Nuclear export of mRNAs is a critical regulatory step in eukaryotic gene expression. The mRNA transcript undergoes extensive processing, and is loaded with a set of RNA-binding proteins (RBPs) to form export-competent messenger ribonucleoprotein particles (mRNPs) in the nucleus. During the transit of mRNPs through the nuclear pore complex (NPC), the DEAD-box ATPase – DDX19 (herein referring to DDX19A and DDX19B) – remodels mRNPs at the cytoplasmic side of the NPC, by removing a subset of RNA-binding proteins to terminate mRNP export. This requires the RNA-dependent ATPase activity of DDX19 and its dynamic interactions with Gle1 and Nup214. However, the regulatory mechanisms underlying these interactions are unclear. We find that DDX19 gets covalently attached with a small ubiquitin-like modifier (SUMO) at lysine 26, which enhances its interaction with Gle1. Furthermore, a SUMOylation-defective mutant of human DDX19B, K26R, failed to provide a complete rescue of the mRNA export defect caused by DDX19 depletion. Collectively, our results suggest that SUMOylation fine-tunes the function of DDX19 in mRNA export by regulating its interaction with Gle1. This study identifies SUMOylation of DDX19 as a modulatory mechanism during the mRNA export process.

This article has an associated First Person interview with the first author of the paper.

mRNA transcripts undergo 5′ capping, splicing and 3′ polyadenylation in the nucleus before they are exported into the cytoplasm for translation (Stewart, 2019; Wende et al., 2019; Xie and Ren, 2019). Association of various RNA-binding proteins (RBPs) with mRNAs during and after transcription, forming nuclear messenger ribonucleoprotein particles (mRNPs), facilitates nuclear export and determines the fate of the mRNAs in terms of their stability, cytoplasmic localization and translation (Carmody and Wente, 2009; Dreyfuss et al., 2002; Maniatis and Reed, 2002). The bulk export of mRNAs is mediated by an evolutionary conserved heterodimeric transport receptor complex, NXF1–NXT1 (human homologs of Mex67p–Mtr2p in yeast) (Katahira, 2015; Katahira et al., 1999; Stutz and Izaurralde, 2003). An RNA-dependent ATPase, UAP56 (also known as DDX39B), facilitates nuclear assembly of NXF1–NXT1 heterodimer onto the mRNPs with the help of a multisubunit complex, THO, and the RBP ALY (also known as ALYREF) (Wende et al., 2019). The RBPs UAP56 and ALY, together with the THO complex, constitute the TRanscription EXport complex (TREX) (Wende et al., 2019). The export factor NXF1–NXT1 mediates the translocation of mRNPs through the nuclear pore complex (NPC) (Schmitt et al., 1999) and dissociates from the mRNPs at the cytoplasmic side of the NPC through the action of a DEAD-box ATPase, DDX19 (herein referring to DDX19A and DDX19B; Dbp5 in yeast), thus terminating the export process (Lund and Guthrie, 2005).

Nuclear export of mRNPs through the NPCs can be broadly divided into three distinct steps (Ben-Yishay et al., 2016) – initial surveillance and tethering of mRNPs at the nuclear basket by specific nucleoporins such as Tpr and Nup153 (Aksenova et al., 2020; Li et al., 2021; Xie and Ren, 2019); their translocation through the NPC by binding of NXF1–NXT1 with FG-containing nucleoporins (Grüter et al., 1998; Schmitt et al., 1999; Sträßer et al., 2000); and their remodeling and release at the cytoplasmic face of the NPC by the helicase DDX19 (Hodge et al., 2011; Montpetit et al., 2011; Noble et al., 2011; Rajan and Montpetit, 2021).

Human DDX19 is an RNA-dependent ATPase required for the terminal step of mRNA export (Lin et al., 2018; Schmitt et al., 1999). DDX19B consists of a short N-terminal extension [NTE, amino acids (aa) 1–90] with an auto-inhibitory α-helix (AIH, aa 54–68), followed by two RecA-like domains, referred to as the N-terminal domain (NTD) and C-terminal domain (CTD) (Fig. 1A) (Lin et al., 2018; Napetschnig et al., 2009). The AIH positions itself in a cleft formed between the NTD and CTD to keep the protein in a closed conformation (Collins et al., 2009; Lin et al., 2018). To attain an open conformation, for the ADP/ATP exchange and activation of DDX19, the N-terminal AIH needs to be removed from the cleft (Collins et al., 2009; Lin et al., 2018). Multiple dynamic interactions of DDX19 with Nup214 and Gle1 appear to contribute to the conformational changes required for the DDX19 catalytic cycle during mRNP remodeling (Rajan and Montpetit, 2021; Xie and Ren, 2019). The exact sequence of events in DDX19-mediated mRNP remodeling is unclear. However, biochemical and structural evidence support a model wherein, at the NPC, DDX19–Gle1 interaction is critical for DDX19 to attain an open conformation for the loading of ATP and/or RNA (Collins et al., 2009; Lin et al., 2018). Binding of ATP and RNA enables DDX19 to switch to the closed conformation for ATP hydrolysis, consequently remodeling and releasing the mRNPs from the NPC into the cytoplasm (Carmody and Wente, 2009; Folkmann et al., 2011). Subsequent binding of Nup214 to DDX19 may facilitate ADP release and prime DDX19 for the next round of mRNP remodeling (Rajan and Montpetit, 2021). The remodeling step releases the RBPs bound to the mRNAs required for their transit through the NPC, such as NXF1–NXT1, thereby terminating the export process (Carmody and Wente, 2009; Stewart, 2010).

Fig. 1.

DDX19 gets SUMOylated. (A) Schematic representation of human DDX19B. DDX19 has distinct domains [N-terminal extension (NTE); RecA-like 1 (N-terminal domain, NTD); RecA-like 2 (C-terminal domain, CTD)]. The NTE consists of an auto-inhibitory α-helix (AIH), preceding, which there is N-terminal (N-term) region. Amino acids are numbered. (B) HEK293T cells were co-transfected with GFP-tagged version of SUMO1G (non-conjugatable form) or SUMO1GG (mature, conjugatable form) and HA-tagged DDX19B. The cells were subjected to co-immunoprecipitation (co-IP) using GFP-specific antibody and the IP samples were analyzed by western blotting (WB) using indicated antibodies. Molecular mass markers are as indicated. Higher molecular mass bands indicative of SUMOylated DDX19 have been highlighted. # represents cross-reaction with antibody heavy chain. (C) HEK293T cells were co-transfected with GFP–DDX19B and Myc–SUMO1 or Myc–SUMO2 and subjected to co-IP with GFP-specific antibody. The IP samples were analyzed by WB using indicated antibodies. SUMOylated species of DDX19 have been highlighted. (D) Endogenous DDX19 was immunoprecipitated from HeLa cells and immunoblotted for SUMO1 and SUMO2/3 species using specific antibodies. Asterisk (*) indicates SUMOylated species of DDX19. # represents cross-reaction with antibody heavy chain. (E) HeLa cells treated with control (siControl) or DDX19 (siDDX19) siRNA were subjected to WB using indicated antibodies. GAPDH was used as loading control. (F) GST–DDX19B was subjected to in vitro SUMOylation using purified SUMO conjugation system in the presence (+) or absence (−) of ATP for indicated time points in hours (h). Asterisk (*) indicates SUMOylated species of DDX19. Blots shown are representative of three or four experiments.

Fig. 1.

DDX19 gets SUMOylated. (A) Schematic representation of human DDX19B. DDX19 has distinct domains [N-terminal extension (NTE); RecA-like 1 (N-terminal domain, NTD); RecA-like 2 (C-terminal domain, CTD)]. The NTE consists of an auto-inhibitory α-helix (AIH), preceding, which there is N-terminal (N-term) region. Amino acids are numbered. (B) HEK293T cells were co-transfected with GFP-tagged version of SUMO1G (non-conjugatable form) or SUMO1GG (mature, conjugatable form) and HA-tagged DDX19B. The cells were subjected to co-immunoprecipitation (co-IP) using GFP-specific antibody and the IP samples were analyzed by western blotting (WB) using indicated antibodies. Molecular mass markers are as indicated. Higher molecular mass bands indicative of SUMOylated DDX19 have been highlighted. # represents cross-reaction with antibody heavy chain. (C) HEK293T cells were co-transfected with GFP–DDX19B and Myc–SUMO1 or Myc–SUMO2 and subjected to co-IP with GFP-specific antibody. The IP samples were analyzed by WB using indicated antibodies. SUMOylated species of DDX19 have been highlighted. (D) Endogenous DDX19 was immunoprecipitated from HeLa cells and immunoblotted for SUMO1 and SUMO2/3 species using specific antibodies. Asterisk (*) indicates SUMOylated species of DDX19. # represents cross-reaction with antibody heavy chain. (E) HeLa cells treated with control (siControl) or DDX19 (siDDX19) siRNA were subjected to WB using indicated antibodies. GAPDH was used as loading control. (F) GST–DDX19B was subjected to in vitro SUMOylation using purified SUMO conjugation system in the presence (+) or absence (−) of ATP for indicated time points in hours (h). Asterisk (*) indicates SUMOylated species of DDX19. Blots shown are representative of three or four experiments.

Biochemical studies have revealed that DDX19 requires Gle1 for its activation (Dossani et al., 2009; Lin et al., 2018; Montpetit et al., 2011; Weirich et al., 2006). Gle1 may activate DDX19 primarily by removing the NTE containing the AIH from the cleft between NTD and CTD (Lin et al., 2018). Interestingly, deletion analyses have shown that removal of NTE from the intact protein makes DDX19 insensitive to Gle1-mediated ATPase activation, suggesting a critical role for the NTE in Gle1-dependency (Lin et al., 2018). However, whether and how DDX19 NTE contributes to Gle1-mediated activation of DDX19 is unclear. Here, we show that DDX19 gets SUMOylated in vitro and in cells at lysine 26 present in the NTE. Furthermore, we find that SUMOylation enhances the interaction between DDX19 and Gle1, and modulates the mRNA export process.

DDX19 can undergo SUMO modification

DDX19 is an RNA-dependent ATPase with multiple domains (Fig. 1A). During the mRNP remodeling cycle at the NPC, DDX19 interacts with different proteins and undergoes extensive structural changes. The NPC is known to harbor both SUMO modification and de-modification machineries (Palancade and Doye, 2008) – the SUMO E3 ligase Nup358 (also known as RANBP2) at the cytoplasmic side (Pichler et al., 2002), and the SUMO proteases SENP1 and SENP2 at the nuclear side (Chow et al., 2012; Goeres et al., 2011; Hang and Dasso, 2002; Zhang et al., 2002). Given this, we hypothesized that mRNA remodeling cycle of DDX19 might be regulated by SUMOylation. Initially, to test whether DDX19 gets SUMOylated, HA–DDX19B was co-expressed with GFP–SUMO1G (a non-conjugatable SUMO control) or GFP–SUMO1GG (a mature, conjugatable SUMO version) in HEK293T cells. The proteins modified with GFP–SUMO were immunoprecipitated (IP) with GFP-specific antibody. The IP samples were probed for the presence of SUMOylated form of HA–DDX19. High molecular mass bands that cross-reacted with HA antibody were detected only when mature SUMO1 was present and not in the presence of non-conjugatable SUMO1 (Fig. 1B), indicating that DDX19B gets SUMOylated in cells. Two distinct SUMO1-positive bands in HA probing were detected (Fig. 1B), indicating that DDX19 might be SUMOylated at least at two positions. To test the paralog-specificity of DDX19 SUMOylation, GFP–DDX19B was co-transfected with Myc control, Myc–SUMO1 or Myc–SUMO2 in HEK293T cells. Immunoprobing of the GFP–DDX19B IP samples with specific antibodies indicated that DDX19B could be modified by both SUMO1 and SUMO2 (Fig. 1C).

We further tested whether endogenous DDX19 is SUMOylated in cells. Immunoprobing of HeLa cell lysate with DDX19-specific antibodies identified two bands; one at ∼55 kDa, corresponding to endogenous full-length DDX19, and a higher molecular mass band of ∼80 kDa (Fig. 1D). These bands were also detected in the endogenous DDX19 IP samples when probed with DDX19 antibodies (Fig. 1D). Furthermore, the DDX19 IP samples showed higher molecular mass immunoreactive bands when probed with SUMO1 and SUMO2/3 antibodies (Fig. 1D). Moreover, siRNA-mediated depletion of DDX19 showed a decrease in the intensity of both the bands, indicating that they corresponded to DDX19 (Fig. 1E). SUMOylation of endogenous DDX19 was also verified in HEK293T cells (Fig. S1). An in vitro SUMOylation assay using bacterially expressed and purified GST–DDX19B, His–SUMO E1 (SAE1/SAE2), His–Ubc9 and His–SUMO1GG (Fig. S2) confirmed the SUMOylation of DDX19 (Fig. 1F). Collectively, these results suggest that DDX19 is SUMOylated in vitro and in cells.

DDX19 is SUMOylated at K26

Towards identifying the lysine (K) residue(s) that get SUMOylated in DDX19, the SUMO site prediction software Joint Advanced SUMOylation Site and SIM Analyser (JASSA) (Beauclair et al., 2015) with a ‘high cut-off’ setting was used. This software predicted K26, K29 and K31 as potential SUMOylation sites (Fig. 2A). All three predicted residues are highly conserved among vertebrates (Fig. 2A). As all the three lysine residues predicted were located within the N-terminal region, a GFP-tagged DDX19B fragment encompassing amino acids 1–309 residues was used for initial experiments to identify the SUMOylation site(s). These lysine residues were mutated individually to arginine (R) to identify the SUMOylation site(s) in DDX19. A triple mutant, wherein all the three lysine residues were mutated, was also generated. Co-immunoprecipitation assays using HEK293T cells co-expressing individual (K26R, K29R or K31R) or a triple (K26/K29/K31R) mutant version of GFP–DDX19B fragment (1–309 aa) and Myc–SUMO1 showed that K26 is the major SUMOylation site (Fig. 2B). Moreover, the K26R single mutation was sufficient to abolish SUMOylation of full-length DDX19 (Fig. 2C). Interestingly, both the SUMO1-positive bands were significantly reduced in the DDX19B-K26R mutant as compared to in wild-type DDX19B (DDX19B-WT), indicating that K26 is critical for the SUMOylation of DDX19. This was further confirmed by an in vitro SUMOylation assay (Fig. 2D). Collectively, our data suggest that K26 is an important SUMOylation site in DDX19.

Fig. 2.

DDX19B is SUMOylated at lysine 26. (A) Top panel, shows a DDX19B schematic highlighting the predicted lysine (K) residues at the N-terminal region (in red). Amino acid numbers are indicated. Bottom panel is a multiple alignment of the amino acid sequence surrounding the predicted lysine residues (in bold, underlined and highlighted in yellow) derived from human (h), mouse (m) and rat (r) DDX19A and DDX19B isoforms. The predicted lysine residues are in bold and underlined. (B) Top panel depicts the version of DDX19B fragment used in this experiment and the lysine residues targeted for mutation to arginine (R) in the N-terminal region. Bottom panel, HEK293T cells were co-transfected with GFP-tagged wild-type (WT) DDX19B N-terminal fragment (1-309 aa) or different indicated mutants (single or triple mutant) and Myc–SUMO1. The cells were subjected to co-IP with GFP-specific antibody and were analyzed by western blotting (WB) with anti-Myc antibody to monitor the SUMOylation status. SUMOylated DDX19 species have been indicated by asterisk (*). (C) HEK293T cells were co-transfected with HA-tagged version of WT or K26R mutant of full-length hDDX19B and GFP–SUMO1G or GFP–SUMO1GG. Cells were subjected to co-IP with anti-GFP antibody and WB with indicated antibodies. SUMOylated species of DDX19 are highlighted. # represents cross-reaction with antibody heavy chain. (D) In vitro SUMOylation assay for GST–DDX19B-WT and K26R mutant. DDX19 was identified by WB with GST antibody. Asterisk (*) indicates SUMOylated species of DDX19. Blots shown are representative of three or four experiments.

Fig. 2.

DDX19B is SUMOylated at lysine 26. (A) Top panel, shows a DDX19B schematic highlighting the predicted lysine (K) residues at the N-terminal region (in red). Amino acid numbers are indicated. Bottom panel is a multiple alignment of the amino acid sequence surrounding the predicted lysine residues (in bold, underlined and highlighted in yellow) derived from human (h), mouse (m) and rat (r) DDX19A and DDX19B isoforms. The predicted lysine residues are in bold and underlined. (B) Top panel depicts the version of DDX19B fragment used in this experiment and the lysine residues targeted for mutation to arginine (R) in the N-terminal region. Bottom panel, HEK293T cells were co-transfected with GFP-tagged wild-type (WT) DDX19B N-terminal fragment (1-309 aa) or different indicated mutants (single or triple mutant) and Myc–SUMO1. The cells were subjected to co-IP with GFP-specific antibody and were analyzed by western blotting (WB) with anti-Myc antibody to monitor the SUMOylation status. SUMOylated DDX19 species have been indicated by asterisk (*). (C) HEK293T cells were co-transfected with HA-tagged version of WT or K26R mutant of full-length hDDX19B and GFP–SUMO1G or GFP–SUMO1GG. Cells were subjected to co-IP with anti-GFP antibody and WB with indicated antibodies. SUMOylated species of DDX19 are highlighted. # represents cross-reaction with antibody heavy chain. (D) In vitro SUMOylation assay for GST–DDX19B-WT and K26R mutant. DDX19 was identified by WB with GST antibody. Asterisk (*) indicates SUMOylated species of DDX19. Blots shown are representative of three or four experiments.

SUMOylation of DDX19 occurs in a Nup358-dependent manner

The DDX19-mediated remodeling occurs at the cytoplasmic side of the NPC. The nucleoporin Nup358, a SUMO E3 ligase, is also positioned at the cytoplasmic side. We tested the possibility of Nup358 mediating the SUMOylation of DDX19. Consistent with this idea, depletion of Nup358 resulted in decreased level of endogenous SUMOylated DDX19 (Fig. 3A). Moreover, a co-immunoprecipitation assay confirmed an interaction between HA–DDX19B and endogenous Nup358 (Fig. 3B). To delineate the region of Nup358 involved in DDX19 interaction, different fragments of Nup358 were individually co-expressed with HA–MBP control or HA–DDX19B in HEK293T cells, followed by immunoprecipitation using HA-specific antibody. The results showed that the N-terminal region of Nup358 (Nup358-N) specifically interacts with DDX19 (Fig. 3C).

Fig. 3.

DDX19 SUMOylation is dependent on Nup358. (A) HeLa cells were transfected with control (siControl) or Nup358 (siNup358) specific siRNA and were analyzed for the extent of Nup358 depletion and SUMOylation of endogenous DDX19 by western blotting (WB) (left panels). Vinculin and GAPDH were used as loading controls for Nup358 and DDX19 levels, respectively. SUMOylated form of DDX19 is indicated by asterisk (*). Quantitative data representing the relative level of SUMOylated DDX19 normalized to the total DDX19 is plotted. Data are presented as mean±s.e.m. for n=3 independent experiments. ***P<0.001 (two-tailed unpaired Student's t-test). (B) HEK293T cells were transfected with HA–DDX19B, and endogenous Nup358 was immunoprecipitated using specific antibodies. The IP samples were immunoprobed with indicated antibodies. Arrow indicates HA–DDX19B, and # represents cross-reaction with antibody heavy chain. (C) HEK293T cells were transfected with HA–DDX19B or HA–MBP control and individual GFP-tagged fragments of Nup358 (Nup358-N, Nup358-M or Nup358-C) as mentioned in the schematic diagram (top panel, LRR, Leucine-Rich Region; R, RanGTP-binding domain; ZnF, Zinc Finger domains; IR, internal repeats; CHD, cyclophilin homology domain). The cells were subjected to IP with HA-specific antibody. The IP samples were probed with indicated antibodies (bottom panel). α-Tubulin was used as loading control. (D) In vitro SUMOylation of GST–DDX19 was performed using SUMO conjugation system with (+) or without (−) Nup358 internal repeat (IR, schematic diagram shown in top panel) (Magre et al., 2019) that possesses SUMO E3 ligase activity in the presence (+) or absence (−) of ATP as indicated (bottom panel). SUMOylated DDX19 species have been shown by asterisk (*). Blots shown are representative of three or four experiments.

Fig. 3.

DDX19 SUMOylation is dependent on Nup358. (A) HeLa cells were transfected with control (siControl) or Nup358 (siNup358) specific siRNA and were analyzed for the extent of Nup358 depletion and SUMOylation of endogenous DDX19 by western blotting (WB) (left panels). Vinculin and GAPDH were used as loading controls for Nup358 and DDX19 levels, respectively. SUMOylated form of DDX19 is indicated by asterisk (*). Quantitative data representing the relative level of SUMOylated DDX19 normalized to the total DDX19 is plotted. Data are presented as mean±s.e.m. for n=3 independent experiments. ***P<0.001 (two-tailed unpaired Student's t-test). (B) HEK293T cells were transfected with HA–DDX19B, and endogenous Nup358 was immunoprecipitated using specific antibodies. The IP samples were immunoprobed with indicated antibodies. Arrow indicates HA–DDX19B, and # represents cross-reaction with antibody heavy chain. (C) HEK293T cells were transfected with HA–DDX19B or HA–MBP control and individual GFP-tagged fragments of Nup358 (Nup358-N, Nup358-M or Nup358-C) as mentioned in the schematic diagram (top panel, LRR, Leucine-Rich Region; R, RanGTP-binding domain; ZnF, Zinc Finger domains; IR, internal repeats; CHD, cyclophilin homology domain). The cells were subjected to IP with HA-specific antibody. The IP samples were probed with indicated antibodies (bottom panel). α-Tubulin was used as loading control. (D) In vitro SUMOylation of GST–DDX19 was performed using SUMO conjugation system with (+) or without (−) Nup358 internal repeat (IR, schematic diagram shown in top panel) (Magre et al., 2019) that possesses SUMO E3 ligase activity in the presence (+) or absence (−) of ATP as indicated (bottom panel). SUMOylated DDX19 species have been shown by asterisk (*). Blots shown are representative of three or four experiments.

Next, we wanted to test whether Nup358 can act as SUMO E3 ligase for DDX19 under in vitro conditions. Previous studies had shown that in vitro SUMOylation of substrates can be achieved even in the absence of E3 ligase by using a relatively higher concentration of the E2 enzyme Ubc9 in the reaction (Pichler et al., 2004). However, under limiting amount of Ubc9, the SUMOylation reaction depends on the presence of E3 ligase such as Nup358-IR (Pichler et al., 2004). To examine whether DDX19 SUMOylation depended on Nup358-IR in vitro, we set up the reaction with a 10-fold lower amount of Ubc9 as compared to the one where no E3 ligase was used (Figs 1F and 2D; and see Materials and Methods section for details). Under this condition, DDX19 SUMOylation was detected in the presence, but not in the absence, of Nup358-IR (Fig. 3D), indicating that Nup358-IR can act as an E3 ligase in vitro. Collectively, these data support the conclusion that SUMOylation of DDX19 is dependent on Nup358.

SUMOylated DDX19 is enriched at the nuclear envelope

Next, we investigated the functional relevance of DDX19 SUMOylation. In many instances, SUMOylation can dictate the subcellular distribution of the target proteins (Flotho and Melchior, 2013). Moreover, DDX19 has been shown to be present in the nucleus, nuclear envelope (NE) and cytoplasm (Rajan and Montpetit, 2021). To test whether SUMOylation affected the localization of DDX19, HeLa cells were fractionated. The nuclear and cytoplasmic fractions were analyzed for the presence of DDX19 by immunoblotting. Interestingly, SUMO-modified DDX19 was relatively enriched in the nuclear fraction as compared to the unmodified protein (Fig. 4A). The presence of the nuclear (lamin A/C) and cytoplasmic (vinculin) markers in the respective fractions confirmed the authenticity of fractionation (Fig. 4A). However, the nuclear fraction contains proteins present both within the nucleus and at the NE. To test whether SUMOylation targeted DDX19 differentially to the NE, nucleoplasmic and NE contents were subfractionated from the nuclear fraction and were monitored for the presence of DDX19. The results showed that SUMOylated DDX19 is relatively enriched at the NE (Fig. 4B).

Fig. 4.

SUMOylation enriches DDX19 at the nuclear envelope and enhances its interaction with Gle1. (A) HeLa cells were subjected to nuclear and cytoplasmic fractionation. Relative abundance of endogenous DDX19 was monitored in these fractions by western blotting (WB). Unmodified and SUMO-modified DDX19 are highlighted with an arrow and asterisk (*), respectively. Lamin A/C and vinculin were used as markers for nuclear and cytoplasmic fractions, respectively. (B) HeLa cells were subjected to nucleoplasmic, nuclear envelope (NE) and cytoplasmic fractions as described in Materials and Methods. The presence of endogenous DDX19 was examined by WB in different indicated fractions. The fractionation was verified by immunoblotting with nucleoplasmic (RNA polymerase II), NE [SUMOylated RanGAP1 (SUMO–RanGAP1)] and cytoplasmic (vinculin) markers as indicated. Unmodified and SUMO-modified DDX19 are shown by arrow and asterisk (*), respectively. Blots shown in A and B are representative of three or four experiments. (C) HEK293T cells were co-transfected with GFP–MBP control or GFP–Nup214 and HA–DDX19B-WT or HA–DDX19B-K26R mutant constructs. The cells were subjected to IP with anti-GFP antibody and the IP samples were analysed for the presence of indicated proteins with specific antibodies. Actin was used as loading control. The relative levels of HA–DDX19B (WT or K26R) co-immunoprecipitated with GFP-Nup214 were quantitated and plotted. Data are presented as mean±s.e.m. for n=3 independent experiments. (D) HEK293T cells were transfected with GFP–Gle1B and HA–MBP control, HA–DDX19B-WT or DDX19B–K26R mutant constructs. Cells were then subjected to IP using HA-specific antibodies and the IP samples were analyzed for the presence of indicated proteins by WB (left panel). The relative levels of GFP-Gle1B co-immunoprecipitated with HA-tagged proteins were quantitated and plotted. Data are presented as mean±s.e.m. for n=5 independent experiments. **P<0.01; n.s., not significant (two-tailed unpaired Student's t-test).

Fig. 4.

SUMOylation enriches DDX19 at the nuclear envelope and enhances its interaction with Gle1. (A) HeLa cells were subjected to nuclear and cytoplasmic fractionation. Relative abundance of endogenous DDX19 was monitored in these fractions by western blotting (WB). Unmodified and SUMO-modified DDX19 are highlighted with an arrow and asterisk (*), respectively. Lamin A/C and vinculin were used as markers for nuclear and cytoplasmic fractions, respectively. (B) HeLa cells were subjected to nucleoplasmic, nuclear envelope (NE) and cytoplasmic fractions as described in Materials and Methods. The presence of endogenous DDX19 was examined by WB in different indicated fractions. The fractionation was verified by immunoblotting with nucleoplasmic (RNA polymerase II), NE [SUMOylated RanGAP1 (SUMO–RanGAP1)] and cytoplasmic (vinculin) markers as indicated. Unmodified and SUMO-modified DDX19 are shown by arrow and asterisk (*), respectively. Blots shown in A and B are representative of three or four experiments. (C) HEK293T cells were co-transfected with GFP–MBP control or GFP–Nup214 and HA–DDX19B-WT or HA–DDX19B-K26R mutant constructs. The cells were subjected to IP with anti-GFP antibody and the IP samples were analysed for the presence of indicated proteins with specific antibodies. Actin was used as loading control. The relative levels of HA–DDX19B (WT or K26R) co-immunoprecipitated with GFP-Nup214 were quantitated and plotted. Data are presented as mean±s.e.m. for n=3 independent experiments. (D) HEK293T cells were transfected with GFP–Gle1B and HA–MBP control, HA–DDX19B-WT or DDX19B–K26R mutant constructs. Cells were then subjected to IP using HA-specific antibodies and the IP samples were analyzed for the presence of indicated proteins by WB (left panel). The relative levels of GFP-Gle1B co-immunoprecipitated with HA-tagged proteins were quantitated and plotted. Data are presented as mean±s.e.m. for n=5 independent experiments. **P<0.01; n.s., not significant (two-tailed unpaired Student's t-test).

SUMOylation of DDX19 enhances its interaction with Gle1

It has been shown that SUMOylation, similar to other post-translational modifications, can alter the protein interaction network and intracellular localization of the target protein (Flotho and Melchior, 2013). Two well-characterized interactors of DDX19 are Nup214 and Gle1 (Rajan and Montpetit, 2021). The effect of DDX19 SUMOylation on its interaction with Nup214 and Gle1 was tested. GFP–Nup214 was co-expressed with HA–DDX19B-WT or HA–DDX19B-K26R mutant in HEK293T cells, and a co-immunoprecipitation assay was performed. There was no discernible difference in the interaction of WT or SUMOylation-defective mutant with Nup214, indicating that SUMOylation might not affect the interaction of DDX19 with Nup214 (Fig. 4C). Interestingly, a significant reduction in the interaction of the SUMOylation-defective DDX19 mutant with Gle1B was observed as compared to with DDX19B-WT (Fig. 4D), indicating that SUMOylation enhances the ability of DDX19 to interact with Gle1 in cells.

SUMOylation fine-tunes the mRNA export function of DDX19

Next, we investigated whether SUMOylation regulates the function of DDX19 during the mRNA export process. Transfection of U2OS cells with DDX19A/B-specific siRNA (siDDX19) resulted in a significant depletion of endogenous DDX19 (Fig. 5A). We monitored the mRNA export process in DDX19-deficient cells by analyzing the ratio of the oligo(dT) signals in the nucleus over the total intensity in that cell. As expected, the relative ratio increased significantly when endogenous DDX19 was depleted (Fig. 5B,C), indicating an mRNA export blockage. The defect in mRNA export was rescued completely by expressing siRNA-resistant version of human wild-type DDX19, but not the DDX19 SUMO-defective mutant (K26R). Interestingly, the DDX19 SUMO defective mutant partially rescued the defect in mRNA export caused by DDX19 depletion (Fig. 5B,C). However, the incomplete rescue of the mRNA export dysfunction in DDX19-deficient cells by the DDX19 SUMO mutant indicates that SUMOylation has a modulatory role in the functioning of DDX19 during mRNA export.

Fig. 5.

SUMO-defective K26R mutation interferes with the mRNA export function of DDX19. (A) U2OS cells were transfected with siControl or siDDX19 and the extent of depletion was monitored by western blotting (WB) using DDX19 antibody. GAPDH was used as loading control. Blot shown is representative of three experiments. (B) Confocal microscopic images showing the relative distribution of poly(A) RNA in cells. U2OS cells were first transfected with HA–MBP control or the siRNA-resistant version of HA–DDX19-WT or HA–DDX19-K26R (red) mutant construct and then with siControl or siDDX19, as indicated. Cells were then fixed and in situ oligo(dT) (green) staining was performed as described in the Materials and Methods section. HA-tagged protein expression was monitored by immunostaining with HA-specific antibody (red). Hoechst 33342 dye (blue) was used to stain the DNA. Scale bars: 20 µm. (C) Quantitative data for the experiments performed in B. Ratio of oligo(dT) intensity in the nucleus relative to the total intensity per cell was calculated and plotted (n=60 cells for each of siControl+HA-MBP-control, siDDX19+ HA-MBP-control, siDDX19+HA-DDX19-WT and siDDX19+HA-DDX19-K26R conditions, derived from four independent experiments). Data represented as mean±s.d. ***P<0.001; n.s., not significant (Mann–Whitney test).

Fig. 5.

SUMO-defective K26R mutation interferes with the mRNA export function of DDX19. (A) U2OS cells were transfected with siControl or siDDX19 and the extent of depletion was monitored by western blotting (WB) using DDX19 antibody. GAPDH was used as loading control. Blot shown is representative of three experiments. (B) Confocal microscopic images showing the relative distribution of poly(A) RNA in cells. U2OS cells were first transfected with HA–MBP control or the siRNA-resistant version of HA–DDX19-WT or HA–DDX19-K26R (red) mutant construct and then with siControl or siDDX19, as indicated. Cells were then fixed and in situ oligo(dT) (green) staining was performed as described in the Materials and Methods section. HA-tagged protein expression was monitored by immunostaining with HA-specific antibody (red). Hoechst 33342 dye (blue) was used to stain the DNA. Scale bars: 20 µm. (C) Quantitative data for the experiments performed in B. Ratio of oligo(dT) intensity in the nucleus relative to the total intensity per cell was calculated and plotted (n=60 cells for each of siControl+HA-MBP-control, siDDX19+ HA-MBP-control, siDDX19+HA-DDX19-WT and siDDX19+HA-DDX19-K26R conditions, derived from four independent experiments). Data represented as mean±s.d. ***P<0.001; n.s., not significant (Mann–Whitney test).

Here, we found that DDX19 gets SUMOylated. This is consistent with the earlier proteomic analysis of SUMOylated proteins identifying DDX19 as a target (Hendriks and Vertegaal, 2016; Mojsa et al., 2021). We discovered a role for SUMOylation in modulating the mRNA export function of DDX19 by enhancing and/or stabilizing its interaction with Gle1. Previous studies have highlighted the possible mechanism by which Gle1, in association with Nup42, activates DDX19. Structural studies of DDX19 with and without different binding factors such as ADP, ATP, Gle1 and Nup214 have shed light on the mechanism of its ATPase activation (Collins et al., 2009; Hodge et al., 2011; Lin et al., 2018; Montpetit et al., 2011; Napetschnig et al., 2009; Noble et al., 2011). Moreover, DDX19 has an ATP-dependent RNA-binding and RNA-dependent ATPase activity (Stewart, 2010), which are required for mRNP remodeling at the cytoplasmic face of the NPC to terminate the mRNA export process (Rajan and Montpetit, 2021). Consistent with this, previous in vitro studies have shown that the ATP-bound form of DDX19 (or Dbp5) can interact with mRNAs, which subsequently triggers the ATPase activity (Ledoux and Guthrie, 2011).

In contrast to the yeast DDX19 homolog Dbp5, human DDX19 has an AIH, which binds to the cleft formed between the NTD and CTD to prevent its ATPase and RNA-binding activities (Collins et al., 2009; Lin et al., 2018). Gle1 is a crucial player in activating DDX19 in vitro and in cells (Rajan and Montpetit, 2021). Gle1 binding has been shown to induce conformational changes in order to stimulate the ATPase activity. One of the main functions of Gle1 appears to release the AIH from the groove and/or stabilize a conformation leading to its activation (Lin et al., 2018). However, how this happens at the molecular level is unclear. Interestingly, the NTE harbors an N-terminal (N-term) region (Fig. 1A) preceding the AIH, but whether it plays a direct or indirect role in Gle1-mediated activation of DDX19 is unclear. Previous studies have shown that removal of the N-terminus (1–53 aa) increased the ATPase activity (Collins et al., 2009; Lin et al., 2018), but made it resistant to Gle1-stimulated activation in vitro (Lin et al., 2018). Interestingly, our studies show that SUMO-modification of DDX19 occurs at K26 within the N-terminal region, which could enhance its interaction with Gle1. Based on the current understanding, this finding provides a possible mechanism for the DDX19 activation cycle. We speculate that the SUMOylation of DDX19 at the N-terminus possibly contributes to the release of AIH, in a Gle1-dependent manner. Mechanistically, SUMOylation may help the release of the AIH from the cleft and/or stabilize a conformation that favors an interaction with Gle1, thereby leading to the activation of DDX19.

The finding that the mRNA export defect caused by depletion of endogenous DDX19A and DDX19B, was not completely rescued by the DDX19B-K26R mutant supports the conclusion that SUMOylation plays an important role in the regulation of the DDX19 function in mRNA export (Fig. 5B,C). As expected, DDX19B-WT completely rescued the mRNA export defect caused by depletion of endogenous DDX19A and DDX19B. If SUMOylation is a process that is critical for the activation of DDX19, one would expect that the DDX19B-K26R would not rescue the phenotype at all. Some of the possible explanations that support the partial rescue of bulk mRNA export by DDX19-K26R mutant are as follows. Firstly, SUMOylation is required for DDX19 function, but in case of the DDX19-K26R mutant, there could be other lysine residue(s) that become SUMO-modified, providing a functional replacement for K26 SUMOylation. In support of this, we do detect at least two SUMO1-positive bands for DDX19 (Figs 1B and 2C), and K26R mutation significantly reduces the signal for both the bands (Fig. 2C), consistent with K26 being the major SUMOylation site. However, we cannot rule out the possibility of SUMOylation of another site contributing to the functionality of DDX19. Secondly, it is possible that SUMOylation at K26R is not an absolute requirement for DDX19 to mediate the mRNA export, but could have a modulatory role. In fact, an earlier study showed that interfering with the SUMO pathway had a pronounced effect on the export of a few specific mRNAs as compared to the bulk mRNA export (Zhang et al., 2014). Based on this and our results, it is plausible that SUMOylation of DDX19 at K26R is required for the export of a subset of mRNAs in addition to fine-tuning the bulk mRNA export kinetics. As endogenous SUMO-modified DDX19 was enriched at the NE (Fig. 4B), and the K26R mutation compromised the interaction of DDX19 with Gle1 (Fig. 4D), it is possible that SUMOylation may be involved in the dynamic association of DDX19 at the NPCs through interaction with Gle1. Recently, it was shown that MKRN2, an RNA-binding ubiquitin E3 ligase, physically interacts with Gle1 and negatively regulates the export of a subset of mRNAs (Wolf et al., 2020). Although the molecular mechanism by which MKRN2 exerts this effect in unclear, it points to the possibility of mRNA export regulation by ubiquitylation. Thus, multiple post-translational modifications of key players could be involved in regulating the dynamic process of mRNA export.

The SUMOylation site K26R and the surrounding amino acid sequence are well conserved in human, mouse and rat DDX19 (Fig. 2A). There are potential SUMOylation sites in the zebrafish (z)DDX19 within the NTE as well. Moreover, we found that DDX19 SUMOylation is dependent on Nup358, a nucleoporin possessing SUMO E3 ligase activity that is present at the cytoplasmic side of the NPC (Fig. 3). We had previously shown that zNup358 also harbors SUMO E3 ligase activity (Magre et al., 2019). As Nup358 is evolutionarily conserved in metazoans but absent in yeast (Ciccarelli et al., 2005), it appears that the Nup358-dependent DDX19-SUMOylation could be a metazoan-specific phenomenon. A careful analysis of Dbp5, the yeast homolog of DDX19, indicated at least two potential SUMOylation sites (52PKVE55; K53 and 15LKID18; K16) at its N-terminal region, as predicted by the JASSA program (Beauclair et al., 2015). Therefore, it is possible that SUMOylation of Dbp5 could play role in mRNA export in lower eukaryotes, including yeast. In such case, a SUMO E3 ligase other than Nup358 may be involved in the modification of Dbp5. Interestingly, our results indicate that, at least in higher eukaryotes, DDX19-mediated mRNP remodeling at the NPC might involve a SUMO modification–demodification cycle mediated by the NPC-associated SUMO-conjugation system (Nup358) (Pichler et al., 2004) and deconjugation system [SUMO proteases, SENP1, SENP2) (Palancade and Doye, 2008)].

Recent studies using high-resolution imaging have shown that Nup358 interacts with NXF1 in a DDX19-dependent manner, and this NPC-resident pool of NXF1 has been implicated in the final release of the mRNA into the cytoplasm (Ben-Yishay et al., 2019). Given this, it is interesting to look at the possible interconnections between Nup358-mediated regulation of DDX19 through its SUMOylation and NXF1 function at the cytoplasmic face of NPC during mRNA export.

Plasmids

GFP-hDDX19B-WT was a kind gift from the late Elisa Izaurralde (Max Planck Institute of Developmental Biology, Germany). HA-DDX19B was generated by subcloning the DDX19B open reading frame (ORF) into a modified pCI-neo vector (Promega). Different single (K26R, K29R or K31R) and triple (K26R/K29R/K31R) DDX19B mutants were generated by a PCR-based method and verified by sequencing. GFP–SUMO1G and GFP–SUMO1GG constructs were as described previously (Yadav et al., 2016). Both pCMV-myc-SUMO1 and pCMV-myc-SUMO2 were kind gifts from Michel Goossens, INSERM, France. The Myc empty vector was generated from pCMV-myc-SUMO1 by restriction digestion to release the entire SUMO1 insert followed by endfilling and self-ligation of the vector backbone.

The constructs expressing GST–DDX19-WT and GST–DDX19B-K26R mutant were generated by subcloning the respective ORFs from GFP–DDX19B-WT and GFP–DDX19B-K26R mutant constructs into pGEX-6P vector (GE Healthcare). GFP–DDX19B (1–309 aa) was generated by removing the remaining C-terminal region (310–479 aa) with appropriate restriction digestion. The siRNA-resistant DDX19B-WT and K26R constructs were generated by changing specific nucleotides without altering the amino acid sequence in the siRNA-binding site of DDX19A and DDX19B mRNA by a PCR-based method. The original DNA sequence corresponding to the siRNA-binding is: 5′-TCAACAAGCTGATCAGAAG-3′, and the siRNA-resistant sequence is: 5′-TGAATAAACTCATTAGGAG-3′ (changed nucleotides are underlined).

The ORF for human Nup214 from pCDX-HA-Nup214 (generously provided by Gerald Grosveld, St. Jude Children's Research Hospital, Memphis, USA) was subcloned into pEGFP-C2 (Clontech) to generate GFP–Nup214. The GFP–Gle1B construct was made by PCR amplifying the Gle1B ORF using the cDNA made from HEK293T cells and cloning into pEGFP-C3 vector (Clontech). The construct was verified by sequencing.

Constructs expressing His-tagged Nup358-IR and other proteins used in in vitro SUMOylation assay were described previously (Magre et al., 2019). GFP–MBP and HA–MBP control constructs were reported previously (Sahoo et al., 2017). GFP-tagged version of Nup358-N (1–993 aa), Nup358-M (901–2220 aa) and Nup358-C (2221–3224 aa) constructs were as described previously (Joseph and Dasso, 2008).

Cell culture and transfections

HEK293T, HeLaS3 and U2OS cell lines (ATCC, Manassas, VA, USA) were routinely tested for mycoplasma contamination. Cells were maintained in Dulbecco's modified Eagle's medium (DMEM) (Gibco, #12100-046) supplemented with 10% fetal bovine serum (FBS) (Sigma, #F2442-500ML) and 10 µg/ml antibiotics Ciprofloxacin (Cipla Ltd., India) at 37°C in a humidified incubator maintaining 5% CO2. Trypsin phosphate versene glucose (TPVG; HiMedia, #TCL022) was used for trypsinization of the cells.

Cells were seeded in an 100 mm dish (107 cells/dish), six-well plate (1.5×106 cells/well) and a 24-well plate (2×104 cells/well). HEK293T and U20S cells were transfected with Polyethylenimine (PEI, linear, MW ∼25,000; Polysciences Inc., #23966), following the manufacturer's protocol. siRNA transfection was performed in HeLa and U2OS cells using Lipofectamine RNAiMAX (Invitrogen, #13778150). All siRNAs were used at a final concentration of 40 nM. The siRNAs were designed against the following target sequences: control siRNA 5′-TTCTCCGAACGTGTCACGT-3′ (Joseph et al., 2004), Nup358 siRNA 5′-GGACAGTGGGATTGTAGTG-3′ (Sahoo et al., 2017) and DDX19A and DDX19B siRNA 5′-TCAACAAGCTGATCAGAAG-3′ (Rajakylä et al., 2015).

Immunoprecipitation and western blotting

Cells were washed with chilled tris-buffered saline (TBS) and collected using a scraper into a microcentrifuge tube. For all the SUMOylation-related western blotting and immunoprecipitation experiments, the cell pellets were re-suspended in chilled RIPA lysis buffer (50 mM Tris-HCl pH 8.0, 150 mM NaCl, 1% NP-40, 0.1% SDS, 0.5% sodium deoxycholate, 1 mM EDTA) supplemented with 7.5 mM sodium fluoride (NaF) (Qualigens, #Q27645), 0.75 mM sodium orthovanadate (Na3VO4) (Sigma, #S6508), 1 mM PMSF (USB, #20203), protease inhibitor cocktail (PIC) (Roche, #11714600) and 20 mM N-ethylmaleimide (NEM) (Sigma-Aldrich, #E3876-25G). Lysate was centrifuged at 12,581 g for 20–30 min and the supernatant was collected. An equal volume of the control and test lysates were added to Protein A–Sepharose beads (Invitrogen, #101041) or Protein A/G–agarose beads (BioBharati LifeScience Pvt. Ltd., # BB-PAG001D) pre-bound with control IgG and test IgG for 2 h at room temperature (RT) on a rotospin. In the present study, antibodies used for IP were rabbit anti-GFP [2 μg/IP, in-house (Sahoo et al., 2012)], rabbit anti-Nup358 antibody [10 μg/IP, in-house (Joseph et al., 2004)] and rabbit anti-DDX19 antibody (2 μg/IP, cat. no. A300-547A, Bethyl Laboratories). In case of HA-IP, agarose beads prebound with mouse anti-HA antibody (Sigma, #A2095) were used. For IP, lysates were incubated with the antibody pre-bound to the beads on rotospin at 4°C for 40 min to 2 h. The beads were washed twice with chilled lysis buffer and then once with chilled TBS. Proteins were eluted from the beads by adding 3× SDS dye and heating at 95°C for 6 min.

For the interaction studies involving endogenous Nup358, fragments of Nup358 and HA-DDX19B (Fig. 3B,C), cell lysis was performed in NP-40 buffer (20 mM Tris-HCl pH 8.0, 137 mM NaCl, 10% glycerol, 1% NP-40, 2 mM EDTA). For endogenous Nup358 IP (Fig. 3B), the last wash was performed with RIPA buffer.

Protein samples were separated by SDS-PAGE and transferred onto PVDF membrane (Millipore, #IPVH00010) using Semi-Dry Transfer Apparatus (TE77 semidry transfer unit, GE Healthcare). Blots were probed with primary antibodies in blocking buffer 0.5% BSA in TBS with 0.1% Tween 20 (TBST) for 2 to 5 h at RT or overnight at 4°C. This was followed by three washes with TBST for 3 min each. The blots were then incubated with HRP-conjugated secondary antibody, diluted in TBST containing 0.5% BSA, for 1 h at RT, followed by washing three times with TBST for 3 min each at RT. The membranes were then incubated with the substrate provided by an ECL-Western blotting kit (Thermo Scientific, #32132) or WESTAR ηC 2.0 (Cyanagen, #XLS070,0250) using the manufacturer's instructions, and the images were captured using a IMAGE-QUANT LAS 4000 machine (GE Healthcare). Densitometric analysis of the western blot was done using Multi Gauge V3.0 software.

Antibodies and the dilutions used for western blotting were: mouse anti-GFP (1:5000, Santa Cruz Biotechnology, sc-9996), rabbit anti-DDX19 (1:5000, Bethyl Laboratories, A300-547A), mouse anti-c-Myc (1:3000, Santa Cruz Biotechnology, sc-40), mouse anti-RNA Polymerase II (1:2000, Santa Cruz Biotechnology, sc-5676), mouse anti-RanGAP1 (1:1000, Santa Cruz Biotechnology, sc-28322), rabbit anti-SUMO1 (1:3000, Cell Signaling Technology, 4930S), rabbit anti-SUMO2/3 (1:3000, Cell Signaling Technology, 4971S), mouse anti-GST (1:3000, Santa Cruz Biotechnology, sc-138), mouse anti-lamin A/C (1:3000, Santa Cruz Biotechnology, sc-7292), rabbit anti-HA (1:6000, Cell Signaling Technology, #C29F4), mouse anti-HA (1:5000, BioLegend, 901514), mouse anti-vinculin (1:10000, Sigma, V913), mouse anti-GAPDH (1:5000, Santa Cruz Biotechnology, sc-166574), mouse anti-Nup358 (1:3000, Santa Cruz Biotechnology, sc-74518), mouse anti-tubulin (1,3000, Sigma, T5168) and rabbit anti-Nup358 [1:3000, in-house (Joseph et al., 2004)]. Secondary antibodies used were HRP-rec-Protein A (1:4000, Invitrogen, 101123), anti-rabbit-IgG horseradish peroxidase-linked whole antibody from donkey (1:10,000, GE Healthcare, NA934V) and anti-mouse IgG-horseradish peroxidase-linked whole antibody from donkey (1:10,000, GE Healthcare, NA931V).

Cytoplasm, nucleoplasm and nuclear envelope fractionation

Cell fractionation was performed as previously published (Dreger et al., 2001; Sahoo et al., 2017), with some modifications. Briefly, HeLa cells were resuspended in cytoplasmic extraction buffer (20 mM HEPES, pH 8.0, 10 mM KCl, 2 mM MgCl2, 1 mM DTT) supplemented with 7.5 mM NaF, 0.75 mM Na3VO4, 1 mM PMSF, PIC and 20 mM NEM. Cells were lysed by gently pipetting up and down. Lysate was centrifuged at 15,294 g for 20 min and the supernatant was collected as cytoplasmic fraction. Pellet was washed twice in 1 ml each cytoplasmic extraction buffer and spun at 3824 g for 2 min. For nucleoplasm extraction, pellet was resuspended in 250 µl TP buffer (10 mM Tris-HCl and 10 mM NaH2PO4/Na2HPO4 pH 7.4), heparin sodium salt (250 µg/ml; Sigma-Aldrich, #H3393-500KU), benzonase nuclease (250 units/µl; Sigma-Aldrich, #E1014-25 KU) supplemented with 7.5 mM NaF, 0.75 mM Na3VO4, 10 µg/ml aprotinin, 10 µg/ml leupeptin, 1 mM PMSF and 20 mM NEM. The solution was kept for 1 h on a rotospin (Tarson Products Pvt. Ltd.) at 4°C and was tapped in between. Lysate was sedimented by centrifugation at 10,621 g for 30 min at 4°C. The supernatant collected was nucleoplasmic fraction. The remaining pellet was washed twice in TP buffer and spun at 3824 g for 2 min. The pellet was resuspended in nuclear lysis buffer (20 mM HEPES pH 7.4, 150 mM NaCl, 1.5 mM MgCl2, 2 mM EGTA, 2 mM DTT and 1% Triton X-100) supplemented with PIC, 7.5 mM NaF, 10 µg/ml aprotinin, 10 µg/ml leupeptin, 1 mM PMSF, 0.75 mM Na3VO4 and 20 mM NEM. The resuspended pellet was sonicated using Vibra-Cell sonicator (Sonics & Materials Inc.) for 5 s with a pulse of 1 s ON and 2 s OFF at 30% amplitude. Sonicated sample was centrifuged 15,294 g for 20 min at 4°C. The supernatant was referred to as nuclear envelope fraction. Proteins were separated by SDS-PAGE and analyzed by western blotting. Vinculin was used as cytoplasmic marker, lamin A/C as total nuclear marker, SUMO–RanGAP1 as a nuclear envelope marker and RNA polymerase II as a nucleoplasmic marker. The volume of the individual fractions to be loaded on SDS-PAGE were normalized to the minimum lysis volume and accordingly loaded. Total fraction was reconstituted by mixing all the three individual fractions in a volume proportional to individual fractions.

In vitro SUMOylation assay

The procedure used for in vitro SUMOylation assay was as previously published (Magre et al., 2019), with some modifications. Recombinant proteins required for in vitro SUMO conjugation were expressed and purified in bacterial system and dialyzed in SUMO conjugation buffer (20 mM HEPES, pH 7.5, 110 mM potassium acetate, 2 mM magnesium acetate, 1 mM DTT, 0.05% tween- 20, 1 mM EGTA). A 20 µl in vitro SUMOylation reaction was setup containing 1200 ng of GST–DDX19-WT or K26R mutant, 1500 ng of His–SUMO1GG, 150 ng of His–SAE1/SAE2 (E1), 100 ng of His-Ubc9 (E2) in the presence or absence of 5 mM ATP. The reaction mix was incubated at 30°C for 0 or 2 h and terminated by addition of SDS-loading dye and heating at 95°C for 6 min.

In order to check whether the human (h)Nup358-IR region acted as a SUMO E3 ligase for DDX19B in vitro (Fig. 3D), the above mentioned reaction was set up, except for using an E2 (Ubc9) concentration of 10 ng (instead of 100 ng) with or without 5 ng of IR.

Oligo(dT) staining

The protocol for oligo(dT) staining was as previously published (Sahoo et al., 2017). U2OS cells seeded (2×104 cells/well) on coverslips in 24-well plate were transfected with HA–MBP control, an siRNA-resistant version of HA–DDX19B-WT or a HA–DDX19B-K26R mutant construct for 24 h. Then cells were re-transfected with control or DDX19 siRNA (40 nM). The cells were fixed after 96 h of siRNA transfection with 4% paraformaldehyde (PFA) for 15 min at RT, washed three times with TBS and again re-fixed with chilled methanol for 5 min. This was followed by incubation of the coverslips with 2× saline-sodium citrate (SSC) buffer at RT for 10 min. Hybridization was carried out in a hybridization chamber wherein moistened strips of filter paper were kept on the edges of the stage to avoid drying. The hybridization mix was made with the following components; 10% formamide (Sigma, #F9037-100ML), 10% dextran sulphate (Sigma, # D8906-10G), 0.1 mg Herring sperm DNA (Sigma, #D3159-10G) and 5 ng/µl of FAM-labelled oligo(dT) (Integrated DNA technology) in 2× SSC solution. The coverslips were added with 100 µl of hybridization mix and incubated at 37°C for 3 h. After hybridization, the coverslips were washed thrice with 2X SSC followed by one wash with 0.2× SSC. Coverslips were once again washed with TBS before incubating it with primary antibody. Both primary and secondary antibodies are made in TBS containing 2% normal horse serum (NHS) (Vector laboratories, #S-2000). Primary antibody, anti-mouse HA (1:1000, BioLegend, #16B12), was added to the coverslips and incubated for 30 min. This was followed by three washes with TBS and further incubation of the coverslips with anti-mouse-IgG Alexa Fluor594 (1:10,000, Invitrogen, #A21203) for 30 min. Hoechst 33342 (0.1%) dye was also added along with the secondary antibody to stain the DNA. Cells were again washed three times with TBS and mounted on glass slides using vectashield mounting medium (Vector laboratories, #H-1000) and sealed with a transparent nail paint to avoid dehydration. Confocal images were captured with Olympus Fluoview FV3000 laser scanning confocal microscope using PlanApochromat 60×/1.42 NA oil immersion objective.

Oligo(dT) images were quantified using ImageJ software (Fiji). Integrated fluorescence density of oligo(dT) in nucleus was normalized with the integrated density within the total cell.

Statistical analysis

All the experiments, except oligo(dT) staining, were repeated at least three times for calculating the statistical significance. Values were represented as mean±s.e.m. A two-tailed unpaired Student's t-test was used to calculate the P-value for the western blotting data. For mRNA export assay using oligo(dT) staining, the experiments were repeated four times, and a Mann–Whitney test was performed for statistical analysis. Data represented as mean±s.d. P<0.05 was considered significant. Graphs were plotted using Graph Pad Prism 8.

We are grateful to Eliza Izzaurralde and Gerald Grosveld for sharing reagents. We are thankful to Joseph and Seshadri lab members for insightful discussions and suggestions. Prachi Deshmukh is acknowledged for initial observations on Nup358-DDX19 interaction. We thank the NCCS Bio-Imaging facility, and staff for their help with confocal microscopy. We acknowledge the help from Lizanne Oliveira for editing the manuscript.

Author contributions

Conceptualization: P.B., J.J.; Methodology: P.B., S.M., M.K., J.J.; Formal analysis: P.B., S.M., M.K., J.J.; Investigation: P.B., J.J.; Writing - original draft: P.B., J.J.; Writing - review & editing: P.B., S.M., M.K., J.J.; Supervision: J.J.; Project administration: J.J.; Funding acquisition: J.J.

Funding

The work was partly supported by intramural funding from the National Centre for Cell Science (NCCS) and a from the Department of Biotechnology, Government of India (grant BT/PR27451/BRB/10/1655/2018). Fellowships from Council of Scientific and Industrial Research (CSIR) to P.B. and M.K., and Department of Biotechnology (DBT), Government of India, to S.M. is gratefully acknowledged.

The peer review history is available online at https://journals.biologists.com/jcs/article-lookup/doi/10.1242/jcs.259449.

Aksenova
,
V.
,
Smith
,
A.
,
Lee
,
H.
,
Bhat
,
P.
,
Esnault
,
C.
,
Chen
,
S.
,
Iben
,
J.
,
Kaufhold
,
R.
,
Yau
,
K. C.
,
Echeverria
,
C.
et al. 
(
2020
).
Nucleoporin TPR is an integral component of the TREX-2 mRNA export pathway
.
Nat. Commun.
11
,
4577
.
Beauclair
,
G.
,
Bridier-Nahmias
,
A.
,
Zagury
,
J.-F.
,
Saïb
,
A.
and
Zamborlini
,
A.
(
2015
).
JASSA: A comprehensive tool for prediction of SUMOylation sites and SIMs
.
Bioinformatics
31
,
3483
-
3491
.
Ben-Yishay
,
R.
,
Ashkenazy
,
A. J.
and
Shav-Tal
,
Y.
(
2016
).
Dynamic encounters of genes and transcripts with the nuclear pore
.
Trends Genet.
32
,
419
-
431
.
Ben-Yishay
,
R.
,
Mor
,
A.
,
Shraga
,
A.
,
Ashkenazy-Titelman
,
A.
,
Kinor
,
N.
,
Schwed-Gross
,
A.
,
Jacob
,
A.
,
Kozer
,
N.
,
Kumar
,
P.
,
Garini
,
Y.
et al. 
(
2019
).
Imaging within single NPCs reveals NXF1's role in mRNA export on the cytoplasmic side of the pore
.
J. Cell Biol.
218
,
2962
-
2981
.
Carmody
,
S. R.
and
Wente
,
S. R.
(
2009
).
mRNA nuclear export at a glance
.
J. Cell Sci.
122
,
1933
-
1937
.
Chow
,
K.-H.
,
Elgort
,
S.
,
Dasso
,
M.
and
Ullman
,
K. S.
(
2012
).
Two distinct sites in Nup153 mediate interaction with the SUMO proteases SENP1 and SENP2
.
Nucleus
3
,
349
-
358
.
Ciccarelli
,
F. D.
,
von Mering
,
C.
,
Suyama
,
M.
,
Harrington
,
E. D.
,
Izaurralde
,
E.
and
Bork
,
P.
(
2005
).
Complex genomic rearrangements lead to novel primate gene function
.
Genome Res.
15
,
343
-
351
.
Collins
,
R.
,
Karlberg
,
T.
,
Lehtiö
,
L.
,
Schütz
,
P.
,
van den Berg
,
S.
,
Dahlgren
,
L.-G.
,
Hammarström
,
M.
,
Weigelt
,
J.
and
Schüler
,
H.
(
2009
).
The DEXD/H-box RNA helicase DDX19 is regulated by an α-helical switch
.
J. Biol. Chem.
284
,
10296
-
10300
.
Dossani
,
Z. Y.
,
Weirich
,
C. S.
,
Erzberger
,
J. P.
,
Berger
,
J. M.
and
Weis
,
K.
(
2009
).
Structure of the C-terminus of the mRNA export factor Dbp5 reveals the interaction surface for the ATPase activator Gle1
.
Proc. Natl. Acad. Sci. USA
106
,
16251
-
16256
.
Dreger
,
M.
,
Bengtsson
,
L.
,
Schöneberg
,
T.
,
Otto
,
H.
and
Hucho
,
F.
(
2001
).
Nuclear envelope proteomics: Novel integral membrane proteins of the inner nuclear membrane
.
Proc. Natl. Acad. Sci. USA
98
,
11943
-
11948
.
Dreyfuss
,
G.
,
Kim
,
V. N.
and
Kataoka
,
N.
(
2002
).
Messenger-RNA-binding proteins and the messages they carry
.
Nat. Rev. Mol. Cell Biol.
3
,
195
-
205
.
Flotho
,
A.
and
Melchior
,
F.
(
2013
).
Sumoylation: a regulatory protein modification in health and disease
.
Annu. Rev. Biochem.
82
,
357
-
385
.
Folkmann
,
A. W.
,
Noble
,
K. N.
,
Cole
,
C. N.
and
Wente
,
S. R.
(
2011
).
Dbp5, Gle1-IP6 and Nup159: A working model for mRNP export
.
Nucleus
2
,
540
-
548
.
Goeres
,
J.
,
Chan
,
P.-K.
,
Mukhopadhyay
,
D.
,
Zhang
,
H.
,
Raught
,
B.
and
Matunis
,
M. J.
(
2011
).
The SUMO-specific isopeptidase SENP2 associates dynamically with nuclear pore complexes through interactions with karyopherins and the Nup107-160 nucleoporin subcomplex
.
Mol. Biol. Cell
22
,
4868
-
4882
.
Grüter
,
P.
,
Tabernero
,
C.
,
Von Kobbe
,
C.
,
Schmitt
,
C.
,
Saavedra
,
C.
,
Bachi
,
A.
,
Wilm
,
M.
,
Felber
,
B. K.
and
Izaurralde
,
E.
(
1998
).
TAP, the human homolog of Mex67p, mediates CTE-dependent RNA export from the nucleus
.
Mol. Cell
1
,
649
-
659
.
Hang
,
J.
and
Dasso
,
M.
(
2002
).
Association of the human SUMO-1 protease SENP2 with the nuclear pore
.
J. Biol. Chem.
277
,
19961
-
19966
.
Hendriks
,
I. A.
and
Vertegaal
,
A. C. O.
(
2016
).
A comprehensive compilation of SUMO proteomics
.
Nat. Rev. Mol. Cell Biol.
17
,
581
-
595
.
Hodge
,
C. A.
,
Tran
,
E. J.
,
Noble
,
K. N.
,
Alcazar-Roman
,
A. R.
,
Ben-Yishay
,
R.
,
Scarcelli
,
J. J.
,
Folkmann
,
A. W.
,
Shav-Tal
,
Y.
,
Wente
,
S. R.
and
Cole
,
C. N.
(
2011
).
The Dbp5 cycle at the nuclear pore complex during mRNA export I: Dbp5 mutants with defects in RNA binding and ATP hydrolysis define key steps for Nup159 and Gle1
.
Genes Dev.
25
,
1052
-
1064
.
Joseph
,
J.
and
Dasso
,
M.
(
2008
).
The nucleoporin Nup358 associates with and regulates interphase microtubules
.
FEBS Lett.
582
,
190
-
196
.
Joseph
,
J.
,
Liu
,
S.-T.
,
Jablonski
,
S. A.
,
Yen
,
T. J.
and
Dasso
,
M.
(
2004
).
The RanGAP1-RanBP2 complex is essential for microtubule-kinetochore interactions in vivo
.
Curr. Biol.
14
,
611
-
617
.
Katahira
,
J.
(
2015
).
Nuclear export of messenger RNA
.
Genes (Basel)
6
,
163
-
184
.
Katahira
,
J.
,
Sträßer
,
K.
,
Podtelejnikov
,
A.
,
Mann
,
M.
,
Jung
,
J. U.
and
Hurt
,
E.
(
1999
).
The Mex67p-mediated nuclear mRNA export pathway is conserved from yeast to human
.
EMBO J.
18
,
2593
-
2609
.
Ledoux
,
S.
and
Guthrie
,
C.
(
2011
).
Regulation of the Dbp5 ATPase cycle in mRNP remodeling at the nuclear pore: a lively new paradigm for DEAD-box proteins
.
Genes Dev.
25
,
1109
-
1114
.
Li
,
Y.
,
Aksenova
,
V.
,
Tingey
,
M.
,
Yu
,
J.
,
Ma
,
P.
,
Arnaoutov
,
A.
,
Chen
,
S.
,
Dasso
,
M.
and
Yang
,
W.
(
2021
).
Distinct roles of nuclear basket proteins in directing the passage of mRNA through the nuclear pore
.
Proc. Natl. Acad. Sci. USA
118
,
e2015621118
.
Lin
,
D. H.
,
Correia
,
A. R.
,
Cai
,
S. W.
,
Huber
,
F. M.
,
Jette
,
C. A.
and
Hoelz
,
A.
(
2018
).
Structural and functional analysis of mRNA export regulation by the nuclear pore complex
.
Nat. Commun.
9
,
2319
.
Lund
,
M. K.
and
Guthrie
,
C.
(
2005
).
The DEAD-box protein Dbp5p is required to dissociate Mex67p from exported mRNPs at the nuclear rim
.
Mol. Cell
20
,
645
-
651
.
Magre
,
I.
,
Fandade
,
V.
,
Damle
,
I.
,
Banerjee
,
P.
,
Yadav
,
S. K.
,
Sonawane
,
M.
and
Joseph
,
J.
(
2019
).
Nup358 regulates microridge length by controlling SUMOylation-dependent activity of aPKC in zebrafish epidermis
.
J. Cell Sci.
132
,
jcs224501
.
Maniatis
,
T.
and
Reed
,
R.
(
2002
).
An extensive network of coupling among gene expression machines
.
Nature
416
,
499
-
506
.
Mojsa
,
B.
,
Tatham
,
M. H.
,
Davidson
,
L.
,
Liczmanska
,
M.
,
Branigan
,
E.
and
Hay
,
R. T.
(
2021
).
Identification of SUMO targets associated with the pluripotent state in human stem cells
.
Mol. Cell. Proteomics
20
,
100164
.
Montpetit
,
B.
,
Thomsen
,
N. D.
,
Helmke
,
K. J.
,
Seeliger
,
M. A.
,
Berger
,
J. M.
and
Weis
,
K.
(
2011
).
A conserved mechanism of DEAD-box ATPase activation by nucleoporins and InsP6 in mRNA export
.
Nature
472
,
238
-
242
.
Napetschnig
,
J.
,
Kassube
,
S. A.
,
Debler
,
E. W.
,
Wong
,
R. W.
,
Blobel
,
G.
and
Hoelz
,
A.
(
2009
).
Structural and functional analysis of the interaction between the nucleoporin Nup214 and the DEAD-box helicase Ddx19
.
Proc. Natl. Acad. Sci. USA
106
,
3089
-
3094
.
Noble
,
K. N.
,
Tran
,
E. J.
,
Alcázar-Román
,
A. R.
,
Hodge
,
C. A.
,
Cole
,
C. N.
and
Wente
,
S. R.
(
2011
).
The Dbp5 cycle at the nuclear pore complex during mRNA export II: Nucleotide cycling and mRNP remodeling by Dbp5 are controlled by Nup159 and Gle1
.
Genes Dev.
25
,
1065
-
1077
.
Palancade
,
B.
and
Doye
,
V.
(
2008
).
Sumoylating and desumoylating enzymes at nuclear pores: underpinning their unexpected duties?
Trends Cell Biol.
18
,
174
-
183
.
Pichler
,
A.
,
Gast
,
A.
,
Seeler
,
J. S.
,
Dejean
,
A.
and
Melchior
,
F.
(
2002
).
The Nucleoporin RanBP2 has SUMO1 E3 ligase activity
.
Cell
108
,
109
-
120
.
Pichler
,
A.
,
Knipscheer
,
P.
,
Saitoh
,
H.
,
Sixma
,
T. K.
and
Melchior
,
F.
(
2004
).
The RanBP2 SUMO E3 ligase is neither HECT- nor RING-type
.
Nat. Struct. Mol. Biol.
11
,
984
-
991
.
Rajakylä
,
E. K.
,
Viita
,
T.
,
Kyheröinen
,
S.
,
Huet
,
G.
,
Treisman
,
R.
and
Vartiainen
,
M. K.
(
2015
).
RNA export factor Ddx19 is required for nuclear import of the SRF coactivator MKL1
.
Nat. Commun.
6
,
5978
.
Rajan
,
A. A. N.
and
Montpetit
,
B.
(
2021
).
Emerging molecular functions and novel roles for the DEAD-box protein Dbp5/DDX19 in gene expression
.
Cell. Mol. Life Sci.
78
,
2019
-
2030
.
Sahoo
,
P. K.
,
Murawala
,
P.
,
Sawale
,
P. T.
,
Sahoo
,
M. R.
,
Tripathi
,
M. M.
,
Gaikwad
,
S. R.
,
Seshadri
,
V.
and
Joseph
,
J.
(
2012
).
Wnt signalling antagonizes stress granule assembly through a Dishevelled-dependent mechanism
.
Biol. Open
1
,
109
-
119
.
Sahoo
,
M. R.
,
Gaikwad
,
S.
,
Khuperkar
,
D.
,
Ashok
,
M.
,
Helen
,
M.
,
Yadav
,
S. K.
,
Singh
,
A.
,
Magre
,
I.
,
Deshmukh
,
P.
,
Dhanvijay
,
S.
et al. 
(
2017
).
Nup358 binds to AGO proteins through its SUMO -interacting motifs and promotes the association of target mRNA with miRISC
.
EMBO Rep.
18
,
241
-
263
.
Schmitt
,
C.
,
Von Kobbe
,
C.
,
Bachi
,
A.
,
Panté
,
N.
,
Rodrigues
,
J. P.
,
Boscheron
,
C.
,
Rigaut
,
G.
,
Wilm
,
M.
,
Séraphin
,
B.
,
Carmo-Fonseca
,
M.
et al. 
(
1999
).
Dbp5, a DEAD-box protein required for mRNA export, is recruited to the cytoplasmic fibrils of nuclear pore complex via a conserved interaction with CAN/Nup159p
.
EMBO J.
18
,
4332
-
4347
.
Stewart
,
M.
(
2010
).
Nuclear export of mRNA
.
Trends Biochem. Sci.
35
,
609
-
617
.
Stewart
,
M.
(
2019
).
Polyadenylation and nuclear export of mRNAs
.
J. Biol. Chem.
294
,
2977
-
2987
.
Sträßer
,
K.
,
Baßler
,
J.
and
Hurt
,
E.
(
2000
).
Binding of the Mex67p/Mtr2p heterodimer to FXFG, GLFG, and FG repeat nucleoporins is essential for nuclear mRNA export
.
J. Cell Biol.
150
,
695
-
706
.
Stutz
,
F.
and
Izaurralde
,
E.
(
2003
).
The interplay of nuclear mRNP assembly, mRNA surveillance and export
.
Trends Cell Biol.
13
,
319
-
327
.
Weirich
,
C. S.
,
Erzberger
,
J. P.
,
Flick
,
J. S.
,
Berger
,
J. M.
,
Thorner
,
J.
and
Weis
,
K.
(
2006
).
Activation of the DExD/H-box protein Dbp5 by the nuclear-pore protein Gle1 and its coactivator InsP6 is required for mRNA export
.
Nat. Cell Biol.
8
,
668
-
676
.
Wende
,
W.
,
Friedhoff
,
P.
and
Sträßer
,
K.
(
2019
).
Mechanism and regulation of co-transcriptional mRNP assembly and nuclear mRNA export
.
Adv. Exp. Med. Biol.
1203
,
1
-
31
.
Wolf
,
E. J.
,
Miles
,
A.
,
Lee
,
E. S.
,
Nabeel-Shah
,
S.
,
Greenblatt
,
J. F.
,
Palazzo
,
A. F.
,
Tropepe
,
V.
and
Emili
,
A.
(
2020
).
MKRN2 physically interacts with GLE1 to regulate mRNA export and zebrafish retinal development
.
Cell Rep.
31
,
107693
.
Xie
,
Y.
and
Ren
,
Y.
(
2019
).
Mechanisms of nuclear mRNA export: a structural perspective
.
Traffic
20
,
829
-
840
.
Yadav
,
S. K.
,
Magre
,
I.
,
Singh
,
A.
,
Khuperkar
,
D.
and
Joseph
,
J.
(
2016
).
Regulation of aPKC activity by Nup358 dependent SUMO modification
.
Sci. Rep.
6
,
34100
.
Zhang
,
H.
,
Saitoh
,
H.
and
Matunis
,
M. J.
(
2002
).
Enzymes of the SUMO modification pathway localize to filaments of the nuclear pore complex
.
Mol. Cell. Biol.
22
,
6498
-
6508
.
Zhang
,
H.
,
Mahadevan
,
K.
and
Palazzo
,
A. F.
(
2014
).
Sumoylation is required for the Cytoplasmic Accumulation of a subset of mRNAs
.
Genes (Basel)
5
,
982
-
1000
.

Competing interests

The authors declare no competing or financial interests.

Supplementary information