The protein kinase Akt acts as a coat adaptor in endocytic recycling

Abstract

Coat proteins have a central role in vesicular transport by binding to cargoes for their sorting into intracellular pathways. Cargo recognition is mediated by components of the coat complex known as adaptor proteins^1-3. We previously showed that Arf-GAP with coil-coil, ANK repeat and PH domain-containing protein 1 (ACAP1) functions as an adaptor for a clathrin coat complex that has a function in endocytic recycling^4-6. Here, we show that the protein kinase Akt acts as a co-adaptor in this complex, and is needed in conjunction with ACAP1 to bind to cargo proteins to promote their recycling. In addition to advancing the understanding of endocytic recycling, we uncover a fundamentally different function in which a kinase acts, as Akt in this case is an effector rather than a regulator in a cellular event.

Extended Data Fig. 1. Further characterizing cargo binding interactions.

a, b, c, Pulldown interactions from 2 independent experiments are quantified for studies shown in Figure 1a (a), in Figure 1b (b), and in Figure 1c (c). Data available in Source Data Extended Data Fig 1a, 1b, 1c. d, e, Pulldown studies followed by Coomassie staining to detect soluble forms Akt bound to GST-α5 on beads (d) or soluble Akt and ACAP1 bound to GST-α5 on beads (e); n= 2 independent experiments. Input shows 10% of soluble components used in the incubation. Data available in Unprocessed Blots Extended Data Fig 1d, 1e. f, A soluble complex containing GST-α5, Akt, and ACAP1 is detected by incubating the three components in solution, followed by isolation using glutathione beads, and then Coomassie staining to assess complex formation; n= 2 independent experiments. Input shows 10% of soluble components used for the incubation. Data available in Unprocessed Blots Extended Data Fig 1f. g, h, i, Pulldown interactions from 2 independent experiments are quantified for studies shown in Figure 1d (g), in Figure 1e (h), and in Figure 1f (i). Data available in Source Data Extended Data Fig 1g, 1h, 1i. j, k, Primary data for biolayer interferometry measurements that quantify the interaction between GST-α5 and the Akt kinase domain shown in Figure 1g (j), and between the Akt kinase domain and the carboxyl portion of ACAP1 shown in Figure 1h (k); n= 2 independent experiments. l, A model for how Akt and ACAP1 act as co-adaptors in binding to the α5β1 integrin heterodimer.

Extended Data Fig. 2. Akt having a non-kinase role in integrin recycling.

a, Efficiency of siRNA against Akt and its rescues in HeLa cells that express the ACAP1 mutant (S554D), as detected by immunoblotting of whole cell lysates; n= 2 independent experiments. Data available in Unprocessed Blots Extended Data Fig 2a. b, Representative primary images from the integrin recycling assay shown in Figure 1i; n= 3 independent experiments. The colocalization of endosomal β1 with Rab11, a marker of the recycling endosome, is assessed; β1 (red), Rab11 (green), bar = 10 um.

Extended Data Fig. 3. Further supporting a non-kinase role of Akt in integrin recycling.

a, Cell migration of the S554D-expressing HeLa cells as assessed through the transwell-based assay; n= 3 independent experiments, with each experiment examining three fields of transwell membranes. Primary images are shown on left, bar = 200 um. Quantitation is shown on right, mean +/− SD, *p=1.57×10⁻²⁰, NS p=0.125, paired two-tailed student’s t-test. Data available in Source Data Extended Data Fig 3a. b, c, Integrin recycling assay assessing the effect of treating the S554D-expressing HeLa cells (b), or control HeLa cells (c), with the Akt kinase inhibitor GDC0068; n= 3 independent experiments. Quantitation is shown on left, mean +/− SD, with statistics performed for the 5-minute time point, NS p=0.248 (for analysis in b), *p=1.54×10⁻²⁵ (for analysis in c), paired two-tailed student’s t-test. Primary images, assessing the colocalization of endosomal β1 with cellubrevin (a marker of the recycling endosome), are shown on right, β1 (red), Cbv (green), bar = 10 um. Data available in Source Data Extended Data Fig 3b, 3c.

Extended Data Fig. 4. Further characterizing the effect of expressing a mutant α5 integrin.

a, Primary images from the integrin recycling assay shown in Figure 2d; n= 3 independent experiments. The colocalization of endosomal β1 with Rab11 is assessed; β1 (red), Rab11 (green), bar = 10 um. b, Cell migration of HeLa cells that express different α5 forms as assessed through the transwell-based assay; n= 3 independent experiments, with each experiment examining three fields of transwell membranes. Primary images are shown on left, bar = 200 um. Quantitation is shown on right, mean +/− SD, *p=2.08×10⁻²⁴, paired two-tailed student’s t-test. Data available in Source Data Extended Data Fig 4b. c, Representative primary images for the colocalization study shown in Figure 2g; n= 3 independent experiments. The colocalization of endosomal β1 with Rab11 is assessed, β1 (red), Rab11 (green), bar = 10 um.

Extended Data Fig. 5. Further supporting a non-kinase role of Akt in TfR recycling.

a, Primary images from the TfR recycling assay shown in Figure 3a; n= 3 independent experiments. The colocalization of endosomal Tf with Rab11 is assessed; Tf (red), Rab11 (green), bar = 10 um. b, TfR recycling assay examining the effect of siRNA against Akt, and also rescue using wild-type (WT) or kinase-dead (K179M) Akt in HEK293 cells; n= 3 independent experiments, with each experiment examining 10 cells. Quantitation is shown above, mean +/− SD, with statistics performed on the 25-minute time point, *p=8.86×10⁻²⁷, NS p=0.557, paired two-tailed student’s t-test. Primary images along with line scans are shown below, Tf (red), cellubrevin (green), bar = 10 um. Data available in Source Data Extended Data Fig 5b.

Extended Data Fig. 6. Further characterizing how Akt acts in the endocytic transport of TfR.

a, TfR internalization assay examining the effect of siRNA against Akt, n= 3 independent experiments, with each experiment examining 10 cells. Quantitation is shown above, mean +/− SD, with statistics performed for the 5-minute time point, NS p=0.538, paired two-tailed student’s t-test. Primary images are shown below, Tf (red), EEA1 (a marker of the early endosome, green), bar = 10 um. Data available in Source Data Extended Data Fig 6a. b, TfR recycling assay examining the effect of siRNA against Akt, and also rescues using various forms of Akt in HeLa cells; n= 3 independent experiments, with each experiment examining 10 cells. Quantitation is shown above, mean +/− SD, with statistics performed on the 25-minute time point, *p=2.68×10⁻²⁷, NS p=0.303, paired two-tailed student’s t-test. Primary images along with line scans are shown below, Tf (red), Rab11 (green), bar = 10 um. Data available in Source Data Extended Data Fig 6b.

Extended Data Fig. 7. Further characterizing how Akt acts in TfR recycling.

a, TfR recycling assay examining the effect of treating HeLa cells with the Akt kinase inhibitor GDC0068; n= 3 independent experiments, with each experiment examining 10 cells. Quantitation is shown on left, mean +/− SD with statistics performed on the 30-minute time point, NS p=0.222, paired two-tailed student’s t-test. Primary images along with line scans are shown on right; Tf (red), cellubrevin (green), bar = 10 um. Data available in Source Data Extended Data Fig 7a. b, Efficiency of siRNA against Akt1 and siRNA against Akt2 in HeLa cells, as assessed by immunoblotting of whole cell lysates; n= 2 independent experiments. Actin level confirms similar levels of cells examined. Data available in Unprocessed Blots Extended Data Fig 7b. c, d, TfR recycling assay examining the effect of siRNA against different isoforms of Akt in HeLa cells; n= 3 independent experiments, with each experiment examining 10 cells. Quantitation is shown in (c), mean +/− SD, with statistics performed on the 25-minute time point, *p=1.72×10⁻²⁴ (control vs si-Akt1), *p=1.53×10⁻³⁰ (control vs si-Akt2), *p=1.78×10⁻⁴³ (control vs si-Akt1/si-Akt2), paired two-tailed student’s t-test. Data available in Source Data Extended Data Fig 7c. Primary images along with line scans are shown in (d), Tf (red), Rab11 (green), bar = 10 um.

Extended Data Fig. 8. Further supporting Akt acts as a co-adaptor with ACAP1 in cargo binding.

a, Schematic showing the sequence of the TfR cytoplasmic domain, and the portions covered by the N19 and the NΔ19 constructs. b, Pulldown studies titrating increasing level of different fusion proteins of Akt and ACAP1 for their binding to the TfR cytoplasmic domain; n= 2 independent experiments. Left panel compares binding by Akt-ACAP1 heterodimer and Akt homodimer. Right panel compares binding by Akt-ACAP1 heterodimer and ACAP1 homodimer. Data available in Unprocessed Blots Extended Data Fig 8b. c, d, Co-precipitation studies examining the effect of siRNA against Akt on the association of endosomal TfR with ACAP1 and Akt in HeLa cells (c), and siRNA against ACAP1 on the association of endosomal TfR with Akt and ACAP1 in HeLa cells (d); n= 2 independent experiments. Biotin-labeled Tf was internalized for 2 hours to label the endosomal pool of TfR. Immunoblotting of whole cell lysates confirms the efficiency of siRNA treatment. Data available in Unprocessed Blots Extended Data Fig 8c, 8d.

Extended Data Fig. 9. Further characterizing membrane recruitment of Akt.

a, Isolating a membrane fraction from HeLa cells enriched for the recycling endosome using a sucrose gradient established through equilibrium centrifugation; n= 2 independent experiments. Fractions enriched for the recycling endosome were identified by tracking cellubrevin and internalized Tf (which bound to endosomal TfR), and not surface Tf (which bound to surface TfR). Data available in Unprocessed Blots Extended Data Fig 9a. b, c, d, Recruitment studies showing that Akt alone cannot be recruited to endosomal membrane (b), while ARF6 alone in its activate form can be recruited to endosomal membrane (c), and clathrin recruitment to endosomal membrane requires ARF6 with either Akt or ACAP1 (d); n= 2 independent experiments. Cellubrevin tracks endosomal membrane. Data available in Unprocessed Blots Extended Data Fig 9b, 9c, 9d. e, Recruitment study examining the relative levels of ARF6, Akt, ACAP1, and clathrin recruited to liposomes; n= 2 independent experiments. Input shows the total amount of each component added for the incubation. Data available in Unprocessed Blots Extended Data Fig 9e.

Extended Data Fig. 10. Further characterizing endogenous Akt at the recycling endosome.

a, Confocal microscopy examining the colocalization of endogenous Akt with endosomal Tf in HeLa cells; n= 3 independent experiments, with each experiment examining 10 cells. Primary images are shown, Akt (green), Tf (red), bar = 10 um. b, Confocal microscopy examining the effect of a more denaturing fixative (containing methanol and acetone) on the ability to detect endogenous clathrin at the recycling endosome in HeLa cells; n= 3 independent experiments, with each experiment examining 10 cells. Primary images along with line scan are shown on left for the colocalization of clathrin with endosomal Tf, clathrin (green), Tf (red), bar = 10 um. Quantitation is shown on right, mean +/− SD, NS p=0.106, paired two-tailed student’s t-test. Data available in Source Data Extended Data Fig 10b. c, Confocal microscopy examining the colocalization of different forms of ARNO with endosomal Tf in HeLa cells; n= 3 independent experiments, with each experiment examining 10 cells. Primary images along with line scan are shown on left for the colocalization of endogenous Akt with endosomal Tf, ARNO (green), Tf (red), bar = 10 um. Quantitation is shown on right, mean +/− SD, NS p=0.686, paired two-tailed student’s t-test. Data available in Source Data Extended Data Fig 10c. d, Confocal microscopy examining the effect of serum stimulation on Akt localization at the recycling endosome in HeLa cells; n= 3 independent experiments, with each experiment examining 10 cells. Primary images along with line scan are shown above for the colocalization of endogenous Akt with endosomal Tf, Akt (red), Tf (green), bar = 10 um. Quantitation is shown below, mean +/− SD, NS p=0.325, paired two-tailed student’s t-test. Data available in Source Data Extended Data Fig 10d.

Figure 1.. Akt binds to integrin and ACAP1.

a, b, c, Pulldown studies examining Akt and/or ACAP1 binding to integrin tails, as assessed by immunoblotting; n= 2 independent experiments. Input shows 15% of soluble components used in the incubation. Data available in Unprocessed Blots Fig 1a, 1b, 1c. d, e, f, Pulldown studies examining interactions through immunoblotting: Akt binding to GST-α5 (d), soluble ACAP1 binding to Akt forms as GST fusions (e), or Akt binding to ACAP1 forms as GST fusions (f); n= 2 independent experiments. In (d), input shows 50% of soluble components used in the incubation. Data available in Unprocessed Blots Fig 1d, 1e, 1f. g, h, Biolayer interferometry quantifying the interaction between GST-α5 and the Akt kinase domain (g), or between the Akt kinase domain and the carboxyl portion of ACAP1 (h); n= 2 independent experiments. Double reference subtracted binding curves generated from 6 data points were analyzed to obtain binding constant (K_D) values using globally fit 1:1 Langmuir binding model, which are expressed as mean +/− standard error (SEM). Data available in Source Data Fig 1g, 1h. i, Integrin recycling assay performed on HeLa cells that express the S554D mutant of ACAP1, examining the effect of siRNA against Akt, and rescue using wild-type (WT) or kinase-dead (K179M) form of Akt; n= 3 independent experiments, with each experiment examining 10 cells. Mean +/− standard deviation (SD), with statistics performed for the 5-minute time point, *p=1.87×10⁻³¹, NS (non-significant) p=0.112, paired two-tailed student’s t-test. Data available in Source Data Fig 1i. j, Co-precipitation study examining the interaction of transfected myc-tagged ACAP1 with endogenous Akt in HeLa cells; n= 2 independent experiments. Data available in Unprocessed Blots Fig 1j. k, Co-precipitation study examining the effect of siRNA against Akt on the formation of a complex that contains endosomal β1, ACAP1, and Akt in the S554D-expressing HeLa cells; n= 2 independent experiments. The mutant ACAP1 is myc-tagged and rescues are done using myc-tagged forms of Akt. Thus, immunoblotting using an anti-myc antibody detects both ACAP1 and Akt. Data available in Unprocessed Blots Fig 1k.

Figure 2.. Akt recognizes a recycling sorting signal in the α5 tail.

a,b, Pulldown studies examining Akt binding to truncation or point mutants of the α5 tail as GST fusions; n= 2 independent experiments. Truncation forms are defined with respect to the sequence shown in Figure 2c; FL (full-length), N10 (first 10 aa), N20 (first 20 aa), N14 (first 14 aa), N17 (first 17 aa), NΔ14 (lacking the first 14 aa), NΔ10 (lacking the first 10 aa), FL/4–5AA (full-length with residues 4 and 5 replaced by alanines), FL/6–7AA (full-length with residues 6 and 7 replaced by alanines), FL/8–10AA (full-length with residues 8, 9, 10 replaced by alanines). Data available in Unprocessed Blots Fig 2a, 2b. c, A schematic showing residues of the α5 tail, with numbering starting with the most membrane proximal residue. Residues underlined are required for Akt binding. d, Integrin recycling assay performed on HeLa cells that express different α5 forms; n= 3 independent experiments, with each experiment examining 10 cells. Mean +/− SD, with statistics performed for the 5-minute time point, *p=7.8×10⁻²⁶, paired two-tailed student’s t-test. Data available in Source Data Fig 2d. e, Co-precipitation studies examining the effect of expressing the mutant α5 on the formation of a complex containing endosomal β1, ACAP1 and Akt in HeLa cells; n= 2 independent experiments. Data available in Unprocessed Blots Fig 2e. f, Co-precipitation studies examining α5β1 heterodimerization in HeLa cells that express either wild-type or mutant α5; n= 2 independent experiments. Data available in Unprocessed Blots Fig 2f. g, Confocal microscopy examining the colocalization of the α5β1 integrin (as tracked through β1) and Rab11 (a marker of the recycling endosome) in HeLa cells that express either wild-type or mutant α5; n= 3 independent experiments, with each experiment examining 10 cells. Mean +/− SD, NS p=0.402, paired two-tailed student’s t-test. Data available in Source Data Fig 2g.

Figure 3.. Akt acts in the cargo sorting of TfR to promote its recycling.

a, TfR recycling assay examining the effect of siRNA against Akt, and rescue using wild-type (WT) or kinase-dead (K179M) Akt in HeLa cells; n= 3 independent experiments, with each experiment examining 10 cells. Mean +/− SD, with statistics performed on the 25-minute time point, *p=1.69×10⁻³⁸, NS p=0.161, paired two-tailed student’s t-test. Data available in Source Data Fig 3a. b, Pulldown study examining Akt binding to the TfR cytoplasmic domain; n= 2 independent experiments. Input shows 10% of the soluble amount used in the incubation. Data available in Unprocessed Blots Fig 3b. c, Pulldown study examining the effect of titrating increasing level of a α5 cargo peptide (wild-type or mutant that cannot bind Akt), on Akt binding to the TfR cytoplasmic domain on beads; n= 2 independent experiments. Data available in Unprocessed Blots Fig 3c. d, e, f, Pulldown studies examining Akt binding to truncation form of the TfR cytoplasmic domain (d), to point mutants of the TfR-N19 construct (e), or to the point mutants of the TfR-NΔ19 construct (f); n= 2 independent experiments. Data available in Unprocessed Blots Fig 3d, 3e, 3f. g, h, Pulldown studies titrating increasing level of Akt for binding to TfR-N19 or TfR-NΔ19 (g), or increasing level of ACAP1 for binding to these two TfR constructs (h); n= 2 independent experiments. Data available in Unprocessed Blots Fig 3g, 3h.

Figure 4.. Akt links clathrin to cargoes and mediates clathrin recruitment to endosomal membrane.

a, Pulldown studies examining clathrin binding to domains of Akt; n= 2 independent experiments. Binding was assessed by either immunoblotting (left panel) or Coomassie staining (right panel). Data available in Unprocessed Blots Fig 4a, 4b. b, c, Pulldown studies examining the effect of having Akt (b) or Akt and ACAP1 (c) being present on clathrin binding to cargoes; n= 2 independent experiments. Data available in Unprocessed Blots Fig 4b, 4c. d, e, Co-precipitation studies examining the effect of siRNA against Akt on the association of clathrin with ACAP1 in HeLa cells (d) or siRNA against ACAP1 on the association of clathrin with Akt in HeLa cells (e); n= 2 independent experiments. Data available in Unprocessed Blots Fig 4d, 4e. f, g, Recruitment studies showing Akt recruitment to endosomal membrane requiring activated ARF6 (f), or ACAP1 recruitment to endosomal membrane requiring activated ARF6 (g); n= 2 independent experiments. Membrane fraction is tracked by the presence of cellubrevin. Data available in Unprocessed Blots Fig 4f, 4g. h, Recruitment studies examining clathrin recruitment to endosomal membrane with respect to the presence of ARF6 (left panel), or the additional presence of Akt and ACAP1 (right panel); n= 2 independent experiments. Membrane fraction is tracked by the presence of cellubrevin. Data available in Unprocessed Blots Fig 4h.

Figure 5.. Akt at the recycling endosome exists mostly within assembled complexes.

a, Confocal microscopy examining the colocalization of endogenous Akt with the endosomal pool of transferrin (Tf) in HeLa cells; n= 3 independent experiments, with each experiment examining 10 cells. Primary images along with line scan are shown, Akt (red), Tf (green), bar = 10 um. b, Subcellular fractionation of HeLa cells assessing the relative distribution of endogenous Akt in the total membrane versus the cytosol fraction; n= 2 independent experiments. Actin tracks cytosol, while cellubrevin tracks endosomal membrane. Data available in Unprocessed Blots Fig 5b. c, Confocal microscopy examining the effect of a more denaturing fixative (containing methanol and acetone) in detecting endogenous Akt at the recycling endosome in HeLa cells; n= 3 independent experiments, with each experiment examining 10 cells. Primary images along with line scan are shown on left for the colocalization of Akt with endosomal Tf, Akt (red), Tf (green), bar = 10 um. Quantitation is shown on right, mean +/− SD, *p=1.18×10⁻²², paired two-tailed student’s t-test. Data available in Source Data Fig 5c. d, e, An antibody-based assay to detect Akt assembled on endosomal membrane; n= 2 independent experiments. Akt along with ACAP1, and clathrin (d), or Akt alone (e) is recruited to endosomal membrane in an ARF6-dependent fashion. An Akt antibody is then incubated with the membrane, either before or after membrane solubilization with detergent, followed by isolation of the antibody to detect associated Akt. Data available in Unprocessed Blots Fig 5d, 5e. f, Confocal microscopy examining the effect of overexpressing either wild-type (WT) or catalytic-dead (E156K) ARNO on the ability to detect endogenous Akt at the recycling endosome in HeLa cells; n= 3 independent experiments, with each experiment examining 10 cells. Primary images along with line scan are shown on left for the colocalization of Akt with endosomal Tf, Akt (red), Tf (green), bar = 10 um. Quantitation is shown on right, mean +/− SD, *p=1.57×10⁻²⁴ (no ARNO vs ARNO), *p=1.56×10⁻³⁷ (ARNO WT vs ARNO E156K), paired two-tailed student’s t-test. Data available in Source Data Fig 5f.