Behaviour of the NE (nuclear envelope) during open mitosis has been explored extensively in metazoans, but lack of native markers has limited similar investigations in plants. In the present study, carried out using living synchronized tobacco BY-2 suspension cultures, the non-functional NE marker LBR (lamin B receptor)–GFP (green fluorescent protein) and two native, functional NE proteins, AtSUN1 [Arapidopsis thaliana SUN (Sad1/UNC84) 1] and AtSUN2, we provide evidence that the ER (endoplasmic reticulum)-retention theory for NE membranes is applicable in plants. We also observe two apparently unique plant features: location of the NE-membrane components in close proximity to chromatin throughout division, and spatially distinct reformation of the NE commencing at the chromatin surface facing the spindle poles and concluding at the surface facing the cell plate. Mobility of the proteins was investigated in the interphase NE, during NE breakdown and reformation, in the spindle membranes and the cell plate. A role for AtSUN2 in nuclear envelope breakdown is suggested.
- live imaging
- nuclear envelope
- protein dynamics
- Sad1/UNC84 (SUN) domain protein
The NE (nuclear envelope) is a dual membrane structure comprising the ONM (outer nuclear membrane) and INM (inner nuclear membrane), perforated by nuclear pores. Although this basic structure is present in all eukaryotes, there is significant variation in its properties and behaviour between kingdoms. In particular, lower eukaryotes undergo a ‘closed’ mitosis in which the envelope remains largely intact, whereas in higher eukaryotes, including plants and higher metazoans, the NE breaks down and reforms in a controlled manner . In metazoans, phosphorylation of chromatin-binding INM components and lamins, which form a supportive meshwork structure underlying the INM, disrupts binding interactions to chromatin, affecting the integrity of the NE structure during NEBD (NE breakdown). At the end of division, dephosphorylation allows these interactions to be re-established and for the NE to reform around the decondensing chromatin . Two models haven been proposed as to the fate of the NE membranes and components during open mitosis . The vesiculation model is supported by earlier research using cell-free Xenopus egg extract to study NE reformation, and suggests that NE membranes and protein components are present as vesicles in the dividing cell, which are recruited to the decondensing chromatin in late anaphase to re-assemble into the NE network. The favoured ER (endoplasmic reticulum)-retention model stems from more recent research using living intact cells, and proposes that during NEBD the NE membranes and protein components become part of the mitotic ER network and re-emerge from this network at late anaphase to reform the NE around decondensing chromatin [1–4]. The molecular mechanisms and components essential to the mitotic dynamics of the NE are well researched in metazoans, but remain mainly obscure in plants. This is partly due to a lack of known proteins and their functions involved in these processes. Although cyclin, cyclin-dependent kinases and aurora kinases are involved in mitotic progression, their impact on NEBD and NE reformation remains unknown [5,6]. Well-documented chromatin-binding INM proteins such as the LBR (lamin B receptor) or LEM (Lap-Emerin-Man) domain proteins are absent in plants. This is also true for lamins. However, recent research of the dynamics of two lamin-like proteins, NMCP (nuclear matrix constituent protein) 1 and NMCP2, suggests that these also disassemble at the beginning of mitosis in Apium graveolens cells, but are distributed and reassemble differently . The NMCP1 associates with the mitotic spindle, whereas NMCP2 is present in the mitotic cytoplasm and assembles later at the reforming NE than NMCP1 . Whether NMCP1 and 2 function similarly to lamins, however, remains unknown. The same study suggests that in fixed A. graveolens cells, dye-stained membranes, including the NE, are present in vesicles, supporting the vesiculation model . On the other hand, studies with the human-derived plant NE marker LBR–GFP (green fluorescent protein) showed that NE membranes in living BY-2 cells distributed to mitotic membranes, supporting the ER-retention model [8,9]. The same marker was used previously to show that in HeLa cells the NE membranes migrate into the mitotic ER and re-emerge to form the new NE . The study also found that although LBR–GFP was strongly immobilized in the interphase NE by binding interactions, it moved relatively freely in the mitotic membranes, indicative of unbound, free movement. Until recently a lack of native plant NE proteins ruled out examining the fate and function of NE intrinsic components in mitosis. In the present study, we address this issue by using the newly characterized plant SUN (Sad1/UNC84) domain proteins . The SUN domain proteins are highly conserved across kingdoms and are a key component of bridging complexes that span the nuclear periplasm to link the nucleoskeleton and chromatin with the cytoskeleton [11–13]. The Arabidopsis thaliana SUN domain proteins AtSUN1 and AtSUN2 were found to be localized at the NE in interphase cells and provide the first evidence of a functional nucleo-cytoskeletal bridging complex in plants and a unique opportunity to study NE membrane protein targeting and mobility . A NLS (nuclear localization signal) was found to be involved in their NE targeting, and a C-terminal coiled-coil domain required for binding interactions between the two SUN domain proteins. They were found to be characteristic, functional NE components  and as such are ideal to study mitotic division.
The aims of the present study were to gain insights into the fate of NE membranes and their functional components in plant cell division. To do this, we observed the localization and mobile behaviour of the plant SUN domain proteins in interphase and mitosis together with the non-functional LBR–GFP NE marker. Our findings support the ER-retention model and show that, in contrast with metazoans, mitotic membranes remain in close proximity to mitotic chromatin and that NE reformation is spatially organized in dividing plant cells. Moreover, we observed changes in the mobility of AtSUN2–YFP (yellow fluorescent protein) during mitosis, indicating alterations in binding interactions.
MATERIALS AND METHODS
The construction of all AtSUN1- and AtSUN2-based fluorescent protein fusion constructs as well as LBR–GFP has been published previously [8,10]. The pBIB::H2B-YFP  was used as a template to make H2B–CFP (cyan fluorescent protein). Forward primer FH2B (5′-ATGGCGAAGGCAGATAAGAAACC-3′) and reverse primer RH2B (5′-AGAACTCGTAAACTTCGTAACCG-3′) were used to amplify the H2B-coding sequence, and primers FH2BGW (5′-GGGGACAAGTTTGTACAAAAAAGCAGGCTTCCCGCCAATGGCGAAGGCAGATAAGAAACC-3′) and RH2BGW (5′-GGGGACCACTTTGTACAAGAAAGCTGGGTCAGAACTCGTAAACTTCGTAACCG-3′) were used to add Gateway® flanking recombination sequences. Using Gateway®-cloning technology, the H2B-coding sequence was first inserted into the entry vector pDONR207 and then moved to the binary expression vector pK7CWG2 to make pK7CWG2::H2B (H2B–CFP). The vector was introduced into Agrobacterium tumefaciens GV3101::pMP90 using heat-shock transformation .
Stable transformation of BY-2 cells
Co-cultivation adapted from previously published protocols  was used to generate stable transformed BY-2 cell lines. Agrobacteria carrying either pCambia1300::AtSUN1–YFP, pCambia1300::AtSUN2–YFP, pCambia1300::YFP–AtSUN1ΔN, pCambia1300::YFP–AtSUN1ΔCC, pCambia1300::YFP–AtSUN2, pVKH18En6::LBR–GFP or pK7CWG2::H2B were used to stably transform BY-2 cells. Wild-type BY-2 cells were grown at 25°C in the dark with shaking at 130 rev./min and subcultured every 7 days . For transformation, 3-day-old cells were used. Liquid overnight agrobacteria cultures were centrifuged at 6000 g for 5 min at 21°C and washed three times with YEB medium containing 200 μM acetosyringone. Bacteria were incubated in a final wash for 1 h before transformation. Bacteria and BY-2 cells were co-cultivated for 72 h at 25°C in the dark without shaking on solid BY-2 medium-containing plates. BY-2 cells were then washed three times with liquid BY-2 medium containing carbenicillin and timentin. Finally, cells were plated on solid BY-2 medium plates containing carbenicillin, timentin and selection antibiotics. H2B–CFP was selected with kanamycin, and the other constructs were selected with hygromycin B. Cells were incubated at 25°C in the dark without shaking for approximately 4 weeks until calli had grown. Calli were screened with a Leica MZFLIII stereomicroscope for fluorescence and selected calli were transferred to liquid BY-2 medium including selection antibiotics to establish stable expressing BY-2 suspension cell lines. Stable expressing cell lines were grown and cultured in a similar manner to wild-type cell lines . For double-transformed cell lines, BY-2 cells were first transformed with one construct and, once stable expressing suspension cells were established, these were used for a second transformation with the other construct.
Synchrony of stable transformed BY-2 cells
Synchrony of BY-2 cells was based on a previously published protocol . In total, 7 ml of 7-day-old cells were transferred to 50 ml of fresh medium and aphidicolin was added to a working concentration of 5 μg/ml. The cells were incubated for 24 h at 25°C with shaking in the dark. The aphidicolin was washed out using a sintered glass funnel. The cells were washed ten times with 50 ml of BY-2 medium and suspended in 50 ml of fresh BY-2 medium after the last wash. Cells were returned to the incubator and sampled for analysis between 4 and 11 h after. At least two independent synchrony experiments were carried out for each cell line.
Cells were mounted on to glass microscope slides covered in a thin film of BY-2 medium containing 0.7% low-melting agarose and sealed with a coverslip. Confocal imaging was carried out as described previously . Acquisition and analysis of FRAP (fluorescence recovery after photobleaching) data was as described previously, except that no latrunculin B treatment was used [10,18]. Approximately 30 FRAP experiments were carried out for each data set [10,18]. Average values±S.D. are displayed. For movie acquisition, samples were scanned every 30 s for approximately 45–60 min. The images were converted to movie files using RAD video tools.
Subcellular localization of SUN domain proteins in dividing BY-2 cells
To follow the dynamics of the SUN domain proteins and the NE in living dividing cells, stable transformed BY-2 cell lines were created that co-expressed either histone H2B fused to the fluorescent protein CFP (H2B–CFP) and AtSUN1–YFP, or H2B–CFP and AtSUN2–YFP. Histone H2B–CFP is an A. thaliana chromatin marker  and was found to localize to the nucleoplasm labelling chromatin in BY-2 cells (Figures 1 and 2, and Supplementary Movies S1–S3 at http://www.BiochemJ.org/bj/435/bj4350661add.htm).
As previously reported, AtSUN1–YFP and AtSUN2–YFP localize to the NE in interphase with low levels of fluorescent protein also being present in the ER in some cells (Figure 1) . As the cells entered mitosis, changes in the distribution of the two SUN domain proteins were already noticeable. In prophase, as condensing chromatin was observed, AtSUN1–YFP and AtSUN2–YFP fluorescence at the NE began to decrease, and the fusion proteins to accumulate at opposing sides of the NE, presumably where the spindle poles were forming, as well as in the ER (Figure 1 and Supplementary Movie S3). The SUN-labelled NE started to lose its shape (Supplementary Movie S3) and in prometaphase a structured NE was no longer discernible (Figure 1 and Supplementary Movies S1–S3). Instead, the two SUN domain proteins were localized in mitotic ER membranes and accumulated in membranes at the spindle (spindle membranes). In particular, AtSUN1–YFP was observed to be present in tubules and also puncta that surrounded condensing chromatin, and were in close proximity to it (Figure 1 and Supplementary Movie S2). AtSUN2–YFP was not observed in puncta, but also surrounded chromatin, although AtSUN2–YFP fluorescence appeared diffuse (Figure 1 and Supplementary Movie S3). As condensed chromosomes aligned in the division zone in metaphase, both SUN proteins accumulated in spindle membranes close to the spindle poles (Figure 1, and Supplementary Movies S1 and S3). However, some AtSUN1–YFP and AtSUN2–YFP fluorescence was observed in tubules stretching through the division zone with tubule tips appearing in close proximity to the chromosomes (Figure 1). In some cells expressing AtSUN1–YFP, the fusion protein was also found in puncta associated with tubules and spindle membranes similar to the prometaphase puncta. As the sister chromatids became separated and migrated to the spindle poles in anaphase, both AtSUN1–YFP and AtSUN2–YFP were observed accumulating on decondensing chromatin facing the spindle poles (Figure 1, and Supplementary Movies S1 and S3). Some tubules were also seen traversing the division zone, and in the transition to telophase both SUN fusion proteins appeared in the forming cell plate (Figure 1, and Supplementary Movies S1 and S3). In telophase the NE containing AtSUN1–YFP and AtSUN2–YFP reformed around decondensing chromatin, with the SUN proteins first present on the side of the chromatin facing the spindle poles, then enveloping the sides of the chromatin and lastly surrounding the chromatin facing the growing cell plate (Figure 1 and Supplementary Movie S1). In cytokinesis both AtSUN1–YFP and AtSUN2–YFP were present in the expanding NE and in the cell plate and phragmoplast.
Effects of mutations on subcellular localization of SUN domain proteins in dividing BY-2 cells
Previously, we have reported that deleting the N-terminus or coiled-coil domain from SUN domain proteins affects their mobile behaviour in the interphase NE caused by disruption of binding interactions . In the present study we examined the effect of these mutations on the distribution of SUN domain proteins in mitotic division. For this we generated stable transformed BY-2 cell lines co-expressing either H2B–CFP and YFP–AtSUN1ΔN, where the N-terminus is deleted, or H2B–CFP and YFP–AtSUN1ΔCC, where the coiled-coil domain is deleted. A third cell line was generated that co-expressed H2B–CFP and YFP–AtSUN2. Although this fusion protein is full-length and not mutated, the fusion of YFP to the N-terminus affects the functionality of the protein altering its mobility and interactions . In the two AtSUN1 mutant cell lines, no general effect on the localization of the SUN proteins was observed in comparison with the non-mutated AtSUN1–YFP cell line (Supplementary Figure S1 at http://www.BiochemJ.org/bj/435/bj4350661add.htm). However, no puncta were observed in prometaphase and metaphase of YFP–AtSUN1ΔN- and YFP–AtSUN1ΔCC-expressing cells (Supplementary Figure S1). Interestingly, the N-terminal YFP fusion affected the NE localization of AtSUN2 in interphase BY-2 cells, as only weak and indistinct NE labelling was observed in YFP–AtSUN2-expressing cells (Figure 2). In prophase, however, the fusion protein aggregated specifically at the NE (Figure 2, asterisk). The localization of YFP–AtSUN2 in the remaining phases of the division was similar to that of AtSUN2–YFP (Supplementary Figure S2 at http://www.BiochemJ.org/bj/435/bj4350661add.htm).
Mobility of AtSUN1–YFP and AtSUN2–YFP in mitotic membranes
Examining the mobility of proteins in the membrane gives information on the binding interactions of these proteins. Slow mobility indicates strong, anchoring binding interactions, whereas high mobility suggests that the protein is unbound, moving unhindered. FRAP was used to quantify the mobile fraction and half-time of AtSUN1–YFP, AtSUN2–YFP and LBR–GFP in the interphase NE (Figures 3A and 3B) and mitotic membranes (Figures 3C and 3D, and Supplementary Figure S3 at http://www.BiochemJ.org/bj/435/bj4350661add.htm) of BY-2 cells (Tables 1 and 2, and Supplementary Table S1 at http://www.BiochemJ.org/bj/435/bj4350661add.htm). The mobile fraction denotes the percentage of fusion protein found moving, and the half-time is an indicator of how fast this movement is.
In the 3-day-old interphase NE, LBR–GFP was very mobile with a mobile fraction of 85.98±8.19% and a half-time of 1.31±0.72 s (Figure 3B and Supplementary Table S1). The mobile fraction and half-time of LBR–GFP remained unchanged during NEBD in spindle membranes, during NE reformation and in the cell plate, indicating that LBR–GFP remains unbound and freely mobile in these membrane systems (Supplementary Table S1 and Supplementary Figure S3).
The AtSUN1–YFP had a mobile fraction of 72.66±5.89% and a half-time of 9.83±1.75 s in the interphase NE, indicating that over 70% of the fusion protein was mobile and less than 30% was immobilized in the membrane (Figure 3B and Table 1). The mobile fraction of AtSUN1–YFP appeared to remain similar during mitosis, as similar mobile fractions were observed for the protein in the metaphase spindle membrane and the reforming NE. In the cell plate, however, the mobile fraction of AtSUN1–YFP was reduced in comparison with the interphase NE indicating that less of the protein was mobile (Figure 3B and Table 1). A similar half-time as in the interphase NE was observed in the reforming NE and the cell plate. However, in the metaphase spindle membranes AtSUN1–YFP appeared to move faster as the half-time was significantly decreased (5.10±0.83 s, P<0.001).
AtSUN2–YFP had a mobile fraction of 37.82±8.39% and a half-time of 6.02±1.62 s in the interphase NE, indicating that approximately 37% of the fusion protein population was mobile and approximately 63% was immobile in the membrane (Figure 3B and Table 2). During mitosis, the mobile behaviour of AtSUN2–YFP was found to vary significantly. During NEBD, the mobile fraction remained low (31.79±15.41%), similar to that of the interphase NE, but it showed a higher rate of movement, as indicated by a lower half-time of 3.44±2.18 s (Table 2). This suggests that while AtSUN2–YFP retains its interphase interactions during prophase NEBD, changes occur as ER retention commences. In metaphase spindle membranes, AtSUN2–YFP had almost twice the mobile fraction (63.46±7.97%) of AtSUN2–YFP in interphase and during NEBD (Table 2). It also had a shorter half-time, indicating that it is much more mobile in metaphase spindle membranes, suggesting that it no longer is part of binding interactions at the interphase NE and during NEBD. As most of the spindle membranes are no longer surrounding chromatin it may be hypothesized that the interphase/prophase interactions of AtSUN2–YFP include binding of chromatin or other nuclear components. This is reinforced by the finding that the mobile fraction of AtSUN2–YFP decreased in the reforming NE, where it was similar to that of interphase and NEBD (Figure 3D and Table 2). The halftime of AtSUN2–YFP in the telophase-reforming NE was higher than during NEBD and similar to that in interphase. In the cell plate, AtSUN2–YFP had a higher mobile fraction (53.57±10.29%; Table 2) indicating fewer or different interactions than found at the NE. However, as still almost half of the protein is immobilized in the cell plate, this may suggest that it is also associated in protein complexes there and may fulfil a function.
In higher eukaryotes the NE breaks down in open cell division and considerable interest has been shown in the fate of the NE and its components, particularly in metazoans. Although work using a Xenopus egg extract system suggested vesiculation of the membrane, more recently the ER-retention model has been favoured. In this, the NE is absorbed into the mitotic ER network and re-assembles from this around decondensing chromatin at the end of division [2,4,19]. Whether the ER-retention model holds true for the fate of plant NE membranes is still debated. Using both light and electron microscopy, it has been shown that ER-derived mitotic membranes are present in the mitotic spindle of plant cells [20,21]. Previously, the non-functional plant NE marker LBR–GFP was shown to be localized in mitotic membranes in dividing BY-2 cells [8,9] providing evidence for the ER-retention model. On the other hand, Kimura et al.  found that in fixed dividing A. graveolens cells, an ER membrane marker, also present in the NE, marked punctate, vesicular structures suggesting vesiculation. Although it is possible that in different plant species the NE and ER assume different morphologies in cell division, it is also known that chemical fixation and permeabilization required for immunostaining affects membrane morphology. To avoid this, and in order to establish whether the ER-retention model was also applicable to plants in addition to metazoans, we used living tobacco BY-2 cells co-expressing native NE membrane and chromatin markers fused to fluorescent proteins to study the fate of the NE membranes and the functionality of plant SUN domain proteins during mitotic division.
A GFP construct based on human LBR was previously shown to be a plant NE marker that is correctly targeted to the plant NE but is not functional there [8,18]. The two plant proteins AtSUN1 and AtSUN2 used in the present study are functional NE intrinsic proteins that belong to the SUN domain family, which is conserved across kingdoms . Yeast and animal SUN domain proteins are essential for several mitotic and meiotic functions, including anchorage of centromeres and centrosomes, as well as decondensation of chromatin at the end of mitosis [13,22,23]. They are also thought to be one of the first INM proteins recruited to the decondensing chromatin during NE reformation . In the present study, we investigated the localization and mobility of AtSUN1 and AtSUN2 in mitotic membranes to gain first insights into their functionality during cell division.
The results of the present study provide strong evidence that the ER-retention model applies to plants as well as to metazoans. Both SUN domain proteins and the LBR–GFP remain in the NE until prophase, upon NEBD they migrate to mitotic ER membranes and accumulate in the reforming NE at the end of division (Figure 1) [8,9]. This redistribution resembles that of metazoan NE intrinsic proteins. The model is therefore likely to be widely applicable to higher eukaryotes undergoing open cell division.
In contrast with metazoans, however, we observe that membranes containing NE proteins remain in close proximity to chromatin throughout division. This may be a plant-specific feature. In metazoans, the NE proteins evenly distribute throughout the mitotic ER, and metaphase chromosomes are devoid of membranes [2,24]. Our observations show that both functional SUN domain proteins, as well as LBR–GFP and mutated SUN domain proteins, localize to spindle membranes and to tubules traversing the division zone in metaphase, when most of the spindle membranes accumulate at the spindle poles (Figure 1, Supplementary Figures S1 and S2, and Supplementary Movies S1–S3) . The presence of the non-functional LBR–GFP and SUN mutants in the tubules, as well as the high mobility of the AtSUN1–YFP and AtSUN2–YFP in the spindle membranes, suggests that the fusion proteins are not necessarily functional in these tubules. However, the presence of the tubules and membranes in close proximity to chromatin throughout mitosis may suggest functional connections. Immunostaining of dividing tomato cells has previously revealed the presence of the NE-specific LCA (lycopersicon Ca2+-ATPase) in spindle-associated membranes, and it was suggested that these membranes are involved in calcium signalling [6,25]. On the other hand, it could be hypothesized that the tubules remain in close proximity to regions of metaphase chromatin, where protein interactions will later initiate NE reformation. In metazoans, mitotic ER tubules containing INM proteins contact decondensing chromatin in anaphase to initiate NE reformation [2,4]. Güttinger et al.  suggest that peripheral and core chromatin areas attract different INM proteins and chromatin decondensation plays an essential part in regulating NE reformation . Although the role of chromatin and the regulation of both NEBD and NE reformation remain to be elucidated in plants, our observations suggest a functional relationship between mitotic membranes and chromatin in NE reformation. This is corroborated by the finding that NE reformation is spatially organized. Both AtSUN1–YFP and AtSUN2–YFP first aggregate at the surface of chromatin facing the spindle pole, followed by accumulation around the sides and finally localization of the proteins on chromatin facing the cell plate. This directionality has also not been observed in metazoans [2,4,26] and may be a plant-specific feature, commensurate with the suggested presence of specific chromatin domains, and suggests a tight spatial regulation of NE reformation. Similar observations were made for NE-associated NMCP1 and NMCP2 in A. graveolens cells and found that NMCP1 also first assembled on the chromatin surface facing the spindle in late anaphase and at telophase completely surrounded the decondensing chromatin . It is interesting to note that although AtSUN1–YFP and AtSUN2–YFP aggregated in the reforming NE in late anaphase, LBR–GFP did not . Although it is present in membranes surrounding the decondensing chromatin and in the reforming NE, it does not appear to be strongly accumulated in the reforming NE, as observed for the two SUN proteins (Figure 1) . This suggests that AtSUN1 and AtSUN2 are recruited to the reforming NE by specific interactions with either chromatin or other specifically targeted nuclear components such as NMCP1. The accumulation of LBR–GFP in the fully reformed NE is probably due to import of the protein  and may therefore suggest that the NE is transport-competent at that stage.
By investigating the mobility of the non-functional LBR–GFP in dividing BY-2 cells, we found that it remains highly mobile throughout division (Supplementary Figure S3 and Supplementary Table S1) and is not involved in binding interactions. Comparisons with AtSUN1–YFP and AtSUN2–YFP mobility are difficult to make, however, because of the varying expression levels between the three constructs (Supplementary Table S2 at http://www.BiochemJ.org/bj/435/bj4350661add.htm).
The results of the present study show that the dynamics of AtSUN2–YFP are affected by its localization in different mitotic membranes. In the interphase NE, during NEBD and NE reformation AtSUN2–YFP remains tightly anchored to the membranes, whereas in the spindle membranes and the cell plate the protein is much more mobile. As in the interphase NE, during NEBD and NE reformation AtSUN2–YFP is in close proximity to chromatin and nucleoplasmic components, it suggests that interactions with these immobilize the protein. A possible role for AtSUN2 in prophase and during NEBD is suggested by the finding that YFP–AtSUN2 accumulates at the NE specifically during prophase. What function this may serve remains to be elucidated. In Dictyostelium, SUN1 is required for centromere–centrosome associations , and generally yeast and animal SUN domain proteins have been found to associate with chromatin [12,13,28], so a role in chromatin organization prior to NEBD could be envisaged for AtSUN2. The low signal of YFP–AtSUN2 at the interphase NE suggests that targeting of the protein is inefficient, and previous work has shown that an N-terminal NLS is required for effective NE targeting of the protein . The N-terminal YFP fusion was observed to affect the functionality of the AtSUN2 N-terminus , indicating that reduced NE targeting of YFP–AtSUN2 in interphase is due to an impaired NLS. The accumulation of YFP–AtSUN2 in prophase is therefore either due to an NLS-independent targeting mechanism or because of an increased permeability of the NE due to the breakdown of NPCs (nuclear pore complexes). In metazoan cells it has been shown that disassembly of NPCs precedes NEBD so that kinases and other components involved in NEBD can accumulate in the nucleoplasm to cause dissociation of interactions at the nuclear periphery and changes in chromatin organization [1,4].
Finally, all three NE proteins were localized at the cell plate and in phragmoplast membranes between reforming NE and cell plate. Previously, other plant NE components including Ran, RanGAP, WITs [WPP (tryptophan-proline-proline)-interacting tail-anchored protein] and WIPs (WPP-interacting protein) have been found to localize to the cell plate, and a role for the Ran cycle in cytokinesis has been hypothesized . The presence of LBR–GFP, a non-functional protein, in the cell plate [8,9] suggests non-specific membrane traffic to be responsible for the presence of NE proteins at the cell plate. However, AtSUN2–YFP appears to have a significant immobile fraction at the cell plate (Figure 3B and Table 2) indicating interactions and therefore may be functional. It is interesting to observe that so far only in plants has spatial NE reformation from proximal to the spindle to proximal to the cell plate been observed, and it is intriguing to speculate whether this is associated with the presence of NE proteins in the cell plate. Among other factors such as kinases, phosphatases, Ran and chromatin remodelling components, the amount of available membrane is known to affect NE reformation  and it is possible that the cell plate and phragmoplast form membrane reservoirs that affect the distribution of membranes in the plant cell. Thus the SUN domain proteins might be involved in such regulation.
In conclusion, we have obtained convincing evidence for the ER-retention model for NE components in living plant cells by showing for the first time that the functional NE proteins AtSUN1 and AtSUN2 relocate to the mitotic membranes in plant cell division. Furthermore, we have demonstrated that in plants some membranes remain in close proximity to chromatin and that the reformation of the NE around decondensing chromatin is spatially organized starting proximal to the spindle and concluding proximal to the cell plate. Both of these observations appear to be unique to plants. We have also shown that the dynamic properties of AtSUN2–YFP change during mitosis and that the protein may be involved in interactions with nuclear components that render it immobile in interphase and are retained during NEBD and NE reformation. A potential role for AtSUN2 in NEBD is hypothesized owing to its immobility at this stage and its specific accumulation in the prophase NE. In addition, its reduced mobility in the cell plate may indicate a potential functional involvement.
Note added in proof (received 14 March 2011)
We wish to draw readers attention to a paper by Yoshihisda Oda and Hiroo Fukuda  entitled “Dynamics of Arabidopsis SUN proteins during mitosis and involvement of nuclear shaping” which was published after the final revision of this paper.
The experimental work was carried out by Katja Graumann. Experimental design and writing of the manuscript was done by David Evans and Katja Graumann.
This work was supported by the Leverhulme Trust [grant number F/00382/H].
We thank Anne Kearns (Oxford Brookes University) for her help with stable BY-2 transformations and culturing of BY-2 cells. We also thank Dr Dennis Francis (School of Biosciences, University of Cardiff, Cardiff, U.K.) for providing wild-type BY-2 cell lines and his advice on synchronizing BY-2 cells.
Abbreviations: CFP, cyan fluorescent protein; ER, endoplasmic reticulum; FRAP, fluorescence recovery after photobleaching; GFP, green fluorescent protein; INM, inner nuclear membrane; LBR, lamin B receptor; NE, nuclear envelope; NEBD, NE breakdown; NLS, nuclear localization signal; NMCP, nuclear matrix constituent protein; NPC, nuclear pore complex; SUN, Sad1/UNC84; WPP, tryptophan-proline-proline; YFP, yellow fluorescent protein
- © The Authors Journal compilation © 2011 Biochemical Society