Role of scaffolds in MAP kinase pathway specificity revealed by custom design of pathway-dedicated signaling proteins.

BACKGROUND: Signal transduction pathways with shared components must be insulated from each other to avoid the inappropriate activation of multiple pathways by a single stimulus. Scaffold proteins are thought to contribute to this specificity by binding select substrates. RESULTS: We have studied the ability of scaffold proteins to influence signaling by the yeast kinase Ste11, a MAPKKK molecule that participates in three distinct MAP kinase pathways: mating, filamentation, and HOG. We used protein fusions to force Ste11 to associate preferentially with a subset of its possible binding partners in vivo, including Ste5, Ste7, and Pbs2. Signaling became confined to a particular pathway when Ste11 was covalently attached to these scaffolds or substrates. This pathway bias was conferred upon both stimulus-activated and constitutively active forms of Ste11. We also used membrane-targeted derivatives of the mating pathway scaffold, Ste5, to show that stimulus-independent signaling initiated by this scaffold remained pathway specific. Finally, we demonstrate that loss of pathway insulation has a negative physiological consequence, as nonspecific activation of both the HOG and mating pathways interfered with proper execution of the mating pathway. CONCLUSIONS: The signaling properties of these kinase fusions support a model in which scaffold proteins dictate substrate choice and promote pathway specificity by presenting preferred substrates in high local concentration. Furthermore, insulation is inherent to scaffold-mediated signaling and does not require that signaling be initiated by pathway-specific stimuli or activator proteins. Our results give insight into the mechanisms and physiological importance of pathway insulation and provide a foundation for the design of customized signaling proteins.

MAP kinases bite back.

Recent studies in a model system challenge our understanding of how signal transmission through a MAP kinase cascade proceeds and how signaling specificity may be achieved.

Role of Cdc42p in pheromone-stimulated signal transduction in Saccharomyces cerevisiae.

CDC42 encodes a highly conserved GTPase of the Rho family that is best known for its role in regulating cell polarity and actin organization. In addition, various studies of both yeast and mammalian cells have suggested that Cdc42p, through its interaction with p21-activated kinases (PAKs), plays a role in signaling pathways that regulate target gene transcription. However, recent studies of the yeast pheromone response pathway suggested that prior results with temperature-sensitive cdc42 mutants were misleading and that Cdc42p and the Cdc42p-PAK interaction are not involved in signaling. To clarify this issue, we have identified and characterized novel viable pheromone-resistant cdc42 alleles that retain the ability to perform polarity-related functions. Mutation of the Cdc42p residue Val36 or Tyr40 caused defects in pheromone signaling and in the localization of the Ste20p PAK in vivo and affected binding to the Ste20p Cdc42p-Rac interactive binding (CRIB) domain in vitro. Epistasis analysis suggested that they affect the signaling step at which Ste20p acts, and overproduction of Ste20p rescued the defect. These results suggest that Cdc42p is in fact required for pheromone response and that interaction with the PAK Ste20p is critical for that role. Furthermore, the ste20DeltaCRIB allele, previously used to disrupt the Cdc42p-Ste20p interaction, behaved as an activated allele, largely bypassing the signaling defect of the cdc42 mutants. Additional observations lead us to suggest that Cdc42p collaborates with the SH3-domain protein Bem1p to facilitate signal transduction, possibly by providing a cell surface scaffold that aids in the local concentration of signaling kinases, thus promoting activation of a mitogen-activated protein kinase cascade by Ste20p.

The role of Far1p in linking the heterotrimeric G protein to polarity establishment proteins during yeast mating.

Heterotrimeric guanosine triphosphate (GTP)-binding proteins (G proteins) determine tissue and cell polarity in a variety of organisms. In yeast, cells orient polarized growth toward the mating partner along a pheromone gradient by a mechanism that requires Far1p and Cdc24p. Far1p bound Gbetagamma and interacted with polarity establishment proteins, which organize the actin cytoskeleton. Cells containing mutated Far1p unable to bind Gbetagamma or polarity establishment proteins were defective for orienting growth toward their mating partner. In response to pheromones, Far1p moves from the nucleus to the cytoplasm. Thus, Far1p functions as an adaptor that recruits polarity establishment proteins to the site of extracellular signaling marked by Gbetagamma to polarize assembly of the cytoskeleton in a morphogenetic gradient.

Membrane recruitment of the kinase cascade scaffold protein Ste5 by the Gbetagamma complex underlies activation of the yeast pheromone response pathway.

In the Saccharomyces cerevisiae pheromone response pathway, the Gbetagamma complex activates downstream responses by an unknown mechanism involving a MAP kinase cascade, the PAK-like kinase Ste20, and a Rho family GTPase, Cdc42. Here we show that Gbetagamma must remain membrane-associated after release from Galpha to activate the downstream pathway. We also show that pheromone stimulates translocation of the kinase cascade scaffold protein Ste5 to the cell surface. This recruitment requires Gbetagamma function and the Gbetagamma-binding domain of Ste5, but not the kinases downstream of Gbetagamma, suggesting that it is mediated by Gbetagamma itself. Furthermore, this event has functional significance, as artificial targeting of Ste5 to the plasma membrane, but not intracellular membranes, activates the pathway in the absence of pheromone or Gbetagamma. Remarkably, although independent of Gbetagamma, activation by membrane-targeted Ste5 requires Ste20, Cdc42, and Cdc24, indicating that their participation in this pathway does not require them to be activated by Gbetagamma. Thus, membrane recruitment of Ste5 defines a molecular activity for Gbetagamma. Moreover, our results suggest that this event promotes kinase cascade activation by delivering the Ste5-associated kinases to the cell surface kinase Ste20, whose function may depend on Cdc42 and Cdc24.

AKR1 encodes a candidate effector of the G beta gamma complex in the Saccharomyces cerevisiae pheromone response pathway and contributes to control of both cell shape and signal transduction.

Mating pheromones of Saccharomyces cerevisiae control both signal transduction events and changes in cell shape. The G beta gamma complex of the pheromone receptor-coupled G protein activates the signal transduction pathway, leading to transcriptional induction and cell cycle arrest, but how pheromone-dependent signalling leads to cell shape changes is unclear. We used a two-hybrid system to search for proteins that interact with the G beta gamma complex and that might be involved in cell shape changes. We identified the ankyrin repeat-containing protein Akr1p and show here that it interacts with the free G beta gamma complex. This interaction may be regulated by pheromone, since Akr1p is excluded from the G alpha beta gamma heterotrimer. Both haploid and diploid cells lacking Akr1p grow slowly and develop deformed buds or projections, suggesting that this protein participates in the control of cell shape. In addition, Akr1p has a negative influence on the pheromone response pathway. Epistasis analysis demonstrates that this negative effect does not act on the G beta gamma complex but instead affects the kinase cascade downstream of G beta gamma, so that the kinase Ste20p and components downstream of Ste20p (e.g., Ste11p and Ste7p) are partially activated in cells lacking Akr1p. Although the elevated signalling is eliminated by deletion of Ste20p (or components downstream of Ste20p), the growth and morphological abnormalities of cells lacking Akr1p are not rescued by deletion of any of the known pheromone response pathway components. We therefore propose that Akr1p negatively affects the activity of a protein that both controls cell shape and contributes to the pheromone response pathway upstream of Ste20p but downstream of G beta gamma. Specifically, because recent evidence suggests that Bem1p, Cdc24p, and Cdc42p can act in the pheromone response pathway, we suggest that Akr1p affects the functions of these proteins, by preventing them from activating mating-specific targets including the pheromone-responsive kinase cascade, until G beta gamma is activated by pheromone.

Saccharomyces cerevisiae cells execute a default pathway to select a mate in the absence of pheromone gradients.

During conjugation, haploid S. cerevisiae cells find one another by polarizing their growth toward each other along gradients of pheromone (chemotropism). We demonstrate that yeast cells exhibit a second mating behavior: when their receptors are saturated with pheromone, wild-type a cells execute a default pathway and select a mate at random. These matings are less efficient than chemotropic matings, are induced by the same dose of pheromone that induces shmoo formation, and appear to use a site near the incipient bud site for polarization. We show that the SPA2 gene is specifically required for the default pathway: spa2 delta mutants cannot mate if pheromone concentrations are high and gradients are absent, but can mate if gradients are present. ste2 delta, sst2 delta, and far1 delta mutants are chemotropism-defective and therefore must choose a mate by using a default pathway; consistent with this deduction, these strains require SPA2 to mate. In addition, our results suggest that far1 mutants are chemotropism-defective because their mating polarity is fixed at the incipient bud site, suggesting that the FAR1 gene is required for inhibiting the use of the incipient bud site during chemotropic mating. These observations reveal a molecular relationship between the mating and budding polarity pathways.

Simian virus 40 minichromosomes as targets for retroviral integration in vivo.

We present a method for studying multiple retroviral integration events into a small DNA target in vivo. Episomal simian virus 40 (SV40) genomes established by infection of CV-1 cells served as integration targets during subsequent infection with murine leukemia virus (MLV). Using a PCR-based assay for the abundance and distribution of integration events, nonrandom integration of MLV DNA into SV40 DNA is detectable as early as 4 hr and reaches a maximum level by 8 hr after MLV infection. The level of integration but not the distribution of integration sites is sensitive to the stage in the SV40 life cycle at which MLV infection is performed. Using a temperature-sensitive tumor (T) antigen mutant SV40 strain, we observed that active replication of the target DNA is not required for efficient integration in vivo. The distribution of integration sites in vivo is closely approximately by in vitro reactions with isolated SV40 minichromosomes as integration targets. However, the degree of bias between the most and least favored sites is greater in vivo than in vitro.

Fv-1 restriction and its effects on murine leukemia virus integration in vivo and in vitro.

We have investigated the mechanisms by which alleles at the mouse Fv-1 locus restrict replication of murine leukemia viruses. Inhibition of productive infection is closely paralleled by reduced accumulation of integrated proviral DNA as well as by reduced levels of linear viral DNA in a cytoplasmic fraction. Nevertheless, viral DNA is present at nearly normal levels in a nuclear fraction, and total amounts of viral DNA are only mildly affected in restrictive infections, suggesting a block in integration to account for reduced levels of proviral DNA. However, integrase (IN)-dependent trimming of 3' ends of viral DNA occurs normally in vivo during restrictive infections, demonstrating that not all IN-mediated events are prevented in vivo. Furthermore, viral integration complexes present in nuclear extracts of infected restrictive cells are fully competent to integrate their DNA into a heterologous target in vitro. Thus, the Fv-1-dependent activity that restricts integration in vivo may be lost in vitro; alternatively, Fv-1 restriction may prevent a step required for integration in vivo that is bypassed in vitro.

Nucleosomes, DNA-binding proteins, and DNA sequence modulate retroviral integration target site selection.

Integration of retroviral DNA can serve as a paradigm for cellular functions that are affected by the packaging of DNA into chromatin. We have used a novel polymerase chain reaction-based assay to survey DNA and chromatin for the precise distribution of many integration sites. Integration into naked DNA targets is non-uniform, implying a nucleotide sequence bias. In chromatin, integration occurs preferentially at positions where the major groove is on the exposed face of the nucleosomal DNA helix, generating a 10 bp periodic spacing of preferred sites. Chromatin assembly enhances the reactivity of many sites, so that integration occurs most frequently at sites in nucleosomal, rather than nucleosome-free, regions of minichromosomes. In contrast, integration is prevented in a region occupied by a site-specific DNA-binding protein. Comparisons of integration events mediated by viral nucleoprotein complexes or by two different retroviral integrases show that the integration machinery also affects target site selection.

Retroviral integration into minichromosomes in vitro.

We describe here the use of chromatin as a target for retroviral integration in vitro. Extracts of cells newly infected with murine leukemia virus (MLV) provided the source of integration activity, and yeast TRP1ARS1 and SV40 minichromosomes served as simple models for chromatin. Both minichromosomes were used as targets for integration, with efficiencies comparable with that of naked DNA. In addition, under some reaction conditions the minichromosomes behaved as if they were used preferentially over naked DNAs in the same reaction. Mapping of integration sites by cloning and sequencing recombinants revealed that the integration machinery does not display a preference for nucleosome-free, nuclease-sensitive regions. The distributions of integration sites in TRP1ARS1 minichromosomes and a naked DNA counterpart were grossly similar, but in a detailed analysis the distribution in minichromosomes was found to be significantly more ordered: the sites displayed a periodic spacing of approximately 10 bp, many sites sustained multiple insertions and there was sequence bias at the target sites. These results are in accord with a model in which the integration machinery has preferential access to the exposed face of the nucleosomal DNA helix. The population of potential sites in chromatin therefore becomes more limited, in a manner dictated by the rotational orientation of the DNA sequence around the nucleosome core, and those sites are used more frequently than in naked DNA.

Biosynthesis of the reverse transcriptase of hepatitis B viruses involves de novo translational initiation not ribosomal frameshifting.

Retroviruses and many other types of genetic elements replicate by reverse transcription of RNA. Although structurally and biologically very diverse, such elements carry conserved polymerase genes (pol) that encode proteins required for reverse transcription. In most cases, the pol gene is preceded by an overlapping gene encoding one or more nucleocapsid proteins, in a different reading frame. Because both coding regions are represented in a single mRNA, the question arises of how the reverse transcriptase in the alternative reading frame is expressed. In retroviruses and retrotransposons it is expressed as a nucleocapsid-polymerase fusion protein by ribosomal frameshifting during translation of the overlapping region. We have examined the mechanism of polymerase biosynthesis in another family of animal viruses that use reverse transcription, the hepatitis B viruses. Genetic and biochemical studies reveal that these viruses do not use ribosomal frameshifting to generate this enzyme, but instead direct translation initiation at an internal initiation (AUG) codon in the polymerase gene.

The design, synthesis, and crystallization of an alpha-helical peptide.

Twelve- and sixteen-residue peptides have been designed to form tetrameric alpha-helical bundles. Both peptides are capable of folding into amphiphilic alpha-helices, with leucyl residues along one face and glutamyl and lysyl residues along the opposite face. Four such amphiphilic alpha-helices are capable of forming a noncovalently bonded tetramer. Neighboring helices run in antiparallel directions in the design, so that the complex has 222 symmetry. In the designed tetramer, the leucyl side chains interdigitate in the center in a hydrophobic interaction, and charged side chains are exposed to the solvent. The designed 12-mer (ALPHA-1) has been synthesized, and it forms helical aggregates in aqueous solution as judged by circular dichroic spectroscopy. It has also been crystallized and characterized by x-ray diffraction. The crystal symmetry is compatible with (but does not prove) the design. The design can be extended to a four-alpha-helical bundle formed from a single polypeptide by adding three peptide linkers.