A new model for nuclear envelope breakdown

Abstract

Nuclear envelope breakdown was investigated during meiotic maturation of starfish oocytes. Fluorescent 70-kDa dextran entry, as monitored by confocal microscopy, consists of two phases, a slow uniform increase and then a massive wave. From quantitative analysis of the first phase of dextran entry, and from imaging of green fluorescent protein chimeras, we conclude that nuclear pore disassembly begins several minutes before nuclear envelope breakdown. The best fit for the second phase of entry is with a spreading disruption of the membrane permeability barrier determined by three-dimensional computer simulations of diffusion. We propose a new model for the mechanism of nuclear envelope breakdown in which disassembly of the nuclear pores leads to a fenestration of the nuclear envelope double membrane.

Figure 1

Organization of the immature starfish oocyte. The large GV (nucleus) is positioned near the oocyte surface at the animal pole (arrow). The centrosomes are located in the small region between the GV and the animal pole. The GV has a circular cross section when viewed along the animal-vegetal axis but is often slightly flattened when viewed perpendicularly to that axis. Approximately one-half of the cytoplasmic space is occupied by the yolk platelets, which are oval-shaped organelles with dimensions of 1–2 μm. The GV breaks down during meiotic maturation: see movie “gvbdt.mov” at the Molecular Biology of the Cell web site. Bar, 50 μm.

Figure 2

Dextran (70 kDa) entry during GVBD. Rhodamine dextran (70 kDa) was injected into the cytosol and imaged by a two-photon microscope at 2.5-s intervals. This image sequence was begun 10–15 min after addition of the maturation hormone 1-MA. Timing relative to the first image is shown on each image (minutes:seconds). There is an initial phase of slow entry followed by a massive entry from one region of the nuclear envelope at the time of GVBD (between g and h). The wave front develops a concave shape. See movie “2pgvbd.mov” at the Molecular Biology of the Cell web site. The small dark circle to the right of the GV is an oil drop introduced by microinjection; another oil drop out of the plane of focus accounts for the dark region beneath the GV. Bar, 50 μm. Graph, from this image sequence, the average fluorescence intensity of a 30- × 30-μm region at the center of the GV was determined. The time points corresponding to those shown in the image sequence are indicated by the lower case letters.

Figure 2

Dextran (70 kDa) entry during GVBD. Rhodamine dextran (70 kDa) was injected into the cytosol and imaged by a two-photon microscope at 2.5-s intervals. This image sequence was begun 10–15 min after addition of the maturation hormone 1-MA. Timing relative to the first image is shown on each image (minutes:seconds). There is an initial phase of slow entry followed by a massive entry from one region of the nuclear envelope at the time of GVBD (between g and h). The wave front develops a concave shape. See movie “2pgvbd.mov” at the Molecular Biology of the Cell web site. The small dark circle to the right of the GV is an oil drop introduced by microinjection; another oil drop out of the plane of focus accounts for the dark region beneath the GV. Bar, 50 μm. Graph, from this image sequence, the average fluorescence intensity of a 30- × 30-μm region at the center of the GV was determined. The time points corresponding to those shown in the image sequence are indicated by the lower case letters.

Figure 3

Permeability of the GV envelope in interphase and in the period before GVBD. (A) Fluorescein dextran (10 kDa) was injected into the cytoplasm of an immature oocyte and was allowed to equilibrate across the GV envelope. Fluorescence in the GV interior was partially photobleached, and the recovery of fluorescence due to diffusion through nuclear pores was imaged. Bar, 50 μm. (B) The average fluorescence intensity in a small region in the GV was determined. The data were fit well by an exponential recovery, corresponding to a permeability coefficient of ∼0.15 μm/s. (C) Permeability of the GV envelope for 70-kDa dextran in the period before GVBD. The permeability coefficient is not constant and was determined for each time point by measuring the slope of the line from a graph like that in Figure 2 and then dividing by the gradient of 70-kDa dextran across the GV envelope. The average value for the permeability coefficient just before GVBD was 0.04 μm/s. The similarity of this value to the permeability of 10-kDa dextran through nuclear pores in the immature oocyte is consistent with entry of 70-kDa through altered nuclear pores with an increased pore size cutoff.

Figure 3

Permeability of the GV envelope in interphase and in the period before GVBD. (A) Fluorescein dextran (10 kDa) was injected into the cytoplasm of an immature oocyte and was allowed to equilibrate across the GV envelope. Fluorescence in the GV interior was partially photobleached, and the recovery of fluorescence due to diffusion through nuclear pores was imaged. Bar, 50 μm. (B) The average fluorescence intensity in a small region in the GV was determined. The data were fit well by an exponential recovery, corresponding to a permeability coefficient of ∼0.15 μm/s. (C) Permeability of the GV envelope for 70-kDa dextran in the period before GVBD. The permeability coefficient is not constant and was determined for each time point by measuring the slope of the line from a graph like that in Figure 2 and then dividing by the gradient of 70-kDa dextran across the GV envelope. The average value for the permeability coefficient just before GVBD was 0.04 μm/s. The similarity of this value to the permeability of 10-kDa dextran through nuclear pores in the immature oocyte is consistent with entry of 70-kDa through altered nuclear pores with an increased pore size cutoff.

Figure 4

Kinetics of a nuclear pore GFP chimera during GVBD. RanGAP is a peripheral protein of the nuclear pore. RanGAP-GFP was expressed by mRNA injection, and the following day, 70-kDa rhodamine dextran was injected into the oocytes. RanGAP-GFP (left) was imaged alternately with 70-kDa rhodamine dextran (right) during maturation at 7-s intervals so that each fluorescent marker was imaged at 14-s intervals (see MATERIALS AND METHODS for more details). As shown by 70-kDa dextran entry, GVBD occurred at ∼507 s. RanGAP-GFP fluorescence at the nuclear rim just before GVBD at 493 s has decreased noticeably from the 0-s time point. In this oocyte, the 70-kDa dextran begins to enter at the sides of the GV rather than in the animal pole region. See movie “rangap.mov” at the Molecular Biology of the Cell web site. Bar, 50 μm. The graph shows fluorescence of RanGAP-GFP in the nuclear rim and of 70-kDa dextran in the GV interior. RanGAP-GFP fluorescence decreases significantly before GVBD and seems to decrease with approximately the same time course as the first phase of 70-kDa dextran entry. This is evidence that the nuclear pores are being disassembled before GVBD and that 70-kDa dextran enters through nuclear pores whose cutoff has been increased.

Figure 4

Kinetics of a nuclear pore GFP chimera during GVBD. RanGAP is a peripheral protein of the nuclear pore. RanGAP-GFP was expressed by mRNA injection, and the following day, 70-kDa rhodamine dextran was injected into the oocytes. RanGAP-GFP (left) was imaged alternately with 70-kDa rhodamine dextran (right) during maturation at 7-s intervals so that each fluorescent marker was imaged at 14-s intervals (see MATERIALS AND METHODS for more details). As shown by 70-kDa dextran entry, GVBD occurred at ∼507 s. RanGAP-GFP fluorescence at the nuclear rim just before GVBD at 493 s has decreased noticeably from the 0-s time point. In this oocyte, the 70-kDa dextran begins to enter at the sides of the GV rather than in the animal pole region. See movie “rangap.mov” at the Molecular Biology of the Cell web site. Bar, 50 μm. The graph shows fluorescence of RanGAP-GFP in the nuclear rim and of 70-kDa dextran in the GV interior. RanGAP-GFP fluorescence decreases significantly before GVBD and seems to decrease with approximately the same time course as the first phase of 70-kDa dextran entry. This is evidence that the nuclear pores are being disassembled before GVBD and that 70-kDa dextran enters through nuclear pores whose cutoff has been increased.

Figure 5

Simulation of 70-kDa entry at the time of GVBD. Left, simulation of entry through a hole with a fixed diameter of 45 μm. Right, simulation of entry through a hole that opens at a constant rate and is complete at 35 s. Timing of the images relative to the beginning of the opening in the simulation are shown. The simulations were done in three dimensions, and the equatorial plane is shown. The simulation with the expanding hole fits the experimental data (Figure 2) much better than the simulation with the fixed hole. See movies “fixed.mov” and “spreading.mov” at the Molecular Biology of the Cell web site. Bar, 50 μm.

Figure 6

A new model for nuclear envelope (NE) breakdown. Active MPF enters the nucleus and begins a stepwise process of nuclear pore (np) disassembly. At some point, perhaps corresponding to loss of the central plug, larger molecules can diffuse through the pore; this is proposed to correspond to the first phase of 70-kDa dextran entry. The nuclear pores continue to be disassembled until the membrane hole in which they are located becomes destabilized, and the holes can begin to expand. In some unknown manner, the expansion of the holes is propagated throughout the nuclear envelope, resulting in a fenestrated membrane; this is proposed to correspond to the second phase of 70-kDa dextran entry. In this model, the nuclear envelope membranes remain continuous and are resorbed into the ER during mitosis. Nuclear lamina disassembly is required for normal mitosis and occurs at approximately the same time, but its role in the disruption of the nuclear envelope membrane barrier is uncertain.