Protein Prenylation: Genes, Enzymes, Targets, and Functions

Annu. Rev. Genet. 1992.30:209-37Copyright © I~ Annual Reviews Inc. All rights reserved

PROTEIN PRENYLATION:GENES, ENZYMES, TARGETS,AND FUNCTIONS

William R. Schafer and Jasper RineDepartment of Molecular and Cell Biology, University of California, Berkeley,California 94720

KEY WORDS: Prenylation, posttranslational modification, isoprenoid, Ras proteins, pal-mitoylation

CONTENTS

INTRODUCTION .......................................... 209GENES ENCODING PRENYLATED PROTEINS ...................... 210

The Ras Superfamily ...................................... 210Lipopeptide Pheromones .................................... 213Nuclear Lamins ......................................... 215G-proteins ............................................ 216

PROTEINS INVOLVED IN PRENYL-DEPENDENT PROCESSING .......... 217Protein Prenyltransferases ................................... 218Proteases ............................................. 221Reversible Modifications .................................... 223

PRENYL-MEDIATED INTERACTIONS ........................... 225Prenyl-protein Receptors .................................... 225Lipid Interactions ........................................ 227Indirect Prenyl-dependent Targeting ............................ 229

FUTURE PERSPECTIVES .................................... 230

INTRODUCTION

The regulation, functional activation, and intracellular targeting of biologicalmacromolecules are often mediated through the covalent attachment ofspecific chemical moieties. The posttranslational modification of specificproteins by phosphorylation, glycosylation, acetylation, fatty acylation, andmethylation has been studied extensively, and a variety of critical biologicalfunctions have been ascribed to these particular side groups. In contrast tothese familiar types of chemical modification, protein prenylation is a process

209

0066-4197/92/1215-0209502.00

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.

http://www.annualreviews.org/aronline

210 SCHAFER & RINE

which, until recently, was poorly understood. This article reviews the currentunderstanding of the biological role for prenylation in the function of specificeukaryotic proteins, focusing on genetic studies of specific prenylated proteinsand the enzymes that catalyze protein prenylation. In addition, we discusspossible models for how prenyl groups mediate specific protein functions,such as intracellular protein targeting. We also speculate on the possiblerelevance of protein prenylation to the medical problems of oncogenesis andhypercholesterolemia.

Prenylation refers to the covalent modification of a molecule by theattachment of a lipophilic isoprenoid group (reviewed in 26, 29, 46). Iso-prenoids are a diverse family of lipophilic molecules based on a repeatingfwe-carbon structure. Famesyl diphosphate, a 15-carbon molecule, andgeranylgeranyl diphosphate, a 20-carbon molecule, are the isoprenoid com-pounds most relevant to protein prenylation. Other biologically importantisoprenoid molecules include cholesterol and other steroids, carotenoids,terpenes, and dolichols. A variety of small molecules are prenylated, includingquinones (coenzyme Q), porphyrins (heme a and chlorophyll), amino acids(dimethylallyl tryptophan), and purines (cytokinins). In addition, transfer RNA molecules are prenylated at a specific adenine residue. Allisoprenoids are synthesized from the six-carbon precursor mevalonate througha common biosynthetic pathway (47).

Many specific proteins are posttranslationally modified with prenyl groups.Most, if not all, prenylated proteins are modified by the attachment of eithera 15-carbon farnesyl or a 20-carbon geranylgeranyl group (117) in a thioetherlinkage to a cysteine residue. In most organisms, geranylgeranylation is amore common modification than farnesylation, although the relative propor-tions of farnesylcysteine and geranylgeranylcysteine in total cellular proteinvary somewhat between organisms and cell types (37). It is possible that othertypes of prenyl modification of proteins may exist; for example, a mammalianprotein appears to be covalently modified with a prenylated adenine moiety(39). Prenylated proteins are found in a variety of cellular compartments,including the nucleus, the cytosol, and membrane-bound organelles (85, 129).

GENES ENCODING PRENYLATED PROTEINS

Most of the prenylated proteins whose identities have been determined belongto one of of four protein families: small GTP-binding proteins, lipopeptidepheromones, nuclear lamins, and trimeric G-proteins. The importance ofprenylation for the functions of these proteins is discussed below.

The Ras SuperfamilyMost of the known prenylated proteins are members of the Ras superfamilyof low-molecular-weight guanine nucleotide-binding proteins. These proteins


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


PROTEIN PRENYLATION 211

can be divided into four smaller protein families: the Ras, Rho, Rap, andRab/Ypt families (23). These proteins participate in a variety of cellularfunctions, including control of cell growth and differentiation, cytokinesis,and membrane trafficking. All small GTP-binding proteins contain cysteineresidues at or near the carboxyl-terminus that appear to serve as targets forposttranslational prenylation. At least three distinct types of prenyl-dependentprocessing are involved in the maturation of these proteins, each of whichaffects a distinct subset of Ras-related proteins with a distinct carboxyl-ter-minal motif.

THE CAAX MOTIF The Ras proteins are plasma membrane-localized mole-cules that regulate cell differentiation and proliferation in mammalian andyeast cells. Ras proteins contain a carboxyl-terrninal sequence called a CaaXmotif, with a cysteine followed by two aliphatic residues and a carboxyl-ter-minal "X" residue, which can be C, S, M, Q, or A (135). Ras proteins undergofarnesylation of the cysteine residue (21, 54), proteolytic removal of the threeamino acids distal to the cysteine, and methylation of the carboxyl-terminus(35, 50). Some Ras proteins are also palmitoylated at a cysteine residue nearthe famesylated cysteine (19). Ras proteins that lack this second cysteineresidue are not palmitoylated but contain instead a region of basic amino acidsin the hypervariable domain near the carboxyl-terminus (54). Ras is synthe-sized on soluble cytoplasmic ribosomes (50, 130), and is localized to theplasma membrane independently of the secretory pathway.

The localization and functional activity of Ras proteins depends on thepresence of the farnesyl group. Genetic or pharmacological inhibition ofprotein prenylation blocks the membrane association and biological activityof both human and yeast Ras proteins (54, 124). However, since RAS2 proteinfrom Saccharomyces cerevisiae lacking the modifiable cysteine retains somefunctional activity when overexpressed in vivo (34), it appears that thefunction of prenylation is primarily to direct the polypeptide to the correctintracellular location. Proteolysis and methylation appear to play lesser rolesin mediating the association of Ras proteins with membranes. Whereas humanRas proteins lacking these modifications exhibit a reduced affinity formembranes (52), mutant K-ras alleles that prevent proteolysis and methylationdo not block the membrane association or the transforming activity of theK-ras protein in vivo (71). Similarly, yeast mutations that cause a defect carboxymethylation do not block Ras localization or functional activity (63,90). In addition to the farnesyl group, either a palmitoylation site or a polybasicregion is required for efficient localization to the plasma membrane and isimportant for the transforming activity of oncogenic mutant proteins (55).

THE CAAL MOTIF A second group of small GTP-binding proteins contain acarboxyl-terminal motif similar to a CaaX box, called a CaaL motif, with a


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


212 SCHAFER & RINE

leucine residue in the carboxyl-terminal position. This group includes manyproteins implicated in cytokinesis, such as the Rap and Rho protein families.Many of these proteins are known to be geranylgeranylated, including G25K(143), RhoA (70), Ral, Rac (74), Rapla (18), and Raplb (72). these proteins are also known to undergo carboxyl-terminal proteolysis andcarboxymethylation (18, 70, 72). Geranylgeranylation of these proteins widely thought to mediate association with intracellular membranes; however,only in the cases of Rapl and Rho has this hypothesis been supportedexperimentally.

Rapla and Raplb are a pair of closely related genes that are members ofthe Ras family of small GTP-binding proteins (108). Overexpression of Rap 1 (also known as K-rev-1) results in suppression of the transforming activity activated Ras alleles (76). Unlike Ras proteins, which are localized to theplasma membrane, Rap 1 proteins are associated with the Golgi complex (13).The posttranslational processing at the carboxyl-terminus of the Rapl b proteinhas been investigated in detail. The carboxyl-terminus, which has the sequenceK-A-R-K-K-S-S-C-Q-L-L, undergoes four modifications: (a) geranylger-anylation of the cysteine residue; (b) proteolytic removal of the three terminalamino acids; (c) methylation of the new carboxyl-terminus, and (d) phosphor-ylation of the distal serine residue by A-kinase (72). A prenylated carboxyl-terminus is required for the association of Raplb with membranes. Prenylationand phosphorylation are both required for interaction of Raplb with a factorknown as GDS, which stimulates GDP/GTP exchange (56), since prenylated,phosphorylated peptides corresponding to the Raplb carboxyl-terminus cancompete with intact Raplb for interaction with GDS, whereas other peptideslacking either modification can not (131).

The Rho proteins are also localized to the Golgi (91), and also appear require geranylgeranylation for association with membranes and with ex-change factors. Mature RhoA protein binds to membranes, and associateswith two GDP/GTP exchange factors, the stimulatory factor GDS and theinhibitory factor GDI. In contrast, bacterially expressed RhoA protein, whichlacks carboxyl-terminal processing, is defective in all three of these interac-tions (59). Prenylation may be generally involved in mediating interactionsbetween other small GTP-binding proteins and exchange factors (7).

OTHER MOT1FS A third group of small GTP-binding proteins contain thecarboxyl-termina] sequences C-C, C-X-C, or C-C-X-X. These proteins includemost of the Rab/Ypt proteins, which are involved in regulating intracellulartrafficking, and are present both in the cytoplasm and associated with distinctmembrane compartments (9). Several of these proteins, including mammalianRablb and Rab3a and YPT1, YPT3, and YPT5 of Schizosaccharomycespombe, have been shown to be geranylgeranylated in vivo (73, 102); genetic


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.



and biochemical evidence indicates that others, including the mammalianRablb, Rab2, and Rab5 (75), and YPTI of S. cerevisiae (119), aregeranylgeranylated as well. Using a variety of approaches, including mutationof the terminal cysteine residues (97), depletion of intracellular isoprenoidpools (73, 75), expression in bacterial cells (7), and analysis of geranyl-geranyltransferase mutants (119), geranylgeranylation has been shown to essential for association of these proteins with membranes.

Although proteins containing C-C, C-X-C, and C-C-X-X motifs are allgeranylgeranylated, some evidence indicates that the modifications directedby these sequences may differ in important respects. Mammalian Rab3A,which contains the C-X-C motif, is geranylgeranylated at both carboxyl-ter-minal cysteine residues and is carboxymethylated (38). This pattern modification may be common to all C-X-C proteins; another protein contain-ing this sequence motif, the YPT5 protein of S. pombe, is also geranyl-geranylated and carboxymethylated (102). In contrast, proteins with carboxyl-terminal C-C motifs may be modified differently. The YPTI protein of S.cerevisiae, with a C-C motif, is palmitoylated, and this palmitoylation isdependent on the presence of the carboxyl-terminal cysteines (97). However,no methylation of the C-C proteins YPT1 and YPT3 of S. pombe is detectedunder the labeling conditions used to detect methylation of Y PT5 (102). Thus,C-C proteins may be geranylgeranylated and palmitoylated, but not methyl-ated. The sites of these modifications are undetermined. Since, in vitro, theyeast and human methyltransferases appear to methylate essentially anyprenylcysteine with a free carboxyl (63, 132), the apparent lack of methylated carboxyl-terminus suggests that the carboxyl-terminal cysteine ofC-C proteins may not be prenylated. Thus, a reasonable hypothesis compatiblewith these data is that the carboxyl-terminal cysteine is palmitoylated, andthat the adjacent cysteine is geranylgeranylated. A novel carboxyl-terminalsequence, C-C-P, is found in Rah, a mammalian small GTP-binding proteinrelated to Rho proteins (99). It is tempting to speculate that this may representa new prenylation motif.

Lipopeptide Pheromones

Haploid fungal cells secrete mating pheromones that trigger cell fusion withcells of the opposite mating type to form diploid zygotes. Several fungal matingpheromones are lipopeptides, consisting of an oligopeptide to which a prenylgroup is covalently attached in a thioether linkage to a carboxyl-terminalcysteine residue. The most extensively characterized is the mating pheromonea-factor of yeast. Yeast contain two a-factor genes, MFA1 and MFA2, whichencode polypeptide precursors of 36 and 38 amino acids, respectively, eachwith a carboxyl-terminal CaaX sequence (15, 93). The a-factor precursorundergoes extensive processing independently of the classic secretory pathway


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


214 SCHAFER & RINE

before being secreted as the mature 12-amino acid lipopeptide (134). Analysisof the structure of mature a-factor reveals that four modifications occur in theprocessing of the precursor to yield active a-factor: (a) proteolytic removalof a 21-amino acid pro sequence to yield the mature amino terminus of thepeptide, (b) attachment of the famesyl group to a cysteine residue three aminoacids from the carboxyl-terminus, (c) proteolytic removal of the three aminoacids distal to that cysteine, and (d) methylation of the newly generatedcarboxyl-terminus (5).

The four steps in a-factor maturation occur in a defined order. In vitrostudies of a-factor precursor processing have demonstrated that farnesylationis a precondition for carboxyl-terminal proteolysis (8) and methylation (61).Furthermore, cells depleted for mevalonate in vivo accumulate an a-factorspecies with the electrophoretic mobility characteristic of the a-factor primaryprecursor (124, 133), suggesting that farnesylation is a precondition foramino-terminal proteolysis. Thus, farnesylation is the obligatory first step ina-factor precursor processing, followed by proteolysis and methylation at thecarboxyl-terminus, and proteolysis at the amino terminus. It is not knownwhether amino-terminal proteolysis and carboxyl-terminal proteolysis occurin a fixed order. Farnesylation and/or proteolysis is also apparently requiredfor the secretion of a-factor, since unfarnesylated, unproteolyzed a-factorprecursors accumulate in the cytosol (124), whereas unmethylated a-factor readily secreted (90). The farnesyl group is also important for the ability a-factor to induce growth arrest and morphological alteration of yeast cellsof the a mating type (5).

Interesting variations on this process are seen in the processing of otherpheromones. Rhodotorucine A, a lipopeptide pheromone from the basidio-mycetous jelly fungus Rhodosporidium toruloides, is an 11-amino acid peptidewith a carboxyl-terminal cysteine covalently bound to a farnesyl group in athioether linkage (69). In contrast to a-factor, the rhodotorucine A carboxyl-terminal cysteine is not methylated. Three genes, RHA1, RHA2 and RHA3,encode rhodotorucine A precursors consisting of three to five tandem copiesof the rhodotorucine A peptide (2). These peptide sequences are separated a four-amino acid spacer peptide, (C)-T-V-A/S-K. The rhodotorucine precursor is apparently cleaved to produce multiple pheromone peptides, aprocess reminiscent of a-factor processing in S. cerevisiae and neuropeptideprocessing in mammals. Maturation of rhodotorucine A may involve cleavageof the large precursor to produce multiple peptides containing canonical CaaXsequences, which are then processed in a manner similar to a-factor precursor.Alternatively, farnesylation of internal cysteine residues could occur prior toproteolytic maturation. In either case, rhodotorucine A provides a precedentfor prenylation of internal cysteine residues (see also Proteases below)


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.



The processing of the tremerogen A-10 pheromone, produced by the jellyfungus Tremella mesenterica, is another variation on the maturation processof a-factor. Tremerogen A-10 is synthesized as a large precursor of unknownprimary sequence. The secreted pheromone is a farnesylated, carboxy-methylated, 10-amino acid lipopeptide with an apparent molecular weight of1.5 kd. Interestingly, the farnesyl group contains an alcohol group at position30 (122). Compactin, an inhibitor of isoprenoid biosynthesis, inhibits matu-ration of tremerogen A-10 precursor, resulting in the accumulation of a 28-kdspecies that is presumably unfarnesylated. However, tunicamycin, an inhibitorof glycosylation, and monensin, an inhibitor of Golgi transit, also inhibitprocessing, and result in the accumulation of lower molecular weightintermediates, implying that the precursor is glycosylated and processed inthe Golgi (96). Tremerogen A-10 is therefore a novel case of a farnesylatedpolypeptide that, unlike a-factor and the small GTP-binding proteins, is alsoprocessed in the secretory pathway. Since farnesylation is an early processingstep that may occur in the cytoplasm, prenylation may mediate entry oftremerogen A-10 into the endoplasmic reticulum.

In addition to these examples, a wide variety of evolutionarily divergentfungal species secrete prenylated pheromones. For example, h" cells of thefission yeast S. pombe secrete a nine-amino acid methylated lipopeptideknown as M-factor, which is encoded by two genes, MFml and MFm2. Theprecursors encoded by these genes contain CaaX sequences, and appear toundergo the same posttranslational modifications as a-factor (32). A numberof mating factors from jelly fungi are prenylated peptides; some, such astremerogen a-13 from T. mesenterica, contain a thioether-linked farnesylgroup, whereas others, such as A-9291-I from T. brasiliensis, contain anoxidized farnesyl group (67, 69, 122). In addition, the mating type of thebasidiomycetous smut fungus Ustilago maydis is determined in part by whichof two genes, mfal or mfa2, is present at the A mating-type locus; both mfaland tufa2 encode short polypeptides that contain canonical CaaX sequences(14). Thus, prenylated pheromones may be ubiquitous among the fungi. It unclear whether prenylated peptide hormones are found in other organisms.

Nuclear Lamins

Lamins are nuclear proteins that are a major component of the karyoskeleton.Lamins are thought to form intermediate filaments that are associated withthe inner surface of the nuclear membrane and mediate nuclear breakdownand reformation during mitosis. Lamin proteins fall into three types, A-type,B-type, and C-type. B-type lamins are always associated with the nuclearmembrane, whereas A- and C-type lamins are found in a soluble form duringmitosis (42). Both A- and B-type lamins undergo posttranslational processing


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


216 SCHAFER & RINE

involving prenylation, and both contain a carboxyl-terminal CaaX sequence,In B-type lamins, posttranslational processing resembles the processing ofRas proteins. The cysteine residue of the CaaX sequence is farnesylated (139),and the three amino acids distal to that cysteine are removed (137). There also evidence that lamin B is carboxymethylated, although the methylationappears to be transient, and actually may play a role in regulating thedisassembly of the nucleus during mitosis (25). As in the cases of Ras anda-factor, prenylation appears to be the first step in processing of lamin B, anda precondition for other processing steps (11,137). Since the presence of theCaaX sequence is necessary for nuclear membrane targeting of lamin Bmolecules (58), prenylation and perhaps proteolysis and methylation are alsonecessary for the targeting of the lamin proteins to the nuclear membrane.Depletion of nonsterol isoprenoid pools in animal cells causes cells to arrestspecifically at the G1/S transition in the cell cycle (111). it has been proposed(46) that this requirement could in fact reflect a requirement for lamins at thisstage of the cell cycle.

The processing of lamin A is more complex. The lamin A precursor isinitially modified with a famesyl group to produce an intermediate known asprelarnin A. Subsequently, however, the 18 carboxyl-terminal amino acids,including the farnesylated CaaX sequence, are removed by an endoprotease;thus, the mature form of lamin A that is assembled into the nuclear laminais not prenylated (I 1). Mevalonate-starved cells accumulate unfarnesylatedlamin A precursor in nuclear inclusions, and this precursor is rapidly processedand assembled into the lamina upon addition of mevalonate. On the basis ofthis observation, it has been suggested that farnesylation of lamin precursorsoccurs not in the cytosol, but in the nucleus (82). Interestingly, thecarboxyl-terminal region of the lamin A precursor is apparently not requiredfor assembly of lamin A into the lamina. This has led to the speculation thatthe prenylated lipopeptide released in lamin A processing may itself have abiological function (82).

G-proteins

G-proteins are heterotrimeric GTP-binding proteins that mediate signaltransduction between transmembrane receptors and intracellular second mes-sengers. The ~/ subunits of trimeric G-proteins contain either CaaX or CaaLconsensus sequences. Several of these ~/subunits are known to be prenylated.The ~ subunit of bovine transducin, the G-protein involved in visual signaltransduction, contains a farnesyl group covalently linked to the carboxyl-ter-minal cysteine residue (44, 80). Transducin ~/is also modified by proteolysisand carboxymethylation. In addition, STE18, the ~/subunit of the G-proteininvolved in yeast mating, requires a prenyl modification, probably farnesyl,for membrane localization and functional activity (41). A number of uniden-


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


tified G~/ subunits from bovine brain and from rat tissue culture cells aregeranylgeranylated (84, 100, 142). In the G~/ subunits of bovine brain, theprenylated cysteine is apparently carboxymethylated (142).

The famesylation of bovine transducin is of particular interest becauseseveral other proteins involved in visual signal transduction in the rod outersegment (ROS) membrane are prenylated. For example, the cGMP phos-phodiesterase, which responds to activated transducin by degrading thesecond-messenger cGMP, is a heterotetramer composed of an or, 13, and two~/ subunits. The ot subunit of this enzyme is farnesylated and carboxy-methylated, whereas the [3 subunit is geranylgeranylated, but apparently notmethylated (4). In addition, rhodopsin kinase, which is involved in theattenuation of rhodopsin activation, is also famesylated and car-boxymethylated (66). Rhodopsin itself contains a novel type of prenylmodification: the covalent attachment of 11-cis-retinal, a molecule derivedfrom carotenoids, to a lysine residue of the opsin protein. Rhodopsin is alsopalmitoylated, a modification common to many prenylated proteins (104).The involvement of several prenylated proteins in a. single signal transductionpathway suggests that this modification may mediate colocalization of andinteraction between these proteins.

At least two Got subunits contain carboxyl-terminal sequences that resemblethe CaaX motif. However, neither of these proteins is prenylated in vivo (84,123). Since mutant versions of these tx subunits containing the sequenceC-V-L-S in place of the normal sequence C-G-L-F are prenylated (68), it hasbeen suggested that the glycine residue prevents prenylation of these proteins.However, Ras mutant proteins with the sequence C-G-L-F are capable ofundergoing prenylation in vivo (71). Therefore, it remains unclear why thissequence does not direct prenylation of the Got subunit.

PROTEINS INVOLVED IN PRENYL-DEPENDENTPROCESSING

The maturation of prenylated proteins involves many variations on a singletheme, rather than a single series of processing reactions. All knownprenylated proteins are initially modified by the attachment of a farnesyl orgeranylgeranyl group in a thioether linkage to a cysteine residue at or nearthe carboxyl-terminus. Subsequently, many of these proteins are furthermodified by proteolytic processing. Finally, the carboxyl-terminal regions ofmany prenylated proteins are targets for reversible modifications such aspalmitoylation, methylation, and phosphorylation. The specific proteins thatdirect these modifications are critical for mediating the common and divergentbiological properties of the processed carboxyl-termini.


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


218 SCHAFER & RINE

Protein Prenyltransferases

BIOCHEMICAL STUDIES Several enzymes, termed protein prenyltransferases,which catalyze the attachment of a specific prenyl moiety to specific proteinprecursors, have been identified and characterized. At least three distinctprotein prenyltransferase activities are present in yeast and mammalian cells.Protein farnesyltransferases (PFT) are soluble enzymes that catalyze thefarnesylation of human K-ras (86, 112), a-factor precursor of yeast (125),and RAS2 of yeast (48). The substrates for these enzymes are the farnesyldonor farnesyl diphosphate (FPP) and a protein precursor containing an intactCaaX sequence. The farnesyltransferase in rat brain has been purified tohomogeneity. The enzyme is a heterodimer composed of nonidentical o~ and13 subunits; the genes encoding these proteins have recently been sequenced(113). A second activity, termed protein geranylgeranyltransferase I (PGT has also been identified in soluble extracts. This enzyme, which is chromato-graphically (22, 144) and genetically (40) separable from PFT, catalyzes geranylgeranylation of peptides containing CaaL sequences, usinggeranylgeranyl diphosphate (GGPP) as a substrate. This enzyme has beenonly partially purified. However, since antisera directed against the mamma-lian PFT ot subunit can immunodeplete PGT I activity from crude extracts, ithas been hypothesized that mammalian PGT I is also a heterodimer with anct subunit that is shared with PFT (127). A third enzyme, PGT II, catalyzesthe geranylgeranylation of protein precursors with C-C carboxyl-terminalmotifs (98). This enzyme is chromatographically separable from both PFTand PGT I. Activities have also been identified in mammalian cells thatcatalyze the geranylgeranylation of C-X-C and C-C-X-X motifs (60, 75).Although both activities are separable from PFT and PGT I, it is not clearwhether either of these activities is distinct from PGT II. In addition, theobservation that farnesylation of prelamin A apparently occurs in the nucleus,whereas Ras farnesylation occurs in the cytoplasm, may suggest the existenceof a distinct nuclear farnesyltransferase (82). In this regard, it has beenreported that farnesyltransferase activity from bovine extracts fractionates astwo peaks on gel filtration chromatography; perhaps one of these peakscorresponds to the putative lamin farnesyltransferase (86).

The substrate specificities of the protein prenyltransferase enzymes areclearly critical for the attachment of the correct lipid modification to specificprenylated proteins. Current evidence suggests that the specificities of PFTand PGT I are largely determined by the particular amino acids present inthe CaaX/CaaL motif itself. Comparison of the sequences of knownfarnesylated and geranylgeranylated proteins led to the hypothesis that thecarboxyl-terminal amino acid of the protein precursor determined whethera particular protein was farnesylated or geranylgeranylated. According to


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.



this hypothesis, proteins containing S, C, M, Q, or A in this position aresubstrates for PFT, whereas proteins with L or F are substrates for PGT I.This hypothesis has been supported most convincingly for PFT in competitionexperiments using synthetic CaaX-related peptides (114). The in vitrospecificity of PGT I also conforms largely to in vivo predictions (98).Interestingly, both PFr and PGT I can act weakly in vitro on the preferredsubstrates of the other enzyme (144).

The PGT II enzyme shows stalking substrate specificity. Intriguingly, thesequences required for geranylgeranylation by PGT II, unlike PFT and PGTI, are not exclusively located at the carboxyl-terminus (98). Geranylgeranyla-tion by PGT II is not competed by peptides corresponding to the carboxyl-termini of PGT II substrates, nor are these carboxyl-terminal peptides usedas substrates by the enzyme. Furthermore, protein precursors containing C-Ccarboxyl-termini are not weakly prenylated by PFF or PGT I. Thus, therecognition of protein substrates by PGT II may be mechanistically dissimilarto substrate recognition by PFT and PGT I. The internal sequences requiredto direct geranylgeranylation by PGT II have not been identified.

GENETIC STUDIES In S. cerevisiae, studies of the genes encoding proteinprenyltransferase subunits have been instrumental in revealing many of theroles of protein prenylation. The yeast genes RAM1 (125) and RAM2 (48) required for PFT activity. Bacterially expressed RAM1 and RAM2 proteins,upon mixing, can reconstitute active yeast PFT enzyme; thus, these proteinsapparently comprise the subunits of the yeast heterodimer (57). The proteinsequences of RAM1 and RAM2 are homologous to the mammalian PFT 13and ot subunits, respectively (77). Mutations in RAM1 have been isolatedindependently by several groups based on several phenotypes: suppression ofRAS2 activation (43, 109), a defect in the production of mature a-factor (138),and interference with the function of the G’~ subunit STE18 (94). Thus, theyeast PFT is required for the functions of all the known yeast farnesylatedproteins (it is worth noting that lamins have not been identified in yeast cells).Mutations in RAM2 also suppress RAS2 activation and block production ofactive a-factor (57).

Yeast contains two additional genes that share sequence homology withRAM1. One of these is known as CDC43 or CALl. Mutations in this genewere identified on the basis of two distinct phenotypes, call mutants arrestspecifically at the G2/M transition unless grown in the presence of highlevels of calcium (106). These mutants are suppressed by overexpressionof yeast (RHO2), which contains the carboxyl-terminal sequence C-I-V-L,suggesting RHO proteins as possible targets of the CDC43 enzyme (Y.Ohya, personal communication). The cdc43 mutants, in contrast, display atemperature-sensitive defect in eytokinesis and the establishment of cell


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


220 SCHAFER & RINE

polarity (1). Mutations in CDC42 and RSR1, which encode proteins thatcontain the carboxyl-terminal sequence C-T-I-L, also cause this phenotype,suggesting that these proteins are also CDC43 substrates (12). Theseobservations suggest that CDC43/CAL1 encodes the [3 subunit of a PGTI-type protein prenyltransferase. Experiments using polypeptide substratescontaining CaaL sequences demonstrate that CDC43/CAL1 is indeed requiredfor PGT I activity (40).

Two yeast genes may encode PGT Ict subunits. The observation thatmammalian PFT and PGT I contain a shared o~ subunit suggested that RAM2might bc the tx subunit for yeast PGT I, and the observation that coexpressionof CDC43 and RAM2 in bacterial cells results in the production of active PGTI enzyme supports this hypothesis (J. B. Gibbs, personal communication).However, the ram2-1 point mutation caused only a twofold decrease in PGTI activity in vitro (40). This may suggest that multiple yeast genes encodePGTot subunits. Interestingly, a second gene, known as MAD2, has beenisolated with sequence similarity to rat PFTet and RAM2. mad2 gain-of-func-tion mutations were first identified on the basis of a defect in the regulationof the G2/M transition (81). Loss-of-function mutations in this gene causearrest at G2/M, a phenotype similar to that of call mutants (R. Li, personalcommunication). Furthermore, the MAD2 protein sequence contains calcium-binding motifs. These genetic data suggest that CDC43/CAL1 may formheterodimers with both RAM2 and MAD2. Thus, a subunits, like [3 subunits,may be shared between protein prenyltransferases. The CDC43-RAM2heterodimer may prenylate proteins involved in cytokinesis and cell polarity,such as CDC42, whereas CDC43-MAD2 may prenylate proteins such asRHO1 and RHO2 that are involved in the control of passage through G2/M.The cdc43 and call point mutations may specifically affect interaction of theCDC43 protein with RAM2 or MAD2, respectively. Intriguingly, these dataalso imply that the activity of the CDC43-MAD2 enzyme may be regulatedin a cell cycle-specific manner by intracellular calcium signals.

A third gene that encodes a protein with similarity to [3 subunits, knownas BET2, was identified on the basis of mutations that affect the secretorypathway. Since these mutations interact genetically with mutations in YPT1and SEC4, both of which contain the carboxyl-terminal sequence C-C, BET2is thought to encode the [3 subunit of the yeast PGT II (119). bet2 mutationsresult in a defect in the membrane localization of both YPT1 and SEC4proteins. In addition, cell extracts of bet2 mutants lack PGT II activitymeasured using bacterially expressed YPT1 precursor (98). No promisingcandidates for a PGT II ot subunit have been identified, ram2 mutants displayno detectable defect in PGT II activity (98), and no genetic interactions havebeen observed between mad2 and bet2 mutations (R. Li, personal communi-cation). Thus, the ot subunit for this enzyme may be encoded by an as yet


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.



unidentified gene. Alternatively, the inability of carboxyl-terminal peptidesto compete for PGT II activity in vitro suggests that PGT II may differ inreaction mechanism, and perhaps subunit structure, from PFT and PGT I.Preliminary evidence is consistent with a heterotrimeric structure for theRab3A PGT II (60). Therefore, the putative other subunit(s) of PGT II mightnot be homologous to previously characterized ~x subunits.

Some prenylated proteins may be processed by multiple protein pre-nyltransferases. For example, yeast RAS2 protein, which is primarily far-nesylated by the RAM1-RAM2 farnesyltransferase, also appears to bemodified to a lesser, degree by a second prenyltransferase activity. Yeast cellscarrying a deletion of RAM1 are viable, and a small but detectable amountof RAS2 protein in these strains is in the membrane-associated, mature form(125). In contrast, a deletion of RAM2 is lethal, and no membrane-associatedRAS2 is detected in ram2 mutant strains (57). These data suggest that RAM 1-independent, RAM2-dependent activity is capable of weakly prenylat-ing RAS2 in vivo. The [3 subunit of this enzyme may be CDC43, sinceoverexpression of CDC43 can suppress the growth defect of a raml mutant.In addition, raml cdc43 double mutants are inviable, but can be rescued byoverexpression of RAS2 (C. Trueblood & J. Rine, unpublished data). SinceRAM2 and CDC43 appear to comprise a PGT I enzyme, some RAS2 proteinmay be geranylgeranylated in vivo; determination of the nature of the prenylmoiety of RAS2 in ram1 strains should address this possibility.

Taken together, these genetic data suggest a model in which the [3 subunitsof prenyltransferases confer protein substrate specificity, whereas et subunitsprovide catalytic function. This model is consistent with biochemical evidencethat PFTot, but not PFT[3, binds the peptide substrate (113). Thus, sharing I~ subunits may serve to generate divergent prenyltransferase specificities,whereas sharing of o~ subunits may generate multiple prenyltransferaseisozymes with common specificity but divergent regulatory properties.

Proteases

The maturation of many prenylated proteins involves specific proteolyticprocessing. In particular, farnesylated and geranylgeranylated proteins withcarboxyl-terminal CaaX sequences are further modified by the proteolyticremoval of the three carboxyl-terminal amino acids, a modification necessaryfor subsequent carboxymethylation. A carboxyl-terminal protease activityactive on Ras precursors is present in mammalian cell membranes (8, 52).Thus, whereas famesylation of Ras precursors occurs in the cytosol, proteol-ysis most likely occurs at a membrane organelle, perhaps the plasmamembrane. The carboxyl-terminal protease is an endopeptidase (8), andrequires a farnesylated substrate. The residue one-amino acid from thecarboxyl-terminus (i.e. the a2 residue of the CalazX sequence) appears to


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


222 SCI-IAFER & RINE

important for the specificity of the mammalian carboxyl-terminal protease, asmutations in this residue can prevent proteolysis, but not farnesylation, ofK-ras in vivo (71). Yeast extracts also contain a membrane-associatedcarboxyl-terminal protease activity (8, 62). Like the mammalian enzyme, thisenzyme is an endopeptidase that acts only on prenylated substrates. Themembrane-associated protease is not encoded by any known gene requiredfor a-factor or RAS2 processing. Geranylgeranylated peptides are alsoproteolyzed by a membrane-associated protease; it is presently unclear whetherthis activity is distinct from the famesyl-dependent membrane protease (8).

In addition to this membrane protease, a carboxyl-terminal protease activeon a-factor is present in the soluble fraction of a yeast extract (8, 62), andhas been partially purified (62). This soluble enzyme is a processiveexopeptidase that is not farnesyl-dependent. In fact, this enzyme can removethe farnesylated cysteine itself unless it is further modified by carboxymethyla-tion. Therefore, it is not clear that this enzyme is relevant to the processingof prenylated proteins in vivo.

The maturation of lamin A, a-factor, and M-mating pheromone of S. pombealso involves a second type of proteolytic processing. These proteins requireendoproteases that cleave the polypeptide precursor at a site amino-terminalto the prenylated cysteine residue, resulting in the release of a prenylatedoligopeptide. In vivo, the amino-terminal proteases involved in lamin A anda-factor maturation are active only on farnesylated precursors (11, 124).Presently, little is known about these enzymes. A yeast gene, STE19, hasbeen identified that is required for the production of active a-factor, and isnot deficient in the farnesyltransferase, carboxyl-terminal protease, or methyl-transferase activities (M. N. Ashby & J. Rine, unpublished). The amino-ter-minal proteolysis of a synthetic substrate based on a-factor can be catalyzedin vitro by a crude yeast extract (88). Using this assay, it should be possibleto determine whether STE19 encodes the amino-terminal protease enzyme.

In addition to the proteases involved in the maturation of prenylatedproteins, there are at least two examples of prenyl-dependent degradativeproteases. Strains of R. toruloides of the a-mating type produce an en-doprotease that cleaves the prenylated mating factor rhodotorucine A (95).This cleavage is necessary for metabolism of and response to rhodotorucineA by a cells. Similarly, S. cerevisiae strains of o~-mating type produce a cellsurface-associated endoprotease that degrades a-factor (89). This proteaseinactivates the a-factor pheromone molecule and promotes recovery fromgrowth arrest. Both enzymes specifically degrade famesylated peptides;peptide analogues that lack the farnesyl group, or are modified with a differentalkyl group, are poor substrates for these enzymes. The sxal gene from S.pombe may encode a prenyl-dependent degradative protease; sxal mutantsare hypersensitive to the M-factor lipopeptide pheromone (65). The possibility


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.



that prenyl-dependent degradative proteases may play a role in the turnoverof prenylated proteins in other organisms remains unexplored.

The primary sequence of the rhodotorucine A precursor suggests theexistence of a protease that releases short oligopeptides, which are furtherprocessed to produce farnesylated pheromone molecules (2). The details this process are unclear; it is not known, for example, whether proteolysisoccurs before or after famesylation, or whether the earlier step is a precon-dition for the later step. At present, rhodotorucine A provides a novel exampleof prenyl modification of cysteine residues located in an internal region ofthe primary precursor. If the prenylation of internal cysteine residues is ageneral biological phenomenon, the details of this type of proteolyticmaturation will acquire particular biological significance.

Reversible ModificationsThe carboxyl-termini of prenylated proteins are often targets for additionalmodifications that, unlike proteolysis and prenylation, are reversible. Inparticular, palmitoylation has been implicated in the functions of severalprenylated proteins, including some Ras, Rap, and Rab proteins. Amongprenylated proteins, palmitoylation has been studied most extensively for thehuman Ras proteins. The palmitoyl modification of these proteins has a fasterturnover time than the proteins themselves; thus, Ras proteins appear to bepalmitoylated and depalmitoylated in a dynamic fashion (50).

A palmitoyltransferase activity has been identified in mammalian cellextracts that can palmitoylate cytosolic N-ras precursors, using palmitoyl-CoAas a substrate (49). This activity is associated with Golgi-rich membranefractions. Bacterially expressed Ras precursors are poor substrates, implyingthat the palmitoyltransferase may require a farnesylated substrate. Since themature N-ras protein is associated with the plasma membrane, and palmitoyla-tion is dynamic, the apparent lack of palmitoyltransferase activity in theplasma membrane is perhaps surprising. Possibly this palmitoyltransferaseactivity may not be relevant to Ras processing in vivo. It is interesting tonote, however, that yeast YPT1 is palmitoylated and localized to the Golgi(97, 128). It is possible, therefore, that this palmitoyltransferase is actuallyinvolved in palmitoylation of Rab/YPT proteins. Clearly, mutants in thepalmitoyltransferase gene would be invaluable in identifying the relevanttargets of this enzyme.

Consideration of the mechanism of the palmitoylation of rhodopsin, anintegral membrane protein, may provide insight into the mechanism of Raspalmitoylation. Rhodopsin is posttranslationally modified by a thioester-linkedpalmitoyl moiety at the rod outer segment, an organelle derived from theplasma membrane (104). Palmitoylation of rhodopsin, like Ras, appears be dynamic. Surprisingly, the palmitoylation of rhodopsin may occur by a


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


224 SCHAFER & RINE

nonenzymatic mechanism (103). Although rod outer segments can transferpalmitate from palmitoyl-CoA to rhodopsin, this activity is insensitive toboiling, and copurifies with rhodopsin itself. Energetically, it is possible thatboth the transfer of a palmitoyl moiety from palmitoyl-CoA to cysteine andthe subsequent hydrolysis of the cysteine thioester may not require an enzymefor catalysis. However, since palmitoyl is the only fatty acyl moiety that istransferred to rhodopsin, some mechanism must exist to provide this speci-ficity. Precedent exists for specific nonenzymatic modification of proteins;for example, the maturation of rhodopsin involves the nonenzymatic covalentattachment of a specific carotenoid molecule, ll-cis-retinal, to a specificamino acid of the opsin polypeptide (145). The parallels between palmitoyla-tion of rhodopsin and Ras suggest that Ras may become palmitoylated anddepalmitoylated at the plasma membrane by a similar nonenzymatic mecha-nism.

A second modification common to many prenylated proteins is methylation.For example, Ras proteins, B-type lamins, some trimeric G-proteins, andsome Rab proteins contain methylated carboxyl-termini. Whereas for Rasproteins the half-life of this modification is identical to that of the proteinitself (50), for several other prenyiated proteins, such as lamin B, the methylgroup is turned over more rapidly (25), suggesting that carboxymethylationcan be reversible. Reversible methylation of prenylated carboxyl-terminiwould presumably require the existence of a methyltransferase and amethylesterase enzyme. In yeast, a membrane-associated methyltransferasehas been identified that is encoded by a single gene, STE14 (61, 63). STE14encodes an integral membrane protein required for the methylation of a-factorand RAS2. Mutants defective in this gene are unable to produce active a-factor,but display no other obvious phenotypes. Thus, methylation may have littleimportance for Ras function, at least in yeast. It is worth noting, however,that proteins such as lamin B and "C-X-C" Rab proteins, for whichmethylation may be a reversible, regulatory modification, have not beenidentified in S. cerevisiae. Methyltransferase activity has also been detectedin crude membrane fractions from mammalian cells (132), and in rod outersegment (ROS) membranes (107). Interestingly, ROS membranes contain active methylesterase activity (107). Since methylation appears to play important role in the interaction of transducin with the activated form ofrhodopsin, these two enzymes may provide an important function in regulatingvisual signal transduction by controlling the methylation state of proteins suchas transducin, cGMP phosphodiesterase, and rhodopsin kinase.

The phosphorylation of Raplb protein provides an intriguing example of areversible modification mediating the effects of an intracellular secondmessenger on the biological properties of a prenylated protein. Raplb isphosphorylated on a specific carboxyl-terminal serine residue by cAMP-de-


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


pendent protein kinase (A-kinase) (56). This modification is important efficient interaction between the carboxyl-terminus of Raplb and the GDPexchange factor GDS (131). GDS stimulates both the exchange of GTP forGDP bound to Raplb and the dissociation of Raplb from intracellularmembranes. Thus, phosphorylation of Raplb by A-kinase results in anincrease in the GTP-bound versus the GDP-bound form of the protein, andtransfer of the protein molecule from a membrane location to the cytosol.Interestingly, Ras proteins from both yeast and humans are phosphorylatedon specific serine residues; in human N-Ras, the site of phosphorylationhas been localized to the carboxyl-terminus (27, 121). In RAS proteins yeast, only the membrane-bound form is phosphorylated. Perhaps phosphor-ylation also regulates the interaction of Ras proteins with guanine nucleotideexchange factors.

PRENYL-MEDIATED INTERACTIONS

Although considerable progress has been made recently in identifying theenzymes that direct protein prenylation and their protein targets, the ways inwhich prenyl modifications function remain largely unknown. In particular,although a primary function of prenylation for many modified proteins is todirect intracellular localization, the mechanisms of these targeting processes,and the cellular factors required for them, have not been determined. Basedon recent data (Figure 1), however, it is possible to speculate upon howprenyl-dependent protein targeting might occur. At least three general typesof models for prenyl-mediated protein targeting can be proposed.

Prenyl-protein Receptors

Prenylated carboxyl-termini could be directed to intracellular compartmentsby specifically localized membrane-bound receptors. This type of mechanismis likely to be responsible for the plasma membrane localization of myristoy-lated Src proteins (115). The putative myristoyl-Src receptor appears recognize both the lipid moiety and the amino acids in the protein amino-ter-minus (116). Similarly, hypothetical prenyl-protein receptors would mostlikely recognize both the prenyl moiety and some specific amino acid sequenceor protein structure, since proteins with identical lipid modifications can belocalized to different subcellular compartments.

The principle of interaction between prenylated polypeptides and specificreceptors has been clearly established by work on fungal mating factors andtheir receptors. The prenylated mating pheromone a-factor interacts with aspecific transmembrane receptor, STE3, to induce growth arrest and morpho-logical differentiation in cells of the ot mating type (51). Furthermore,experiments using synthetic analogues of a-factor have demonstrated that the


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


226 SCHAFER & RINE

Figure 1 Models for prenyl-dependent protein targeting. A. Prenyl-protein receptor model. Amembrane-localized receptor specifically recognizes lipid and protein structure. B. Dumb lipidmodel. The lipid group promotes nonspecific affinity for membranes; specific targeting is mediatedby protein-protein interactions. C. Smart lipid model. Specific lipid-lipid interactions mediateprotein localization; bars indicate boundaries of a membrane region of specific lipid composition.D. Secondary interaction model. A prenyl-dependent secondary modification mediates proteintargeting.

receptor is highly specific for both the peptide sequence and the posttranslatio-nal lipid modifications. Peptide analogues lacking an alkylated cysteine areessentially inactive. If the famesyl group is replaced with a differentthioether-linked alkyl group, the activity of the resulting peptide is reduced;some analogues, such as S-geranyl a-factor, retain as much as 50% of theactivity of authentic a-factor, whereas S-methyl a-factor is only 0.2% as activeas authentic a-factor in some assays (5, 87). Methylation is also importantfor the activity of a-factor; unmethylated a-factor analogues are generallyinactive in inducing a-factor arrest (87). a-factor is inactivated by proteolysis(89); therefore, the peptide must also be important for recognition by STE3.The effects of the pheromone rhodotorucine A of R. toruloides and thepheromone tremerogen A-10 of T. mesenterica, whose activities are also likelyto be receptor-mediated, also require both the peptide sequence and the lipidmodification for biological activity (136).

Interactions between the prenylated carboxyl-terminus of a-factor and aspecifically localized transmembrane protein may also be involved in theextracellular secretion of the pheromone. The secretion of farnesylateda-factor is known to require the activity of a transmembrane glycoprotein,STE6. STE6 was identified on the basis of mutations that rendered yeast cellsof the a-mating type mating-deficient (118). ste6 mutants are capable ofsynthesizing mature a-factor; however, this a-factor is not secreted (79). Thesequence of STE6 predicts that it encodes a membrane-spanning protein with


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.



significant sequence similarity to bacterial membrane transporters and, inparticular, mammalian P-glycoproteins (78, 92). Thus, STE6 is likely toencode a protein that pumps processed a-factor across the plasma membrane.It is not known whether STE6 protein specifically recognizes the farnesylgroup of a-factor. However, since mutations that block famesylation ofa-factor prevent its secretion by STE6, whereas a mutation that blocksmethylation does not (90), either farnesylation or one of the proteolyticmodifications must be essential for export by STE6.

Mammalian cells contain genes that are highly homologous to STE6, theMDR (multidrug resistance) or P-glycoprotein genes. Overexpression of theMDR1 gene in transformed cells results in resistance to a wide range ofstructurally divergent cytotoxic compounds, including vinblastin, vincristine,actinomycin, colchicine, and adriamycin (3, 28). This phenotype results froman increased efflux of drug from the treated cells, implying that the MDR1protein is pumping toxins out of the cell (45). Other MDR1 homologues areinvolved in translocation of MHC peptides into the ER lumen (36). One mightalso speculate that an MDR-like protein might mediate the apparent prenyl-dependent entry of tremerogen A-10 precursor into the secretory pathway.Although the idea of prenyl-mediated intracellular trafficking mediated byMDR homologues is reasonable, it is important to recognize that no MDRhomologue, including STE6, has been directly demonstrated to require aprenyl modification for membrane translocation at a peptide substrate.

Another candidate for a prenyl-protein receptor that mediates proteintargeting is the lamin B receptor, a 58-kd nuclear membrane protein. Thisreceptor mediates the association of lamin B with the nuclear membrane, aprocess that also depends on farnesylation (141). The gene for the lamin receptor, which has been cloned from chicken, is predicted to encode apolytopic integral membrane protein (140). However, it is presently unknownwhether recognition by the lamin B receptor specifically requires the farnesylmodification.

Lipid Interactions

A second possible mechanism for prenyl-mediated protein targeting isassociation of prenylated proteins with cellular membranes by interactionsbetween the prenyl group and membrane lipids. Models of this type includethe "dumb lipid" model, in which the prenyl group merely provides anonspecific lipophilic glue that provides affinity for membranes, but does notmediate targeting to a specific membrane organelle, and the "smart lipid"model, in which the lipid modification directs the protein to a specific localethrough a nonreceptor-based mechanism. A type of smart lipid model hasbeen proposed to explain the localization of GPI-linked proteins to plasma-membrane sites called caveolae (6). Clustering of GPI-linked proteins, such


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


228 SCHAFER & RINE

as the folate receptor in caveolae, is disrupted by drugs, such as filipin andnystatin, that bind cholesterol (120). Thus, it has been suggested that thespecific lipid composition of caveolae attracts GPI-linked proteins throughspecific lipid interactions.

A lipid interaction model is consistent with much of the information on thelocalization ofprenylated Ras-superfamily proteins. Clearly, farnesylation andthe other carboxyl-terminal modifications serve to increase the affinity of Rasproteins for membranes. For K-ras proteins synthesized in vitro, famesylation,proteolysis, and methylation all have the cumulative effect of increasing themembrane affinity of the Ras protein (52). Similarly, the membrane affinitiesof mutant H-ras proteins lacking palmitoylation sites and K-ras proteinslacking a polybasic region are markedly less than wild-type forms of theseproteins (53). Some experiments with mutant Ras proteins suggest thatprenylation is a nonspecific membrane-localization signal. For example, K-rasmutant proteins with a CaaL sequence identical to Rapla are geranyl-geranylated rather than farnesylated, yet are still localized to the plasm~membrane rather than the Golgi (53). Similarly, the biological activity Rapl a, a Golgi-associated protein that suppresses Ras activation, is unaffectedby replacing the geranylgeranyl modification with farnesyl (30). In addition,the yeast YPT1 protein, a member of the Rab family that contains thegeranylgeranylation sequence C-C at its carboxyl-terminus, retains biologicalactivity when this sequence is altered to produce a CaaX farnesylationsequence (97). These results suggest that targeting of these proteins to specificmembrane compartments does not involve the prenyl group, but rather theamino acid sequence of the proteins.

The carboxyl-termini of small GTP-binding proteins contain a hypervariableregion that may direct endomembrane targeting. Mutations that affect thepolybasic region of the hypervariable domain of geranylgeranylated K-ras donot affect membrane affinity in vitro, nor membrane partitioning in vivo.However, these mutant proteins are targeted neither to the plasma membrane,like K-ras, nor the Golgi, like Rapl a, but to many membranes in an apparentlynonspecific manner (53). The hypervariable domain of geranylgeranylatedRab proteins also appears to be critical to the localization of these proteins tospecific membrane compartments (24). Thus, in K-ras, Rapl, and at leastsome Rab proteins, prenylation, proteolysis, and methylation appear topromote nonspecific membrane association, whereas the hypervariable regiondirects localization to specific membranes.

Although the above results suggest that the farnesyl modification of K-rasis a dumb lipid, experiments with H-ras suggest that the prenyl moiety mayprovide more than nonspecific lipophilicity. Mutation of the carboxyl-terminalresidue of H-ras to leucine (30, 53), or deletion of the CaaX sequence andinsertion of an amino-terminal myristoylation sequence (20), results in the


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.



production of mutant H-ras proteins that are either geranylgeranylated ormyristoylated, rather than famesylated. Both mutant proteins are membraneassociated; for the geranylgeranylated form, the site of localization is theplasma membrane (53). In addition, activated forms of these proteins werecapable of inducing cellular transformation. However, the nonactivated formsof these proteins behaved differently from farnesylated H-ras; geranylgerany-lated H-ras exhibits a dominant negative phenotype (30), and myristoylatedH-ras has a transforming phenotype (20). Interestingly, H-ras, but not K-ras,is palmitoylated. Perhaps the interactions between the farnesyl group, thepalmitoyl group, and the membrane lipids serve to orient the protein in a waythat is important for wild-type, but not transforming, function. In this regard,note that palmitoylation has been suggested to be important for orientingrhodopsin in the rod outer segment membrane (103).

Indirect Prenyl-dependent Targeting

prenylation itself need not play a direct role in mediating protein targeting;in some cases, prenylation may instead be a precondition for a secondarymodification that actually directs protein localization. The processing of laminA provides a clear case of an indirect effect of prenylation on proteinlocalization. Famesylation of prelamin A is clearly required for the assemblyof normal lamin A protein into the nuclear lamina, since inhibition of thisprocess results in accumulation of prelamin A in nuclear inclusions (11).However, famesylation per se is not required for this process. Rather, theessential step is removal of the prelamin A carboxyl-terminus, which, in turn,requires farnesylation, as lamin A precursors lacking the carboxyl-terminalamino acids are assembled into the lamina in the complete absence ofprenylation (82). Prenyl-dependent proteolysis may also be essential for thesecretion of a-factor of yeast. Although the requirements for STE6-mediatedexport of a-factor have not been determined, it is possible that the criticalprocessing step required for this process is proteolysis of the amino-terminalpro region of the a-factor precursor.

Reversible prenyl-dependent modifications may be important for thelocalization and function of prenylated proteins. For example, methylationappears to be required for translocation of transducin ~’~ from the cytosol tothe ROS membrane, where it interacts with transducin tx and rhodopsin (105).Activation of rhodopsin with light stimulates the membrane localization ofmethylated transducin 13~/, a stimulation that requires transducin or. Thus, theeffect of methylation on the membrane-targeting of transducin 13~/appears toresult, at least in part, from enhanced protein-protein interactions betweentransducin ~x and 13~/. The existence of several proteins involved in visualsignal transduction, including rhodopsin kinase and cGMP phosphodiesterase,that are similarly farnesylated and carboxymethylated, suggests that reversible


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


230 SCHAFER & RINE

methylation may be critical for the interaction of all these proteins withrhodopsin and with one another. The example of transducin also suggests asimilar role for methylation and palmitoylation in the cycling of Rab and Rhoproteins between the cytosol and cellular membranes; this possibility remainsto be explored.

These models for prenyl-dependent protein targeting are not mutuallyexclusive. For example, the interaction of geranylgeranylated Raplb with itsGDS requires the geranylgeranyl group, the carboxyl-terminal peptide se-quence, and the phosphorylated serine residue. This interaction is correlatedwith transit of Raplb from the Golgi membrane to the cytosol (56). Thus,this process involves a protein-prenyl interaction that is analogous to the prenylreceptor model, yet also requires a secondary modification. The targeting ofmany prenylated proteins may likewise involve mechanisms that containelements of several of the models described here.

FUTURE PERSPECTIVES

Although much has been learned about protein prenylation in a relatively shortperiod of time, many critical questions remain unanswered. Clearly, theprecise functional role played by the prenyl group in protein localization andfunction is only beginning to be understood, and understanding this role istherefore a major focus of current research. In addition, the identification andcharacterization of new protein prenyltransferases and other enzymes involvedin prenyl-dependent modifications remains an important area of research, sincethe existence of new enzymes with novel substrate specificities appears likely.In particular, the combinatorial sharing of protein prenyltransferase subunitscould provide both diverse protein and lipid specificities and differentialregulation of protein prenylation. The identification of new prenylated proteinsmay also provide conceptual breakthroughs, particularly if new prenylmodifications or new modification sites are identified.

Understanding protein prenylation may provide insight into two relatedbiological phenomena of medical significance: oncogenesis, and cholesterolmetabolism. The Ras proteins are well known for their oncogenic potential;activated Ras genes have been found in a variety of types of human tumors,including colorectal tumors, lung adenocarcinomas, and pancreatic tumors(10). Prenylation has been shown to be important for the functions of Rasproteins, since inhibition of prenylation suppresses the phenotypes of consti-tutively active Ras proteins in yeast (109). Furthermore, genetic and pharma-cological inhibition of Ras processing in animal cells likewise suppressesRas-induced transformation (31, 64). Thus, a greater understanding of thecatalytic mechanism of protein prenyltransferases may aid in the design ofPFT inhibitors with potential utility for cancer treatment. Inhibitors ofisoprenoid biosynthesis may also be useful for cancer therapy; several studies


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.



have indicated that inhibitors of mevalonate synthesis suppress Ras-inducedcell transformation (110, 126, 33).

Protein prenylation may also be important for the regulation of HMG-CoA

reductase (HMGR), the rate-limiting step in cholesterol biosynthesis. TheHMGR inhibitor mevinolin is commonly used to treat hypercholesterolemia,

since mevinolin treatment results in increased synthesis of the LDL receptor(17, 83). However, depletion of nonsterol isoprenoids also causes an increase

in the translation of the HMGR message and a decrease in HMGR degradation,resulting in abnormally high levels of HMG-CoA reductase protein (16, 101).Thus, in patients undergoing treatment with mevinolin, the cholesterolbiosynthetic potential is actually elevated. Understanding the mechanism oftranslational and half-life control may provide insights into how this increase

in HMGR synthesis could be prevented. Prenylated proteins could mediateboth processes, although other models involving mevalonate-binding proteinsor prenylated mRNA are also plausible. Further studies should clarify themechanism of translational regulation of HMGR, and perhaps providepractical insights into more effective treatment of hypercholesteremia.

ACKNOWLEDGMENTS

We thank the many colleagues who provided reprints, recent preprints,unpublished observations, and stimulating discussions, and regret omissionsdemanded by space limitations. In addition, we thank the members of ourlaboratory for many suggestions and insights. Work in the authors’ laboratorywas supported by grants from the National Institutes of Health (GM35827),the California Tobacco-related Disease Research Program, and by a postdoc-toral fellowship from Merck, Inc.

Literature Cited

1. Adams, A. E. M., Johnson, D. I.,Longnecker, R. M., Sloat, B. F.,Pringle, J. R. 1990. CDC42 andCDC43, two additional genes involvedin budding and the establishment ofcell polarity in the yeast Saccharomycescerevisiae. J. Cell Biol. 111:131-42

2. Akada, R., Minomi, K., Kai, J.,Yamashita, I., Miyakawa, T., et al.1989. Multiple genes coding for pre-cursors of rhodotomcine A, a famesylpeptide mating pheromone of the ba-sidiomycetous yeast Rhodosporidiumtoruloides. Mol. Cell. Biol. 9:3491-98

3. Akayima, S., Fojo, A., Hanover, J.A., Pastan, I., Gottesman, M. M.1985. Isolation and genetic character-ization of human KB cell lines resistantto multiple drugs. Somat. Cell Mol.Genet. 11:117-26

4. Anant, J. S., Ong, O. C., Xie, H.,Clarke, S., O’Brien, P. J., et al. 1992.In vivo differential prenylation of ret-inal cyclic GMP phosphodiesterasecatalytic subunits. J. Biol. Chem.267:687-90

5. Anderegg, R. J., Betz, R., Carr, S.A., Crabb, J. W., Duntze, W. 1988.Structure of the Saccharomycescerevisiae mating hormone a- factor:identification of S-farnesyl cysteine asa structural component. J. Biol. Chem.263:18236-40

6. Anderson, R. G. W., Kamen, B. A.,Rothberg, K. G., Lacey, S. W. 1992.Potocytosis: sequestration and transportof small molecules by caveolae. Science255:410-11

7. Araki, S., Kaibuchi, K., Sasaki, T.,Hata, Y., Talai, Y. 1991. Role of the


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


232 SCHAFER & RINE

C-terminal region of smg p25A in itsinteraction with membranes and theGDP/GTP exchange protein. Mol. Cell.Biol. 11:1438-47

8. Ashby, M. N., King, D., Rine, J.1992. Endoproteolytic processing of afamesylated peptide in vitro. Proc.Natl. Acad. Sci. USA 89:4613-17

9. Balch, W. E. 1990. Low molecularweight GTP-binding proteins (LMGPs).involved in vesicular transport: binaryswitches or biological transducers?Trends Biochem. Sci. 15:469-72

10. Barbacid, M. 1987. The ras oncogene.Annu. Rev. Biochem. 56:779-828

11. Beck, L. A., Hosick, T. J., Sinensky,M. 1990. Isoprenylation is requiredfor the processing of the lamin Aprecursor. J. Cell Biol. 110:1489-99

12. Bender, A., Pfingle, J. R. 1989. Multi-copy suppression of the cdc24 buddingdefect in yeast by CDC42 and threenewly identified genes including theras-related gene RSR1. Proc. Natl.Acad. Sci. USA 86:9976-80

13. Beranger, F., Goud, B., Tavitan, A.,de Gunzberg, J. 1991. Association ofthe Ras-antagonistic Rapl/Krev-1 pro-teins with the Golgi complex. Proc.Natl. Acad. Sci. USA 88:1606-10

14. Bolker, M., Urban, M., Kahmann, R.1992. The a mating type locus of U.maydis specifies cell signalling com-ponents. Cell 68:441-50

15. Brake, A. J., Brenner, C., Najaxian, R.,Laybourne, P., Merryweather, J. 1985.Structure of genes encoding precursorsof the yeast peptide mating pheromonea-factor. In Protein Transport and Se-cretion, ed. M. J. Gething. Cold SpringHarbor, NY: Cold Spring Harbor Lab.

16. Brown, M. S., Faust, J. R., Goldstein,J. L., Kaneko, I., Endo, A. 1978.Induction of 3-hydroxy-3-methylglutaryl coenzyme A reductasein human fibroblasts incubated withcompaetin (ML-236B), a competitiveinhibitor of the reductase. J. Biol.Chem. 253:1121-28

17. Brown, M. S., Goldstein, J. L. 1986.A receptor-mediated pathway for cho-lesterol homeostasis. Science 232:34-47

18. Buss, J. E., Quilliam, L. A., Kato,K., Casey, P. J., Solski, P. A., etal. 1991. The COOH-terminal domainof the RaplA (Krev-1) protein isoprenylated and supports transforma-tion by an H-ras:RaplA chimeric pro-tein. Mol. Cell. Biol. 11:1523-30

19. Buss, J. E., Sefton, B. W. 1986.Direct identification of palmitic acidas the lipid attached to p21ras. Mol.Cell. Biol. 6:116-22

20. Buss, J. E., Solski, P. A., Schaeffer,J. R., MacDonald, M. J., Der, C. J.1989. Activation of the cellular proto-oncogene product p21ras by additionof a myristylation signal. Science 243:1600-3

21. Casey, P. J., Solski, P. A., Der, C.J., Buss, J. E. 1989. p21ras is modifiedby a farnesyl isoprenoid. Proc. NatlAcad. Sci. USA 86:8323-27

22. Casey, P. J., Thissen, J. A., Moomaw,J. F. 1991. Enzymatic modification ofproteins with a geranylgeranyl iso-prenoid. Proc. Natl. Acad. Sci. USA88:8361-65

23. Chardin, P. 1988. The ras superfamilyproteins. Biochimie 70:865-68

24. Chavrier, P., Gorvel, J-P., Stelzer,E., Simons, K., Gruenberg, J., et al.1991. Hypervariable C-terminal domainof rab proteins acts as a targetingsignal. Nature 353:769-72

25. Chelsky, D., Olson, J. F., Koshland,D. J. 1987. Cell cycle-dependentmethyl esterification of lamin B. J.Biol. Chem. 262:4303-9

26. Clarke, S. 1992. Protein isoprenylationand methylation at carboxyl-terminalcysteine residues. Annu. Rev. Biochem.61:355-86

27. Cobitz, A. R., Yim, E. H., Brown,W. R., Perou, C. M., Tamanoi, F.1989. Phosphorylation of RASI andRAS2 proteins in Saccharomycescerevisiae. Proc. Natl. Acad. Sci.USA 86:858-62

28. Comwell, M. M., Safa, A. R., Felsted,R. L., Gottesman, M. M., Pastan, I.1986. Membrane vesicles from multi-drug-resistant human cancer cells con-tain a specific 150- to 170-kDa proteindetected by photoaffinity labeling.Proc. Natl. Acad. Sci. USA 83:3847-50

29. Cox, A. D., Der, C. J. 1992. TheRas/cholesterol connection: implica-tions for Ras oncogenicity. Crit. Rev.Oncogen. In press

30. Cox, A. D., Hisaka, M. M., Buss,J. E., Der, C. J. 1992. Specificisoprenoid modification is required forfunction of normal, but not oncogenicras protein. Mol. Cell. Biol. In press

31. Crowell, P. L., Chang, R. R., Ren,Z., Elson, C. E., Gould, M. N. 1991.Selective inhibition of isoprenylationof 21-26-kDa proteins by the anti-carcinogen d-limonene and its metab-olites. J. Biol. Chem. 266:17679-85

32. Davey, J. 1992. Mating pheromonesof the fission yeast Schizosaccharomy-ces pombe: purification and structuralcharacterization of M-factor and iso-lation and analysis of two genes en-


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


coding the pheromone. EMBO J. 11:951-60

33. Defeo-Jones, D., McAvoy, E. M., Jones,R. E., Vuocolo, G. A., Haskell, K. M.,Wegrzyn, R. J., et al. 1991. Lovastatinselectively inhibits ras activation of the12-O-tetradecanoylphorbol- 13-acetetateresponse element in mammalian cells.Mol. Cell. Biol. 11:2307-10

34. Deschenes, R. J., Broach, J. R. 1987.Fatty acylation is important hut notessential for Saccharomyces cerevisiaeRAS function. Mol. Cell. Biol. 7:2344-51

35. Deschenes, R. J., Stirmnel, J. B., Clarke,S., Stock, J., Broach, J. R. 1989. RAS2protein of Saccharomyces cerevisiae ismethyl-esterified at its carboxyl termi-nus. J. Biol. Chem. 264:11865--73

36. Deverson, E. V., Gow, I. R., Coadwell,J., Monaco, J. J., Butcher, G. W.,et al. 1990. MHC class II regionencoding proteins related to the multi-drug resistance family of transmem-brane transporters. Nature 348:738-41

37. Epstein, W. W., Lever, D., Leining,L. M., Bruenger, E., Rilling, H. C.1991. Quandtation of prenylcysteinesby a selective cleavage reaction. Proc.Natl. Acad. Sci. USA 88:9668 70

38. Farnsworth, C. C., Kawata, M.,Yoshida, Y., Takai, Y., Gelb, M. H.,et al. 1991. C-terminus of the smallGTP-binding protein smg p25A con-tains two geranylgeranylated cysteineresidues and a methyl ester. Proc.Natl. Acad. Sci. USA 88:6196-200

39. Faust, J. R., Dice, J. F. 1991. Evidencefor isopentyladenine modification on acell cycle-regulated protein. J. Biol.Chem. 226:9961-70

40. Finegold, A. A., Johnson, D. I., Farns-worth, C. C., Gelb, M. H., Judd, S.R., et al. 1991. Geranylgeranyl;proteintransferase of Saccharomyces cere-visiae is specific for Cys-Xaa-Xaa-Leumotif proteins and requires the CDC43gene product, but not the DPRI geneproduct. Proc. Natl. Acad. Sci. USA.88:4448-52

41. Finegold, A. A., Schafer, W. R.,Rine, J., Whiteway, M., Tamanoi, F.1990. Common modifications of tri-meric G proteins and ras proteins:involovement of polyisoprenylation.Science 249:165-68

42. Franke, W. W. 1987, Nuclear laminsand cytoplasmic intermediate filamentproteins: a growing muitigene family.Cell 48:3-4

43. Fujiyama, A., Matsumoto, K., Tam-anoL F. 1987. A novel yeast mutantdefective in the processing of ras pro-teins: assessment of the effect of the

mutation on processing steps. EMBOJ. 6:223-28

44. Fukada, Y., Takao, T., Ohguro, H.,Yoshizawa, T., Akino, T., et al. 1990.Famesylated ~/-subunit of photoreeeptorG protein indispensible for GTP-bind-ing. Nature 346:658-60

45. Gerlach, J. H., Endicott, J. A.,Juranka, P. F., Henderson, G., Sa-rangi, F., et al. 1986. Homology be-tween P-glycoprotein and a bacterialhaemolysin transport protein suggestsa model for muitidrug resistance. Na-ture 324:485-89

46. Glomset, J. A., Gelb, M. H., Farnsworth,C. C. 1990. Prenyl proteins in eu-karyotic cells: a new type of membraneanchor. Trends Biol. Sci. 15:139-42

47. Goldstein, J. L., Brown, M. S. 1990.Regulation of the mevalonate pathway.Nature 343:425-30

48. Goodman, L. E., Judd, R. R., Farns-worth, C. C., Powers, S., Gelb, M.H., et al. 1990. Mutants of Saccharo-myces cerevisiae defective in thefarnesylation of ras proteins. Proc.Natl. Acad. Sci. USA 87:9665-69

49. Gutierrez, L., Magee, A. I. 1991. Char-acteri.z~tion of an acyltransferase actingon p21 -ras protein in a cell-free system.Biophys. Biochem. Acta 1078:147-54

50. Gutierrez, L., Magee, A. I., Marshall,C. J., Hancock, J. F. 1989. Post-trans-lational processing of p21ras is two-stepand involves carboxyl-methylationand carboxyl-terminal proteolysis.EMBO J. 8:1093-98

51. Hagen, D. C., McCaffrey, G., Spra-gue, G. F. Jr. 1986. Evidence theyeast STE3 gene encodes a receptorfor the peptide pheromone a-factor:gene sequence and implications for thestructure of the presumed receptor.Proc. Natl. Acad. Sci. USA 83:1418-22

52. Hancock, J. F., Cadwallader, K., Mar-shall, C. J. 1991. Methylation andproteolysis are essential for efficientmembrane binding of prenylated p21K-ras(B). EMBO J. 10:641--46

53. Hancock, J. F., Cadwallader, K.,Paterson, H., Marshall, C. J. 1991.A CAAX or a CAAL motif and asecond signal are sufficient for plasmamembrane targeting of ras proteins.EMBO J. 10:4033-39

54. Hancock, J. F., Magee, A. I., Childs,J. E., Marshall, C, J. 1989. All rasproteins are polyisoprenylated but onlysome are palmitoylated. Cell 57:1167-77

55. Hancock, J. F., Patterson, H., Mar-shall, C. J. 1990. A polybasic domainor palmitoylation is required in additionto the CAAX motif to localize p21ras


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


234 SCHAFER & RINE

to the plasma membrane. Cell 63:133-39

56. Hata, Y., Kaibuchi, K., Kawamura,S., Hiroyoshi, M., Shiratald, H., etal. 1991. Enhancement of the actionsof smgp21 GDP/GTP exchange proteinby the protein ldnase A-catalyzed phos-phorylation of smgp21. J. Biol. Chem.266:6571-77

57. lie, B., Chert, P., Chert, S-Y., Vancura,K. L., Micbaelis, S., et al. 1991. RAM2,an essential gene of yeast, and RAM1encode the two polypeptide componentsof the farnesyltransferase that pre-nylates a-factor and Ras proteins. Proc.Natl. Acad. Sci. USA 88:11373-77

58. Holtz, D., Araki Tanaka, R., Hartwig,J., McKeon, F. 1989. The CaaX motifof lamin A functions in conjugationwith the nuclear localization signal totarget assembly to the nuclear envelope.Cell 59:969-77

59. Hori, Y., Kikuchi, A., Isomura, M.,Katayama, M., Miura, Y., et al. 1991.Post-translational modifications of theC-terminal region of the rho proteinare important for its interaction withmembranes and the stimulatory andinhibitory GDP/GTP exchange pro-teins. Oncogene 6:515-22

60. Horiuchi, H., Kawata, M., Katayama,M., Yoshida, Y., Musha, T., et al.1991. A novel prenyltransferase for asmall GTP-binding protein having aC-terminal cys-ala-cys structure. J.Biol. Chem. 266:16981-84

61. Hrycyna, C. A., Clarke, S. 1990.Farnesyl cysteine C-terminal methyl-transferase activitity is dependent uponthe STE14 gene product in Sac-charomyces cerevisiae. Mol. Cell. Biol.10:5071-76

62. Hrycyna, C. A., Clarke, S. 1992.Maturation of isoprenylated proteins inSaccharomyces cerevisiae : multiple ac-tivities catalyze the cleavage of thethree carboxyl- terminal amino acidsfrom farnesylated substrates in vitro.J. Biol. Chem. In press

63. HITcyna, C. A., Sapperstein, S. K.,Clarke, S., Michaelis, S. 1991. TheSaccharomyces cerevisiae STE14 geneencodes a methylWansferase that me-diates C-terminal methylation of a-factor and RAS proteins. EMBO J.10:1699-709

64. Huang, S., Axelrod, D. E. 1991.Altered post-translational processing ofp21 ras oncoprotein in a transforma-tion-suppressed cell line. Oncogene6:1211-18

65. Imai, Y., Yamamoto, M. 1992. Schiz-osac~haromyces pombe sxal + andsxa2 encode putative proteases in-

volved in the mating response. Mol.Cell. Biol. 12:1827-34

66. Inglese, J., Glickman, J. F., Lorenz,W., Caron, M. G., Leikowitz, R. J.1992. Isoprenylation of a protein ki-nase. J. Biol. Chem. 267:1422-25

67. Ishibashi, Y., Sakagami, Y., Isogai,A., Suzuki, A. 1984. Structure oftremerogens A-9291-I and A-9291-VIII: peptidyl sex hormones ofTremella brasiliensis. Biochemistry 23:1399-404

68. Jones, T. L. Z., Spiegel, A. M. 1990.Isoprenylation of an inhibitory G pro-tein ct subunit occurs only upon mu-tagenesis of the carboxyl terminus. J.Biol. Chem. 265:19389-92

69. Kamiya, Y., Sakurai, A., Tamura, S.,Takahashi, N., Abe, K., et al. 1978.Structure of rhodotorucine A, a novellipopeptide, inducing mating tube for-mation in Rhodosporidium toruloides.Biochem. Biophys. Res. Commun. 83:1077-83

70. Katayama, M., Kawata, M., Yoshida,Y., Horiuchi, H., Yamamoto, T., etal. 1991. The post-translationally mod-ified C-terminal structure of bovineaortic smooth muscle rhoA p21. J.Biol. Chem. 266:12639-45

71. Kato, K., Cox, A. D., Hisaka, M.M., Graham, S. M., Buss, J. E., ctal. 1992. Isoprenoid addition to rasprotcin is the critical modification forits membrane association and trans-forming activity. Proc. Natl. Acad.Sci. USA. In press

72. Kawata, M., Farnsworth, C. C.,Yoshida, Y., Gclb, M. H., Glomsct,J. A., et al. 1990. Posttranslationallyprocessed structure of the human plate-let protein smg p21B: evidence forgcranylgcranylation and caboxyl mcth-ylation of the C-terminal cysteinc.Proc. Natl. Acad. Sci. USA 87:8960-64

73. Khosravi-Far, R., Lutz, R. J., Cox,A. D., Conroy, L., Bourne, J. R., ctal. 1991. Isoprenoid modification oftab proteins terminating in CC or CXCmotifs. Proc. Natl. Acad. Sci. USA88:6264-68

74. Kinsella, B. T., Fa’dman, R. A., Maltese,W. A. 1991. Carboxyl-terminal isopren-ylation of ras-related GTP-bindingproteins encoded by racl, rac2, andralA. J. Biol. Chem. 266:9786-94

75. Kinsella, B. T., Maltese, W. B. 1991.rab GTP-binding proteins implicatedin vesicular transport are isoprenylatedin vitro at cysteines within a novelcarboxyl-terminal motif. J. Biol. Chem.266:8540-44

76. Kitayama, H., Sugimoto, Y., Matsu-zaki, Y., Ikawa, Y., Noda, M. 1989.


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.



A ras-related gene with transformationsuppressor activity. Cell 56:77-84

77. Kohl, N. E., Diehl, R. E., Schaber,M. D., Rands, E., Soderman, D. D.,et al. 1991. Structural homology amongmammalian and Saccharomyces cere-visiae isoprenyl-protein transferases. J.Biol. Chem. 266:18884-88

78. Kuchler, K., Sterne, R. E., Thomer,J. 1989. Saccharomyces cerevisiaeSTE6 gene product: a novel pathwayfor protein export in eukaryoac cells.EMBO J. 8:3973-84

79. Kuchler, K., Thorner, J. 1990. Mem-brane translocation of proteins withouthydrophobic signal peptides. Curt.Opin. Cell Biol. 2:617-24

80. Lai, R. K., Perez-Sala, D., Canada,F. J., Rando, R. R. 1990. The ~/subunit of transducin is farnesylated.Proc. Natl. Acad. Sci. USA 87:7673-77

81. Li, R., Murray, A. W. 1991. Fecdbackcontrol of mitosis in budding yeast.Cell 66:519-31

82. Lutz, R. J., Trujillo, M. A., Denham,K. S., Wegner, L., Sinensky, M.1992. Nucleoplasmic localization ofprelamin A: implications for prenyla-tion dependent lamin A assembly intothe nuclear lamina. Proc. Natl. Acad.Sci. USA 87:3000-4

83. Ma, P. T. S., Gil, G., Sudhof, T.C., Billheimer, D. W., Goldstein, J.L., et al. 1986. Mevinolin, an inhibitorof cholesterol synthesis, induces mRNAfor low density lipoprotein receptor inlivers of hamsters and rabbits. Proc.Natl. Acad. Sci. USA 83:8370-74

84. Maltese, W. A., Robishaw, J. D.1990. Isoprenylation of C-temxinal cys-teine in a G-protein "~ subunit. J. Biol.Chem. 265:18071-74

85. Maltese, W. A., Sheridan, K. M. 1987.Isoprenylated proteins in cultured cells:subcellular distribution and changes re-lated to altered morphology and growtharrest induced by mevalonate depriva-tion. J. Cell. Physiol. 133:471-81

86. Manne, V., Roberts, D., Tobin, A.,O’Rourke, E., De Virgilio, M., et al.1990. Identification and preliminarycharacterization of protein-eysteinefarnesyltransferase. Proc. Natl. Acad.Sci. USA 87:7541-45

87. Marcus, S., Caldwell, G. A., Miller,D., Xue, C-B., Nalder, F., et al.1991, Significance of C-terminal cys-teine modifications to the biologicalactivity of the Saccharomycescerevisiae a-factor mating pheromone.Mol. Cell. Biol. 11:3603--12

88. Marcus, S., Caldwell, G. A., Xue, C-B.,Naider, F., Becker, J. M. 1990. Totalin vitro maturation of the Sacchar-

omyces cerevisiae a-factor lipopeptidemating pheromone. Biochem. Biophys.Res. Comm. 172:1310-16

89. Marcus, S., Xue, C-B., Nalder, F.,Becket, J. M. 1991. Degradation ofa-factor by a Saccharomyces cerevisiaea-mating-type-specific endopeptidase:evidence for a role in recovery of cellsfrom GI arrest. Mol. Cell. Biol. 11:1030--39

90. Mart, R. E. S., Blair, L. C., Thorner,J. 1990. Saccharomyces cerevesiaeSTE14 gene is required for C-terminalmethylation of a-factor mating phero-mone. J. Biol. Chem. 265:20057-60

91. McCaffrey, M., Johnson, J. S., Goud,B., Myers, A. M., Rossier, J., et al.1991. The small GTP-binding proteinRholp is localized on the Golgi ap-paratus and post-Golgi vesicles inSaccharomyces cerevisiae. J. Cell Biol.115:309-19

92. McGrath, J. P., Varshavsky, A. 1989.The yeast STE6 gene encodes a ho-mologue of mammalian multidrug re-sistance P-glycoprotein. Nature 340:400-4

93. Michaelis, S., Herskowitz, I. 1988.The a-factor pheromone of Sac-charomyces cerevisiae is essential formating. Mol. Cell. Biol. 8:1309-18

94. Miyajima, I., Nakayama, N.,Nakafuku, M., Kaziro, Y., Aral, K.,et al. 1988. Suppressors of gpal mu-tation cause sterility in Saccharomycescerevisiae. Genetics 119:797 804

95. Miyakawa, T., Nishihara, M.,Tsuchiya, E., Fukui, S. 1982. Roleof metabolism of the mating pheromonein sexual differentiation of the hetero-basidiomycete Rhodosporidium tor-uloides. J. Bacteriol. 151:1184-94

96. Miyakawa, T., Tabata, M., Tsuehiya,E., Fukui, S. 1985. Biosynthesis andsecretion of tremerogen A-10, poly-prenyl peptide mating pheromone ofTremella mesenterica. J. Biochem.147:489-93

97. Molenaar, C. M. T., Prange, R.,Gallwitz, D. 1988. A carboxyl-terminalcysteine residue is required for palmiticacid binding and biological activity ofthe ras-related yeast YPT1 protein.EMBO J. 7:971

98. Moores, S. L., Schaber, M. D.,Mosser, S. D., Rands, E., O’Hara,M. B., et al. 1991. Sequence depen-dence of protein isoprenylation. J. Biol.Chem. 266:14603-10

99. Morimoto, B. H., Chuang, C-C., Kosh-land, D. E. 1991. Molecular cloningof a member of a new class of low-molecular-weight GTP-binding pro-teins. Genes Dev. 5:2386-91


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


236 SCHAFER & RINE

100. Mumby, S. M., Casey, P. J., Gilman,A. G., Gutowski, S., Sternweiss, P.C. 1990. G protein ~/subunits containa 20-carbon isoprenoid. Proc. Natl.Acad. Sci. USA 87:5873-77

101. Nakanishi, M., Goldstein, J. L., Brown,M. S. 1988. Multivalent control of3-hydroxy-3-methylglutaryl coenzymeA reductase. Mevalonate-derived prod-uct inhibits translation of mRNA andaccelerates degradation of enzyme. J.Biol. Chem. 263:8929--37

102. Newman, C. M. H., Giannakouros,T., Hancock, J. F., Fawell, E. H.,Armstrong, J., et al. 1992. Post-trans-lational processing of Schizosaccharo-myces pombe YPT proteins. J. Biol.Chem. In press

103. O’Brien, P. J., St. Jules, R. S., Reddy,T. S., Bazan, N. G., Zatz, M. 1987.Acylation of disc membrane rhodopsinmay be nonenzymatic. J. Biol. Chem.262:5210-15

104. O’Brien, P. J., Zatz, M. 1984. Acyla-tion of bovine rhodopsin by [31-I]pal-mitic acid. J. Biol. Chem. 259:5054-57

105. Ohguro, H., Fukada, Y., Takao, T.,Shimonishi, Y., yoshizawa, T., et al.1991. Carboxyl methylation and farn-esylation of transducin ~/-subunit syn-ergistically enhance its coupling withmetarhodopsin II. EMBO J. 10:3669-74

106. Ohya, Y., Ohsumi, Y., Anraku, Y.1984. Genetic study of the role ofcalcium ions in the cell division cycleof Saccharomyces cerevisiae: a cal-cium-dependent mutant and its triflou-perazine-dependent pseudorevertants.Mol. Gen. Genet. 193:389-94

107. Perez-Sala, D., Tan, E. W., Canada,F. J., Rando, R. R. 1991. Methylationand demethylation reactions of guaninenucleotideobinding proteins of retinalrod outer segments. Proc. Natl. Acad.Sci. USA 88:3043-46

108. Pizon, V., Chardin, P., Lerosey, I.,Olofsson, B., Tavitian, A. 1988. HumancDNAs rap1 and rap2 homologous tothe Drosophila gene Dras3 encode pro-teins closely related to ras in the"effector" region. Oncogene 3:201-4

109. Powers, S., Michaelis, S., Broek, D.,Santa Anna-A., S., Field, J., et al.1986. RAM, a yeast gene required fora functional modification of RAS pro-teins and for production of matingpheromone a-factor. Cell 47:413-22

110. Qiu, M-S., Pitts, A. F., Winters, T.,R., Green, S. H. 1991. Ras isoprenyla-tion is required for ras-induced but notfor NGF-induced neuronal differentia-

tion of PCI2 cells. J. Cell Biol. 115:795-808

111. Quesney-Huneeus, V., Wiley, M. H.,Siperstein, M. D. 1979. Essential rolefor mevalonate synthesis in DNA rep-lication. Proc. Natl. Acad. Sci. USA76:5056-60

112. Reiss, Y., Goldstein, J. L., Seabra,M. C., Casey, P. J., Brown, M. S.1990. Inhibition of purified p21rasfamesyl: protein transferase by Cys-AAX tetrapeptides. Cell 62:81

113. Reiss, Y., Seabra, M. C., Armstrong,S. A., Slaughter, C. A., Goldstein,J. L., 9I al. 1991. Nonidentical subunitsof p21n’ras farnesyltransferase. J. Biol.Chem. 266:10672-77

114. Reiss, Y., Stradley, S. J., Gierasch,L. M., Brown, M. S., Goldstein, J.L. 1991. Sequence requirement forpeptide recognition by rat brain p21rasfarnesyl;protein transferase. Proc. Natl.Acad. Sci. USA 88:732-36

115. Resh, M. D. 1989. Specific and sat-urable binding of pp60v-src to plasmamembranes. Cell 58:281-86

116. Resh, M. D., Ling, H. 1990. Identi-fication of a 32K plasma membraneprotein that binds to the myristoylatedamino-terminal sequence of p6Ov-src.Nature 346:84-86

117. Rilling, H. C., Breunger, E., Epstein,W. W., Crain, P. F. 1990. Prenylatedproteins: the structure of the isoprenoidgroup. Science 247:318-20

118. Rine, I. 1979. Regulation and trans-position of cryptic mating-type genesin Saccharomyces cerevisiae. PhD the-sis. Univ. Oregon, Eugene, OR

119. Rossi, G., Jiang, Y., Newman, A. P.,Ferro-Novick, S. 1991. Bet2, a putativecomponent of a prenyltransferase, isrequired for membrane attachment ofYptl and Sec4. Nature 351:158-61

120. Rothberg, K. G., Heuser, J. E.,Donzell, W. C., Ying, Y-S., Glenney,J. R., et al. 1992. Caveolin, a proteincomponent of caveolae membranecoats. Cell 68:673-82

121. Saikumar, P., Ulsh, L. S., Clanton,D. J., Shih, T. Y. 1988. Novel phos-phorylation of c-ras p21 by proteinkinases. Oncog. Res. 3:213-22

122. Sakagami, Y., Yoshida, M., Isogal, A.,Suzuki, A. 1981. Peptidal sex hormonesinducing conjugation tube formation incompatible mating-type cells of Trem-ella mesenterica. Science 212:1525-27

123. Sanford, J., Codina, J., Bimbaumer, L.1991. 3~-subunits of G-proteins, but nottheir a- or I~-subunits, are polyisopren-ylated. J. Biol. Chem. 266:9570-79

124. Schafer, W. R., Kim, R., Sterne, R.,


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


PROTEIN PREN YLATION 237

Thorner, J., Kim, S-H., et al. 1989.Genetic and pharmacological suppres-sion of oncogenic mutations in RASgenes of yeast and humans. Science245:379-85

125. Schafer, W. R., Trueblood, C. E.,Yang, C-C., Mayer, M. P., Rosenberg,S., et al. 1990. Enzymatic couplingof cholesterol intermediates to a matingpheromone precursor and to the Rasprotein. Science 249:1133-39

126. Schulz, S., Nyce, J. W. 1991. Inhi-bition of protein isoprenylation andp21ras membrane association bydehydroepiandrosterone in human ad-enocarcinoma cells in vitro. CancerRes. 51:6563-67

127. Seabra, M. C., Reiss, Y., Casey, P.J., Brown, M. S., Goldstein, J. L.1991. Protein farnesyltransferase andgeranylgeranyltransferase share a com-mon alpha subunit. Cell 65:429-34

128. Segev, N., Mulholland, J., Botstein,D. 1988. The yeast GTP-binding YPT1protein and a mammalian counterpartare associated with the secretion ma-chinery. Cell 52:915-24

129. Sepp-Lorenzino, L., Azrolan, N.,Coleman, P. S. 1989. Cellular distri-bution of cholesterogenesis-linkedphosphoisoprenylated proteins in pro-liferating cells. Proc. Natl. Acad. Sci.USA 82:3257-61

130. Shih, T. Y., Weeks, M. O., Gross,P., Dhar, R., Oroszlan, S., et al.1982. Identification of a precursor inthe biosynthesis of the p21 transformingprotein of Harvey murine sarcomavirus. J. Virol. 42:253-61

131. Shirataki, H., Kaibuchi, K., Hiroyoshi,M., Isomura, M., Araki, S., et al.1991. Inhibition of the action of thestimulatory GDP/GTP exchange proteinfor smgp21 by the geranylgeranylatedsynthetic peptides designed from itsC-terminal region. J. Biol. Chem. 266:20672-77

132. Stephenson, R. C., Clarke, S. 1990.Identification of a C-terminal proteincarboxyl methyltransferase in rat livermembranes utilizing a synthetic farn-esyl cysteine-containing peptide sub-strate. J. Biol. Chem. 265:16248-54

133. Sterne, R. 1989. A novel pathway forhormone biogenesis processing and se-cretion of the mating pheromone a-factor by Saccharomyces cerevisiae.PhD thesis. Univ. Calif., Berkeley,CA

134. Sterne, R. E., Thorner, J. 1987. Pep-tide hormone export by a novel path-way: yeast a-factor mating pheromoneis a lipopeptide. J. Cell Biol. 105:80a

135. Taparowsky, E., Shimizu, K., Gold-farb, M., Wigler, M. 1983. Structureand activation of the human N-rasgene. Cell 34:581-86

136. Tsuchiya, E., Fukui, S., Kamiya, Y.,Sakagami, Y., Fujino, M. 1978. Re-quiremems of chemical structure forhormonal activity of lipopeptidyl fac-tors inducing sexual differentiation invegetative cells of heterobasidiomycet-ous yeasts. Biochem. Biophys. Res.Commun. 85:459-63

137. Vorburger, K., Kitten, G. T., Nigg,E. A. 1989. Modification of nuclearlamin proteins by mevalonic acid occursin reticulocyte lysates and requires thecysteine residue of the C-terminalCXXM motif. EMBO J. 8:4007-13

138. Wilson, K. L., Herskowitz, I. 1987.STE16, a new gene required for pher-omone production by a ceils of Sac-charomyces cerevisiae. Genetics 115:441 49

139. Wolda, S. L., Glomset, J. A. 1988.Evidence for modification of lamin Bby a product of mevalonic acid. J.Biol. Chem. 263:5997-6000

140. Worman, H. J., Evans, C. D., Blobel,G, 1990. The lamin B receptor of thenuclear envelope inner membrane: apolytopic protein with eight potentialtransmembrane domains. J. Cell Biol.111:1535-42

141. Worman, H. J., Yuan, J., Blobel, G.,Georgatos, S. D. 1988. A lamin Breceptor in the nuclear envelope. Proc.Natl. Acad. Sci. USA 85:8531-34

142. Yamane, H. K., Farnsworth, C. C.,Xie, H., Howald, W., Fung, B. K-K.,et al. 1990. Brain G protein ~/subunitscontain an all-trans-geranylgeranyl-cys-teine methyl ester at their carboxyltermini. Proc. Natl. Acad. Sci. USA87:5868-72

143. Yamane, H. K., Farnsworth, C. C.,Xie, hi. Y., Evans, T., Howald, W.N., et al. 1991. Membrane-bindingdomain of the small G protein G25Kcontains an S-(all-trans-geranylgeranyl)cysteine methyl ester at its carboxylterminus. Proc. Natl. Acad. Sci. USA88:286-91

144. Yokoyama, K., Goodwin, G., Ghorn-sashchi, F., Glomset, J. A., Gelb, M.H. 1991. A protein geranylgeranyl-transferase from bovine brain: implica-tions for protein prenylation specificity.Proc. Natl. Acad. Sci. USA 88:5302-6

145. Zhukovsky, E. A., Robinson, P. R.,Oprian, D. D. 1991. Transducin acti-vation by rhodopsin without a covalentbond to the 11-cis-retinal chromophore.Science 251:558-60


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.


Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.

Ann

u. R

ev. G

enet

. 199

2.26

:209

-237

. Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Med

ical

Res

earc

h C

ounc

il-C

ambr

idge

UK

on

11/2

1/06

. For

per

sona

l use

onl

y.

Documents

Protein Prenylation: Genes, Enzymes, Targets, and Functions