The PDZ domain is one of the most common protein-protein interaction modules in metazoans and appears to have an important role in organization of cellular signal transduction. As part of our continuing Signal Transduction and Cellular Organization series, Baruch Harris and Wendell Lim review the structure and function of PDZ domains, highlighting their roles in supramolecular signalling assemblies. PDZ domains recognize short peptide motifs present at the C-termini of certain proteins and occasionally structurally related internal motifs. Specificity appears to be provided primarily by the first and third residues (P0 and P-2) in the 5-residue PDZ-binding motif, which interact with two distinct pockets in the peptide-binding groove of the PDZ domain. Significantly, multiple PDZ domains are often present within the same protein, and this allows assembly of multimolecular complexes. The Drosophila protein INAD, for example, contains five PDZ domains and appears to serve as an essential scaffold for assembly of signalling molecules involved in G-protein-receptor-coupled phototransduction in the fly eye.

A-kinase-anchoring proteins (AKAPs) are scaffold proteins that bind to cAMP-dependent kinase (PKA) and target it to particular subcellular locations and substrates. AKAPs can thus compartmentalize PKA signalling, localizing the enzyme to structures such as the centrosome and mitotic chromosomes. Two articles in this issue now provide the first evidence for regulated interactions between PKA and AKAPs, revealing that phosphorylation of the PKA regulatory subunit RIIα at Thr54 regulates PKA-AKAP interactions at mitosis.

On p. 3243, Cathrine Carlson and co-workers show that phosphorylation of Thr54 by Cdc2 at mitosis releases RIIα from centrosomes by disrupting its association with AKAP450. They demonstrate that wild-type RIIα is bound to the centrosome during interphase but released at mitosis or following treatment with Cdc2 in vitro. A non-phosphorylatable Thr54Glu mutant, by contrast, remains attached to the centrosome throughout the cell cycle and cannot be released by Cdc2. The authors conclude that it is the RIIα-AKAP450 interaction that is disrupted, since Cdc2 reduces the affinity of RIIα for AKAP450 in vitro. Significantly, Carlson and co-workers also find that renucleation of microtubules from the centrosome is compromised in cells expressing Thr54Glu RIIα, which indicates that dissociation of PKA at mitosis is important for spindle formation.

On p. 3255, Philippe Collas and co-workers report that phosphorylation of Thr54 of RIIα also regulates its association with the chromatin-bound AKAP, AKAP95. They show that wild-type RIIα (but not the Thr54Glu mutant) associates with chromatin at mitosis by binding to AKAP95, which correlates with Thr54 phosphorylation. The authors also find that disruption of RIIα-AKAP95 interactions or depletion of RIIα from chromatin causes premature chromosome decondensation and, furthermore, that pseudophosphorylated RIIα can promote re-condensation of prematurely decondensed chromosomes. Phosphorylation of Thr54 thus appears to function as a cell-cycle-dependent molecular switch that controls not only spindle assembly but also chromatin dynamics at mitosis.

The small GTPase Ran is central to nuclear import and export, directly interacting with nuclear transport receptors such as importin β and modulating their association with cargo. Recent work indicates that Ran is also important for mitotic spindle formation and nuclear envelope assembly. But what are its targets in these processes? Markus Künzler and Ed Hurt discuss work that is beginning to shed light on the proteins with which Ran interacts. In addition to the factors that regulate its activity (RanGAP1/Rna1p and RCC1/Prp20p), Ran interacts with several other proteins. These include NTF2 (which regulates Ran nuclear import), importin-β-family proteins, nuclear pore components, RanBP1-RanBP3, Mog1, Yrb30p and RanBPM. A requirement for chromatin-bound RCC1 appears to be common to the three known processes in which Ran is involved. Intriguingly, importin β is a key Ran target during spindle formation, as well as nuclear import. Ran might therefore use not only a common regulator but also a common target to perform its various functions.

Signalling by integrins plays an important role in regulation of cell adhesion. A wide variety of signalling molecules are known to be activated by integrins, but only recently has PKA - typically associated with G-protein-coupled receptor signalling - been implicated. John Whittard and Steven Akiyama have investigated the role of PKA in regulation of adhesion by integrin β1. They show that treatment of Ht-1080 cells with the anti-β1-integrin antibody 12G10 increases cAMP levels and PKA activity and that 12G10-induced cell-cell adhesion is blocked by PKA-specific inhibitors (but not inhibitors of PKC or tyrosine kinases). The authors also find that the PKA type II regulatory subunit translocates from the cytoplasm to 12G10-stimulated integrins. Moreover, they demonstrate that integrin clustering and F-actin polymerization, both of which are required for cell-cell adhesion, are dependent on PKA. Whittard and Steven Akiyama conclude that PKA is critical for cell adhesion signalling by β1 integrin and hypothesize that integrin β1 resides in a signalling complex coupled to the Gαs protein, which activates the cAMP/PKA pathway.

Lamin-associated polypeptide 2β (LAP2β) is a highly conserved inner nuclear membrane (INM) protein that is important for assembly of the nuclear envelope (NE). It can also bind to chromatin and chromatin-associated factors and might therefore have a broader role in nuclear organization. Amos Simon and co-workers now reveal another facet of LAP2β function: it can act as a transcriptional repressor! They show that LAP2β interacts with a novel nuclear matrix component, mGCL (the mammalian homologue of the Drosophila protein Germ-cell-less), and that the LAP2β-mGCL complex can repress transcriptional activity of the transcription factor E2F-DP. In fact, LAP2β is able to repress E2F-DP activity even in the absence of mGCL. Simon and co-workers conclude that LAP2β is part of a novel spatial mechanism for transcriptional repression. LAP2β could, for example, sequester E2F-DP at the NE away from target promoters; alternatively, it might recruit DNA-bound E2F-DP to the NE before stimulating formation of a repressive chromatin structure.