Cell-cycle-resolved τSTED nanoscopy reveals nanoscale reorganization of TGN46-positive structures

Abstract

The Golgi apparatus is a highly organized membrane system whose structure is remodeled during cell-cycle progression, but nanoscale changes in individual Golgi subdomains during interphase remain incompletely understood. Here, we used Fucci-compatible fluorescence lifetime-based STED (τSTED) nanoscopy to examine cell-cycle-associated reorganization of cis-Golgi and trans-Golgi network (TGN)-associated structures in fixed HeLa cells. By combining Fucci-SA reporters with red/far-red Golgi labeling, large-field sequential imaging, Abberior STAR RED nanobody labeling, and reproducible field-wise lifetime-trajectory-based τSTED reconstruction, we correlated Fucci-defined cell-cycle state with nanoscale Golgi morphology in individual cells. Automated morphometry showed that GM130-positive cis-Golgi and TGN46-positive trans/TGN-associated distribution domains were relatively compact in G1 cells, whereas TGN46-positive structures exhibited an expanded spatial distribution in G2-enriched cells. Optimized τSTED imaging resolved this expansion as dispersed submicron TGN46-positive structures rather than uniform enlargement of the Golgi apparatus. These findings reveal cell-cycle-associated spatial reorganization of TGN46-positive trans/TGN-associated structures and demonstrate the utility of cell-cycle-resolved super-resolution imaging for analyzing organelle architecture.

Key words: Golgi apparatus, cell cycling, super-resolution microscopy, τSTED, Fucci

Graphical Abstract

Introduction

The Golgi apparatus is a highly organized and dynamic organelle composed of stacked membrane cisternae with distinct structural and functional polarity. In mammalian cells, individual Golgi stacks are laterally connected to form the Golgi ribbon, in which cis-, medial-, trans-Golgi, and trans-Golgi network (TGN) domains coordinate protein and lipid modification, sorting, and trafficking (Makhoul et al., 2019; Tie et al., 2018). During cell-cycle progression, the Golgi apparatus undergoes regulated structural remodeling. In particular, during the G2/M transition, the Golgi ribbon is unlinked and subsequently fragmented to ensure proper inheritance by daughter cells (Colanzi et al., 2007; Lucocq et al., 1987, 1989; Misteli and Warren, 1994; Shima et al., 1998; Sütterlin et al., 2002; Tang and Wang, 2013; Wei and Seemann, 2017). Although these global changes in Golgi organization have been extensively studied, how specific sub-Golgi domains, especially cis-Golgi and trans/TGN-associated structures, are reorganized at the nanoscale during defined interphase stages remains less well understood (Makhoul et al., 2019; Mascanzoni et al., 2024; Tie et al., 2018).

A major obstacle has been the difficulty of correlating cell-cycle state with nanoscale organelle morphology in the same cell. Electron microscopy provides ultrastructural resolution and has been fundamental for defining mitotic Golgi fragments (Lucocq et al., 1987, 1989), but it is not readily compatible with fluorescence-based cell-cycle classification of many individual cells within the same specimen. Conversely, Fucci-based imaging enables single-cell identification of cell-cycle state (Sakaue-Sawano and Miyawaki, 2014), but conventional confocal microscopy cannot resolve Golgi structures below the diffraction limit. By contrast, τSTED nanoscopy can overcome this resolution barrier (Hell and Wichmann, 1994). Time-gated STED is a distinct implementation that uses temporal detection windows to improve effective resolution and contrast (Vicidomini et al., 2011); it should not be conflated with τSTED, which uses fluorescence-lifetime information in image reconstruction. However, long-duration, large-area, multicolor STED imaging remains technically demanding because image quality depends strongly on fluorophore photophysics, labeling geometry, depletion efficiency, photobleaching, and mechanical stability during acquisition.

Recent image-based analyses have further shown that subcellular organization itself contains information about cell-cycle state. Nagao et al. used convolutional neural network-based classification of conventional fluorescence images and showed that nuclear, GM130-positive Golgi, and microtubule-associated features can contribute to discrimination between G1/S and G2 phases (Nagao et al., 2020). This work highlighted the value of cell morphology as an informative readout of cell-cycle state. The present study builds on this concept from the complementary direction: once cell-cycle state is defined independently by Fucci reporters, we ask how specific Golgi subdomains are reorganized at super-resolution scale. This framing is particularly relevant for understanding the relationship between organelle structure and function during interphase cell-cycle progression.

Fluorescence lifetime-based STED nanoscopy, or τSTED, offers a strategy to improve nanoscale contrast by incorporating fluorescence lifetime information into STED imaging. In τSTED, phasor analysis of fluorescence-lifetime data can be used to identify fluorophore-associated components, validate their spatial correspondence, and define a lifetime trajectory for τSTED reconstruction (Alvarez et al., 2021; Digman et al., 2008). However, τSTED performance is highly fluorophore-dependent. Therefore, successful application of τSTED to densely packed organelles such as the Golgi apparatus requires not only spectral compatibility with cell-cycle reporters, but also careful optimization of fluorophores, labeling probes, phasor-component identification and lifetime-trajectory definition, and image acquisition conditions (Alvarez et al., 2021; Digman et al., 2008; Pleiner et al., 2015; Sograte-Idrissi et al., 2020).

Here, we examined whether cis-Golgi and trans/TGN-associated structures show distinguishable nanoscale organization during Fucci-defined cell-cycle progression in fixed HeLa cells. To address this question, we established a Fucci-compatible τSTED workflow with field-wise lifetime-trajectory-based reconstruction in which Golgi markers were first imaged in red and far-red channels, followed by confocal acquisition of Fucci fluorescence in the same cells. We used GM130 as a cis-Golgi marker, based on its established localization as a cis-Golgi matrix protein (Nakamura et al., 1995), and TGN46 as a trans/TGN-associated marker, based on the well-characterized TGN localization and trafficking behavior of the TGN38/TGN46 family (Humphrey et al., 1993; Lujan et al., 2024; Luzio et al., 1990; Ponnambalam et al., 1996; Prescott et al., 1997; Reaves et al., 1993). Using this optimized workflow, we show that TGN46-positive trans/TGN-associated structures display an expanded nanoscale distribution in G2-enriched cells, appearing as dispersed submicron structures rather than as simple enlargement of the Golgi apparatus.

Results

Cell-cycle-resolved τSTED imaging enables nanoscale analysis of Golgi and TGN organization

To correlate cell-cycle state with nanoscale Golgi morphology in the same fixed cells, we first designed a spectrally compatible imaging workflow using Fucci-SA HeLa cells and red/far-red τSTED dyes. The Fucci-SA cells used in this study express Kusabira Orange 2 (KO2) and Azami Green 1 (AG1) as cell-cycle reporters, following the principle of fluorescent ubiquitination-based cell-cycle indicators (Sakaue-Sawano and Miyawaki, 2014). For τSTED imaging of Golgi markers, we selected Alexa Fluor Plus 594 for the red channel and Abberior STAR RED for the far-red channel. Comparison of the emission spectra showed that the Fucci fluorophores were sufficiently separated from the red and far-red τSTED channels under our detection settings, allowing Fucci-based cell-cycle classification and τSTED imaging of Golgi markers to be performed in the same fixed specimens with minimal spectral crosstalk (Fig. 1A).

Fig. 1

Fucci-compatible τSTED workflow for cell-cycle-resolved Golgi nanoscopy

(A) Spectral compatibility of Fucci-SA reporters and the red/far-red imaging channels used for Golgi-marker imaging. (B) Sequential imaging workflow. Golgi markers were first acquired in τSTED mode using red/far-red excitation and a 775 nm STED depletion laser, followed by confocal acquisition of Fucci fluorescence from the same field after switching off the STED laser. (C) Representative whole-field phasor plot showing the Golgi-marker-associated fluorescence-lifetime component and the lifetime trajectory defined along this component. The trajectory assigns higher intensity to the longer-lifetime side and lower intensity to the shorter-lifetime side. (D) Representative validation of the initial phasor ROI by overlaying its corresponding mask on the τSTED image. The validated component and lifetime trajectory were used for field-wide τSTED reconstruction and were not optimized for individual cells. (E) Representative merged field showing Golgi-marker τSTED signals and Fucci fluorescence used for cell-cycle classification. Spectral colors in the optical scheme indicate wavelength ranges, whereas microscopy images are displayed in pseudocolors assigned for channel visibility and matched across image panels.

We then established a sequential imaging workflow. Golgi markers labeled with Alexa Fluor Plus 594 and Abberior STAR RED were first imaged in τSTED mode using 594 nm and 633 nm excitation together with a 775 nm STED depletion laser. After τSTED acquisition, the STED laser was switched off, and Fucci fluorescence was acquired in confocal mode using 488 nm and 561 nm excitation in the same field of view (Fig. 1B). This acquisition order enabled high-resolution imaging of Golgi-marker-positive structures before retrospective cell-cycle assignment by Fucci fluorescence in the same fixed cells.

Large-field τSTED imaging enables analysis of multiple Fucci-defined cells under matched optical conditions

To analyze multiple cells under identical imaging conditions, we acquired large-field τSTED images of 8192 × 8192 pixels with a pixel size of 15 nm. Each field of view contained approximately 15–25 cells, allowing cells at different Fucci-defined cell-cycle states to be imaged under the same optical and image-processing conditions. Large-area scans typically required more than 1 h and were performed using a microscope installed in a highly vibration-isolated environment to support stable long-duration acquisition.

Representative large-field images are shown in Fig. 2. Two reciprocal labeling configurations were used (Supplementary Table S1): GM130 labeled with Abberior STAR RED nanobody together with TGN46 labeled with Alexa Fluor Plus 594 IgG (Fig. 2A), and GM130 labeled with Alexa Fluor Plus 594 IgG together with TGN46 labeled with Abberior STAR RED nanobody (Fig. 2B). In both configurations, τSTED images of Golgi markers and confocal Fucci images were acquired from the same field of view. Cells showing red-only Fucci fluorescence and cells showing green-only Fucci fluorescence could be identified within the same large-field images and subjected to subsequent Golgi morphometric analysis.

Fig. 2

Large-field τSTED imaging of Fucci-defined G1 and G2-enriched cells

(A, B) Representative 8192 × 8192 pixel fields acquired under reciprocal labeling configurations. In (A), GM130 was labeled with Abberior STAR RED nanobody and TGN46 with Alexa Fluor Plus 594 IgG. In (B), GM130 was labeled with Alexa Fluor Plus 594 IgG and TGN46 with Abberior STAR RED nanobody. Golgi-marker images and Fucci fluorescence were acquired sequentially from the same field of view. Boxed regions indicate representative G1 and G2-enriched cells shown at higher magnification. Scale bars are indicated in the figure.

Quantitative analysis was not based on a single large image. For each reciprocal labeling condition, two independently prepared coverslips were examined, and four large-field τSTED images were acquired per condition (eight fields in total). Depending on the field, 1–10 eligible cells contributed to the analysis. To avoid cells imaged near the upper or lower surface of the specimen, analysis was restricted to cells in which the nucleus was clearly visible at the selected focal plane. Because the Fucci-SA system used here does not sharply resolve S phase, cells with red-only Fucci fluorescence were classified as G1-phase cells, whereas cells with green-only Fucci fluorescence were operationally treated as a G2-enriched population. All cells meeting these criteria within the analyzed fields were included; cells with mixed, weak, or ambiguous Fucci signals were excluded.

Phasor-based lifetime selection in far-red nanobody τSTED reproducibly resolves Golgi-associated nanoscale structures

We next optimized τSTED processing for Golgi nanoscopy. Phasor analysis provides a fit-free graphical representation of fluorescence-lifetime components and can distinguish fluorophore-derived signals from background components (Digman et al., 2008). Because weak signals may not produce a clearly defined distribution in the phasor plot, large fields were scanned slowly at low excitation intensity to improve the signal-to-noise ratio. For each 8192 × 8192 pixel field, a phasor plot was generated from the photon-arrival-time information acquired across the entire field. The Abberior STAR RED-derived component appeared at a similar position across independent fields, and a lifetime trajectory was defined along this component (Fig. 1C). An initial phasor region of interest (ROI) was used to identify the fluorophore-associated component and was validated by overlaying the corresponding mask on the τSTED image (Fig. 1D). Following this validation, the trajectory was configured to assign higher intensity to the longer-lifetime side and lower intensity to the shorter-lifetime side, and the τSTED lifetime image was generated accordingly.

In the Abberior STAR RED channel, this procedure reproducibly identified a Golgi-associated phasor component at a similar position and with a similar distribution pattern across independent fields (Fig. 1C). FLIM-mode reanalysis of representative fields with the highest signal-to-noise ratios indicated lifetime ranges of 2.097–2.370 ns for GM130 and 1.971–2.248 ns for TGN46 labeled with Abberior STAR RED nanobodies. The trajectory endpoints varied by several percent among fields. These values are presented as reference ranges because τSTED reconstruction and conventional FLIM lifetime estimation are handled differently in the software. Reconstruction based on the defined lifetime trajectory extracted Golgi-marker-associated fluorescence from all cells within each field, indicating that the trajectory represented a field-wide Golgi-associated fluorescence component rather than a locally optimized subset (Fig. 1D). This observation is consistent with the principle that τSTED uses fluorescence-lifetime information to improve STED image contrast and suppress uncorrelated background signals (Alvarez et al., 2021).

In contrast, Alexa Fluor Plus 594 was sufficiently bright for conventional STED imaging but did not produce a stable and reproducible Golgi-associated phasor component under our τSTED conditions. Alexa Fluor Plus 594 was therefore used primarily for conventional time-gated STED imaging to visualize two-color localization and compare marker distributions, whereas nanoscale texture analysis was restricted mainly to the Abberior STAR RED channel.

To further improve nanoscale localization, we used nanobody-based labeling in the Abberior STAR RED channel. Because conventional IgG-based labeling introduces a relatively large distance between the target antigen and fluorophore, linkage error can affect the apparent morphology of structures near the 30-nm scale. Previous studies have shown that nanobody-based labeling can reduce linkage error and improve super-resolution localization accuracy (Pleiner et al., 2015; Sograte-Idrissi et al., 2020). Therefore, the optimized far-red imaging condition used for nanoscale interpretation combined Abberior STAR RED with nanobody-based labeling.

Automated morphometry of GM130-positive and TGN46-positive distribution domains

We next asked whether cis-Golgi-associated and trans/TGN-associated structures show different cell-cycle-associated spatial behaviors. GM130 was used as a marker for cis-Golgi-associated structures, based on its characterization as a cis-Golgi matrix protein (Nakamura et al., 1995), whereas TGN46 was used to label trans/TGN-associated structures, based on the established TGN localization and trafficking behavior of the TGN38/TGN46 family (Humphrey et al., 1993; Lujan et al., 2024; Luzio et al., 1990; Ponnambalam et al., 1996; Prescott et al., 1997; Reaves et al., 1993). Because the Golgi apparatus has a distinct cis-to-trans polarity, we analyzed the spatial distributions of GM130-positive and TGN46-positive domains independently in Fucci-defined G1-phase and G2-enriched cells.

To quantify the spatial extent of Golgi-marker-positive structures, we developed an automated ImageJ/Fiji macro (Supplementary Method S1). Briefly, each τSTED image was processed by rolling-ball background subtraction with a radius of 50 pixels (750 nm) to reduce diffuse background, nonspecific fluorescence, and isolated punctate signals. The image was then smoothed using a Gaussian filter with σ = 2 pixels to suppress pixel-scale noise and prevent small punctate signals from being detected as independent Golgi domains. After automatic thresholding and binary mask generation, the boundary of the marker-positive distribution domain was detected, and its major-axis length was measured automatically.

This value should not be interpreted as the absolute size, area, volume, or membrane content of the Golgi apparatus. Rather, it represents an index of the spatial extent over which GM130-positive or TGN46-positive structures are distributed within each cell. We therefore refer to this measurement as the major-axis length of the Golgi-marker-positive distribution domain.

TGN46-positive trans/TGN-associated structures show expanded spatial distribution in G2-enriched cells

In Fucci-defined G1-phase cells, GM130-positive cis-Golgi structures and TGN46-positive trans/TGN-associated structures occupied relatively compact spatial domains. Automated quantification showed no clear difference in major-axis length between the GM130-positive and TGN46-positive distribution domains in G1-phase cells, indicating that both marker-positive domains were distributed within comparable spatial ranges under these conditions.

We then compared these distributions in G2-enriched cells. Under optimized Abberior STAR RED nanobody τSTED conditions, the major-axis length of the TGN46-positive distribution domain was significantly increased in G2-enriched cells compared with G1-phase cells (Fig. 3). By contrast, expansion of GM130-positive cis-Golgi structures was less pronounced under the same analysis conditions. The cell-cycle phase × Golgi subdomain interaction showed a strong trend but did not reach the conventional threshold for statistical significance (F (1,112) = 3.75, P = 0.0555). Together, these results suggest that TGN46-positive trans/TGN-associated structures may undergo more pronounced cell-cycle-associated spatial reorganization than GM130-positive cis-Golgi structures, although the difference between the two subdomains was not statistically conclusive.

Fig. 3

Quantification of GM130-positive and TGN46-positive distribution domains

Major-axis length of the marker-positive distribution domain was measured after background subtraction, smoothing, thresholding, and binary-mask generation. Half-violin plots show the distribution of cell-level measurements, with individual data points and mean + SD. (A) Quantification for the GM130-Abberior STAR RED nanobody/TGN46-Alexa Fluor Plus 594 IgG configuration. (B) Quantification for the GM130-Alexa Fluor Plus 594 IgG/TGN46-Abberior STAR RED nanobody configuration. The TGN46-positive distribution domain measured under the optimized Abberior STAR RED nanobody τSTED condition was expanded in G2-enriched cells. P values were calculated using unpaired two-tailed Student’s t-tests. n represents the number of individual cells analyzed. ns, not significant. (C, D) Representative high-magnification τSTED images of TGN46-positive structures in Fucci-defined G1-phase (C) and G2-enriched (D) cells. Detected-object boundaries are color-coded by equivalent diameter: cyan, small (≤100 nm); magenta, intermediate (>100 to ≤500 nm); and red, large (>500 nm). (E) Total area of TGN46-positive structures per cell. (F) Total number of TGN46-positive structures per cell. (G) Number of TGN46-positive structures classified by equivalent diameter as small (≤100 nm), intermediate (>100 to ≤500 nm), or large (>500 nm). Only objects comprising at least 3 pixels were included. Individual points represent single cells; horizontal bars and error bars indicate mean ± SEM. G1, n = 29 cells; G2-enriched, n = 8 cells. Statistical significance was assessed using two-tailed Mann–Whitney U tests. Y-axes are logarithmic. ns, not significant; *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.

The G2-associated expansion of the TGN46-positive distribution domain was detected under the optimized Abberior STAR RED nanobody condition but was not detected as a significant increase in the Alexa Fluor Plus 594 IgG channel (Fig. 3). This result should not be interpreted as evidence that TGN46-positive redistribution is absent in the Alexa Fluor Plus 594 conditions. Rather, it suggests that detection of small dispersed TGN46-positive structures depends on the optimized far-red τSTED workflow, including Abberior STAR RED labeling, nanobody-based detection, and reproducible field-wise lifetime-trajectory-based τSTED reconstruction.

Three-way ANOVA supports imaging-condition-dependent detection of cell-cycle-associated Golgi/TGN expansion

To evaluate whether the optimized dye-based τSTED condition affected detection of Golgi-marker-positive distribution domains and their cell-cycle-associated changes, we performed a three-way ANOVA using cell-cycle phase, Golgi subdomain, and imaging/dye condition as factors (Fig. 4). The analysis showed significant main effects of cell-cycle phase (F (1,112) = 51.66, P = 7.83 × 10^–11), Golgi subdomain (F (1,112) = 4.26, P = 0.041), and imaging/dye condition (F (1,112) = 6.57, P = 0.0117) on major-axis length. These main effects indicate overall differences after averaging across the other factors and therefore do not, by themselves, identify the condition in which the cell-cycle-associated change occurred. The interaction between cell-cycle phase and imaging/dye condition was significant (F (1,112) = 4.20, P = 0.0427), demonstrating that the magnitude of the G1-to-G2-enriched change depended on the imaging condition. Consistent with the condition-stratified comparisons in Fig. 3, the increase was most clearly detected under the optimized Abberior STAR RED nanobody τSTED condition. The cell-cycle phase × Golgi subdomain interaction was not significant (F (1,112) = 3.75, P = 0.055), whereas the Golgi subdomain × imaging/dye condition interaction (F (1,112) = 1.00, P = 0.320) and the three-way interaction (F (1,112) = 1.61, P = 0.208) were not significant. The significant cell-cycle phase × imaging/dye condition interaction indicates that the nanoscale manifestation of this remodeling is detected more effectively under the optimized imaging condition. Thus, the analysis supports imaging-condition-dependent detection. The non-significant cell-cycle phase × Golgi subdomain interaction did not establish that TGN46-positive structures undergo more pronounced cell-cycle-associated reorganization than GM130-positive structures.

Fig. 4

Three-way ANOVA of major-axis length measurements under different imaging conditions

(A) Experimental design for the three-way ANOVA. Major-axis length was analyzed using cell-cycle phase, Golgi subdomain, and imaging/dye condition as fixed factors. (B) ANOVA summary table. Significant main effects were detected for all three factors; these effects describe overall differences averaged across the other factors. (C) The cell-cycle phase × imaging/dye condition interaction was significant (F (1,112) = 4.20, P = 0.0427), demonstrating that the magnitude of the G1-to-G2-enriched change depended on the imaging condition. The cell-cycle phase × Golgi subdomain interaction showed a strong trend but did not reach statistical significance (F (1,112) = 3.75, P = 0.0555), suggesting a possible difference in cell-cycle-associated reorganization between the GM130- and TGN46-positive domains. The Golgi subdomain × imaging/dye condition interaction (F (1,112) = 1.00, P = 0.320) and the three-way interaction (F (1,112) = 1.61, P = 0.208) were not significant. n represents individual cells; the nesting of cells within fields was not modeled in this fixed-factor ANOVA. ns, not significant; *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.

τSTED resolves the expanded TGN46-positive domain as spatially separated submicron elements

High-resolution inspection of Abberior STAR RED nanobody τSTED images revealed that the increased major-axis length of the TGN46-positive distribution domain in G2-enriched cells did not reflect uniform enlargement of the Golgi apparatus. Instead, TGN46-positive structures appeared as multiple discrete submicron elements distributed over a broader cytoplasmic area. An example corresponding to a high-value measurement in Fig. 3B showed spatially separated TGN46-positive elements that expanded the measured distribution domain (Fig. S1). These peripheral elements were most clearly detected under the optimized τSTED condition. Because TGN38/TGN46-family proteins are known to cycle between the TGN, the cell surface, and endosomal/recycling pathways (Lujan et al., 2024; Ponnambalam et al., 1994; Prescott et al., 1997; Reaves et al., 1993), these spatially separated elements were interpreted conservatively as TGN46-positive trans/TGN-associated structures rather than as a single continuous TGN membrane domain.

Representative high-magnification images suggested a morphological shift from relatively compact TGN46-positive elements in G1-phase cells to more numerous and spatially dispersed elements in G2-enriched cells (Fig. 3C and D). To visualize the size classification directly in these images, object boundaries were overlaid in cyan for small structures (≤100 nm), magenta for intermediate structures (>100 to ≤500 nm), and red for large structures (>500 nm). To quantify the morphological basis of this difference, we measured the total TGN46-positive area and the number of discrete TGN46-positive structures per cell. Both measurements were significantly increased in G2-enriched cells compared with G1-phase cells (Fig. 3E and F). We next classified individual structures by equivalent diameter as small (≤100 nm), intermediate (>100 to ≤500 nm), or large (>500 nm). The number of TGN46-positive structures increased significantly in all three size classes in G2-enriched cells (Fig. 3G), indicating that the increase was distributed across a broad range of object sizes rather than being attributable only to enlargement of pre-existing large structures.

Parallel analysis of GM130-positive structures showed a significant increase in total marker-positive area but no significant change in the total number of structures (Fig. S2A and B). Size-stratified analysis revealed an increase only in large GM130-positive structures (>500 nm), with no significant increase in the small or intermediate classes (Fig. S2C). Thus, the G2-associated increase in small and intermediate TGN46-positive structures was not mirrored by the GM130-positive cis-Golgi compartment and represents a morphological feature preferentially associated with the TGN46-positive trans/TGN domain.

In this study, we refer to these structures as TGN46-positive fragments or TGN46-positive trans/TGN-associated structures. We do not assign them definitively to fragmented TGN membranes, TGN-derived transport carriers, or endosomal recycling intermediates. Their precise structural identity will require further analysis using additional markers and complementary ultrastructural approaches. Nevertheless, these observations support the conclusion that TGN46-positive trans/TGN-associated structures become spatially dispersed in G2-enriched cells.

Together, these results show that TGN46-positive trans/TGN-associated structures undergo cell-cycle-associated nanoscale reorganization, with detectable expansion of their spatial distribution in G2-enriched cells and resolution of this phenotype as dispersed submicron structures by the optimized Fucci-compatible τSTED workflow. This approach provides a foundation for future quantitative analyses of organelle nanoarchitecture across defined cell-cycle states.

Discussion

In this study, we used Fucci-compatible τSTED nanoscopy to visualize cell-cycle-associated nanoscale reorganization of Golgi subdomains in individual fixed HeLa cells. The main structural observation was an expanded and dispersed spatial organization of TGN46-positive trans/TGN-associated structures in G2-enriched cells under the optimized imaging condition, whereas the corresponding change in GM130-positive cis-Golgi structures was less pronounced. Together with the near-significant cell-cycle phase × Golgi subdomain interaction, this pattern suggests preferential reorganization of TGN46-positive structures. The correlative workflow combined Fucci-based cell-cycle identification with nanoscale imaging of Golgi markers, allowing Golgi and TGN-associated morphology to be compared between Fucci-defined G1-phase and G2-enriched cells within asynchronous populations.

A key methodological conclusion from this study is that detecting cell-cycle-associated changes in densely packed Golgi/TGN structures requires integrated optimization of fluorophores, labeling probes, phasor analysis, lifetime-trajectory definition, and acquisition stability, rather than simply the use of STED hardware. In the Abberior STAR RED channel, phasor plots generated separately from entire 8192 × 8192 pixel fields showed a reproducible Golgi-associated component at a similar position across independent fields. Because weak signals may not produce a clearly defined phasor distribution, the large fields were scanned slowly at low excitation intensity to improve the signal-to-noise ratio. An initial phasor ROI and its corresponding image mask were used to verify the fluorophore-associated component, after which a lifetime trajectory was defined along that component. The trajectory assigned higher intensity to the longer-lifetime side and lower intensity to the shorter-lifetime side. The similar phasor positions and reference lifetime ranges across fields supported the reproducibility of this procedure, while reconstruction of Golgi-marker fluorescence across all cells in each field reduced the risk of subjective cell-specific optimization (Alvarez et al., 2021; Digman et al., 2008).

In contrast, Alexa Fluor Plus 594 provided sufficient fluorescence intensity for STED imaging but did not yield a stable and reproducible Golgi-associated phasor component under our τSTED conditions. Therefore, the Alexa Fluor Plus 594 channel was useful for two-color localization and distribution-level comparison, but nanoscale texture interpretation was primarily restricted to the Abberior STAR RED channel. This distinction should not be interpreted as a general limitation of Alexa Fluor Plus 594 as a STED dye. Rather, it indicates that, under the present imaging and lifetime-trajectory conditions, Abberior STAR RED was better suited for extracting nanoscale Golgi-associated signals. In addition, nanobody-based labeling was important for resolving small Golgi-marker-positive structures. At an effective resolution close to 30 nm, the linkage error introduced by conventional IgG-based labeling can substantially affect apparent morphology. By reducing the distance between the target epitope and the fluorophore, nanobody labeling likely improved the detectability of small TGN46-positive structures (Pleiner et al., 2015; Sograte-Idrissi et al., 2020).

Using this optimized workflow, we compared GM130-positive cis-Golgi structures and TGN46-positive trans/TGN-associated structures in Fucci-defined G1 cells and G2-enriched cells. In G1-phase cells, GM130-positive and TGN46-positive distribution domains occupied similarly compact spatial regions. In contrast, the condition-stratified comparison showed an expansion of the TGN46-positive distribution domain in G2-enriched cells under Abberior STAR RED nanobody τSTED conditions. This increase in major-axis length should not be interpreted as uniform enlargement of the entire Golgi apparatus. Instead, high-resolution inspection indicated that TGN46-positive structures were dispersed as submicron fragments over a broader cytoplasmic area. This observation is consistent with pre-mitotic Golgi reorganization before the G2/M transition, including Golgi ribbon unlinking and mitotic Golgi fragmentation described in previous studies (Colanzi et al., 2007; Lowe et al., 1998; Lucocq et al., 1987, 1989; Mascanzoni et al., 2024; Misteli and Warren, 1994; Shima et al., 1998; Sütterlin et al., 2002, 2001; Tang and Wang, 2013; Wei and Seemann, 2017). The preferential expansion observed for the TGN46-positive domain, together with the near-significant cell-cycle phase × Golgi subdomain interaction, suggests that pre-mitotic remodeling may be spatially non-uniform across Golgi polarity and may be manifested more prominently in trans/TGN-associated membrane organization than in GM130-positive cis-Golgi organization. This possibility is biologically relevant because it implies that the sorting and trafficking region of the Golgi may undergo detectable spatial redistribution before generalized mitotic fragmentation. However, the subdomain difference did not reach the conventional threshold for statistical significance and should therefore be interpreted as suggestive rather than definitive. The object-based analysis further refined this interpretation. TGN46-positive structures increased in number across all three size classes, whereas the increase in GM130-positive area occurred without a significant increase in total object number and was confined to the large-object class. The increase in TGN46-positive structures across all size classes, together with the TGN46-specific increase in small and intermediate structures, supports remodeling of the trans/TGN-associated membrane system rather than uniform enlargement of Golgi structures.

Several limitations should be considered when interpreting these morphometric data. First, the major-axis length measured in this study is not a direct measurement of Golgi volume, membrane content, or absolute TGN size. It is an index of the spatial extent over which thresholded Golgi-marker-positive structures are distributed within each cell. Second, TGN46 is enriched in the trans-Golgi network but is also known to cycle between the TGN, the cell surface, and endosomal/recycling pathways (Lujan et al., 2024; Luzio et al., 1990; Ponnambalam et al., 1996, 1994; Prescott et al., 1997; Reaves et al., 1993). This prior work provides a plausible context for spatially separated TGN46-positive elements, while not by itself defining their precise identity. Therefore, the TGN46-positive fragments observed in G2-enriched cells cannot be assigned definitively to fragmented TGN membranes, TGN-derived transport carriers, or endosomal recycling intermediates based on the present data alone. Additional markers and complementary ultrastructural approaches will be required to define the precise identity of these structures. Third, because the Fucci-SA system used here does not sharply resolve S phase, our analysis was restricted to cells with unambiguous red-only or green-only Fucci signals, operationally classified as G1-phase and G2-enriched cells, respectively. Fourth, cells were nested within large-field images and therefore shared staining, acquisition, τSTED reconstruction, and processing conditions. To reduce the influence of any single staining or acquisition instance, each labeling condition was sampled across four independently acquired large fields from two separately prepared coverslips. Thus, each comparison incorporated cells exposed to multiple field-level technical conditions rather than repeated measurements from a single image. This sampling design reduces dependence on any one field and broadens the field-to-field variation represented in the dataset. Nevertheless, field was not included as a random effect and the number of eligible cells varied among fields; consequently, the cell-level analyses cannot formally partition cell-to-cell from field-to-field variation, and the statistical inference should be interpreted accordingly. Finally, the present study examined only HeLa/Fucci cells and did not include a medial-Golgi marker. Whether the observed behavior extends to medial cisternae or other cell types remains to be determined.

Although the optimized Abberior STAR RED nanobody τSTED images revealed nanoscale texture and local signal depressions within Golgi-marker-positive structures, we did not assign these features to bona fide Golgi fenestrations in the present study. Previous super-resolution studies have shown that the Golgi contains nanometer-scale spatial organization and that τSTED microscopy can resolve nanoscale Golgi-associated domains (Frye et al., 2020; Tie et al., 2018; Wong-Dilworth et al., 2023). However, rigorous quantitative interpretation of fenestration-like features will require dedicated object-detection algorithms, validation against the effective point spread function and labeling geometry, and complementary approaches such as correlative light and electron microscopy. The workflow established here provides a basis for such future analyses. More broadly, this cell-cycle-resolved τSTED platform should be applicable not only to Golgi architecture but also to other dynamic cellular structures, including the endoplasmic reticulum, mitochondria, microtubules, and actin cytoskeleton. Thus, this study provides evidence for cell-cycle-associated nanoscale reorganization of TGN46-positive trans/TGN-associated structures and establishes a methodological foundation for linking organelle structure to cell-cycle state by super-resolution nanomorphometry.

Materials and Methods

HeLa/Fucci cell culture

HeLa/Fucci cells (RCB2812; RIKEN BioResource Research Center Cell Bank, Tsukuba, Japan) were used in this study. Cells were cultured in high-glucose Dulbecco’s modified Eagle’s medium (DMEM; 048-30275; FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) supplemented with 10% fetal bovine serum (FBS; 10270-106; Gibco/Thermo Fisher Scientific, Waltham, MA, USA, or 35-015-CV; Corning, Corning, NY, USA), MEM non-essential amino acids (11140-050; Gibco/Thermo Fisher Scientific), and penicillin-streptomycin (15140-122; Gibco/Thermo Fisher Scientific). Cells were maintained at 37°C in a humidified incubator with 5% CO₂.

Immunofluorescence staining for τSTED microscopy

Cells were fixed and immunostained as follows. The antibody labeling configurations used for each imaging condition are summarized in Supplementary Table S1. Briefly, 10% formalin was prepared by diluting formalin (16223-55; Nacalai Tesque, Kyoto, Japan) in cytoskeleton-stabilizing buffer (CSB; 137 mM NaCl, 5 mM KCl, 1.1 mM Na₂HPO₄, 0.4 mM KH₂PO₄, 4 mM NaHCO₃, 2 mM MgCl₂, 5.5 mM glucose, 2 mM EGTA, and 5 mM PIPES [pH 6.1]). Cells were cultured on coverslips (No. 1S HT; Matsunami Glass, Osaka, Japan) for 2 days and fixed with 10% formalin supplemented with 0.05% glutaraldehyde (071-02031; FUJIFILM Wako Pure Chemical Corporation) for 10 min at room temperature.

All subsequent steps were performed at room temperature unless otherwise stated. Cells were rinsed with CSB and permeabilized with 0.5% Triton X-100 in CSB for 90 s. After rinsing, cells were incubated with 1.5 mg/mL sodium borohydride (NaBH₄) in CSB for 20 min to quench glutaraldehyde-derived autofluorescence and then blocked with 3% bovine serum albumin (BSA) in CSB for 30 min. Cells were incubated with primary antibody (1:100 dilution) solution, followed by fluorescently labeled secondary antibody (1:200 dilution) or nanobody solution (1 μg/mL) in the dark. Details of the antibodies used are provided in Supplementary Table S2. After washing with CSB, coverslips were mounted on glass slides with ProLong Glass Antifade Mountant (P36982; Thermo Fisher Scientific). Abberior STAR RED and Alexa Fluor Plus 594 were used as fluorophores for double staining. The labeling configurations and antibody information are summarized in Supplementary Tables S1 and S2, respectively.

Microscopic observation

STED and confocal microscopy were performed using a TCS SP8 STED 3X microscope (Leica Microsystems, Wetzlar, Germany) equipped with a 775-nm depletion laser for τSTED imaging and a white-light laser (470–670 nm) for excitation. Before quantitative acquisition, the microscope was allowed to equilibrate for several hours until thermal drift had stabilized. Imaging was then conducted over 2–4 consecutive days with the microscope maintained in an equilibrated state to minimize drift during nanometer-scale measurements. The excitation and detection settings used for τSTED imaging are summarized in Fig. 1B. To improve the signal-to-noise ratio, line accumulation was set to eight, and the STED laser power was set to 30%. Raw τSTED images were acquired with a pixel size of 15 nm using an HC PL APO CS2 100×/1.40 oil-immersion objective (Leica Microsystems).

τSTED image reconstruction was performed using the τSTED module in LAS X software (Leica Microsystems). Phasor plots were generated separately for each large field from the photon-arrival-time information acquired across the entire image. Because weak signals may not form a clearly defined distribution in the phasor plot, large fields were scanned slowly at low excitation intensity to improve the signal-to-noise ratio. The fluorophore-associated component appeared at a similar position across independent fields. An initial phasor region of interest (ROI) was selected around this component, and the corresponding mask was overlaid on the τSTED image to verify that it represented Golgi-marker-associated fluorescence throughout the field while excluding diffuse background and unrelated fluorescence components. A lifetime trajectory was then defined along the validated fluorophore-associated component. The trajectory was configured to assign higher intensity to the longer-lifetime side and lower intensity to the shorter-lifetime side, and the τSTED lifetime image was generated accordingly. The trajectory was defined for each entire field and was not optimized for individual cells or local regions. In representative fields with high signal-to-noise ratios, FLIM-mode reanalysis indicated lifetime ranges of 2.097–2.370 ns for GM130 and 1.971–2.248 ns for TGN46 labeled with Abberior STAR RED nanobodies; the trajectory endpoints varied by several percent among fields. These values are presented as reference ranges because τSTED reconstruction and conventional FLIM lifetime estimation are handled differently in the software. Representative phasor plots, trajectory definition, mask validation, and corresponding reconstructed images are shown in Fig. 1C and D. This lifetime-trajectory-based τSTED reconstruction is distinct from temporal detection gating in gated STED.

After τSTED image acquisition, the STED laser was switched off, and confocal Fucci images were acquired at the same positions. Confocal Fucci images and τSTED images of the Golgi apparatus were merged into four-channel composite images using ImageJ version 1.54g (National Institutes of Health, Bethesda, MD, USA). The distribution of GM130- or TGN46-positive structures in individual cells was quantified using a custom ImageJ macro (Supplementary Method S1).

Image analysis and major-axis measurement and object-based morphometry

The spatial extent of Golgi-marker-positive structures was quantified using an automated ImageJ/Fiji macro. For each labeling condition, cells were sampled from four large fields acquired from two independently prepared coverslips. Within each field, all cells with a clearly visible nucleus at the selected focal plane and an unambiguous red-only or green-only Fucci signal were analyzed; mixed or ambiguous Fucci-positive cells were excluded. Images were processed by rolling-ball background subtraction with a radius of 50 pixels (750 nm), followed by Gaussian smoothing (σ = 2 pixels) to suppress pixel-scale noise and isolated puncta. After thresholding and binary-mask generation, the boundary of the marker-positive distribution domain was detected and the major-axis length was measured. This measurement was interpreted as an index of the spatial extent of marker-positive distribution within each cell, not as the physical size of a single continuous Golgi object. For object-based morphometry, connected components in the thresholded binary masks were analyzed. To minimize detection of pixel-scale noise while accounting for the effective optical resolution, only objects with an area of at least 3 pixels (675 nm² at 15 nm/pixel; equivalent diameter approximately 29 nm) were included. Objects were classified by equivalent diameter as small (≤100 nm; 3–35 pixels), intermediate (>100 to ≤500 nm; 36–873 pixels), or large (>500 nm; ≥874 pixels). Total marker-positive area, total object number, and the number of objects in each size class were calculated for each cell. For the representative images in Fig. 3C and D, the boundaries of the detected objects were overlaid using the same size classes: cyan for small objects (≤100 nm), magenta for intermediate objects (>100 to ≤500 nm), and red for large objects (>500 nm).

Statistical analysis

Statistical analyses were performed using OriginPro. Pairwise comparisons shown in Fig. 3 were performed using unpaired two-tailed Student’s t-tests without correction for multiple comparisons. The sample size (n) represents the number of individual cells analyzed, with at least 10 cells analyzed for each condition. Cells were nested within four large-field images independently acquired from two separately prepared coverslips for each labeling condition, with 1–10 eligible cells contributed by an individual field. Sampling across multiple fields and coverslips reduced dependence on any single staining or acquisition instance and incorporated field-to-field technical variation into the analyzed cell population. However, field was not included as a random effect and the number of cells per field varied; therefore, the cell-level tests do not formally separate cell-to-cell from field-to-field variation. To evaluate the effects of cell-cycle phase, Golgi subdomain, and imaging/dye condition on major-axis length, a three-way ANOVA was performed with these three variables as fixed factors. Significant interactions were interpreted before the corresponding main effects because an interaction indicates that the effect of one factor depends on the level of another. P<0.05 was considered statistically significant. Comparisons of total marker-positive area, total object number, and size-class-specific object counts between G1-phase and G2-enriched cells were performed using two-tailed Mann–Whitney U tests. For the object-based analysis, n = 29 G1-phase and 8 G2-enriched cells for TGN46, and n = 11 G1-phase and 10 G2-enriched cells for GM130. Statistical significance symbols used in the figures are defined as follows: ns, not significant; *P<0.05; **P<0.01; ***P<0.001; ****P<0.0001.

Author Declaration Statements

Funding

This work was supported in part by JST-CREST (grant number JPMJCR21E3 to K Kato and JPMJCR21N6 to TM), the Human Glycome Atlas Project, MEXT/JSPS KAKENHI (JP24H00599 to K Kato and JP25K09591 to K Katoh), the ExCELLS Advanced Co-creation Platform (23EXC601 and 25EXC603), and ExCELLS “Golgi Atlas Project” (to K Kato) and the ExCELLS Research for Young Scientists (25-Y11 and 26-Y3 to ST) and the Joint Research by ExCELLS (program number 22EXC307, 23EXC304 and 24EXC312 to K Katoh) and ExCELLS Themed Research (Seeds Discovery) (program number 25EXC204 and 26EXC208 to TM).

Conflict of Interest Statement

The authors declare that they have no conflicts of interest regarding the contents of this article.

Data Availability Statement

Any additional information required to reanalyze the data reported in this paper is available from Koichi Kato ([e-mail]).

Author Contribution Statement

Conceived and designed experiments: K Katoh and K Kato

Performed the imaging analysis: K Katoh

Performed the data analysis: K Katoh, ST, and TM

Wrote the manuscript: K Katoh and K Kato

Ethics Approval and Consent to Participate

Ethics approval is not required in this study.

Patient Consent for Publication

Not applicable. This study did not involve patients or identifiable patient data.

Acknowledgments

We thank Mrs. Ayako Kojima and Dr. Masakazu Namihira (AIST) for help in keeping samples in good conditions.

Data Availability Statement

The supporting information for this article is available in J-STAGE Data.

• 1) Raw τSTED Images of GM130- and TGN46-Positive Golgi Structures in G1- and G2-Enriched HeLa Cells, with Corresponding Raw Confocal FUCCI Images

References

•  Alvarez,  L.,  Schwarz,  U.,  Friedrich,  L.,  Foelling,  J.,  Hecht,  F., and  Roberti,  M.J. 2021. TauSTED: Pushing STED beyond its limits with lifetime. Nature Methods, Application Note.
•  Colanzi,  A.,  Hidalgo Carcedo,  C.,  Persico,  A.,  Cericola,  C.,  Turacchio,  G.,  Bonazzi,  M.,  Luini,  A., and  Corda,  D. 2007. The Golgi mitotic checkpoint is controlled by BARS-dependent fission of the Golgi ribbon into separate stacks in G2. EMBO J., 26: 2465–2476.
•  Digman,  M.A.,  Caiolfa,  V.R.,  Zamai,  M., and  Gratton,  E. 2008. The phasor approach to fluorescence lifetime imaging analysis. Biophys. J., 94: L14–L16.
•  Frye,  K.,  Renda,  F.,  Fomicheva,  M.,  Zhu,  X.,  Gong,  L.,  Khodjakov,  A., and  Kaverina,  I. 2020. Cell cycle-dependent dynamics of the Golgi-centrosome association in motile cells. Cells, 9: E1069.
•  Hell,  S.W. and  Wichmann,  J. 1994. Breaking the diffraction resolution limit by stimulated emission: Stimulated-emission-depletion fluorescence microscopy. Opt. Lett., 19: 780–782.
•  Humphrey,  J.S.,  Peters,  P.J.,  Yuan,  L.C., and  Bonifacino,  J.S. 1993. Localization of TGN38 to the trans-Golgi network: Involvement of a cytoplasmic tyrosine-containing sequence. J. Cell Biol., 120: 1123–1135.
•  Lowe,  M.,  Rabouille,  C.,  Nakamura,  N.,  Watson,  R.,  Jackman,  M.,  Jämsä,  E.,  Rahman,  D.,  Pappin,  D.J., and  Warren,  G. 1998. Cdc2 kinase directly phosphorylates the cis-Golgi matrix protein GM130 and is required for Golgi fragmentation in mitosis. Cell, 94: 783–793.
•  Lucocq,  J.M.,  Berger,  E.G., and  Warren,  G. 1989. Mitotic Golgi fragments in HeLa cells and their role in the reassembly pathway. J. Cell Biol., 109: 463–474.
•  Lucocq,  J.M.,  Pryde,  J.G.,  Berger,  E.G., and  Warren,  G. 1987. A mitotic form of the Golgi apparatus in HeLa cells. J. Cell Biol., 104: 865–874.
•  Lujan,  P.,  Garcia-Cabau,  C.,  Wakana,  Y.,  Lillo,  J.V.,  Rodilla-Ramírez,  C.,  Malhotra,  V.,  Salvatella,  X.,  Garcia-Parajo,  M.F., and  Campelo,  F. 2024. Sorting of secretory proteins at the trans-Golgi network by TGN46. Elife, 12: RP91708.
•  Luzio,  J.P.,  Brake,  B.,  Banting,  G.,  Howell,  K.E.,  Braghetta,  P., and  Stanley,  K.K. 1990. Identification, sequencing and expression of an integral membrane protein of the trans-Golgi network (TGN38). Biochem. J., 270: 97–102.
•  Makhoul,  C.,  Gosavi,  P., and  Gleeson,  P.A. 2019. Golgi dynamics: The morphology of the mammalian Golgi apparatus in health and disease. Front. Cell Dev. Biol., 7: 112.
•  Mascanzoni,  F.,  Ayala,  I.,  Iannitti,  R.,  Luini,  A., and  Colanzi,  A. 2024. The Golgi checkpoint: Golgi unlinking during G2 is necessary for spindle formation and cytokinesis. Life Sci. Alliance, 7: e202302469.
•  Misteli,  T. and  Warren,  G. 1994. COP-coated vesicles are involved in the mitotic fragmentation of Golgi stacks in a cell-free system. J. Cell Biol., 125: 269–282.
•  Nagao,  Y.,  Sakamoto,  M.,  Chinen,  T.,  Okada,  Y., and  Takao,  D. 2020. Robust classification of cell cycle phase and biological feature extraction by image-based deep learning. Mol. Biol. Cell, 31: 1346–1354.
•  Nakamura,  N.,  Rabouille,  C.,  Watson,  R.,  Nilsson,  T.,  Hui,  N.,  Slusarewicz,  P.,  Kreis,  T.E., and  Warren,  G. 1995. Characterization of a cis-Golgi matrix protein, GM130. J. Cell Biol., 131: 1715–1726.
•  Pleiner,  T.,  Bates,  M.,  Trakhanov,  S.,  Lee,  C.T.,  Schliep,  J.E.,  Chug,  H.,  Böhning,  M.,  Stark,  H.,  Urlaub,  H., and  Görlich,  D. 2015. Nanobodies: Site-specific labeling for super-resolution imaging, rapid epitope mapping and native protein complex isolation. Elife, 4: e11349.
•  Ponnambalam,  S.,  Girotti,  M.,  Yaspo,  M.L.,  Owen,  C.E.,  Perry,  A.C.,  Suganuma,  T.,  Nilsson,  T.,  Fried,  M.,  Banting,  G., and  Warren,  G. 1996. Primate homologues of rat TGN38: Primary structure, expression and functional implications. J. Cell Sci., 109(Pt 3): 675–685.
•  Ponnambalam,  S.,  Rabouille,  C.,  Luzio,  J.P.,  Nilsson,  T., and  Warren,  G. 1994. The TGN38 glycoprotein contains two non-overlapping signals that mediate localization to the trans-Golgi network. J. Cell Biol., 125: 253–268.
•  Prescott,  A.R.,  Lucocq,  J.M.,  James,  J.,  Lister,  J.M., and  Ponnambalam,  S. 1997. Distinct compartmentalization of TGN46 and beta 1,4-galactosyltransferase in HeLa cells. Eur. J. Cell Biol., 72: 238–246.
•  Reaves,  B.,  Horn,  M., and  Banting,  G. 1993. TGN38/41 recycles between the cell surface and the TGN: Brefeldin A affects its rate of return to the TGN. Mol. Biol. Cell, 4: 93–105.
•  Sakaue-Sawano,  A. and  Miyawaki,  A. 2014. Visualizing spatiotemporal dynamics of multicellular cell-cycle progressions with fucci technology. Cold Spring Harb. Protoc., 2014: pdb.prot080408.
•  Shima,  D.T.,  Cabrera-Poch,  N.,  Pepperkok,  R., and  Warren,  G. 1998. An ordered inheritance strategy for the Golgi apparatus: Visualization of mitotic disassembly reveals a role for the mitotic spindle. J. Cell Biol., 141: 955–966.
•  Sograte-Idrissi,  S.,  Schlichthaerle,  T.,  Duque-Afonso,  C.J.,  Alevra,  M.,  Strauss,  S.,  Moser,  T.,  Jungmann,  R.,  Rizzoli,  S.O., and  Opazo,  F. 2020. Circumvention of common labelling artefacts using secondary nanobodies. Nanoscale, 12: 10226–10239.
•  Sütterlin,  C.,  Hsu,  P.,  Mallabiabarrena,  A., and  Malhotra,  V. 2002. Fragmentation and dispersal of the pericentriolar Golgi complex is required for entry into mitosis in mammalian cells. Cell, 109: 359–369.
•  Sütterlin,  C.,  Lin,  C.Y.,  Feng,  Y.,  Ferris,  D.K.,  Erikson,  R.L., and  Malhotra,  V. 2001. Polo-like kinase is required for the fragmentation of pericentriolar Golgi stacks during mitosis. Proc. Natl. Acad. Sci. USA, 98: 9128–9132.
•  Tang,  D. and  Wang,  Y. 2013. Cell cycle regulation of Golgi membrane dynamics. Trends Cell Biol., 23: 296–304.
•  Tie,  H.C.,  Ludwig,  A.,  Sandin,  S., and  Lu,  L. 2018. The spatial separation of processing and transport functions to the interior and periphery of the Golgi stack. Elife, 7: e41301.
•  Vicidomini,  G.,  Moneron,  G.,  Han,  K.Y.,  Westphal,  V.,  Ta,  H.,  Reuss,  M.,  Engelhardt,  J.,  Eggeling,  C., and  Hell,  S.W. 2011. Sharper low-power STED nanoscopy by time gating. Nat. Methods, 8: 571–573.
•  Wei,  J.-H. and  Seemann,  J. 2017. Golgi ribbon disassembly during mitosis, differentiation and disease progression. Curr. Opin. Cell Biol., 47: 43–51.
•  Wong-Dilworth,  L.,  Rodilla-Ramirez,  C.,  Fox,  E.,  Restel,  S.D.,  Stockhammer,  A.,  Adarska,  P., and  Bottanelli,  F. 2023. STED imaging of endogenously tagged ARF GTPases reveals their distinct nanoscale localizations. J. Cell Biol., 222: e202205107.

Abbreviations

AG1

Azami Green 1

ANOVA

analysis of variance

BSA

bovine serum albumin

CSB

cytoskeleton-stabilizing buffer

DMEM

Dulbecco’s modified Eagle’s medium

FBS

fetal bovine serum

FLIM

fluorescence lifetime imaging microscopy

Fucci

fluorescent ubiquitination-based cell-cycle indicator

KO2

Kusabira Orange 2

STED

stimulated emission depletion

TGN

trans-Golgi network

τSTED

fluorescence lifetime-based stimulated emission depletion microscopy