The nuclear envelope protein TMEM209 is an integral component of the nuclear pore complex and interacts with Nup210

ABSTRACT

A highly curved membrane region connecting the inner and the outer nuclear membrane serves as a platform where nucleoporins with one or more transmembrane domains promote anchoring of the nuclear pore complex to the nuclear envelope. In mammalian cells, three transmembrane nucleoporins, Nup210, POM121 and NDC1, are inserted at this site. Here, we characterize TMEM209, which had initially been identified as a protein concentrated at the nuclear envelope, as a fourth transmembrane nucleoporin. Proximity labeling revealed that TMEM209 is present close to proteins of the inner nuclear membrane and to other nucleoporins. TMEM209 localized to the nuclear pore complex in immunofluorescence microscopy and biochemically interacted with Nup210 via a region containing its two transmembrane domains. TMEM209 depletion impaired cell growth and delayed entry into S, G2 and M phases of the cell cycle. Conversely, its overexpression specifically dissociated Nup210 from the nuclear envelope. Together, these findings establish TMEM209 as a novel transmembrane nucleoporin that cooperates with Nup210 in cell cycle progression and cell proliferation.

Summary: TMEM209 is a novel nucleoporin containing a single transmembrane domain that interacts with Nup210.

INTRODUCTION

The nuclear envelope (NE) separates the cytoplasm from the nuclear compartment. It comprises the inner nuclear membrane (INM), the outer nuclear membrane (ONM), the intermembrane or perinuclear space, and the nuclear pore complexes (NPCs), large structures at the sites where INM and ONM fuse to form a conduit for nucleocytoplasmic exchange of macromolecules. In metazoan cells, a nuclear lamina underneath the INM with the intermediate filament proteins lamin A/C and lamin B1 and B2 as key components provides stability to the nucleus (Burke and Stewart, 2013; Maurer and Lammerding, 2019) and is also involved in regulatory functions, for example, in gene expression (Odell and Lammerding, 2023; van Steensel and Belmont, 2017). NPCs control active transport of proteins and RNA–protein complexes between the nucleus and the cytoplasm in both directions (Wing et al., 2022) and also allow passive diffusion of smaller molecules between the two compartments (Mohr et al., 2009; Timney et al., 2016). In human cells, they are composed of multiple copies of ∼30 different proteins called nucleoporins, and reach a mass of ∼120 MDa (for a review, see Hoelz et al., 2016; Petrovic et al., 2025). In recent years, high-resolution structures of NPC subcomplexes and of entire NPCs from different species have been obtained (Akey et al., 2022; Bley et al., 2022; Kim et al., 2018; Kosinski et al., 2016; Mosalaganti et al., 2018, 2022; Petrovic et al., 2022; Singh et al., 2024; von Appen et al., 2015), allowing a classification of nucleoporins according to their localization with respect to the plane of the NE. Asymmetric nucleoporins are only found on the cytoplasmic side of the NPC (e.g. Nup214 and Nup358) or in the nuclear basket (e.g. Nup153 and TPR). Symmetric nucleoporins (e.g. Nup98 and Nup107) localize to inner or outer nuclear rings on either side of the NPC. Nucleoporins containing one or more transmembrane domains (TMDs) anchor the entire complex in the NE, as they are inserted into the highly curved membrane patch that connects the INM and the ONM. In addition to these transmembrane nucleoporins (TM-Nups), several nucleoporins have the capacity to interact with membranes via amphipathic helices (Drin et al., 2007; Hamed and Antonin, 2021), further contributing to the association of NPCs with the NE, to their correct alignment and, possibly, the membrane curvature at the nuclear pore. Of the ∼30 nucleoporins, about ten contain characteristic phenylalanine-glycine (FG) motifs that are important for the interaction of soluble nuclear transport factors with the NPC (Kehlenbach et al., 2023). These FG-Nups, in particular Nup98, also play a major role in the formation of the permeability barrier that restricts the free passage of molecules between the nucleus and the cytoplasm (Hülsmann et al., 2012). In the fully assembled NPC, between one and eight copies of each of the individual nucleoporins interact to form a spoke-like structure, eight of which are then combined in the functional complex with its characteristic eight-fold symmetry along the axis perpendicular to the plane of the NE. Hence, nucleoporins typically occur in copy numbers of eight (Nup42 and GLE1) or multiples of eight (all other nucleoporins) (Ori et al., 2013; Petrovic et al., 2025).

The first TM-Nups that had been identified were Nup210 (initially termed gp190 or gp210; Gerace et al., 1982; Wozniak et al., 1989) and POM121 (pore membrane protein of 121 kDa) (Hallberg et al., 1993). Nup210 is an integral membrane protein with a single TMD close to its C-terminal end that projects most of its mass into the perinuclear space (Greber et al., 1990; Wozniak et al., 1989). Only the last ∼60 C-terminal amino acid residues beyond the TMD are exposed to the NPC side of the NE. In Pom152, the yeast ortholog of Nup210, this luminal region folds into immunoglobulin (Ig)-like domains (Upla et al., 2017), and such structures have also been observed in human NPCs, where up to eight copies of Nup210 are present in a single spoke (Mosalaganti et al., 2022). For NPC targeting of Nup210, its transmembrane region and a short C-terminal stretch seem to be sufficient (Wozniak and Blobel, 1992). In Caenorhabditis elegans and in Xenopus laevis, Nup210 has been reported to affect NE breakdown (Galy et al., 2008). Interestingly, Nup210 is absent from certain cell types (D'Angelo et al., 2012; Olsson et al., 2004) and even knockout mice are viable (van Nieuwenhuijze et al., 2018), showing that the protein is not absolutely required for NPC biogenesis or function.

The transmembrane protein POM121 has a single TMD close to its N-terminus and is the only TM-Nup that contains several FG-motifs (Hallberg et al., 1993). Its large C-terminal part faces the nuclear side of the membrane (Söderqvist and Hallberg, 1994) and also contains regions that are required for integration into NPCs (Söderqvist et al., 1997; Stavru et al., 2006b). In nuclear assembly reactions in a Xenopus system, the depletion of POM121 has been reported to inhibit NE formation, whereas Nup210 seems to be dispensable (Antonin et al., 2005). In mammalian cells, by contrast, even a co-depletion of POM121 and Nup210 did not drastically affect NPC assembly (Stavru et al., 2006b), suggesting that other TM-Nups could take over redundant functions of the two proteins.

NDC1, the mammalian ortholog of yeast Ndc1p, was identified as a third TM-Nup, comprising six TMDs with both termini oriented towards the pore side of the NE (Mansfeld et al., 2006; Stavru et al., 2006a). It is the only known TM-Nup with a clear sequence conservation between yeast and higher eukaryotes. Depletion of NDC1 from HeLa cells results in NPC assembly defects. Similar effects were also observed in homozygous ndc1^−/− (known as npp-22 in worms) C. elegans mutants lacking the nucleoporin (Stavru et al., 2006a). These worms, however, were viable, again pointing to redundant functions of individual TM-Nups.

The INM is contiguous with the ONM and the ER membrane, and at least small proteins are thought to freely diffuse in the plane of the lipid bilayer between the three membrane systems (Katta et al., 2014; Ungricht et al., 2015; Zuleger et al., 2011). Nevertheless, many proteins are enriched at the INM because they can be retained at specific binding sites, for example, at proteins of the nuclear lamina. Such proteins have largely been discovered in proteomic screens (Cheng et al., 2019, 2023; Korfali et al., 2012; Malik et al., 2010; Schirmer et al., 2003), and for many of them, their precise function remains unknown. TMEM209, initially referred to as NET31, was identified in an early screen for NE-concentrated proteins (Schirmer et al., 2003) and later confirmed in other studies to be enriched at the NE (Cheng et al., 2023; Malik et al., 2010). In proximity to its N-terminus, this 63-kDa protein contains two predicted TMDs. Likewise, PNET1, a plant ortholog of TMEM209, contains two such TMDs and has been recently shown to be incorporated into NPCs in Arabidopsis thaliana (Fang et al., 2025). In human cells, TMEM209 interacts with the nucleoporin Nup205 in co-immunoprecipitation experiments (Fujitomo et al., 2012). Here, we investigate the topology, the subcellular localization and the interaction partners of TMEM209 in detail. We identify TMEM209 as the fourth bona fide TM-Nup in human cells. TMEM209 interacts with Nup210 and affects progression through the cell cycle and, together with Nup210 and/or other TM-Nups, cell proliferation.

RESULTS

TMEM209 is exposed to the nucleoplasm

TMEM209 was identified as a protein concentrated at the NE (Schirmer et al., 2003), its precise localization and its membrane topology, however, have not been investigated in detail. Close to its N-terminal end, TMEM209 has two predicted transmembrane domains, connected by a short linker sequence. To investigate the topology of TMEM209, we fused an HA tag to either the N- or the C-terminus of the protein and subjected transfected cells to a differential permeabilization approach. Digitonin is a mild detergent that binds cholesterol and readily permeabilizes the plasma membrane, leaving internal membranes largely intact (Adam et al., 1990). Hence, only epitopes that are exposed to the cytoplasm should be detected upon digitonin permeabilization. As shown in Fig. 1A, both N- and C-terminally HA-tagged versions of TMEM209 could be detected with antibodies against the tag, suggesting that both ends of the protein are exposed to the cytoplasm in cells overexpressing HA–TMEM209 or TMEM209–HA, respectively. Depending on the expression level, the tagged proteins were observed at the level of the NE and/or the ER. As a control, we used antibodies against the nuclear proteins lamin A/C, which could only be detected upon permeabilization with Triton X-100, that is under conditions that also lead to permeabilization of the nuclear membranes. With digitonin, by contrast, epitopes were not accessible for the anti-lamin A/C antibodies. Fig. 1B shows the predicted TmAlphaFold structure (Dobson et al., 2023; Jumper et al., 2021) of TMEM209 and its topology, with both ends of the protein facing the cytoplasmic side of the ER, or, upon translocation of the protein to the INM, the nucleoplasm. The short linker sequence between the two TMDs thus resides in the ER or the perinuclear space. According to the prediction, an α-helix with an amphipathic character (amino acids 239–257 of TMEM209) aligns with the surface of the membrane, similar to nucleoporins that associate with the NE via such helices (Hamed and Antonin, 2021).

Fig. 1.

Topology of TMEM209. (A) HeLa cells were transfected with constructs coding for HA–TMEM209 or TMEM209–HA and subjected to indirect immunofluorescence using digitonin or Triton X-100 for permeabilization and anti-HA- or anti-lamin A/C antibodies. Scale bars: 10 µm. Note that the transfection efficiency and the expression level for TMEM209–HA were low. Major differences compared to HA–TMEM209 regarding the localization pattern of the two proteins, however, were not observed. Images representative of four experimental repeats. (B) Model of TMEM209 inserted into a membrane, as predicted by the TmAlphaFold database (Dobson et al., 2023). The structured region (top) faces the cytoplasm or, after translocation of the protein to the INM, the nucleoplasm. The arrows indicate the N-terminal (red) and C-terminal ends (blue) of TMEM209, respectively. The green arrow points to the short region between the transmembrane domains, which is expected to localize in the lumen of the ER or in the intermembrane space between INM and ONM. Note the extended α-helix of the first TMD and the α-helical region predicted to associate with the cytoplasmic/nucleoplasmic leaflet of the membrane. Model confidence: the average predicted local distance difference test (pLDDT) score for TMEM209 is 68.38 (low) with 29.8% and 29.4% of the protein with a ‘very high’ (pLDDT>90) or ‘high’ (pLDDT >70) level of confidence, respectively (compare with Fig. S6). The evaluation result of TmAlphaFold was “good” with a qValue (TMDET result) of 70.18 and a CCTOP reliability level of 84.2. See also Fig. 7, where truncated versions of TMEM209 are used.

Next, we addressed the question of whether tagged versions of TMEM209 could reach regions of the NE exposed to its nuclear side, that is the INM and/or nucleoplasmic NPC regions. Cells were co-transfected with constructs coding for HA–FRB–TMEM209 and the nuclear reporter protein APEX2–dEGFP–cNLS–FKBP12. The FRB and the FKBP12 cassettes are expected to dimerize in the presence of rapamycin (Chen et al., 1995). If the protein of interest (here HA-FRB-TMEM209) actually resides at the INM and/or regions of the NPC exposed to the nucleoplasm, the soluble reporter protein (here APEX2–dEGFP–cNLS–FKBP12) should relocate from the nuclear interior to the periphery upon addition of the drug. Indeed, 30 min after the addition of rapamycin, the green reporter protein was detected at the nuclear rim (Fig. 2A), suggesting a partial localization of HA–FRB–TMEM209 to nucleoplasmically exposed regions of the NE. Our nuclear reporter protein with its APEX2 moiety (i.e. an ascorbate peroxidase that generates biotin-phenoxyl radicals from biotin phenol in the presence of H₂O₂; Hung et al., 2016) allows a precisely controlled biotinylation of proteins that are in close proximity to HA–FRB–TMEM209 only in the presence of rapamycin. Such proteins can be identified by stable isotope labeling with amino acids in cell culture (SILAC) and mass spectrometry, enabling the identification of proximity and interaction partners of TMEM209 at the INM or the NPC. Fig. 2B shows the result of our established RAPIDS (rapamycin- and APEX-dependent identification of proteins by SILAC; see James et al., 2022, 2019 for details of the method) approach. Several proteins of the INM [emerin (EMD) and the lamina associated polypeptide 1 (LAP1, TOR1AIP1)] and the nuclear lamina [lamin A (LMNA), lamin B (LMNB1, LMNB2)], as well as nucleoplasmically exposed nucleoporins (Nup153, Nup155, AHCTF1), were found to be enriched in the biotinylated (eluted) fraction of proteins upon treatment of cells with rapamycin (for mass spectrometry details see Table S1, which lists the results of two RAPIDS experiments). Nup205, a suggested interaction partner of TMEM209 (Fujitomo et al., 2012), was identified as well, albeit at a low level of significance. To confirm several of these proteins as proximity partners of TMEM209 with favored biotinylation in the presence of rapamycin, samples were analyzed by western blotting (Fig. 2C). For Nup153, lamin A/C, lamin B1 and emerin, the blots show higher protein levels upon treatment of cells with rapamycin. As an alternative to the rapamycin experiments, we used proximity ligation assays (PLAs; Söderberg et al., 2006) to monitor the relationship of HA–TMEM209 to emerin and lamin B1, proteins of the INM and the nuclear lamina, respectively. With this approach, many more PLA dots surrounding the nucleus were observed compared to in control experiments, confirming the localization of HA–TMEM209 to the nucleoplasmically exposed regions of the NE (Fig. S1).

Fig. 2.

RAPIDS identifies proximity partners of TMEM209. HeLa cells were transfected with constructs coding for APEX2–dEGFP–cNLS–FKBP12 and HA–FRB–TMEM209 and were treated with or without rapamycin. (A) After indirect immunofluorescence detecting the HA and the GFP tags, cells were analyzed by confocal microscopy. Scale bars: 10 µm. Images representative of four experimental repeats. (B) Cells were subjected to RAPIDS. The upper left quadrant of the scatter plot shows normalized log₂ ratios of proteins eluted from Neutravidin beads in single forward [heavy medium (H) without rapamycin; light medium (L), with rapamycin; x-axis] and reverse (heavy medium, with rapamycin; light medium, without rapamycin; y-axis) experiments. Magenta, proteins that were significant in all reactions; green, proteins that were significant in only one reaction. See also Table S1, which lists the results of the experiment in B and an experimental replicate. (C) Total cell lysates (total) and proteins eluted from Neutravidin beads (elution) from forward and reverse reactions were analyzed by SDS-PAGE followed by Western blotting using antibodies against Nup153, lamin B1, lamin A/C, TMEM209 and emerin. L/H, ±indicate light and heavy media, with or without rapamycin. Blot representative of two experimental repeats.

TMEM209 associates with the nuclear pore complex

Our RAPIDS results suggested that TMEM209 occurs in close proximity to nucleoporins. To further investigate the relationship between TMEM209 and the NPC, we performed high- and super-resolution microscopy. First, we transfected cells with a construct coding for a V5-tagged version of TMEM209 and subjected them to confocal microscopy, detecting the V5 tag and Nup358. The two proteins colocalized at the level of the NE, suggesting a physical interaction of TMEM209 with the NPC (Fig. 3A). Furthermore, V5–TMEM209 was detected at extranuclear structures that also contained Nup358. In all likelihood, these structures are annulate lamellae, a cytoplasmic reservoir of NPC precursors that contain a subset of nucleoporins, including Nup358 (Wu et al., 2001). Next, we used two-color STED microscopy to analyze the relation of HA–TMEM209 to nucleoporins as detected with the mAb414 antibody. mAb414 binds Nup358, Nup214, Nup153 and Nup62, FG-Nups that are also modified by O-linked N-acetylglucosamine (Davis and Blobel, 1986). As shown in Fig. 3B,C, HA–TMEM209 and FG-nucleoporins showed a similar staining pattern on the surface of the nucleus with partially overlapping peaks in the line profiles of the two fluorophores. Finally, we used STED microscopy and a segmentation approach to more quantitatively monitor the co-occurrence of HA–TMEM209 and an individual nucleoporin, Nup153, in a defined segment (Fig. 3D,E). In a control experiment with Nup153 and Nup210 as established nucleoporins, we found that ∼80% of the analyzed segments (i.e. NPCs) detected with anti-Nup153 antibodies were also detected with antibodies against Nup210 (Fig. S2). A very similar percentage (77.4%) of double-positive segments (NPCs) was observed with anti-Nup153- and anti-HA-antibodies (Fig. 3E). At high signal densities, colocalization in two channels could result from a random overlap of specific or unspecific antibody signals. We therefore repeated the analyses by rotating the images in one channel. In both experiments, the rotation resulted in a largely reduced signal overlap (Fig. 3E, from 77.4% to 48.8%; Fig. S2, from 80.3% to 40.1%), suggesting a non-random colocalization of Nup210 and Nup153 (Fig. S2) and also of HA–TMEM209 and Nup153 (Fig. 3E).

Fig. 3.

TMEM209 associates with the NPC. (A) HeLa cells were transfected with a construct coding for V5–TMEM209 and subjected to indirect immunofluorescence and confocal microscopy, detecting the V5 tag and Nup358. Scale bar: 10 µm. Images representative of two experimental repeats. (B–F) HeLa cells were transfected with a construct coding for HA–TMEM209 and subjected to indirect immunofluorescence and STED microscopy. (B) For immunofluorescence, the mAb414 and anti-HA antibodies were used. Scale bar: 10 µm. (C) RGB line profile of a region (yellow bar) in B. Signal resolutions were in the range of 80 nm and 125 nm for the HA- and the mAb414-signal, respectively. (D) For immunofluorescence, antibodies against Nup153 and the HA tag were used. Scale bar: 10 µm. (E) The level of colocalization of Nup153 and HA–TMEM209 was analyzed for 7226 NPCs from seven individual nuclei of different images as defined by Nup153 staining. (F) As a control, images used for the analysis in E were rotated by 90° in one of the two channels. This rotation resulted in a reduction of double-positive NPCs (from 77.4% to 48.8%), suggesting that only a portion of the original signals overlapped by chance. See also Fig. S2 for control.

TMEM209 interacts with Nup210

Our results described so far showed that TMEM209 localizes to the INM and/or the NPC and also associates with selected proteins of these structures. We therefore asked whether this association can also be detected by biochemical means. To facilitate the analysis, we generated a stable HEK FLp-In cell line expressing HA–TMEM209 in a tetracycline-inducible manner. Tetracycline concentrations were chosen to reach similar expression levels of endogenous TMEM209 and the HA-tagged version of the protein. Cells were treated with or without tetracycline and subjected to immunoprecipitation reactions using the zwitterionic detergent CHAPS for cell lysis and antibodies against the HA tag. We then analyzed potential interaction partners for co-precipitation, together with HA–TMEM209. Of the previously identified proximity partners (Fig. 2), low levels of co-precipitating emerin were detected (Fig. 4A), whereas Nup153 and Nup155 were not found in this fraction (data not shown). We also analyzed the western blots for binding of the two transmembrane nucleoporins, Nup210 and POM121. Indeed, low levels of POM121 and higher levels of Nup210 were found to co-immunoprecipitate with HA–TMEM209 from lysates of tetracycline-treated cells (Fig. 4A). In an independent experiment, we then transfected HA–TMEM209-expressing cells with constructs coding for GFP or Nup210–GFP. As shown in Fig. 4B, endogenous TMEM209 as well as HA–TMEM209 induced by tetracycline was co-immunoprecipitated with Nup210–GFP using antibodies against the GFP tag, but not with GFP alone. Finally, we used an antibody against TMEM209 to precipitate the endogenous protein. With this antibody, but not with an IgG control antibody, Nup210 co-precipitated with TMEM209 (Fig. 4C). We also performed immunoprecipitation experiments with digitonin as an alternative detergent for cell lysis. Under these conditions, not only Nup210, but also low amounts of Nup205 co-precipitated with endogenous TMEM209 (Fig. S3). In addition to experiments with CHAPS- or digitonin-containing lysis buffers, we also used the more stringent detergent Nonidet P-40 for cell lysis. Under these conditions, Nup210 did not co-immunoprecipitate with TMEM209 (data not shown), suggesting that the membrane environment influences the interaction. Together with the immunofluorescence localization, these results demonstrate a close association of TMEM209 with NPCs, in particular with the transmembrane nucleoporins POM121 and Nup210, suggesting that TMEM209 is a bona fide nucleoporin itself.

Fig. 4.

Co-immunoprecipitation of Nup210 with TMEM209. HEK Flp-In T-REx 293 cells were stably transfected with a construct coding for HA–TMEM209 and expression was induced with or without tetracycline (±tet). (A) Total cell lysates were subjected to immunoprecipitation (IP) reactions using antibodies against the HA tag. (B) Stable HEK cells were transiently transfected with a construct coding for GFP or Nup210–GFP, and subjected to immunoprecipitation reactions using antibodies against the GFP tag. In A and B, precipitated proteins (A, IP:HA; B, IP:GFP and IP:Nup210–GFP) were analyzed by SDS-PAGE, followed by western blotting using antibodies against the HA tag, Nup210, POM121, lamin B1, lamin B2, emerin and GAPDH (A) or the GFP tag and TMEM209 (B). For controls, 3% of the lysate was loaded. Note that the levels of endogenous TMEM209 and HA–TMEM209 in induced cells (+tet) are similar. IP experiments were performed three times with very similar results. (C) Total HeLa cell lysates were subjected to immunoprecipitation reactions using antibodies against TMEM209 and precipitated proteins were analyzed by SDS-PAGE, followed by western blotting. Note that compared to the condition in A, a larger proportion of Nup210 (input versus elution) co-precipitated with TMEM209. Compare with Fig. S3. Blot in C representative of two experimental repeats.

Effects of depletion and overexpression of TMEM209

Proteins of the INM and transmembrane nucleoporins in particular play important roles in the biogenesis of the NPC. We therefore asked the question of whether either the depletion of TMEM209 or its overexpression would result in changes of NPC composition and/or general cellular functions. First, we used two different siRNAs for the depletion of TMEM209 in HeLa cells. As shown in Fig. 5A,B, total TMEM209 levels were reduced to ∼21% and ∼45% of the control values, respectively. The more efficient siRNA was then used to analyze the effects of reduced TMEM209 levels on staining patterns of proteins of the NE. Neither the INM protein emerin, nor lamin B1 or any of the nucleoporins tested here (Nup358, Nup214, Nup88, Nup153, TPR, Nup98, Nup210 and POM121) were detectably affected by the siRNA treatment (Fig. 5C). Owing to a lack of reliable antibodies, we could not analyze NDC1 in these assays.

Fig. 5.

Knockdown of TMEM209 has no detectable effect on other nucleoporins. HeLa cells were treated with siRNAs against TMEM209. (A) Cell lysates were analyzed by SDS-PAGE, followed by western blotting, detecting TMEM209 and GAPDH. (B) Quantitative analysis of results as shown in A. The bars show the relative levels of TMEM209 upon treatment with two different siRNAs against TMEM209 compared to a control siRNA. Error bars indicate the mean±s.d. (three independent experiments). ****P<0.0001 (one-way ANOVA with Tukey's multiple comparison test). (C) Cells treated with control siRNAs or siRNA #1 against TMEM209 were analyzed by indirect immunofluorescence and confocal microscopy, detecting Nup358, Nup214, Nup88, Nup153, TPR, Nup98, Nup210, Pom121, lamin B1 and emerin. Scale bars: 10 µm. Images representative of two to three experimental repeats.

Next, we used human HAP1 cells lacking TMEM209 and compared them to the parental cells (Fig. S4A,B). Even in the complete absence of TMEM209, the NE was not compromised, as the staining patterns of the analyzed proteins were very similar in control and knockout cells. These results suggested that in these cultured cells, a possible function of TMEM209 at the NE or the NPC could, at least in part, be taken over by other transmembrane proteins, for example, Nup210, NDC1 or POM121.

We next asked the question of whether increased levels of TMEM209 affected the protein composition of the NE or NPC, and transfected HeLa cells and HAP1 cells with a construct coding for HA–TMEM209. For the analysis, we used antibodies against nucleoporins in different regions of the NPC for immunofluorescence labeling (cytoplasmic Nup358 and Nup214, nuclear Nup153 and TPR, the FG-Nup Nup98 of the central channel, and the transmembrane Nups POM121 and Nup210). Strikingly, of all nucleoporins analyzed, only Nup210 showed a clear phenotype in both cell lines – in cells with ectopic expression of HA–TMEM209, the characteristic nuclear rim observed for Nup210 in non-expressing control cells was abolished (Fig. 6). To control for tag- or vector-specific effects, we analyzed different versions of TMEM209 and also the unrelated protein HA–VAPB as a control. Loss of Nup210 from the NE was observed upon expression of all versions of TMEM209, but not in cells expressing HA–VAPB (Fig. S5). We also analyzed the stable HEK cells described above, which typically express much lower levels of HA–TMEM209 compared to transiently transfected HeLa cells. At low expression levels of TMEM209 (comparable to that for the endogenous protein), changes in Nup210 localization were not observed (data not shown). Nevertheless, the effect on Nup210 seen in highly expressing HeLa or HAP1 cells (Fig. 6) was specific, as other nucleoporins, including POM121, were not affected.

Fig. 6.

Overexpression of TMEM209 leads to a loss of Nup210 from the nuclear envelope. HeLa cells (A) or HAP1 cells (B) were transfected with a construct coding for HA–TMEM209, subjected to indirect immunofluorescence using antibodies against the HA tag and Nup358, Nup214, Nup153, Nup98, TPR, POM121 or Nup210, as indicated and analyzed by confocal microscopy. Arrowheads point to cells with reduced levels of Nup210 at the NE. Scale bars: 10 µm. Images of HeLa cells are representative of two experimental repeats, images of HAP1 cells are of one (Nup98, TPR) or at least two repeats.

To further characterize this effect, we ectopically overexpressed truncated versions of HA–TMEM209, as depicted schematically in Fig. 7A. These versions were based on AlphaFold predictions (compare Fig. 1B), with two transmembrane domains, followed by an unstructured region and a structured C-terminal domain. A construct lacking the two transmembrane domains (HA–TMEM209 ΔTMD) resulted in an HA protein that localized diffusely in the cytoplasm, without obvious effects on endogenous Nup210 (Fig. 7B). Two of the C-terminally truncated proteins, HA–TMEM209 1–138 and HA–TMEM209 1–307, resulted in similar phenotypes as the full-length protein (HA–TMEM209 1–561), with a rim around the nucleus and reduced levels of Nup210 at the NE. Another version that lacked only 13 amino acid residues at the very C-terminal end (HA–TMEM209 1–548) exhibited a very peculiar phenotype – the HA protein itself was largely nuclear, forming foci at the NE that also sequestered endogenous Nup210 (Fig. 7B). Other nucleoporins, by contrast, were not found in these HA–TMEM209 1–548 foci (Fig. 7C). The TM-Nup POM121, in particular, did exhibit some intranuclear foci, which were also observed in control cells and never colocalized with HA–TMEM209 1–548. Finally, we performed co-immunoprecipitation experiments using full-length and truncated versions of HA–TMEM209. Endogenous Nup210 co-precipitated with all versions except HA–TMEM209 ΔTMD, although the highest levels were observed with the full-length protein (Fig. 7D). Hence, one or both transmembrane domains of TMEM209 and/or a portion of the unstructured region are sufficient for Nup210 interaction. The C-terminal domain, on the other hand, is not required, although the lack of the very C-terminal end of TMEM209 did affect the subcellular localization of Nup210. We speculate that the single TMD of Nup210 is also involved in the interaction with TMEM209. Indeed, a computed prediction using AlphaFold3 (Abramson et al., 2024) suggested such an interaction as shown in Fig. S6. Together, our findings establish TMEM209 as a novel transmembrane nucleoporin – it can reach the INM and the NPC, where it is present in close proximity to other membrane proteins and nucleoporins. Furthermore, overexpression of TMEM209 specifically affected a single nucleoporin, Nup210, that is, one of only three known TM-Nups that had been described previously. Depending on the version of HA–TMEM209, Nup210 was removed from the NE and/or was sequestered in HA–TMEM209-containing foci.

Fig. 7.

Analysis of TMEM209 truncation mutants. (A) Scheme of HA-tagged TMEM209 versions. Blue, TMDs at the N-terminal end. Green, largely unstructured region; red, structured region of TMEM209 (compare Fig. 1B). (B) HeLa cells were transiently transfected with HA-tagged versions of TMEM209 as depicted in A and analyzed by indirect immunofluorescence and confocal microscopy, using antibodies against the HA tag and Nup210. The stars indicate a cluster of three cells with very low expression levels of HA–TMEM209. The settings in the green channel were adjusted accordingly. Scale bar: 10 µm. (C) Cells were transfected with a construct coding for HA–TMEM209 1–548 and analyzed by indirect immunofluorescence and confocal microscopy, using the mAb414 antibody and antibodies against the proteins of interest (POI) Nup358, Nup153, POM121, Nup155 and Nup210, as indicated. Scale bars: 10 µm (5 µm in enlarged frames). Note that only Nup210 colocalized with HA–TMEM209 in nuclear aggregates. Images in B and C representative of at least two experimental repeats, except for the Nup155 staining in C, which was only performed once. (D) HeLa cells were transfected with constructs coding for versions of HA-tagged TMEM209 as indicated or with an empty vector (HA control) and cell lysates were subjected to immunoprecipitation (IP) reactions using antibodies against the HA tag. Precipitated proteins were analyzed by SDS-PAGE followed by western blotting, using antibodies against Nup210 and the HA tag. 3% of the cell lysate was loaded for comparison. The experiment was performed at least three times for individual HA–TMEM209 fragments, with very similar results.

TMEM209 affects the cell cycle and proliferation

TMEM209 is not absolutely required for cell growth, as, for example, HAP1 cells lacking the protein are viable. Similarly, it has been described that cells lacking Nup210, POM121 or NDC1 are viable, although cell growth can be compromised in the absence of more than one transmembrane nucleoporin (Stavru et al., 2006a,b). Furthermore, plant cells lacking PNET1, the homolog of TMEM209, show reduced growth (Fang et al., 2025). We therefore decided to analyze the effects of reduced concentrations of TMEM209 on proliferation in detail. We first used HAP1 cells lacking TMEM209 and compared them to the parental cell line. As shown in Fig. 8A, TMEM209-knockout cells grew somewhat slower than the control cells. Whether knockout of TMEM209 per se slows cell cycle progression is uncertain, as different HAP1 cell strains can vary in their ploidy status (Beigl et al., 2020). We therefore used the two strains of HAP1 cells to also deplete the transmembrane nucleoporins Nup210 and POM121 by mean of RNAi. Depletion of the latter had no effect on proliferation in control cells or in the TMEM209-knockout cells, whereas the depletion of Nup210 reduced the proliferation of control cells. In TMEM209-knockout cells, this effect was aggravated and proliferation ceased after about 72 h of siRNA treatment (Fig. 8B). We also subjected these cells to indirect immunofluorescence to investigate potential effects on individual nucleoporins or INM proteins. In this analysis, no obvious phenotype was observed in HAP1 cells lacking both TMEM209 and Nup210 (Fig. S7).

Fig. 8.

Effects of TMEM209 on cell proliferation. (A) 10⁵ HAP1 parental and TMEM209-knockout cells were seeded in individual six-well plates, removed from the wells by trypsinization after the indicated time points and counted. (B) Cells as in A were transfected with control siRNAs or siRNAs against POM121 or Nup210. In A and B, the graphs show the number of cells after 24, 48, 72 and 96 h. Error bars indicate mean±s.d. (three independent experiments). (C,D) HeLa cells were treated with control siRNAs or siRNAs against TMEM209 and subjected to a synchronization protocol. (C) Cells were fixed at the indicated time points after the release from the thymidine block and analyzed by indirect immunofluorescence and confocal microscopy, using antibodies against phosphorylated histone H3 (pHistoneH3) and tubulin. Scale bar: 20 µm. Images representative of two experimental repeats. (D) Cells as in C were analyzed by flow cytometry and relative cell numbers in the individual stages of the cell cycle (G1 phase, S phase, G2/M phase) were plotted over time. Note that in one set of experiments, we determined the cell cycle stage distribution from 0 to 6 h (indicated by the dashed lines) after the release from the thymidine block, with a total of two measurements. Another set measured the distribution between 7 and 14 h after the release, including the 0-h time point. A total of two measurements were performed for the 7- and 14-h time points, whereas a total of three measurements was performed for 8–13-h timepoints. Error bars mean±s.e.m. (two to three independent experiments).

To analyze the effects of TMEM209 on proliferation in more detail, we synchronized control HeLa cells and TMEM209-depleted cells using a double-thymidine block and followed their progression through the cell cycle upon release of the block. Cells were first analyzed by fluorescence microscopy, detecting phosphorylated histone H3 (pHistoneH3) as a late-G2/mitotic marker and tubulin to monitor the cellular shape. In control cells, pHistoneH3 was detected ∼9 h after release from the thymidine block. In TMEM209-knockout cells, the appearance of pHistoneH3 was delayed by ∼2 h, with fluorescent signals still visible 13 h after release (Fig. 8C). Concomitantly, the depleted cells exhibited a different morphology; in particular between 9 and 11 h after release, the control cells were rather round, as expected for mitotic cells, whereas the depleted cells largely retained their elongated shape. We then subjected cells to cell cycle analysis by flow cytometry. In unsynchronized cells, only minor differences between control cells and TMEM209-depleted cells were observed (data not shown). Upon synchronization and release, however, clear differences became apparent; the depleted cells needed a longer period of time to leave G1 and remained in S phase for an extended period of time before entering G2/M-phase 2 to 3 h later compared to the control cells, that is, in good agreement with the results described above for phosphorylated histone H3. This result indicates the knockdown of TMEM209 in HeLa cells slows cell cycle progression.

DISCUSSION

TMEM209 is a transmembrane nucleoporin

We confirm TMEM209 as a protein with N- and C-terminal ends exposed to the cytoplasm or the nuclear volume and a short stretch between the TMDs that projects into the ER or the perinuclear space, respectively. It remains to be investigated whether the predicted α-helix (amino acids 239–257, Fig. 1B) contributes to membrane association and/or curvature, as shown for other nucleoporins (Hamed and Antonin, 2021). TMEM209 is in close proximity to INM proteins and also to several nucleoporins, as seen by RAPIDS and PLAs. The established TM-Nups Nup210 and POM121 interact with endogenous and ectopic TMEM209 in co-immunoprecipitation assays, although they were not identified by RAPIDS as specific proximity partners. Perhaps these proteins do not contain accessible amino acid residues that could be targeted by the short-lived biotin phenoxyl radicals as they are generated by the ascorbate peroxidase APEX2 in RAPIDS. Indeed, POM121 and Nup210 contain only a few tyrosine, tryptophan, histidine and cysteine residues (amino acids that are prone to biotinylation; Trinkle-Mulcahy, 2019), adjacent to their TMDs. We also identified Nup205, a nucleoporin that has been shown previously to interact with TMEM209 (Fujitomo et al., 2012) as a proximity (Fig. 2B) or interaction partner (Fig. S3), albeit at rather low levels of significance in both assays. Together with our immunofluorescence microscopy data showing a characteristic NPC staining pattern of TMEM209, the co-immunoprecipitation data thus establish the protein as a genuine nucleoporin. Our findings are in line with recent publications on TMEM209. After its original identification as a nuclear membrane protein (Schirmer et al., 2003), TMEM209 was described as an NE-enriched protein in several studies. It has been reported to have a particularly high NE-enrichment score (i.e. the protein is enriched at the NE compared to the ER; Cheng et al., 2023) and suggested to be a novel TM-Nup in the same study. The plant ortholog of TMEM209, PNET1, has been identified by BioID2 as a TM-Nup that is in close proximity to several plant nucleoporins (Tang et al., 2020) and was recently shown to interact with proteins of the NPC scaffold in a phosphorylation-dependent manner (Fang et al., 2025). In trypanosomes, the lamina-associated protein 59 (LAP59) is proposed to be ortholog of TMEM209 and PNET1, as they all share the same domain structure (Butterfield et al., 2024). LAP59 has been shown in pull-down assays to interact with LAP333, a protein with Ig-like domains and structural homologies to Nup210 (Butterfield et al., 2024). Finally, TMEM209 interacts with Nup214 and many other nucleoporins in affinity capture experiments (Serganov et al., 2022). Together, these results establish TMEM209 as a bona fide nucleoporin. Besides Nup210, POM121 and NDC1, TMEM209 is the fourth TM-Nup in metazoans, together with its orthologs in plants and trypanosomes. In the yeast Saccharomyces cerevisiae, Pom34p was identified as a TM-Nup with two TMDs that interacts with other nucleoporins (Miao et al., 2006; Rout et al., 2000). There are no other obvious structural homologies between Pom34p and TMEM209. Nevertheless, they might fulfill similar functions, as Pom34p can also interact with the yeast TM-Nup Pom152, which shares many structural features with Nup210 (Upla et al., 2017).

In a recent cryo-electron tomography study of human NPCs, TMEM209 has not been assigned (Mosalaganti et al., 2022) and its exact position in the NPC remains unclear. Based on our biochemical data, an association with Nup210 appears likely. Of particular interest here is the TMEM209 mutant that lacks the last 13 C-terminal amino acid residues (TMEM209 1–548). Overexpression of this mutant resulted in a specific co-association with Nup210 in non-NPC structures, leaving POM121 unaffected. Other membrane-associated versions of TMEM209 resulted in a partial or total loss of Nup210 from the NE, suggesting that in TMEM209-overexpressing cells, the two proteins can already interact at the level of the ER, so that a portion of newly synthesized Nup210 is unable to reach its final destination, the NPC. In addition, Nup210 that has already been incorporated into an NPC could dissociate from the complex and diffuse, via the ONM, to the ER to be sequestered there by newly synthesized overexpressed TMEM209. Indeed, the residence time of Nup210–EGFP at the NPC is extremely short (∼5 min, compared to many hours for POM121 and most other nucleoporins; Eriksson et al., 2004; Rabut et al., 2004). The dynamics of TMEM209 have not been analyzed so far and we do not know its relative distribution between the NE and the NPC. We speculate that TMEM209 has a short residence time at the NPC as well, perhaps explaining its late description as an integral component of the complex. Based on our findings, we suggest that the TMDs of TMEM209 and Nup210 are crucial for their interaction. This interpretation is in line with the results of our co-immunoprecipitation experiments (Figs 4 and 7D). Only ∼60 amino acid residues of Nup210 are exposed to the cytoplasm or the nuclear region, where they could potentially interact with TMEM209. Together with its TMD, that C-terminal tail of Nup210 is sufficient to localize the protein to the pore membrane (Wozniak and Blobel, 1992), suggesting that this region is also involved in TMEM209 binding. Of note, a recently published tool for prediction of protein–protein interactions (Zhang et al., 2025) identifies Nup210 as a prominent binding partner of TMEM209, with interacting TMDs in one of the suggested models (http://prodata.swmed.edu/humanPPI/), similar to the model shown in Fig. S6. Interestingly, serine residues close to the C-terminus of Nup210 have been shown to be phosphorylated during mitosis, a modification that could interfere with binding to other nucleoporins (Favreau et al., 1996). For LAP59, the short stretch between the TMDs has been suggested to interact with the perinuclear region of LAP333 (Butterfield et al., 2024) and it will be interesting to investigate whether the corresponding regions are also relevant for TMEM209–Nup210 interactions.

Possible functions of TMEM209

TMEM209 is not essential for proliferation of cultured cells, as HAP1 cells grow in its absence. Furthermore, knockout mice are viable, although they are reported to be “significantly impacted” at different levels (e.g. the immune system and the nervous system) by the International Mouse Phenotyping Consortium (IMPC, https://www.mousephenotype.org/). In our siRNA-knockdown experiments and also in HAP1 cells lacking TMEM209, we did not observe major NPC defects, suggesting that, in cultured cell models, a loss of TMEM209 can be compensated for by other TM-Nups. As shown for other TM-Nups, cells (and animals) seem to tolerate the lack of individual proteins and only the combined loss of two or more TM-Nups can lead to more severe phenotypes (D'Angelo et al., 2012; Funakoshi et al., 2011; Galy et al., 2008; Mansfeld et al., 2006; Stavru et al., 2006a,b). In yeast, even a double mutant (pom34Δ pom152Δ) of two TM-Nups is viable (Miao et al., 2006). Pom34 and Pom152 form a complex with yeast Ndc1 (Onischenko et al., 2009) and depletion of Ndc1 from cells also lacking Pom34 and Pom152 resulted in defective NPCs (Madrid et al., 2006). Defective NPC biogenesis could be explained by the role of NDC1 in the recruitment of membrane vesicles to chromatin towards the end of mitosis (Rasala et al., 2008). In human HAP1 cells lacking TMEM209, the depletion of POM121 did not lead to reduced cell growth. The depletion of Nup210 in the HAP1 knockout strain, by contrast, resulted in a clear growth defect, as the cells did not further proliferate 2 to 3 days after siRNA treatment. Our cell cycle analyses in HeLa cells further support a role for TMEM209 in proliferation, as in depleted cells, the G1-S-phase transition and the duration of S-phase were affected, leading to a delayed entry into G2-phase and mitosis. These observations are in line with results from plant cells, where PNET1 mutants showed higher percentages of cells in G1- and S-phase compared to the wild type (Fang et al., 2025). Those authors further describe PNET1 as a protein predominantly expressed in actively dividing cells that is required for rapid cell growth and speculate that phosphorylation of PNET1 could regulate disassembly and/or reassembly of NPCs in the course of mitosis. Human TMEM209 contains several phosphorylation sites (Johnson et al., 2023), which might affect binding to other nucleoporins. In human lung cancer cells, the depletion of TMEM209 inhibited proliferation, whereas its overexpression promoted cell growth (Fujitomo et al., 2012). The authors also reported a co-immunoprecipitation of TMEM209 and Nup205, providing an early hint to TMEM209 being a component of the NPC. Furthermore, TMEM209 was shown to promote proliferation of hepatocellular carcinoma cells (Fang et al., 2024).

Together, we identify TMEM209 as a novel TM-Nup with a role in cell proliferation. Like other TM-Nups, it is not essential, but seems to cooperate with Nup210, POM121 and/or NDC1, for example during NPC assembly or disassembly. To better understand the function of TMEM209, it will be important to explore its dynamics with respect to its interaction partners and its distribution at the NE throughout the cell cycle.

MATERIALS AND METHODS

Plasmids and molecular cloning

PCR was used to amplify DNA fragments from cDNA or from available plasmids using Phusion^TM High-Fidelity DNA Polymerase (2 U/μl) and forward and reverse primers as listed in Table S2. Standard molecular biology methods were used for restriction digests and cloning. For the Gibson assembly method, 50 ng of vector DNA was mixed with a 3-fold molar excess of insert DNA and 10 μl 2× Gibson Assembly Master Mix (New England Biolabs) in a total volume of 20 μl. After incubation for 30 min at 50°C, E. coli (DH5α) were used for transformation. Plasmids (see Table S3 for all plasmids) were confirmed by sequencing (Eurofins Genomics).

Purification of proteins

GST–TMEM209 238–561 was expressed overnight at 18°C in E. coli BL21 DE3 cells after induction with 0.5 mM IPTG at an optical density of 0.6. Cells were harvested by centrifugation at 5020 g for 20 min and resuspended in 35 ml GST lysis buffer (25 mM Tris-HCl pH 8.3, 200 mM NaCl, 5 mM MgCl₂, 3% glycerol, 2 mM β-mercaptoethanol, 1% Triton X-100, 0.1 mM PMSF, 1 μg/ml each of aprotinin, leupeptin and pepstatin). Cell disruption was performed on the EmulsiFlex-C3. The lysates were cleared by centrifugation using a JA30.50TI rotor (Beckman Coulter) at 18,000 g at 4°C for 45 min. Lysates were incubated for 2 h at 4°C with 2 ml Glutathione Sepharose 4 Fast Flow, equilibrated in GST standard buffer (25 mM Tris-HCl pH 8.3, 200 mM NaCl, 5 mM MgCl₂, 3% glycerol, 2 mM β-mercaptoethanol, 0.1 mM PMSF, 1 μg/ml each of aprotinin, leupeptin and pepstatin) for 10 min. Beads were washed four times with 50 ml GST washing buffer (25 mM Tris-HCl pH 8.3, 500 mM NaCl, 5 mM MgCl₂, 3% glycerol, 2 mM β-mercaptoethanol, 0.1 mM PMSF, 1 μg/ml each of aprotinin, leupeptin and pepstatin) in a gravity-low column and proteins were eluted in two steps using 5 ml GST elution buffer (25 mM Tris/HCl pH 8.3, 200 mM NaCl, 5 mM MgCl₂, 3% glycerol, 2 mM β-mercaptoethanol, 50 mM reduced glutathione and protease inhibitors as above).

MBP–TMEM209 238–561 was expressed for 3 h 30 min at 37°C in E. coli BL21 DE3 cells after inducing with 0.5 mM IPTG at an OD₆₀₀ of 0.6. Cell harvesting and disruption was performed as described above using MBP lysis buffer (50 mM Tris-HCl pH 7.4, 200 mM NaCl, 5 mM MgCl₂, 3% glycerol, 2 mM β-mercaptoethanol, 1% Triton X-100 and protease inhibitors as above). For protein purification, lysates were incubated at 4°C for 2 h with 2 ml Amylose Resin High Flow (New England Biolabs) equilibrated in MBP standard buffer (50 mM Tris-HCl pH 7.4, 200 mM NaCl, 5 mM MgCl₂, 3% glycerol, 2 mM β-mercaptoethanol and protease inhibitors as above). Beads were washed four times with 50 ml MBP washing buffer (50 mM Tris-HCl pH 7.4, 500 mM NaCl, 5 mM MgCl₂, 3% glycerol, 5 mM ATP, 2 mM β-mercaptoethanol and protease inhibitors as above) and eluted in two steps using 5 ml MBP elution buffer (50 mM Tris-HCl pH 7.4, 200 mM NaCl, 5 mM MgCl₂, 3% glycerole, 20 mM D-maltose 2 mM β-mercaptoethanol and protease inhibitors as above).

Purified proteins were dialyzed against GST or MBP standard buffers without protease inhibitors, using a 6 ml Amicon^® Ultra Centrifugal Filter 10 kDa MWCO, frozen in liquid nitrogen and stored at −80°C.

Generation of antibodies

Rabbits were immunized with purified GST–TMEM209 238–561 (Eurogentech). For purification of antibodies, MBP–TMEM209 238–561 was dialyzed against coupling buffer (0.1 M NaHCO₂ pH 8.3, 0.5 M NaCl) and coupled to cyanogen bromide (CnBr) activated Sepharose 4B (Sigma-Aldrich) for 90 min at room temperature. Excess ligand was removed by five repeats of alternating washing steps with acidic (0.1 M NaOAc pH 4.0, 0.5 M NaCl) and basic (0.1 M Tris-HCl pH 8.0, 0.5 M NaCl) washing buffers. Antisera were diluted 10-fold in PBS (10 mM Na₂HPO₄, pH 7.5, 1.8 mM KH₂PO₄, 137 mM NaCl, 2.7 mM KCl) and incubated overnight at 4°C with beads that had been equilibrated in PBS. After washing with 0.5 M NaCl, antibodies were eluted in 500 µl fractions using an acidic elution buffer (0.2 M NaOAc pH 2.7, 0.5 M NaCl in PBS) and immediately neutralized using 100 µl 1 M Tris base. Antibody-containing fractions were pooled and concentrated using a 6 ml Amicon^® 30.000 kDa MWCO centrifugal filter using sterile PBS for buffer exchange.

Cultivation and treatment of mammalian cells

HeLa P4 cells (Charneau et al., 1994) were obtained from the NIH AIDS Reagent Program. Cells were grown in Dulbecco's modified Eagle's medium (DMEM, Thermo Fisher Scientific) supplemented with 10% (v/v) fetal bovine serum (FBS, Thermo Fisher Scientific), 100 U ml⁻¹ penicillin, 100 µg ml⁻¹ streptomycin and 2 mM l-glutamine (Thermo Fisher Scientific) at 37°C in the presence of 5% CO₂ and tested for contamination by mycoplasma on a regular basis. HEK Flp-In™ T-REx™ 293 cells (Thermo Fisher Scientific) were grown in tetracycline-free medium. HAP1 cell lines (obtained from Horizon Discovery) were cultured in Iscove's Modified Dulbecco's Medium (IMDM, Thermo Fisher Scientific).

For SILAC, HeLa P4 cells were cultivated in DMEM for SILAC (Thermo Fisher Scientific) for five consecutive passages to incorporate isotope-labeled amino acids. DMEM for SILAC was supplemented with 10% FBS dialyzed against PBS containing 2 mM L-glutamine and 100 μg/ml of penicillin and streptomycin. For heavy isotope medium, 0.4 mM ¹³C₆¹⁵N₂-L-lysine and 0.2 mM ¹³C₆¹⁵N₄-L-arginine (Silantes, Munich, Germany) and for light isotope medium, 0.4 mM ¹²C₆¹⁴N₂-L-lysine and 0.2 mM ¹²C₆¹⁴N₄-L-arginine (Sigma-Aldrich) were supplemented. Furthermore, 0.173 mM L-proline (Sigma-Aldrich) was added to heavy and light isotope medium.

Transfections were performed following the calcium phosphate method. In brief, 4–12 µg plasmid DNA was first mixed with 300 µl CaCl₂ (250 mM) and subsequently with the same volume of HEPES buffer (50 mM HEPES, pH 6.98, 250 mM NaCl, and 1.5 mM NaHPO₄). After 20 min of incubation at room temperature, the transfection mix was added to the cells.

Rapamycin- and APEX-dependent identification of proteins by SILAC

RAPIDS was performed as previously described (James et al., 2022, 2019) with minor adaptations. In brief, HeLa P4 cells were cultured in 10-cm dishes using SILAC medium as described above. Cells were transfected with constructs coding for HA–FRB–TMEM209 and APEX2–dGFP–cNLS–FKBP12 and grown for 24 h. Next, 500 µM biotin-phenol (Iris Biotech, Marktredwitz, Germany) with or without 200 nM rapamycin (Enzo Life Sciences) was added to the plates, followed by an incubation of 30 min at 37°C and 5% CO₂ before adding 1 mM H₂O₂ for 1 min at room temperature. After two washing steps with quenching buffer (5 mM Trolox, 10 mM NaN3 and 10 mM sodium ascorbate in PBS) and another washing step with PBS to stop the biotinylation reaction, cells were lysed using RIPA buffer [50 mM Tris-HCl pH 7.4, 5 mM Trolox, 0.5% (w/v) sodium deoxycholate, 150 mM NaCl, 0.1% (w/v) sodium dodecyl sulfate, 1% (v/v) Triton X-100, 1 mM PMSF, 10 mM sodium azide, 10 mM sodium ascorbate, 1 μg/ml each of aprotinin, leupeptin and pepstatin] and lysates were cleared by centrifugation at 16,000 g for 30 min. The protein concentration was measured using the Pierce 660 nm Protein Assay (Thermo Fisher Scientific) and lysates were adjusted to equal concentrations in a total volume of 5 ml using RIPA buffer.

For western blot analysis, 2 ml lysate from each condition was incubated in two tubes with 130 µl Neutravidin beads (Thermo Fisher Scientific) each to enrich biotinylated proteins at 4°C overnight on a rotation wheel. Beads were washed for 8 min each with RAPIDS buffer I [50 mM HEPES pH 7.4, 0.1% (w/v) sodium deoxycholate, 1% (v/v) Triton X-100, 500 mM NaCl, 1 mM EDTA], RAPIDS buffer II [50 mM Tris-HCl pH 8.0, 250 mM LiCl, 0.5% (v/v) Nonidet P-40, 0.5% (w/v) sodium deoxycholate, 1 mM EDTA] and RAPIDS buffer III [50 mM Tris-HCl pH 7.4 and 50 mM NaCl; two washing steps] at 4°C. For the elution of biotinylated proteins, 100 μl hot 4× SDS sample buffer supplemented with 5 mM desthiobiotin was used to pool beads from both tubes.

For mass spectrometry analysis, the forward and reverse reactions were prepared by combining 3 ml of the respective lysates and were divided into six tubes per reaction, each containing 130 µl Neutravidin beads. Binding and washing steps were performed as described for western blot samples. Biotinylated proteins were eluted and pooled from three tubes using 100 μl hot 4× SDS sample buffer supplemented with 5 mM desthiobiotin, resulting in two samples of forward and reverse reactions. Lysates of cells that had been grown in light isotope medium with and heavy isotope medium without rapamycin are defined as the forward reaction, while the conditions for the reverse reaction are switched. Two technical replicates were performed per RAPIDS experiment.

Mass spectrometry

Samples were separated on 4–12% NuPAGE Novex Bis-Tris Minigels (Invitrogen). Gels were stained with Coomassie Blue, and each lane sliced into 11–12 equidistant bands. After washing, gel slices were reduced with dithiothreitol (DTT), alkylated with 2-iodoacetamide and digested in-gel with trypsin (sequencing grade, Promega, Madison Wisconsin) overnight (Atanassov and Urlaub, 2013). The resulting peptide mixtures were then extracted, dried in a SpeedVac, reconstituted in 2% acetonitrile/0.1% formic acid/ (v/v) and analyzed by nanoLC-MS/MS on a hybrid quadrupole/orbitrap mass spectrometer (Q Exactive, Thermo Fisher Scientific, Dreieich, Germany) as described previously (James et al., 2019). Raw data were processed using MaxQuant Software version 1.5.7.4 (Max Planck Institute for Biochemistry, Martinsried, Germany). Proteins were identified against the human reference proteome (v2017.02) along with a set of common lab contaminants. The Arginine R10 and Lysine K8 labels, including the ‘Re-quantify’ option, were specified for relative protein quantification. Perseus Software version 1.6.15.0 (Max Planck Institute for Biochemistry, Martinsried, Germany) was used for statistical evaluation of relative protein quantification values, and a two-sided significance B test was performed using normalized log2 ratios. For the analysis, a Benjamini–Hochberg correction was applied and a threshold value of 0.05 was chosen.

The MS proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository (Perez-Riverol et al., 2025) with the dataset identifier PXD069494.

Generation of stable cell lines

HEK293 Flp-InTM T-REx 293 cells (Thermo Fisher Scientific) were grown on six-well plates and co-transfected with pcDNA5-HA-TMEM209 and pOG44 (Flp-Recombinase expression, Thermo Fisher Scientific) at a ratio of 1:3, using X-tremeGENE™ 9 DNA (Sigma-Aldrich), following the instructions of the manufacturer. After 24 h, 100 μg/ml proteaSEycin B (Thermo Fisher Scientific) and 10 μg/ml blasticidin S (AppliChem) were added to start the selection process. Transgene expression was induced by the addition of tetracycline (0.1 µg/ml) and further incubation for 24 h.

Cell growth analysis

HAP1 parental and TMEM209-knockout cells were seeded in six-well plates. After 24 h, cells were transfected with control siRNA or siRNA targeting Nup210 or POM121 (see below). After 24 h, the medium was changed to fresh IMDM. Cells were counted using the Countess 3 automated cell counter (Thermo Fisher Scientific).

Cell synchronization

30,000 HeLa cells were seeded in 24-well plates and treated with siRNAs after 18 h. After 6 h, the medium was exchanged to DMEM containing 2 mM thymidine to block the cell cycle in G1/S phase. Cells were released from the thymidine block after 16 h by washing twice with warm PBS and a second transfection of siRNA was performed 2 h later. After 6 h of transfection, a second thymidine block was applied as before. Cells were collected for further analyses at selected time points after release from the second thymidine block.

Flow cytometry

Trypsinized cells were collected in 400 µl PBS and pelleted by centrifugation at 800 g for 3 min. After a second washing step with PBS, cells were resuspended in 500 µl cold PBS, transferred for fixation into FACS tubes filled with 2 ml ice-cold 70% ethanol and incubated for up to 5 days at 4°C. Cells were then centrifuged at 4°C at 4000 g for 5 min, washed twice in 500 μl PBS and resuspended in 500 μl PBS containing 0.1 mg/ml RNase A (Macherey-Nagel). After incubation overnight at 4°C, propidium iodide (Thermo Fisher Scientific) was added to a concentration of 0.03 μg/μl. Flow cytometry was performed using a BD FACSCanto II (Becton-Dickinson) and the FACS Diva 6.1.1 software. FlowJo Software (v10.8) was used for data evaluation.

RNAi-mediated knockdowns

Cells were seeded in 24-well plates and after 24 h, transfected with siRNA (siRNA non-targeting: 5′-UGGUUUACAUGUCGACUAA-3′; siRNA TMEM209#1: 5′-GCGAGUAGCAAGUCAUAUA-3′; TMEM209#3: 5′-AUAAUAAGUUGGCGAGCUU-3′; siRNA Nup210: 5′-CCGUGACGGUUUACUAUGA-3′; siRNA POM121: 5′-GCCAUCCAUCCUAUCUUU-3′) using Lipofectamin^® RNAiMAX (Thermo Fisher Scientific) according to the protocol of the manufacturer. The next day, medium was changed to remove the transfection reagent, and samples were collected after different time points.

Indirect immunofluorescence

Transfected or non-transfected cells were grown on coverslips and fixed with 3.7% formaldehyde diluted in PBS for 12 min at room temperature. Permeabilization was performed with 0.5% Triton X-100 in PBS at room temperature or with 0.0025% digitonin in PBS at 4°C for 5 min. Cells were washed three times with PBS, followed by 20 min of blocking with 3% BSA or 0.2% fish gelatin (Sigma-Aldrich) in PBS. Antibodies (Table S4) were diluted in blocking reagent and applied to cells on coverslips for one hour at room temperature, followed by washing steps and a 1-h incubation with Alexa Fluor^® secondary antibodies (Thermo Fisher Scientific). After washing, coverslips were mounted on microscopy glass slides using Moviol mounting media containing 1 µg/ml DAPI (Sigma-Aldrich) for DNA staining.

Proximity ligation assay

PLAs were performed using the Duolink^® in situ proximity ligation assay kit (Sigma-Aldrich). Cells were grown on coverslips and transfected as described above. Fixation was performed using 4% formaldehyde diluted in PBS and cells were incubated for 10 min at room temperature. Cells were permeabilized with 0.5% Triton X-100 in PBS for five minutes at room temperature and blocked for 30 min at 37°C with Duolink^® blocking solution. The coverslips were incubated with primary antibodies (Table S4) diluted in Duolink^® antibody diluent at room temperature for one hour, followed by two washing steps with PLA wash buffer A (0.01 M Tris-HCl pH 8.0, 0.15 M NaCl, 0.05% Tween-20) for 5 min. Cells were then incubated with PLA probes anti-mouse PLUS and anti-rabbit MINUS, diluted at a ratio of 1:5 in antibody diluent in a final volume of 40 μl, for 1 h at 37°C. After repeated washing steps as above, the PLA-ligation mix was added, and the coverslips were incubated at 37°C for 30 min. After washing, they were treated with the amplification mix for 100 min at 37°C, washed again twice with wash buffer B (0.2 M Tris-HCl pH 7.5, 0.1 M NaCl) and once with PLA wash buffer A. To identify proteins of interest, cells were counterstained using indirect immunofluorescence. Coverslips were mounted using Duolink^® mounting media containing DAPI.

Microscopy and image processing

Confocal microscopy

Cells were analyzed using either an LSM510 confocal laser scanning microscope with a 100× Plan-Neofluar 1.3 M27 oil immersion lens or an LSM780 confocal laser scanning microscope with a 63×/1.4 oil immersion lens with immersion oil Immersol™ 518 F (Zeiss, Germany). Image analysis was performed using the Zen System 3.1 (blue edition) software. In addition, microscopy images were taken using the Stellaris5 DMI8 confocal microscope (Leica) using either the HC PL APO CS2 63×/1.4 Oil or the HC PL FLUOTAR 10×/0.3 DRY objective. Photons were detected using the HyD S detector 1–3 (Leica) according to their respective emission wavelength and images were analyzed using the LAS-X software. Image processing was performed using Fiji Software (Schindelin et al., 2012). Images were processed by applying background subtraction using the rolling ball method and a Gaussian filter for image deblurring.

STED microscopy

For STED microscopy, images were captured on a Leica TCS SP8 STED DMI6 microscope using an HC PL APO CS2 100×/1.3 Oil objective and ImmersolTM 518 F oil immersion. Following dual-color immunofluorescence labeling, Abberior STAR 580 and STAR 635P fluorophores (see Table S4) were excited at 580 and 635 nm, respectively. STED depletion occurred at 775 nm, and photons were detected between 590–626 nm and 659–708 nm with a HyD S detector, respectively. Further acquisition parameters were: pixel size 16.23×16.23 nm, pixel dwell time 400 ns, scanning speed 600 Hz, 32× line averaging. Image processing and analysis was done using Fiji software. Line profiles were created using the RGB Profiler and signal resolutions were obtained from full-width at half-maximum values. For signal segmentation, Nup153 and Nup210 STED nanocscopy signals were analyzed for local maximum intensity positions in order to detect NPC center of mass positions. Nuclear HA-TMEM209 STED nanoscopy signals were segmented with the Phansalker method, a local thresholding technique, and subsequent Watershed processing. Based on these metrics, colocalization between HA–TMEM209 and NPCs was defined as HA segments occurring within a 60-nm radius around NPC center positions as identified by Nup153 signals, corresponding to NPC diameters of 120 nm. As a control for the colocalization analysis, images were rotated by 90° in one of the two channels.

Immunoprecipitation

For co-immunoprecipitation experiments, 1.5×10⁶ HeLa or HA–TMEM209-expressing HEK293 Flp-In T-Rex-cells were seeded in 10 cm dish plates. After 24 h, HeLa cells were transfected with plasmid DNA, whereas HEK293 cells were induced with 0.1 µg/ml tetracycline (Sigma-Aldrich) and, if necessary, transfected with a plasmid encoding Nup210–GFP. The next day, cells were lysed in 1 ml IP buffer [0.3% CHAPS (AppliChem), 20 mM Tris-HCl pH 7.4, 300 mM NaCl, 5 mM EDTA, 1× c0mplete protease inhibitor mixture (Roche Applied Science) and 1 mM DTT] and incubated for 20 min at 4°C. As an alternative to CHAPS, digitonin (1%; Serva) was used as a detergent for cell lysis. To reduce the viscosity, the lysate was passed 5 times through a 27-gauge (3/4-inch) needle. Lysates were centrifuged at 16,000 g for 30 min at 4°C and the total protein concentration of the supernatant was determined using the Pierce^® BCA protein assay kit. For immunoprecipitation, 40 µl of HA magnetic beads (Thermo Fisher Scientific) or GFP selector beads (NanoTec Biotechnologies) were equilibrated for 10 min in IP buffer. Protein lysates were loaded onto beads and incubated for 60–80 min at 4°C in a rotation wheel. Washing steps were performed five times for 3 min using IP buffer containing 0.3% CHAPS or 0.1% digitonin. Proteins were eluted with 60 µl of 1×NuPAGE LDS sample buffer containing 50 mM DTT at 95°C for 10 min.

Western blot analyses

Protein samples were loaded on 4–12% NuPAGE Bis-Tris gels (Thermo Fisher Scientific) or 8% Bis-Tris gels (1.4 M Bis-Tris-HCl pH 6.8, 8% polyacrylamide) and separated via gel electrophoresis using NuPAGE^® 1× MES SDS running buffer or 1× MOPS (50 mM Tris-HCl pH 7.8, 50 mM MOPS, 0.1% SDS) running buffer with 200 mM Tris-HCl pH 9.0 anode buffer, respectively. Proteins were transferred to a PVDF membrane (Millipore) in a Mini Trans-Blot^® Cell (Bio-Rad) using transfer buffer (25 mM Tris, 0.193 M glycine, 0.02% SDS, 20% ethanol) for 90 min at 100 V or 16 h at 20 V. Membranes were blocked for 1 h at room temperature in blocking buffer [5% milk powder (Sigma-Aldrich) in TBS (24.8 mM Tris-HCl pH 7.4, 137 mM NaCl, 2.7 mM KCl)] and washed twice with TBS containing 0.01% (v/v) Tween 20 (TBS-T). Membranes were incubated overnight at 4°C with primary antibodies (Table S4) in TBS-T, washed with TBST-T and incubated with IRDye© secondary antibodies [Table S4; 1:15,000, in TBS-T containing 0.01% (v/v) SDS] for 1 h at room temperature. The Odyssey^®SA Infrared Imaging System (LI-COR) was used for signal detection and images were analyzed using the ImageStudioLite 5.25 Software (LI-COR). See Fig. S8 for Western blot transparencies.

Peer review history

The peer review history is available online at https://journals.biologists.com/jcs/lookup/doi/10.1242/jcs.264534.reviewer-comments.pdf