In This Issue

By cloning a murine homologue to a yeast protein involved in endocytosis and characterizing its intracellular localization, Engqvist-Goldstein et al. (page 1503) have begun to uncover the molecular machinery linking endocytosis to the actin cytoskeleton in mammalian cells. Though the functional connection between the actin cytoskeleton and endocytosis has been firmly established in yeast, the endocytic role of actin has not been determined as clearly in mammalian cells.

After identifying a homologue to the yeast Sla2 gene in a murine EST database, Engqvist-Goldstein and colleagues cloned the corresponding cDNA and characterized the mouse protein, designated mHip1R. Sequence alignments and secondary structure predictions verify that mHip1R is a member of the Sla2 protein family, of which the yeast protein is suggested to link that actin cytoskeleton to endocytosis, and a human member binds to huntingtin, the protein that is mutated in Huntington disease. The talin-like domain of mHip1R binds to F-actin in vitro, and immunofluorescence shows that the protein colocalizes with F-actin and markers for receptor-mediated endocytosis. The team is now searching for additional functional parallels between the yeast and mammalian endocytic systems.

Beginning on page 1519, Sumi et al. present data showing that the kinase LIMK2 acts as a downstream effector of cytoskeletal rearrangements mediated by Rho and Cdc42, but not Rac. The results complement earlier findings showing that LIMK1, a kinase related to LIMK2, is a downstream effector of Rac, suggesting a model in which the activation of distinct effectors leads to the diverse effects of Rho-subfamily GTPases.

In searching for substrates of LIMKs, the researchers discovered that both LIMK1 and LIMK2 phosphorylate cofilin, a protein that depolymerizes actin filaments. Previous work had shown that LIMK1 acts on cofilin downstream of Rac, and the team reasoned that LIMK2 might be regulated by other members of the Rho subfamily. Wild-type LIMK2, but not a kinase-dead mutant, can phosphorylate cofilin and induce the formation of stress fibers and focal complexes, and expression of activated Rho and Cdc42, but not Rac, stimulates LIMK2 phosphorylation of cofilin. Conversely, expression of the kinase-dead mutant inhibits the Rho- and Cdc42-mediated induction of stress fibers and filopodia, but does not affect the Rac-induced formation of lamellipodia. The researchers suggest that LIMK1 and LIMK2 are downstream effectors of distinct Rho-subfamily GTPases. The team is now using knockout mice to study the role of LIMK2 in mammalian development.

In characterizing the stress granules (SGs) that form in mammalian cells in response to environmental stresses like heat shock, Kedersha et al. (page 1431) discovered that phosphorylation of the translation initiation factor eIF-2α is necessary and sufficient to induce SG formation. The team also found that the SG-associated RNA-binding proteins TIA-1 and TIAR act downstream of eIF-2α phosphorylation in forming SGs.

In earlier work, the same lab identified RNA-binding proteins, including TIA-1 and TIAR, that coaggregate with poly(A)⁺ RNA at mammalian SGs. Reasoning that the SGs were linked to translation control, the researchers examined the role of eIF-2α phosphorylation, the primary determinant of stress-induced translational arrest. A phospho-mimetic mutant of eIF-2α induces SG formation, while a non-phosphorylatable eIF-2α mutant prevents the formation of SGs. In addition, a TIA-1 mutant lacking its RNA-binding domains acts as a trans-dominant inhibitor of SG formation, suggesting that TIA-1 acts downstream of eIF-2α phosphorylation to sequester RNAs in the SGs. Paul Anderson, senior author on the paper, adds that TIA-1 and TIAR may also play a more general role in the cell, regulating the translation of specific mRNAs under conditions other than stress. The researchers are now trying to characterize that phenomenon and the molecular interactions between TIA-1, TIAR, and eIF-2α.

Starting on page 1561, Collins et al. demonstrate for the first time that a single adaptor protein can bifurcate proliferative and migratory signaling pathways. Shc, a protein composed of multiple protein–protein interaction domains, was known to play a role in both processes, but the new report shows that differential use of Shc's phosphotyrosine interacting domains determines whether the protein transduces a migratory or proliferative signal.

Under limiting growth factor conditions, Shc stimulates haptotactic cell migration without affecting anchorage-dependent proliferation, but in the presence of growth factors, Shc is required for DNA synthesis and loses its influence on cell migration. The researchers used mutant forms of Shc to determine that tyrosine phosphorylation is necessary for both activities, but the phosphotyrosine-interacting domains of the protein play distinct roles: the PTB domain regulates migration and the SH2 domain is selectively required for proliferation.

The results are consistent with a model in which the PTB domain targets Shc to areas of the membrane, such as integrin-containing focal contacts, where it can participate in the cell migration signaling pathway. When growth factors are present, Shc may be recruited to growth factor receptors and utilize its SH2 domain to transduce mitogenic signals.

Doria et al. (page 1379) found that Eps15 and Eps15R, proteins involved in endocytosis, are also involved in the Rev nuclear export pathway. Eps15 and Eps15R are part of a network of proteins that interact through EH domains and are involved in controlling endocytosis. EH-containing proteins were known to interact with Hrb and Hrb1, proteins in the Rev export pathway, but the new data demonstrate that Eps15 and Eps15R play functional roles in Rev-mediated export.

Using a reporter gene linked to a Rev-response element, the researchers were able to determine relative levels of activity in the Rev nucleocytosolic export pathway. Antisense inhibition of Eps15 reduces the activity of the Rev pathway, and cotransfection experiments show that Eps15 and Eps15R synergize with Hrb and Hrb1 to enhance the function of Rev in the export pathway. The synergistic effect requires the EH-mediated interaction between Eps15 and Hrb. The proteins appear to colocalize in the cytoplasm, an unexpected site of action for Hrb. One possibility is that the Hrb-Eps15–containing complex is involved in protecting proteins from degradation, a model the team is now testing.