Validation of mNG-β1 and HiBit-β1 cells.

(A) Cartoon showing the peptide tag complementation systems used to label endogenous Gβ1 subunits. (B) SDS-PAGE of HiBit-β1 and mNG-β1 cell lysates; the predicted molecular weights of the edited gene products are 38.9 and 41.1 kilodaltons (KDa), respectively; representative of 3 independent experiments. (C) In permeabilized nucleotide-depleted cells BRET between dopamine D2R-Nluc receptors and mNG-β1-containing heterotrimers increases in response to dopamine (DA; 100 μM) and reverses after addition of GDP (100 μM); mean ± 95% CI; n=27 replicates from 2 independent experiments. (D) In intact cells BRET between HiBit-β1 and the Gβγ sensor memGRKct-Venus increases after stimulation of D2R dopamine, β2AR adrenergic, or M3R acetylcholine receptors with DA (100 μM), isoproterenol (Iso; 10 μM) and acetylcholine (Ach; 100 μM), respectively. Signals reversed when receptors were blocked with haloperidol (10 μM), ICI 118551 (10 μM) or atropine (10 μM); mean ± 95% CI; n=16 replicates from 4 independent experiments.

Endogenous G proteins are abundant on the plasma membrane but not large organelles.

(A) A single field of view of mNG-β1 cells at three magnifications; scale bars are 40 μm, 20 μm and 10 μm. (B) mNG-β1 does not colocalize with expressed markers of the endoplasmic reticulum (ER; PTP1b), mitochondria (MT; MOA) or medial Golgi apparatus (GA; GalT); intensity line profiles depict absolute fluorescence intensity in each channel; scale bars are 2 μm.

Endogenous G proteins colocalize with markers of endosomes and lysosomes.

(A) mNG-β1 colocalizes with expressed markers of early endosomes (EE; FYVE and rab5a), recycling endosomes (RE; rab11a), late endosomes (LE; rab7a) and lysosomes (lyso; LysoView 633); intensity line profiles depict absolute fluorescence intensity in each channel; scale bars are 5μm. (B) Mean mNG-β1 fluorescence intensity line profiles drawn across the plasma membrane (PM) and FYVE-positive vesicles; mean ± 95% CI; n=40 vesicles/cells. (C) mNG-β1 signal/background ratios for regions of interest surrounding the plasma membrane (PM; n=99), FYVE-positive (n=125) and rab5a-positive (n=56) early endosomes, and rab7a-positive (n=26) late endosomes; horizontal lines represent the median. (D) Bystander net BRET signals between HiBit-β1 and Venus-tagged markers of the plasma membrane (PM), endoplasmic reticulum (ER), mitochondria (MT), early endosomes (FYVE and rab5a), recycling endosomes (rab11a) and late endosomes (rab7a); horizontal lines represent the median; n=5-7 independent experiments.

Constitutive G protein endocytosis is inefficient.

(A) mNG-β1 colocalizes with newly internalized endocytic vesicles labeled with FM4-64 and CellMask Deep Red (arrowheads); scale bar is 2 μm. (B) A fluorescence intensity line profile for mNG-β1, FM4-64 and CellMask normalized to the peak value of each label at the plasma membrane (PM). (C) Mean mNG-β1, FM4-64 and CellMask fluorescence intensity line profiles drawn across vesicles, normalized to fluorescence intensity at the plasma membrane for each label; mean ± 95% CI; n=45 vesicles/cells.

Receptor activation does not change G protein endocytosis.

(A) mNG-β1 colocalizes with newly internalized endocytic vesicles labeled with FM4-64 and SNAP-tagged β2 adrenergic receptor (β2AR) labeled with Alexa Fluor 674; scale bar is 5 μm. Cells were stimulated with 10 μM isoproterenol for 15 minutes to induce β2AR internalization. (B) A fluorescence intensity line profile for mNG-β1, FM4-64 and β2AR normalized to the peak value of each label at the plasma membrane (PM). (C) Mean mNG-β1, FM4-64 and β2AR fluorescence intensity line profiles drawn across vesicles, normalized to fluorescence intensity at the plasma membrane for each marker; mean ± 95% CI; n=91 vesicles/cells. (D) Normalized peak mNG-ý1/FM4-64 did not differ between vesicles that contained receptors (R; n=91) and vesicles formed by constitutive endocytosis (no R; n=45); n.s., not significant, P=0.20, unpaired t-test. (E) Bystander BRET between HiBit-β1 and markers of early endosomes (EE), recycling endosomes (RE) and late endosomes (LE) was unchanged after 30 minutes of receptor activation with isoproterenol (Iso; 10 μM), dopamine (DA; 100 μM) or acetylcholine (Ach; 100 μM) compared to the DPBS vehicle alone (control); mean ± SD, n=4 independent experiments; no agonist-treated group was significantly different from the control, paired t-test with a false discovery rate (FDR) of 1%.

Receptor-mediated accumulation of cyclic AMP (cAMP) is similar in HiBit-β1, mNG-β1 and parent cell lines.

Activation of endogenous β adrenergic receptors with isoproterenol (Iso; 10 μM) produces similar increases in cAMP as indicated by the Nluc-EPAC-VV cAMP sensor in (A) parental HEK and HiBit-β1 cells and (B) parental mNG2(1-10) and mNG-β1 cells; mean ± 95% CI of 20 replicates from n=5 independent experiments.

G proteins are not abundant on the endoplasmic reticulum (ER).

Exemplary images of mNG-β1 cells coexpressing the ER marker mRuby2-PTP1b. Intensity line profiles depict fluorescence intensity normalized to the maximum value in each channel; scale bars are 5 μm.

G proteins are not abundant on mitochondria (MT).

Exemplary images of mNG-ý1 cells coexpressing the MT marker mRuby2-MOA. Intensity line profiles depict fluorescence intensity normalized to the maximum value in each channel; scale bars are 5 mm.

G proteins are not abundant on the medial Golgi apparatus (GA).

Exemplary images of mNG-β1 cells coexpressing the GA marker mRuby2-Golgi-7 (GalT). Intensity line profiles depict fluorescence intensity normalized to the maximum value in each channel; scale bars are 5 μm.

G proteins colocalize with the early endosome marker FYVE on some endosomes.

Exemplary images of mNG-β1 cells coexpressing the early endosome marker pmCherry-2xFYVE (FYVE). Intensity line profiles depict fluorescence intensity normalized to the maximum value in each channel; scale bars are 5 μm.

G proteins colocalize with the early endosome marker rab5a on some endosomes.

Exemplary images of mNG-β1 cells coexpressing the early endosome marker mRuby2-rab5a (rab5a). Intensity line profiles depict fluorescence intensity normalized to the maximum value in each channel; scale bars are 5 μm.

G proteins colocalize with the recycling endosome marker rab11a in a perinuclear region.

Exemplary images of mNG-β1 cells coexpressing the recycling endosome marker mCherry-rab11a (rab11a). Intensity line profiles depict fluorescence intensity normalized to the maximum value in each channel; scale bars are 10 μm.

G proteins colocalize with the late endosome marker rab7a on many endosomes.

Exemplary images of mNG-β1 cells coexpressing the late endosome marker mCherry-rab7a (rab7a). Intensity line profiles depict fluorescence intensity normalized to the maximum value in each channel; scale bars are 5 μm.

G proteins are abundant on lysosomes.

(A) Exemplary images of mNG-β1 cells stained with LysoView 633 (LV633); scale bars are 5 μm. (B) Exemplary images of mNG-β1 cells incubated overnight with 10,000 m.w. CF 640 dextran (640 dex); scale bars are 5 μm.

G proteins colocalize with the lysosome marker LAMP1.

Exemplary images of mNG-ý1 cells coexpressing the lysosome marker LAMP1-mScarlet; scale bars are 5 mm.

G proteins colocalize with the trans-Golgi marker TGNP.

Exemplary images of mNG-ý1 cells coexpressing the trans-Golgi marker mCherry-TGNP. Intensity line profiles depict fluorescence intensity normalized to the maximum value in each channel; scale bars are 5 mm.

Constitutive endocytosis of G proteins is inefficient.

(A) Images of mNG-β1 cells immediately after and 10 minutes after staining with FM4-64. (B) An absolute fluorescence intensity line profile for mNG-β1 and FM4-64; scale bar is 2 μm. In this example absolute fluorescence intensity at the plasma membrane (PM) was similar for the two labels.

mNG-HRas ct is less abundant on endocytic vesicles than the plasma membrane.

(A) mNG-HRas ct colocalizes with newly internalized endocytic vesicles labeled with FM4-64 (arrowheads); scale bar is 2 μm. (B) A fluorescence intensity line profile for mNG-HRas ct and FM4-64 normalized to the peak value of each label at the plasma membrane (PM). (C) Mean mNG-HRas ct and FM4-64 fluorescence intensity line profiles drawn across vesicles, normalized to fluorescence intensity at the plasma membrane for each label; mean ± 95% CI; n=78 vesicles/cells.

Transient translocation of endogenous HiBit-β1 from the plasma membrane to intracellular compartments during activation.

(A) Bystander BRET between HiBit-β1 and a Venus-tagged marker of the plasma membrane (PM) decreased, whereas BRET between HiBit-β1 and a Venus-tagged markers of the endoplasmic reticulum (ER), early endosomes (EE), recycling endosomes (RE) and late endosomes (LE) increased during receptor activation with isoproterenol (Iso; 10 μM), dopamine (DA; 100 μM) or acetylcholine (Ach; 100 μM) compared to the DPBS vehicle alone (control); mean ± SD, n=5 independent experiments; all agonist-treated groups were significantly different from the control, paired t-test with a false discovery rate (FDR) of 1%. (B) Time course of bystander BRET between HiBit-β1 and a Venus-tagged marker of the ER during activation of D2R with DA (100 μM), followed by block of D2R with the antagonist haloperidol (Hal; 10 μM); mean ± 95% CI of 24 replicates from n=4 independent experiments.

G protein abundance on endosomes after GPCR and G protein activation.

Bystander BRET between HiBit-β1 and markers of early endosomes (EE), recycling endosomes (RE) and late endosomes (LE) was largely unchanged after 5 minutes (A) or 15 minutes (B) of receptor activation with isoproterenol (Iso; 10 μM), dopamine (DA; 100 μM) or acetylcholine (Ach; 100 μM) compared to the DPBS vehicle alone (control); mean ± SD, n=4 independent experiments; only two agonist-treated groups (indicated with *) were significantly different from the control, paired t-test with a false discovery rate (FDR) of 1%.