Organisation of human ER-exit sites: requirements for the localisation of Sec16 to transitional ER

Abstract

The COPII complex mediates the selective incorporation of secretory cargo and relevant machinery into budding vesicles at specialised sites on the endoplasmic reticulum membrane called transitional ER (tER). Here, we show using confocal microscopy, immunogold labelling of ultrathin cryosections and electron tomography that in human cells at steady state, Sec16 localises to cup-like structures of tER that are spatially distinct from the localisation of other COPII coat components. We show that Sec16 defines the tER, whereas Sec23-Sec24 and Sec13-Sec31 define later structures that precede but are distinct from the intermediate compartment. Steady-state localisation of Sec16 is independent of the localisation of downstream COPII components Sec23-Sec24 and Sec13-Sec31. Sec16 cycles on and off the membrane at a slower rate than other COPII components with a greater immobile fraction. We define the region of Sec16A that dictates its robust localisation of tER membranes and find that this requires both a highly charged region as well as a central domain that shows high sequence identity between species. The central conserved domain of Sec16 binds to Sec13 linking tER membrane localisation with COPII vesicle formation. These data are consistent with a model where Sec16 acts as a platform for COPII assembly at ERES.

Fig. 1.

Localization of GFP-Sec16A, COPII subunits and ERGIC components. (A,B) HeLa cells transfected with GFP-Sec16A fixed after 22 hours, stained with antibodies to the COPII subunit Sec24C and the ERGIC marker ERGIC-53. (C,D) HeLa cells transfected with GFP-Sec16A fixed after 22 hours and stained with antibodies to the COPII subunit Sec31A and the COPI subunit β′-COP. The images were taken using high resolution laser-scanning confocal microscopy, and images underwent iterative deconvolution. (E,F) HeLa cells fixed and stained with antibodies to Sec23A and Sec31A, showing complete colocalisation between COPII proteins. Enlarged images from the panels above are shown in B,D,F. Scale bars: 5 μm (A,C,E); 100 nm (B,D,F).

Fig. 2.

Sec16A and Sec31A immunoelectron microscopy. (A-C) Localisation of Sec16A and Sec31A by immunogold labelling of ultrathin cryosections. Sec16A (6 nm gold particles) and Sec31A (10 nm gold particles) were immunolocalised on cryosections of HeLa cells. (A) Examples of measurement of distance between Sec16A and Sec31A. (B) Quantification of distance of gold grains from three independent experiments, taken from 74 images of juxtanuclear areas of HeLa cells. (C) Colocalisation with Sec31A, membrane association and cup-shaped labelling pattern for Sec16A. Scale bars: 200 nm. (D-J) Electron tomography of Sec16-labelled structures. (D) Sec16A immunolocalisation on a 210 nm cryosection in electron tomography. Image shows four merged orthoslices. (E) Example from the transparency projection. ER membrane in yellow, Sec16A label in red. (F) Magnified image showing association of Sec16A with membrane structures from a different tomogram. (G,H) Rotated magnifications of one Sec16 cluster; (I,J) Rotated magnifications of a distinct cluster (both from the data set in D,E). Scale bar: 200 nm.

Fig. 3.

Fluorescence photobleaching of GFP-Sec16A and GFP-Sec23A at ERES. (A-C) FRAP using GFP-Sec23A, (D-F) or GFP-Sec16A. Panels show the region of interest used for photobleaching (circled) for GFP-Sec23A (A) and GFP-Sec16A (D). (B,E) An example plot from a single ERES is shown; the insets show these data fitted to a single exponential. For GFP-Sec23A (C) and for GFP-Sec16A (F) data from ERES were averaged and plotted (error bars show s.d., n=21 for GFP-Sec16A and n=27 for GFP-Sec23A); insets show these data fitted to a single exponential. Individual recovery rates (G) and immobile fractions (H) were plotted. Data shown represent the mean of at least 16 independent experiments (horizontal bar). (I) HeLa cells expressing GFP-Sec16A were repeatedly bleached within a region of interest indicated in panel 2 (t=10 seconds), at 37°C followed by imaging the entire field of view at low laser power. (J) Mean fluorescence intensities (±s.d.) of nine ERES are plotted. This example is representative of a data set of six different cells, and for time courses of both 12 minutes and 30 minutes. After repeated bleaching, GFP-Sec16A-labelled ERES outside the ROI showed a slow loss in fluorescence (black circles) relative to no loss in fluorescence as seen for ERES of neighbouring cells (grey squares). F, fluorescence intensity; F(av), average fluorescence intensity, both in arbitrary units.

Fig. 4.

Association of Sec23-Sec24 and Sec13-Sec31 with ERES membranes is not required to maintain the association of Sec16. (A-D) HeLa cells were transfected with Myc-TBC1D20 and fixed after 22 hours. High expressers were immediately identified using GM130 staining because TBC1D20 expression results in loss of the Golgi complex (Haas et al., 2007) and by subsequent immunolabelling with anti-Myc antibodies (asterisks). Arrow indicates a cell in which the Golgi is completely disrupted with no apparent effect on juxtanuclear ERES (see Haas et al., 2007). Transfection of the catalytically inactive Myc-TBC1D20-R105A had no effect on Golgi organisation or COPII localisation. Cells were co-stained with antibodies to GM130 and (A) Sec16A, (B) Sec24C, (C) Sec31A and (D) COPI. Scale bar: 20 μm. Cells expressing Sec16A and either wild-type TBC1D20, or a catalytically inactive R105A point mutant were used for FRAP experiments; cells were confirmed to express Myc-TBC1D20 by immunofluorescence after FRAP; 97% of all cells transfected with GFP-Sec16A were also transfected with Myc-TBC1D20. (E) Plot of the half-life of recovery of fluorescence for GFP-Sec16A in cells expressing Myc-TBC1D20 or Myc-TBC1D20-R105A. (F) Plot of immobile population of GFP-Sec16A in the presence of TBC1D20 wild type and mutant expressers. The bar represents the mean with individual data points shown. Data shown represent the mean of at least nine independent experiments.

Fig. 5.

Requirements for the localisation of GFP-Sec16A in HeLa cells. The top of the figure shows a schematic of constructs generated and summary of findings; lettering refers to both the schematic and micrographs below. Amino acid numbers are shown relative to the full Sec16A (KIAA0310) coding sequence, GFP is indicated by the green bar. Lettering refers to the panels below. (A) Full-length GFP-Sec16A (1-2357) localises to tER. Constructs incorporating regions of Sec16A (see schematic summary) were expressed as GFP-fusion proteins in HeLa cells, fixed and imaged to determine tER localisation. Constructs generated include the following amino acids of Sec16A fused to GFP (B) 1382-2357, (C) 1-1254, (D) 1-2056, (E) 1-1615+, (F) 1009-2357, (G) 1019-1890, (J) 1019-1433 and (M) 1434-1890. (G-I) GFP-Sec16A (1019-1890) efficiently targets to exit sites (G) and shows close apposition to puncta labelled with Sec24C (arrows in the enlargements in G). High expression of GFP-Sec16A (1019-1890) leads to loss of endogenous Sec16A (H) as well as Sec24C (I). (J-L) Sec16A (1019-1433) is not targeted to tER membranes. Low expression of the fusion protein has no effect on Sec24C (J); however, high expression leads to a disruption of Sec16A (K) and Sec24C (L). (M-O) Sec16A (1434-1890) is not targeted to tER membranes and does not cause significant disruption of endogenous Sec16A (O) or Sec24C (N) when highly overexpressed. Scale bars: 20 μm.

Fig. 6.

Fluorescence recovery after photobleaching of GFP-Sec16A (1019-1890). HeLa cells were transfected with GFP-Sec16A (1019-1890) and FRAP was performed on single ERES (A-J) and high-expressing cells (non-ERES) (K-N). Single ERES were found to fall into three populations: one showing recovery (A-D), a second recovering at a slow linear rate (E-G), and a third population showing no recovery (H-J). (A,E,H) Recovery curves of single ERES; insets show data fitted to a single exponential. (B,F,I) Average recovery curves; insets show data fitted to a single exponential. (C,G,J) For each single ERES, images were taken before the bleach (t=0 seconds), immediately after the bleach (t=3 seconds) and at the end of the time course (t=25.7 seconds). Scale bar: 10 μm. (D) Half-life of GFP-Sec16A (1019-1890). Each point represents the half-life of a single ERES, the bar represents the mean. (K-M) Photobleaching of cells highly expressing GFP-Sec16A (1019-1890). The recovery curve of a single ROI and the resulting curve fit (inset) is shown in K. Images acquired before, immediately after the bleach and at the end of the time course are shown in M. Scale bar: 10 μm. In L, the average recovery curve of 23 experiments as well as the curve fit (inset) is displayed. (N) Half-life of GFP-Sec16A (1019-1890) in high-expressing cells. Each point represents the half-life of a single ROI, the bar represents the mean. F, fluorescence intensity; F(av), the mean. Error bars show s.d. See also Table 1.

Fig. 7.

Interaction of Sec16A CCD and Sec13. (A) Triplicate spots of yeast co-transformed with pGADT7-T (left) or pGADT7-Sec16A CCD (right) and COPII subunits as indicated were plated onto quadruple dropout medium and assayed for growth after 4 days. A positive interaction is seen only between large T antigen (pGADT7-T) and p53 (positive control) and between Sec16A CCD and Sec13. All colonies grow on double dropout medium (not shown).

Fig. 8.

Schematic of an ERES. At steady state, Sec16 localises to tER (green), Sec23-Sec24 (and also Sec13-Sec31) predominantly localise to post-ER structures (red), and ERGIC-53 to the pre-VTC and ERGIC (blue). These data demonstrate steady-state localisations; functionally considerable overlap will occur (i.e. Sec16 can bind directly to Sec23-Sec24, and Sec23-Sec24 mediates the export of cargo from the ER including ERGIC-53).