Targeting proteins to membranes, using signal sequences for lipid modifications.

Changing an existing lipid or appending a lipid to a cytosolic protein has emerged as an important technique for targeting proteins to membranes and for constitutively activating the membrane-bound protein. The potential for more precise or regulated interactions of lipidated proteins in membrane subdomains suggests that this method for membrane targeting will be of increasing usefulness.

Oxidative modification of H-ras: S-thiolation and S-nitrosylation of reactive cysteines.

The reactive cysteines in H-ras are subject to oxidative modifications that potentially alter the cellular function of this protein. In this study, purified H-ras was modified by thiol oxidants such as hydrogen peroxide (H(2)O(2)), S-nitrosoglutathione, diamide, glutathione disulphide (GSSG) and cystamine, producing as many as four charge-isomeric forms of the protein. These results suggest that all four reactive cysteines of H-ras are potential sites of regulatory modification reactions. S-nitrosylated and S-glutathiolated forms of H-ras were identified by protocols that depend on separation of alkylated proteins on electrofocusing gels. S-nitrosoglutathione could S-nitrosylate H-ras on four cysteine residues, while reduced glutathione (GSH) and H(2)O(2) mediate S-glutathiolation on at least one cysteine of H-ras. Either GSSG or diamide S-glutathiolated at least two cysteine residues of purified H-ras. Iodoacetic acid reacts with three cysteine residues. In intact NIH-3T3 cells, wild-type H-ras was S-glutathiolated by diamide. Similarly, cells expressing a C118S mutant or a C181S/C184S double mutant of H-ras were S-glutathiolated by diamide. These results suggest that H-ras can be S-glutathiolated on multiple thiols in vivo and that at least one of these thiols is normally lipid-modified. In cells treated with S-nitrosocysteine, evidence for both S-nitrosylated and S-glutathiolated H-ras was obtained and S-nitrosylation was the predominant modification. These results show that oxidative modification of H-ras can be extensive in vivo, that both S-nitrosylated and S-glutathiolated forms may be important, and that oxidation may occur on reactive cysteines that are normally targeted for lipid-modification reactions.

Analysis of function and regulation of proteins that mediate signal transduction by use of lipid-modified plasma membrane-targeting sequences.

It is now established that the function of many signaling molecules is controlled, in part, by regulation of subcellular localization. For example, the dynamic recruitment of normally cytosolic proteins to the plasma membrane, by activated Ras or activated receptor tyrosine kinases, facilitates their interaction with other membrane-associated components that participate in their full activation (e.g., Raf-1). Therefore, the creation of chimeric proteins that contain lipid-modified signaling sequences that direct membrane localization allows the generation of constitutively activated variants of such proteins. The amino-terminal myristoylation signal sequence of Src family proteins and the carboxy-terminal prenylation signal sequence of Ras proteins have been widely used to achieve this goal. Such membrane-targeted variants have proved to be valuable reagents in the study of the biochemical and biological properties of many signaling molecules.

Murine guanylate-binding protein: incomplete geranylgeranyl isoprenoid modification of an interferon-gamma-inducible guanosine triphosphate-binding protein.

Farnesylation of Ras proteins is necessary for transforming activity. Although farnesyl transferase inhibitors show promise as anticancer agents, prenylation of the most commonly mutated Ras isoform, K-Ras4B, is difficult to prevent because K-Ras4B can be alternatively modified with geranylgeranyl (C20). Little is known of the mechanisms that produce incomplete or inappropriate prenylation. Among non-Ras proteins with CaaX motifs, murine guanylate-binding protein (mGBP1) was conspicuous for its unusually low incorporation of [(3)H]mevalonate. Possible problems in cellular isoprenoid metabolism or prenyl transferase activity were investigated, but none that caused this defect was identified, implying that the poor labeling actually represented incomplete prenylation of mGBP1 itself. Mutagenesis indicated that the last 18 residues of mGBP1 severely limited C20 incorporation but, surprisingly, were compatible with farnesyl modification. Features leading to the expression of mutant GBPs with partial isoprenoid modification were identified. The results demonstrate that it is possible to alter a protein's prenylation state in a living cell so that graded effects of isoprenoid on function can be studied. The C20-selective impairment in prenylation also identifies mGBP1 as an important model for the study of substrate/geranylgeranyl transferase I interactions.

Mutation of Ha-Ras C terminus changes effector pathway utilization.

In PC12 cells, Ha-Ras modulates multiple effector proteins that induce neuronal differentiation. To regulate these pathways Ha-Ras must be located at the plasma membrane, a process normally requiring attachment of farnesyl and palmitate lipids to the C terminus. Ext61L, a constitutively activated and palmitoylated Ha-Ras that lacks a farnesyl group, induced neurites with more actin cytoskeletal changes and lamellipodia than were induced by farnesylated Ha-Ras61L. Ext61L-triggered neurite outgrowth was prevented easily by co-expressing inhibitory Rho, Cdc42, or p21-activated kinase but required increased amounts of inhibitory Rac. Compared with Ha-Ras61L, Ext61L caused 2-fold greater Rac GTP binding and phosphatidylinositol 3-kinase activity in membranes, a hyperactivation that explained the numerous lamellipodia and ineffectiveness of Rac(N17). In contrast, Ext61L activated B-Raf kinase and ERK phosphorylation more poorly than Ha-Ras61L. Thus, accentuated differentiation by Ext61L apparently results from heightened activation of one Ras effector (phosphatidylinositol 3-kinase) and suboptimal activation of another (B-Raf). This surprising unbalanced effector activation, without changes in the designated Ras effector domain, indicates the Ext61L C-terminal alternations are a new way to influence Ha-Ras-effector utilization and suggest a broader role of the lipidated C terminus in Ha-Ras biological functions.

S-Nitrosocysteine increases palmitate turnover on Ha-Ras in NIH 3T3 cells.

Ha-Ras is modified by isoprenoid on Cys(186) and by reversibly attached palmitates at Cys(181) and Cys(184). Ha-Ras loses 90% of its transforming activity if Cys(181) and Cys(184) are changed to serines, implying that palmitates make important contributions to oncogenicity. However, study of dynamic acylation is hampered by an absence of methods for acutely manipulating Ha-Ras palmitoylation in living cells. S-nitrosocysteine (SNC) and, to a more modest extent, S-nitrosoglutathione were found to rapidly increase [(3)H]palmitate incorporation into cellular or oncogenic Ha-Ras in NIH 3T3 cells. In contrast, SNC decreased [(3)H]palmitate labeling of the transferrin receptor and caveolin. SNC accelerated loss of [(3)H]palmitate from Ha-Ras, implying that SNC stimulated deacylation and permitted subsequent reacylation of Ha-Ras. SNC also decreased Ha-Ras GTP binding and inhibited phosphorylation of the kinases ERK1 and ERK2 in NIH 3T3 cells. Thus, SNC altered two important properties of Ha-Ras activation state and lipidation. These results identify SNC as a new tool for manipulating palmitate turnover on Ha-Ras and for studying requirements of repalmitoylation and the relationship between palmitate cycling, membrane localization, and signaling by Ha-Ras.

Transient palmitoylation supports H-Ras membrane binding but only partial biological activity.

H-Ras is >95% membrane-bound when modified by farnesyl and palmitate, but <10% membrane-bound if only farnesyl is present, implying that palmitate provides major support for membrane interaction. However the direct contribution of palmitate to H-Ras membrane interaction or the extent of its cooperation with farnesyl is unknown, because in the native protein the isoprenoid must be present before palmitate can be attached. To examine if palmitates can maintain H-Ras membrane association despite multiple cycles of turnover, a nonfarnesylated H-Ras(Cys186Ser) was constructed, with an N-terminal palmitoylation signal, derived from the GAP-43 protein. Although 40% of the GAP43:Ras(61Leu,186Ser) protein (G43:Ras61L) partitioned with membranes, the chimera had less than 10% of the transforming activity of fully lipidated H-Ras(61Leu) in NIH 3T3 cells. Poor focus formation was not due to incorrect targeting or gross structural changes, because G43:Ras61L localized specifically to plasma membranes and triggered differentiation of PC12 cells as potently as native H-Ras61L. Proteolytic digestion indicated that in G43:Ras61L both the N-terminal and the two remaining C-terminal cysteines of G43:Ras61L were palmitoylated. A mutant lacking all three C-terminal Cys residues had decreased membrane binding and differentiating activity. Therefore, even with correct targeting and palmitates at the C-terminus, G43:Ras61L was only partially active. These results indicate that although farnesyl and palmitate share responsibility for H-Ras membrane binding, each lipid also has distinct functions. Farnesyl may be important for signaling, especially transformation, while palmitates may provide potentially dynamic regulation of membrane binding.

A non-farnesylated Ha-Ras protein can be palmitoylated and trigger potent differentiation and transformation.

Ha-Ras undergoes post-translational modifications (including attachment of farnesyl and palmitate) that culminate in localization of the protein to the plasma membrane. Because palmitate is not attached without prior farnesyl addition, the distinct contributions of the two lipid modifications to membrane attachment or biological activity have been difficult to examine. To test if palmitate is able to support these crucial functions on its own, novel C-terminal mutants of Ha-Ras were constructed, retaining the natural sites for palmitoylation, but replacing the C-terminal residue of the CAAX signal for prenylation with six lysines. Both the Ext61L and ExtWT proteins were modified in a dynamic fashion by palmitate, without being farnesylated; bound to membranes modestly (40% as well as native Ha-Ras); and retained appropriate GTP binding properties. Ext61L caused potent transformation of NIH 3T3 cells and, unexpectedly, an exaggerated differentiation of PC12 cells. Ext61L with the six lysines but lacking palmitates was inactive. Thus, farnesyl is not needed as a signal for palmitate attachment or removal, and a combination of transient palmitate modification and basic residues can support Ha-Ras membrane binding and two quite different biological functions. The roles of palmitate can therefore be independent of and distinct from those of farnesyl. Reciprocally, if membrane association can be sustained largely through palmitates, farnesyl is freed to interact with other proteins.

Murine GBP-2: a new IFN-gamma-induced member of the GBP family of GTPases isolated from macrophages.

We have cloned a new member of the interferon (IFN)-induced guanylate-binding protein (GBP) family of GTPases, murine GBP-2 (mGBP-2), from bone marrow-derived macrophages. mGBP-2 is located on murine chromosome 3, where it is linked to mGBP-1. With the identification of mGBP-2 there are now two human and two murine GBPs. Like other GBPs, mGBP-2 RNA and protein are induced by IFN-gamma. In addition, mGBP-2 shares with the other GBPs important structural features that distinguish this family from other GTPases. First, mGBP-2 contains only two of the three consensus sequences for nucleotide binding found within the classic GTP binding regions of other GTPases. A second amino acid motif found in mGBP-2 is a potential C-terminal site for isoprenoid modification, called a CaaX sequence. mGBP-2 is prenylated, as detected by [3H]mevalonate incorporation, when expressed in COS cells and preferentially incorporates the C-20 isoprenoid geranylgeraniol. Surprisingly, despite having a functional CaaX sequence, mGBP-2 is primarily cytosolic. GBP proteins are very abundant in IFN-exposed cells, but little is known about their function. mGBP-2 is expressed by IFN-gamma-treated cells from C57Bl/6 mice, whereas mGBP-1 is not. Thus, the identification of mGBP-2 makes possible the study of GBP function in the absence of a second family member.

Novel determinants of H-Ras plasma membrane localization and transformation.

Although it is well-established that modification of Ras by farnesol is a critical step for its membrane association and transforming activity, the contribution of other C-terminal sequences and palmitate modification to Ras localization and function remains unclear. We have characterized H-Ras mutant proteins with alterations in the palmitoylated cysteines or in sequences flanking these residues. We found that non-palmitoylated proteins were impaired not only in membrane association but also in transforming activity. Mutations which drastically altered residues adjacent to the palmitoylated cysteine did not abolish palmitoylation. However, despite continued lipid modification the mutant proteins failed to bind to plasma membranes and instead accumulated on internal membranes and, importantly, were not transforming. Addition of an N-terminal myristoylation signal to these defective mutants, or to proteins entirely lacking the C-terminal 25 residues restored both plasma membrane association and transforming activity. Thus, H-Ras does not absolutely require prenylation or palmitoylation nor indeed its hypervariable domain in order to interact with effectors that ultimately cause transformation. However, in this native state, the C-terminus appears to provide a combination of lipids and a previously unrecognized signal for specific plasma membrane targeting that are essential for the correct localization and biological function of H-Ras.

Prenylation of an interferon-gamma-induced GTP-binding protein: the human guanylate binding protein, huGBP1.

Interferons (IFN) and lipopolysaccharide (LPS) cause multiple changes in isoprenoid-modified proteins in murine macrophages, the most dramatic being the expression of a prenyl protein of 65 kDa. The guanylate binding proteins (GBPs) are IFN-inducible GTP-binding proteins of approximately 65 kDa that possess a CaaX motif at their C-terminus, indicating that they might be substrates for prenyltransferases. The human GBP1 protein, when expressed in transfected COS-1 cells, incorporates radioactivity from the isoprenoid precursor [3H]mevalonate. In addition, huGBPs expressed from the endogenous genes in IFN-gamma-treated human fibroblasts or monocytic cells were also found to be isoprenoid modified. IFN-gamma-induced huGBPs in HL-60 cells were not labeled by the specific C20 isoprenoid, [3H]geranylgeraniol, but did show decreased isoprenoid incorporation in cells treated with the farnesyl transferase inhibitor BZA-5B, indicating that huGBPs in HL-60 cells are probably modified by a C15 farnesyl rather than the more common C20 lipid. Differentiated HL-60 cells treated with IFN-gamma/LPS showed no change in the profile of constitutive isoprenylated proteins and the IFN-gamma/LPS-induced huGBPs remained prenylated. Despite being prenylated, huGBP1 in COS cells and endogenous huGBPs in HL-60 cells were primarily (approximately 85%) cytosolic. Human GBPs are thus among the select group of prenyl proteins whose synthesis is tightly regulated by a cytokine. HuGBP1 is an abundant protein whose prenylation may be vulnerable to farnesyl transferase inhibitors that are designed to prevent farnesylation of Ras proteins.

Rat p67 GBP is induced by interferon-gamma and isoprenoid-modified in macrophages.

The guanylate binding proteins, GBPs, are a family of interferon-induced GTP-binding proteins that include the rat p67. We report here that rat p67, for which interferon regulation had not previously been demonstrated, is induced by IFN-gamma and also by LPS in both cultured bone marrow-derived macrophages and microglia. The basal level of rat p67 in macrophages is low but increases dramatically between 2 and 4 hours after treating cells with either IFN-gamma or LPS. It then remains elevated over the next 24 hours. Rat p67 is isoprenoid modified. The isoprenoid modification was detected in p67 isolated both from primary IFN-gamma-activated macrophages and when the gene for p67 was transfected into COS cells. This is the first demonstration of in vivo prenylation of a GBP. The interferon regulation and prenylation of rat p67 point toward this protein being significant in the functions of both activated macrophages and microglia.

Farnesyl transferase inhibitors: the successes and surprises of a new class of potential cancer chemotherapeutics.

Proteins that transmit abnormal growth signals offer enticing points of intervention for the treatment of cancer. The discovery that isoprenoid attachment is required for the aberrant biological activity of oncogenic Ras proteins has provided just such a target.

Induction of a prenylated 65-kd protein in macrophages by interferon or lipopolysaccharide.

Treatment of murine bone marrow-derived macrophages with interferon-gamma (IFN-gamma) and/or lipopolysaccharide (LPS) resulted in changes in the abundance of a number of prenylated proteins. The most significant change involved a protein of 65 kd (p65) that became one of the most abundant prenylated proteins following treatment. The 65-kd protein was induced by agents that stimulate macrophage activation (IFNs or LPS) but not by cytokines that promote macrophage proliferation, such as granulocyte-macrophage colony-stimulating factor (GM-CSF), M-CSF, or interleukin-3. The majority of p65 was localized to subcellular fractions containing internal and plasma membranes but was not detected in nuclear membranes. The farnesyltransferase inhibitor BZA-5B caused a dramatic decrease in p65 prenylation, suggesting that this protein may be modified by the C15 isoprenoid farnesyl. These observations provide the first direct evidence that interferons and LPS cause changes in the abundance of specific isoprenoid-modified proteins in macrophages.

Targeting proteins to membranes using signal sequences for lipid modification.

Covalent attachment of lipids appears to be an important mechanism by which many proteins interact with membranes. As we learn more about how lipids and adjacent amino acids participate in addressing proteins to specific membranes within the cell, it should be possible to design more elegant and precise membrane targeting systems that can be used to guide proteins to functionally relevant destinations.

Polylysine domain of K-ras 4B protein is crucial for malignant transformation.

Previous studies have shown that posttranslational modifications are required for both oncogenic K-ras 4B protein membrane binding and transforming activity. In addition, Hancock et al. [Hancock, J. F., Patterson, H. & Marshall, C. J. (1990) Cell 63, 133-139] found that a polylysine domain contained at the C terminus of K-ras 4B was also absolutely essential for K-ras 4B membrane binding but, surprisingly, neither the polylysine domain nor membrane binding was required for transforming activity. We have performed similar studies, but our results are distinctly different. Our studies indicate that the polylysine domain is crucial for K-ras 4B transforming activity. Moreover, we demonstrate that although the polylysine domain increases K-ras 4B membrane binding, significant amounts of membrane binding can occur in the absence of this domain. Finally, while our studies are consistent with the notion that membrane binding is required for K-ras 4B transforming activity, we show that membrane binding, in and of itself, is not sufficient for efficient K-ras 4B transforming activity.

The carboxyl-terminal CXXX sequence of Gi alpha, but not Rab5 or Rab11, supports Ras processing and transforming activity.

Although the heterotrimeric Gi alpha subunit terminates in an apparent CXXX prenylation signal (CGLF), it is not modified by isoprenylation. To determine if the Gi alpha CXXX sequence can signal prenylation when placed at the carboxyl termini of normally prenylated proteins, we have characterized the processing and biological activity of chimeric oncogenic Ras proteins that terminate in the Gi alpha CXXX sequence (Ras/Gi alpha). Surprisingly, these chimeras were prenylated both in vivo and in vitro, demonstrated significant membrane association, exhibited transforming activity, and induced transcriptional transactivation from Ras-responsive elements. We then extended these studies to determine if, unlike the CC or CXC carboxyl-terminal sequences of other Rab proteins, the carboxyl-terminal CXXX sequences of the Ras-related Rab5 and Rab11 proteins represent conventional CXXX prenylation signals that can support Ras processing and transforming activity. Unexpectedly, these Ras/Rab chimeras were nonprenylated, were cytosolic, and lacked detectable transforming or transcriptional transactivation activity. Taken together, these results suggest that the context within which a CXXX sequence occurs may also critically control the modification of a protein by prenylation, and that the Rab5 and Rab11 carboxyl termini do not possess conventional CXXX sequences. Instead, their CCXX and CCXXX motifs may represent additional classes of protein prenylation signals.

Differential prenylation of proteins as a function of mevalonate concentration in CHO cells.

The incorporation of [5-3H]mevalonate into prenylated proteins and polyisoprenoid lipids has been determined as a function of mevalonate concentration in Chinese hamster ovary (CHO) cells that are inhibited in mevalonate synthesis. The relative incorporation of mevalonate into the different end products of isoprenoid metabolism was markedly dependent upon the concentration of mevalonate in the medium. The synthesis of cholesterol was dominant at higher concentrations of mevalonate while higher molecular weight isoprenoids were favored at the lower concentrations. The relative incorporation of mevalonate into the different prenylcysteines of prenylated proteins was dependent upon mevalonate concentration with geranylgeranylcysteine being the principal product at higher concentrations. At low levels of mevalonate farnesylcysteine synthesis predominated and geranylcysteine was detected. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of proteins from CHO cells that had been radiolabeled at different concentrations of [3H]mevalonate had different patterns on fluorography with relatively few proteins labeled at low concentrations. A study of this effect on the prenylcysteines of a specific protein, Ras, showed considerably less sensitivity to mevalonate concentration than bulk protein. These results indicate that the specific proteins that are prenylated depend upon the availability of the isoprenyl diphosphate substrates.

Structure and biological effects of lipid modifications on proteins.

Both the prevalence of lipid modifications of proteins and their importance for protein function and cellular localization have been widely observed. The advances made during the past year in defining the enzymology of lipid addition and in understanding the biological consequences of these modifications on protein function are discussed.

Prenoids and palmitate: lipids that control the biological activity of Ras proteins.

Ras proteins can be modified by two types of lipids--an isoprenoid and the fatty acid palmitate. These lipids help the otherwise cytoplasmic Ras protein to interact with the plasma membrane of a cell. The biological consequences of this association between Ras and membranes are dramatic, and can alter a cell's behavior from normal growth into malignancy. The scope and limits of our knowledge of the steps, structures and enzymes involved in this molecular transformation from soluble inactivity to membrane-bound potency are offered below. The prospects of regulating Ras function by controlling its intracellular location provides a tantalizing opportunity to translate research into a novel therapeutic reality.