20:  20.4 Receptor Tyrosine Kinases and Ras

Ligand Binding Leads to Autophosphorylation of RTKs

All RTKs comprise an extracellular domain containing a ligand-binding site, a single hydrophobic transmembrane α helix, and a cytosolic domain that includes a region with protein-tyrosine kinase activity. Binding of ligand causes most RTKs to dimerize; the protein kinase of each receptor monomer then phosphorylates a distinct set of tyrosine residues in the cytosolic domain of its dimer partner, a process termed autophosphorylation (Figure 20-21). Autophosphorylation occurs in two stages. First, tyrosine residues in the phosphorylation lip near the catalytic site are phosphorylated. This leads to a conformational change that facilitates binding of ATP in some receptors (e.g., the insulin receptor) and binding of protein substrates in other receptors (e.g., FGF receptor). The receptor kinase activity then phosphorylates other sites in the cytosolic domain; the resulting phosphotyrosines serve as docking sites for other proteins involved in RTK-mediated signal transduction.

As described later, the subunits of some RTKs, including the insulin receptor, are covalently linked. Although these receptors exist as dimers or tetramers even in the absence of ligand, binding of ligand is required for autophosphorylation to occur. Presumably, ligand binding induces a conformational change that activates the kinase.

The phosphotyrosine residues in activated RTKs interact with adapter proteins containing SH2 or PTB domains. These proteins couple the activated receptors to other components of the signal-transduction pathway but have no intrinsic signaling properties. Before examining the structure and function of adapter proteins, we discuss the role of Ras, the other key signaling molecule in pathways triggered by activation of RTKs. As we will see later, several membrane-associated enzymes that function in signal transduction also bind to specific phosphotyrosines in activated RTKs.

Ras and G_α Subunits Belong to the GTPase Superfamily of Intracellular Switch Proteins

Ras is a GTP-binding switch protein that, like the G_α subunits in different G proteins, alternates between an active on state with a bound GTP and an inactive off state with a bound GDP. As discussed in the previous section, G_α is directly coupled to GPCRs and transduces signals to various effectors such as adenylyl cyclase. In contrast, Ras is not directly linked to RTKs.

Activation of both Ras and G_α is triggered by hormone binding to an appropriate cell-surface receptor. Ras activation is accelerated by a protein called guanine nucleotide – exchange factor (GEF), which binds to the Ras · GDP complex, causing dissociation of the bound GDP (Figure 20-22). Because GTP is present in cells at a higher concentration than GDP, GTP binds spontaneously to “empty” Ras molecules, with release of GEF. In contrast, formation of an active G_α · GTP complex does not require an exchange factor.

Hydrolysis of the bound GTP deactivates both Ras and G_α. The average lifetime of a GTP bound to Ras is about 1 minute, which is much longer thanthe lifetime of G_α · GTP. The reason for this difference is that the deactivation of Ras, unlike the deactivation of G_α, requires the assistance of another protein: a GTPase-activating protein (GAP), which binds to Ras · GTP and accelerates its intrinsic GTPase activity by a hundredfold (see Figure 20-22). Mammalian Ras proteins have been studied in great detail because mutant Ras proteins are associated with many types of human cancer. These mutant proteins, which bind but cannot hydrolyze GTP, are permanently in the “on” state and cause neoplastic transformation (Chapter 24).

The differences in the cycling mechanisms of Ras and G_α are reflected in their structures. Ras (≈170 amino acids) is smaller than G_α proteins (≈300 amino acids), but its three-dimensional structure is similar to that of the GTPase domain of G_α. Recent structural and biochemical studies show that G_α also contains another domain, a helical domain that apparently functions like GAP to increase the rate of GTP hydrolysis by G_α (see Figure 20-19). In addition, the direct interaction between an activated receptor and G_α · GDP promotes release of GDP and binding of GTP, so that a separate exchange factor is not required.

Both G_sα and Ras are members of a family of intracellular GTP-binding switch proteins collectively referred to as the GTPase superfamily. This family also includes other G_α subunits (e.g., G_iα), the Rab proteins which regulate fusion of vesicles within cells (Section 17.10), and the Rho family proteins which regulate the actin cytoskeleton (Section 18.2). The many similarities between the structure and function of Ras and G_sα, and the identification of both proteins in all eukaryotic cells, indicate that a single type of signaltransducing GTPase originated very early in evolution. The gene encoding this protein subsequently duplicated and evolved to the extent that cells today contain a superfamily of such GTPases, comprising perhaps a hundred different intracellular switch proteins. These related proteins control many aspects of cellular growth and metabolism.

An Adapter Protein and GEF Link Most Activated RTKs to Ras

The first indication that Ras functioned downstream from RTKs in a common signaling pathway came from experiments in which cultured fibroblast cells were induced to proliferate by treatment with a mixture of platelet-derived growth factor (PDGF) and epidermal growth factor (EGF). Microinjection of anti-Ras antibodies into these cells blocked cell proliferation. Conversely, injection of a constitutively active mutant Ras protein (i.e., Ras^D), which hydrolyzes GTP very inefficiently and thus persists in the active state, caused the cells to proliferate in the absence of the growth factors. These findings are consistent with studies showing that addition of fibroblast growth factor (FGF) to fibroblasts leads to a rapid increase in the proportion of Ras present in the GTP-bound active form.

But how does binding of a growth factor (e.g., EGF) to an RTK (e.g., the EGF receptor) lead to activation of Ras? Two cytosolic proteins — GRB2 and Sos — provide the key links (Figure 20-23). An SH2 domain in GRB2 binds to a specific phosphotyrosine residue in the activated receptor. GRB2 also contains two SH3 domains, which bind to and activate Sos. GRB2 thus functions as an adapter protein for the EGF receptor. Sos functions as a guanine nucleotide – exchange protein (GEF), which helps convert inactive GDP-bound Ras to the active GTP-bound form.

Genetic analysis of mutants blocked at particular stages of differentiation have provided considerable insight into RTK signaling pathways. Most of these genetic studies were done in the worm C. elegans and in the fly Drosophila. Mutants in these species in which development of specific cells is blocked were particularly useful in elucidating the pathway from an RTK to Ras shown in Figure 20-23. To illustrate the power of this experimental approach, we consider development of a particular type of cell in the compound eye of Drosophila.

The compound eye of the fly is composed of some 800 individual eyes called ommatidia (Figure 20-24a). Each ommatidium consists of 22 cells, eight of which are photosensitive neurons called retinula, or R cells, designated R1 – R8 (Figure 20-24b). An RTK called Sevenless (Sev) specifically regulates development of the R7 cell. In flies with a mutant sevenless (sev) gene, the R7 cell in each ommatidium does not form (Figure 20-24c). Since the R7 photoreceptor is necessary for flies to see in ultraviolet light, mutants that lack functional R7 cells are easily isolated.

During development of each ommatidium, a protein called Boss (Bride of Sevenless) is expressed on the surface of the R8 cell. This membrane-bound protein binds to the Sev RTK on the surface of the neighboring R7 precursor cell, signaling it to develop into a photosensitive neuron (Figure 20-25a). In mutant flies that do not express a functional Sev RTK, interaction between the Boss and Sev proteins cannot occur, and no R7 cells develop (Figure 20-25b).

To identify intracellular signal-transducing proteins in the Sev RTK pathway, investigators produced mutant flies expressing a temperature-sensitive Sev protein. When these flies were maintained at a permissive temperature, all their ommatidia contained R7 cells; when they were maintained at a nonpermissive temperature, no R7 cells developed. At a particular intermediate temperature, however, just enough of the Sev RTK was functional to mediate normal R7 development. The investigators reasoned that at this intermediate temperature, the signaling pathway would become defective, and thus no R7 cells would develop, if the level of another protein involved in the pathway was reduced. A recessive mutation affecting such a protein would have this effect because, in diploid organisms like Drosophila, a heterozygote containing one wild-type and one mutant allele of a gene will produce half the normal amount of the gene product; hence even if such a recessive mutation is in an essential gene, the organism will be viable. However, a fly carrying a temperature-sensitive mutation in the sev gene and a second mutation affecting another protein in the signaling pathway would be expected to lack R7 cells at the intermediate temperature.

By use of this screen, researchers identified genes encoding three important proteins in the Sev pathway (see Figure 20-23):

• A Ras protein exhibiting 80 percent identity with its mammalian counterparts

• A GEF called Sos (Son of Sevenless) exhibiting 45 percent identity with its mouse counterpart

• An SH2-containing adapter protein exhibiting 64 percent identity to human GRB2

These three proteins have been found to function in other RTK signal-transduction pathways initiated by ligand binding to different receptors and used at different times and places in the developing fly. Recessive lethal mutations in these essential genes can be identified by the strategy described here much more easily than by the procedure described in Chapter 8 (see Figure 8-10).

In subsequent studies, researchers introduced a mutant ras^D gene into fly embryos carrying the sevenless mutation. As noted earlier, the ras^D gene encodes a Ras protein that has reduced GTPase activity and hence is present in the active GTP-bound form even in the absence of a hormone signal. Although no functional Sev RTK was expressed in these double-mutants (sev⁻; ras^D), R7 cells formed normally, indicating that activation of Ras is sufficient for induction of R7-cell development (Figure 20-25c). This finding is consistent with the results with cultured fibroblasts described earlier.

SH2 Domain in GRB2 Adapter Protein Binds to a Specific Phosphotyrosine in an Activated RTK

To identify proteins that associate with phosphotyrosine residues in the cytosolic domain of activated EGF receptors, scientists used an expression cloning strategy. cDNAs synthesized from mRNAs isolated from human brain-stem tissue were inserted into a λgt11 expression vector, which then was plated on a lawn of E. coli cells (see Figure 7-21). When the resulting cDNA library was screened using a fragment of phosphorylated human EGF receptor as the probe, two cDNA clones were identified. One encoded a subunit of PI-3 kinase that contains an SH2 domain and the other encoded GRB2, a homolog of the SH2-containing adapter protein identified in the Drosophila Sev pathway. Thus GRB2 and its Drosophila homolog are adapter proteins that function downstream from RTKs but upstream of Ras in both flies and mammalian cells.

GRB2 and similar adapter proteins bind to different phosphotyrosine residues on RTKs via the conserved SH2 domain. This domain derived its full name, the Src homology 2 domain, from its homology with a region in the prototypical cytosolic tyrosine kinase encoded by src. The three-dimensional structures of SH2 domains in different proteins are very similar. Each binds to a distinct sequence of amino acids surrounding a phosphotyrosine residue. The unique amino acid sequence of each SH2 domain determines the specific phosphotyrosine residues it binds. The SH2 domain of the Src tyrosine kinase, for example, binds strongly to any peptide containing the critical core sequence of phosphotyrosine – glutamic acid – glutamic acid – isoleucine (Figure 20-26a). These four amino acids make intimate contact with the peptide-binding site in the Src SH2 domain. Binding resembles the insertion of a two-pronged “plug” — the phosphotyrosine and isoleucine residues of the peptide — into a two-pronged “socket” in the SH2 domain. The two glutamic acids fit snugly onto the surface of the SH2 domain between the phosphotyrosine socket and the hydrophobic socket that accepts the isoleucine residue. Variations in the nature of the hydrophobic socket in different SH2 domains allow them to bind to phosphotyrosines adjacent to different sequences, accounting for differences in their binding specificity.

Activated RTKs also can recruit signaling molecules through a different domain called the phosphotyrosinebinding (PTB) domain. While SH2-binding specificity is largely determined by residues C-terminal to the phosphotyrosine, PTB-binding specificity is determined by specific residues five to eight residues N-terminal to the phosphotyrosine residue.

Sos, a Guanine Nucleotide – Exchange Factor, Binds to the SH3 Domains in GRB2

In addition to one SH2 domain, which binds to phosphotyrosine residues in RTKs, GRB2 contains two SH3 domains, which bind to Sos, a guanine nucleotide – exchange factor. SH3 domains, which contain ≈55 – 70 residues, are present in a large number of proteins involved in intracellular signaling. Although the three-dimensional structures of various SH3 domains are similar, their specific amino acid sequences differ. SH3 domains selectively bind to proline-rich sequences in Sos and other proteins; different SH3 domains bind to different proline-rich sequences.

Proline residues play two roles in the interaction between an SH3 domain in an adapter protein (e.g., GRB2) and a proline-rich sequence in another protein (e.g., Sos). First, the proline-rich sequence assumes an extended conformation that permits extensive contacts with the SH3 domain, thereby facilitating interaction. Second, a subset of these prolines fit into binding “pockets” on the surface of the SH3 domain (Figure 20-26b). Several nonproline residues also interact with the SH3 domain and are responsible for determining the binding specificity. Hence the binding of peptides to SH2 and SH3 domains follows a similar strategy: certain residues provide the overall structural motif necessary for binding, and neighboring residues confer specificity to the binding.

Following hormone-induced activation of an RTK (e.g., the EGF receptor), a complex containing the activated RTK, GRB2, and Sos is formed on the cytosolic face of the plasma membrane (see Figure 20-23). Formation of this complex depends on the dual binding ability of GRB2. Receptor activation thus leads to relocalization of Sos from the cytosol to the membrane, bringing Sos near to its substrate, membrane-bound Ras · GDP. Biochemical and genetic studies indicate that the C-terminus of Sos inhibits its nucleotideexchange activity and that GRB2 binding relieves this inhibition. Binding of Sos to Ras · GDP leads to changes in the conformation of two regions of Ras, switch I and switch II, thereby opening the binding pocket for GDP so it can diffuse out (Figure 20-27). Because GTP is present in cells at a concentration some 10 times higher than GDP, GTP binding occurs preferentially, leading to activation of Ras. The activation of Ras and G_sα thus occur by similar mechanisms: a conformational change induced by binding of a protein — Sos and an activated GPCR, respectively — that opens the protein structure so bound GDP is released to be replaced by GTP. As we discuss in the next section, binding of GTP to Ras, in turn, induces a specific conformation of switch I and II that allow Ras · GTP to activate downstream effector molecules.

Several other proteins, including GAP, bind to specific phosphotyrosines in activated RTKs. This binding localizes GAP close to Ras · GTP, so it can promote the cycling of Ras (see Figure 20-22); exactly how GAP interacts with Ras and perhaps other components of the RTK-Ras pathway is unclear.

SUMMARY

• Receptor tyrosine kinases (RTKs), which bind to peptide/protein hormones, may exist as dimers or dimerize during binding to ligands.

• Ligand binding leads to activation of the kinase activity of the receptor and autophosphorylation of tyrosine residues in its cytosolic domain (see Figure 20-31). The activated receptor also can phosphorylate other protein substrates.

• Ras is an intracellular GTPase switch protein that acts downstream from most RTKs. Like G_sα, Ras cycles between an inactive GDP-bound form and active GTP-bound form. Ras cycling requires the assistance of two proteins, GEF and GAP, (see Figure 20-22), whereas G_sα cycling does not.

• Unlike GPCRs, which interact directly with an associated G protein, RTKs are linked indirectly to Ras via two proteins, GRB2 and Sos (see Figure 20-23).

• The SH2 domain in GRB2, an adapter protein, binds to specific phosphotyrosines in activated RTKs. The two SH3 domains in GRB2 then bind Sos, a guaninenucleotide exchange factor, thereby bringing Sos close to membrane-bound Ras · GDP and activating its exchange function.

• Binding of Sos to inactive Ras causes a large conformational change that permits release of GDP and binding of GTP.

• Normally, Ras activation and the subsequent cellular response is induced by ligand binding to an RTK. However, in cells that contain a constitutively active Ras, the cellular response occurs in the absence of ligand binding.