Mature peroxisomes fractionate as a homogeneous species, but various subsets of peroxisomal proteins can be found in other biochemical fractions. Now Titorenko et al. (page 29) show that these subsets represent precursors in peroxisome formation. They characterize six distinct peroxisome species in the yeast Yarrowia lipolytica, from the most immature (P1) to the most mature (P6), and find that an early step in peroxisome formation involves vesicular fusion.

By pulse–chase, proteins appear first in the P1 and P2 fractions, before progressing sequentially through the other stages. P1 and P2 have distinct protein compositions, and they fuse in vitro to form a set of vesicles that appear identical, by several measures, to the P3 species formed in vivo.

As the peroxisomes mature, sequential import provides them with an increasing fraction of the proteins present in mature peroxisomes. The various species of peroxisomes are clearly distinct, suggesting that the vesicles are held up by a quality control step before rapidly transiting from one state to another. Senior author Richard Rachubinski says he does not know the reason for this complexity, but he is now looking at what mechanisms might regulate peroxisome vesicle fusion and conversion.

Muscles contain some very large proteins; on this much there is consensus. But the exact relationships between the large proteins remain controversial. The dialogue continues on page 101, with a report from Hakeda et al. about a fly protein called Kettin.

The mother of all muscle proteins is titin, weighing in at up to 3.7 Mda. Single titin molecules span half sarcomeres from the Z disc (where muscle sarcomeres join one another) to the M line (the midpoint of a sarcomere).

Hakeda et al. complete the sequence of the gene for the related but smaller Kettin protein, which was originally identified in insect flight muscle. Kettin’s predicted size is 540 kD. It has 35 of the repeats (an immunoglobulin domain followed by a spacer) seen in titin, but lacks titin’s kinase and PEVK elasticity domains. Flies lacking Kettin die as embryos; weaker kettin mutations cause movement failures as larvae or, in heterozygotes, flightlessness as adults.

Kettin is necessary for the formation and maintenance of the Z disc. The uniform spacer sizes in a central kettin region and the region’s total predicted length suggest that it may bind the repeated structure of actin filaments as Kettin spans the Z disc.

The 5′ end of the Kettin gene was also isolated by Machado et al. (Machado, C., C.E. Sunkel, and D.J. Andrew. 1998. J. Cell Biol. 141:321–333) as part of a proposed Drosophila titin gene (D-Titin) whose product is involved in both muscle function and chromosomal structure. The size and repeated structure of D-Titin has delayed complete cloning and mapping efforts, but Machado et al. found that probes from two regions (one shared with Kettin and one not) both identified a Mda-sized protein, and gave identical in situ patterns in embryonic muscle tissue. Thus Kettin and D-Titin may be different splice products of the same gene. Inconsistencies between the two groups remain, but the release of the complete Drosophila sequence should soon settle the issue.

When electrons leak out of the respiratory chain to form superoxide (O2), superoxide dismutase defuses the superoxide to produce oxygen and hydrogen peroxide (H2O2). But H2O2 still has some kick to it. On page 7, Czubryt et al. show that H2O2 inhibits nuclear import by activating the MAP kinase pathway.

Whereas free radicals like superoxide can inhibit insoluble components of the import apparatus, H2O2 works on a soluble factor. The effect is prevented by either a tyrosine kinase inhibitor or an inhibitor of the MAP kinase kinase MEK1; it can be reproduced by addition of an activated (phosphorylated) version of the MAP kinase ERK2. Reduced import after H2O2 treatment may be a consequence of the observed increase in cytoplasmic Ran protein, as nucleo-cytoplasmic cycling of Ran is required for general nuclear import. Perhaps, suggest the authors, the ERK2 pathway is regulating the GTP/GDP state of Ran. Conversion between these states is required for cycling between cytoplasm and nucleus.

H2O2 has now been implicated in signaling from some receptors, and it is elevated in conditions such as atherosclerosis and reperfusion after ischemia (when metabolism is ramped up after the return of a blood and oxygen supply). Inhibition of nuclear import may contribute to these pathological processes by increasing anything from apoptosis to cell proliferation; the next step will be to identify the critical transcription factors whose import may be altered.

The cell cycle has remained a somewhat neglected field for worm biologists, but on page 73, Lorson et al. show that LIN-5 is a spindle component involved in multiple mitotic processes. The protein is new both to worm researchers and the cell cycle field in general.

Worms mutant in lin-5 limp through embryogenesis with maternally supplied RNA and protein, but in larval divisions there is a failure of chromosome movements to and then from the metaphase plate. When the LIN-5 supply is obliterated early with RNA-mediated interference, spindle orientation and movement is defective as early as the first division, although chromosome movements only fail after two to four successful divisions.

LIN-5 is predicted to be a novel coiled-coil protein. Its localization at the centrosome requires microtubules; it is also localized to spindle microtubules and the cell cortex. LIN-5 is not required for centrosome duplication or spindle formation. Lorson et al. suggest that LIN-5 may anchor motor proteins necessary for spindle and chromosome movement, although subtle defects in spindle formation and spindle–cortex anchoring remain possible explanations for defects in lin-5 worms.

There is still no proof that a mitotic checkpoint exists in worms, as anti-spindle drugs are kept at bay by the worm’s cuticle. LIN-5 is now a candidate for participating in a checkpoint, as lin-5 mutants enter and exit mitosis on schedule, even in the absence of chromosome movement. Partial loss of function mutants show a delay in mitosis, possibly because residual LIN-5 transmits a mitotic checkpoint signal.

The link from cadherin adhesion molecules to actin, through α- and β-catenin, has been well defined. In contrast, p120 catenin has been seen as a weakly associated protein of unknown importance. Thoreson et al. show on page 189 that p120 is an abundant junctional component whose function is a prerequisite for cell compaction. The expression of E-cadherin is shown to be both necessary and sufficient to recruit p120 to the membrane, so the fraction of p120 at this location, >90%, can be presumed to be cadherin-associated.

Thoreson et al. use triple alanine mutations to narrow down the p120-binding site within the juxtamembrane domain of E-cadherin. When the site is mutated to uncouple p120 and E-cadherin, cells form contacts, but the apposition of cells is discontinuous, and adhesion assays show that simple pipetting disperses groups of cells.

The weakness of adhesion also blocks compaction. Normally, actin bundles in single cells insert into cadherin junctions and contract, thus pulling cells together in a tight ball. Senior author Albert Reynolds believes that, in the cells with mutant E-cadherin, compaction does not occur because p120 has not adequately carried out the prerequisite step of E-cadherin clustering. Barry Gumbiner and colleagues have shown that a large deletion around the p120-binding site prevents E-cadherin clustering. Reynolds is now investigating how p120 induces the clustering.