Thus far three kinases and three arrestin proteins 
have been defined by biochemical, functional, and 
molecular cloning studies. Dr. Lefkowitz and his col- 
leagues have cloned and sequenced full-length 
cDNAs for rhodopsin kinase, the enzyme that regu- 
lates rhodopsin function in the visual cycle, as well 
as the two distinct isoforms of |8ARK (/3-adrenergic 
receptor kinase), ;8ARK1 and (8ARK2, that regulate 
the function of the /5-adrenergic receptor and other 
G protein-coupled receptors. Only the conforma- 
tionally active forms of the various G protein- 
coupled receptors are substrates for the kinases. 
These enzymes phosphorylate the receptors on their 
carboxyl-terminal cytoplasmic tails. This leads to 
the binding of an arrestin protein to the receptor 
that thereby sterically interdicts signal transduction 
to the G protein, providing a means for rapid 
quenching of the stimulus. 
In both the rhodopsin kinase and (3 ARK systems, 
the enzymes are predominantly cytosolic but rap- 
idly translocate to the membrane upon stimulation 
of the receptor. Until recently the molecular basis 
for this stimulus-orchestrated compartmentation of 
the enzymes was unknown. Inspection of the se- 
quences of the cloned enzymes reveals that rhodop- 
sin kinase, but neither of the (SARKs, possesses a 
CAAX box at its carboxyl terminus (cysteine/ali- 
phatic/aliphatic/any amino acid). This carboxyl- 
terminal signature sequence specifies a series of 
three interrelated post-translational modifications. 
These include isoprenylation of the cysteine by a 
C-15 farnesyl or C-20 geranylgeranyl moiety, fol- 
lowed by proteolysis of the last three amino acids 
and then carboxyl methylation of the now-terminal 
cysteine residue. 
To test the role of rhodopsin kinase isoprenyla- 
tion in its translocation. Dr. Lefkowitz and his col- 
leagues prepared and expressed three different 
cDNAs, including the wild-type rhodopsin kinase, 
which ends with the residues CVLS, specifying far- 
nesylation; another rhodopsin kinase ending in 
CVLL, which specifies geranylgeranylation; and a 
third in which the cysteine was mutated to serine, 
which is not isoprenylated at all. When the cDNAs 
were expressed in COS-7 cells, it could be docu- 
mented that they specified the production of en- 
zymes that were isoprenylated or not in the ex- 
pected fashion. Moreover, whereas the CI 5 and C20 
forms of the enzyme showed approximately equal 
biological activity to phosphorylate rhodopsin, the 
nonprenylated form of the enzyme was markedly 
impaired in this activity (<30% activity). Moreover, 
Dr. Lefkowitz and his colleagues were also able to 
measure the photon-induced translocation of the 
enzyme in an assay developed for this purpose uti- 
lizing rhodopsin-containing rod outer segment 
membranes. Only the farnesylated form of rhodop- 
sin kinase attached to the membrane-bound rhodop- 
sin when stimulation with light was provided. Thus 
the geranylgeranylated form was largely membrane 
bound to begin with, and the nonprenylated form, 
which was not membrane bound in the absence of 
light, also failed to translocate even upon illumina- 
tion. These data suggest that the isoprenylation of 
rhodopsin kinase plays a crucial role in mediating 
its light-induced translocation to rhodopsin- 
containing membranes and in supporting its biologi- 
cal activity. 
The /3ARK does not contain a CAAX box and hence 
is not prenylated. However, Dr. Lefkowitz and his 
colleagues observed that if they prepared chimeric 
forms of /3ARK containing CAAX boxes, which there- 
fore could be prenylated, their enzymatic activity 
increased by about threefold, similar to the results 
obtained with rhodopsin kinase. Moreover, the iso- 
prenylated forms of the enzyme were better able to 
associate with receptor-containing vesicles. These 
data suggested that perhaps isoprenylation might in 
some way be involved in the translocation of jSARK 
as well. 
The 7 subunit of the heterotrimeric G proteins, 
with which the receptors interact, are geranylgeran- 
ylated and are thought to play a role in anchoring 
the complex of the G proteins in the cell mem- 
brane. The laboratory showed that bovine brain /?7 
subunits have a very marked stimulatory effect on 
the enzymatic activity of |8ARK1 and |8ARK2 (> 10- 
fold), which is apparent when their ability to phos- 
phorylate membrane receptors is examined but not 
when their ability to phosphorylate simple peptides 
is measured. Moreover, the 0y subunit physically 
interacts with the (8ARK enzyme and mediates its 
attachment to receptor-containing membranes. The 
sites of interaction on the /3ARK enzyme are located 
in the last 200 amino acids, as documented by physi- 
cal interaction studies between I3y and GST fusion 
proteins containing the relevant sequences of |8ARK. 
These new results indicate an important role for 
isoprenylation of proteins in mediating the stimu- 
lus-driven translocation of G protein-coupled re- 
ceptor kinases. For rhodopsin, the isoprenyl group 
is found on the rhodopsin kinase, whereas for /3ARK 
it is donated by the 187 subunit of G proteins. These 
167 subunits become available only when they are 
dissociated from by interaction of G protein with 
receptor. Thus even as the G protein is being acti- 
vated to mediate receptor signaling, it is providing 
the crucial 187 subunit that helps to target (3ARK to 
CELL BIOLOGY AND REGULATION 87 
