In this issue

Integrins at the cell surface play a critical role in control of cell proliferation, transmitting adhesion-dependent signals without which many cells undergo G1 phase cell-cycle arrest or even apoptosis. A multitude of signalling pathways can be activated by integrins. However, whether all of these are important for cell cycle control - rather than cell migration or cytoskeletal rearrangement, for example - has remained unclear. In this context, Martin Schwartz and Richard Assoian discuss the roles of integrin signalling pathways that converge on G1 cyclins and cyclin-dependent kinases(CDKs). Recent work, for instance, indicates that increased expression and stability of cyclin D1 depend on sustained activation of ERK and its translocation to the nucleus and that ERK is downstream of focal adhesion kinase (FAK) in this pathway. Activation of PI 3-kinase and Rac following integrin engagement also appears to be linked to cyclin D protein levels. Schwartz and Assoian propose that the integrin—ERK—cyclin-D1 pathway represents a paradigm for signalling mechanisms underlying anchorage-dependent growth, which only a decade ago were a black box.

Members of the SWI/SNF family form ATP-dependent chromatin-remodelling complexes that play important roles in transcription, DNA replication, DNA repair and recombination. These complexes appear to modulate histone-DNA interactions and thereby facilitate interaction of other proteins with their binding sites on DNA. Gernot Längst and Peter Becker discuss the functions of ISWI - one of a subfamily of SWI2/SNF2-related proteins. ISWI powers several nucleosome-remodelling machines, including NURF and CHRAC, and has homologues in yeast, Drosophila and mammals. These remodelling complexes appear to be able to reposition nucleosomes by facilitating `sliding' of histone octamers. This can either improve the regularity of nucleosome arrays (and thereby facilitate chromatin assembly) or reposition them to allow proteins such as transcription factors access to the DNA. The authors discuss the mechanism of ISWI-dependent nucleosome sliding in the context of three different models: spooling, twisting and bulging.

Intermediate filament (IF) proteins are encoded by a large multigene family. This allows metazoans to generate a host of structurally related IFs that have functional characteristics appropriate to the cell type/compartment in which they exist. These include the type I/II IFs (keratins), type III IFs(e.g. desmin and vimentin), type IV IFs (e.g. neurofilament proteins), type V IFs (nuclear lamins), and eye lens proteins. Michael Hesse, Thomas Magin and Klaus Weber have searched the human genome sequence for IF genes, focusing on keratin genes in particular. They have identified 65 IF genes, of which 49 encode keratins. The latter all map to the two known keratin gene clusters on chromosomes 12 and 17, whereas the remaining genes are not clustered. Interestingly, the authors detected a large number of keratin pseudogenes and gene fragments, which are spread throughout the genome and in 90% of cases are keratin 8/18 related. Hesse and co-workers' analysis should prove particularly useful for studies of fragility syndromes and other diseases associated with mutations in IF genes.

The cyclin-dependent kinase (CDK) family comprises a group of Ser/Thr kinases related to the cell cycle regulator Cdc2. Although some CDKs have roles in cell cycle regulation, many have other functions. CDK7, CDK8 and CDK9, for example, are implicated in control of transcription through phosphorylation of the RNA polymerase II (Pol II) C-terminal domain (CTD). Jonathon Pines and co-workers have identified and characterized a novel human CDK - CrkRS - which they show is ubiquitously expressed and evolutionarily conserved. CrkRS is the largest CDK-related kinase identified to date (1490 residues) and contains an arginine/serinerich (RS) domain similar to that found in spliceosome components. The authors find that the kinase co-localizes with splicing factors and Pol II in nuclear speckles. Furthermore, they demonstrate that it can phosphorylate both the splicing factor ASF and the Pol II CTD. Their findings thus raise the exciting possibility that CrkRS is a novel, conserved CTD kinase that directly couples transcription and RNA splicing.

Tumour invasion and metastasis depend on interactions between tumour cells and the extracellular matrix. These interactions are mediated by cell-surface integrins, which are linked to signalling pathways that regulate cytoskeletal activity, cell migration and cell proliferation. Integrin αvβ3, in particular, has been linked with invasion by several types of malignant tumour. Dan Ramos and co-workers have investigated the role of this integrin in metastasis of melanoma cell lines. They show that cells that expressβ3 integrin are highly metastatic in comparison with those that do not and that antisense β3 integrin constructs suppress metastatic colonization. Furthermore, they demonstrate that the level of phosphorylation of focal adhesion kinase (FAK) at Tyr397 is higher in the β3-expressing cells. FAK-related non-kinase (FRNK), which functions as a dominant negative FAK, can suppress this FAK phosphorylation as well as tumour invasion. Similarly, expression of activated SRC increases invasiveness, but expression of a kinase-dead SRC mutant inhibits it. Ramos and co-workers therefore conclude that integrin αvβ3 promotes tumour metastasis by activating FAK and SRC.

In seeking to provide good mentoring for graduate students and post-docs,do we sometimes fail to teach them to think for themselves? Sabretooth, a new contributor to this column, is concerned that modern mentoring is not giving young scientists the chance to make mistakes and learn from them. In a memo to the members of his lab, he urges his students and post-docs to try to solve problems themselves, since he won't be around when they set up their own labs.