Cell biology: How does the nucleus get its membrane?

Summary

Nuclear shape and size depend on nuclear membrane availability through an unknown process. Using asymmetric cell division, Roubinet et al suggest that nuclear membrane is derived from the endoplasmic reticulum and that limiting nuclear membrane expansion can affect cell fate.

The idea that form affects function is plausible, but, in biology, seldom understood. The relationship between altered nuclear morphology, often observed in cancer cells, and cellular transformation is largely unknown¹, and the mechanisms by which lobulated nuclei of certain hematopoietic cells affect cell function are only starting to emerge². In both cases, the irregular shape of the nucleus is accompanied by a relative increase in nuclear surface area. Most cells have spherical or ovoid nuclei; how is the amount of nuclear membrane in these cells regulated so that the nucleus has a smooth appearance under the microscope? Where does the nuclear membrane come from, and what process or processes regulate the amount of membrane allocated to the nucleus so that it “fits” just right? In a new study published in this issue of Current Biology, Roubinet et al³ utilized a developmental system in which cells undergo an unusual form of nuclear division. This allowed them to determine a likely source of nuclear membrane and to conclude that restricting nuclear membrane expansion can affect nuclear function.

The nucleus is surrounded by a nuclear envelope that is made of two membranes, an inner and an outer nuclear membrane, that separate the nuclear compartment from the cytoplasm (Figure 1). Embedded in these membranes are hundreds to thousands of nuclear pore complexes that allow selective transport of proteins and RNA between the two compartments. The outer nuclear membrane is continuous with the rest of the ER and they share many of the same proteins; in fact, the lumen between the two nuclear membranes and the lumen of the ER are one and the same (for simplicity, from here on the term ER will refer to all ER membranes except the nuclear membranes). In animal cells, underlying the inner nuclear membrane is an intermediate filament meshwork, called the nuclear lamina, that is made of lamins and contributes to nuclear rigidity, chromosome organization and nuclear function⁴.

Figure 1: The nuclear envelope.

The nuclear envelope is composed of two membranes (in blue; the inner nuclear membrane (INM) and outer nuclear membrane (ONM)), nuclear pore complexes (NPC; in purple) and a nuclear lamina (in red). Not shown are many inner and outer nuclear membrane proteins that are also part of the nuclear envelope. The ONM is continues with the ER, a single tube of which is shown connected to the nuclear envelope. Green arrows: indicate a hypothetical flow of membrane components from the ER to the nuclear membranes. Orange arrows: indicate a hypothetical force applied on the nuclear membranes when nuclear protein concentration increases.

Until recently, it was thought that the nuclear membrane is inherited during cell division only in cells that undergo “closed mitosis”, which occurs in many fungi. During closed mitosis the nuclear membrane expands to accommodate chromosomes segregation, and the nuclear and cytoplasmic compartments do not mix⁴. This type of mitosis is contrasted with “open mitosis”, common to all metazoans, where the nuclear envelope breaks down due to lamina and nuclear pore complex disassembly, thereby abolishing the permeability barrier imposed by the nuclear membranes during interphase. It was generally assumed that the during open mitosis the nuclear membrane itself also disappears. However, more recent studies suggest that the at least in some cases, nuclear membranes become highly fenestrated but remains otherwise intact around the mitotic spindle (for example, see Rahman et al⁵ and references therein). Nonetheless, even if nuclear membrane fenestration is a ubiquitous mitotic strategy, the nuclear envelopes of the daughter nuclei, where examined, form by the assembly of new membrane structures around the two segregated chromosomes, and not by the collapse of the original nuclear membrane onto the chromosome masses⁶. Roubinet et al now show that in fly neural stem cells undergoing asymmetric cell division, the nuclear envelope of the daughter cells, including the nuclear lamina, is inherited from the parent cell³. In this system, the nuclear membranes become fenestrated then elongate as the chromosomes separate during anaphase. Nuclear pore complexes disassemble but, surprisingly, the nuclear lamina does not (although its amount is reduced by nearly 2-fold). In fact, the persistence of the lamina was found to be essential for this nuclear division to occur. This observation is remarkable for several reasons: first, it implies that metazoans can employ a range of strategies that determine the fate of the nuclear membranes during mitosis, far more than we appreciated. Second, nuclear lamina disassembly had been the hallmark of metazoan mitosis; how do these cells manage to undergo mitosis with the nuclear lamina present? Are these lamins refractory to mitotic phosphorylation that in other cell types leads to lamina disassembly, and does lamina structure change as the chromosome segregate? And finally, why do these cells go through the trouble of undergoing this kind of mitosis? One of the striking features of this cell division is that following mitosis, the two daughter cells and their nuclei differ in size: the self-renewing neuroblast is the larger cell with a large nucleus, while the differentiating ganglion mother cell (GMC) is smaller and has a proportionately smaller nucleus. Furthermore, Roubinet et al found that the sibling nuclei also differed in histone modifications, and specifically histone H3 lysine 4 di-methylation (H3K4me2), which was more abundant in the GMC nucleus³. This was the case even if nuclear division occurred in the same cytoplasm (namely when cytokinesis was inhibited but nuclear division proceeded), suggesting that nuclear size may ultimately affect gene expression. Thus, in this system, regulating nuclear size may be critical for cell fate.

It has long been known that there is a correlation between nuclear size and cell size^7,8, but the mechanism that leads to this correlation is still not understood, and it is quite possible that different cell types abide by different rules. The field is currently divided on whether the amount of nuclear membrane ultimately determines nuclear size^9,10. Regardless, without membrane expansion nuclear size cannot increase after cell division or, in fungi, during mitosis. In fact, in the yeast Schizosaccharomyces japonicus, in which mitotic nuclear membrane expansion does not occur, the nuclear membrane ruptures to allow proper chromosome segregation¹¹. Roubinet et al showed that the difference between the two nuclei immediately after the fly neural stem cell asymmetric division was around 4.5-fold, and this difference increased to 37-fold as the neuroblast, but not the GMC, increased in both cell and nuclear size during telophase/G1³. What allowed nuclear membrane expansion in the neuroblast, but not in the GMC?

In principle, the membrane for an expanding nucleus could come from at least two, non-mutually exclusive, sources: membrane could be “pulled” from the ER¹² to which the outer nuclear membrane is connected (Figure 1, green arrows), or phospholipids can be synthesized locally by enzymes present on either of the nuclear membrane. That phospholipids must be synthesized for mitotic nuclear expansion to occur has been shown in budding yeast¹³, and the enzyme that synthesizes one of the main membrane phospholipids, phosphatidylcholine, localizes to the nuclear membrane during yeast exponential growth¹⁴. However, the relative contribution of phospholipid synthesis in the ER vs. the nuclear membrane to nuclear envelope expansion has been more difficult to ascertain. In favor of the ER as the source of membrane components for nuclear envelope expansion, Roubinet et al found a correlation between nuclear size and the amount of ER in the cell after the asymmetric neural stem cell division. Mukherjee et al¹⁵ reached a similar conclusion using sea urchin early embryos, but they argued that the only relevant ER pool is the peri-nuclear ER rather than the entire cellular ER. Furthermore, in fly neural stem cells under conditions where the nucleus divided asymmetrically but cytokinesis was delayed, the rate of nuclear membrane expansion was similar for two nuclei, indicating that nuclear size increase is not nucleus-autonomous, but depends on one or more cytoplasmic constituents³.

These observations are consistent with the ER as a source of membrane for nuclear expansion, by either providing the environment for phospholipid synthesis, or simply serving as a reservoir of readymade membrane. But how is the amount of membrane added to the nucleus regulated? Specifically, is there a mechanism that controls how much membrane the ER allocates to the nucleus? The ER is organized as an interconnected network of membrane sheets and tubules, but how much of the ER membrane is available for the nuclear membrane - if this is indeed how it works- is not known. Movement of membrane components from the ER into the nuclear membrane would have to overcome forces that keep the ER in this network configuration. One possibility is that import of proteins into the nucleus generates pressure on the nuclear membrane¹⁶ (Figure 1, orange arrows), causing the nuclear membrane to expand by drawing membrane from the ER. In this case, nuclear membrane expansion would be a passive outcome of nuclear import. Conversely, there may be mechanisms, still unknown, that regulate the flow of membrane from the ER to the nuclear envelope, for example by limiting the number of contacts between the nuclear membrane and the rest of the ER, by trapping the ER in a conformation that is less conducive for membrane flow to the nucleus¹², or perhaps even a more outlandish hypothetical mechanism that regulates membrane flow at site of ER/nuclear membrane junctions. It is also possible that membrane can be removed from the nucleus, as needed, although it isn’t clear what the signal for excess membrane might be. Indeed, when budding yeast cells are delayed in mitosis, their nuclear membrane continues to expand despite the block to chromosome segregation¹⁷. There are examples, however, of pathways that could serve to reduce nuclear membrane: the budding yeast phospholipid-diacylglycerol acyltransferases, Lro1, localizes to the nucleus and can promote the conversion of phospholipid-derived fatty acid into intranuclear lipid droplets¹⁸. Vesicle trafficking has also been implicated in promoting removal of excess nuclear membrane¹⁹ but the effect may be indirect by controlling lipid metabolism proteins²⁰. Whatever the mechanism, it will be interesting to determine whether and how nuclear membrane allocation is altered in hematopoietic and cancer cells.