Arrestin interactions with G protein-coupled receptors. Direct binding studies of wild type and mutant arrestins with rhodopsin, beta 2-adrenergic, and m2 muscarinic cholinergic receptors.

Arrestins play an important role in quenching signal transduction initiated by G protein-coupled receptors. To explore the specificity of arrestin-receptor interaction, we have characterized the ability of various wild-type arrestins to bind to rhodopsin, the beta 2-adrenergic receptor (beta 2AR), and the m2 muscarinic cholinergic receptor (m2 mAChR). Visual arrestin was found to be the most selective arrestin since it discriminated best between the three different receptors tested (highest binding to rhodopsin) as well as between the phosphorylation and activation state of the receptor (> 10-fold higher binding to the phosphorylated light-activated form of rhodopsin compared to any other form of rhodopsin). While beta-arrestin and arrestin 3 were also found to preferentially bind to the phosphorylated activated form of a given receptor, they only modestly discriminated among the three receptors tested. To explore the structural characteristics important in arrestin function, we constructed a series of truncated and chimeric arrestins. Analysis of the binding characteristics of the various mutant arrestins suggests a common molecular mechanism involved in determining receptor binding selectivity. Structural elements that contribute to arrestin binding include: 1) a C-terminal acidic region that serves a regulatory role in controlling arrestin binding selectivity toward the phosphorylated and activated form of a receptor, without directly participating in receptor interaction; 2) a basic N-terminal domain that directly participates in receptor interaction and appears to serve a regulatory role via intramolecular interaction with the C-terminal acidic region; and 3) two centrally localized domains that are directly involved in determining receptor binding specificity and selectivity. A comparative structure-function model of all arrestins and a kinetic model of beta-arrestin and arrestin 3 interaction with receptors are proposed.

Polypeptide variants of beta-arrestin and arrestin3.

Retinal arrestin (S-antigen) inactivates the phototransduction cascade by binding to light-activated phosphorylated rhodopsin and thereby "arresting" coupling to the G protein transducin. beta-Arrestin (beta arr), a ubiquitous arrestin homolog, acts analogously to desensitize the beta 2-adrenergic receptor by disrupting Gs receptor interaction. In an attempt to identify additional "arrestins" which might regulate the multitude of G protein-coupled receptors, we have isolated two bovine brain cDNAs which encode polypeptide variants of an arrestin homolog which we have designated arrestin3 (arr3). The open reading frames of these two cDNAs are identical except that the long form, arr3L, contains an 11-amino-acid insert between residues 361 and 362. Arr3 is more closely related to bovine beta arr (78% identity) than to bovine visual arrestin (56% identity). Polymerase chain reaction amplification of RNA and immunoblotting of lysates with an arr3-specific antibody suggest that the short form, arr3S, is the major form of arr3 in all bovine tissues and that it is most abundant in the spleen. Furthermore, polymerase chain reaction amplification of beta arr mRNA indicates that in several tissues (lung, liver, spleen, and pituitary), the major form of beta arr lacks 8 amino acids which are present in brain beta arr. Immunoblotting with an antibody which recognizes beta arr and arr3 with equal sensitivity demonstrates that beta arr (either the long or the short polypeptide) is the major arrestin in all (non-photoreceptor bearing) tissues examined. These observations suggest that in some tissues, as many as four arrestin homolog variants may play a role in the regulation of G protein-coupled receptors.

O-linked glycoproteins of the nuclear pore complex interact with a cytosolic factor required for nuclear protein import.

Mediated import of proteins into the nucleus requires cytosolic factors and can be blocked by reagents that bind to O-linked glycoproteins of the nuclear pore complex. To investigate whether a cytosolic transport factor directly interacts with these glycoproteins, O-linked glycoproteins from rat liver nuclear envelopes were immobilized on Sepharose beads via wheat germ agglutinin or specific antibodies. When rabbit reticulocyte lysate (which provides cytosolic factors required for in vitro nuclear import) was incubated with the immobilized glycoproteins, the cytosol was found to be inactivated by up to 80% in its ability to support mediated protein import in permeabilized mammalian cells. Inactivation of the import capacity of cytosol, which was specifically attributable to the glycoproteins, involves stoichiometric interactions and is likely to involve binding and depletion of a required factor from the cytosol. This factor is distinct from an N-ethylmaleimide-sensitive receptor for nuclear localization sequences characterized recently since it is insensitive to N-ethylmaleimide. Cytosol inactivation is suggested to be caused by at least two proteins of the glycoprotein fraction, although substantial capacity for inactivation can be attributed to protein bound by the RL11 antibody, consisting predominantly of a 180-kD glycosylated polypeptide. Considered together, these experiments identify a novel cytosolic factor required for nuclear protein import that directly interacts with O-linked glycoproteins of the pore complex, and provide a specific assay for isolation of this component.