Dual control of formin-nucleated actin assembly by the chromatin and ER in mouse oocytes

Summary

The first asymmetric meiotic cell divisions in mouse oocytes is driven by formin 2 (FMN2)-nucleated actin polymerization around the spindle. In this study, we investigated how FMN2 is recruited to the spindle peripheral ER and how its activity is regulated in mouse meiosis I (MI) oocytes. We show that this process is regulated by the Ran GTPase, a conserved mediator of chromatin signal, and the ER-associated protein VAPA. FMN2 contains a nuclear localization sequence (NLS) within a domain (SLD) previously shown to be required for FMN localization to the spindle periphery. FMN2 NLS is bound to the importin α1/β complex, and disruption of this interaction by Ran^GTP is required for FMN2 accumulation in the area proximal to the chromatin and the MI spindle. The importin-free FMN2 is then recruited to the surface of ER around the spindle through binding of the SLD with the ER membrane protein VAPA. We further show that FMN2 is autoinhibited through an intramolecular interaction between the SLD with the C-terminal formin homology 2 (FH2) domain that nucleates actin filaments. VAPA binding to SLD relieves the autoinhibition of FMN2, leading to localized actin polymerization. This dual control of formin-mediated actin assembly allows actin polymerization to initiate the movement of the meiotic spindle toward the cortex, an essential step in the maturation of the mammalian female gamete.

In brief

The first asymmetric meiotic division is essential for the formation of a functional mature female gamete. Wang et al. demonstrates that dual control of formin 2-mediated actin assembly by Ran GTPase signalling and an ER membrane protein restricts actin polymerization from the ER surface at the spindle periphery. This initiates the meiotic spindle migration from the center of the oocyte to the cortex.

Graphical Abstract

Introduction

De novo nucleation of actin filaments is a key step in the temporal and spatial remodeling of the actin cytoskeleton, which is critical for complex cell functions, such as cell morphogenesis, movement, division, and adhesion ^1,2. Formins are a family of conserved eukaryotic proteins that control the nucleation and elongation of actin filaments through the conserved formin homology 1 and 2 (FH1 and FH2) domains ^3–5. The sequences flanking FH1 and FH2 are usually involved in the spatial localization and regulation of the formin activity ^3,6. The best understood are Diaphanous-related formins (DRFs), which assume an autoinhibited conformation through intramolecular interactions between the C-terminal Diaphanous Auto-regulatory Domain (DAD) and the N-terminal Diaphanous Inhibitory Domain (DID). This autoinhibition can be released upon binding to the GTP-bound form of Rho family GTPases ^7–10. Non-DRF formins such as the mammalian FMN1 and FMN2 have no DAD domain. The only well-known binding partner of FMN2 is Spire, a protein also possessing actin nucleation activity that binds to the formin Spire interaction (FSI) motif at the very C-terminal ends of the FMN2 ^11–15. Although existing data suggest that Spire and FMN2 collaborate in actin polymerization, how their activities are regulated remain poorly understood.

FMN2 is highly expressed in human and mouse oocytes and the central nervous system, and its mutations are associated with infertility, intellectual disability, and several cancers ^16–21. In oocytes, FMN2 is a key regulator of the first asymmetric meiotic (MI) cell division during oocyte maturation. A critical step in MI division, which depends on FMN2-mediated actin polymerization, is the movement of the meiotic spindle from a central to a cortical location in the oocyte ^17,22–24. Fmn2^−/− oocytes cannot correctly position the spindle to extrude the first polar body and thus fail to produce a fertile gamete ¹⁷. Previous studies from us and others showed that FMN2 is localized to the mouse oocyte cortex before germinal vesicle (GV – oocyte nucleus) break down (GVBD). In addition to cortical localization, FMN2 gradually accumulated around the MI spindle after GVBD ^22,25,26. The spindle-peripheral FMN2 nucleates F-actin to initiate MI spindle migration ²⁴, but the mechanism of FMN2 localization and activation at the spindle periphery remained unclear.

Previous studies also showed that meiotic maturation is controlled by Ran^GTP, which forms a spatial gradient emanating from the meiotic chromatin due to chromatin-associated guanine nucleotide exchange factor RCC1 ^27–31. In this study, we showed that Ran^GTP and its effector importin regulate FMN2-mediated actin nucleation by promoting the accumulation of FMN2 at the surface of ER vesicles that accumulate around the MI spindle. Furthermore, we showed that FMN2, while does not possess a DAD domain, is nonetheless autoinhibited. The actin nucleation activity of FMN2 can be activated upon binding to the ER-membrane protein VAPA (vesicle-associated membrane protein-associated protein A). This dual regulation by Ran and VAPA ensures the activation of FMN2 and the production of the actin-driven forces to occur at the ER membrane around the spindle. Consequently, the FMN2-dependent actin nucleation enables the MI spindle to migrate cortically to enable asymmetric MI cell division.

Results

Ran^GTP is required for spindle-peripheral FMN2 localization and spindle migration.

The Ran^GTP gradient is known to regulate a variety of cellular processes in the proximity of chromosomes ^30–34 (Figure S1A). Here, we tested whether Ran^GTP also regulates the localization of FMN2 to the spindle periphery during MI. To block Ran^GTP production, we expressed a GDP-locked mutant Ran^T24N, which acts as an inhibitor of Ran guanine nucleotide exchange factor RCC1 ^35,36. Ran^T24N prevented the localization of FMN2-AcGFP to the spindle periphery (Figures 1A, 1B, S1B, and S1C). We have previously shown that the FMN2 targeting to the oocyte spindle periphery is mediated through the spindle-periphery localization domain (SLD, aa 275–734 of FMN2) ²⁴. As expected, Ran^T24N also abolished SLD-mCherry’s ability to localize to the spindle periphery (Figures S1D and S1E). Live imaging of oocytes expressing Ran^T24N further confirmed that the MI spindle was static and failed to migrate to the cortex (Figures 1C–1F; Video S1). We also expressed RanGAP1 (Ran GTPase Activating Protein 1), which converts cytoplasmic Ran^GTP into Ran^GDP (Figure S1A), and found that it also disrupted spindle peripheral localization of FMN2 and SLD and spindle migration (Figures 1, S1D, and S1E; Video S1). These results together revealed that Ran^GTP regulates the recruitment of FMN2 to the spindle periphery and is required for spindle movement towards the cortex.

Figure 1. Ran^GTP is required for spindle peripheral recruitment of FMN2 and MI spindle migration.

(A) Representative images of FMN2 localization in control, Ran^T24N-expressing, and RanGAP1-expressing oocytes. The bottom shows fluorescence intensity profiles of FMN2 along a thick line (Figures S1B and S1C) crossing the mid-zone of the MI spindle, smoothened curves are shown.

(B) Quantification of FMN2 spindle periphery intensity ratio (FMN2 intensity at spindle periphery/FMN2 intensity outside spindle periphery) in control, Ran^T24N-expressing, and RanGAP1-expressing oocytes. All boxes plot defines the 25th and 75th percentiles, with a line at the median and error bars (whiskers) defining the 10th and 90th percentiles. Points below and above the whiskers are drawn as individual points. Number of oocytes is specified in brackets.

(C) Representative time-lapse images of spindle/chromosomes migration in control, Ran^T24N-expressing, and RanGAP1-expressing oocytes. The panel on the far right of each row shows the trajectory of chromosome movement. White dashed spheres mark initial chromosome positions.

(D) Chromosomes were tracked in oocytes (see Methods) as shown in (C) and the chromosomes movements were plotted. Number of oocytes is specified in brackets.

(E) The speed of spindle/chromosomes migration was analyzed from the plots in (D). Number of oocytes is specified in brackets.

(F) Quantification of the percentage of oocytes that underwent spindle migration in control, Ran^T24N-expressing, and RanGAP1-expressing oocytes. Data are from three independent experiments. Number of oocytes is specified in brackets.

Scale bars, 10 μm (for all images). ***P< 0.001; ****P< 0.0001

See also Figure S1, Video S1.

Importin α/β mediate the Ran^GTP regulation of FMN2 localization to the spindle periphery

A common mechanism by which Ran^GTP regulates downstream processes is by binding to importin β, a component of importin complex, which sequesters an NLS-containing protein with a key downstream function, thereby triggering a release of this NLS-containing protein from the importin complex ^37,38. Examination of FMN2 SLD amino acid sequence indicated the existence of a putative NLS (aa 441–449) (Figure 2A). To determine whether this is a functional NLS, we fused it to an Enhanced Yellow Fluorescent Protein (EYFP) (EYFP-NLS) and expressed in the oocytes arrested in the germinal vesicle stage which precedes MI. This experiment showed that the FMN2 NLS sequence was sufficient to drive the accumulation of EYFP in the nucleus, which was prevented by Ran^T24N expression (Figures 2B and 2C). To test whether FMN2 localization is also regulated by importin β, we expressed a Ran^GTP‐ resistant mutant of importin β (importin β^{71– 876}), which functions as a dominant active importin β in mouse oocytes ^29,39. Importin β^71–876 expression markedly reduced the localization of FMN2 and SLD to the spindle periphery (Figures 2D, 2E, S2A, and S2B). We next mutated the NLS motif in SLD at three positively charged residues (K444A/K445A/R447A, SLD^{NLS mut}), which should disrupt importin binding to NLS. SLD^{NLS mut} was resistant to mislocalization caused by importin β^71–876 (Figures S2A and S2B), suggesting that NLS mediates the inhibitory effect of importin β^71–876 and that SLD can localize in the absence of importin-NLS interaction.

Figure 2. Spindle peripheral recruitment of FMN2 is regulated by Ran^GTP and importin β-NLS interaction.

(A) Domain organization of mouse FMN2 including a putative nuclear localization sequence (NLS) site within SLD.

(B) Localization of EYFP and EYFP fused putative NLS (EYFP-NLS) in GV oocytes. White dashed spheres mark nuclear (GV) positions.

(C) Nucleocytoplasmic ratio of EYFP in EYFP-expressing, EYFP-NLS-expressing, and EYFP-NLS and Ran^T24N-coexpressing GV oocytes. Number of oocytes is specified in brackets.

(D) Representative images of FMN2 localization in control, and Importin β⁷¹⁻⁸⁷⁶ (Impβ⁷¹⁻⁸⁷⁶) expressing oocytes. The fluorescence intensity trace of FMN2 is shown on the bottom as in Figure 1A.

(E) Quantification of FMN2 spindle periphery intensity ratio (FMN2 intensity at spindle periphery/FMN2 intensity outside spindle periphery) in control, and importin β⁷¹⁻⁸⁷⁶-expressing oocytes. Number of oocytes is specified in brackets.

Scale bars, 10 μm (for all images). ****P< 0.0001.

See also Figure S2.

Importin β can bind NLS-containing cargos directly or through adaptor proteins such as importin α isoforms ^40–43. We used a recombinant SLD protein to test its binding to importins. SLD was expressed with a cleavable glutathione transferase (GST) fusion tag. After tag cleavage, the purified protein was immobilized on N-hydroxysuccinimide activated magnetic beads and used as bait to pull down bound proteins from mouse oocyte extracts, which were then identified by using quantitative multi-dimensional protein identification technology (MudPIT) analysis (Figure S2C). Three importin proteins, importin α1 (KPNA2), importin α8 (KPNA7), and importin β (KPNB1), were pulled down by SLD, with importin α1 being the most abundant among the three (Figure 3A). Direct binding of importin α1^ΔIBB (importin α1 lacking the importin β-binding domain) to GST-SLD was confirmed in vitro by using purified recombinant proteins (Figure 3B). Importin α8 can also bind to GST-SLD but much more weakly than importin α1^ΔIBB, whereas importin β^71–876 did not show direct binding to SLD, indicating that importin α1 is likely to mediate the association of importin β to SLD in vivo (Figure 3B and S2D). Indeed, importin α1^ΔIBB expression via mRNA injection disrupted FMN2 spindle periphery localization (Figures 3C and 3D) and spindle migration (Figures 3E–3H; Video S1). These results demonstrate that Ran^GTP regulates FMN2 spindle periphery localization by preventing binding of the importin α1 and β complex to NLS.

Figure 3. FMN2 is a direct cargo of importin α1.

(A) Mass spectrometry analysis revealed 3 importin proteins associated with SLD in mouse oocyte extracts. The protein abundance was scored by the distributed normalized spectral abundance factor (dNSAF).

(B) GST pulldown assay using purified proteins to test the direct interactions between SLD and importins. Same molar concentration (0.66 μM) of His-tagged importin proteins as indicated was pulled down by GST-SLD-loaded (50 μg/ml or 0.66 μM) beads. Importins were shown by bands in the bottom panel. Estimated position for importin β⁷¹⁻⁸⁷⁶ is indicated with a dashed box.

(C) Representative images of FMN2 localization in control, and importin α1^ΔIBB (Impα1^ΔIBB) expressing oocytes. The fluorescence intensity trace of FMN2 is shown on the bottom as in Figure 1A.

(D) Quantification of FMN2 spindle-periphery intensity ratio (FMN2 at spindle periphery intensity/FMN2 outside spindle periphery intensity) in control or importin α1^ΔIBB-expressing oocytes. Number of oocytes is specified in brackets.

(E) Representative time-lapse images of spindle/chromosomes migration in control, importin β^71-876, importin α1^ΔIBB-expressing oocytes. The panel on the far right shows the trajectory of chromosome movement.

(F) Chromosomes were tracked in oocytes as shown in (E) and chromosomes movements were plotted. Number of oocytes is specified in brackets.

(G) The speed of spindle/chromosomes migration was analyzed from the plots in (F). Number of oocytes is specified in brackets.

(H) Quantification of the percentage of oocytes that underwent spindle migration in control, importin β⁷¹⁻⁸⁷⁶-expressing, importin α1^ΔIBB-expressing oocytes. Data are from three independent experiments. Number of oocytes is specified in brackets.

Scale bars, 10 μm (for all images). ****P< 0.0001.

See also Figure S2, Data S1, and Video S1.

The ER protein VAPA interacts with SLD and is required for FMN2 localization

Previous studies showed that FMN2 not only accumulates around the spindle but also is bound to the cytoplasmic surface of ER vesicles ^24,25. Our quantitative model demonstrated that the ER association of FMN2 is important for the production of dynamic pushing force via actin polymerization and symmetry breaking of this force to initiate spindle migration ²⁴. While Ran^GTP and importin drive FMN2 accumulation around the spindle, a factor that causes the association of FMN2 with ER should be elucidated. Since FMN2 does not have a transmembrane domain, this association is likely to be mediated through an ER membrane protein. Among the 504 proteins identified by MudPIT that bind specifically to SLD but not the FMN2 cortical localization domain (CLD) ²⁴ (Figure 4A), 26 were known ER membrane proteins (Figure 4A; Table S1). VAPA, an integral ER membrane protein known to mediate ER contacts with multiple organelles, had the highest association scores with SLD (Figure 4A). We first confirmed that VAPA associates homogeneously with the ER, as labelled with AcGFP-tagged Sec61β and BFP-tagged KDEL (Figures S3A-S3C). FMN2 or SLD indeed colocalizes with VAPA at the spindle periphery in mouse MI oocytes (Figures 4B, and S3D-S3F). To confirm the interaction between SLD and VAPA, we first performed co-immunoprecipitation of SLD-EGFP with full-length mCherry-taged VAPA, which were co-expressed in HEK293 cells (Figure S3G). We thus confirmed the interaction between VAPA and SLD by pull-down assay in somatic cells (Figure S3G). A direct interaction between FMN2 and VAPA was further confirmed by in vitro binding between a purified (His)₆-tagged cytosolic portion of VAPA (aa 8–226, VAPA^ΔTM) and GST-SLD (Figure 4C). To gain insight into the domain(s) of VAPA involved in binding SLD, we purified its N-terminal major sperm protein domain (MSP, aa 8–131), and central domain (CD, aa 132–226) (Figure S3H). Interestingly, neither the MSP domain nor the central domain was sufficient to bind SLD (Figure S3I), suggesting that the entire cytoplasmic domain of VAPA is required for binding SLD or that this interaction requires some structural features that are lost in the smaller VAPA fragments.

Figure 4. VAPA recruits FMN2 to ER surface by binding to SLD.

(A) Venn diagram summarizing the number of proteins that interact with CLD or SLD of FMN2 (left panel), which were identified via mass spectrometry analysis of the pull-downs from mouse oocyte extracts. 26 out of 504 proteins, which exclusively interact with SLD, are ER membrane proteins. Top 10 abundantly present ER membrane proteins in the SLD-binding partners in mouse oocytes are shown on the right panel. The protein abundance was scored and sorted by the distributed normalized spectral abundance factor (dNSAF).

(B) Representative images of MI oocyte expressing VAPA-mCherry and FMN2-AcGFP, showing that VAPA co-localizes with FMN2 at spindle periphery. Line scan and colocalization analysis (Pearson coefficient) of VAPA and FMN2 are shown at the bottom.

(C) GST pulldown assay by using GST-SLD (50 μg/ml or 0.66 μM) as bait to pull down His-VAPA^ΔTM (0.66 μM).

(D) Schematic diagram for generating phase-separated fluorescent foci (Fluoppi foci). Ash tag and tetrameric fluorescence protein hAG are fused to bait (VAPA) and prey (SLD) proteins, respectively. Interaction between bait and prey proteins produces fluorescent foci inside cells via phase separation.

(E) Fluoppi assay demonstrating an interaction between SLD and VAPA in mouse oocyte. Foci of SLD-hAG were observed by confocal microscopy.

(F-H) Quantification of Fluoppi foci number (H) size (I) and intensity (J) in different groups (Ash + SLD-hAG, VAPA-Ash + SLD-hAG, VAPA-Ash + SLD-hAG + importin α1^ΔIBB).

Scale bars, 10 μm (for all images). **P< 0.01; ****P< 0.0001.

See also Figure S3, Data S1, and Table S1.

Pre-incubating importin α1^ΔIBB with SLD prevented SLD from binding to VAPA^ΔTM and vice versa (Figures S3J and S3K), suggesting that importin α1 and VAPA compete for binding to SLD. If so, releasing importin α1 from SLD by Ran^GTP should allow FMN2 to bind VAPA in vivo (Figure S3L). To validate the interaction between VAPA and SLD in MI oocytes, we used fluorescent-based technology to detect protein-protein interactions (FluoPPI) ^44,45. Fluoppi involves fusing one of the potential binding partners with the tetrameric green fluorescent protein Azami Green (hAG) and fusing the counter part with Assembly helper (Ash) tag. Binding of partners would result in fluorescent puncta whereas a lack of binding is indicated by diffused fluorescence (Figure 4D). Using this method, we observed prominent fluorescence puncta in oocytes expressing SLD-hAG and VAPA-Ash (Figure 4E), supporting the interaction between SLD and VAPA in live oocytes. As expected, the puncta number and size were both reduced by importin α1^ΔIBB expression (Figure 4E–4H).

To test the importance of VAPA for FMN2 localization and spindle migration, we used the “Trim-Away” method ^46,47 of antibody-mediated degradation of endogenous VAPA in mouse oocytes. After Trim-Away using an anti-VAPA antibody (see Methods), VAPA signal was diminished at the spindle periphery (Figures S4A-S4C). VAPA depletion significantly disrupted FMN2 but not ER spindle-peripheral localization (Figures 5A, 5B, S4D, and S4E) and reduced the spindle-peripheral F-actin accumulation in oocytes (Figures 5C and 5D). Live imaging of these oocytes showed that VAPA Trim-Away markedly reduced both the speed and the distance of chromosomes migration toward the cortex (Figures 5E–5H; Video S2). These results demonstrate that VAPA-driven recruitment of FMN2 to the ER membrane is important for actin polymerization at the spindle periphery and spindle migration.

Figure 5. VAPA is required for FMN2-mediated actin polymerization around the spindle and spindle migration.

(A) Representative images of FMN2 localization in control, and VAPA Trim-Away oocytes. The fluorescence intensity trace of FMN2 is shown on the bottom as in Figure 1A.

(B) Quantification of FMN2 spindle-periphery intensity ratio (FMN2 intensity at spindle periphery/FMN2 intensity outside spindle periphery) in control and VAPA Trim-Away oocytes. Number of oocytes is specified in brackets.

(C) Representative images of phalloidin staining of F-actin after VAPA Trim-Away. The accumulation of F-actin surrounding the spindle was reduced in VAPA Trim-Away oocytes. The bottom panels show fluorescence intensity profiles of F-actin along a thick line crossing the mid-zone of the MI spindle, smoothened curves are shown.

(D) Quantification of F-actin spindle-periphery intensity ratio (F-actin intensity at spindle periphery/F-actin intensity outside spindle periphery) in control and VAPA Trim-Away oocytes. Number of oocytes is specified in brackets.

(E) Representative time-lapse images of spindle/chromosomes migration in control, VAPA Trim-Away oocytes. The panel on the far right shows the trajectory of chromosome movement.

(F) Chromosomes were tracked in oocytes as shown in e and the chromosomes movements were plotted. Number of oocytes is specified in brackets.

(G) The speed of spindle/chromosomes migration was analyzed from the plots in (F). Number of oocytes is specified in brackets.

(H) Quantification of the percentage of oocytes that underwent spindle migration in control, VAPA Trim-Away oocytes. Data are from three independent experiments. Number of oocytes is specified in brackets.

Scale bars, 10 μm (for all images). ****P< 0.0001.

See also Figure S4 and Video S2.

VAPA regulates autoinhibition of actin nucleation by FMN2

DRFs are known to be autoinhibited through interaction of the DAD domain with the DID domain. Interestingly, SLD of FMN2 can bind the FH2 domain located near the C-terminus of FMN2 (Figure 6A), raising the possibility that SLD works as an autoinhibitory domain that controls the actin polymerization activity of FH2 (Figure 6B). We therefore tested the ability of purified SLD to inhibit the actin polymerization activity of FMN2’s FH2 domain in a pyrene-actin assembly assay. Indeed, SLD inhibited FH2’s ability to stimulate actin polymerization in a dose-dependent manner, whereas SLD alone did not affect actin polymerization (Figures 6C, 6D, S5A, and S5B). As an independent assay, we observed fluorescent actin filament assembly using total internal reflection fluorescence (TIRF) microscopy in the presence and absence of SLD and FH2 (Figures 6E and 6F). TIRF assays showed that SLD inhibited actin nucleation by FH2 (Figures 6E and 6F). These data suggest that the interaction between SLD and FH2 confers the autoinhibition of FMN2’s actin nucleation activity.

Figure 6. SLD-FH2 interaction confers autoinhibition of FMN2.

(A) GST pull-down assay by using GST-tagged SLD (50 μg/ml or 0.66 μM) to pull down His-tagged FH2 (0.66 μM).

(B) Schematic of hypothesized FMN2 autoinhibition based on intramolecular interaction between SLD and FH2 and relief of autoinhibition by VAPA binding to SLD.

(C) Pyrene-actin polymerization assays showing a dose-dependent inhibition of FH2 by SLD. FH2 was used at 50 nM for all conditions except the actin alone control. SLD, 0.25, 1, 1.25, 1.75, and 2 μM. Actin was used at 1.5 μM with 5% being pyrene labeled.

(D) Actin polymerization rate (arbitrary unit) was calculated by using the slops of pyrene fluorescence curves at 60 s – 120 s.

(E) TIRF actin polymerization assays with actin alone (500 nM in all cases), actin + FH2 (25 nM), and actin + FH2 (25 nM) + SLD (0.8 μM). Representative images at t = 0 s and t = 90 s were shown.

(F) Actin seeds number was quantified at t = 90 s. Each dot represents the seeds number from the whole 150 μm × 150 μm field of view. Scale bar, 20 μm. Statistics from one-way ANOVA, with **P< 0.01, ****P< 0.0001; ns, not significant.

See also Figure S5 and Data S1.

We next asked whether VAPA activates FMN2 by relieving this autoinhibition (Figure 6B). Increasing VAPA^ΔTM concentration progressively inhibited the binding of FH2 by SLD in a pulldown assay (Figure 7A). Furthermore, VAPA^ΔTM rescued the actin nucleation activity of FH2 in the presence of SLD in both the pyrene-actin assay and the TIRF assay (Figures 7B-7E, S5C-S5E; Video S3). In the absence of SLD, VAPA^ΔTM did not pulldown FH2 or increase actin nucleation by FH2 but rather slightly negatively affected actin polymerization (Figures S5F-S5H). These results support our hypothesis that VAPA binding to SLD not only helps localize FMN2 to spindle peripheral ER but also activates local actin polymerization by relieving FMN2 from an autoinhibited state. It is worth noting that, even though importin α1^ΔIBB at higher concentration also disrupted binding of FH2 by SLD in a pull-down assay (Figure S5I), its stimulatory activity on actin nucleation by FH2 in the presence of SLD was much weaker than VAPA (Figures 7D, 7E, S5A, S5B, and S5J-S5M).

Figure 7. Relief of FMN2 autoinhibition by VAPA.

(A) GST pulldown assay showing competition between VAPA and FH2 for binding to SLD. The same amount of GST-tagged SLD (50 μg/mL or 0.66 μM) was immobilized on the beads first and then His-tagged VAPA^ΔTM of varied concentrations ((0, 0.5, 1, and 2) × 0.66 μM) and His-tagged FH2 (0.66 μM) were added to the reaction.

(B and C) Pyrene-actin assembly assays showing stimulation of actin assembly by FH2, inhibition of the stimulation by SLD, and its restoration by VAPA ^{Δ TM} (B). Quantification of actin polymerization rate at 60 s - 180 s is shown in (C). Actin was used at 1.5 μM with 5% being pyrene labeled. SLD was used at 2 μM. VAPA^ΔTM was used at 0.5 μM. See also Figure S5L for different concentrations of VAPA^ΔTM were used.

(D and E) TIRF actin polymerization assays showing VAPA^ΔTM but not importin α1^ΔIBB rescued the actin nucleation activity of FH2 in the presence of SLD (D). Representative images at t = 0 s and t = 90 s were shown. The quantification of actin seeds number at t = 90 s is shown in (E). Concentrations of proteins used: actin 500 nM, FH2 25 nM, SLD 0.8 μM, VAPA^ΔTM 0.4 μM, importin α1^ΔIBB 0.4 μM. Scale bar, 20 μm. Each dot represents the seeds number from the whole 150 μm × 150 μm field of view.

(F) Schematic model of FMN2 recruitment to the spindle periphery and activated by VAPA.

Scale bars, 10 μm (for all images). Statistics from one-way ANOVA, with *P< 0.05, **P< 0.01, ****P< 0.0001; ns, not significant.

See also Figure S5, Data S1, and Video S3.

Discussion

Collectively, the results above suggest a model of FMN2 localization and activation controlled by both the meiotic chromatin and ER (Figure 7F). First, FMN2 is freed from importin binding at the spindle periphery by the high Ran^GTP concentration generated by the chromatin-associated Ran GEF. This enables the importin-free FMN2 to interact with VAPA and to be recruited to ER surface. The VAPA binding also concomitantly relieves the autoinhibition of FMN2 by disrupting the interaction between SLD and FH2, leading to localized actin polymerization. Our previous work showed that the actin bundles nucleated from the ER surface push against the surrounding mitochondria to initiate spindle movement ²⁴.

It is known that Ran GTPase mediated chromatin signalling acts as a ‘genome-positioning system’ (GPS) in the mitotic cells ^37,48 and controls cortical polarity during oocyte meiotic division ^28,30,31,34. Here we have uncovered another, previously unknown function of Ran GTPase that regulates mouse oocyte MI spindle migration that relies on FMN2. Our proteomics analysis and in vitro pull-down assays confirmed that importin α1 directly interacts with FMN2 SLD. We also showed that NLS-disrupting mutation significantly increased FNN2 SLD spindle periphery localization and confers resistance to mislocalization caused by importin β^71–876 (Figures S2A and S2B). These data suggested that Ran-importin signalling pathway directly regulates FMN2 to the spindle periphery through FMN2’s NLS motif. Interestingly, we also found that high-level inhibition of Ran-importin signalling could disrupt ER localization (Figures S6A-S6D), suggesting that Ran has dual functions regulating ER localization around the spindle and recruitment of FMN2 to spindle periphery ER, and these functions require different thresholds of Ran^GTP. Inhibition of Ran-importin signalling also moderately reduced cortical localization FMN2 (Figures 1A, 2D, 3C, and S6E). While this pool of FMN2 is not required for spindle migration ²⁴, this observation may suggest that an increased pool of importin-bound FMN2 also interferes with FMN2 localization to the cortex.

The mechanism by which FMN2 activation is regulated as revealed by our study follows a similar principle as the regulation of DRFs yet involves different molecular players and interactions. First, the autoinhibition involves a portion of FMN2 (SLD) N-terminal to FH1-FH2 that has no sequence similarity to the DAD or DID domain of DRFs. Furthermore, relieving the autoinhibition of these formins involves different small GTPases. It is well known that Rho family GTPases activate DRFs by disrupting the autoinhibited conformation to stimulate actin polymerization typically associated with the plasma membrane ^5,6,10,49,50. In the case of FMN2, autoinhibition is conferred through an intramolecular interaction different from that in DRFs, and a different small GTPase, Ran, controls FMN2 activation by enabling binding of the SLD by the ER membrane protein VAPA, thereby disrupting autoinhibition. This two-tiered regulation restricts formin-mediated actin polymerization not only from the ER surface but also in the periphery of the spindle.

More detailed structure-function analysis will be needed to further elucidate the mechanism of FMN2 autoinhibition and its reversal by VAPA. Moreover, the potency of sub-stoichiometric concentration of VAPA^ΔTM in relieving the FH2 inhibition by SLD is intriguing (Figures 7B, 7C, and S5D). While a trivial explanation may be that only a fraction of the recombinant SLD was competent to inhibit FH2 and bind VAPA, an alternative possibility is that VAPA has a higher binding affinity with SLD when SLD is in complex with FH2 than with free SLD, which would enable VAPA to dissociate SLD-FH2 in a semi-catalytic manner. It is also interesting that although importin α1 prevented VAPA binding to SLD likely through competition, importin α1 did not have a significant effect on FH2 nucleation, suggesting that the effects of importin and VAPA binding to SLD on the activity of FH2 are different.

The molecular players studied here in mouse oocytes, namely FMN2, VAPA, and Ran/importin, are all present in many somatic cell types. It will be interesting to investigate whether the regulatory mechanisms uncovered in this study are also important for actin assembly in the somatic cell cytoplasm. VAPA has been mostly studied in the context of organelle tethering ^51–55. Perhaps there is a hidden role for actin in this function of VAPA, whereby VAPA stimulates actin polymerization to control the extent or duration of the contact between organelles such as ER and mitochondria. In addition, since nuclear actin filament has recently shown to be present in mouse zygotes and important for efficient DNA damage repair and embryonic development ^56,57, it will be interesting to determine whether Ran/importin-mediated nuclear import of FMN2 plays a role in nuclear F-actin assembly during early embryo development.

STAR METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Rong Li (mbihead@nus.edu.sg).

Materials availability

All unique/stable reagents and plasmids generated in this study are available from the Lead Contact with a completed Materials Transfer Agreement.

Data and code availability

All data reported in this paper will be shared by the lead contact upon request.

This paper does not report original code.

Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Animals

All mouse care and use were approved by the Institutional Animal Care and Use Committee at the National University of Singapore or Johns Hopkins University. Mice were housed under a 12-h light/12-h dark cycle in individually ventilated cages with continuous access to food and water. Mice were fed a diet containing low fiber (5%), protein (20%), and fat (5–10%). Mouse rooms were maintained at 30–70% relative humidity and a temperature of 18–26 °C (64–79 °F) with at least 10 room air changes per hour. Animals were housed grouped (maximum 5 females per cage). Euthanasia of mice was performed by carbon dioxide asphyxiation, followed by cervical dislocation. Fully-grown GV oocytes were isolated from 8 to 10 weeks old female CD1 mice in M2 medium supplemented with 0.2 mM 3-isobutyl-1-methylxanthine (IBMX) (Sigma-Aldrich). To induce resumption of meiosis, oocytes were incubated in IBMX-free M16 medium.

METHOD DETAILS

mRNA transcription and oocyte microinjection

Capped mRNA was synthesized from linearized plasmid templates using the T7 or T3 mMESSAGE mMACHINE Transcription Kit (Ambion), polyadenylated (Poly(A) Tailing Kit, Ambion), and purified with the RNeasy MinElute Cleanup Kit (QIAGEN). Approximately 5 to 10 pl of capped mRNAs were microinjected into the cytoplasm of oocytes using a micromanipulator (IM-300 microinjector, Narishige). After injection, the oocytes were cultured at 37°C with 5% CO2 in M16 medium supplemented with 0.2 mM IBMX for at least 3 hours to allow protein expression. Oocytes were microinjected with the following RNAs: YPet-H2B ²⁷, H2B-mCherry ²⁷, mCherry-Trim21 ⁴⁶, EYFP, EYFP-NLS, VAPA-mCherry (rat VAPA constructs obtained from addgene, #18874), SLD-mCherry ²⁴, SLD-EGFP ²⁴, SLD^{NLS mut}-mCherry at 500 ng/μl; FMN2-AcGFP²⁴, Ran^{T24N 29}, RanGAP1 ²⁷, Importin α1^ΔIBB (constructs obtained from addgene, #26678), Importin β^{71−876 29}, SLD-hAG (hAG-tag obtained from MBL Life Science, AM-8011M), VAPA-Ash (Ash-tag obtained from MBL Life Science, AM-8011M) at 800 ng/μl. For RanGAP1 and Importin β⁷¹⁻⁸⁷⁶ overexpression, the mRNA concentration was used at 2000 ng/μl.

Confocal microscopy

For live oocyte imaging, images were acquired with Zeiss LSM980 or LSM780 microscope using a 40× C-Apochromat 1.2–numerical aperture water-immersion objective and maintained oocytes at 37°C with 5% CO2. For fixed oocytes, images were acquired using the Zeiss LSM980 or LSM880 microscopes and processed after acquisition using ZEN. In some images, shot noise was reduced with a Gaussian filter.

Immunofluorescence

For immunofluorescence staining, the oocytes were fixed for 30 min in 4% paraformaldehyde in phosphate-buffered saline (PBS) and then transferred to a membrane permeabilization solution (0.5% Triton X-100) for 20 min. After blocking with 2% bovine serum albumin in PBS, the oocytes were stained with the first antibody overnight at 4°C and then subjected to a secondary antibody at room temperature for 1 hour. Primary antibodies used were mouse anti-VAPA (sc-293278, Santa Cruz Biotechnology; 1:50), rabbit anti–VAPA (HPA009174, Sigma; 1:100). For secondary antibodies, Alexa Fluor 488–labeled anti-mouse (Invitrogen), Alexa Fluor 488–labeled anti-rabbit (Sigma-Aldrich), Alexa Fluor Plus 647–labeled anti-mouse (Invitrogen) antibodies were used. F-actin was stained with Alexa Fluor 488–labeled phalloidin (#8878, Cell Signaling Technology). DNA was stained with Hoechst 33342 (10 μg/ml).

VAPA Trim-Away

The VAPA knockdown by the Trim-Away method was performed as previously described with some modifications^46,47. In brief, anti-VAPA antibody was concentrated using Amicon Ultra-0.5 100 KDa centrifugal filter devices (Millipore) and replaced the buffer with PBS. Mouse GV oocytes were injected with anti-VAPA antibody (final concentration of 1 μg/μl) and anti-mouse antibody (final concentration of 5 μg/μl) and kept 1 h in IBMX-containing M16 medium. Then GV oocytes were injected with Trim21 mRNA (1000 ng/μl) for 3–5 hours to allow protein expression. For the control experiment, normal mouse immunoglobulin G (IgG) antibody was used instead of anti-VAPA antibody.

G-actin purification and biotin-actin preparation

G-actin was extracted generally following the established protocol ⁵⁸. In short, rabbit muscle acetone powder was rehydrated in G-buffer (5 mM Tris pH = 8.0, 0.2 mM ATP (I2481C50, GoldenBio), 0.1 mM CaCl₂, 0.5 mM dithiothreitol (DTT, DTT25, GoldenBio)) overnight. The crude mix was centrifugated at 15,000 rpm using Ti45 rotor (Beckman) for 30 min. The supernatant was collected and filtered by double-layered cheesecloth. To polymerize actin, 50 mM KCl and 2 mM MgCl₂ were added into the supernatant and gently stirred for 1 hour then KCl powder was added to a final concentration of 0.8 M following 30 min more stirring to get rid of actin-binding proteins. The solution was then pooled and ultracentrifuged at 35,000 rpm using Ti45 rotor for 2 hours. The supernatant was discarded and the pellet was rinsed with G-buffer and then homogenized using Dounce homogenizer. The solution then was loaded into dialysis tubing with 10 kDa cutoff (68100, Thermo Scientific) and dialyzed 4 times with around 12 hours each time against G-buffer. The supernatant was later collected and ultracentrifuged at 55,000 rpm using Ti70 rotor (Beckman) for 3 hours. The supernatant was further gel filtrated by using HiLoad 16/600 Superdex75 column (GE Healthcare). The peaks were checked by coomassie blue staining (Data S1) and pooled and keep at 4 °C and were consumed in 1 month. All steps are performed at 4 °C.

To label actin with biotin, the crude solution after homogenization was dialyzed against G-buffer without DTT overnight twice. 2 × labeling buffer (50 mM Imidazole pH = 7.5, 200 mM KCl, 0.3 mM ATP, 4 mM MgCl₂) was added into actin solution along with 12–15 fold molar excess of NHS-dPEG4-biotin (QBD10200, Sigma) and mixed gently overnight. Then actin was pelleted down, dialyzed, gel filtrated, and pooled as non-labeled actin. All steps were also performed at 4 °C and biotin-actin was flash-frozen by liquid N₂ and kept at –80 °C.

Plasmids used in protein production/co-immunoprecipitation (co-IP), and protein purification

GST tagged SLD, CLD were cloned into pGEX-6P-1 vector. GST tagged VAPA^ΔTM was cloned into pGEX-T vector. His tagged FH2 (G1129-T1578, from FMN2 construct, originally gift from Claire-Waterman lab) and VAPA^ΔTM (addgene #13395) was cloned into pET28C vector. His-TrxA tagged FH2 was a gift from Marie-France Carlier lab ¹⁵ (Q1134-T1578, modified pGEX-6P-1 vector). His tagged importin α1^ΔIBB (addgene #26678) and importin α8 (addgene #26683) was cloned into pET28C vector. His tagged importin β⁷¹⁻⁸⁷⁶ was from Petr Kalab ²⁹. His tagged importin α1^ΔIBB variant with two I27 domains (parental plasmid was a gift from Dr. Yan Jie lab) was inserted after His tag was built from His tagged importin α1^ΔIBB. His-VAPA^MSP (aa 8–131) and His-VAPA^CD (aa 132–226) were truncated from His-VAPA^ΔTM. VAPA-mCherry and SLD-EGFP were cloned into pEGFP vector for somatic cell expression.

Proteins were produced in BL21(DE3) or BL21(DE3) Rosetta T1R. After transformation, single colonies were picked and cultured (250 rpm, 37 °C) in 20 mL LB with proper antibiotics overnight as seeds. On the second day, bacteria seeds were added into 1 l LB with antibiotics and cultured at 37°C until OD600 reached around 0.8. 1 ml 0.4 M IPTG (I2481C50, GoldenBio) was added to induce protein production overnight at 20°C. Cells were pelleted at 6,000 rpm, 4 °C for 15 min. Cell pellets were resuspended in 30 mL lysis buffer (10 mM Tris pH = 7.0, 250 mM NaCl, 0.5 mM EDTA, 1 mM DTT and 5% glycerol (v/v) for GST tagged proteins and 20 mM HEPES pH = 7.4, 500 mM NaCl and 20 mM Imidazole for His tagged proteins) and lysed by using French Press. The lysates were then ultracentrifuged at 40,000 rpm using Ti45 rotor at 4°C for 30 min followed with filtration by using 0.22 μm pore size filter. Then the clear lysates were loaded onto GST/His tag affinity columns (GE Healthcare) and gradient elution was performed with proper elution buffers (additional 20 mM reduced GSH or 0.5 M imidazole in lysis buffers for GST/His tagged proteins separately). Eluted fractions were checked by Coomassie blue staining and fractions of targeted proteins were collected and size-exclusion gel filtration was further performed with gel filtration buffers (25 mM Tris pH = 7.4, 500 mM NaCl, 4 mM DTT, and 0.5 mM EDTA if not announced). Specifically, His tagged VAPA^ΔTM used 150 mM (instead of 500 mM) NaCl and His-TrxA tagged FH2 used 20 mM Tris pH = 7.4, 75 mM KCl and 1 mM DTT for gel filtration. Eluted fractions were checked by Coomassie blue staining again and peak fractions from target proteins were pooled and concentrated (0.2–1 ml). 10% glycerol was added into the concentrated proteins and protein concentrations were measured by Nanodrop 2000/2000c (Thermo Scientific) at A280. For some proteins not suitable for A280 absorption measurement, Bradford assays were used instead. Proteins were aliquoted into small volumes and flash-frozen by liquid N₂ and then kept at −80°C.

For His-VAPA^MSP/CD purification, gel filtration was skipped and the eluted fractions were dialyzed in gel filtration buffer with 50% glycerol at 4°C overnight and then flash frozen and kept at −80°C.

GST pulldown assay, co-IP, and western blot

Glutathione magnetic agarose beads (78601, Thermo Scientific) were first washed with pulldown buffer (20 mM HEPES pH = 7.4, 300 mM NaCl, 1 mM DTT, 5 mM MgCl₂, 0.1% TritonX-100, 5% glycerol (v/v)) once and the tubes were decanted by using magnetic rack and vacuum aspiration. Pulldown buffer and 2 mg/ml BSA in water were added into the tubes in 1:1 (v:v) and GST-tagged proteins were added into the tubes with clean beads and immobilized on the beads by end-to-end nutating at 4°C for 1 hr. After immobilization, beads were washed with pulldown buffer 3 times, and 1:1 pulldown buffer with BSA was added again. Later, the prey proteins were added and incubated with beads for 1 hour at 4°C, end-to-end nutating as well. After incubation, beads were washed by pulldown buffer for 4 times, and proteins were eluted by adding 15 μl GST elution buffer on ice for 10 min. 5 μl 4 × sample buffers were added into eluted proteins and boiled at 100°C for 5 min and 10 μl of samples were used to run the gradient gel (4–15%) at 200 V, 25 min.

For co-IP, mCherry, EGFP, VAPA-mCherry, and SLD-EGFP were expressed in HEK 293T cells for 24 hr, lysed in 500 μL IP buffer (PBS, 0.5% TritonX-100, 1 mM DTT) and sonicated for 8 times. The suspension was cleared by tabletop centrifuge for 10 min at 4°C at full speed, 4 μL of the supernatant was kept for inputs and the rest were transferred to new tubes with RFP-Trap magnetic agarose beads (rtma-20, ChromoTek) added. The beads were incubated with the cell lysis for 1 hr at 4°C, end-to-end nutating. After incubation, beads were washed with IP buffer for 5 times, eluted by 2 × sample buffer then loaded into the gradient gel (4–15%) and ran at 200 V, 25 min. For both GST pulldown and co-IP, after running the gels, proteins were transferred onto nitrocellulose membrane (1704158, BioRad) by using Turbo Transfer system (1704150, BioRad). Membranes were first blocked by using Intercept® (PBS) Blocking Buffer (927–70001, LI-COR) for 30 min at room temperature. Primary antibodies, including rabbit anti-GST (A-5800, Thermo Scientific, 1:2000), mouse anti-His₆ (MA1–21315, Thermo Scientific, 1:2000), rabbit anti-RFP (600-401-379, Thermo Fisher, 1:1000) and mouse anti-GFP (11814460001, Sigma-Aldrich, 1:1000) were added in blocking buffer with 0.1% Tween20 and incubated with the membranes for 1 hour at room temperature or 4°C overnight. Membranes then were washed with PBST (PBS with 0.5% Tween20) 3 times per 5 min. Fluorescently labeled anti-rabbit (680) and anti-mouse (800) secondary antibodies (926–68021 and 926–32211, LI-COR) were incubated with membranes in 1: 15,000 in blocking buffer with 0.1% Tween20 and 0.1% SDS for 1 hour at room temperature. The membranes were lastly washed with PBST 4 times and imaged by Odyssey CLx imaging system (LI-COR) at 700/800 wavelength (Data S1).

FMN2 SLD and CLD pulldowns from mouse oocytes and MudPIT analysis

GST-tagged FMN2 truncates (GST-SLD and GST-CLD) expressed in bacteria were first loaded onto the affinity column and the column was washed with 3 column volume of cleaving buffer (50 mM Tris pH = 8.0, 150 mM NaCl, 1 mM EDTA, 1 mM DTT). After washing, 4% PreScission Protease (27084301, GE Healthcare) was added to the cleaving buffer and incubated at 4°C overnight. The flow-through was collected and dialyzed against 50 mM borate (pH = 8.5) to get rid of amine groups from the buffers used (e.g. Tris). The purified SLD and CLD were then crosslinked on NHS-activated magnetic bead (88826, Thermo Scientific) and quenched following the manufacturer’s instruction. The protein-coupled beads were then washed 3 times with PBS and kept in PBS at 4°C before use.

Around 2,000 mouse oocytes were lysed in lysis buffer (20 mM Tris-HCl pH = 7.5, 100 mM NaCl, 1 mM EDTA, 1 × protease inhibitor cocktail (G6521, Promega), 1 mM PMSF, 1 mM DTT and 5% Triton-X100) along with sonication. CLD/SLD-coupled beads were mixed with oocyte extracts and incubated for 30 min at 4°C. The beads were then washed 3 times using washing buffer (20 mM Tris-HCl pH = 7.5, 100 mM NaCl, 1 mM EDTA, 0.3% Triton-X100) and proteins bound were eluted by using high salt buffer (20 mM Tris-HCl pH = 7.5, 1 M NaCl, 1 mM EDTA).

The eluted proteins were TCA precipitated, urea-denatured, reduced, alkylated, and digested with endoproteinase Lys-C (Roche) and modified trypsin (Roche), and then submitted for MudPIT analysis⁵⁹. To identify ER membrane proteins, the gene list encoding ER membrane proteins was obtained from Mouse Genome Informatics (http://www.informatics.jax.org/).

Pyrene assay

10 μM actin with 5% pyrene-labeled (AP05-A, Cytoskeleton) was converted to Mg²⁺-ATP actin on ice for around 3 min. Proteins were added into G-buffer and mixed with Mg²⁺-ATP actin right before the addition of 10 × KME buffer (10 mM MgCl₂, 10 mM EGTA, and 0.5 M KCl), reaching a final volume at 120 μl (actin working/final concentration 1.5 μM). Reading by BioTek Cytation5 (Figures 6C, 6D, S5A-S5D, and S5M) or ThermoFisher Varioskan LUX (Figures 7B, 7C, and S5J-S5L) was initiated right after adding 10 × KME buffer with excitation at 365 and emission at 407 nm for 35 min, 15 sec per frame, at 25°C. The slopes from the first 60 s-120 s (BioTek machine) or 60 s-180 s (Thermofisher machine) with R² > 0.9 were used for calculating actin polymerization rate. In a few cases when fitting was not ideal, additional data points from 60 − 15 s or 120/180 + 15 s were included for better fitting.

TIRF assay

24 mm × 50 mm No 1.5 glass coverslips were first washed with 20% sulfuric acid overnight at room temperature and rinsed with a huge amount of sterile water. Then the coverslips were coated with 2 mg/ml methoxy-PEG-silane (MPEG-SIL-2000, Laysan Bio) and 2 μg/ml biotin-PEG-silane (Biotin-PEG-SIL-3400, Laysan Bio) in 80% ethanol (pH = 2.0) at 45 °C overnight. Afterward, coverslips were rinsed with sterile water thoroughly and dried by N₂ and kept at −80°C for longterm storage. Before use, coverslips were attached to chamber slides (80608, ibidi). The flow cells were first blocked by blocking buffer (20 mM HEPES pH = 7.5, 1 mM EDTA, 50 mM KCl, and 1% BSA) for 1 min and followed with 1 min 0.1 mg/ml streptavidin (N7021S, NEB) incubation in buffer with 20 mM HEPES (pH = 7.5), 1 mM EDTA, 50 mM KCl, and 10% glycerol (v/v). Lastly, the flow cells were washed with 1 × TIRF buffer (10 mM imidazole, 50 mM KCl, 1 mM MgCl2, 1 mM EGTA, 0.4 mM ATP, 40 mM DTT, 15 mM glucose, 100 μg/ml glucose oxidase (G7141, Sigma), and 0.5% 400 cP methylcellulose (M0262, Sigma)) for 4 times. Before starting, non-actin proteins were mixed first (protein mix since) and put on ice. Right after protein mixing, dark actin mixed with 20% ATTO488-actin (8153–04, Hypermol) and 0.5% biotin-actin (actin mix since) was converted to Mg²⁺-ATP form on ice for around 3 min. After actin conversion, protein mix was mixed with actin mix (reaction mix since), and the reaction mixture was added into 2 × TIRF buffer in 1:1 ratio and injected into the flow cell and image acquiring was started simultaneously (actin working concentration 0.5 μM). TIRF experiments were done on Nikon ECLIPSE Ti inverted microscope (Nikon Instruments) with an Apo TIRF 100 × / 1.49 lens (Nikon Instruments) and an EMCCD camera (Hamamatsu Photomics). During imaging, Perfect Focus System was on to keep the focal plane, and signals were captured every 15 seconds for 10 min. MetaMorph (Molecular Devices) was used to control the machines. Actin seed numbers at 0 and 90 seconds from the 150 μm × 150 μm field of view were manually counted in ImageJ.

QUANTIFICATION AND STATISTICAL ANALYSIS

Image analysis

For tracking spindle migration, chromosomes movement was tracked by using Imaris (Bitplane). First, for the position of chromosomes, the fluorescence images of chromosomes were applied with a Gaussian filter to remove noise using ImageJ. Next, we segmented the chromosomes by applying a threshold and tracked the chromosomes’ center of mass during spindle movement using Imaris. For averaging the chromosomes movements in different oocytes, we temporally aligned the different data sets to the time just before anaphase I occur. In oocytes, where the chromosomes did not segregate, the data sets were aligned to the average time point corresponding to the anaphase I occur in the same group. The distance of the spindle to the alignment position was calculated for each time point and plotted over time. To calculate average migration speed, only the displacements projected in the direction pointing from initial to the final location were used.

The fluorescence intensity profiles for FMN2, SLD, and F-actin were measured in ImageJ using a line width of 100 pixels crossing the mid-zone of the MI spindle in a 2D plane as shown in Figures S1B and S1C. To measure spindle periphery and outside spindle periphery fluorescence intensity, we used Brush Selection Tool in ImageJ with a selection width of 25 pixels. We manually selected the spindle periphery or outside spindle periphery as indicated in Figure S1B and measured the mean intensity of the selected area. Colocalization analysis was performed on dual-color confocal images that were captured from live oocytes. All processing and analysis were performed with Colocalization Finder plugins written for ImageJ, and Pearson’s correlation coefficient was used to calculate double fluorescence correlation coefficients.

Statistics

The statistical analysis was performed with GraphPad Prism software. Comparisons of data were performed by Student’s t-test (always two-tailed) or one-way analysis of variance (ANOVA) with Turkey’s multiple comparison test, P values < 0.05 was considered significant. All boxes plot defines the 25th and 75th percentiles, with a line at the median and error bars defining the 10th and 90th percentiles. Violin plot shows median (horizontal lines) and quartiles (dot lines).

Supplementary Material

3. Video S1. Representative video of spindle/chromosomes migration in control, Ran^T24N, RanGAP1, importin β⁷¹⁻⁸⁷⁶, and importin α1^ΔIBB expressing oocytes. Related to Figures 1 and 3.

z-projection of five sections, 5 μm apart; time interval: 15 min; movie length: 3 hr. White circle is chromosomes’ center of mass that was used for tracking. Magenta: YPet-H2B (to label chromosomes).

4. Video S2. Representative video showing VAPA Trim-Away prevented spindle/chromosomes migration. Related to Figures 5.

z-projection of five sections, 5 μm apart; time interval: 15 min; movie length: 3 hr. White circle is chromosomes’ center of mass that was used for tracking. Magenta: YPet-H2B (to label chromosomes).

5. Video S3. Representative time-lapse TIRF movies of actin polymerization. Related to Figure 7.

Time interval: 15 s; movie length: 90 s. Concentrations of proteins used: actin 500 nM, FH2 25 nM, SLD 0.8 μM, VAPA^ΔTM 0.4 μM, importin α1^ΔIBB 0.4 μM.

Key Resources Table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
mouse anti-VAPA	Santa Cruz Biotechnology	sc-293278
rabbit anti–VAPA	Sigma	HPA009174
Alexa Fluor 488–labeled anti-mouse	Invitrogen	A28175
Alexa Fluor 488–labeled anti-rabbit	Invitrogen	A27034
Alexa Fluor Plus 647–labeled anti-mouse	Invitrogen	A28181
GST Tag Polyclonal Antibody	Thermo Scientific	A-5800
6x-His Tag Monoclonal Antibody	Thermo Scientific	MA1-21315
Fluorescently labeled anti-rabbit	LI-COR	926-68021
Fluorescently labeled anti-mouse	LI-COR	926-32211
IgG control antibody	Santa Cruz	sc-2025
Chemicals, Peptides, and Recombinant Proteins
3-isobutyl-1-methylxanthine (IBMX)	Sigma-Aldrich	I7018
M16 medium	Sigma	M7292
Alexa Fluor 488–labeled phalloidin	Cell Signaling Technology	8878
Hoechst 33342	ThermoFisher	H3570
Dithiothreitol	GoldenBio	DTT25
NHS-dPEG4-biotin	Sigma	QBD10200
IPTG	GoldenBio	I2481C50
Intercept® (PBS) Blocking Buffer	LI-COR	927-70001
PreScission Protease	GE Healthcare	27084301
NHS-activated magnetic bead	Thermo Scientific	88826
Protease inhibitor cocktail	Promega	G6521
Endoproteinase Lys-C	Roche	11420429001
Pyrene-labeled actin	Cytoskeleton	AP05-A
Methoxy-PEG-silane	Laysan Bio	MPEG-SIL-2000
Biotin-PEG-silane	Laysan Bio	Biotin-PEG-SIL-3400
Chamber slides	ibidi	80608
Streptavidin	NEB	N7021S
Glucose oxidase	Sigma	G7141
400 cP methylcellulose	Sigma	M0262
ATTO488-actin	Hypermol	8153-04
Critical Commercial Assays
mMESSAGE mMACHINE T7 kit	ThermoFisher	AM1344
mMESSAGE mMACHINE T3kit	ThermoFisher	AM1348
RNeasy MinElute Cleanup Kit	QIAGEN	74204
Experimental Models: Organisms/Strains
CD1 mice	Charles River Laboratories or InVivos Pte Ltd	Strain Code 022
Recombinant DNA
YPet-H2B	27	N/A
H2B-mCherry	27	N/A
mCherry-Trim21	46	N/A
EYFP-NLS	This work	N/A
VAPA-mCherry	This work	N/A
mCherry	This work	N/A
SLD-mCherry	24	N/A
SLD-EGFP	24	N/A
SLDNLS mut-mCherry	This work	N/A
FMN2-AcGFP	24	N/A
RanT24N	29	N/A
RanGAP1	27	N/A
Importin α1ΔIBB	Addgene	26678
Importin β71−876	29	N/A
hAG-tag	MBL Life Science	AM-8011M
Ash-tag	MBL Life Science	AM-8011M
SLD-hAG	This work	N/A
VAPA-Ash	This work	N/A
His-VAPAΔ™	Addgene	13395
His-VAPAMSP	This work	N/A
His-VAPACD	This work	N/A
GST-SLD	This work	N/A
GST-CLD	This work	N/A
GST-VAPAΔ™	This work	N/A
His-FH2	This work	N/A
His-TrxA-FH2	15	N/A
His-importin α1ΔIBB	This work	N/A
His-importin α8	Addgene	26683
His-importin β71−876	29	N/A
His-2I27-importin α1ΔIBB	This work	N/A
Software and Algorithms
ImageJ	NIH	N/A
Imaris	Bitplane	N/A
Prism	GraphPad	N/A
ZEN	Zeiss	N/A
MetaMorph	Molecular Devices	N/A
Other
Zeiss LSM780 microscope	Zeiss	N/A
Zeiss LSM980 microscope	Zeiss	N/A
ECLIPSE Ti inverted microscope	Nikon	N/A
Turbo Transfer system	BioRad	1704150
Odyssey CLx imaging system	LI-COR	9140
Cytation5	BioTek	N/A
Varioskan LUX	ThermoFisher	VL0000D0
Amicon Ultra-0.5 100 KDa centrifugal filter	Millipore	UFC510008
Ti70 rotor	Beckman	337922
Ti45 rotor	Beckman	339160
HiLoad 16/600 Superdex75 column	GE Healthcare	GE28-9893-33
Dialysis tubing, 10 kDa cutoff	Thermo Scientific	68100
GSTrap™ HP Columns	GE Healthcare/Cytiva	17528101
HisTrap FF	GE Healthcare/Cytiva	17531901
Glutathione magnetic agarose beads	Thermo Scientific	78601
Nitrocellulose membrane	BioRad	1704158

Highlights.

1. FMN2 is an NLS-containing formin protein bound to importin α1/β complex.

2. Ran GTPase regulates FMN2 recruitment to spindle periphery during oocyte meiosis I.

3. Importin dissociation by Ran allows VAPA to anchor FMN2 to ER membrane.

4. VAPA binding relieves FMN2 autoinhibition to nucleate actin assembly.