Fig 4.18  Structural differences in the switch regions of GDPand GTP-bound Ras.Structures of the N-terminal residues of H-Ras complexed with GDP and with GppNHp (an analogue of GTP). The switch regions are indicated. The nucleotides are depicted as spheres. The green sphere is Mg2 . (4q2163–65,106 and 5p2166).

helix ( 2-helix) (Figure 4.18 and Table 4.5). The conformational transitions in these two switch regions are coupled so that binding of GTP brings about an ordered helix →coil transition at the N-terminus (closest to the nucleotide) of switch II. Subsequent changes to the 2-helix effectively reorganize the components of the effector-binding interface. All this is reversed when the bound GTP is hydrolysed to GDP.

Post-translational modifications

Beyond residue 165, the sequences become highly divergent. However, at the C-terminal extremities there is again homology in the form of Caax (Cysteine, aliphatic, aliphatic, anything: CVLS in human H-Ras, CVVM in N- Ras and CIIM in K-Ras: see Table 4.4). The Caax motif is post-translationally modified by the addition of a farnesyl extension (C15, 3 isoprene units: see

Figure 4.12) through stable thioether linkage to the cysteine. This is followed by carboxymethylation, cleavage of the final three amino acids, and methylation of the newly exposed C-terminus (now cysteine). A second lipid modification occurs by palmitoylation of another cysteine residue within the C-terminal hypervariable region.

These modifications ensure strong attachment to the internal surface of the plasma membrane. When the post-translational prenylation reaction is prevented, either by mutation of the Caax sequence or by farnesyl

transferase inhibitors, the normal functions of Ras are suppressed.107 The nonfarnesylated Ras accumulates in the cytoplasm and as a result, it scavenges the associated kinase Raf (see page 334), preventing it from coming into contact with the membrane.108

K-Ras-2B is different. It is not palmitoylated but instead there is a positive charge cluster which is thought to assist its association

with negatively charged phospholipids on the inner surface of the membrane.

Farnesyl transferase inhibitors such as simvastatin, lovastatin, and pravastatin are used to suppress cholesterol biosynthesis in patients

having a history of stroke or heart disease. These drugs inhibit the conversion of 3-hydroxy- 3-methylglutaryl CoA to mevalonate (catalysed by HMG CoA reductase). This is the key step in the pathway to the formation of isoprene units which then polymerize to form C10 (geranyl), C15 (farnsesyl), C20 (geranylgeranyl), and C30 (squalene). The last of these is the precursor in the biosynthesis of cholesterol. Farnesyl transferase inhibitors are also potential anticancer agents.

GTPases everywhere!

There are hundreds of monomeric GTPases. With few exceptions they have been identified by screening cDNA libraries for Ras-like proteins and their functions have necessarily been assigned (and are still being assigned) subsequent to their discovery. This is an ongoing activity. The normal strategy is to express (or introduce) mutant forms, expected to possess altered functions which might be revealed as an altered phenotype. Thus, the sequence comprising the effector domain of Ras was established long before any effectors were identified. By comparison, the first discovered heterotrimeric G proteins (Gs, Gi, Go, Gt (transducin), etc.) were isolated by classical purification strategies seeking proteins that bind and hydrolyse GTP and which interact with appropriate receptors and downstream effectors known to be affected by GTP.

To date, representative functions have been assigned for members of all the groups of Ras-related proteins in vertebrates and also in other eukaryotic organisms such as yeasts, flies, nematodes, and slime moulds (Table 4.6). Some of the Ras-like proteins appear to have more than one function, possibly determined by the cells in which they are expressed or by their subcellular localization. Doubtless, more will become evident. Here we concentrate on the Ras proteins. Others are considered at appropriate points in the text.

Mutations of Ras that promote cancer

As with the heterotrimeric G proteins, the proportion of Ras in the activated

form is given by the ratio kdiss/kcat (see Equation 4.1). Some of the oncogenic viral gene products (vRas) differ from the wild-type cellular forms by single

point mutations which inhibit the GTPase activity (reducing kcat) and this

Table 4.6  Functions of Ras-related proteins

prolongs the active state. For this reason, a higher proportion of the activated GTP-bound form exists in cells that have been transformed by vRas. The most common mutations promoting the transformed phenotype occur at residues 12, 13, and 61 (see Figure 4.15).

Other mutations promoting spontaneous activation of Ras are those which enhance the rate of dissociation (kdiss) of bound nucleotide (GDP). In the unmodified protein, this occurs at a very slow rate indeed (about 10 5 mol s 1 per mol of protein), but there are numerous mutations which increase this rate significantly. Many of these are at or in the close vicinity of the residues interacting with the nucleotide. In particular, substitution of Ala-146 increases the rate of guanine nucleotide exchange 1000-fold. As with the GTPase mutants, these mutations ensure that a higher proportion of the protein exists in the activated GTP-bound form. In consequence these too, are transforming mutations.

Functions of Ras

While the identification of receptors and their catalytic activities generally preceded the discovery of the heterotrimeric G proteins that link their operations, the proteins operating immediately upstream or downstream of Ras at first eluded definition. Even so, a place for Ras in the regulation of mitogenesis was not in doubt. Cellular expression of the transformed ras genes,109,110 or direct introduction of their products into single cells by

microinjection,111,112 promotes cell detachment and loss of contact inhibition, characteristics of the transformed phenotype (see page 306). Conversely, microinjection of antibodies (to neutralize the native cellular Ras), or of a dominant inhibitory mutant form (S17N) prevents normal growth and cell division.113–115 However, it should not be assumed that mitogenesis is the exclusive outcome of activating Ras. Other cell types undergo differentiation as the consequence of Ras activation. In PC12 cells the introduction of the Ras oncogene product provides a signal for neurite outgrowth.116,117

How Ras controls all this has been far from clear. It is true to say that Ras lies at the centre of a network of interacting pathways, is activated and modulated directly and indirectly by several receptors, and in turn, has its influence on

a large number of downstream processes (see Figure 4.19 and Chapter 12). Indeed, it is amazing that a protein as small as Ras can interact with so many other proteins. Among functions directly linked to Ras are the activation of the protein kinase Raf and the activation of phosphatidylinositol 3-kinase (PI 3-kinase, see Chapter 18). It is the first member of the chain sequence of

phosphorylating enzymes that leads to the activation of extracellular receptor kinase (ERK) and thence to the transcription of genes controlled through the serum response element, SRE (Chapter 12).

S17N: Serine replaces asparagine at position 17.

PC12 cells are adrenal medullary cell line derived from a phaeochromocytoma.