Regulation of nuclear envelope dynamics is an important example of the universal phenomena of membrane fusion. The signalling molecules involved in nuclear membrane fusion might also be conserved during the formation of both pronuclear and zygote nuclear envelopes in the fertilised egg. Here, we determine that class-I phosphoinositide 3-kinases (PI3Ks) are needed for in vitro nuclear envelope formation. We show that, in vivo, PtdIns(3,4,5)P3 is transiently located in vesicles around the male pronucleus at the time of nuclear envelope formation, and around male and female pronuclei before membrane fusion. We illustrate that class-I PI3K activity is also necessary for fusion of the female and male pronuclear membranes. We demonstrate, using coincidence amplified Förster resonance energy transfer (FRET) monitored using fluorescence lifetime imaging microscopy (FLIM), a protein–lipid interaction of Rab7 GTPase and PtdIns(3,4,5)P3 that occurs during pronuclear membrane fusion to create the zygote nuclear envelope. We present a working model, which includes several molecular steps in the pathways controlling fusion of nuclear envelope membranes.
The mechanisms by which nuclear envelopes are formed are not completely understood. It is generally agreed that, whether during mitosis, or during formation of the male pronuclear envelope around the sperm nucleus after fertilisation, the bulk of the nuclear envelope is derived from endoplasmic reticulum (ER) membranes that undergo fusion to enclose the chromatin (Burke and Ellenberg, 2002; Poccia and Larijani, 2009; Prunuske and Ullman, 2006). Fusion of complete male and female pronuclear envelopes to create the zygote nuclear envelope also occurs in many organisms (Byrne et al., 2014; Collas and Poccia, 1996b). Similarly, but less commonly, complete nuclear envelopes can form around individual chromosomes creating karyomeres, whose envelopes can fuse together to create complete nuclear envelopes around the full chromosome set (Domart et al., 2012).
On the basis of experiments on the in vitro and in vivo formation of the male pronuclear envelope, as well as in vivo data on zygote nuclear envelope formation and karyomere fusion, we have previously presented a common mechanistic model in which fusion is controlled upstream by a Src-family kinase (SFK1) that activates a phosphatidylinositide-specific phospholipase C (PLCγ) that catalyses the hydrolysis of PtdIns(4,5)P2, resulting in the production of a fusogenic lipid, diacylglycerol (DAG) (Byrne et al., 2012, 2007, 2009a, 2014; Poccia and Larijani, 2009). The in vitro assays have been well documented over many years in several of our publications (Collas and Poccia, 1996a, 1998; Dumas et al., 2010), and the necessary controls, such as time-resolved Förster resonance energy transfer (FRET) analyses of nuclear envelope membrane fusion and electron micrographs, showing complete formation of the male pronucleus upon the GTP hydrolysis have been demonstrated.
These events are associated in the sea urchin with a subset of membrane vesicles (MV1) that had been isolated from egg cytoplasm that is highly enriched in SFK1, PLCγ and phosphorylated forms of phosphatidylinositol (Byrne et al., 2007, 2009b; Collas and Poccia, 1996a). Fusion with MV2 vesicles, which contain an ER marker, depends on and is initiated by the MV1 vesicles in male pronuclei (Collas and Poccia, 1996a; Dumas et al., 2010). Fusion requires GTP hydrolysis, but the GTPases responsible have not been identified. Male pronuclear envelope formation has also been shown to require phosphoinositide 3-kinase (PI3K) activity on the basis of inhibition by the relatively non-specific inhibitors wortmannin and LY29400 (Larijani et al., 2001). Connections with these potential upstream events remain to be elucidated.
Fusion in many membranous compartments of the cell has been linked to Rab GTPases (Stenmark, 2009). Members of the Rab family, often in conjunction with partner phosphoinositides, are fairly specific for different compartments, thus serving as identity markers. Rab proteins act as molecular switches as they swap between their GTP- and GDP-bound forms. Rabs generally participate in membrane trafficking and other events, and assist in the recruitment of various effector molecules. In fusion reactions, they can recruit tethering factors that specifically bind together the membranes to be fused. RAB escorts present Rabs to geranylgeranyl transferases for prenylation that permits their membrane recruitment from cytosol.
SFKs are recruited from the cytosol to the cytosolic face of membranes through palmitoylation (Sato et al., 2009). Changes in palmitoyl state can result in alternative localisations. Src is found in a large pool in late endosomes and/or lysosomes. Src that lacks palmitoyl groups but that has been myristoylated can exchange between plasma membrane and late endosomes and/or lysosomes. It is unknown from which compartment SFK1-enriched MV1 vesicles arise.
Rab7 (of which there are two isoforms in mammals, Rab7a and Rab7b) is a monomeric GTPase that has been linked to SFK activity. The endocytic pathway is important for sorting proteins to be degraded in lysosomes or recycled to other membranous compartments in the cell. Late endosomes and/or lysosomes are associated with Rab7, which in conjunction with the ESCRT pathway, are needed for correct placement of Src within these organelles (Tu et al., 2011). ESCRT-III complexes have been implicated in the sealing of nuclear envelopes at a post-mitotic stage (Olmos et al., 2015; Vietri et al., 2015). Rab7 activity can be modified through interaction with PTEN, a phosphatase that prevents Rab7 from being recruited to late endosomes by GDP dissociation inhibitor protein (Shinde and Maddika, 2016). Coincidence detection of Rab7 and PtdIns(3)P by the ER protein protrudin promotes the formation of contact sites with late endosomes (Raiborg et al., 2015).
Both PtdIns(3)P and PtdIns(3,4,5)P3 are much more abundant in MV1 than in MV2. Interestingly, Jethwa et al. (2015) have recently shown that PtdIns(3,4,5)P3 is enriched in both the nuclear envelope and early endosomal membranes.
Because nuclear envelope formation in sea urchin depends on PI3Ks of unknown class and on GTP hydrolysis by unidentified GTPases, we tested whether class-I kinases are required for nuclear envelope formation and examined where PtdIns(3,4,5)P3 is located and whether Rab7 interacts with PtdIns(3,4,5)P3.
Here, using the specific class-I PI3K inhibitors PIK-75 and BYL719, we report that class-I PI3Ks are needed for in vitro nuclear envelope formation. Using a GRP-1 plextrin homology (PH) domain specific for PtdIns(3,4,5)P3 detection, we show that this phospholipid is transiently located in vesicles around the male pronucleus at the time of nuclear envelope formation, and around male and female pronuclei in vivo before membrane fusion. Pronuclear envelope fusion is also blocked by specific class-1 PI3K inhibitors (PIK-75 and BYL719). Finally, using time-resolved coincidence amplified-FRET monitored by performing fluorescence lifetime imaging microscopy (FLIM), we show, for the first time, a protein–lipid interaction of Rab7 and PtdIns(3,4,5)P3. We discuss models to account for the roles of Rab7 and PtdIns(3,4,5)P3 in nuclear envelope formation and fusion.
To determine the upstream effectors of the fusion machinery [PLC, SFK1 PtdIns(4,5)P2 and DAG] that are implicated in nuclear envelope formation (Byrne et al., 2007, 2009a, 2014; Poccia and Larijani, 2009), we set out to test the involvement of the class-I PI3Ks and their product PtdIns(3,4,5)P3. These experiments were performed initially in cell-free assays and afterwards in situ using intact fertilised sea urchin (Paracentrotus lividus) eggs.
Class-I PI3Ks are required for nuclear envelope formation in vitro
Using the cell-free assay, we tested the class-I-PI3K-specific inhibitors PIK-75 and BYL719, which block the kinase activity through competitive inhibition at the substrate-binding site and at the ATP-binding sites, respectively. The reaction was terminated after 1 h of incubation with the ATP-generating system; here membrane vesicles were bound to chromatin, but they did not fuse. Upon addition of GTP, the nuclear envelope formed a continuous membrane around the nuclei. The addition of the two PI3K-specific inhibitors resulted in the malformation of the nuclear membrane (Fig. S1A). As mentioned above, membrane fusion in these in vitro assays has been previously demonstrated by time-resolved FRET (Dumas et al., 2010). Therefore, the fluorescent assays were used here to quantify nuclear envelope formation. The assays were performed with and without inhibitor by scoring 100 nuclei for each condition. Upon addition of GTP, more than 70% of the nuclei had complete nuclear membranes. In all cases, 0.35 μl of DMSO was added after the first incubation as a vehicle control. Two different concentrations of LY 294002, a general PI3K inhibitor, were used as positive controls, and the PIK-75 inhibitor that is specific for class-I PI3Ks was added with the same DMSO volume as in the controls. After a 2-h incubation with GTP, the number of nuclei with a continuous nuclear envelope were scored. A significant decrease in nuclear envelope formation was observed with PIK-75 and BYL19 (Fig. 1).
These results indicate that PtdIns(3,4,5)P3 formation was a requirement for nuclear envelope formation. To determine localisation of PtdIns(3,4,5)P3, we used the PH domain of general receptor for phosphoinositides 1-associated scaffold protein (Grp1), which recognises PtdIns(3,4,5)P3 selectively (Ferguson et al., 2000; Klarlund et al., 2000). The Grp1 PH domain was co-purified with a GST tag, and its in vitro affinity was tested, as described previously (Jethwa et al., 2015). PtdIns(3,4,5)P3 was added in vitro to the cell-free assay. In these experiments, the reaction was terminated at different stages, firstly after the ATP addition, secondly 30 min after the GTP addition and, thirdly, two hours after GTP addition when the assembly was completed. Fig. S1B shows the punctate localisation of PtdIns(3,4,5)P3 both during binding of the membrane vesicles and their fusion. The addition of PIK-75 resulted in a reduction of the fluorescence intensity of the punctate localisation of PtdIns(3,4,5)P3. The inhibition did not reduce the number of punctate regions but it did reduce the fluorescence intensity.
These results demonstrate the involvement of both class-I PI3Ks and PtdIns(3,4,5)P3 in the in vitro formation of the nuclear envelope.
PtdIns(3,4,5)P3 is localised at the male and female pronuclei in situ
Because PtdIns(3,4,5)P3 was necessary for the in vitro formation of the nuclear membrane, we used the same probe, GST–Grp1-PH, to further explore its role in the various membrane fusion events that take place in the fertilised sea urchin egg. We specifically focussed on the female and the male pronuclear fusion. Fig. 2 shows immunofluorescence experiments that were performed using this probe. PtdIns(3,4,5)P3 membrane vesicles were localised at the egg cortex during the first stages of fertilisation, whereas at the later stages (7 min), a recruitment of the PtdIns(3,4,5)P3 vesicles at the male pronucleus and, at 10–20 min, at the female pronucleus was observed. At 30 min, following the fusion of the male and female pronuclei, the PtdIns(3,4,5)P3 vesicles once again showed a cortical localisation.
Inhibition of class-I PI3Ks abrogates continued fusion of male and female pronuclei
To investigate the role of class-I PI3Ks in pronuclear fusion, we used, as in the in vitro experiments, PIK-75. PIK-75 was added to P. lividus eggs 4 min after fertilisation. Fertilisation was terminated by fixing the eggs at different time points before the first cell division. Vehicle controls were performed by adding the same volume of DMSO alone. The effect of the inhibitor in situ was characterised by measuring the distance between the male pronucleus and the female pronucleus when treated with 1 μM PIK-75 or without inhibitor at different times post fertilisation (Fig. 3). Fusion was scored as the merging of the male and female chromatin DNA stains. At 15 min, when there is an accumulation of PtdIns(3,4,5)P3 vesicles, the male and female pronuclei had not yet fused. It was only two minutes later, at 17 min post fertilisation, that the male pronucleus and the female pronucleus were in close contact and mostly fused. However, in eggs that had been treated with 1 μM PIK-75, a substantially lower percentage had fused although most were in proximity. The quantification demonstrates that the inhibition of class-I PI3Ks prevented the continued fusion of the male and female pronuclei but not the accumulation of PtdIns(3,4,5)P3 vesicles, hence demonstrating the requirement of class-I PI3Ks for this fusion event.
PtdIns(3,4,5)P3 and Rab7 vesicles colocalise in the pronuclear region
We postulated that, as in the maturation of late endosomes in mammalian cells, the small GTPase Rab7 would be in the same compartment as SFKs (Stenmark, 2009) and that PtdIns(3,4,5)P3 would be involved in the recruitment of Rab7, where it might lead to the recruitment of SFK1 to activate PLCγ (Byrne et al., 2012). To test this hypothesis, we performed a series of immunofluorescence experiments to determine the localisation of PI3K and its substrate and product PtdIns(4,5)P2 and PtdIns(3,4,5)P3, respectively (Fig. 4A–B). The immunofluorescence experiments also showed the localisation of Rab7 and PtdIns(3,4,5)P3 (Fig. 4C), and of SFK and Rab7 (Fig. 4D). All colocalisations were quantified in whole eggs as well as in the regions of the pronuclei at the 7- and 15-min time points. The box and whisker plots show a significant colocalisation of PI3K and PtdIns(4,5)P2 at 7 min in the pronuclear region. Rab7 and SFK1 colocalisation were significant in the whole egg at 7 min, whereas significant differences were not observed between 7 and 15 min. The most significant colocalisation was between Rab7 and PtdIns(3,4,5)P3 at 15 min in the pronuclear region where PtdIns(3,4,5)P3 vesicles were recruited (Fig. 2) for the fusion of the male and female pronuclei.
In addition to indicating that Rab7 and SFK1 might share the same vesicle in the MV1 population, these results illustrate for the first time the localisation and quantified colocalisation of PtdIns(3,4,5)P3 and Rab7 during zygote nuclear membrane fusion.
PtdIns(3,4,5)P3 and Rab7 interact in vesicles around the forming zygote nuclear envelope
The quantified colocalisations between PtdIns(3,4,5)P3 and Rab7 vesicles indicated that these effectors were within the same vicinity. To determine whether PtdIns(3,4,5)P3 and Rab7 actually interacted during zygote nuclear membrane fusion, we exploited coincidence amplified time-resolved FRET (a-FRET) (Byrne et al., 2014). This method has been described in detail previously (Byrne et al., 2014; Veeriah et al., 2014). Briefly, Rab7 was labelled using a primary monoclonal antibody conjugated to an anti-chicken F(ab′)2–Atto488 (the donor). The PtdIns(3,4,5)P3 labelling was performed with GST–Grp1-PH, an antibody against GST that had been conjugated to a secondary anti-rabbit F(ab′)2 that had been labelled with horseradish peroxidase (HRP) and Alexa-Fluor-594 (the acceptor). The main difference from previously published a-FRET methods was that here, for the first time, we exploited a-FRET to determine protein–lipid interactions instead of protein–protein interactions.
Fig. 5 shows box and whisker plots of the lifetime distributions of the donor alone, of the donor and acceptor, and of the donor in the presence of the ‘false acceptor’ control. All the lifetime measurements were acquired using an oil immersion 60.3× objective, a wide-field microscope and multiple frequency domain FLIM. In unfertilised eggs, the lifetime distributions did not significantly change (Fig. 5A). However, at 15 min post fertilisation, donor lifetime decreased significantly in the pronuclear region (Fig. 5B). The lifetime was recovered to the donor lifetime with the ‘false acceptor’ control (Rab7+HRP). The calculated FRET efficiencies between unfertilised and 15 min fertilisation were 0% and 10%, respectively.
These data indicate that Rab7 and PtdIns(3,4,5)P3 were not just co-localised but were actually interacting during male and female pronuclear fusion. This interaction could explain the reason for the accumulation of PtdIns(3,4,5)P3 vesicles at the 10-min time point (Fig. 2).
Recently, Byrne et al. (2012) have shown, using time-resolved FRET, the spatially restricted and direct molecular interaction of PLCγ and SFK1 during the formation of the male pronuclear envelope. They have shown that during GTP initiation of nuclear membrane precursor fusion, PLCγ is phosphorylated on residue Tyr783, which is dependent on SFK1 activity. Transient phosphorylation of Tyr783 was detected on the surface of the male pronuclei in vivo during formation of the nuclear envelope. Kinetic analysis led to a model in which vesicles enriched in SFK1, PLCγ and PtdIns(4,5)P2 are bound to the nuclei under conditions in which SFK1 is inactive. GTP hydrolysis then removes the inhibition, allowing SFK1 to phosphorylate PLCγ, which hydrolyses the PtdIns(4,5)P2 to locally produce fusogenic DAG, leading to membrane fusion and nuclear envelope formation. After initiation of fusion, SFK1 and PLCγ remain on the nuclear envelope but no longer directly interact.
Byrne et al. (2014) have shown more recently that a similar mechanism accounts for the fusion of the new male pronuclear envelope with the already formed female pronuclear envelope during the first embryonic cell cycle to create the zygote nuclear envelope. Once again, PLCγ and SFK1 were shown, using time-resolved FRET, to directly interact at the site of nuclear membrane fusion where PtdIns(4,5)P2 levels are lower.
In karyomere fusion, alteration of DAG levels through microinjection of DAG-modifying enzymes that either deplete its precursor or convert it to phosphatidic acid slow or block karyomere nuclear envelope fusion. In addition, these treatments block or delay mitotic nuclear membrane reformation in both sea urchin and mammalian cells (Domart et al., 2012).
Taken together, these data suggest a conserved mechanism for nuclear envelope formation and membrane fusion. Despite the abundant evidence suggesting a new role for DAG and its upstream control by SFK1 and PLCγ in nuclear envelope fusion events, the nature of the small GTPase and PI3K requirements remain unclear, as does the mechanism for fusion propagation within the MV2 population, which creates the complete nuclear envelope. Recently, Jethwa et al. (2015) have shown that PtdIns(3,4,5)P3 is enriched in both the nuclear envelope and early endosomal membranes. By using the selective inhibitor PIK-75, which binds to the p100α subunit (encoded by PIK3CA) of class-I PI3Ks, they showed that depletion of PtdIns(3,4,5)P3 lowered activation of targeted PKB/Akt in the early endosomal and nuclear envelope compartments, suggesting that PtdIns(3,4,5)P3 resides in endomembranes such as the early endosomes and the nuclear envelope.
On the basis of experiments shown here, we suggest that PtdIns(3,4,5)P3 also has a role in recruiting the fusion machinery necessary for the formation of the zygote nuclear membrane that results from fusion of male and female pronuclei. It should be noted that, in this model, formation of PtdIns(3,4,5)P3 is not transient as in the case of the mammalian plasma membrane. In sea urchin, vesicles stably enriched in PtdIns(3,4,5)P3 and PtdIns(4,5)P2 are recruited to the sites of fusion. The quantitative colocalisation experiments showed an enhancement of vesicles containing Rab7 and SFK1, suggesting a role for both of these effectors during the zygote nuclear envelope formation. The localisation of PtdIns(3,4,5)P3-containing vesicles around the female pronucleus and the prevention of zygote nuclear envelopes by class-1 PI3K inhibitors indicate an important role for PtdIns(3,4,5)P3 during the male and female pronuclei fusion. Furthermore, the spatiotemporal FRET results show a direct interaction of PtdIns(3,4,5)P3 and Rab7 during zygote nuclear membrane formation. The in situ involvement of PtdIns(3,4,5)P3 and Rab7 vesicles in a membrane fusion event is a new finding.
Colocalisation of SFK1 and Rab7 strongly suggests that they co-exist in MV1 vesicles along with PtdIns(3,4,5)P3, PtdIns(4,5)P2 and PLCγ. The function of the minor membrane fraction MV1 in initiation of fusion of male pronuclear membrane precursors (Dumas et al., 2010) suggests an analogous role in male and female pronuclear envelope fusion as well.
MV1 vesicles are located mostly in the egg cortex before fertilisation where they might aid cortical vesicle and/or plasma membrane fusion; they might also initiate fusion of male pronuclear envelope precursors of the MV2 class within a few minutes of fertilisation (Byrne et al., 2012; Dumas et al., 2010). Later MVIs are found to be associated with the male and female pronucleus envelopes as they fuse (Byrne et al., 2014; this paper). Thus, MV1 could represent a relatively stable assembly that comprises a regulated mobile nuclear envelope fusion platform.
A working model for control of pronuclear envelope fusion is shown in Fig. 6. MV1 components, comprising Rab7, SFK1 and PLCγ, are preassembled in a stable fusion platform in MV1 vesicles that are highly enriched in PtdIns(3,4,5)P3 and PtdIns(4,5)P2. The platform is inactive because SFK1 is inhibited by a putative C-terminal Src-kinase (Csk) protein, a non-receptor tyrosine kinase that phosphorylates an inhibitory tyrosine site in the N-terminal region of Src-family kinases.
MV1 vesicles contain mixed populations of PtdIns(3,4,5)P3, PtdIns(4,5)P2, SFK1, PLCγ and Rab7. Vesicles containing only some of these components stably associate with other vesicles that carry the missing phosphoinositides or proteins; these vesicles associate in close enough proximity to allow for the FRET interactions that have been reported (Byrne et al., 2014; this paper). The maintenance of fusogenic properties requires class-I PI3K activity, probably to keep the concentration of PtdIns(3,4,5)P3 elevated in the face of potential 3-phosphatase activity.
GTP can locally initiate the fusion pathway. GTPase activity removes the block to SFK activity. It is not known how many steps occur between GTPase and SFK activation, or whether multiple GTPases are active. SFK1 phosphorylation of PLCγ then initiates local elevation of the fusogenic lipid DAG, which initiates fusion of the two pronuclear envelopes to form a zygote nuclear envelope.
The possibility of pre-assembling an inactive membrane fusion complex in vesicles transported to the site at which fusion will take place would offer a new addition to a repertoire of fusion mechanisms that are available in the cell. Involvement of Rab7 and the maintenance of PtdIns(3,4,5)P3 represent additional upstream components to the machinery that is required for nuclear membrane fusion events.
MATERIALS AND METHODS
P. lividus sea urchins were purchased from Macrae and company, or collected from the coast of Armintza (Basque Country). 3′3-Dihexyloxacarbocyanine iodide (DiOC6), Hoechst 33342, ProLong antifade reagent, and chemically competent Escherichia coli MAX efficiency DH5α, peroxidase suppressor and tyramide signal amplification (TSA) kit with Alexa-Fluor-594 tyramide were from Invitrogen Life Technologies. Anti-PI3K p85 antibody (# 86714; 1:100), anti-Rab7 antibody (# 50533; 1:100), donkey anti-mouse IgG conjugated to Alexa-Fluor-488 (# 150105; 1:200), goat anti-chicken IgY conjugated to Alexa-Fluor-647 (# 150171; 1:500), goat polyclonal antibody to GST conjugated with DyLight650 (# 117497; 1:500) and anti-GST antibody (# 9085; 1:500) were from Abcam. AffiniPure F (ab′) 2 IgG anti-mouse (# 715-006-150; 1:500) and peroxidase AffiniPure F(ab′)2 fragment donkey anti-rabbit IgG (# 711-036-152; 1:500) were purchased from Jackson ImmunoResearch. Inhibitors used were: LY 294002 from Cayman Chemicals and class-I-specific PIK-75 and BYL719 from Selleck Chemicals. Membrane lipid strips and PI(4,5)P2 Grip (PLC-δ1-PH) were purchased from Echelon. 3-Amino-1,2,4-triazole (ATA), Triton X-100 (TX-100), paraformaldehyde (PFA), adenosine 5′-triphosphate disodium salt hydrate (ATP), guanosine 5′-triphosphate sodium salt hydrate (GTP), phosphocreatine disodium salt hydrate (CP), phosphocreatine kinase (CPK), poly-L-lysine solution, saponin and fatty-acid-free bovine serum albumin (FAF-BSA) were from Sigma-Aldrich. Poly-L-lysine-coated coverslips were from BD Biosciences. PD-10 dye removal columns and PD MiniTrap G-25 were purchased from GE Healthcare Life Sciences. inositol 1,3,4,5-tetrakisphosphate (IP4) was purchased from Cell Signals. All other materials (salts and organic solvents) were of analytical grade.
Recombinant protein purification and specificity testing
E. coli cells were transformed with pGEX plasmid containing GST–Grp1-PH. A resulting colony was used to inoculate a Luria broth (LB) culture, which was grown at 37°C, with shaking at 200 rpm. Protein expression was induced at OD600 0.6 with 1 mM IPTG at 18°C overnight. Cells were harvested by centrifugation at 11,000 g for 20 min at 4°C, resuspended in distilled water and treated with 10 mg of lysozyme. The suspensions were frozen overnight at −20°C and, the following day, resuspended in 10× lysis buffer [500 mM Tris-HCl, pH 7.4, 10% (v/v) Triton X-100, 20 mM EDTA, 10 mM DTT, 10 mM AEBSF] before probe sonication for 5 min. After centrifugation at 12,000 g for 20 min at 4°C to remove cell debris, the supernatant was loaded onto 0.5 ml of slurry of glutathione Sepharose beads that had been pre-equilibrated in 50 mM Tris-HCl, pH 7.4, 140 mM NaCl, 2.7 mM KCl. The sample was incubated for 15 min at 4°C. Two washes with equilibration buffer (pre-equilibration buffer supplemented with 2 mM DTT) and two further washes with equilibration buffer supplemented with 500 mM NaCl were performed. Recombinant GST-tagged protein was eluted by adding 0.5 ml of elution buffer (50 mM Tris-HCl, pH 7.4, 20 mM glutathione, 250 mM NaCl, 2 mM DTT, 20% glycerol), vortexing and incubating for 10 min on ice. Aliquots of the eluted fractions were analysed by SDS-PAGE on a 4–12% NuPAGE pre-cast gel (Invitrogen). The eluted protein was quantified by Bradford assay. To corroborate that the purified protein was the protein of interest, a western blot was performed with an anti-GST antibody.
The construct was also identified by protein mass spectrometry. For this, it was first detected by Coomassie Blue staining, and the band of the expected molecular mass (44 kDa) was excised, diced and sent for analysis (in collaboration with the Protein Analysis Facility, Clare Hall, Cancer Research UK). Excised protein bands were subjected to in-gel trypsin digestion using a Perkin-Elmer Janus Automated Workstation. Peptide mixtures were acidified to contain 0.1% formic acid and injected onto a nanoACQUITY UPLC instrument (Waters Corporation) coupled to an LTQ-Orbitrap XL (Thermo Fisher Scientific) through an Advion Biosciences Nanomate. Peptides were eluted over a 30 min gradient [5–40% (v/v) acetonitrile]. Peak lists were extracted using Mascot distiller and searched with Mascot v.2.4.1 (Matrix Science) against the UniProt reference database. Oxidation of methionine was entered as a variable modification and the search tolerances were 5 ppm and 0.8 Da for peptides and fragments, respectively. Searches were compiled in Scaffold 4.1.1.
GST–Grp1-PH probe specificity was demonstrated using dot blots, as described previously (Jethwa et al., 2015). Commercial lipid strips were incubated with blocking buffer (3% FAF-BSA, 0.1% Tween-20 PBS) for 1 h at room temperature. 1 µg/ml of GST–Grp1-PH in blocking buffer was added and incubated for 1 h at room temperature. Excess buffer was washed with PBS-T (five washes, 5 min each). Then, anti-GST antibody (1:2500) was added in blocking buffer, incubated for 1 h at room temperature and washed as before. Finally, anti-mouse HRP-conjugated antibody (1:5000) in blocking buffer was added and incubated for 1 h at room temperature. Washes were performed as above and rinsed with PBS. Blots were photo-activated with a 1:1 mixture of the Amersham ECL reagents and exposed to photographic AGFA Cronex 5 Medical X-Ray Film in a Kodak BioMax MS cassette for various periods of time. The films were developed using the IGP Compact2 Developer.
Nuclei and egg extracts
Gamete shedding and collection of P. lividus took place as described previously (Byrne et al., 2009b) through intracoelomic injection of 0.5 M KCl. Sea urchin sperm was permeabilised with 0.1% Triton X-100, and fertilised egg cytoplasmic extract (S10) was prepared by twice vigorously homogenising the egg suspension 10 min after fertilisation with a syringe with a 22-gauge, 1.5-inch needle and centrifugation to remove the yolk (upper layer) and the pigment pellet (Byrne et al., 2009b). The S10 fraction, which includes cytosol and cytoplasmic membrane vesicles, can be subsequently fractionated using a sucrose gradient to separate the different membrane vesicle populations (MV1, MV2, etc.) that contribute to nuclear envelope formation (Byrne et al., 2009b).
Cell-free assembly of the male pronucleus
De-membraned sperm nuclei were mixed with the S10 fraction and an ATP-generating system (1:1:1 ATP:creatine phosphate:creatine phosphate kinase ratio) and incubated for 1 h at room temperature. Nuclei were decondensed, and membrane vesicles were bound to chromatin after this time. Then, GTP (1 or 5 mM final concentration) was added and incubated for 2 h at room temperature to initiate nuclear envelope formation. The reaction was stopped by diluting the samples approximately sevenfold with cold lysis buffer (10 mM Hepes, pH 8.0, 250 mM NaCl, 5 mM MgCl2, 110 mM glycine, 250 mM glycerol, 1 mM DTT, 1 mM AEBSF). For analysis, samples were settled in poly-L-lysine-coated coverslips and fixed with 3.7% PFA in isolation buffer.
For inhibitor studies of nuclear envelope assembly, LY 294002, PIK-75 or BYL719 were added before the GTP addition and incubated for 15 min.
Sea urchin egg time series
Eggs were collected in Millipore-filtered sea water (MPSW) and dejellied by passing them twice through a Nytex mesh of the appropriate pore size (210 µm) (Foltz et al., 2004). Subsequently, the eggs were fertilised and left on a rotator shaker. At various time points, post-fertilisation eggs were fixed by mixing 1 ml of the egg suspension with 1 ml of 7.4% PFA in cold isolation buffer (80 mM Pipes, pH 7.2, 5 mM EGTA, 5 mM MgCl2, 1 M glycerol). Fixation was terminated after 1 h of gentle shaking. The eggs were centrifuged (150 g, 4°C, 1 min) and washed twice with PBS. Finally, eggs were resuspended in 0.05% azide in PBS and stored at 4°C.
For inhibitor studies, inhibitors were added to eggs 3–4 min after fertilisation. Aliquots were fixed and stored as described above.
For confocal imaging of the cell-free assay, nuclei were labelled by adding Hoechst 33342 (2 µg/ml), and membranes were labelled with DiOC6 (2 µM), both in TBS. If GST–Grp1-PH was added, the sample was blocked with 3% FAF-BSA in TBS for 15 min before the recombinant protein was added in 3% FAF-BSA in TBS, and incubated for 1 h. The excess was washed with TBS, followed by labelling with Dylight650-conjugated anti-GST antibody (1:500) for 1 h. Secondary antibody alone was incubated as a control with nuclei to check the absence of signal.
Intact eggs were blocked and permeabilised before antibody addition by incubating the cells with 3% FAF-BSA and 0.5–1% saponin in TBS for 15–50 min. Then, the required antibodies were introduced into the buffer (3% FAF-BSA 0.5% saponin TBS), and samples were incubated for 1 h on an overhead mixer and washed after each incubation.
After labelling and settling onto poly-L-lysine-coated coverslips, ProLong reagent was used to mount slides. Image acquisition was performed on a Zeiss 710 upright confocal microscope run by Zen software (2009), with a 63×1.4NA oil immersion objective and the custom filter set-up for the probes as indicated; or an inverted confocal fluorescence microscope Nikon D-ECLIPSE C1 (Nikon Inc., Melville, N.Y.) with a 60×1.4 NA oil immersion objective.
Inhibitors used in the cell-free assay
In the cell-free experiments, when inhibitors were added, the formation of the nuclear envelope of at least 100 nuclei was scored in three independent experiments using a wide-field inverted Eclipse-Ti microscope (Nikon) PL APO 60×1.4 NA oil immersion objective. Because inhibitors were added in DMSO, the corresponding vehicle controls were performed.
Quantification of colocalisation
The threshold for every image was subjected to the algorithm moments (ImageJ software). This method keeps the moments of the original image in each thresholded result (Tsai, 1985).
The regions of interest were selected, and the colocalisation was calculated in the following manner. The colocalisation of the various proteins and phospholipids under investigation were analysed only in a region of interest around the female pronucleus, in unfertilised eggs, or in female and male pronuclei in fertilised eggs. This was performed in order to mainly focus on processes occurring within the vicinity of the nuclear envelope. The colocalisation measured in these regions of interest was expressed as a percentage of the total number of pixels in each region of interest – i.e. the ratio of pixels in two different channels (corresponding to the two fluorescent labels) relative to the total number of pixels in the regions of interest. These were plotted as box and whisker plots, and the statistics were performed using a non-parametric Mann–Whitney test. This quantification was performed for at least ten eggs for each condition.
Data acquisition for the measurement of FRET
Fixed eggs were blocked and permeabilised as described above. Cell autofluorescence was quenched with 2 mg/ml of sodium borohydrate (NaBH4) in PBS for 2 min, and endogenous peroxidase activity was suppressed by incubating the cells with the specific reagent for 20 min. Proteins were labelled with anti-Rab7 antibody followed by anti-chicken F(ab′)2–Atto-488, and lipids were labelled with GST–Grp1-PH and then an anti-GST antibody followed by anti-rabbit F(ab′)2 conjugated to HRP. All antibody incubations took place in 3% FAF-BSA with 0.5% saponin in PBS for 1 h. To create an enhanced acceptor signal, eggs were treated with TSA–Alexa-Fluor-594 for 10 min at room temperature (Byrne et al., 2014; Veeriah et al., 2014), as per the manufacturer's instructions, with the enhancement reaction being terminated by PBS washes (150 g, 1 min). Eggs were additionally labelled with 2 μg/ml Hoechst 33342 and mounted in ProLong Gold.
A multiple frequency domain FLIM lifetime imaging microscope from Lambert Instruments was used, and samples were excited with the 473-nm diode laser. Images were acquired on an Eclipse-Ti microscope (Nikon) PL APO 60×1.4 NA oil-immersion objective. FRET efficiency (Ef) was determined using the following formula: Ef (%)=[1−(tDA/tD)]×100; where tD is donor lifetime and tDA is the donor plus acceptor lifetime. Only samples with donor intensity at least two times higher than the autofluorescence intensity were included for FRET efficiency calculations.
Conjugation of Atto-488 to anti-chicken F(ab′)2 fragments
100 μg of anti-chicken F(ab′)2 antibody was conjugated to 25 μg of chromophore in a 100-μl volume adjusted to pH 8.5 with 50 mM sodium borate buffer. Atto-488 (donor chromophore) was conjugated to an anti-chick F(ab′)2 fragment for 1 h in the dark at 20°C with shaking. Free chromophore was separated from antibody–dye conjugate on a PD Minitrap™ G-25 column (GE Healthcare). The dye-to-protein ratio of the chromophore–antibody conjugate was calculated as per the manufacturer's instructions. Conjugates had a dye-to-protein ratio of <3:1.
Statistical analysis for the box and whisker plots was performed using the Origin software. Boxes represent the 25–75% percentiles of the data set, and the whiskers represent 100% of the population. Individual data points are shown as diamonds. Statistical significance between the groups was calculated with Mann–Whitney test. For other graphs, statistical analysis was performed in SigmaPlot subjecting data to an unpaired t-test. Differences were considered statistically significant when P≤0.05.
We thank Gloria de Las Heras and Unai Lorenzo for their help with the image software analysis.
M.G.L., R.D.B. and B.L. designed the experiments; M.G.L., D.L.P. and B.L. wrote the manuscript; M.G.L. performed the experiments; A.A. provided reagents and financial support; all authors reviewed and approved the manuscript prior to submission.
We also acknowledge the support of the Ikerbasque, Basque Foundation for Science to B.L.; the Ministerio de Economía y Competitividad to A.A. [MINECO-FEDER; BFU2015-66306-P (to A.A.) and BFU2015- 65625-P (to B.L.)]; Eusko Jaurlaritza (the Basque Government) [IT838-13] to B.L.; and a Faculty Research Award of the Axel Schupf ‘57 Fund for Intellectual Life and Senior Sabbatical Fellowship, Amherst College  to D.P. M.G.L. was a doctorate fellow supported by Eusko Jaurlaritza.
The authors declare no competing or financial interests.