Life with Two Microtubule Motors

By eliminating motors that work in either antagonistic or redundant pairs, Cottingham et al. find that Saccharomyces cerevisiae can function with just two microtubule-based motors (page 335). Budding yeast does not require microtubules for protein trafficking, so the two motors define a minimal set for constructing and then using a mitotic spindle.

Cottingham et al. start by rounding out the roles for motors in nuclear positioning. The kinesins Kar3p and Kip3p work with a dynein (Dyn1p) to oppose another kinesin (Kip2p). The removal of Kip2p compensates for the loss of dynein plus either Kar3p or Kip3p; the remaining motor opposes a non-microtubule force (possibly actin-based streaming) that disrupts nuclear positioning.

But Kar3p and Kip3p also have an essential role in the nucleus, where they work together to promote spindle assembly. For Kar3p, at least, this function opposes the BimC class of kinesins (Cin8p and Kip1p). (The antagonism of a similar pair, the human Kar3p homologue HSET and the BimC motor Eg5, is described by Mountain et al. on page 351.)

The minimal set has one motor from each class, either Kar3p or Kip3p paired with Cin8p. An important function of Kar3p and Kip3p appears to be microtubule depolymerization, and indeed a strain with only Cin8p and Kar3ts can grow at high temperatures in the presence of a microtubule-depolymerizing drug. Senior author Andrew Hoyt says that such economy “is a little unexpected.” It suggests that a single BimC motor can build and elongate a spindle, and that Kar3p is sufficient both for spindle positioning and for microtubule shortening, which is required for spindle functions such as anaphase A.

Dynactin as a Centrosome Anchor

Dynactin is an ∼1-MDa assembly of proteins that links dynein to different cargoes. On page 307, Eckley et al. report the cloning of the final four dynactin subunits, which contain several possible cargo-binding motifs. And on page 321, Quintyne et al. propose that dynactin has a dynein-independent function of anchoring microtubules at the centrosome.

Quintyne et al. overexpress various subunits and fragments of dynactin. This causes a loss of focus at the centrosome and, in most cases, fragmentation of the Golgi apparatus. Loss of microtubule organization is an expected result of interfering with dynein, as dynein may retrieve microtubules and transport centrosomal components to the centrosome.

Overexpression of two dynactin constructs, designated as class C, suggest an additional function. The overexpression has no effect on the Golgi, so dynein and the dynactin–cargo linkage are still functional. But microtubules are disorganized. The class C constructs may be selectively disrupting a microtubule-anchoring function of dynactin. This anchoring would counteract ejection forces generated by any dynein at the centrosome, whereas relaxation of the anchoring could lead to the more dispersed microtubule organization seen in neurons and epithelial cells.

Translation regulatory factors have been studied in flies and plants, but the report by Kwon et al. on page 247 is the first description of a mammalian translation enhancer.

The relevant 21 nucleotides, the RNA trafficking sequence (RTS), was first identified as being necessary and sufficient for transport of the myelin basic protein (MBP) RNA to the oligodendrocyte periphery. Kwon et al. find that addition of the RTS to an mRNA encoding green fluorescent protein (GFP) increases the amount of translation up to 10-fold. Binding of the hnRNP A2 protein to the RTS is necessary for transport of MBP RNA, and Kwon et al. show that antisense treatment to suppress hnRNP A2 expression reduces RNA transport and abolishes translation upregulation. The RTS has no effect on the rate of translation in a reticulocyte lysate unless hnRNP A2 is also added.

Translation is believed to be turned off during RNA transport, so hnRNP A2 may be capable of switching between several roles depending on the available partner proteins. The “working hypothesis” of senior author John Carson is that “this is going to turn out to be an exciting class of navigator proteins that associates with the specific RNA in the nucleus, and remains associated during subsequent RNA trafficking, sequentially mediating nuclear export, transport, and translation.”

The function of cyclin A in mitosis has been largely ignored in favor of its role in DNA replication. On page 295, Furuno et al. show that cyclin A/Cdk2 controls entry into prophase of mitosis, before cyclin B/Cdk1 takes over at the end of prophase.

Soon after Furuno et al. inject G2 cells with cyclin A/Cdk2, the cells round up and condense their chromosomes, thus skipping G2 in favor of prophase entry. But injection of a cyclin A inhibitor, a fragment of p21, causes early prophase cells to revert to a G2 state. This mimics the response of cells that are irradiated in early prophase, suggesting that p21 inhibition of cyclin A may be the mechanism responsible for this prophase checkpoint. Only when cyclin B enters the nucleus, near the end of prophase, is the cell irreversibly committed to mitosis.

The amyloid precursor protein (APP) is cleaved by α-, β-, and then γ-secretase. Normally, cleavage by γ-secretase produces βA4(1-40), but mutations in presenilin 1 (PS1) associated with familial Alzheimer's disease result in the production of the amyloidogenic βA4(1-42).

This gain of deleterious function has been taken by some to mean that PS1 is γ-secretase, but this leaves a geographical paradox. The cleavage events that produce βA4(1-40) have been localized to the endocytic compartment and plasma membrane, whereas PS1 has been localized, in some studies, to more proximal compartments. Other studies using overexpression have suggested a wider distribution of PS1.

On page 277, Annaert et al. confirm, using immunofluorescence and fractionation of untransfected cells, that the furthest that endogenous PS1 gets into the secretory compartment is the cis-Golgi. Adding an ER retention signal to APP increases the production of βA4(1-42), although how these fragments are first processed by the distal α- and β-secretase is unclear. “The hypothesis that PS1 is γ-secretase may be too simple to explain all the effects,” says first author Wim Annaert. He is therefore looking for PS1 effects on trafficking between the ER and Golgi, which may affect the functioning of the real γ-secretase.

By William A. Wells, 1095 Market St. #516, San Francisco, CA 94103. E-mail: [email protected]