Biochemical Society Transactions
Receptor regulation of G protein palmitoylation
I56
S. M. Mumby*$ and K. H. Muntzt Departments of *Pharmacology and tCell Biology and Neuroscience, University of Texas Southwestern Medical Center at Dallas, 5323 Harry Hines Blvd., Dallas, TX 75235, U.S.A. A major role of G proteins is to couple extracellular messengers to intracellular effector systems. Upon activation by receptor, a heterotrimeric G protein dissociates into a GTP-bound a subunit and a /3y subunit complex, both of which are able to modulate the activity of effectors [l]. These G proteins are positioned at the inner face of the plasma membrane where they are able to interact with membrane-spanning receptors and effectors. The molecular basis for interaction of signal-transducing G proteins with the membrane is not well understood as they lack clear hydrophobic domains which would be anticipated to promote interactions with the lipid bilayer. Covalent lipid modifications of G protein subunits appear to facilitate membrane association and interaction with other components of the signal-transduction systems [2]. The y subunits are prenylated and carboxymethylated at their C-termini. These modifications facilitate association with the /3y complex with membranes and are indispensable for high-affinity interactions of Py with a subunits, receptor, and effector molecules [3-51. Members of the a; subfamily of a subunits are irreversibly modified by myristate amide-linked to the N-terminal glycine residue (Gly-2) [6-91. Myristoylation increases the affinity of a for By and effector and also plays a role in membrane localization [7,10-121. Palmitate is the most recent modification to be identified and this fatty acid is found linked to several a subunits [13-171. Most members of the a , subfamily are tandemly modified by both myristate and palmitate (presumably at Gly-2 and Cys-3 respectively) [ 131. The reversibility of palmitoylation makes it an exciting modification to study since it could potentially serve a regulatory role in signal transduction. One of the best-characterized examples of guanine nucleotide-regulated signal transduction is the hormone-sensitive adenylate cyclase system. Production of the second messenger, cyclic AMP, by the effector enzyme is under dual stimulatory and inhibitory control by G proteins termed G, and GI [ 11,181. Receptors that stimulate adenylate cyclase are coupled to the enzyme by G,, whereas inhibitory receptors interact with Gi. The dynamic Abbreviation used: GPI, glycosyl-phosphatidylinositol. $To whom correspondence should be addressed.
Volume 23
potential of palmitoylation prompted us and others to examine whether receptor stimulation could regulate palmitoylation of a, (the a subunit of G, which stimulates adenylate cyclase activity). When COS cells are incubated with ['Hlpalmitate in the presence of the /3-adrenergic receptor agonist, isoprenaline (isoproterenol), incorporation of radioactivity into both the long- and short-splice variants of endogenous a, were increased relative to untreated cells (Figure la) [ 19-21]. Isoprenaline did not affect synthesis of a,; the agonist caused no change in the incorporation of ["S]methionine into the protein [ 191. The /3-adrenergic antagonist propranolol had no effect on the incorporation of ['Hlpalmitate into a,, but when it was combined with isoprenaline it prevented a change in the incorporation of ['Hlpalmitate into a , [ 19,201. The agonist-induced increase in the incorporation of radioactivity is specific for a, since we observed no change in the incorporation of radioactivity into ai (which is not activated by /?-adrenergic receptors in cells) or into the transferrin receptor, another palmitoylated protein (Figure la) [ 19,201. Support for an effect on activated a , is provided by studies of two mutant forms of the protein in S49 lymphoma cells. Neither the UNC mutant ( a , is uncoupled from receptor) nor the G226A mutant ( a , is unable to dissociate from B y ) is stimulated by isoprenaline to incorporate ['Hlpalmitate (although the wild-type a , is stimulated in these cells) [20,21]. Exposure of cells to forskolin (a direct activator of adenylate cyclase) or to dibutyryl cyclic AMP had no effect on incorporation of ['Hlpalmitate into a, which suggests that the effect of isoprenaline is directly mediated by activation of a,, rather than by cyclic AMP acting as a second messenger [ 19-21]. Pulse-chase experiments have been conducted to distinguish whether the agonist-induced increase in ['Hlpalmitate incorporation into a, (Figure la) could be attributed to an increase in the stoichiometry of palmitoylation (addition of radioactive palmitate to previously unpalmitoylated a,) or to an increase in the turnover of palmitate on a , (loss of unlabelled palmitate from a, allowing replacement by radioactive palmitate) [ 19,211. In these experiments radioactivity in a, was lost more readily in cells incubated with isoprenaline, suggesting that the agonist increases the turnover of palmi-
Regulation of Signal-Transducing Polypeptides
Figure I Time course and pulse-chase of [)H]palmitate incorporation into endogenous a,, and modulation by isoprenaline COS cells (not transfected) were assayed by immunoprecipitation, SDS/PAGE, fluorography and densitometry (a) Time course of incorporation of [)H]palmitate ( I .5 mCi/ml) into endogenous proteins in the presence of I pM isoprenaline ( 0 ) or its absence ( 0 ) .(b) Pulse-chase. After a 15-min incubation with [3H]palmitate,cells were rinsed and incubated with medium containing 30pM unlabelled palmitate and 5% serum with (+ ) or without ( - ) I pM isoprenaline for 20 or 40 min. Film was exposed for 2.5 months. This data is reproduced with permission [ 191. (a) l i m e course
'1.....;
a,short
I
500
.-3 c
f
Y
.-
Y
4
lransferrin receptor
0
100
v)
6 1
I
I
l
I
I
I
5 10 15 20 25 30
I
I
I
I
I
10 20 30 40 50 60
5 10 15 20 25 30
l i m e (min)
Time (min)
l i m e (min)
(b) Pulse chase
a, short
a, long
rn .z 3
200
2
150
0 In
100
100
100
8
50
50
50
c
E .-
c
200
200 150
Isoprenaline. . .
-
..
0
l i m e (min).
receptor
250
-
+
20
-
+
40
- -+
-+
- -+
20
40
0
0
tate in this protein (Figure lb). The most obvious mechanistic hypothesis is that the activated a subunit is the preferred substrate for a palmitoylthioesterase. If receptor or a downstream mediator were to activate such an enzyme directly, then the enzyme must have specificity for a given receptoractivated a subunit, since changes in a, were observed but not in aior the transferrin receptor. Although this mechanism would suggest a net decrease in the level of palmitoylation of a,, firm conclusions cannot be drawn since the stoichiometries have not been measured. It will be of interest to learn whether receptor-mediated changes in the palmitoylation of a influence its interactions with receptors and/or effectors or their distribution within the cell. A correlation has been drawn between receptor activation of an epitope-tagged a, (expressed in S49 cells), modulation of its palmitoylation, and release of the protein from membranes [21]. In contrast, we have not yet been able to demonstrate release of endogenous a subunits
20
1 -+ 40
from membranes by activation (S. M. Mumby and K. H. Muntz, unpublished work). Myristate was previously thought to dictate membrane association of a , and a, since a Gly-2 to Ala mutation of these proteins prevented them from being myristoylated and targeted them to the cytoplasmic fraction of COS cells [7,12]. The contribution of myristate to the localization of the proteins became unclear when it was later realized that palmitate was also missing from these Gly-2 mutant proteins [ 191. Apparently myristoylation must precede palmitoylation. Myristoylation of dually acylated a subunits may be necessary for substrate recognition by a protein palmitoyltransferase; alternatively, myristoylation may be necessary for association with membranes, wherein protein palmitoyltransferase activities have been detected [22]. Mutagenesis of Cys-3 prevents palmitoylation of a,, a;, or a,, presumably because Cys-3 is the site of palmitoylation of these proteins [ 14- 16,191. Since myristoylation is unaffected in Cys-3 to Ala
I995
I57
Biochemical Society Transactions
I58
mutants the loss of a single fatty acid (palmitate) could be tested. Overexpression of a,, or a ; alone resulted in distribution of some of the expressed Cys-3 to Ala mutant protein to the soluble fraction of cells, whereas the wild-type protein is almost entirely present in the particulate fraction (Figure 2, left-hand panel) [ 19,231. Since the Cys-3 mutant proteins are not entirely cytosolic (but the Gly-2 mutants are) it appears that both palmitate and myristate contribute to membrane association of these proteins. The role of palmitate in the distribution of a, (which is not myristoylated) in cells has been studied by expression of proteins with Cys-3 mutations and crude fractionation of cells. Wedegaertner et al. [ 151 report that a Cys-3 to Ser mutation of epitope-tagged a, expressed in 293 cells has a cytosolic distribution. They conclude that palmitoylation of Cys-3 is required for membrane attachment of as. In contrast, Degtyarev et a]. and we have found that a Cys-3 to Ala mutation of a , (which is not epitope-tagged) does not change the fractionation of the mutant relative to wild-type protein in either transfected COS (Figure 2, right-hand panel) [ 16,191 or Sfc) insect cells infected with recombinant baculovirus [ 191. T h e majority of the overexpressed a , could be extracted from pellets with 1% (w/v) sodium cholate, indicating that the protein was not simply denatured and aggregated [ 191. It is not clear whether the difference in the observations made by Wedegaertner et a]. is due to the amino acid
Figure 2
Distribution of wild-type and non-palmitoylated Cys-3 -Ala (C3A) mutant a subunits in fractionated COS cells Fraction PI i s the pellet from a 1000 g centrifugation, P2 is the 200000 g pellet, and S2 i s the 200000 supernatant fraction (a) lmmunoprecipitation of a subunits from transfected COS cells that had been incubated with [35S]methionine(IOOpCi/ml) for 60min) Control cells were transfected with the pCMV5 vector The A569 antibodies recognize endogenous a , and expressed wild-type and C3A mutant a , Wild-type a. was largely confined t o the membrane-containing pellets (lanes 7 and 8),whereas a significant amount of the non-palmitoylated mutant C3A protein was present in the cytosolic S2 fraction (lane 6) Both wild-type and mutant C3A a , proteins were confined to the pellets (lanes 13, 14, 16 and 17), which were extracted and immunoprecipitated with 584 antibodies Data i s reproduced with permission [ 191. Expression Fraction
vector
a C3A
-LA PI PZ
sz
PI PZSZ PI PZ
vector
sz
a,C3A
--A
PI PZ
sz
10 11 12 13 14 15 16 17
18
PI
PZ
sz
PI PZ
sz
-
a1 \
a0
1
Volume 23
2 3 4 5 6 7 8 9
replacement, the presence of the epitope tag, the cell type, or other differences. The fact that Degtyarev et al. and we do not see an effect on the Cys-3 mutation on the distribution of a, between particulate and soluble fractions of cells does not necessarily mean that the mutation (or palmitoylation) has no effect on the ultrastructural localization of the protein. Subtle differences could exist which would not be detected by simple separations of particulate from soluble fractions. For instance, partitioning of a into or out of plasma membrane specializations would not be evident in such crude fractionations. Wedegaertner et al. have demonstrated that epitope-tagged a,, when expressed in HEK293 cells, couples the a,-adrenergic receptor (also introduced by transfection) to the production of cyclic AMP [ 151. The epitope-tagged a, with the Cys-3 to Ser mutation lacks this coupling activity. In an effort to attribute this loss of coupling either to a deficiency in interaction with receptor or effector they introduced the Cys-3 to Ser mutation into the context of a constitutively active a,(caused by Arg-201 to Cys mutation). The double mutant protein retains the ability to stimulate its effector in the absence of agonist, although it is somewhat less active than the single mutation Arg-201 to Cys protein expressed at similar levels. The results are consistent with the Cys-3 to Ser mutation causing a loss of interaction of G, with the receptor as well as a reduction in efficiency of interaction of G, with effector. It is not known if these changes are due to the loss of palmitate or to the introduction of serine. W e predict that palmitoylation of G protein a subunits would influence their interactions with receptors since the N-terminus of a,, interacts with a mastoparan derivative. Mastoparan (a wasp venom peptide) activates G proteins by promoting GDPIGTP exchange in a manner that closely resembles that of receptor. ["'I-Tyr-3, Cysis ll]Mastoparan, when covalently bound to a,,, cross-linked to Cys-3 [24]. If mastoparans interact with a subunits at a site similar to that utilized by activated receptors (as we assume), it seems likely that palmitoylation of Cys-3 would necessarily influence its interaction with receptor Parenti et al. and others have noted that the N-terminal sequences and fatty acylation properties of most members of the a , subfamily are strikingly similar to those of some members of the Src-related family of tyrosine kinases [ 14,19,25]. These proteins (a,,, a,, a,, ~.56''~,and p59kn) are both . . myristoylated and palmitoylated at adjacent sites [ 13,26-281. Individual replacement of the two Cys residues within the first 10 amino acids of p5W
Regulation of Signal-Transducing Polypeptides
and p56'" with Ser indicates that Cys-3 is a major determinant of palmitoylation and association of the kinases with glycosyl-phosphatidylinositol (GP1)anchored proteins [29]. This association may provide the signal by which cross-linking of GPI-anchored proteins (which are inserted into only the outer leaflet of the plasma membrane) can activate T cells. GPI-anchored proteins are clustered in microinvaginations of the plasma membrane termed caveolae (plasmalemmal vesicles) and are relatively insoluble in mild detergent such as Triton X- 100 [ 301. Detergent-insoluble complexes have been isolated from cells and analysed morphologically and biochemically. Electron microscopic examination reveals that the complexes contain vesicles resembling caveolae [25,311. These vesicles can be decorated by immunogold with antibodies for caveolin (a protein marker for caveolae) and for G proteins [25]. Western immunoblotting of the complexes confirms that these proteins are enriched in the preparations and that GPI-linked proteins and Src-related kinases are present [25,3 11. Shenoy-Scaria et al. have recently shown that the partitioning of p56'ck into Triton X- 100-insoluble complexes depends on Cys-3 (and presumably palmitoylation) [29]. The importance of Cys-3 in the distribution of the kinases is underscored by a mutation in ~ 6 0 "that ~ causes a gain of function. ~ Cys Replacement of Ser-3 in wild-type ~ 6 0 "with allowed the protein to incorporate ['Hlpalmitate, interact with a GPI-linked protein, and to partition into Triton X- 100-insoluble complexes (unlike the unpalmitoylated wild-type protein) [20]. It was concluded that the partitioning of palmitoylated Src-related kinases into Triton X- 100-insoluble complexes represents a distribution of these proteins to caveolae. The tandem acylation and the Triton X- 100 fractionation similarities between a, and the kinases have led us to propose that dynamic palmitoylation of a subunits may regulate a distribution into and out of membrane specializations such as caveolae. W e are using immunocytochemical techniques to study, in greater detail, the distribution of G proteins within the cells. We find that G protein a and /3 subunits are distributed in a punctate immunofluorescent pattern consistent with labelling of the plasmalemma and internal membranes. The punctate pattern of a , staining at the plasmalemma was confirmed by immunofluorescent labelling of plasma membranes isolated on coverslips by sonication (Figure 3). W e propose that the punctate pattern represents a clustered distribution of a , at the plasma membrane. T o determine whether the
F i g u r e :I
Ga, immunolabelling to the internal surface of the plasma membrane of a cultured renal epithelial cell (MA I 04) Cells grown on coverslips were sonicated under conditions that removed the upper membrane and intracellular components leaving the internal surface of the lower membrane exposed [32,33] The membrane was fixed in 3 7 % para formaldehyde and incubated in affinity purified C-terminal peptide antibodies selective for a and a ?, followed by a fluor escein-labelled second antibody The labelling pattern on the internal surface of the plasma membrane IS punctate Scale bar= IOpm
clustered distribution could be attributed to association with a, with identifiable structures such as caveolae or coated pits, cells were double-labelled with antibodies to a, and to either caveolin or clathrin (a structural protein of coated pits). In MA104 renal epithelial cells, there was considerable (though not complete) co-localization of a, with cavoelin. but not with clathrin. Quantitative immunoelectron microscopy confirmed that a , was distributed on the plasmalemma and at intracellular membranes of both MA104 cells and fibroblasts (results not shown). Although we have observed convincing immunogold labelling of caveolae with a caveolinspecific antibody, we have not observed consistent labelling of these structures with our a,- or psubunit-reactive antibodies. Since the distribution of a, is clearly not restricted to caveolae we propose that a subunits may partition into and out of these or other such membrane specializations. Dynamic palmitoylation is an attractive means by which such redistribution among membrane compartments could be regulated. Receptor-regulated palmitoylation that is restricted to particular subdomains of the plasmalemma could contribute an additional level of organization in the signalling cascade. Such restrictive organization could provide the specificity of G protein interactions with appropriate receptors
I995
I59
Biochemical Society Transactions
and effectors that is observed in whole cells but is often lost in experiments involving reconstitution of purified proteins into phospholipid vesicles. I60
We thank Alfred G. Gilman for his support and Christaine Kleuss for critical reading of the manuscript. Technical assistance was skillfully provided by Laurence Cooke, Dennis Bellotto, Helen Aronovich, and Rebecca Burkett. This work was supported by American Heart Association grants 90144 (to K.H.M.) and 91R-977 (from the Texas Affiliate to S.M.M.), US. Public Health Service grants HL45326 and HL17669 (to K.H.M.), GM50515 (to S.M.M.), and GM34497 (to A.G.G.) and American Cancer Society Grant BE30-0 (to A.G.G.). 1 Hepler, J. R. and Gilman, A. G. (1992) Trends Biochem. Sci. 17,383-387 2 Spiegel, A. M., Backlund, P. S., Jr., Butrynski, J. E., Jones, T. L. Z. and Simonds, W. F. (1991) Trends Biochem. Sci. 16,338-341 3 Iiiiguez-Lluhi, J., Kleuss, C. and Gilman, A. G. (1993) Trends Cell Biol. 3,230-236 4 Ohguro, H., Fukada, Y., Takao, T., Shimonishi, Y., Yoshizawa, T. and Akino, T. (1991) EMBO J. 10, 3669-3674 5 Wildman, D. E., Tamir, H., Leberer, E., Northup, J. K. and Dennis, M. (1993) Proc. Natl. Acad. Sci. U.S.A. 90,794-798 6 Buss, J. E., Mumby, S. M., Casey, P. J., Gilman, A. G. and Sefton, B. M. (1987) Proc. Natl. Acad. Sci. U.S.A. 84,7493-7497 7 Mumby, S. M., Heuckeroth, R. O., Gordon, J. I. and Gilman, A. G. (1990) Proc. Natl. Acad. Sci. U.S.A. 87, 728-732 8 Neubert, T. A., Johnson, R. S., Hurley, J. B. and Walsh, K. A. (1992) J. Biol. Chem. 267,18274- 18277 9 Kokame, K., Fukada, Y., Yoshizawa, T., Takao, T. and Shimonishi, Y. (1992) Nature (London) 359, 749-752 10 Linder, M. E., Pang, I.-H., Duronio, R. J., Gordon, J. I., Sternweis, P. C. and Gilman, A. G. (1991) J. Biol. Chem. 266,4654-4659 11 Taussig, R., Ifiiuez-Lluhi, J. and Gilman, A. G. (1993) Science 261,218-221 12 Jones, T. L. Z., Simonds, W. F., Merendino, J. J., Jr., Brann, M. R. and Spiegel, A. M. (1990) Proc. Natl. Acad. Sci. U.S.A. 87,568-572 13 Linder, M. E., Middleton, P., Hepler, J. R., Taussig, R., Gilman, A. G. and Mumby, S. M. (1993) Proc. Natl. Acad. Sci. U.S.A. 90,3675-3679
Volume 23
14 Parenti, M., Vigano, M. A., Newman, C. M. H., Milligan, G. and Magee, A. I. (1993) Biochem. J. 291, 349-353 15 Wedegaertner, P. B. Chu, D. H., Wilson, P. T., Levis, M. J. and Bourne, H. R. (1993) J. Biol. Chem. 268, 25001-25008 16 Degtyarev, M. Y., Spiegel, A. M. and Jones, T. L. Z. (1993) Biochemistry 32,8057-8061 17 Veit, M., Nurnberg, B., Spicher, K., Harteneck, C., Ponimaskin, E., Schultz, G. and Schmidt, M. F. G. (1994) FEBS Lett. 339,160-164 18 Gilman, A. G. (1987) Annu. Rev. Biochem. 56, 6 15-649 19 Mumby, S. M., Kleuss, C. and Gilman, A. G. (1994) Proc. Natl. Acad. Sci. U.S.A. 91,2800-2804 20 Degtyarev, M. Y., Spiegel, A. M. and Jones, T. L. Z. (1993) J. Biol. Chem. 268,23769-23772 21 Wedegaertner, P. B. and Bourne, H. R. (1994) Cell 77, 1063- 1070 22 Magee, A. I. (1990) J. Cell Sci. 97,581-584 23 Degtyarev, M. Y., Spiegel, A. M. and Jones, T. L. 2. (1994) FASEB J. 8, A1431 24 Higashijima, T. and Ross, E. M. (199 1) J. Biol. Chem. 266,12655-12661 25 Chang, W.-J., Ying, Y.-S., Rothberg, K. G., Hooper, N. M., Turner, A. J., Gambliel, H. A., De Gunzburg, J., Mumby, S. M., Gilman, A. G. and Anderson, R. G. W. (1994) J. Cell Biol. 126, 127-138 26 Paige, L. A., Nadler, M. J. S., Harrison, M. L., Cassady, J. M. and Geahlen, R. L. (1993) J. Biol. Chem. 268, 8669-8674 27 Shenoy-Scaria, A. M., Timson, G. I,. K., Kwong, J., Shaw, A. S. and Lublin, D. M. (1993) Mol. Cell. Biol. 13.6385-6392 28 Alland, L., Peseckis, S. M., Atherton, R. E., Berthiaume, L. and Resh, M. D. (1994) J. Biol. Chem. 269,1670 1- 16705 29 Shenoy-Scaria, A. M., Dietzen, D. J., Kwong, J., Link, D. C. and Lublin, D. M. (1994) J. Cell Biol. 126, 353-363 30 Anderson, R. G. W. (1993) Proc. Natl. Acad. Sci. USA. 90,10909- 10913 31 Sargiacomo, M., Sudol, M., Tang, 2. L. and Lisanti, M. P. (1993) J. Cell Biol. 122,789-807 32 Muntz, K. H., Sternweis, P. C., Gilman, A. G. and Mumby, S. M. (1992) Mol. Biol. Cell 3,49-61 33 Moore, M. S., Mahaffey, D. T., Brodsky, F. M. and Anderson, R. G. W. (1987) Science 236,558-563 Received 5 September 1994