still miss the full picture of the mechanisms governing the formation, organization, and functionality of the peripheral heterochromatin vis-à-vis INM proteins. That being said, there have been many fresh insights into important roles of INM proteins like lipid synthesis, mechanosensing, and functional collaborations with the ESCRT-III and Torsin machineries. Understanding the molecular principles governing these functions remains an exciting challenge for the future. Finally, a large number of INM proteins, especially those expressed in a tissue-specific manner, remain uncharacterized. This demands attention in light of nuclear envelopathies, many of which are characterized by tissue-specific defects. We emphasize that convergence of fundamental research driven toward unraveling the mechanistic basis of INM protein functions with translational research directed toward understanding disease etiology will provide a great promise toward developing future therapeutic strategies.