The protooncogene Wnt-1 is a secreted glycoprotein capable of both acting as a growth factor and modulating gene transcription and thus cell fate. Its link to tumorigenesis has been assumed to involve either or both of these functions. But on page 87, Chen et al. report that Wnt-1 can increase cell survival by reducing apoptosis in response to chemotherapeutic drugs.

Chen et al. find that cells producing Wnt-1 show less than half the amount of apoptotic death seen in wild-type cells in response to treatment with the antimicrotubule chemotherapeutic drugs vincristine and vinblastine. Wnt-1 signaling normally inhibits the constitutive phosphorylation and degradation of β-catenin, allowing the β-catenin to enter the nucleus and activate gene expression by interaction with members of the Lef/Tcf family of transcription factors. In the current paper, full levels of apoptosis are restored in the presence of a dominant-negative Tcf, but various other signaling pathways that are stress-induced or potentially apoptosis-inhibiting are not affected by Wnt-1 expression.

Candidate downstream targets of Lef/Tcf that would affect apoptosis, such as Fas, Bcl, and IAP proteins, are not turned on by Wnt-1 expression. Identifying the relevant Lef/Tcf target or targets is crucial. But just as important will be coming up with drugs to inhibit the Lef/Tcf pathway. Such drugs should eliminate the apoptosis-resistance of tumor cells that have abnormally activated Wnt signaling, and thus improve the effectiveness of existing chemotherapies.

Anaphase is initiated in yeast when the anaphase-promoting complex (APC) targets Pds1p for destruction, thus freeing up Pds1p's binding partner Esp1p so that it can cleave the cohesin protein Scc1p. This relieves sister-chromatid cohesion, and the chromosome separation of anaphase A can proceed.

Anaphase B, the lengthening of the spindle, could also result directly from this release of cohesive tension. But on page 27, Jensen et al. report that Esp1p is required even in the absence of sister cohesion, for some function that induces spindle elongation. Yeast lacking both Scc1p and Esp1p function show premature sister separation (because Scc1p is lacking), but the spindle does not elongate (because Esp1p is lacking).

Jensen et al. find that Esp1p moves from the chromatin to the spindle around the time of the metaphase to anaphase transition. Pds1p is needed for both the nuclear and, independently, spindle localization of Esp1p, although the viability of a strain lacking all Pds1p suggests that some Esp1p can function without the help of Pds1p. The movement of Esp1p from chromatin to spindle may also be related to an APC-induced proteolysis event.

Much of the spindle-localized Esp1p is at the midzone, where microtubules from opposite poles interdigitate. Motors acting on these interdigitating microtubules may drive spindle-lengthening during anaphase B, and Esp1p may regulate this process. The first experiment will be to test whether Esp1p's protease activity is necessary for its spindle function. If so, the search for Esp1p's spindle substrate will be on.

Two papers in this issue add to our understanding of nuclear export. Holaska et al. (page 127) find that calreticulin, a chaperone previously reported to function in the endoplasmic reticulum (ER), also acts in the nucleoplasm as a nuclear export receptor for steroid hormone receptors. Black et al. look at a later step in the export pathway in their report on NXT1, which they propose is involved in delivering export complexes to a site on the cytoplasmic side of the nuclear pore complex (NPC) where receptor and substrate can be released (page 141).

Holaska et al. study export of a model substrate, protein kinase inhibitor (PKI), from nuclei of permeabilized cells. Export continues in the absence of the general export receptor, Crm1, and purification of the remaining activity yields calreticulin. Holaska et al. speculate that calreticulin may be retrotranslocated out of the ER, as are certain bacterial toxins. Calreticulin has a 400-fold higher affinity for nuclear export signals than for the folding intermediates that it encounters in the ER.

Calreticulin is known to block the binding of several nuclear hormone receptors to DNA, and Holaska et al. find that calreticulin (but not Crm1) can promote nuclear export of the glucocorticoid receptor (GR). The in vivo export of GR absolutely requires calreticulin. (This is an interesting twist on GR nuclear import, which requires Hsp90 binding to GR to induce a conformation capable of drug binding and thus nuclear entry.) The DNA-binding domain of GR acts as a calreticulin-dependent nuclear export signal.

At limiting concentrations of either Crm1 or calreticulin, export is dependent on the GTP-bound form of the GTPase Ran. In the cytoplasm, hydrolysis of GTP to yield RanGDP is thought to liberate transported proteins.

Black et al. suggest that NXT1 may transport export complexes to a site where this hydrolysis can occur. They find that Crm1 and RanGTP cooperate to export protein to a NPC site that is accessible to cytoplasmic antibodies. NXT1 added at this stage has no effect, but NXT1 included from the start of the export reaction leads to displacement of the complex from the NPC. NXT1 does not enhance the binding of Crm1 to the nuclear export substrate, or of RanGTP to Crm1, so Black et al. suggest that it is functioning at a later stage to deliver the complex to a site where RanBP1 can act to expose RanGTP to the actions of cytoplasmic RanGAP. The resulting GTP hydrolysis should free both receptor and substrate. This theory may be strengthened by future electron microscopy to localize the proteins acting at different stages.

On page 165, Garin et al. identify 140 proteins from phagosomes, adding to an earlier proteomics description in The Journal of the nuclear pore complex (Rout et al., 148:635–652). These focused proteomics efforts allow the identification of low abundance proteins, and simultaneously give protein localization information.

Confidence in the proteins' localization is high in this case because the phagosomes were formed after engulfment of latex beads, which allow the phagosomes to be purified well away from any contaminating organelles. Some proteins known to be associated with other organelles are, however, amongst those identified. Contamination is not a likely explanation because the most abundant proteins in these other organelles are not present in the phagosome sample.

Garin et al. suggest other explanations. Proteins from the ER, for example, may arrive at the phagosome when ER membranes are used for membrane homeostasis. Phagosome membranes returning to the plasma membrane may be replenished using ER membranes.

The lipid raft protein flotillin-1 is another surprise. Although it could just be coming along for the ride during phagocytosis, an increase in its abundance during phagosome maturation suggests that lipid rafts and membrane specialization may have a role in phagosome function. The abundance of different hydrolases also varies during phagosome maturation, suggesting a sequential maturation process.

The 140 proteins include a number of apoptosis-related proteins. The cell may monitor the state of affairs inside the phagosome, and induce apoptosis of the phagocytic cell if the phagocytosed organism is not being brought under control. This proposed apoptosis control is just one of the new aspects of phagosome biology revealed by this study.

By William A. Wells, 1012 Sanchez Street, San Francisco, CA 94114. E-mail: