Nuclear Pore Complexes

Macromolecules enter and exit the nucleus by trafficking through NPCs. The complexity of these large, 40–60 MDa protein complexes has for decades limited their detailed structural analysis. More recently, however, structural biologists have taken advantage of the intrinsic modularity of the NPC, and are making progress toward deciphering this nanomachine in detail. Structures of several of the stable building blocks of the NPC have been solved, and several models for their assembly have been proposed (Brohawn et al. 2009; Elad et al. 2009). To approximate the molecular architecture of the yeast NPC, a diverse set of biophysical and proteomic data was collected and used to model localization of the NPC's 456 constituent proteins. The proposed structure shows that roughly half of the NPC is composed of a core scaffold, with structural similarity to vesicle-coating complexes. This scaffold coats the curved surface of the nuclear envelope membrane. The selective barrier for transport is formed by disordered repeat regions of proteins (FG-repeats) that line the inner face of the scaffold. Accordingly, the NPC consists of only a few structural modules and is organized in a 16-fold repetition of ‘columns’ (Alber et al. 2007).

The FG-nucleoporins (FG-Nups) coordinate and potentially regulate translocation through the NPC The extensive repeats of phenylalanine-glycine (FG) in each FG-Nup directly bind to shuttling transport receptors moving through the NPC. The exact mechanism of the path through the pore and its biochemistry and biophysics have not been clarified, but much progress has been made recently to refine our understanding of the roles of the FG-Nups. Several elegant approaches show that selective pores can be reconstituted from defined components in vitro and that the selectivity is largely due to FG-Nups that form hydrogels that reseal upon translocation of molecules through the pore (Frey and Gorlich 2007; Frey et al. 2006; Jovanovic-Talisman et al. 2008). For a recent review on the different models of nuclear pore transversion, see Walde and Kehlenbach (2010).

Until recently, NPCs were viewed as relatively static structures, embedded in the NE and controlling the molecular trafficking between nucleus and cytoplasm. Now, it has become evident that nuclear pore complexes are instead highly dynamic and are involved in diverse cellular processes, which range from the organization of the cytoskeleton to gene expression. Enzymatic activities at the NPC regulate nucleocytoplasmic transport and use the NPC as a regulatory scaffold, while nucleoporins may regulate gene expression by contacting chromatin. Thus, discriminating between the effects of individual nucleoporins on transport, scaffolding, or gene expression is a major challenge in understanding the role of the NPC (D'Angelo and Hetzer 2008; Xylourgidis and Fornerod 2009).

Plant Nuclear Pore Structure

For many years, plant biologists have relied on images of yeast and vertebrate NPCs informing about the high-resolution organization of the NPC. Reviews about the plant NPC tended to cite the early study by Roberts and Northcote (Roberts and Northcote 1970) as the most recent high-resolution ultrastructure work on the plant NPC This situation has now changed with the 2009 publication by Fiserova et al. (Fiserova et al. 2009) providing an in-depth view of the tobacco BY2 cell and onion root cell NE and NPC structure and organization. The authors used in-lens feSEM (field emission electron scanning microscopy) to visualize the NPC from both the nuclear and the cytoplasmic surface of the NE, and have compared NPC ultrastructure in quiescent and dividing BY2 cells. They show that the plant NPC closely resembles the known yeast and vertebrate NPCs and that its diameter falls somewhat between those (plant ∼105 nm, Xenopus laevis ∼110–120 nm, yeast ∼95 nm). Plant NPCs appear to be surprisingly densely spaced (∼ 50 NPCs per µm²compared to 60 NPCs per µm² for X. laevis oocytes, considered very rich in NPCs). Interestingly, the NPCs are not randomly distributed but rather aligned in rows, similar to other higher eukaryotes, but different from yeast (Belgareh and Doye 1997; Maeshima et al. 2006).

Plant Nucleoporins

In earlier studies, several candidate plant nucleoporins (Nups) were described. A 100 kD carrot nuclear matrix protein associates closely with the NPC, and is recognized by antibodies against mammalian and yeast nucleoporins (Scofield et al. 1992), but the identity of this protein remains unknown. Eight proteins associated with the tobacco NE were identified through their N-acetylglucosamine (GlcNAc) modification (Heese-Peck et al. 1995). The association of these proteins with the plant NPC and their potential function have not been tested. More recently, several proteins with significant similarity to animal and yeast nucleoporins have been identified in forward genetic screens for diverse pathways. In addition, reverse genetic approaches with Nup homologs have revealed impairment of flowering-time regulation and overall plant development. Some of the phenotypes are shared with mutants in components of the protein import and RNA export machinery (Table 1).

Putative Arabidopsis Nup160 and Nup96 were identified as suppressors of auxin-resistant 1 (axr1), SUPRESSOR OF AXR1 1 and 3 (SAR1 and SAR 3) (Parry et al. 2006). In sar1 and sar3, the morphological and molecular phenotypes of axr1 are partially restored. Among other defects in nucleocytoplasmic trafficking found in sar1 and sar3, nuclear import of the repressor IAA17 was affected. This suggests that reduced nuclear import of transcriptional Aux/IAA repressors might counteract their over-accumulation based on reduced proteolysis in the axr1 background. A defect in Nup160/SAR1 was also found to sensitize plants to chilling stress and disrupt acquired freezing tolerance (Dong et al. 2006).

R proteins are intracellular immune sensors, typically of the NB (nucleotide binding)-LRR (leucine-rich repeat) type, that act as regulatory signal transduction switches upon sensing isolate-specific pathogen effectors. A gain-of-function mutation in the Arabidopsis SNC1 R gene leads to constitutive SNC1 activity and constitutive defense response. Suppressors of snc1 (modifiers of snc1, mos) include MOS3 (Nup96), which is identical to SAR3 (Zhang and Li 2005), MOS6 (Importin alpha 3) (Palma et al. 2005) and MOS7, (Nup88) (Cheng et al. 2009).

Table 1.

Categories of characterized plant NE/NPC-associated proteins

(continued)

Two loss-of-function alleles of putative Nup133 and Nup85 affect nodulation and mycorrhiza in Lotus japonicus (Kanamori et al. 2006; Saito et al. 2007). The two Lotus Nups are required for nucleus-associated calcium spiking during symbiotic signal transduction, suggesting a role in transducing a calcium signal at the nuclear pore. Calcium-mediated structural changes have been shown at the X. laevis NPC and a function for the NPC as a gated ion channel has been proposed (Bustamante 2006; Stoffler et al. 1999). Currently, the molecular mechanism underlying the Nup-dependent calcium spiking in Lotus is not known.

An additional Nup mutant with an extremely early-flowering phenotype was identified in a reverse genetic study targeting Arabidopsis nuclear long coiled-coil proteins (Xu et al. 2007b). NUA (Nuclear Pore Anchor) is a long coiled-coil protein located at the inner nuclear envelope, consistent with the association with the inner nuclear basket of its human and its yeast homologs Tpr (Translocated Promoter Region) and Mlp1/Mlp2 (Myosin-like proteins 1 and 2), respectively (Cordes et al. 1997; Strambio-de-Castillia et al. 1999; Xu et al. 2007b). NUA was also identified in a screen for mutants that suppress the expression of the floral repressor FLC (Flowering Locus C) (Jacob et al. 2007). In yeast, the TREX-2 complex, involved in mRNA export, is associated with the NPC through binding the nucleoporin Nup1. Orthologs of several TREX-2 components, as well as Nup1 have recently been identified in Arabidopsis (Lu et al. 2010).

While the plant nuclear pore proteome still awaits full identification, it is interesting that already now likely plant-specific components have been found. A group of putative plant specific Nups are the members of the WIP and WIT protein families. They share a function of anchoring RanGAP to the NE and a structure with a coiled-coil protein interaction domain adjacent to a C-terminal transmembrane domain (CC-TMD proteins). Both the WIP and the WIT proteins families likely reside at the nuclear pore and have no counterparts outside the land plant lineage (Xu et al. 2007a; Zhao et al. 2008).

The Ran Cycle

The small GTP-binding protein Ran is a crucial component of nuclear import and export. GTPase activating protein (RanGAP) and Ran binding proteins 1 and 2 (RanBP1 and RanBP2) are localized to the cytoplasmic side of the NPC, whereas the nucleotide exchange factor for Ran, named RCC1, is localized inside the nucleus. These localizations are thought to establish a gradient of high RanGDP in the cytoplasm and high RanGTP in the nucleus, which determines the directionality of nucleocytoplasmic transport. RanGTP dissociates imported cargo from import receptors, but stabilizes complexes between export receptors and their cargo. After an export receptor/cargo complex reaches the cytoplasm, it in turn is dissociated due to hydrolysis of Ran-bound GTP to GDP ((Stewart 2007) and therein). Ran has been cloned from several plant species and plant Ran proteins suppress the pim1 mutation in Schizosaccharomyces pombe, demonstrating that they are functional (Ach and Gruissem 1994; Merkle et al. 1994). Overexpression of wheat Ran1 in rice and Arabidopsis was shown to cause meristem changes, alterations in mitotic progress, and an altered sensitivity to auxin (Wang et al. 2006). Overexpression of rice Ran2 in rice and Arabidopsis was shown to cause hypersensitivity to salinity and osmotic stress (Zang et al. 2010). The molecular causes of these phenomena and their connection to nucleocytoplasmic trafficking are not known. In addition, Arabidopsis Ran2 has been found in a yeast two-hybrid screen as an interactor of a phragmoplastin-interacting protein and was shown to be localized in the perinuclear region with the highest concentration at the nuclear envelope (Ma et al. 2007). In another yeast two-hybrid screen, Arabidopsis Ran3 was found to interact with a methyl CpG-binding protein, and both proteins were found localized in the nucleus during interphase and primarily around chromatin during mitosis (Yano et al. 2006). Arabidopsis RanBP1a and RanBP1b are about 60% similar to mammalian and yeast RanBP proteins and antisense of Arabidopsis RanBP1c renders transgenic roots hypersensitive to auxin and alters auxin-induced root growth and development by arresting mitotic progress (Cho et al. 2008; Haasen et al. 1999a; Kim et al. 2001; Kim and Roux 2003). Arabidopsis has two copies of RanGAP, and Arabidopsis and Medicago RanGAPs were shown to complement the yeast RanGAP mutant rna1 (Pay et al. 2002). Nicotiana benthamiana and potato RanGAP2 have been found to interact with the NB-LRR resistance protein Rx of Potato Virus X and found required for extreme resistance to Potato virus X, as discussed in more detail below (see The Role of the NPC in Plant Innate Immunity). No plant homologs of RCC1 have been identified yet.

The domain organization of RanGAP differs significantly across kingdoms, with metazoan and plant RanGAPs containing specific, distinct NE targeting domains at their C-terminus and N-terminus, respectively (Rose and Meier 2001). Metazoan Ran-GAPs have a C-terminal, sumoylated, unique domain that binds the nucleoporin RanBP2/Nup358. No RanBP2/Nup358 homolog appears to exist in plants. Plant RanGAPs share a plant-specific N-terminal WPP domain. The WPP motif within the WPP domain is required for the NE targeting. The WPP domain of plant Ran-GAPs has sequence similarity with a small plant protein family named WPP-domain proteins. Other than the WPP-domain that resembles the N-terminus of RanGAP, this group of proteins has no other functionally recognizable domains. WPP-domain proteins are also associated with the NE (Patel et al. 2004).

Two families of proteins were identified that target Arabidopsis RanGAP to the NE. WPP-domain interacting proteins (WIPs) were identified in a yeast two-hybrid screen using Arabidopsis RanGAP1 as bait. They contain a coiled-coil domain and a C-terminal putative transmembrane domain (TMD). This domain structure was named CC-TMD for coiled-coil-transmembrane domain. No WIP-like proteins have been identified from non-plant species (Xu et al. 2007a). Using Arabidopsis WPP1 and WPP2 as baits for tandem-affinity purification coupled with mass spectrometry, a second, related class of CC-TMD proteins was identified (WIT1 and WIT2). Like WIP1, 2a and 3, WIT1 and WIT2 are necessary for RanGAP association with the NE in root tip cells (Zhao et al. 2008). Neither WIP nor WIT proteins appear to be conserved outside the plant lineage, indicating that RanGAP targeting to the NE has evolved at least twice, in plants and animals, utilizing different domains and interaction partners.

Protein Nuclear Import

Proteins to be imported into the nucleus generally contain small peptide sequence motifs recognized by import receptors. The best understood signal is the classical nuclear localization signal (NLS), which typically consists of one or two clusters of basic amino acid residues. In animals and yeast, nuclear import receptors (karyopherins) of the importin β type (also called karyopherin β) are the main nuclear import receptors. While importin β type karyopherins can directly bind to many NLS-containing cargos, often importin α adaptors bind the cargo protein and importin β then binds to importin α. Additional nuclear import signals include the M9 domain (recognized by transportin) and the IBB domain. For a more thorough overview of non-plant karyopherins and nuclear localization signals, see the excellent recent reviews in the field (Lange et al. 2007; Stewart 2007; Wagstaff and Jans 2009).

Homologs of both importin α and importin β have been identified in plants. The Arabidopsis importin α1 homolog binds in vitro to all three types of animal NLS (Hicks and Raikhel 1993; Hicks et al. 1995). Arabidopsis has eight putative importin α genes, seven of which code for proteins containing all functional domains, and 17 importin β-like nuclear transport receptors, whereas the human genome encodes at least 21 importin β-like proteins (Kanneganti et al. 2007; Merkle 2009). Characterized plant importin β-like proteins (including exportins) include SAD2 (RanBP7/8) (Zhao et al. 2007), rice imp β (Jiang et al. 1998), Exportin 2 (Haasen and Merkle 2002), Transportin 1 (Ziemienowicz et al. 2003), Exportin 1 (Haasen et al. 1999b), Exportin T (Hunter et al. 2003) and Exportin 5 (Bollman et al. 2003; Park et al. 2005). For the adaptors, MOS6 (Palma et al. 2005) and IMPα-4 (Bhattacharjee et al. 2008) have been investigated. Interestingly, all plant importin α proteins appear to fall into the imp α1-like clade, while in metazoans, imp α1, α2, and α3 clades exist (Mason et al. 2009). The functional significance of the loss of imp α2 and α3 and the expansion of the α1 clade in plants are currently not known. It will be interesting to investigate potential cargo specificity of the eight Arabidopsis importin α proteins.

One striking difference between plant and animal nuclear import is that permeabilized plant protoplasts cannot be depleted of cytoplasmic factors involved in nuclear import. Also, nuclear transport in plant systems is not inhibited by low temperature, as is nuclear import in HeLa cells (Hicks et al. 1996; Merkle et al. 1996). This experimental difference, which caused plant importins to be analyzed in mammalian in vitro import systems rather than plant systems (Hubner et al. 1999), might point to fundamental differences in the organization or mechanisms of plant nuclear import receptors in vivo. Interestingly, an Arabidopsis importin α co-localizes in the cytoplasm with microtubules and microfilaments, and is redistributed when these cytoskeletal structures are depolymerized (Smith and Raikhel 1998).

Ran-Dependent Nuclear Export

In general, the nuclear protein export machinery appears to be well conserved between species. As mentioned above, the Arabidopsis genome contains 17 genes that code for importin β-like nuclear transport receptors and some of these might function in protein export. The best-characterized protein exporter from Arabidopsis is Exportin 1, the homolog of the vertebrate and yeast nuclear export receptors CRM1/XPO1. There are two genes in Arabidopsis that produce two very similar proteins (XPO1a, XPO1b) with high sequence similarity. Arabidopsis XPO1 interacts with RanGTP and confers nuclear export of proteins that contain a leucine-rich nuclear export signal (NES; Haasen et al. 1999b). The HIV Rev NES is recognized by Arabidopsis XPO1 and confers nuclear export of a plant protein. XPO1a and XPO1b both contain the conserved cysteine residue that is the site of modification by leptomycin B (LMB) (Haasen et al. 1999b; Kudo et al. 1999). If this residue is mutated, LMB sensitivity is lost (Haasen et al. 1999b; Merkle 2009). Exportin 1 genes are essential for the development and function of gametophytes in Arabidopsis (Blanvillain et al. 2008).

The second Arabidopsis exportin that was functionally characterized is Exportin 2, a homolog of CAS (Cellular Apoptosis Susceptibility Protein). CAS is the importin α exporter. Of the eight Arabidopsis importin α homologs, four (Impα1–4) interact with Arabidopsis CAS, which also binds to Ran, indicating functional conservation of this function in plants (Haasen and Merkle 2002).

Two Arabidopsis exportins were identified in forward genetic screens. Exportin-t (Xpo-t), the Arabidopsis ortholog of the export receptor for tRNAs, was identified in genetic screens for mutants in different developmental pathways (Li and Chen 2003), and was designated PAUSED (PSD). Arabidopsis PSD/Xpo-t shares high sequence similarity with its vertebrate and yeast orthologs, interacts with Ran and complements a temperature-sensitive allele of Los1, the gene encoding yeast Xpo-t (Hunter et al. 2003). A putative PAUSED gene was also reported from rice (Yao et al. 2008).

The exportin 5 gene was also identified in a genetic screen for developmental mutants and was designated HASTY (HST), because of the accelerated change from juvenile to adult phase in the mutant (Bollman et al. 2003). Human Exportin 5 is the export receptor for double stranded RNA (dsRNA) molecules including precursor microRNAs (pre-miRNAs). HST binds Ran, binds dsRNA, and is involved in miRNAs nuclear export (Bollman et al. 2003; Park et al. 2005), again indicating conservation in plants.

Apart from these five proteins that were experimentally characterized as exportins, there are three more proteins of the Arabidopsis importin β family that might also act as nuclear export receptors, based on similarity to human proteins (Merkle 2009). One of them is similar to human Exportin 4, the export receptor for eukaryotic translation initiation factor eIF5A (Lipowsky et al. 2000), one to Exportin 7 (Mingot et al. 2004), and one to Importin 13/RanBP13 that also acts as an export receptor for eIF1A (Mingot et al. 2001). So far, none of these proteins have been experimentally characterized (Merkle 2009).

mRNA Nuclear Export

Proteins involved in mRNA export are unrelated to the karyopherins discussed above and to the Ran cycle. In humans, TAP (for Tip-Associated Protein)/NXF1 (for Nuclear RNA Export Factor 1) and p15/NXT1 (for NTF2-like Export Factor 1) form heterodimers to facilitate export of mRNA out of the nucleus (Cullen 2003; Rodriguez et al. 2004). In yeast, the respective proteins are Mex67 and Mtr2 (mRNA Transport Regulator 2). TAP/Mex67 and p15/Mtr2 have NTF2-like domains named after NTF2 (Nuclear Transport Factor 2), the nuclear import receptor for RanGDP (Ribbeck et al. 1998; Smith and Raikhel 1998; Suyama et al. 2000). In contrast to NTF2, p15 does not form homodimers and does not interact with RanGDP, but binds to TAP (Katahira et al. 2002; Wiegand et al. 2002). TAP has an NTF2-like domain for binding p15 and an RMM RNA-binding domain, as well as leucine-rich repeats (LRRs), and the so-called TAP-C domain. This C-terminal domain is, together with the NTF2-like domain, important for binding to FG repeat-containing nuclear pore proteins (Fribourg et al. 2001; Levesque et al. 2006). Directionality of nuclear export of mRNA is not driven by the Ran gradient but the RanGTP gradient is needed for the mRNA export factors (Izaurralde et al. 1997). There is no obvious homolog of TAP in the Arabidopsis genome. Arabidopsis contains three genes encoding NTF2-like proteins, two of which interact with Ran (termed NTF2a and NTF2b). The third one, termed NTL for NTF2-like does not interact with Ran (Zhao et al. 2006b). The function of this protein is uncharacterized to date, and it is a candidate for the Arabidopsis homolog of vertebrate p15/NXT1.

Recently, components of the THO-TREX complex have been identified in plants, indicating that this part of the pathway from transcription to mRNA export appears to be conserved. A mutant of the gene for TEX1, a THO-TREX complex subunit, was found in a screen for siRNA biogenesis, suggesting an involvement in processing or transport of a long RNA precursor (Yelina et al. 2010). In an unrelated screen for suppressors of the erecta locus, the homolog of another THO complex component, Hpr1 was found, and several other THO complex subunits were identified, and found essential in Arabidopsis (Furumizu et al. 2010). A homolog of Thp1, a component of the TREX-2 complex was found in another unrelated screen, and the loss-of-function mutant shows retention of nuclear mRNA (Lu et al. 2010). These new findings indicate that plant TREX-like activities are functioning in RNA processing/export and - as mentioned above - connected to the nuclear pore via AtNUP1.

Proteins binding to the 5′ CAP as well as to the 3′ poly(A) tail have also been implicated in mRNA export, for example the poly(A)-binding protein (PABP) Pab1. Yeast has one and humans have three genes encoding PABPs, while Arabidopsis contains a small gene family of eight members (Belostotsky 2003). Arabidopsis PAB3 binds to RNA and PAB2, 3, and 5 rescue the yeast pab1Δ mutant, suggesting that PABP function in mRNA biogenesis and export is conserved (Chekanova and Belostotsky 2003). Arabidopsis has homologs of the large and the small subunit of the Cap-binding complex (CBC, termed CBC80 and CBC20, respectively) (Kmieciak et al. 2002). Mutations in CBP80 confer abscisic acid (ABA)-hypersensitivity, suggesting that mRNA processing factors act as negative regulators for ABA signaling. Interestingly, another RNA-binding protein is implicated in ABA signaling (Razem et al. 2006), as well as CBC20 (Papp et al. 2004). Proteomic studies of the Arabidopsis nucleolus identified SR proteins as well as components of the exon-junction complex that functions in mRNA export and surveillance (Pendle et al. 2005). In addition, Arabidopsis also contains four homologs of vertebrate ALY/REF proteins (Uhrig et al. 2004), RNA binding proteins that recruit TAP/p15 to the mRNA (Stutz et al. 2000; Stutz and Izaurralde 2003).

Mutants in several nuclear pore-associated proteins in Arabidopsis have defects in mRNA export, as indicated by a strong increase of intranuclear signal in in situ hybridizations with an oligo-dT probe. They include the nuclear basket protein NUA (Tpr in mammals, Mlp1/2 in yeast, Megator in Drosophila), Nup160, Nup96 and AtNUP1 (Lu et al. 2010; Jacob et al. 2007; Parry et al. 2006; Xu et al. 2007b; Zhang and Li 2005). In addition, the cryophyte/los4-2, a temperature-sensitive allele of a DEAD-box RNA helicase, leads to nuclear poly(A) RNA accumulation (Gong et al. 2005). LOS4 is an Arabidopsis homolog of yeast Dbp5, which associates with mRNA early in the nucleus and accompanies it to the cytoplasmic side of the NPC, where it participates in mRNA export. The ATPase activity of Dbp5 is activated by interaction with Gle1, an RNA export factor that is also associated with the cytoplasmic side of the NPC, and inositol polyphosphate IP₆. This activity leads to mRNP re-modeling and probably constitutes a crucial step for dissociation of specific RNP factors to impose directionality on mRNP export (Carmody and Wente 2009; Tran et al. 2007). Whether a similar mechanism also operates in plants is currently not known.

Lamins

A mesh of intermediate filament proteins, the nuclear lamins, lines the mammalian inner nuclear membrane. Lamins bind DNA and chromatin, and are depolymerized during mitosis in response to phosphorylation. Lamins are thought to provide scaffolding structures within interphase nuclei, and to mediate the attachment of chromatin to the nuclear envelope during interphase and chromatin detachment during mitosis (Grant and Wilson 1997). A-type lamins (lamins A and C), encoded by the LMNA gene, are major protein constituents of the mammalian nuclear lamina. The lamina acts as a scaffold for protein complexes that in turn regulate nuclear structure and functions. Interest in these proteins has increased significantly in recent years with the discovery that LMNA mutations cause a variety of human diseases which are collectively termed laminopathies and which include progeroid syndromes and disorders that primarily affect striated muscle, adipose, bone, and neuronal tissues. Recent research effort supports the concept that mammalian lamins and lamina-associated proteins regulate gene expression in health and disease by connecting various signal transduction pathways, transcription factors, and chromatin-associated proteins (Andres and Gonzalez 2009).

Early electron microscopy (EM) studies revealed a structure similar to the vertebrate nuclear lamina in the nuclei of higher plant cells (Galcheva-Gargova and Stateva 1988). Immunohisto-chemical studies suggested the existence of lamin-like proteins in plant nuclei (Li and Roux 1992; McNulty and Saunders 1992; Minguez and Moreno Diaz de la Espina 1993), but no sequence information from the antigens detected in these studies is available, and no lamin-coding genes seem to be present in the fully sequenced plant genomes. Similarly, despite earlier reports of lamin-like proteins in yeast, the fully sequenced Saccharomyces cerevisiae genome also contains no lamin genes (Georgatos et al. 1989; Mewes et al. 1998). It is therefore likely that non-animal eukaryotes have a distinct set of nuclear envelope proteins that functionally replace the lamins.

Somewhat similar, yet significantly larger long coiled-coil proteins do exist in plants and might be candidates for lamin-like proteins. A plant-specific insoluble nuclear protein was identified in carrot (Daucus carota), called Nuclear Matrix Constituent Protein1 (NMCP1), which contains extensive coiled-coil domains and localizes to the nuclear periphery (Masuda et al. 1997). This 134-kD carrot protein fractionated with an insoluble nuclear fraction and localized exclusively to the nuclear periphery. Two NMCP1-related nuclear proteins were described in Arabidopsis, LITTLE NUCLEI1 (LINC1) and LINC2. A mutation in either gene caused reduced nuclear size and changes in nuclear morphology. A double mutant had an additive effect. These results suggest that LINC coiled-coil proteins are important determinants of plant nuclear structure (Dittmer et al. 2007).

Another class of plant proteins with a possible role as nuclear envelope structural proteins are the filament-like plant proteins (FPP1 – FPP7; (Gindullis et al. 2002). Like lamins and the NMCP1 –related proteins, they contain extended coiled-coil domains. Two indirect lines of evidence connect them to the nuclear envelope. First, the tomato homolog LeFPP was identified as a yeast-two hybrid interactor of LeMAF1, a NE-associated protein (Gindullis et al. 1999; Gindullis et al. 2002). Second, a pea protein with similarity to AtFPP3 was identified with an antibody against Lamin B (Blumenthal et al. 2004). Interestingly, the AtFPP family stands out as a plant-specific family of long-coiled coil proteins in a cluster analysis of long coiled-coil proteins of 22 genomes (Rose et al. 2005) and is characterized by four highly conserved sequence motifs of unknown function (Gindullis et al. 2002). The subcellular location of the Arabidopsis proteins is currently not known.

In addition, Fiserova et al. (Fiserova et al. 2009) offer a new view on the plant nuclear lamina. Using feSEM, the authors reveal the meshwork of filaments underlying the inner NE in tobacco BY-2 cells, closely resembling the animal nuclear lamina both in terms of organization and filament thickness. These data should re-ignite the interest in plant lamin-like proteins, especially in light of the numerous new findings from both animals and yeast, indicating that the inner NE has crucial regulatory functions, a role attributed in vertebrates at least in part to the nuclear lamina (Heessen and Fornerod 2007).

Inner Nuclear Envelope Proteins

The number of known inner nuclear envelope (INE) proteins from animals is growing, in part due to the newly awakened interest in such proteins as targets of human genetic diseases (Ellis 2006; Wheeler and Ellis 2008; Worman and Bonne 2007). Such proteins include lamin B receptor (LBR), lamina-associated polypeptide-1 (LAP1), LAP2, emerin, MAN1, otefin, and nurim (Wagner and Krohne 2007). In addition, proteome analyses have added significantly to the list of proteins associated with the nuclear envelope, which are now available for functional investigation (Schirmer and Gerace 2005). Very few of these proteins have an identifiable homolog in the plant databases. These findings, though negative, underscore the possibility that the plant nuclear membrane has a unique protein composition, and that plants evolved unique solutions to nuclear architectural problems such as chromatin organization and nuclear structure.

The characterization of two novel plant INE proteins, AtSUN1 and AtSUN2 has recently been reported (Graumann et al. 2010). The proteins are the Arabidopsis homologs of a group of animal and yeast INE proteins containing a well-conserved SUN (Spindle architecture defective 1/UNC84 homology) domain. In animals, SUN-domain proteins interact in the lumen of the NE with KASH-domain proteins (located at the outer NE) to form protein complexes that connect the nucleus to the cytoplasmic cytoskeleton during interphase. SUN-KASH protein complexes are involved in attaching centrosomes to the nuclear periphery, alignment of homologous chromosomes, and their pairing and recombination in meiosis. They have been implicated in the regulation of apoptosis. the maturation and survival of the germline, nuclear location, and in human diseases such as laminopathies and Emery Dreifuss muscular dystrophy (Burke and Roux 2009; Fridkin et al. 2009; Hiraoka and Dernburg 2009).

Both Arabidopsis SUN proteins share a similar domain layout with their animal counterparts and interact with each other as indicated by fluorescence resonance energy transfer. Confocal microscopy of fluorescent protein fusions and electron microscopy suggest localization at the plant INE. Deletion of either the SUN domain or a nuclear localization signal abolished this localization. It will be interesting to study their in planta functions, and to investigate whether there are plant KASH-domain proteins, or whether SUN-domain proteins have other interaction partners in plants.

Microtubule Nucleation at the Plant Nuclear Envelope

During nuclear division, the genetic material is distributed into two daughter nuclei by the action of the spindle apparatus. Formation of the mitotic spindle originates at microtubule organizing centers (MTOCs), which are the cytoplasmic centrosomes in animal cells and the NE-embedded spindle pole bodies in yeast cells. One of the main differences in cell division between higher plant cells and other eukaryotic cells is the absence of centrosomes or spindle pole bodies. Instead, the plant NE plays an important role in plant microtubule nucleation. A finding that isolated maize nuclei are capable of nucleating microtubules in vitro (Stoppin et al. 1994) gave early evidence for the association of microtubules with the nuclear surface in a plant cell. While the nuclear surface might not be the only site for microtubule nucleation in plants cells, it is especially active in the G2 phase of the cell cycle before NE breakdown (reviewed by Canaday et al. 2000), suggesting that its microtubule-nucleating activity might be under cell-cycle control. Microtubule-associated proteins (MAPs) modulate the nucleating activity of the plant nuclear surface (Stoppin et al. 1996).

Using the microtubule end binding protein AtEB1a fused to green fluorescent protein (GFP), it was shown that at the onset of cell division, a focused pattern of the plus end comets grows away from sites associated with the nuclear periphery. In cells with a preprophase band (PPB), EB1 was found at the nuclear periphery, segregating into two polar caps, perpendicular to the PPB, before nuclear envelope breakdown (NEBD). These polar caps then marked the spindle poles upon NEBD. Interestingly, in cells without PPBs, no prepolarization at the nuclear envelope was seen, and the bipolar spindle only emerged clearly after NEBD. In those cells, both spindle and phragmoplast orientation were more variable, suggesting that the PPB triggers a polarization that is perceived by the early developing spindle poles at the nuclear periphery (Chan et al. 2005).

γ-tubulin is centrally involved in microtubule (MT) nucleation and organization in eukaryotic cells (Wiese and Zheng 2006). Together with five proteins it forms the γ-tubulin ring complex, or TuRC, which binds to and stabilizes MT minus ends (Job et al. 2003). In fungal and animal cells, components of the TuRC concentrate at the MT organizing centers (MTOCs) (Raynaud-Messina and Merdes 2007). γ-tubulin patterns differ somewhat in plant cells, with a punctate distribution pattern on all MT arrays during mitosis (Liu et al. 1995; Liu et al. 1993). However, γ-tubulin along MTs has also been found in spindles and midbodies in animal cells (Lajoie-Mazenc et al. 1994; Luders and Stearns 2007). In contrast, γ-tubulin is highly concentrated in MTOCs like the plastid surface and polar organizers at spindle poles in early land plants (Brown et al. 2004; Shimamura et al. 2004). The large variety of MTOCs in non-vascular plants might inform about the evolution of the typical diffuse acentrosomal plant spindle.

In yeast, the γ-TuRC consist of the spindle pole proteins Spc98p and Spc97p that form a complex with γ-tubulin, which in turn binds to the large coiled-coil protein Spc110p of the spindle pole body at the NE (Job et al. 2003; Knop and Schiebel 1997; Moritz and Agard 2001). Plant homologues of Spc98p have been reported to colocalize with γ-tubulin at the nuclear surface in plant cells (Erhardt et al. 2002). Microtubule nucleation in plant cells depends on the presence of both γ-tubulin and Spc98p, and GFP-labeled plant Spc98p localizes within the nucleus, at the NE and close to the plasma membrane, strengthening the hypothesis of multiple nucleation sites in plant cells (Seltzer et al. 2003).

Vertebrate γ-TuRC contains several proteins not present in yeast cells but conserved in plants. Human 76p, also known as GCP4 (γ-tubulin complex protein 4), was identified as a component of an isolated γ-tubulin complex and localizes to the centrosomes and the spindle poles. It has an orthologue in higher plants and the moss Physcomitrella patens, but appears to be absent in yeast (Fava et al. 1999). Murphy and colleagues reported the finding of the two remaining components of the γ-tubulin complex in mammalian cells, GCP5 and GCP6, and identified the Arabidopsis counterparts of h76p/GCP4 and the GCP5/6 family by sequence comparison (Murphy et al. 2001). It is not understood how GCP4–GCP6 are involved in acentrosomal MT nucleation in plant cells. However, Kong et al. (2010) showed that GCP4 is associated with γ-tubulin in Arabidopsis. Downregulating GCP4 expression caused a reduced γ-tubulin signal at the mitotic spindle and the phragmoplast. MTs did not converge at unified spindle poles, the phragmoplast was disorganized, and cytokinesis was defective, as indicated by the presence of cell wall stubs. These data thus suggest that GCP4 is important for MT nucleation and organization in plant cells, as well (Kong et al. 2010).

Two papers have recently been published that add a novel dimension to the question of plant MTOCs on the NE and bring an unanticipated player into the picture. Mizuno and coworkers (Hotta et al. 2007; Nakayama et al. 2008) have used a biochemical approach to purify and identify a microtubule-organizing activity from tobacco BY2 cell nuclei. Raising an antibody against the MTOC-enriched fraction they identified an isoform of histone H1 and showed that recombinant histone H1 indeed had a microtubule-organizing activity. The same antibody was able to inhibit MT nucleation on isolated nuclei. Interestingly, immunolabeling demonstrated that a fraction of H1 is associated with the nuclear rim, outside of the extension of DAPI-stained chromatin. In a second accompanying paper the group showed that H1 can form ring-shaped complexes with tubulin and that these complexes nucleate and elongate radial microtubules (Hotta et al. 2007). These new data would place histone H1 at the plant NE MTOC and suggest that aspects of its composition radically differ from animal and yeast MTOCs. In addition, this raises the question of where exactly the non-DAPI overlapping H1 is located, if there is any evidence for H1 on the outer surface of the nucleus, and how its location might change at the onset of mitosis and NE breakdown. These are all very exciting questions that can now be further addressed.

Targeting protein for Xklp2 (TPX2) is a key regulator of vertebrate spindle assembly. In vertebrates, both a centrosome-based and a chromosome-based spindle assembly mechanism operate. In the chromosome-based mechanism, the RanGTPase in its GTP-bound form acts as a signal for the location of the chromatin. In the vicinity of a high RanGTP concentration, NLS-containing cargo proteins are released from importin β. TPX2 is activated by release from importin β and induces microtubule nucleation at the kinetochores and around the chromosomes. Activated TPX2 also localizes the mitotic kinase Aurora A to the spindle and activates it. Aurora A, among other functions, activates microtubule nucleation from the centrosomes. In X. laevis oocytes, that operate only with the chromosome-based mechanisms, TPX2 depletion leads to aberrant spindles or block of spindle formation. In HeLa cells, that have centrosomes, TPX2 depletion leads to aberrant centrosomes-based asters that do not connect (e.g., see (Bayliss et al. 2003; Gruss and Vernos 2004; Kawabe et al. 2005).

Vos et al. (Vos et al. 2008) have characterized the role of Arabidopsis TPX2. They show that AtTPX2 is predominantly nuclear during interphase, like vertebrate TPX2, but is actively exported before NE breakdown. AtTPX2 accumulates in the nucleoplasm in interphase and prophase, with a clear elevation at the NE and accumulates in the vicinity of the prospindle upon prospindle formation. AtTPX2 can induce microtubule assembly in X. laevis egg extracts and binds importin α, consistent with its proposed role. Microinjection of anti-TPX2 antibodies into stamen hair cells of Tradescantia virginiana blocks NE breakdown and prospindle formation, consistent with a role of TPX2 in plant prospindle formation.

Rae1/mrnp41 in metazoans, Gle2p in S. cerevisiae, and Rae1 in S. pombe is an mRNA export factor associated with the NPC. It binds polyadenylated mRNA and associates with the nucleoporin Nup98. Rae1 mutants accumulate nuclear poly(A) RNA (Pritchard et al. 1999). In addition, Rae1 has cell-cycle functions and Rae1 mutants are arrested at the G2/M transition (Whalen et al. 1997). Rae1 operates as a mitotic spindle checkpoint component in conjunction with Bub3 and forms a complex with Nup98 and the Cdh1-activated anaphase promoting complex (APC), preventing degradation of Securin before anaphase (Babu et al. 2003; Whalen et al. 1997). Tobacco Rae1 was identified in a functional genomics screen for growth arrest and abnormal leaf development (Lee et al. 2009). Both GFP-tagged and HA-tagged Rae1 was found enriched at the nuclear periphery. Virus-induced gene silencing (VIGS) of Rae1 caused nuclear accumulation of poly(A) RNA. In addition, mitosis was affected, with reduced CDKA activity, reduced expression of cyclin B1, CDKB1-1 and histones H3 and H4. Differentiated leaf cells showed an increased degree of ploidy. Tobacco Rae1 was found associated with mitotic microtubules and capable of binding to microtubules in vitro. Estradiol-inducible RNAi in tobacco BY2 cells affected spindle organization, leading to multipolar spindles and defects in chromosome segregation. Together, these data suggest that plant Rae1 plays a role both in mRNA export and mitosis, too, and that its mitotic role is connected to spindle assembly and function (Lee et al. 2009).

A Connection between the NPC and the Preprophase Band

The plane of cell division in higher plants is defined by the assembly of the preprophase band (PPB), a ring of microtubules and F-actin that appears during G2 phase, and the migration of the premitotic nucleus into the plane defined by the PPB. During mitosis, the site of the former PPB becomes the cortical division zone (CDZ), which remains “marked” in an unknown way and which is thought to guide the phragmoplast and the outwardly growing new plasma membrane. The molecular nature of the CDZ has long been enigmatic, and only recently the first molecular markers have been described (reviewed in Muller et al. 2009).

As discussed above, RanGAP is associated with the outer surface of the NPC in plants and animal interphase cells, where it is involved in the directionality of nucleocytoplasmic transport. During mitosis in vertebrates, the Ran cycle takes on additional roles, like the localized activation of spindle promoting factors such as TPX2. Nothing is currently known about plant Ran cycle functions during cell division. Xu et al. (Xu et al. 2008) have now shown that Arabidopsis RanGAP1 accumulates at the PPB and remains associated with the site of division through mitosis and cytokinesis. RanGAP1 persistence at the CDZ depends on the two mitotic kinesins POK1 and POK2. Inducible depletion of Arabidopsis RanGAP in seedling roots leads to misplaced cell walls similar to other mutants with division plane defects, suggesting an involvement of the Ran cycle in this process.

During animal mitosis and cytokinesis, the general theme is that a high local concentration of RanGTP promotes microtubule growth. This is based on the release from inhibition of NLS-containing spindle assembly factors (such as TPX2). Thus, RanGAP could assist in the disassembly of the PPB by keeping local RanGTP levels low and thereby favoring depolymerization over polymerization of microtubules at this site. In addition to microtubule stability, RanGTP has been shown to affect the polarity of microtubule motor activities. Wilde et al. (Wilde et al. 2001) showed that RanGTP increases plus end-directed motor activity and decreases minus-end directed activity. During plant cytokinesis, the microtubule arrays of the phragmoplast are oriented with the plus ends toward the growing cell plate and deliver vesicles presumably by plus end-directed motor activity (Jurgens 2005). RanGAP1 might therefore cause reduced plus end growth of the phragmoplast microtubules and/or reduced plus-end motor activity. Such a regulatory function could conceivably fine-tune phragmoplast and vesicle delivery dynamics and thereby contribute to the overall precision of the processes underlying cell plate synthesis and positioning.

Secondary Messengers at the NE

Aside from a cytoplasmic source of Ca²⁺ ions that enters the nucleus through the nuclear pore complexes, the nuclear envelope itself represents a depot of Ca²⁺ that can be released into the nucleus. There, it controls various processes, such as gene expression, phosphorylation, dephosphorylation, to name a few (Xiong et al. 2006). In animals, changes in free Ca²⁺ concentration are enabled by Ca²⁺ channels that traverse the nuclear envelope and allow mobilization of accumulated Ca²⁺ from the NE lumen and replenish the Ca²⁺ store within the same compartment (Bkaily et al. 2006; Gerasimenko and Gerasimenko 2004). The channels are activated by secondary messengers such as inositol-1,4,5-triphosphate (IP₃), which is generated by the activity of phosphoinositide-dependent phospholipase C (PI-PLC) (Gerasimenko and Gerasimenko 2004).

In plants, there is both electrophysiological and localization evidence that place Ca²⁺ pumps or channels at the NE (Bunney et al. 2000; Downie et al. 1998). Several physiological responses to environmental stimuli, such as wind, mechanical stimulation or temperature changes include changes in the concentration of Ca²⁺ in the nucleus (van Der Luit et al. 1999; Xiong et al. 2004). The process that has been studied in most detail and includes periodic release and re-uptake of Ca²⁺ in the nucleus is known as calcium spiking. Nod factors, produced by endosymbiotic rhizobia in response to flavonoids, cause rapid influx of Ca²⁺ ions at the plasma membrane of the cells at the root hair tip and subsequent perinuclear Ca²⁺ spiking in legumes (Ehrhardt et al. 1996). CASTOR and POLLUX were identified in mutant screens in Lotus japonicus for nodulation defects and impaired perinuclear Ca²⁺ spiking. Although they were also shown to be localized in chloroplasts, both proteins are targeted to the NE and have electrophysiological properties of cation channels, with a preference for potassium (Charpentier et al. 2008). It was hypothesized that they could either function as counter-ion channels, or change the electric potential across the NE that could possibly activate the still unknown calcium channels. DMH (Doesn't Make Infections 1), the Medicago truncatula homologue of POLLUX is associated with the NE, too (Riely et al. 2007). Interestingly, a homologue of CASTOR and POLLUX is predicted in the Arabidopsis genome (encoded by At5g49960), although its function has yet to be revealed, since Arabidopsis is an asymbiotic species (ImaizumiAnraku et al. 2005). A role for the NE in Ca²⁺ spiking does not end here. In Lotus japonicus, Nup85 and Nup133, both components of the Nup107–Nup160 NPC subcomplex, were shown to be required for Ca²⁺ spiking and nodule formation (Kanamori et al. 2006; Saito et al. 2007).

Another example for the role Ca²⁺ plays in signal transduction at the NE comes from heat stress studies. Liu and colleagues have shown that heat stress is followed by elevated cytoplasmic levels of calcium ions (Liu et al. 2006; Liu et al. 2003), released from intracellular Ca²⁺ depot membranes, such as the endoplasmic reticulum or NE. An increased Ca²⁺ level leads to the increased expression and accumulation of calmodulin (CaM), which has multiple levels of action in response to the heat shock. CaM promotes the DNA-binding activity of the heat shock transcription factor HSF in maize (Li et al. 2004; Sun et al. 2000). In Arabidopsis, it was shown that CaM interacts with a protein phosphatase PP7, which regulates the expression of heat shock protein genes (Liu et al. 2007). CaM also activates calmodulinbinding protein kinase 3 (AtCBK3), which in turn promotes phosphorylation of Arabidopsis HSFA1a and subsequent expression of the heat shock protein genes (Liu et al. 2008).

In legumes, pharmacological studies suggest the involvement of both phospholipase D and phospholipase C upstream of Ca²⁺ spiking (den Hartog et al. 2003; Engstrom et al. 2002). Phospholipase D also activates the phospholipids PA (phosphatidic acid) and PIP₂ (phosphatidylinositol 4,5-biphosphate) and causes relocation of PIP₂ to the NE, among other subcellular localizations. These phospholipids are known as mediators of signaling pathways involved in signal transduction during heat, cold, osmotic, salt, pathogen and wounding stress (Mishkind et al. 2009), illustrating that a series of signaling events, mediating stress responses, take place at the NE.

Signal Transduction of External Stimuli—the Role of the NE/NPC

Hormone-Regulated Signal Transduction

The Role of the NPC in Plant Innate Immunity

In an evolving arms race against their microbial and viral pathogens, plants deploy two different, yet converging defense strategies that include two groups of immune receptors. The first class consists of membrane-associated extracellular receptors that recognize pathogen-associated molecular patterns (PAMPs), which set off the PAMP-triggered immunity (PTI) response. The PTI signal is further transmitted to activate the mitogen-activated protein kinase (MAPK) signaling pathways and downstream transcriptional responses, mediated to a large part by WRKY transcription factors. The second class is represented by resistance proteins (R) that specifically recognize the presence of pathogen effector proteins, secreted into the cell by Type III secretion systems, resulting in effector-triggered immunity (ETI). ETI is characterized by the response that results in the programmed cell death at the site of infection, called hypersensitive response (HR), as well as by the systemic signals that provide a long distance defense response of uninfected tissues, named systemic acquired resistance (SAR). A resistance response activated by virulent pathogens on susceptible hosts is considered a basal defense response and includes PTI and weak ETI caused by weak recognition of pathogen effectors (Jones and Dangl 2006).

Salicylic acid (SA) is required for the initiation of SAR and the expression of pathogenesis-related (PR) genes. NPR1 (Nonexpressor of PR genes 1) is one of the major regulators of both basal defense and SAR (Dong 2004). npr1 mutants are characterized by abolished SAR, due to insensitivity to SA, leading to a failure to express PR genes (Cao et al. 1994). NPR1 represents one the first examples of a protein involved in plant defense against pathogens for which it was shown that the function requires the partitioning between the nucleus and the cytoplasm. The localization of NPR1-GFP, which was able to rescue the npr1 mutant, was shown to be different in the absence and the presence of SA. NPR1-GFP signal was found evenly distributed between nucleus and cytoplasm under non-inducing conditions, but was exclusively nuclear after treatment with SA. Nuclear accumulation of NPR1 was found to be necessary, but not sufficient for PR gene expression and SAR (Kinkema et al. 2000). It was demonstrated that NPR1 exists predominantly in the form of oligomers in the cytoplasm and that due to a change in the cellular redox potential, the majority of NPR1 becomes monomeric and moves to the nucleus upon SA induction (Mou et al. 2003). Recently, it was shown that the nuclear function of NPR1 as a co-regulator of the transcription of PR genes is proteasome-dependent (Spoel et al. 2009). This degradation is dependent upon phosphorylation of NPR1 after activation of PR gene transcription. Both phosphorylation and degradation of NPR1 are required for NRP1-mediated activation of target genes and disease resistance. This study clearly demonstrates how different levels of regulation of localization, posttranslational modification and degradation of NPR1 are orchestrated to control its function.

Two types of R proteins share a central nucleotide binding (NB) and a C-terminal leucine-rich repeat (LRR) domains and differ based on the presence of the N-terminal Toll interleukin receptor (TIR) or a coiled-coil (CC) domain. TIR-NB-LRR proteins require EDS1/PAD4 (Enhanced disease susceptibility 1/Phytoalexin deficient 4) and EDS1/SAG101 (Senescence-associated gene 101) complexes as their downstream signaling component (Feys et al. 2005). Interestingly, nucleocytoplasmic trafficking of EDS1 and PAD4 was proposed as one of the possible mechanisms explaining the requirement of EDS1/PAD4 complex partitioning in its defense signaling function (Wiermer et al. 2005). CC-NB-LRR proteins, on the other hand, commonly use the membrane protein NDR1 (Non race-specific disease resistance 1) in their downstream signal transduction pathways (Century et al. 1997). EDS1- and NDR1-dependent pathways converge at the downstream step of SA synthesis, which provides a signal necessary for basal defense and SAR. A gain of function mutation in the Arabidopsis TIR-NB-LRR R gene, SNC1 (Suppressor of npr1-1, constitutive 1) leads to constitutive SNC1 activity and constitutive defense response (Zhang et al. 2003). Suppressors of snc1 (Modifiers of snc1, mos) include MOS6 (importin α3), MOS3 (Nup96/SAR3) and MOS7 (Nup88) (Cheng et al. 2009; Palma et al. 2005; Zhang and Li 2005). Since all three MOS proteins represent either a component of the NPC or of nuclear import, it was proposed that the impaired nucleocytoplasmic transport of several candidates is responsible for the suppression phenotype of snc1. These include EDS1, PAD4, SAG101, the transcription factor bZIP10, NPR1 and SNC1 itself (Wiermer et al. 2007). That this is indeed the case was demonstrated in mos7-1, a partial loss of function mutant in which nuclear accumulation of SNC1 was reduced substantially (Cheng et al. 2009). In addition, both overall and nuclear levels of EDS1 and NPR1 were reduced in the mutant background, suggesting a role of MOS7/Nup88 in the stability of these two proteins. It is possible that both decreased nuclear import and enhanced nuclear export of SNC1 in mos7-1 contribute to its decreased nuclear accumulation.

In the past few years, there has been a burst of publications describing the absolute requirement of nuclear localization or trafficking through the nucleus of several other R proteins for their activity. A fraction of CC-NB-LRR type MLA R proteins from barley was found in the nucleus (Bieri et al. 2004; Shen et al. 2007). Nuclear accumulation of MLA10 was increased upon fungal infection and its function was lost when an NES was added to prevent its nuclear localization. The interaction of MLA10 with WRKY transcription repressors in the nucleus was dependent on the presence of its fungal effector A10. Based on genetic analysis, it was concluded that A10-dependent MLA10/WRKY repressor interaction causes the activation of the defense genes expression by inhibiting WRKY repressor function, established during PTI (Shen et al. 2007; Shen and Schulze-Lefert 2007). The most exciting finding from this study was a demonstration that ETI and PTI converge in the nucleus, acting on the same set of WRKY transcription factors. This indicates that ETI can act as an amplification loop to increase the severity of PTI responses.

The tobacco TIR-NB-LRR R protein N, which provides resistance to Tobacco mosaic virus (TMV) fails to achieve this function when fused to an NES. Although the interaction between N and its viral effector p50 TMV replicase occurs in the cytoplasm, it was proposed that this interaction activates N, enabling its translocation to the nucleus and subsequent interaction with SPL transcription factors, required for TMV disease resistance (Burch-Smith et al. 2007; Shen and Schulze-Lefert 2007). The Arabidopsis TIR-NB-LRR protein RPS4, which confers resistance to Pseudomonas syringae strains expressing the effector AvrRps4, also requires nuclear localization for its function in disease resistance (Wirthmueller et al. 2007). Unlike MLA and N, RPS4 was shown to contain a functional bipartite NLS, necessary for nuclear import and immunity. However, a recent prediction of nuclear localization estimates that many more R proteins contain a nuclear localization signal. This indicates that the function of R proteins in the nucleus might be a rule rather than an exception (Liu and Coaker 2008). Another interesting example is the Arabidopsis TIR-NB-LRR protein RRS1 that induces resistance to Ralstonia solanacearum. RRS1 interaction with the bacterial effector PopP2, which possesses a bipartite NLS, causes its translocation to the nucleus. In addition to nuclear localization, the C-terminal WRKY domain of RRS1 is also required for PopP2-induced ETI (Deslandes et al. 2003). Therefore, RRS1 represents yet another case where R protein function inside the nucleus provides a direct modulation of the transcription of genes involved in defense responses.

RanGAP2 has been shown to specifically interact with a coiled-coil domain of the CC-NB-LRR protein Rx, which confers resistance to PVX (potato virus X), as well as with the coiledcoil domains of Rx-like proteins Rx2 and Gpa2, the last being required for resistance against nematode Globodera pallida (Sacco et al. 2009; Sacco et al. 2007; Tameling and Baulcombe 2007). It has been demonstrated that RanGAP2 was required for both Rx-mediated and Gpa2-mediated resistance to their corresponding pathogens (Sacco et al. 2009; Sacco et al. 2007; Tameling and Baulcombe 2007). Interestingly, neither GAP activity nor NE localization of RanGAP2 seem to be important for this capability, since RanGAPI does not interact with Rx-like proteins and the WPP/AAP mutant of RanGAP2 retains the ability to induce HR in the presence of Rx (Sacco et al. 2007). Although the mechanism by which RanGAP2 plays a role in disease resistance remains unknown, it has been confirmed that it mediates an initial recognition of a pathogen effector through the interaction with a coiled-coil domain of a Rx-like protein, while a LRR domain provides further recognition specificity (Rairdan et al. 2008; Sacco et al. 2009).

A general mechanism of a defense strategy of plants against viruses that includes neither RNA silencing nor gene-for-gene resistance was recently discovered (Carvalho et al. 2008). It was demonstrated that a ribosomal protein rpL10A represents a substrate and a downstream effector of the NSP-interacting kinase (NIK). NIK kinase activity causes translocation of rpL10A to the nucleus, where it either functions to directly induce defense responses or to attenuate general protein synthesis. Although geminivirus nuclear shuttle protein (NSP) inhibits NIK-mediated phosphorylation of rpL10A, it was proposed that this mechanism might act to protect the host against other viruses by inhibiting plant growth and development. As all of the examples mentioned above illustrate, translocation of proteins in and out of the nucleus is becoming an emerging theme in the pathogen resistance field, thus obtaining the level of appreciation it deserves.