A novel ubiquitination factor, E4, is involved in multiubiquitin chain assembly.

Proteins modified by multiubiquitin chains are the preferred substrates of the proteasome. Ubiquitination involves a ubiquitin-activating enzyme, E1, a ubiquitin-conjugating enzyme, E2, and often a substrate-specific ubiquitin-protein ligase, E3. Here we show that efficient multiubiquitination needed for proteasomal targeting of a model substrate requires an additional conjugation factor, named E4. This protein, previously known as UFD2 in yeast, binds to the ubiquitin moieties of preformed conjugates and catalyzes ubiquitin chain assembly in conjunction with E1, E2, and E3. Intriguingly, E4 defines a novel protein family that includes two human members and the regulatory protein NOSA from Dictyostelium required for fruiting body development. In yeast, E4 activity is linked to cell survival under stress conditions, indicating that eukaryotes utilize E4-dependent proteolysis pathways for multiple cellular functions.

Role of the ubiquitin/proteasome system in regulated protein degradation in Saccharomyces cerevisiae.

Selective degradation of proteins in eukaryotes is mediated primarily by the ubiquitin system in conjunction with the 26S proteasome. The yeast Saccharomyces cerevisiae has proved a powerful model system to study protein degradation in vivo. Biochemical and genetic studies complemented by the sequence analysis of the entire yeast genome have identified more than 70 genes presumed to function in the ubiquitin/proteasome system. Moreover, a number of physiological substrates of the ubiquitin system have been identified in yeast which are key regulatory proteins involved in the control of the cell cycle and transcription. In this review we will describe the enzymes effecting ubiquitin-protein conjugation and degradation. In addition we will discuss several targets of this system and describe the cellular functions mediated by this pathway.

Structural requirements for the efficient regulation of the Src protein tyrosine kinase by Csk.

Protein tyrosine kinases of the Src family are negatively regulated by phosphorylation in the C-terminal tail of the molecule. A different protein tyrosine kinase, Csk, is largely responsible for this regulation. The phosphorylated tail of c-Src engages with the SH2 domain in a conformation that is associated with low kinase activity and which involves stabilization by the SH3 domain. Inducible expression of c-Src in fission yeast is lethal unless Csk is coexpressed. Using this assay we present evidence that Src regulation by C-terminal phosphorylation does not require the myristylation signal or the unique domain at the N-terminus of the Src protein. Mutagenesis of the SH3 and SH2 domains of Csk show that neither are necessary in yeast or in vitro for efficient regulation of Src. Mutation of Tyr416 of Src, a site of autophosphorylation common to most protein tyrosine kinases, abolished the ability of Src to arrest growth of phosphorylate endogenous proteins. Tyr416 had the same effect on a shorter form of Src consisting of the kinase domain only, indicating that the mutation affects a property intrinsic to the catalytic domain. The residual activity of full-length Src mutated at Tyr416 is efficiently repressed by Csk action, suggesting that regulation by C-terminal phosphorylation does not act by preventing phosphorylation at Tyr416.

Sequence requirements for binding of Src family tyrosine kinases to activated growth factor receptors.

Activation of growth factor receptor protein tyrosine kinases frequently results in the binding of numerous proteins to their tyrosine-phosphorylated cytoplasmic domains. These interactions involve the SH2 domains of the binding proteins and phosphorylated tyrosines on the receptor molecules, with the specificity of interaction dictated by the amino acid composition surrounding the phosphorylated tyrosine. In the case of the platelet-derived growth factor (PDGF) receptor, the major binding site for Src family tyrosine kinases is in the juxtamembrane domain and includes tyrosine 579 (Mori, S., Rönnstrand, L., Yokote, K., Engström, A., Courtneidge, S. A., Claesson-Welsh, L., and Heldin, C-H. (1993) EMBO J. 12, 2257-2264). To analyze in more detail which amino acids surrounding the phosphorylated tyrosine at position 579 were important for high affinity interaction with Src family kinases, we synthesized a series of phosphopeptides corresponding to this binding site in which single amino acids were individually changed and tested their ability to compete with the PDGF receptor for binding of Fyn. We found that not only the three residues carboxyl-terminal to the phosphorylated tyrosine were important but that also residues at positions -1 and +4 relative to the tyrosine were required. Phosphorylation of both tyrosines 579 and 581 significantly increased competition efficiency. The activated colony stimulating factor-1 (CSF-1) receptor, which is known to associate with Src family kinases, has a sequence in its juxtamembrane region similar to that surrounding Tyr-579 of the PDGF receptor, and a phosphopeptide modeled on this sequence competed the association of Fyn with the receptor in vitro. Furthermore, mutational analysis demonstrated that these sequences were required for the efficient association of Src family kinases with the activated CSF-1 receptor in vivo. Phosphopeptides corresponding to the Src family binding sites of both PDGF and CSF-1 receptors activated Src kinase activity in vitro. These observations support a model in which the enzymatic activity of Src family tyrosine kinases is controlled by intra- and intermolecular interactions of tyrosine-phosphorylated peptides with the SH2 domain of the kinases.

DNA synthesis induced by some but not all growth factors requires Src family protein tyrosine kinases.

The Src family of protein tyrosine kinases have been implicated in the response of cells to several ligands. These include platelet-derived growth factor (PDGF), epidermal growth factor (EGF), and colony stimulating factor type 1 (CSF-1, in macrophages and in fibroblasts engineered to express the receptor). We recently described a microinjection approach which we used to demonstrate that Src family kinases are required for PDGF-induced S phase entry of fibroblasts. We now use this approach to ask whether other ligands also require Src kinases to stimulate cells to replicate DNA. An antibody specific for the carboxy terminus of Src, Fyn, and Yes (anti-cst.1) inhibited Src kinase activity in vitro and caused morphological reversion of Src transformed cells in vivo. Microinjection of this antibody was used to demonstrate that Src kinases were required for both CSF-1 and EGF to drive cells into the S phase. Expression of a kinase-inactive form of Src family kinases also prevented EGF- and CSF-1-stimulated DNA synthesis. However, even though the Src family kinases were necessary for both PDGF- and EGF-induced DNA synthesis in Swiss 3T3 cells, the responses to two other potent growth factors for these cells, lysophosphatidic acid and bombesin, were unaffected by the neutralizing antibodies. Therefore, some but not all growth factors required functional Src family kinases to transmit mitogenic responses.

Palmitoylation of multiple Src-family kinases at a homologous N-terminal motif.

We have recently identified a novel N-terminal cysteine-containing motif which specifies the palmitoylation of several G-protein alpha-subunits [Parenti, Viganó, Newman, Milligan and Magee (1993) Biochem. J. 291, 349-353]. A related motif occurs at the N-terminus of members of the Src family of protein tyrosine kinases except for Src itself and Blk. We have investigated whether the Src, Fyn, Yes and Lck gene products are palmitoylated. Src was not labelled with [3H]palmitate when endogenously expressed in COS cells. In contrast, endogenous Yes immunoprecipitated from COS cells was palmitoylated. Fyn was palmitoylated in insect cells infected with a recombinant baculovirus and the palmitoylation was independent of protein synthesis, suggesting a dynamic turnover of this lipid. Fatty acid analysis indicated that most of the label was incorporated as palmitate. Lck was palmitoylated when expressed by transfection in COS cells. All of these protein tyrosine kinases were also detectably myristoylated in each of the systems tested. Experiments performed with mutants of Lck expressed by transfection in COS cells indicated that cysteines at positions 3 and 5 were both palmitoylation sites and that myristoylation was required for palmitoylation. To confirm that palmitoylation was occurring on cysteines in the N-terminal region of Fyn, site-directed mutagenesis was used to replace the cysteines at positions 3 and 6 with alanine. The resulting protein was not palmitoylated but was still myristoylated when expressed in COS cells. A glycine to alanine mutant at position 2 was also not palmitoylated, showing that myristoylation is a prerequisite for palmitoylation. Our data indicate that Src family members containing the N-terminal cysteine motif are indeed palmitoylated. By analogy with Ras, it is possible that palmitoylation may play an important role in the localization and function of Src family protein tyrosine kinases.

Rapid and efficient purification of Src homology 2 domain-containing proteins: Fyn, Csk and phosphatidylinositol 3-kinase p85.

To analyse the regulation of Src family tyrosine kinases in vitro, we have purified Fyn and Csk, a kinase capable of regulating Fyn activity by phosphorylation, from baculovirus-infected insect cells. The proteins were purified by affinity purification over a phosphotyrosine column. Highly purified proteins were eluted from the resin by a salt gradient and further purified by ion-exchange chromatography. This purification scheme was successfully applied to a third, unrelated protein that also contains the Src homology 2 (SH2) domain, namely the 85 kDa subunit of phosphatidylinositol 3-kinase, indicating that this method is versatile and should prove applicable to any protein with an accessible SH2 domain. The binding of Csk to different phosphopeptides was tested, and specificity for the autophosphorylation site of Fyn was demonstrated. Pure Csk was used to phosphorylate Fyn and down-regulate its kinase activity, and the kinetic parameters of both the active and the repressed forms of Fyn were determined. Repression of Fyn activity by Csk reduced binding of Fyn to phosphopeptides to undetectable levels, supporting the model that predicts an intramolecular interaction of the Fyn SH2 domain with a C-terminal phosphotyrosine residue.

The phosphatidylinositol 3-kinase alpha is required for DNA synthesis induced by some, but not all, growth factors.

The phosphatidylinositol 3-kinase (PI 3-K) becomes activated when quiescent cells are stimulated with a variety of growth factors. We have microinjected antibodies specific for the p110 alpha subunit of the PI 3-K into quiescent fibroblasts and tested their effect on the ability of growth factors to stimulate exit from quiescence and entry into S phase. The antibodies inhibited platelet-derived growth factor-induced DNA synthesis, a result in keeping with previous studies using mutant platelet-derived growth factor receptors. Interestingly, functional PI 3-K was required for the first 6 hr of G1--i.e., until approximately 4 hr before the point at which the cells were committed to make DNA. A second tyrosine kinase receptor, the epidermal growth factor receptor, also required the PI 3-K for efficient signaling. However, colony-stimulating factor 1 (whose receptor is highly related to the platelet-derived growth factor receptor) could induce DNA synthesis in the absence of active PI 3-K, as could two growth factors (bombesin and lysophosphatidic acid) whose receptors are functionally coupled to G proteins. These data, therefore, demonstrate that some, but not all, growth factors require functional PI 3-K.

Generation of a temperature-sensitive cSRC.

We have created retroviruses encoding a temperature-sensitive variant of the cSrc protein, and of its activated allele cSrc(Y527F). Cells expressing the activated allele harboring this mutation displayed a transformed phenotype when grown at 34 degrees, but not at 39 degrees, as quantitated by measuring the susceptibility of the cells to contact inhibition and their ability to grow anchorage-independently at the two temperatures. The level of the mutant protein in cells grown at the high temperature was slightly reduced, and the in vitro kinase activity of the protein was hypersensitive to inactivation by heat treatment: the combination of these two defects probably explains the temperature-sensitive nature of the transformed phenotype. Surprisingly, the pattern of tyrosine phosphorylated proteins differed very little between the two temperatures, suggesting that critical substrates are few and of low abundance. The introduction of the same mutation into the wild-type allele of cSrc, Y527, also rendered the protein thermolabile. These ecotropic retroviruses should prove useful in assessing the function of cSrc in mammalian cells.

Csk inhibition of c-Src activity requires both the SH2 and SH3 domains of Src.

The protein tyrosine kinase c-Src is negatively regulated by phosphorylation of Tyr527 in its carboxy-terminal tail. A kinase that phosphorylates Tyr527, called Csk, has recently been identified. We expressed c-Src in yeast to test the role of the SH2 and SH3 domains of Src in the negative regulation exerted by Tyr527 phosphorylation. Inducible expression of c-Src in Schizosaccharomyces pombe caused cell death. Co-expression of Csk counteracted this effect. Src proteins mutated in either the SH2 or SH3 domain were as lethal as wild type c-Src, but were insensitive to Csk, even though they were substrates for Csk in vivo. Peptide binding experiments revealed that Src proteins with mutant SH3 domains adopted a conformation in which the SH2 domain was not interacting with the tail. These data support the model of an SH2 domain-phosphorylated tail interaction repressing c-Src activity, but expand it to include a role for the SH3 domain. We propose that the SH3 domain contributes to the maintenance of the folded, inactive configuration of the Src molecule by stabilizing the SH2 domain-phosphorylated tail interaction. Moreover, the system we describe here allows for further study of the regulation of tyrosine kinases in a neutral background and in an organism amenable to genetic analysis.

The Src family of protein tyrosine kinases: regulation and functions.

Most of the nine members of the Src family of tyrosine kinases are restricted in their expression, often to cells of the haematopoietic lineage, while some, particularly Src, Fyn and Yes, are more ubiquitously expressed. We have been studying the functions of Src, Fyn and Yes in fibroblasts. We have shown that stimulation of quiescent fibroblasts with platelet-derived growth factor (PDGF) causes Src, Fyn and Yes to become activated, and to associate transiently with the PDGF receptor. To address the role of Src, Fyn and Yes in the response to PDGF, we have used a dominant negative approach, in which cells were engineered to express catalytically inactive forms of Src kinases. These cells were unable to enter S phase in response to PDGF, and we therefore conclude that Src family tyrosine kinases are required in order for the PDGF receptor to transmit a mitogenic signal. It has previously been shown that the kinase activity of Src is negatively regulated by phosphorylation of tyr 527 in its carboxy-terminal tail. A kinase, Csk, that phosphorylates tyr 527 has recently been identified. We expressed Src in yeast to test the model that phosphorylation of tyr 527 represses activity by promoting intramolecular association between the tail and the SH2 domain. Inducible expression of Src in S. pombe caused cell death. Co-expression of Csk counteracted this effect. Src proteins mutated in the SH2 domain were as lethal as wild-type Src, but were insensitive to Csk. We interpret these results in favour of an SH2 domain: phosphorylated tail interaction repressing Src activity. (ABSTRACT TRUNCATED AT 250 WORDS)