A COPII assembly assay using purified COPII components. Digitonin-permeabilized NRK cells were incubated with 1 μg/ml Sar1p, 3 μg/ml Sec23/24p, 7 μg/ml Sec13/31p, 1 mg/ml defatted BSA, 500 μM GTPγS, and an ATP-regenerating system (ATP r.s.) for 10 min at 32°C (A). In B–F, Sec23/24p (B), Sar1p (C), all nucleotides (D), GTPγS only (E), or the ATP r.s. only (F) was omitted from the reaction. Cells were fixed and stained using a rabbit serum against Sec24p. Bar, 10 μm. (G) Quantification (see Materials and Methods), mean ± SEM, of the average peripheral fluorescence per cell (>25 cells) resulting from each omission

Salt-wash inhibition of COPII assembly and restoration by a cytosolic factor. Digitonin-permeabilized NRK cells were washed with 70 mM KOAc (A) or 1 M KCl (B and C) and incubated with purified COPII proteins (as in Figure 1), 500 μM GTPγS, and an ATP-regenerating system (ATP r.s.) in the absence (A and B) or presence (C) of 40 μg/ml HeLa cytosol. After incubation for 10 min at 32°C, cells were fixed and stained using antibodies against Sec24p. Bar, 10 μm. Quantification of the average peripheral fluorescence per cell (>25 cells), mean ± SEM, under each of the indicated conditions is shown (D).

Purification of the restorative activity from cytosol. The monoS peak (see Materials and Methods; Figure 4 and Table 1) was further fractionated on an S200 gel filtration column (A), and the resulting fractions were resolved by SDS-PAGE and silver stained (B). The major copurifying polypeptide at 17 to 20 kDa is indicated by the arrowhead, and contaminating keratins at ∼55–65 kDa, as well as a minor copurifying polypeptide at ∼40 kDa, are indicated by asterisks. The positions at which the gel filtration markers migrated on the column are indicated (A). (C) Digitonin-permeabilized and salt-washed NRK cells incubated with purified COPII proteins (as in Figure 1), 500 μM GTPγS, and an ATP-regenerating system (ATP r.s.) in the absence (control) or presence (+ 2 U/ml the S200 peak) of fraction 14. After 10 min at 32°C, cells were fixed and stained using antibodies against Sec24p.

Both endogenous and recombinant Nm23H2 are localized to an ER-like network along which a subset of COPII assembly sites generated in vitro align. (A–C) Nm23H2 localizes to a tubular/filamentous network that may in part be associated with the ER. NRK cells were digitonin permeabilized, fixed immediately in paraformaldehyde, doubly stained with an Nm23H2 mAb (A, green in merged image C) and a polyclonal antibody against calnexin (B, red in merged image C), and viewed by confocal microscopy. A single optical section from each individual, as well as the merged, channels is shown. Bar, 10 μm. (D–F) A subset of in vitro COPII assembly sites align along Nm23H2 tubules/filaments. Digitonin-permeabilized NRK cells were incubated with HeLa cytosol (4 mg/ml) in the presence of 500 μM GTPγS and an ATP-regenerating system (ATP r.s.) After 10 min at 32°C, cells were fixed and doubly stained using a mAb against Nm23H2 (D, green in merged image F) and a polyclonal antibody against Sec13p (E, red in merged image F). Arrowheads indicate a subset of COPII structures that overlap with or are adjacent to an Nm23H2 tubule/filament in the periphery of the cell. Bar, 5 μm. (G–L) 6-His-Nm23H2 localizes to an ER-like network. Digitonin-permeabilized NRK cells were incubated with rat liver cytosol (3 mg/ml), 0.025 mg/ml 6-His-Nm23H2, 500 μM GTPγS, and an ATP r.s. After 10 min at 32°C, cells were fixed and doubly stained either for the 6-His epitope (G) and Sec13p (H) or the 6-His epitope (J) and calnexin (K). The merged image for 6-His (G), green, and Sec13p (H), red, is shown in I. The merged image for 6-His (J), green, and calnexin (K), red, is shown in L. Bar, 5 μm.

Depletion of Nm23H2 protein reduces COPII assembly in vivo. (A) Equal protein loadings of homogenates from HeLa cells mock-transfected, transfected with a control siRNA, Nm23H2 siRNA #24, or Nm23H2 siRNA #14, as described under Materials and Methods, were subjected to SDS-PAGE and immunoblotting by using antibodies against the indicated proteins. (B–E) Cells mock transfected (B), tranfected with a control siRNA (C), siRNA#24 (D), or siRNA #14 (E) were fixed and stained using affinity-purified antibodies against Sec13p. Projected images from confocal acquisitions are shown. Bar, 10 μm. (F) Images from B–E were captured and quantified as described (Materials and Methods). The fluorescence in all discernable objects per cell (50 cells) is shown, mean + SD.

Depletion of Nm23H2 protein leads to reduced kinetics of GalNacT2-GFP export from the ER. HeLa cells stably expressing GalNacT2-GFP and mock transfected, transfected with a control siRNA, siRNA #24, or siRNA #14, were incubated with 5 μg/ml BFA for 20 min. Thereafter, cells were washed into normal medium and fixed at various times to assess the extent of reemergence of GalNacT2-GFP into post-ER compartments. A representative time course, before the BFA treatment (A and E), and at 0 min (B and F), 40 min (C and G), or 100 min (D and H) after the BFA washout, for mock-transfected cells (A–D) and cells transfected with siRNA #24 (E–H) is shown. Bar, 20 μm. (I) A quantification of the percentage of cells with GalNacT2-GFP in post-ER compartments at each time point for each transfection condition is shown. More than 150 cells were analyzed at each time point. The results represent mean ± SD.

Purified COPII proteins assemble at bona fide ER exit sites. To collapse Golgi proteins into the ER, NRK cells were incubated with 2.5 μg/ml BFA for 20 min, followed by 2.5 μg/ml BFA + 100 μM H89 for 10 min. After five washes into normal medium at RT, cells were digitonin permeabilized and incubated with 1 μg/ml Sar1p, 3 μg/ml Sec23/24p, 7 μg/ml Sec13/31p, 1 mg/ml BSA, 500 μM GTPγS, and an ATP-regenerating system (ATP r.s.) for 20 min at 32°C. Cells were fixed and doubly stained using a rabbit serum against Sec24p (A) and a mAb against giantin (B). The merged image with Sec24p in red and giantin in green is shown (C). Bar, 10 μm.

Purification procedure used to isolate the restorative activity from cytosol.

Recombinant wild-type and catalytically inactive 6-His-Nm23H2 both facilitate COPII assembly in vitro. (A) One, 3, and 9 μg of purified wild-type 6-His-Nm23H2 as well as the catalytically inactive version Nm23H2 (H118C) were resolved by SDS-PAGE and stained with Coomassie Blue. (B) A standard coupled assay that measures the nucleoside diphosphokinase-dependent oxidation of NADH (see Materials and Methods) was performed on wild-type and (H11C) versions of 6-His-Nm23H2. The relative decrease in NADH absorbance over time is plotted for each. (C–G) Both wild-type and (H118C) versions of Nm23H2 facilitate COPII assembly. Digitonin-permeabilized NRK cells washed with low (C) or high (D–F) salt were incubated with purified COPII proteins (as in Figure 1), 500 μM GTPγS, and an ATP-regenerating system (ATP r.s.) in the absence (C and D) or presence of 0.05 mg/ml wild-type 6-His-Nm23H2 (E) or 0.05 mg/ml 6-His-Nm23H2 (H118C). After 10 min at 32°C, cells were fixed and stained using antibodies against Sec24p. Bar, 5 μm. Quantification of the average total Sec24 object fluorescence per cell (>25 cells), mean ± SEM, under each of the indicated conditions is shown (G).

Nm23H2 facilitates the assembly of mammalian cytosolic COPII in vitro. Ammonium sulfate precipitation was used to fractionate Nm23H2 from COPII proteins in rat liver cytosol (see Materials and Methods). (A) The 0–30% fraction and the 30–80% fraction were resolved by SDS-PAGE and probed by immunoblot by using antibodies against Sec13p and Nm23H2. (B–D) Digitonin-permeabilized and salt-washed cells were incubated with unfractionated cytosol (3 mg/ml) (B), the 0–30% fraction (0.3 mg/ml) in the presence of 1 μg/ml yeast Sar1p (C), or the 0–30% fraction (0.3 mg/ml) in the presence of 1 μg/ml yeast Sar1p and 0.1 mg/ml 6-His-Nm23H2 (D). After incubation for 10 min at 32°C, cells were fixed and stained using antibodies against Sec13p. Bar, 5 μm. Quantification of the total fluorescence in Sec13p objects per cell (>25 cells), mean ± SEM, under each of the indicated conditions is shown (E).

Depletion of Nm23H2 protein does not inhibit the association of COPI with the Golgi. HeLa cells stably expressing GalNacT2-GFP were mock-transfected or transfected with Nm23H2 siRNA #24. Cells were fixed and stained either using an antibody against Sec13p (A) or an antibody against βCOP (B) and analyzed by confocal microscopy. Projected images were captured and quantified as described (Materials and Methods). (A) The Sec13p fluorescence in all discernable objects per cell is shown (>25 cells), mean ± SD. (B) The ratio of βCOP fluorescence (rhodamine) in the Golgi region to GalNacT2-GFP fluorescence in the Golgi region of each cell (>25 cells), mean ± SD is shown.