Oligomer-to-monomer transition underlies the chaperone function of AAGAB in AP1/AP2 assembly

Significance

AAGAB is an assembly chaperone governing the assembly of the adaptor complexes 1 and 2(AP1 and AP2). Mutations on the AAGAB gene cause the skin disease punctate palmoplantar keratoderma type 1 (PPKP1). We found that the C-terminal domain (CTD), often missing in the mutant protein in PPKP1 patients, mediates AAGAB homodimerization by forming an antiparallel dimer. Interestingly, CTD also binds and stabilizes the γ subunit in the AP1 complex or the α subunit in the AP2 complex in a 1:1 stoichiometry. Our findings demonstrate a dual role of AAGAB and provide a molecular explanation for disease-causing AAGAB mutations. The oligomerization state transition mechanism may also underlie the functions of other assembly chaperones.

Abstract

Assembly of protein complexes is facilitated by assembly chaperones. Alpha and gamma adaptin-binding protein (AAGAB) is a chaperone governing the assembly of the heterotetrameric adaptor complexes 1 and 2 (AP1 and AP2) involved in clathrin-mediated membrane trafficking. Here, we found that before AP1/2 binding, AAGAB exists as a homodimer. AAGAB dimerization is mediated by its C-terminal domain (CTD), which is critical for AAGAB stability and is missing in mutant proteins found in patients with the skin disease punctate palmoplantar keratoderma type 1 (PPKP1). We solved the crystal structure of the dimerization-mediating CTD, revealing an antiparallel dimer of bent helices. Interestingly, AAGAB uses the same CTD to recognize and stabilize the γ subunit in the AP1 complex and the α subunit in the AP2 complex, forming binary complexes containing only one copy of AAGAB. These findings demonstrate a dual role of CTD in stabilizing resting AAGAB and binding to substrates, providing a molecular explanation for disease-causing AAGAB mutations. The oligomerization state transition mechanism may also underlie the functions of other assembly chaperones.

Formation of protein complexes is facilitated by assembly chaperones that transiently interact with individual subunits and assembly intermediates to prevent aggregation and ensure correct assembly (1). Substrate-specific assembly chaperones have been identified for a number of protein complexes including ribosomes, nucleosomes, proteosomes, and spliceosomes (1).

In clathrin-mediated membrane trafficking, alpha and gamma adaptin-binding protein (AAGAB, also known as p34) is an assembly chaperone required for the formation of the clathrin adaptors AP1 and AP2 (2, 3). Clathrin coats drive the budding of vesicles from the endosome, the plasma membrane, and the trans-Golgi network (4–6). Clathrin relies on adaptors to recruit cargo proteins to vesicle budding sites (7, 8). Two predominant clathrin cargo adaptors are the AP1 complex involved in trafficking from the endosome and the trans-Golgi network and the AP2 complex regulating clathrin-mediated endocytosis (5, 9). Both AP1 and AP2 are heterotetrameric complexes with two large subunits (γ/α and β), one medium subunit (µ), and one small subunit (σ) (10–15).

In AAGAB-assisted AP1 assembly, AAGAB first binds to the γ subunit (gamma adaptin) to form an AAGAB:γ binary complex, which then recruits the σ subunit to form an AAGAB:γ:σ ternary complex. In these complexes, AAGAB stabilizes the γ and σ subunits as well as the γ:σ hemicomplex. Subsequently, β and µ subunits displace AAGAB, leading to the formation of the AP1 complex (2, 3). AAGAB regulates the assembly of the AP2 adaptor through a similar mechanism (2). Without the assistance of the assembly chaperone AAGAB, AP1 and AP2 complexes fail to form, resulting in degradation of their subunits and membrane trafficking defects (2). Autosomal dominant mutations in the AAGAB gene cause punctate palmoplantar keratoderma type 1 (PPKP1, also known as Buschke-Fischer-Brauer disease), a skin disease characterized by lesions on palms and soles (16–20).

It remains unclear how AAGAB recognizes AP1/2 subunits and how AAGAB-AP1/2 interactions promote AP1/2 adaptor assembly. In this work, we uncovered that AAGAB itself dimerizes in the absence of AP1/2, mediated by a highly conserved C-terminal domain (CTD). Without CTD, AAGAB becomes unstable and is degraded in the cell. We solved the crystal structure of the CTD dimer, which revealed a dimerization interface formed by two extended antiparallel alpha helices. Interestingly, AAGAB uses this exact CTD to bind and stabilize AP1γ and AP2α subunits. There is just one copy of AAGAB in the AAGAB:γ and AAGAB:α binary complexes, suggesting that AAGAB dimers dissociate into monomers upon AP1/2 binding. Structure-guided mutations in the CTD region disrupt both AAGAB dimerization and its binding to the γ subunit of AP1 and the α subunit of AP2. Consistent with these in vitro observations, the CTD mutations disrupt the assembly of AP1 and AP2 adaptors and impair clathrin-mediated trafficking in the cell. Since AAGAB mutations in PPKP1 patients usually lead to truncated mutants lacking the CTD, our findings demonstrate that the mutants are unable to engage in AP1γ and AP2α interactions. Taken together, our study unveiled that the AP1/2-binding region of AAGAB oligomerizes prior to AP1/2 association to stabilize the substrate-free state. We suggest that the oligomer-to-monomer transition mechanism also operates in other chaperone-assisted assembly processes.

Results

AAGAB Exists as a Dimer in Its Resting State.

AAGAB is a cytosolic protein with two conserved regions—an N-terminal G protein-like domain (GD, residues 1 to 177) and a CTD (residues 258 to 315) without any known homologous structure (Fig. 1A). The two conserved domains are connected by a ~70-residue non-conserved linker (Fig. 1A). To investigate its biochemical and structural characteristics, we expressed and purified full-length (FL) AAGAB protein from Escherichia coli. The purified protein exhibited excellent purity and homogeneity based on the size-exclusion chromatography (SEC) profile and Coomassie blue-stained sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) gel (Fig. 1 A and B). Interestingly, we observed that FL AAGAB eluted at a position much earlier than its calculated molecular weight (MW) of 34.6 kDa. The elution position at 13.4 mL corresponded to an experimental MW of 144.2 kDa, which suggests an oligomeric assembly (Fig. 1A). Serial dilution of AAGAB at concentrations ranging from 36 to 4 μM did not change AAGAB elution position from the Superdex 200 column (Fig. 1A), indicating that the oligomeric assembly is stable at the lowest concentration examined.

Fig. 1.

AAGAB exists as a dimer in its resting state. (A) Domain architecture of FL AAGAB protein (Top) and SEC profiles of AAGAB protein at three threefold serial dilutions from a Superdex 200 Increase 10/300 column. Starting and ending residue numbers and elution positions of protein standards with known MW are labeled above. The small peak marked by the red asterisk is from a contaminating protein. (B) Coomassie blue-stained gels showing AAGAB proteins before and after crosslinking. New protein bands consistent with dimer and tetramer sizes appear after crosslinking, in addition to the non-crosslinked monomer band. (C) SV-AUC analysis of AAGAB at three different concentrations shows AAGAB (~3.3 S species) as a dimer. The left side peaks marked by the red asterisk are likely from a contaminating protein. (D) SEC-MALS characterizes FL AAGAB as an entity of 57.8 kDa in solution, close to the dimer assembly. Black trace: refractive index (RI) of AAGAB normalized to 1; red trace: calculated MW across the AAGAB peak. (E) Representative immunoblots showing the total expression of 3xFLAG- and V5-tagged AAGAB in AAGAB KO HeLa cells. Control: cells transfected with an empty vector. Same volumes of control and transfected whole cell lysates (WCLs) were loaded into the gel for comparison. (F) Representative immunoblots showing the interaction between 3xFLAG- and V5-tagged AAGAB in the cells. The 3xFLAG-AAGAB protein was immunoprecipitated using anti-FLAG antibodies out of the whole cell lysates from E and the presence of 3xFLAG-AAGAB and V5-AAGAB in the immunoprecipitated products were visualized by immunoblotting.

In parallel, we treated FL AAGAB with a lysine-specific crosslinking reagent and examined the crosslinked proteins using SDS-PAGE in the hope to gain additional information regarding the oligomeric state of AAGAB. As expected, denatured non-crosslinked FL AAGAB migrated as monomers on SDS-PAGE (Fig. 1B). Crosslinked AAGAB, however, exhibited two new bands consistent with dimeric and tetrameric MW (Fig. 1B). Since SEC elution positions may be affected by the shape of the molecule and crosslinking results may be skewed by concentrations of protein and crosslinking reagents, they are not reliable ways to assess the oligomeric states of molecules. Thus, we carried out sedimentation velocity-analytical ultracentrifugation (SV-AUC) and SEC coupled with multi-angle light scattering (SEC-MALS) experiments to resolve the oligomeric state of AAGAB (Fig. 1 C and D). Unexpectedly, SV-AUC and SEC-MALS yielded a MW of 58 ± 2 kDa and 57.8 (± 0.11%) kDa (Fig. 1 C and D and SI Appendix, Table S1), respectively, pointing to a dimer assembly of AAGAB. The c(s) distributions from the AUC data taking from 21, 7, and 2 μM AAGAB proteins were similar, indicating no substantial dissociation at these concentrations. The two-dimensional c(s, f/f₀) distribution of the SV-AUC sample revealed an averaged frictional ratio f/f₀ of 1.852 (SI Appendix, Fig. S1A), suggesting a significant elongated shape of AAGAB and explaining the early elution position observed in SEC. We pursued a third, independent method mass photometry (21, 22) to measure the molecular mass of AAGAB. MW derived from mass photometry was 69 kDa, corresponding to a dimer (SI Appendix, Fig. S1B). Combining AUC, SEC-MALS, and mass photometry results, we conclude that AAGAB exists predominantly as a dimer prior to AP1/2 binding.

Next, we examined whether AAGAB oligomerizes in the cell. AAGAB was N-terminally tagged with either a V5 or 3xFLAG epitope, which does not interfere with AAGAB function in the cell (2). V5- and 3xFLAG-tagged AAGAB proteins were co-expressed in AAGAB knockout (KO) HeLa cells (Fig. 1E). We then performed co-immunoprecipitation using anti-FLAG antibodies. We observed that V5-tagged AAGAB was co-precipitated with 3xFLAG-tagged AAGAB (Fig. 1F), suggesting that AAGAB oligomerizes in the cell. These data are consistent with the in vitro results and further support the notion that AAGAB oligomerizes in its resting state.

AAGAB Dimerization Is Mediated by Its CTD.

To gain insights into the dimerization interface(s), we performed mass spectrometry to analyze the lysine-specific crosslinked FL AAGAB proteins. We identified multiple intermolecular crosslinks between AAGAB molecules. Interestingly, all crosslinked lysines mapped exactly to the short conserved CTD of AAGAB (Fig. 2 A and B and SI Appendix, Table S2). All seven lysines in the CTD were crosslinked to other lysines from another AAGAB molecule, whereas none of the seven lysines in the GD was crosslinked (Fig. 2 A and B and SI Appendix, Table S2). These crosslinking results suggest that the CTD mediates the dimerization of AAGAB.

Fig. 2.

AAGAB dimerization is mediated by its CTD. (A) Crosslinking mass spectrometry (XL-MS) mapped AAGAB interfacial residues solely to the CTD. (B) Sequence analysis of the human AAGAB protein. The amino acids are colored based on their conservation levels. Predicted secondary structural elements are marked on top of the sequence: cylinder: α helix; arrow: β strand. The core CTD (residues 258 to 301) is highlighted in the black outline box. Crosslinked lysines identified in A are marked by red arrowheads. (C) Top: Diagram of FL AAGAB with the GD and the CTD marked; Bottom: SEC profiles of FL AAGAB, AAGAB GD (residues 1 to 177) and AAGAB CTD (residues 252 to 315) on a HiLoad 16/600 Superdex 200 column. Elution positions of protein standards with known MW are marked on the top. (D) SEC-MALS profiles of MBP-CTD (residues 258 to 301) (green) and MBP control (gray). The calculated MW distributions of MBP-CTD (residues 258 to 301) (red) and MBP (orange) are overlaid onto the respective RI traces.

Both the conserved GD and CTD of AAGAB are predicted to fold into ordered structures, whereas the non-conserved linker region in the middle is predicted to be unstructured (Figs. 1A and 2B). We expressed the two structured domains individually in Escherichia coli and purified them in a similar manner as FL AAGAB (Fig. 2C). The GD was monomeric based on its elution position on SEC (Fig. 2C, theoretical 19.7 kDa vs. experimental 23.3 kDa). In contrast, the CTD of AAGAB (residues 252 to 315, monomeric MW 7.2 kDa) exhibited an estimated MW of 19.4 kDa on SEC (Fig. 2C), more in line with an oligomeric state. Based on our secondary structure prediction, the CTD contains a long alpha helix flanked by flexible loops (Fig. 2B). All intermolecularly crosslinked lysines are present in the alpha helical region (Fig. 2A). We next removed the flexible loops and fused the shortened CTD (residues 258 to 301) to maltose-binding protein (MBP), a monomeric protein with a MW of 40.3 kDa and measuring 41.6 kDa from SEC-MALS (Fig. 2D and SI Appendix, Table S1). The shortened CTD of AAGAB (aa 258 to 301) was sufficient to drive the dimerization of the fusion protein (89.1 kDa as measured by SEC-MALS vs. monomeric MW 45.4 kDa) (Fig. 2D and SI Appendix, Table S1), confirming that CTD is a bona fide dimerization domain.

Crystal Structure of AAGAB CTD Reveals an Antiparallel Dimer of Helices.

We next crystallized the core region of AAGAB CTD (residues 258 to 301) and solved its atomic structure using single-wavelength anomalous dispersion of the selenomethionyl protein (Fig. 3A). The crystal structure was refined to 2.1 Å with R_work/R_free = 21.2/25.9% (SI Appendix, Table S3). The SeMet sites enabled us to register all residues without ambiguity. The high-quality main-chain and side-chain electron density allowed us to confirm the assignment with high confidence (SI Appendix, Fig. S2).

Fig. 3.

Crystal structure of AAGAB CTD reveals an antiparallel dimer. (A) Cartoon representation of two symmetry-related AAGAB CTD dimers in two orthogonal views. One dimer is colored in blue, while the other is colored in pale green. The N and C termini are marked for one molecule in each dimer. The two potential dimer interfaces A and B are boxed in red. (B) The interaction details of the dimer interface A. Since the interfaces are symmetrical, only one set of residues involved in dimerization are labeled for clarity. (C) The interface B between two symmetry-related dimers. (D) SEC profiles of the AAGAB-WT, L269R, L279R/V286R, F262R/F266R, and F262R/F266R/L269R (“3R”) mutants from a Superdex 200 Increase 10/300 column. Elution positions of protein standards with known MW are marked on the top. The mutated residues are labeled on the CTD sequence. Green triangles represent residues on the dimer interface A as shown in (B), blue triangles represent residues involved in crystal packing interface B as shown in (C), and orange triangles mark residues involved in both interactions. Both y axes are OD280 (mAU). All SEC profiles are scaled to the y axes on the left side of the chart except WT AAGAB, which is scaled to the y axes on the right.

The crystal structure of AAGAB CTD revealed a dimeric assembly in each asymmetric unit. The CTD of AAGAB folds into a continuous alpha helix with a bend introduced by P280, adopting a boomerang-like configuration (Fig. 3 A and B). Two CTDs run antiparallelly to form a homodimer (Fig. 3B). The two concave sides of the boomerangs interlock each other through an extensive dimer interface, burying a surface area of ~890 Ų on each CTD molecule. The dimer interface (interface A) is mainly mediated by hydrophobic residues and hydrophobic moieties of charged residues, including L269, M272, A276, L279, E282, V286, E289, and W296 (Fig. 3B). We also observed a few polar interactions in this region. E282 forms a hydrogen bond with another E282 from the other chain in the dimer, while simultaneously forming a salt bridge with K285 also from the other chain in the dimer (Fig. 3B). In the crystal lattice, one CTD molecule also interacts with another CTD from the neighboring asymmetric unit in a head-to-head fashion. This secondary interface (interface B), burying 327 Ų surface area on each CTD molecule, is formed by the surface opposite to the primary extensive dimerization interface A on the N-terminal region of CTD. F262 and F266 on the two molecules constitute the core of the predominantly hydrophobic interface, while L265 and L269 reinforce the interface from periphery (Fig. 3C).

To investigate the functional role of the dimerization interfaces observed in the crystal structure, we introduced single (L269R), double (F262R/F266R, L279R/V286R), and triple mutations (F262R/F266R/L269R, or 3R) into the potential dimer interfaces of the FL AAGAB protein. Interestingly, the peak of AAGAB L269R in the SEC profile shifted from the dimer position to a monomer position (Fig. 3D and SI Appendix, Fig. S3A and Table S1), indicating a near complete disruption of dimeric assembly. Double mutant L279R/V286R also shifted its peak elution volume to a monomer position (Fig. 3D and SI Appendix, Fig. S3B and Table S1). In contrast, the F262R/F266R double mutant profile displayed a ~50:50 split between dimers and monomers (Fig. 3D and SI Appendix, Fig. S3C and Table S1). Given F266 is very close to the key residue of L269, the partial disruption of dimerization in F262R/F266R may be attributed to perturbation of L269 environment rather than direct involvement in dimerization. The triple mutant 3R (F262R/F266R/L269R) is completely monomeric (Fig. 3D and SI Appendix, Fig. S3D and Table S1). The quintuple mutant 5R (F262R/F266R/L269R/L279R/V286R) eluted at the same monomeric position as the 3R mutant (SI Appendix, Fig. S4). Since the 3R and 5R mutants are monomeric and stable in solution, we focused on the 3R and 5R mutants in the subsequent studies.

Overall, the dimeric assembly uncovered in the structure agrees well with the biochemical data of FL AAGAB wild-type (WT) and mutants and the CTD domain in solution.

AAGAB interacts with AP1γ Subunit and AP2α Subunit with a 1:1 Molar Ratio.

To regulate the assembly of AP1 and AP2 adaptors, AAGAB first interacts with the γ subunit of AP1 or the α subunit of AP2 to form AAGAB:γ or AAGAB:α binary complexes (3). In these binary complexes, AAGAB stabilizes the γ or α subunit and prepares them for subsequent assembly with other AP1/2 subunits (2). Next, we expressed and purified recombinant AAGAB:γ and AAGAB:α binary complexes from Escherichia coli. The SEC peak elution position of the tag-free AAGAB:γ (trunk domain, residues 2 to 595) binary complex yielded a calculated MW of 149 kDa (Fig. 4A), suggesting a 1:1 or 2:1 molar ratio of the AAGAB:γ binary complex (theoretical MW 101.9 kDa and 136.5 kDa, respectively). The band intensities of γ and AAGAB from the peak fractions, however, are consistent with a molar ratio of 1:1 (Fig. 4B). We therefore conclude that the AAGAB:γ binary complex contains one copy of each protein.

Fig. 4.

AAGAB interacts with AP1γ and AP2α subunits with a 1:1 molar ratio. (A) SEC profile of the AAGAB:AP1γ (trunk domain, residues 2 to 595) binary complex from a Superdex 200 Increase 10/300 column. Elution positions of protein standards with known MW are marked on the top. (B) AAGAB and AP1γ proteins in peak fractions of A were analyzed by SDS-PAGE and stained by Coomassie blue. (C) SEC profile of AAGAB:AP2α (trunk domain, residues 2 to 621) binary complexes from a HiLoad 16/600 Superdex 200 column. Elution positions of protein standards with known MW are marked on the top. (D) AAGAB and AP2α proteins in peak fractions of C were analyzed by SDS-PAGE and stained by Coomassie blue.

When co-expressed in Escherichia coli, His₆-SUMO-AAGAB interacted with GST-tagged α subunit (trunk domain, residues 2 to 621) in a glutathione S-transferase (GST) pull-down assay (2). To further characterize the AAGAB:α binary complex, we removed both the His₆-SUMO and GST tags after purification using tandem Ni-NTA beads and GSTrap affinity chromatography. The purified tag-free AAGAB:α binary complex exhibited excellent homogeneity and purity on SEC and Coomassie blue-stained SDS-PAGE gel (Fig. 4 C and D), confirming that AAGAB formed a stoichiometrically stable complex with the α subunit in solution. Interestingly, the SEC profile of the AAGAB:α binary complex exhibited a MW of 139 kDa, suggesting a 1:1 or 2:1 molar ratio (theoretical MW 104.1 kDa and 138.7 kDa, respectively). The Coomassie blue-stained SDS-PAGE gel of AAGAB:α peak fractions also suggests that the AAGAB:α binary complex adopts a 1:1 stoichiometry. Finally, the SEC-MALS experiments produced a MW of 105.9 kDa for the AAGAB:α binary complex, confirming the 1:1 ratio between the two subunits (SI Appendix, Fig. S5 and Table S1).

Taken together, AAGAB forms a stoichiometric binary complex with the γ subunit of AP1 or the α subunit of AP2 with a 1:1 ratio. These data indicate that AAGAB undergoes a dimer-to-monomer transition when it engages in AP1 and AP2 binding.

CTD Mediates the Interaction of AAGAB with the γ Subunit of AP1 Adaptor and the α Subunit of AP2 Adaptor.

We next sought to pinpoint the region(s) in AAGAB that interacts with AP1γ subunit and AP2α subunit. We co-expressed GST-tagged γ subunit (trunk domain, residues 2 to 595) with FL or truncated AAGAB bearing an N-terminal His₆-SUMO tag in Escherichia coli and evaluated the quantity of soluble AP1γ from the soluble fraction of lysates using GST pull-down (Fig. 5A). Consistent with SEC results (Fig. 4 A and B), co-expression with FL AAGAB strongly increased soluble GST-tagged γ subunit through the formation of the AAGAB:γ binary complex (Fig. 5A). When GD was co-expressed, however, GST-tagged γ subunit was not stabilized even when GD expressed abundantly (Fig. 5A). In contrast, soluble AP1γ subunit was markedly enhanced when co-expressed with an extended version of AAGAB CTD (residues 228 to 315) and was able to pull down the extended AAGAB CTD. These results suggest that AAGAB CTD, but not GD, binds and stabilizes the AP1γ subunit. We further tested the core region of AAGAB CTD (residues 258 to 301). Again, soluble GST-tagged γ was strongly enhanced through a direct interaction with AAGAB CTD (residues 258 to 301) (Fig. 5A). Thus, we conclude that AAGAB CTD directly recognizes and stabilizes the γ subunit of AP1 adaptor.

Fig. 5.

AAGAB CTD directly recognizes and stabilizes AP1γ and AP2α subunits. (A) Coomassie blue-stained gel showing the GST pull-down results of GST-AP1γ (trunk domain, residues 2 to 595) co-expressed with His₆-SUMO tagged FL AAGAB and individual domains in Escherichia coli. Note that there is no specific soluble protein in lane 2 (indicated with a red arrowhead). (B) Coomassie blue-stained gel showing the binding of AP2α subunit (trunk domain, residues 2 to 621) to co-expressed GST-AAGAB CTD (residues 258 to 301) in Escherichia coli. The identity of GST-AAGAB CTD is confirmed by 3C protease treatment to remove the GST tag. (C and D) Coomassie blue-stained gels showing the soluble GST-tagged AP1γ subunit (C) and AP2α subunit (D) when they are expressed alone or co-expressed with His₆-SUMO (HS) tagged AAGAB-WT or 3R mutant in Escherichia coli. WCL, whole cell lysate. SN, supernatant. Elu, elution.

To test whether CTD also binds the α subunit of AP2 adaptor, we co-expressed GST-tagged CTD (residues 258 to 301) and His₆-SUMO tagged α subunit (trunk domain, residues 2 to 621) in Escherichia coli. After Ni affinity chromatography and Ulp1 protease treatment to remove His₆-SUMO from the α subunit, we were able to pull down tagless AP2α subunit bound to GST-AAGAB-CTD (residues 258 to 301) using GST affinity chromatography (Fig. 5B). Thus, AAGAB CTD directly recognizes and stabilizes both the γ and α subunits.

Next, we sought to define the binding interface of the AAGAB-AP1/2 binary complexes. As expected, co-expression with AAGAB-WT substantially increased the levels of soluble AP1γ (Fig. 5C and SI Appendix, Fig. S6A). Interestingly, the levels of soluble AP1γ were not enhanced when it was co-expressed with the AAGAB-3R mutant, which was defective in dimerization, and an AAGAB:γ binary complex was not observed (Fig. 5C and SI Appendix, Fig. S6A). Since expression levels of AAGAB-WT and AAGAB-3R were comparable (SI Appendix, Fig. S7), we conclude that the triple mutations F262R/F266R/L269R, which disrupt AAGAB dimerization, also abolish the interaction between AAGAB and AP1γ subunit. Similarly, AAGAB-WT, but not the AAGAB-3R mutant, bound and stabilized AP2α subunit when the proteins were co-expressed (Fig. 5D and SI Appendix, Figs. S6B and S7B). We observed similar disruption of AP1γ or AP2α interaction with the single mutation L269R (SI Appendix, Fig. S8). Taken together, these data demonstrate that AAGAB CTD directly binds and stabilizes the γ and α subunits, and the AAGAB dimerization interface is also involved in the interactions.

CTD Mutations Disrupt AP1- and AP2-Dependent Membrane Trafficking in the Cell.

To examine the physiological relevance of our structural and biochemical findings, we determined how the CTD mutations affect AAGAB function in the cell. In AAGAB KO cells, AP1 and AP2 adaptors are lost, disrupting both AP1- and AP2-mediated trafficking of membrane proteins (2, 3). As a result, surface levels of HA-GLUT4-GFP reporter and transferrin receptor (TfR) were elevated in AAGAB KO HeLa cells, and the phenotype was rescued by expression of a WT AAGAB gene (Fig. 6 A and B) (2, 3). By contrast, we observed that expression of the AAGAB-3R or -5R mutant did not restore TfR surface levels (Fig. 6 A and B). Likewise, deletion of the CTD (AAGABΔCTD, residues 1 to 258) also abolished the ability of AAGAB to restore TfR surface levels (Fig. 6 A and B). We also examined the effects of the AAGAB mutations on stabilization of endogenous AP1 and AP2 subunits. We observed that expression of WT AAGAB in AAGAB KO cells increased the levels of AP1γ subunit and AP2α subunit by ~threefold over their levels in the AAGAB KO cells, whereas none of the AAGAB mutants was able to enhance their expression (Fig. 6C and SI Appendix, Fig. S9). The AAGAB-3R and -5R mutants increased the levels of AP1σ subunit to the same level as WT AAGAB (Fig. 6C and SI Appendix, Fig. S9), implying that CTD is not involved in σ binding. While 3R and 5R mutants displayed comparable expression levels as WT AAGAB, AAGABΔCTD was expressed at a significantly lower level despite similar mRNA levels (Fig. 6C and SI Appendix, Figs. S9 and S10), suggesting CTD is required for AAGAB stability. Consistently, all tested adaptor subunit levels remained low when AAGABΔCTD was expressed (Fig. 6C and SI Appendix, Fig. S9). These results are consistent with our in vitro data and demonstrate that CTD mediates the binding of AAGAB to AP1γ and AP2α subunits in the cell.

Fig. 6.

CTD mutations disrupt AP1- and AP2-dependent clathrin-mediated trafficking. (A) Representative confocal microscopy images showing surface levels of HA-GLUT4-GFP reporter and TfR in AAGAB KO HeLa cells expressing WT or mutant AAGAB. Surface TfR levels of non-permeabilized cells were labeled using anti-TfR antibodies and Alexa Fluor 568-conjugated secondary antibodies (red). Cellular HA-GLUT4-GFP reporter levels were visualized by GFP fluorescence (green). Nuclei were stained with Hoechst 33342 (blue). Images were captured using a 100× oil immersion objective on a Nikon A1 Laser Scanning confocal microscope. (Scale bars, 10 μm.) (B) Normalized surface levels of TfR in AAGAB KO HeLa cells expressing WT or mutant AAGAB genes. Surface TfR levels were measured by flow cytometry. Cells were disassociated by Accutase and stained with monoclonal anti-TfR antibodies and APC-conjugated secondary antibodies. Allophycocyanin (APC) fluorescence measurements of ~5,000 cells were collected on a CyAn ADP analyzer. Mean APC fluorescence of mutant cells was normalized to that of WT cells. Data are presented as mean ± SD, n = 3. **P < 0.01, ***P < 0.001. P values were calculated using one-way ANOVA. (C) Representative immunoblots showing the expression of the indicated proteins in AAGAB KO HeLa cell expressing WT or mutant AAGAB. Control: cells transfected with an empty vector.

Discussion

Individual subunits and assembly intermediates of a protein complex are often prone to misfolding, aggregation, and degradation. Substrate-specific assembly chaperones like AAGAB stabilize these structures and maintain them at states competent for assembly. In this work, we discovered that in its resting state, AAGAB exists as a homodimer. AAGAB dimerization is mediated by a conserved CTD. Intriguingly, the same CTD interacts with AP1γ and AP2α subunits in a 1:1 molar ratio and is required for stabilizing these subunits both in vitro and in vivo. Moreover, the same hydrophobic residues, in particular L269, are involved in both AAGAB dimerization and interactions with AP1γ or AP2α. We propose a model that AAGAB exists in the dimeric state prior to AP1/2 binding, and dimerization shields the hydrophobic AP1/2-binding interface from the aqueous environment and thus stabilizes the substrate-free AAGAB proteins. To engage in AP1/2 binding, an AAGAB dimer dissociates into monomers, and each of them binds one AP1γ or AP2α subunit using the same CTD, leading to formation of binary complexes (Fig. 7A). This model nicely explains how AAGAB itself is stabilized prior to AP1/2 binding.

Fig. 7.

Model of AAGAB-assisted AP1/2 adaptor assembly. (A) AAGAB itself exists as a dimer in its resting state. An AAGAB dimer dissociates into monomers upon binding to AP1γ or AP2α subunit using its CTD domain. In the AAGAB:γ and AAGAB:α binary complexes, AAGAB stabilizes the AP1/2 subunits and prepare them for subsequent association with other subunits; σ subunit joins in to form an AAGAB:γ/α:σ ternary hemicomplex, possibly through interaction with AAGAB GD; β and μ subunits then displace AAGAB to form the fully assembled AP complexes, while AAGAB returns to the dimer reservoir. Under pathological conditions, AAGAB lacking the CTD domain is unable to interact with and stabilize AP1γ or AP2α subunit, leading to degradation of AP1γ or AP2α subunit, no assembly of AP complexes, and disrupted membrane trafficking. (B) Reported heterozygous PPKP1-causing mutations in the AAGAB coding region are mapped onto AAGAB domain diagram. Nonsense and frameshift mutations are depicted by blue arrows, whereas missense mutations are marked by red arrows.

We identified two potential CTD dimerization interfaces from the crystal structure: interface A that spreads through most of CTD (Fig. 3B) and interface B which is localized to the N terminus of CTD (Fig. 3C). We believe that interface A represents the true dimerization interface in solution for the following reasons: 1) It buries ~890 Ų surface area upon dimer formation and engages nine residues on each AAGAB protomer, almost three times of interface B (327 Ų); 2) the mutation L279R/V286R on interface A completely disrupted AAGAB dimer, while the interface B mutation F262R/F266R only showed partial disruption (Fig. 3D and SI Appendix, Fig. S3); 3) due to the closeness of F266 to the key residue L269, the disruptive effect observed for F262R/F266R may be indirect through perturbation of L269 environment. Nevertheless, the interface B may mediate transient interactions of two AAGAB dimers into a homotetramer, consistent with our crosslinking data (Fig. 1B).

How is the dimer-to-monomer transition of AAGAB achieved? AAGAB may exist in a dimer-monomer equilibrium, with monomeric AAGAB being the binding-competent form. As such, presence of AP1/2 subunits spontaneously drives dissociation of AAGAB dimers. However, given the hydrophobic nature of the interfacial residues, monomeric AAGAB may not be stable, which is consistent with the fact that we did not observe substantial dissociation of AAGAB in serial dilution experiments (Fig. 1 A and C). Thus, we postulate that dissociation of AAGAB dimers into monomers is triggered by and/or coupled to binding to AP1γ/AP2α (Fig. 7A). It is also possible that unidentified cellular factors exist to facilitate AAGAB dissociation into monomers.

Among a total of 36 known distinct heterozygous disease mutations reported in the AAGAB coding region, 33 are nonsense or frameshift mutations that result in deletion of the entire CTD of AAGAB (Fig. 7B) (16, 17, 23). While these point mutations may also impact AAGAB activity through other mechanisms such as nonsense-mediated mRNA decay, our data clearly demonstrate that these AAGAB mutants lose the ability to bind AP1γ and AP2α subunits (Fig. 7A). Heterozygous mutations of AAGAB in PPKP1 patients reduce functional AAGAB proteins by half, and the consequences on clathrin-mediated trafficking appear to be restricted to certain tissues such as the skin. Alteration of membrane protein homeostasis in the skin is likely one of the causes of PPKP1. Notably, mutations in genes encoding AP1 subunits such as AP1B1, AP1S1, and AP1S3 are reported to underlie skin disorders (24–27), indicating malfunction of AP1-mediated cargo trafficking may be the main cause of PPKP1. Intriguingly, one of the three AAGAB missense mutations causing PPKP1 is within CTD (E282K), and the residue E282 lies at the dimer interface of CTD (Fig. 3B).

The bound AAGAB assists in recruitment of small σ subunits and facilitates AP1γ:σ and AP2α:σ interactions (2, 3). Ultimately, AAGAB is replaced by β and μ subunits, resulting in functional heterotetrameric AP1 and AP2 adaptor complexes and return of AAGAB to the reservoir (Fig. 7A). Although we have reached in-depth understanding of the function for AAGAB CTD region, the exact function of its conserved N-terminal GD region remains unknown. The crystal structure of GD from the putative yeast homologue Irc6p was solved and found to adopt a genuine GTPase fold (28). However, Irc6p lacks guanosine-5'-triphosphate (GTP) binding motifs and barely binds to GTP (28). In human AAGAB GD, the GTP binding motifs are also disrupted, indicating it may not possess any GTP binding or hydrolysis capacity either. It is tempting to propose that AAGAB GD interacts with the σ subunits and brings them to the vicinity of the γ or α subunit. Further biochemical and structural studies are needed to uncover the function of GD and reveal the molecular basis of AAGAB:σ interactions.

A future research direction is to establish the structural basis of the interactions between AAGAB and AP1/2 complex subunits. This information will likely reveal certain differences between AAGAB-AP1 and AAGAB-AP2-binding interfaces and explain why AAGAB does not regulate the AP3 adaptor complex (3). Additionally, such structural information could potentially guide us to selectively disrupt AAGAB interaction with one AP complex without affecting the other by introducing well-positioned mutations. In this way, we will be able to dissect the roles of AP1 and AP2 adaptor complexes in PPKP1 pathogenesis. Importantly, should our pathogenesis model prove valid, chemical chaperones that mimic CTD function and stabilize γ or α subunit could potentially rescue their expression, restore clathrin-mediated membrane trafficking, and reinstate membrane protein homeostasis. Such chemical chaperones may represent new therapeutics to treat PPKP1 patients.

Methods

Construction of Expression Plasmids.

DNA sequences encoding FL human AAGAB (residues 1 to 315) and individual domains (residues 1 to 177, 252 to 315, 258 to 301) were generated using a standard PCR-based cloning strategy. They were all inserted between BamHI and SalI sites at the first multiple cloning site in a modified pRSFDuet-1 vector with a His₆-SUMO (small ubiquitin-related modifier) tag at the N terminus. N-terminally His₆-SUMO tagged mouse AP2α (residues 2 to 621, which shares a 99% sequence identity with human AP2α) was generated in a similar method. AAGAB mutants were generated by a two-step PCR-based overlap extension method.

For GST-tagged recombinant protein expression, FL AAGAB (residues 1 to 315), AAGAB CTD (residues 258 to 301), and AP2α (residues 2 to 621) genes were subcloned into the pGEX-6P-1 vector, whereas AP1γ (residues 2 to 595) was subcloned into the pGEX-4T-3 vector.

For MBP tagged recombinant protein expression, AAGAB CTD (residues 258 to 301) was subcloned into the pMal-c2x vector.

For transient expression in mammalian cells, the human AAGAB gene was subcloned into the p3xFLAG 7.1 vector, yielding the p3xFLAG7.1-AAGAB-WT plasmid. p3xFLAG7.1-AAGAB-3R and p3xFLAG7.1-AAGAB-5R mutants were generated using bacterial expression plasmids as PCR templates and subsequently subcloned into the p3xFLAG7.1 vector. p3xFLAG7.1-AAGABΔCTD was made by inserting a stop codon after residue 258 in the plasmid p3xFLAG7.1-AAGAB-WT.

Protein Sequence Analysis.

The sequence analysis of protein FL AAGAB was performed by the online server PredictProtein (29). It not only generated the secondary structure prediction for FL AAGAB, but also reported the conserved regions in AAGAB using the ConSurf server (https://consurf.tau.ac.il) (30).

Expression and Purification of Recombinant Proteins.

All recombinant proteins were expressed in BL21(DE3) Escherichia coli cells in LB medium supplemented with proper antibiotics. The cells were grown at 37 °C and induced by 0.4 mM IPTG when OD₆₀₀ reached a value of 0.6 to 0.8, followed by culturing overnight at 20 °C. Cells were harvested by centrifugation at 4,000× g for 20 min. Cell pellets were lysed by sonication followed by centrifugation at 15,000× g at 4 °C for 60 min to remove cell debris. All His₆-SUMO tagged proteins were purified from the soluble fractions by nickel affinity chromatography, followed by SUMO protease Ulp1 treatment (w:w 1:1,000) overnight at 4 °C to cleave the His₆-SUMO tag. The His₆-SUMO tag was later removed from samples via a second round of nickel affinity chromatography. Proteins were further purified by SEC using a HiLoad 16/600 Superdex 200 prep grade (PG) column (Cytiva #28-9893-35) in the HEPES-buffered saline (HBS) buffer (50 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer pH 7.5, 150 mM NaCl, and 2 mM β-mercaptoethanol). Peak fractions containing the desired proteins were assessed by SDS-PAGE for purity, pooled and concentrated using Amicon Ultra Centrifugal filters (MilliporeSigma) with a 10 kDa or 50 kDa MW cutoff. Concentrated proteins were aliquoted, flash frozen in liquid nitrogen, and stored at −80 °C. The MBP protein and MBP tagged AAGAB CTD (residues 258 to 301) protein were first purified by amylose resin (NEB #8021S), followed by SEC and the purified protein were concentrated and stored in the same manner as all the His₆-SUMO tagged proteins.

The L-SeMet derivatized protein was produced using the feedback-inhibition of methionine synthesis pathway. L-SeMet was added to bacterial cell cultures at IPTG induction to a final concentration of 0.1 mg/mL. SeMet-containing AAGAB CTD was purified in the same way as the native protein. β-mercaptoethanol (5 mM) was maintained throughout protein purification to prevent oxidation of SeMet.

For AAGAB:AP2α binary protein complex purification, GST-AAGAB (Ampicillin resistant) and His₆-SUMO-AP2α (Kanamycin resistant) plasmids were co-transformed into BL21(DE3) Escherichia coli cells to express the complex. The complex was expressed in the same way as individual proteins. After cell lysis and centrifugation, the binary complex was first purified by nickel affinity chromatography, followed by overnight Ulp1 treatment (w:w 1:1,000) at 4 °C to cleave the His₆-SUMO tag. The Ulp1 treated protein sample was loaded directly onto a prepacked GSTrap column (GE Healthcare). Protease 3C was added to the eluted GST-AAGAB:AP2α binary complex to cleave the GST tag, which was removed by a second GSTrap column. The tag-free binary protein complex was finally purified by SEC using a HiLoad 16/600 Superdex 200 PG column (Cytiva #28-9893-35) in the HBS buffer. AAGAB CTD:AP2α binary complex was expressed and purified using a similar method.

Crystallization, Data Collection, Structure Determination, and Refinement.

AAGAB CTD crystals were grown using the hanging-drop vapor-diffusion method by mixing the protein (27 mg/mL in HBS buffer) with an equal volume of reservoir solution containing 30% glycerol, 0.5 M ammonium phosphate (Hampton Research) at 16 °C. The shining diamond like crystals started to show up after overnight incubation and reached full size within a week. All crystals were flash frozen in liquid nitrogen without any additional cryoprotectant.

Diffraction datasets were collected on 22-ID and 22-BM beamlines (SER-CAT) at Advance Photon Source (APS), Argonne National Laboratory and AMX (17-ID-1), and FMX (17-ID-2) beamlines at National Synchrotron Light Source II (NSLS-II), Brookhaven National Laboratory.

The datasets of human AAGAB CTD were indexed, integrated, and scaled using HKL2000 package (31). The crystal belongs to space group P6₁22 (a = b = 47.539 Å, c = 191.394 Å, α = β = 90°, γ = 120°) and contains two molecules per asymmetric unit. The structure was determined by single-wavelength anomalous diffraction (SAD) using data collected at Se-peak wav elength to a resolution of 2.4 Å. The position of selenium atoms was located by the program AutoSol, and the initial model was built by the program AutoBuild, which was later extended to 2.1 Å from a native dataset. Further model improvement was carried out with alternate rounds of refinement using Phenix.refine (32) and model building via COOT (33). The structure refinement was completed with cycles of individual B-factor refinement along with TLS parameters, leading to a final structure with R_work of 21.2% and R_free of 25.9%. The final model is of good stereochemical quality and Ramachandran plot of the main-chain angles showed 100% of the residues found in the favored region. The data collection and refinement statistics for this structure are listed in SI Appendix, Table S3.

Analytical Gel Filtration.

Recombinant FL AAGAB proteins were analyzed on a Superdex 200 increase 10/300 GL column (Cytiva #28-9909-44) with a flow rate of 0.5 mL/min and an injection volume of 0.5 mL. All other protein samples were analyzed on HiLoad 16/600 Superdex 200 PG column (Cytiva #28-9893-35) with a flow rate of 0.8 mL/min and an injection volume of 4 mL. All experiments were performed in the HBS buffer. The columns were calibrated using the Gel Filtration Standard (Bio-Rad #1511901).

Lysine-Specific Crosslinking Coupled with Nano-Liquid Chromatography Mass Spectrometry.

The purified AAGAB protein was crosslinked with a solution of 1:1 BS3-d0: BS3-d4 (ThermoFisher #21590 and #21595) crosslinkers. Crosslinking reaction product was separated by SDS-PAGE. The bands corresponding to dimer and tetramer were cut. Cut gel bands were destained, reduced with dithiothreitol, and digested with trypsin at 37 °C overnight. Tryptic peptides were separated by an Easy Nano LC II system (Thermo Scientific). Mobile phases were water with 0.1% formic acid (A) and acetonitrile with 0.1% formic acid (B). A 3-h gradient (from 5 to 45% B) was performed with a flow rate of 300 nL/min. nLC_eluates were online ionized by nano-electrospray ionization and detected by a Velos LTQ-Orbitrap Mass Spectrometer (Thermo Scientific). Precursor ions were detected in the Orbitrap with a mass resolution of 60 K, while the data-dependent MS2 of the top 10 most abundant precursor ions were carried out in LTQ. The collected .raw files were converted to .mzXML files for crosslinking data analysis with StavroX, an open-access software (34). Search parameters were used as followed: max 4 trypsin miscleavages, dynamic modification of methionine by oxidation, precursor mass accuracy of better than 5 ppm, and fragment ion mass accuracy of better than 0.8 Da. StavroX generated crosslinked peptide list was further manually checked for assignment.

Analytical Ultracentrifugation.

Standard Beckman Epon charcoal-filled dual-sectored centerpieces were sandwiched between sapphire windows in an aluminum cell housing. AAGAB samples were prepared from thawed proteins diluted in “Working Buffer” (20 mM HEPES pH 7.5 and 150 mM NaCl) to different concentrations. The reference sectors were filled with 410 μL of the respective buffer and the sample sectors with 410 μL of the respective AAGAB sample. The assembled cells were placed in a Beckman An50Ti 8-hole rotor, and the rotor was positioned in the Optima XL-I centrifuge. The chamber was evacuated, and the samples allowed to equilibrate at the experimental temperature (20 °C) for 3 h. Centrifugation at 50,000 rpm was then commenced, and it continued overnight. Data were acquired at 280 nm using the absorbance optical system and also with the Rayleigh interferometric optical system (IF). The absorbance data were loaded into SEDFIT and analyzed using the c (s) methodology (35, 36). The sample meniscus position, the frictional ratio, and the time-independent noise were fitted (37). In cases where the c(s,f/f₀) distribution was used (38), the frictional ratios ranged from 1.0 to 3.0, and the resolution in this second dimension was 10, with the regularization level kept identical to that of the s dimension. All figures were rendered using GUSSI (39).

SEC Coupled Multi-Angle Light Scattering (SEC-MALS).

AAGAB protein, mutants, or the binary AAGAB:AP2α complex was injected into a Superdex 200 Increase 10/300 size-exclusion column equilibrated in a buffer containing 25 mM HEPES pH 7.5 and 150 mM NaCl. The column was coupled to a multi-angle light scattering detector (DAWN HELEOS II, Wyatt Technology) and a refractometer (Optilab T-rEX, Wyatt Technology). Data were collected every 0.5 s at a flow rate of 0.5 mL/min at room temperature. Data processing was carried out using the program ASTRA 7.3 (Wyatt Technology). Molar mass and mass distribution of each protein or protein complex were calculated and reported by ASTRA 7.3.

Pull-Down Assay using Co-expressed Recombinant Proteins.

GST-AP1γ and GST-AP2α were expressed individually or in combination with His₆-SUMO-AAGAB-WT or -3R plasmids in Escherichia coli BL21(DE3) cells. Cells were cultured in 500 mL LB medium with corresponding antibiotics. After harvesting, cells were lysed and centrifuged following the same protocol as described above. The cleared cell lysates were equally split for either GST or nickel affinity pull-down. For GST affinity pull-down, the cleared cell lysate was loaded onto a gravity column with 1 mL bed volume of Glutathione agarose resin (Gold Biotechnology, #G-250-5). The resin was extensively washed, and bound proteins were eluted with a GST elution buffer (50 mM Tris-HCl pH 8.0, 150 mM NaCl, and 30 mM L-Glutathione reduced) in 1 mL fractions. For nickel affinity pull-down, the cleared cell lysate was loaded onto a gravity column with 0.5 mL bed volume of Ni-NTA resin (Qiagen, #30210). The resin was extensively washed, and bound proteins were eluted with a Ni-NTA elution buffer (50 mM Tris-HCl pH 8.0, 150 mM NaCl, and 300 mM imidazole) in 0.5 mL fractions. All elution fractions were analyzed by SDS-PAGE.

Flow Cytometry.

HeLa cells were maintained in Dulbecco's Modified Eagle Medium supplemented with 10% FB Essence (Seradigm, #3100-500) and penicillin-streptomycin (MilliporeSigma, #P4333). To stain surface TfR, HeLa cells were washed three times with the KRH buffer (121 mM NaCl, 4.9 mM KCl, 1.2 mM MgSO₄, 0.33 mM CaCl₂, and 12 mM HEPES, pH 7.0). Cells were then chilled on ice and stained with monoclonal anti-TfR antibodies (DSHB, #G1/221/12) at a final concentration of 0.1 μg/mL and APC-conjugated secondary antibodies (Thermo Fisher Scientific, #17-4015-82) at a final concentration of 0.8 μg/mL. After dissociation from plates using Accutase (Innovative Cell Technologies, #AT 104), APC fluorescence of the cells was measured on a CyAn ADP analyzer (Beckman Coulter). Mean APC fluorescence of mutant cells was normalized to that of WT cells. Data from populations of ~5,000 cells were analyzed using the FlowJo software (FlowJo, LLC, v10) based on experiments run in biological triplicates.

Immunoblotting.

To detect proteins in whole cell lysates, cells grown in 24-well plates were lysed in the SDS protein buffer. Protein samples were resolved on 8% Bis-Tris SDS-PAGE, and proteins were detected using primary antibodies and horseradish peroxidase-conjugated secondary antibodies. Primary antibody used in this work included monoclonal anti-FLAG antibodies (MilliporeSigma, #F1804) at a final concentration of 1 μg/mL, polyclonal anti-AP1γ antibodies (Bethyl, #A304-771A) at a final concentration of 1 μg/mL, polyclonal anti-AP2α antibodies (BD Biosciences, #610502) at a final concentration of 1 μg/mL, polyclonal anti-AP1σ antibodies (Bethyl, #A305-396A) at a final concentration of 1 μg/mL, and monoclonal anti-α-tubulin antibodies (DSHB, #12G10) at a final concentration of 43 ng/mL. To detect GST-AP1γ and GST-AP2α levels in bacterial co-expression experiments, Escherichia coli whole cell lysates and supernatants were mixed with SDS protein buffer. The primary antibody used to blot GST-AP1γ was GST tag monoclonal antibody (8-326) (ThermoFisher, #MA4-004). The primary antibody for GST-AP2α was the same for cellular studies.

Immunostaining and Imaging.

Cells grown in a four-compartment 35-mm glass-bottom dish were washed three times with the KRH buffer and fixed using 2% paraformaldehyde. Cells were permeabilized by 0.1% Triton-X100 in PBS. Surface TfR proteins were stained using monoclonal anti-TfR antibodies (DSHB, #G1/221/12) at a final concentration of 0.1 μg/mL and Alexa Fluor 568-conjugated secondary antibodies (Thermo Fisher Scientific, #A11004) at a final concentration of 1 μg/mL. The nuclei were stained with Hoechst 33342 (Thermo Fisher Scientific, #H3570) at a final concentration of 10 μg/mL. Images were captured using a 100× oil immersion objective on a Nikon A1 Laser Scanning confocal microscope and processed using FIJI software (40).

Real-Time Quantitative Reverse Transcription PCR.

Total RNAs were isolated using the rNeasy Mini Kit (Qiagen, #74104), followed by treatment with ezDNAse (Thermo Fisher Scientific, #18091150). First strand complementary DNA synthesis was performed using a SuperScript IV kit (Thermo Fisher Scientific, #18091050). Gene expression was determined by quantitative reverse transcription PCR on Applied Biosystems™ 7500 Fast Real-Time PCR Detection System using SsoAdvanced Universal SYBR Green Supermix (Bio-Rad, #172-5272) with gene-specific primer sets. The cycle threshold values of a candidate gene were normalized to those of GAPDH, a reference gene, and the Δcycle threshold values were calculated. The results were plotted as fold changes relative to the WT AAGAB rescue sample. PCR primers for AAGAB and mutants were as follows: 5′-TGACGATGACAAGCTTATGGCT-3′ (forward) and 5′-CGGAAAATACTGAGGAGCAGC-3′ (reverse). PCR primers for GAPDH were as follows: 5′-GACAGTCAGCCGCATCTTCT-3′ (forward) and 5′-GCGCCCAATACGACCAAATC-3′ (reverse).

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

Coordinates and structural factors for AAGAB CTD have been deposited in the Protein Data Bank with the accession code 7TWD (41).