Palmitoylation regulates raft affinity for the majority of integral raft proteins

Abstract

The physical basis for protein partitioning into lipid rafts remains an outstanding question in membrane biology that has previously been addressed only through indirect techniques involving differential solubilization by nonionic detergents. We have used giant plasma membrane vesicles, a plasma membrane model system that phase separates to include an ordered phase enriching for raft constituents, to measure the partitioning of the transmembrane linker for activation of T cells (LAT). LAT enrichment in the raft phase was dependent on palmitoylation at two juxtamembrane cysteines and could be enhanced by oligomerization. This palmitoylation requirement was also shown to regulate raft phase association for the majority of integral raft proteins. Because cysteine palmitoylation is the only lipid modification that has been shown to be reversibly regulated, our data suggest a role for palmitoylation as a dynamic raft targeting mechanism for transmembrane proteins.

Fig. 1.

LAT enriches in raft phase in NEM derived GPMVs. (A) LAT-GFP segregates away from the nonraft marker rhPE in nGPMVs (induced with 2 mM NEM), in contrast to slight depletion in pdGPMVs (25 mM PFA and 2 mM DTT). (B) Preparation-dependent partitioning is not observed for either (B) the raft-enriched GPI-GFP or (C) the nonraft TfR, but is significantly different for (D) H-ras-GFP (rhPE marker images not included; the marker is enriched in the H-ras-GFP-rich phase in both cases). Average + SD from 7–10 vesicles/condition; representative of two independent experiments. (E) Graphic of LAT constructs with palmitoylated cysteines highlighted by red lines and TMD in the gray region. (F) LAT is enriched in the raft phase of nGPMVs regardless of FP oligomerization or PID (average + SD from three independent experiments; 7–10 vesicles/condition/experiment. *p < 0.05; ** p < 0.01; p > 0.1 between all constructs in nGPMVs). Quantifications from 6–10 vesicles/condition; vesicles shown in all figures are 5- to 10-μm diameter.

Fig. 2.

DTT induces LAT translocation to nonraft phase independent of cross-linking. (A) K_p,raft as a function of GPMV treatment. GPMVs derived with NEM (2 mM ; lanes 1, 3, 4, and 5) or PFA + DTT (25 mM + 2 mM, respectively; lanes 2 and 6), then treated for 2 h at the condition shown (*** p < 0.001; * p < 0.05; ^nsp > 0.05 of 7–10 vesicles/condition and representative of three independent experiments). (B) Representative images of LAT-TMD-mRFP (red) with respect to the raft phase marker CTxB-A488. (C) Western blot against native LAT after pulldown of palmitoylated proteins by acyl-biotinyl exchange. Treatment of lysates with 2 mM DTT leads to ∼50% loss of LAT signal, whereas signal from 20 mM DTT treatment is equivalent to complete depalmitoylation by HAM. (D) Densitometric quantification is average + SD from three independent experiments (** p < 0.01; *** p < 0.001 compared to control).

Fig. 3.

Specific homodimerization enhances raft phase partitioning. (A) Schematic of LAT-IDer inducible dimerization. Representative images show LAT-IDer reverses phase preference from slightly nonraft to highly raft enriched following dimerization. (B) Dimerization enhances raft partitioning of LAT-IDer in both pdGPMVs and nGPMVs, dependent on palmitoylation (average + SD of 7–12 vesicles/condition; *** p < 0.001; p > 0.05 between the two 20 mM DTT-treated conditions; results representative of three independent experiments). Average + SD of 6–11 vesicles/condition; results representative of three independent experiments (*** p < 0.001; ** p < 0.01; ^nsp > 0.05).

Fig. 4.

Effect of palmitoylation-deficient mutants and palmitoylation inhibitor on raft phase partitioning. (A) Loss of palmitoylation at membrane-embedded C26 leads to complete reversal of raft phase partitioning while lack of (juxtamembrane) C29 palmitoylation leads to approximately equal partitioning between the two phases. (B) Pretreatment of cells with the palmitoylation inhibitor 2-BP abrogates LAT raft phase partitioning regardless of isolation agents. Average + SD of 8–11 vesicles/condition; results representative of two independent experiments (*** p < 0.001; ** p < 0.01; ^nsp > 0.05).

Fig. 5.

Majority of integral raft phase proteins in nGPMVs are palmitoylated. (A) Depalmitoylating effect of DTT is observed for most palmitoylated proteins as evidenced by reduction of silver staining of acyl-biotinyl exchanged pulldowns in DTT-treated lysates. (B) Densitometric quantification suggests that treatment with 20 mM DTT removes all S-linked palmitates, at parity with 0.5 M HAM (average plus SD of the seven major bands between 28–51 kDa, representative of three separate experiments; ** p < 0.01; *** p < 0.001). (C) Representative images of total external membrane protein (stained with monomeric fluorescent anti-biotin following nonspecific biotinylation) in GPMVs following treatment with 20 mM DTT followed by 0.1 U/mL PI-PLC. (D) External plasma membrane protein is somewhat depleted from raft phase in nGPMVs; treatment with either DTT or PI-PLC reduces raft phase signal (average + SD of three independent experiments each with 7–14 vesicles/condition; * p < 0.05). (E) Fluorescent quantification of raft protein abundance following removal of palmitoylated TM proteins by 20 mM DTT or GPI-APs by PI-PLC allows quantification of the composition of the nGPMVs (average + SD from three independent experiments). Nonraft outnumber raft proteins ∼2∶1 while palmitoylated proteins comprise > 50% of integral raft proteins.