A dual role of the conserved PEX19 helix in safeguarding peroxisomal membrane proteins

Summary Accurate localization of membrane proteins is essential for proper cellular functioning and the integrity of cellular membranes. Post-translational targeting of peroxisomal membrane proteins (PMPs) is mediated by the cytosolic chaperone PEX19 and its membrane receptor PEX3. However, the molecular mechanisms underlying PMP targeting are poorly understood. Here, using biochemical and mass spectrometry analysis, we find that a conserved PEX19 helix, αd, is critical to prevent improper exposure of the PEX26 transmembrane domain (TMD) to cytosolic chaperones. Furthermore, the αd helix of PEX19 interacts with the cytosolic domain of the PEX3 receptor, thereby triggering PEX26 release at the correct destination membrane. The peroxisome-deficient PEX3-G138E mutant completely abolishes this secondary interaction, leading to lack of PEX3-induced PEX26 release from PEX19. These findings elucidate a dual molecular mechanism that is essential to membrane protein protection and destination-specific release by a molecular chaperone.


INTRODUCTION
Membrane proteins represent $30% of the human proteome and play essential roles in the regulation of cellular functions, including transport of molecules, signal transduction, enzyme activity, and inter-organelle communication. 1 Due to these roles in various physiological processes, proper membrane protein targeting to specific cellular compartments is crucial for overall cellular function and membrane integrity.3][4] Membrane protein targeting to peroxisomes is critical for the formation of the appropriate membrane composition required for the biogenesis and function of mature peroxisomes. 2,5,6Indeed, aberrant membrane protein localization disrupts peroxisome biogenesis, eventually leading to impaired metabolism and developmental abnormalities. 2,51][12][13][14] PEX19 directly captures newly synthesized PMPs in the cytosol, thus preventing PMP aggregation. 15,16PMP-bound PEX19 is recruited to the peroxisomal membrane by the cytosolic domain of PEX3, which interacts with an N-terminal PEX19 domain, the aa helix. 11,12,17In addition, farnesylation of the C-terminal cysteine residue in the PEX19-CaaX motif increases the binding affinity of a PMP-derived peptide. 10Intriguingly, a previous study showed that mutation of four hydrophobic residues to alanine in the amphipathic ad helix located at the NTD of Neurospora PEX19 (PEX19-ad4A) impaired Neurospora PEX26 targeting to the peroxisome.However, the PEX19-ad helix is dispensable for chaperone activity and overall PEX3 binding. 15Thus, how the ad helix of PEX19 regulates PMP targeting and whether the NTD of human PEX19 has additional roles in PMP targeting beyond PEX3 interaction remain enigmatic.
9][20] TAs contain a single transmembrane domain (TMD) near the C-terminus. 21Due to this topology, TMDs of TAs are captured by cytosolic chaperones and targeted to their destined membrane post-translationally. 18,22,23During the post-translational targeting process, hydrophobic TMDs of TAs are at risk of promiscuous interactions with off-pathway chaperones.5][26] PEX26, a peroxisomal TA involved in the import and export of peroxisome matrix proteins, 27,28 is known to be inserted into the peroxisome in a PEX19-dependent manner but is not bound by the ER TA chaperone, Get3/TRC40. 15,16,29However, how PEX19 protects TAs from such off-pathway chaperones is currently unknown.
Here, we use biochemical and in vitro PMP import assays to address how the peroxisomal TA PEX26 is properly targeted.Our results show that the ad helix at the PEX19-NTD prevents PEX26 loss to other chaperones.Furthermore, we find that the cytosolic domain of PEX3 interacts Figure 1.The ad helix of PEX19 is crucial for protecting the TMD of PEX26 from other chaperones (A) The AlphaFold-predicted structure of human PEX19. 30,31Four conserved hydrophobic residues (F125, L129, L133, L136) located within the ad helix of PEX19 15 were highlighted in red (AF-P40855-F1).(B-E) Time courses of PEX26 (1.5 mM) aggregation in the presence of indicated concentrations of PEX19-WT (B), farnesylated PEX19-WT F (C), PEX19-ad4A (D), and farnesylated PEX19-ad4A F (E).After mixing PEX26 with PEX19 proteins for 15 s, the real-time measurement of the optical density of PEX26 at 360 nm was conducted using a UV spectrophotometer.(F and G) PEX19 variants solubilize PEX26 in a dose-dependent manner.The data in (F) were fit to Equation 1, and the K soluble values of PEX19 variants were summarized in (G).
with the ad helix of PEX19, a secondary PEX3-binding site, thus activating PEX26 release from PEX19.The peroxisome-deficient mutant PEX3-G138E completely abolishes both the interaction with the ad helix of PEX19 and PEX3-activated PEX26 dissociation.Therefore, our study reveals that the PEX19-ad helix plays a dual role by shielding peroxisomal TAs from off-pathway chaperones and cooperating with PEX3 to promote TA release to the peroxisomal membrane.

RESULTS
The ad helix of PEX19 prevents PEX26 loss to other chaperones The intrinsically disordered NTD of PEX19 contains several helices 15 (Figure 1A).As previously shown, 15 the hydrophobic residues of both the aa helix, a known PEX3-binding site, and the ad helix are highly conserved (Figure S1).To dissect which molecular steps in peroxisomal TA targeting require the conserved ad helix of human PEX19, we established a turbidity assay that monitors the chaperone activities of PEX19 proteins (Figures 1B-1G).We performed an in vitro farnesylation reaction using the yeast farnesyltransferases, Ram1 and Ram2 10 (Figures S2A  and S2B), and then purified the farnesylated PEX19-WT (denoted as PEX19-WT F ) and PEX19-ad4A (PEX19-ad4A F ; F125A, L129A, L133A, L136A) proteins (Figure 1A).The affinity-purified farnesylated PEX19 proteins exhibit increased electrophoretic mobility in SDS-PAGE compared with non-farnesylated PEX19 10 (Figures S2B-S2D).Irrespective of the farnesylation state, PEX19-WT and PEX19-ad4A had similar sensitivity to limited trypsin digestion (Figures S2E and S2F), suggesting that the ad4A mutation does not cause large structural alteration on the PEX19 protein.As a model PMP, we fused the N-terminal 23Strep-tagged SUMO domain to the PEX26 targeting sequences (237-305 aa) encompassing the TMD and C-terminal charged tail of PEX26. 16The non-cleavable, soluble SUMO domain was previously used to purify detergent-solubilized ER and mitochondrial TAs. 26,32,33We purified the recombinant PEX26 protein solubilized in 0.05% LDAO (N,Ndimethyl-1-dodecanamine-N-oxide) (Figure S2G).
To test whether the ad helix of PEX19 is crucial for the dissociation of PMPs after PMPs are loaded onto PEX19, we monitored the dissociation of PEX26 from PEX19 using the UV crosslinking assay (Figure 1H).We site-specifically incorporated p-benzoyl-l-phenylalanine (Bpa) into the TMD of PEX26 at the L264 residue using amber suppression 34 (Figure S3A), then purified the PEX19-PEX26 Bpa complex (Figures S3B-S3D).35,36 Adding excess CaM to the purified PEX19-PEX26 Bpa complex led to the release of PEX26 Bpa from the complex and its binding to CaM, which facilitated UV-activated Bpa crosslinking of PEX26.This approach allowed us to monitor the direct interaction of PEX26 with two distinct chaperones, PEX19 and CaM (Figures 1H-1J).To avoid the potential effects arising from non-specific interactions of the hydrophobic PEX26 Bpa with plastic surfaces in the tested samples, we focused exclusively on the relative distribution of crosslinked PEX26xPEX19 and PEX26xCaM bands within each sample.
The release of PEX26 from farnesylated PEX19-WT F occurred at a slower rate compared with non-farnesylated PEX19-WT (Figures 1I and  1K, blue vs. black).Surprisingly, irrespective of the farnesylation state of PEX19, PEX26 rapidly dissociated from PEX19-ad4A.In the presence of CaM, 70% of PEX26 was released from PEX19-ad4A within 10 min, whereas only 16% of PEX26 dissociated from PEX19-WT in the same period (Figures 1J and 1K, black vs. magenta).This difference in dissociation kinetics was also observed in the presence of another cytosolic chaperone, SGTA, known to bind to ER TAs 25 (Figures S3E and S3F).In the presence of SGTA, approximately 50% of PEX26 was released from PEX19-ad4A within 20 min.In contrast, PEX26 was scarcely released from PEX19-WT even after 80 min (Figures S3E and S3F).Furthermore, the observed dissociation rates of PEX26 from PEX19-ad4A were strongly accelerated by increasing the CaM concentrations, whereas PEX26 release from PEX19-WT appeared to remain similar (Figures S4A and S4B vs. S4C-S4E).Consistent with a previous study, 24,26 this linear dependency of the chase concentrations on the observed PEX26 dissociation rates for the PEX19-ad4A-PEX26 complex indicates that CaM actively invades and releases PEX26 from this complex (Figures S4C-S4E).In contrast, the unaltered dissociation kinetics of PEX26 in the presence of two different concentrations of CaM suggests that CaM is more likely to function as a passive trap, binding and sequestering released PEX26 from PEX19-WT (Figures S4A and S4B).Taken together, our PEX26 Bpa crosslinking data suggest that the ad helix of PEX19 is crucial in preventing the promiscuous PEX26 handoff from PEX19 to other cytosolic chaperones.Continued (H) Schematic representation of the Bpa crosslinking assay used to monitor PEX26 Bpa dissociation from PEX19 in the presence of a chaser chaperone, CaM; 20 mM CaM was mixed with PEX19$PEX26 Bpa to initiate PEX26 release.The zero time point samples were collected before the addition of CaM.At indicated times (t), aliquots (15 mL) of the reaction were frozen and analyzed by UV crosslinking at À20 C. (I and J) A representative western blot image of PEX26 Bpa dissociation from PEX19-WT (I) and PEX19-ad4A (J) proteins.The total volume (15 mL) of each sample at the indicated time was loaded onto 8% Tricine gels, and the non-crosslinked and crosslinked PEX26 proteins were probed with a Strep antibody.(K) Quantification of PEX26 dissociation from the Bpa crosslinking data in (I) and (J) and their replicates.PEX26 dissociation (%) was calculated as described in STAR methods.All values in (F), (G), and (K) are reported as mean G SD, with n = 3. Error bars are shown but may not be visible in some cases.

The cytosolic domain of PEX3 activates PEX26 release from PEX19
The cytosolic domain of PEX3 interacts with the aa helix of PEX19, thereby facilitating the recruitment of PEX19 to the peroxisomal membrane. 11,12Furthermore, a previous study showed that two hydrophobic residues (L209 and L261) at the base of the Neurospora crassa PEX3-a2 and a3 helices are required for PEX26 insertion into the peroxisome. 15This study suggested that the cytosolic domain of PEX3 plays an additional role in PMP targeting beyond its interaction with the aa helix of PEX19.Using the PEX26 Bpa crosslinking assay established in Figure 1H, we measured PEX26 release from PEX19 in the presence of the cytosolic domains of PEX3-WT and the PEX3-G138E mutant, which is incapable of forming peroxisomes in peroxisome-deficient CHO cells (ZPG208). 37,38The cytosolic domain of PEX3-WT (PEX3DN-WT, 41-373aa), lacking its N-terminal TMD, induced significantly greater PEX26 release from both non-farnesylated PEX19 and farnesylated PEX19 (Figures 2A-2D, black vs. blue).In contrast, the peroxisome-deficient PEX3 mutant, PEX3DN-G138E, 38 was not able to activate PEX26 release from PEX19-WT proteins (Figures 2A-2D, black vs. red).Furthermore, in the case of both the PEX19-ad4A-PEX26 and PEX19-ad4A F -PEX26 complexes, PEX3DN-WT did not alter PEX26 dissociation from PEX19-ad4A proteins (Figures 2E-2H).Together, these data suggest that the cytosolic domain of PEX3 triggers PEX26 release from PEX19, possibly a prerequisite for PMP insertion into the peroxisomal membrane.

The ad helix of PEX19 directly interacts with the cytosolic domain of PEX3
We next considered why PEX19-ad4A and PEX3DN-G138E mutants cannot support PEX3-induced PEX26 release.The simplest explanation is that these mutants disrupt intermolecular interactions between PEX19 and PEX3.We tested this possibility by performing His 6 -PEX19 pulldown experiments at the same concentrations of PEX19 and PEX3 used in the PEX26 Bpa crosslinking experiments (Figure 2).We found that half of the peroxisome-deficient PEX3DN-G138E mutant binds PEX19-WT (Figures S5A and S5B).As expected, only $17% of PEX3DN-W104A, a PEX19-binding deficient mutant, 11,39 was copurified with PEX19-WT (Figures S5A and S5B).In contrast, PEX3DN-WT bound the mutated PEX19-ad4A protein at levels comparable to PEX19-WT (Figures S5A and S5B).Therefore, although PEX19-ad4A lacks the PEX3induced PMP release, it does not account for overall binding to the PEX3 receptor.
A previous study found that the binding affinity of PEX19-aa peptide to the cytosolic domain of PEX3-WT (K D = 40.8nM) is 12-fold lower than that of the full-length PEX19 (K D = 3.4 nM), 11 suggesting that PEX19 contains a second PEX3-binding site that weakly contributes to the overall binding of PEX19 to PEX3 (Figure 3A).We hypothesized that the ad helix of PEX19 provides a secondary PEX3-binding site that induces PEX26 release from PEX19.To test this hypothesis, we genetically incorporated Bpa at the four different hydrophobic residues (F125, L129, L133, L136) located at the ad helix of PEX19 (Figure 1A) and monitored the interaction between PEX19 and PEX3DN-WT using the PEX19-F125 Bpa crosslinking assay (Figures 3B, S5C, and S5D).Of those four residues, only the F125 Bpa residue of PEX19 forms a $115 kDa crosslinked protein with PEX3DN-WT (Figure S5C).The intermolecular crosslink between PEX19-F125 Bpa and PEX3DN-WT formed even after pre-incubation with PEX26, indicating that PEX3DN-WT directly interacts with the ad helix of PEX19 (Figures 3C and 3D).In contrast, no PEX19-F125 Bpa -mediated crosslink was formed in the presence of PEX3DN-G138E, suggesting that PEX3DN-G138E does not interact with the PEX19-ad helix (Figures 3C and 3D).Given that PEX19-ad4A and PEX19-WT bind to PEX3DN-WT equally well (Figures S5A and  S5B), the ad4A mutation in PEX19 likely only disrupts the secondary interaction between the ad helix of PEX19 and PEX3DN-WT.In this case, the primary interaction of PEX19-aa helix with PEX3DN-WT would remain intact.In contrast, PEX3DN-G138E interferes with both the primary interaction with PEX19-aa and the secondary interaction with PEX19-ad (Figures S5A and S5B).
Several experimental data suggest that the ad helix of PEX19 is unlikely to directly interact with PEX26.None of the four hydrophobic residues in the PEX19-ad helix showed a distinct crosslinked band with PEX26 in the absence of PEX3 (Figure S5D).In addition, PEX19-F125 Bpa did not crosslink with PEX26 in the absence or presence of PEX3DN variants (Figures 3C and 3E).In contrast, PEX19-M179 Bpa located in the known PMP-binding a1 helix showed a distinct PEX19-PEX26 crosslinked band (Figures S5E and S5F).Considering the data in Figures 1, S3, and S4, it becomes evident that, although the hydrophobic residues in the ad helix of PEX19 do not directly bind to PEX26, the ad helix of PEX19 could be located proximally to the TMD of PEX26, thereby shielding it from other chaperones.
To further identify the PEX19-F125-binding site on the cytosolic domain of PEX3, we utilized a Bpa-crosslinking-based mass spectrometry approach. 40,41The PEX19-F125 Bpa -PEX3DN-WT crosslink was in-gel digested with trypsin and AspN endoproteinases, and the cleaved peptides were further analyzed by LC-MS/MS (Figures S5G and S5H).Using a database search algorithm pLink 2, 42 we identified the most frequently detected Bpa crosslinked peptide (PEX3 residues T64-R77), which represents over 90% of the identified crosslinked peptides (Figure 3A, highlighted with green).These residues are located in the a1 helix of PEX3, distinct from the primary PEX19-aa-binding site (Figure 3A).
Because mutating all four hydrophobic amino acids in the ad helix of PEX19 to alanine (PEX19-ad4A) abolishes PEX3-activated PEX26 release (Figures 2E-2H), the hydrophobic residues at the a1 helix of PEX3 may interact with the PEX19-F125 residue.Consistent with this Figure 2. The cytosolic domain of PEX3-WT activates PEX26 dissociation from PEX19-WT, whereas PEX3-G138E does not (A-D) (A and C) Representative western blot images of PEX26 Bpa dissociation from PEX19-WT (A) and PEX19-WT F (C).In the absence and presence of 400 nM PEX3DN variants, PEX19-WT$PEX26 Bpa and PEX19-WT F $PEX26 Bpa complexes at the final concentrations of 400 nM PEX19 proteins were incubated with 20 mM CaM.The zero time point samples were collected before the addition of CaM to the PEX19$PEX26 Bpa .At the indicated time points, aliquots of the reactions were frozen and analyzed by UV crosslinking.PEX26, PEX3DN, and CaM were probed in immunoblots with Strep, Flag, and HA antibodies, respectively.(B and D) Quantification of PEX26 dissociation from the Bpa crosslinking data in (A) and (C) and their replicates.(E-H) (E and G) Representative western blot images of PEX26 Bpa dissociation from PEX19-ad4A (E) and PEX19-ad4A F (G). PEX26 Bpa dissociation assays from PEX19-ad4A (E) and PEX19-ad4A F (G) were conducted in the same manner as in (A) and (C).(F and H) Quantification of PEX26 dissociation from the Bpa crosslinking data in (E) and (G) and their replicates.All values in (B), (D), (F), and (H) are reported as mean G SD, with n = 3. Error bars are shown but may not be visible in some cases.
Figure 3.The ad helix of PEX19 interacts with the a1 of PEX3 (A) Crystal structure of PEX19-aa-bound PEX3DN (PDB 3MK4). 12The amino acid residues in the PEX19-F125 Bpa -crosslinked PEX3 peptide identified from MS analysis were highlighted in green.The identified PEX19-F125 Bpa -binding residue M72 in PEX3DN-a1 was shown in red.The known PEX19-aa-binding residue W104 and the disease-causing residue G138 (G138E) in PEX3DN-a2 were shown in purple and magenta, respectively.(B) A schematic representation of PEX19-F125 Bpa crosslinking to PEX3DN variants; 400 nM PEX19-F125 Bpa was incubated in the absence and presence of 100 nM PEX26 at room temperature for 1 min; 400 nM PEX3DN-WT and its variants were added to the reaction and further incubated at room temperature for 5 min.The frozen reaction samples were subjected to UV crosslinking.(C-E) Western blot analysis of PEX19-F125 Bpa crosslinking to PEX3DN variants.The crosslinked proteins were probed using antibodies against PEX19, Flag (PEX3DN), and Strep (PEX26).''*'' represents the SDS-resistant PEX26 dimers in (E).(F) The number of matched peptides for PEX19-F125 Bpa -PEX3 crosslink identified from MS analysis.The M72 residue in the PEX3 peptide ( 64 TCNMTVLSMLPTLR 77 ) was most frequently crosslinked with F125-Bpa in the PEX19 peptide ( 115 VGSDMTSQQEBpaTSCLK 130 ) (also see Figure S5H).(G) Western blot analysis of PEX19-F125 Bpa crosslinking to PEX3DN-WT, PEX3DN-M72K, and PEX3DN-M72E.The reactions were carried out in the same way as in (B).(H) Western blot analysis of PEX19-F125 Bpa crosslinking to the endogenous PEX3 in semi-intact HeLa cells; 200-nM PEX19-F125 Bpa was incubated with semipermeabilized HeLa cells at room temperature for 1 h, after which the cell lysates were subjected to UV crosslinking as described in STAR methods.hypothesis, mass spectrometry analysis showed that the hydrophobic M72 residue at the PEX3-a1 helix is the most frequently observed residue that crosslinks to PEX19-F125 Bpa (Figure 3F).Charge mutations of the M72 residue, PEX3DN-M72K and PEX3DN-M72E, failed to crosslink to the PEX19-F125 Bpa residue (Figure 3G), suggesting that the interaction between PEX3-a1 and PEX19-ad is dominated by inter-helical hydrophobic interactions.In contrast, the overall binding of PEX3DN-M72K to PEX19-WT is comparable with the wild-type protein, PEX3DN-WT (Figures S5A and S5B).
Next, we tested whether this secondary interaction between PEX19 and PEX3 occurs in cells.After washing out the cytosolic fraction (Figure S6A, steps I and II), the purified PEX19-F125 Bpa protein was incubated with the semi-permeabilized HeLa cells.The data of UV crosslinking experiment show that PEX19-F125 Bpa forms a $120 kDa crosslink to the full-length PEX3 protein in the cells (Figure 3H).This observation indicates that the membrane-embedded full-length PEX3 also interacts with the ad helix of PEX19.Collectively, our data suggest that the ad helix of PEX19 serves as a previously obscure secondary PEX3-binding site that mediates substrate dissociation once the TA reaches the peroxisome.
Incubation with the 23Strep-PEX19-WT$GFP-PEX26 complex for 1 h led to an 88.4% PEX26 colocalization rate with PMP70, a peroxisomal maker protein (Figures 4B and 4C).In contrast, the shorter incubation (10 min) with the 23Strep-PEX19-WT$GFP-PEX26 complex was insufficient for the stable targeting of PEX26 into peroxisome (Figures S6C-S6E, blue).Furthermore, when 23Strep-PEX19-Daa$GFP-PEX26 and 23Strep-PEX19-ad4A$GFP-PEX26 complexes were incubated with the semi-permeabilized cells for 1 h, PEX19-Daa and PEX19-ad4A strongly inhibited the PEX26 localization into the peroxisome (Figures 4B and 4C).Similarly, the mean intensities of peroxisome-localized PEX26 were also strongly reduced in both the PEX19-Daa and PEX19-ad4A mutants (Figures 4B and 4D).These different localization efficiencies of PEX26 to peroxisome are unlikely due to different amounts of PEX26 binding to the PEX19 variants (Figure 4A).Pre-incubation of 23Strep-PEX19-WT$GFP-PEX26 complex with excess PEX3DN-WT almost completely abolished the colocalization of PEX26 with PMP70, further supporting the essential role of PEX19-PEX3 interactions in the peroxisomal targeting of PMPs (Figures S6C-S6E).Together, these data highlight the dual role of the PEX19-ad helix in protecting PEX26 from other chaperones and interacting with the cytosolic domain of PEX3 to ensure the efficient localization of PEX26 into the peroxisome.

DISCUSSION
Using biochemical assays and mass spectrometry analysis, we revealed that the ad helix of PEX19 serves as a release modulator for PMPs at the peroxisome.PEX19 rapidly captures PEX26 upon its release from the ribosome (step 1), 15 and farnesylation of PEX19 at the C-terminus enhances PEX26 solubility.The ad helix of PEX19 is essential in preventing PEX26 loss to off-pathway chaperones.Removal of conserved hydrophobic residues in the ad helix of PEX19 (PEX19-ad4A) led to premature PEX26 release to off-pathway chaperones.PEX26-bound PEX19 is recruited to PEX3 via the primary PEX3 binding helix, PEX19-aa (step 2). 11,12The secondary PEX3-binding site, PEX19-ad, further interacts with the cytosolic domain of PEX3, thereby activating PEX26 release to the membrane (step 3).The peroxisome-deficient mutants PEX3-G138E and PEX3-M72K lack this secondary interaction with the ad of PEX19.Finally, PEX26 is inserted into the membrane via unknown mechanisms (step 4).
Despite the functional importance of PEX19 farnesylation in vivo, only a few previous studies showed that farnesylation of PEX19 enhances PMP-binding affinity. 10,43Using a turbidity assay that monitors TA aggregation, we showed that farnesylated PEX19-WT F captures and solubilizes PEX26 more efficiently than non-farnesylated PEX19-WT (Figures 1B-1G).The PEX26 solubilization constant for PEX19-WT F is about 2-fold lower than PEX19-WT (Figures 1F and 1G).This mild enhancement of PEX19 chaperone activity by farnesylation is consistent with a previous study of Neurospora PEX19, which qualitatively assessed chaperone activity using ultracentrifugation. 15 Furthermore, the required time to release $20% PEX26 from PEX19-WT and PEX19-WT F to the external chaperone, CaM, is 30$40 min and 240 min, respectively (Figures 1I-1K).Thus, farnesylation of PEX19 not only helps to capture PMPs more effectively but also prevents TA loss from PEX19 to other cytosolic chaperones.
Membrane protein chaperones appear to possess structural features that protect the hydrophobic TMDs of membrane proteins from offpathway chaperones in the crowded cytosolic environment.Our mechanistic dissections revealed that the ad helix of PEX19-NTD minimizes the loss of PEX26 from PEX19 to external chaperones, such as CaM and SGTA.In the presence of CaM, 60% to 80% of PEX26 was released from PEX19-ad4A proteins within 5 min, whereas only 20% to 30% of PEX26 dissociated from PEX19-WT proteins within 4 h (Figures 1 and 2).PEX19-ad4A lost over 40% of PEX26 to SGTA within 10 min, whereas only less than 4% of PEX26 was released from PEX19 within 80 min (Figures S3E and S3F).Previous studies showed that the a8 helix of the yeast cytosolic chaperone Get3 serves as a lid for the ER TA-binding groove, protecting ER TAs from other cytosolic chaperones. 24,44,45Similarly, although AlphaFold 30,31 predicts that the intrinsically disordered NTD of PEX19 forms an open conformation (Figure 1A), the ad helix may adopt a more compact conformation toward the a1 helix upon TA binding to the PEX19-CTD, thereby shielding the hydrophobic TMDs of peroxisomal TAs. 10,15,46Given the lack of structural information on the PEX19-NTD, this hypothesis would require future experimental testing.
The mechanism activating cargo release from TA-loaded PEX19 to the membrane remains enigmatic.Our results suggest that the secondary interaction between PEX19-ad and PEX3-a1 modulates the dissociation of PEX26 (Figures 2 and 3).Although the PEX19-ad4A mutant binds to the cytosolic domain of PEX3 (Figures S5A and S5B), it loses PEX3-activated PEX26 release, potentially due to lack of the secondary interaction between PEX19-ad and PEX3-a1.Furthermore, the PEX3-G138E and M72K mutants, 38,47 which lead to lack of peroxisomes, completely abolished this secondary interaction between PEX19-ad and PEX3-a1 (Figures 3C and 3G).Together with previous studies, 11,12 these data indicate that the secondary interaction of PEX19-ad with PEX3DN destabilizes the PEX19-TA complex prior to membrane insertion (Figure 2).However, high-affinity binding between PEX19 and PEX3DN is primarily mediated by the N-terminal PEX19-a1 helix. 11,12An analogous mechanism was previously observed in TA targeting to the ER membrane. 48In the guided entry of tail-anchored protein (GET) (E) A proposed model of PEX26 targeting to the peroxisome.PEX19 rapidly captures free PEX26 in the cytosol (step 1).As shown in (A), a majority of PEX19 proteins in the cytosol are farnesylated (asterisk).Farnesylated PEX19 displays an increased binding affinity to PEX26.The ad helix of PEX19 protects against PEX26 loss to off-pathway chaperones (step 2).The aa helix of PEX19 primarily interacts with the cytosolic domain of PEX3.The secondary interaction of PEX19-ad helix with PEX3 destabilizes the PEX19$PEX26 complex, thereby inducing the release of PEX29 to the membrane and further leading to its membrane insertion (steps 3-4).
0][51] However, a secondary interaction between the internal helices of Get2 and Get3 remodels the latter protein to optimize TA release. 48Therefore, multiple interactions between chaperones and their membrane receptor proteins may ensure efficient recruitment of the targeting complex as well as destination-specific release of membrane proteins to their target membranes.
In conclusion, our study revealed a dual role for PEX19-ad in shielding the TMD of TAs from other cytosolic chaperones and in interacting with PEX3 to trigger TA release prior to membrane insertion.In contrast to other chaperones that use ATP hydrolysis and cochaperones to regulate their conformations, PEX19 appears to employ a conserved helix, ad, at its intrinsically disordered NTD for TA targeting.This switchlike ad helix of PEX19 would minimize the exposure of the hydrophobic TMD to other proteins and only allow TA release upon binding to the cytosolic domain of PEX3.Given that PEX19 is known to deliver TA proteins to mitochondria, 19,52 we envision that PEX19 uses a similar mechanism for targeting of mitochondrial TA proteins.Guided by this study, future structural and single-molecule analyses could provide mechanistic insight into how an ATP-independent chaperone dynamically modulates its conformation for proper membrane protein targeting.

Limitations of the study
As a model substrate for PEX19, we used the peroxisomal TA, PEX26, which contains only one TMD near the C-terminus.Given that PEX19 is also known to regulate the peroxisomal targeting of multi-spanning PMPs, 7,9,17 it remains an open question whether the ad helix of PEX19 plays the same dual role in protecting multiple TMDs from other cytosolic chaperones and assisting PMP release to the peroxisomal membrane.
The PMP import assay used in this study contains a significant amount of cytosolic fractions.Although our results suggest that the ad helix of PEX19 is crucial for the peroxisomal localization of TAs, the assay would not be optimal to dissect two molecular steps: PEX26 dissociation before and after PEX19 binding to PEX3.Currently, there is no solid evidence indicating whether PEX19-PEX3 constitutes a minimal targeting complex for peroxisomal TAs or if an additional targeting factor is necessary.Once this information is revealed, biochemical reconstitution of the targeting complex into proteoliposomes will help further elucidate the molecular mechanism underlying the insertion of TAs into the peroxisome (steps 3-4 in Figure 4E).
To express 23Strep-SUMO-PEX26 Bpa , an amber codon (TAG) was introduced in the TMD-encoding sequence replacing the eighteenth amino acid (Leu264) ( 247 FFSLPFKKSLLAALILCLLVV 267 ).23Strep-SUMO-PEX26 Amb and tRNA CUA Opt synthetase 34 plasmids were co-transformed into BL21 Star (DE3) cells.The cells were grown to an optical density at 600 nm (OD 600 ) of 0.3, and the expression of tRNA CUA Opt synthetase was induced with 0.2% arabinose (Sigma).At OD 600 of 0.6, expression of 23Strep-SUMO-PEX26 Bpa was induced with 0.1 mM IPTG and 1 mM Bpa (Bachem) at 37 C for 1.5 h.His 6 -SUMO-PEX19 Bpa variants (F125 Bpa , L129 Bpa , L133 Bpa , L136 Bpa , and M179 Bpa ) were expressed with a similar method as 23Strep-SUMO-PEX26 Bpa , apart from the IPTG concentration, which was 0.5 mM.All proteins were purified in the same way as their non-Bpa proteins.

In vitro farnesylation reaction
In vitro farnesylation reaction was carried out as described previously, with a minor modification. 1030 mM His 6 -PEX19 was incubated with 700 nM farnesyl transferases, Ram1 and Ram2, and 65 mM farnesyl pyrophosphate (Sigma) in the reaction buffer (50 mM Tris-HCl (pH 8.0), 20 mM CH 3 COOK, 5 mM MgCl 2 , 0.1 mM ZnCl 2 , and 10 mM DTT) at room temperature for 1 h.The reaction mixture was diluted 10-fold with Buffer C and then incubated with Ni-NTA agarose resin for 5 min.The farnesylated PEX19 proteins were purified using a standard Ni-NTA purification method.
Limited proteolysis of PEX19 proteins 20 mM PEX19 was incubated with 1.5 mg/mL trypsin (Sigma) at 37 C.At the indicated time, 12 mL of samples were mixed with SDS loading buffer and heated at 95 C for 5 min.Subsequently, 5 mL of quenched samples were loaded onto 12.5% Tris-glycine gels.The gels were stained with SimplyBlue SafeStain (Invitrogen).

Turbidity assay
200 mM PEX26 stored in Buffer B containing 0.05% LDAO was rapidly diluted to a final concentration of 1.5 mM in the assay buffer (20 mM K-HEPES (pH 7.5), 150 mM CH 3 COOK, 1 mM DTT) containing various concentrations of PEX19 proteins.The optical density of PEX26 at 360 nm was measured in real time using a UV spectrophotometer (Optizen Alpha, KLAB).The observed solubility of PEX26 (S obsd ) was calculated from the % change of the OD 360 value at 5 min between PEX26 alone and in the PEX19-containing samples.The data were plotted as a function of PEX19 concentration and fit to Equation 1, S obsd = S Max 3 ½PEX19 K soluble +½PEX19 (Equation 1) in which S Max is the % soluble PEX26 at saturating PEX19 concentrations, and K soluble is the concentration of PEX19 required to reach half of S Max .

Figure 1 .
Figure1.Continued (H) Schematic representation of the Bpa crosslinking assay used to monitor PEX26 Bpa dissociation from PEX19 in the presence of a chaser chaperone, CaM; 20 mM CaM was mixed with PEX19$PEX26 Bpa to initiate PEX26 release.The zero time point samples were collected before the addition of CaM.At indicated times (t), aliquots (15 mL) of the reaction were frozen and analyzed by UV crosslinking at À20 C. (I and J) A representative western blot image of PEX26 Bpa dissociation from PEX19-WT (I) and PEX19-ad4A (J) proteins.The total volume (15 mL) of each sample at the indicated time was loaded onto 8% Tricine gels, and the non-crosslinked and crosslinked PEX26 proteins were probed with a Strep antibody.(K) Quantification of PEX26 dissociation from the Bpa crosslinking data in (I) and (J) and their replicates.PEX26 dissociation (%) was calculated as described in STAR methods.All values in (F), (G), and (K) are reported as mean G SD, with n = 3. Error bars are shown but may not be visible in some cases.

Figure 4 .
Figure 4. PEX19-ad4A significantly reduced the peroxisomal targeting of PEX26 (A) 23Strep-PEX19 pull-down experiments.23Strep-PEX19 variants (WT, D aa, and ad4A) and GFP-PEX26 were co-expressed in HEK293T cells, and the cytosolic fractions were subjected to 23Strep-PEX19 pull-down experiments.I and E denote input and elution, respectively.The I and E fractions were subjected to SDS-PAGE and western blotting with antibodies against Strep and GFP.(B) The 23Strep-PEX19$GFP-PEX26 complexes from the E fractions (A) were used for PEX26 targeting experiments in semi-permeabilized HeLa cells, as shown in Figure S4A.After incubating the 23Strep-PEX19$GFP-PEX26 complexes for 1 h, the cells were fixed and further analyzed by immunofluorescence (scale bar: 10 mm); 5.3-fold enlarged images of the boxed areas were shown separately in the right panel (scale bar: 2 mm).(C) Colocalization rates of PEX26 with the peroxisomal membrane protein (PMP70).A total of 130 cells from three biological replicates were analyzed using LAS X software.The lines indicate the mean values of the colocalization rate of PEX26 for each condition (n = 130).*p < 0.0001 (Student's t test).(D) Mean intensity of colocalized PEX26 with PMP70.The mean intensities of peroxisome-localized PEX26 in (C) were analyzed using LAS X software.Values are reported as mean G SD, with n = 3 (three biological replicates).Error bars are shown but may not be visible in some cases.(E)A proposed model of PEX26 targeting to the peroxisome.PEX19 rapidly captures free PEX26 in the cytosol (step 1).As shown in (A), a majority of PEX19 proteins in the cytosol are farnesylated (asterisk).Farnesylated PEX19 displays an increased binding affinity to PEX26.The ad helix of PEX19 protects against PEX26 loss to off-pathway chaperones (step 2).The aa helix of PEX19 primarily interacts with the cytosolic domain of PEX3.The secondary interaction of PEX19-ad helix with PEX3 destabilizes the PEX19$PEX26 complex, thereby inducing the release of PEX29 to the membrane and further leading to its membrane insertion (steps 3-4).