Noncanonical and reversible cysteine ubiquitination prevents the overubiquitination of PEX5 at the peroxisomal membrane

Abstract

PEX5, the peroxisomal protein shuttling receptor, binds newly synthesized proteins in the cytosol and transports them to the organelle. During its stay at the peroxisomal protein translocon, PEX5 is monoubiquitinated at its cysteine 11 residue, a mandatory modification for its subsequent ATP-dependent extraction back into the cytosol. The reason why a cysteine and not a lysine residue is the ubiquitin acceptor is unknown. Using an established rat liver-based cell-free in vitro system, we found that, in contrast to wild-type PEX5, a PEX5 protein possessing a lysine at position 11 is polyubiquitinated at the peroxisomal membrane, a modification that negatively interferes with the extraction process. Wild-type PEX5 cannot retain a polyubiquitin chain because ubiquitination at cysteine 11 is a reversible reaction, with the E2-mediated deubiquitination step presenting faster kinetics than PEX5 polyubiquitination. We propose that the reversible nonconventional ubiquitination of PEX5 ensures that neither the peroxisomal protein translocon becomes obstructed with polyubiquitinated PEX5 nor is PEX5 targeted for proteasomal degradation.

Fig 1. PEX5(C11K) is polyubiquitinated at the DTM when the REM is blocked.

Radiolabeled PEX5(C11K) was subjected to standard (left panel) or E1/E2-fortified (right panel) single-step in vitro assays in the presence of either ATP or ADPNP, and in the absence or presence of NDPEX14 (lanes “-” and “+NDPEX14,” respectively). After treatment with NEM, reactions were centrifuged to isolate organelle (lanes “P”) and soluble (lanes “S”) proteins. Samples were analyzed by reducing SDS-PAGE/autoradiography. Note that recombinant NDPEX14, in addition to interacting strongly with soluble PEX5, also has some tendency to adsorb to membrane lipids [49]. This is probably the reason why more PEX5 is found in the organelle fractions in assays containing NDPEX14. Lanes I, ³⁵S-PEX5(C11K) protein used in the assays; P5(C11K), Ub-P5(C11K) and oligo/polyUb-P5(C11K) indicate unmodified, mono-, and oligo/polyubiquitinated PEX5(C11K), respectively; “o”, PEX5(C11K) proteins of unknown identity (ubiquitinated?) that are occasionally detected in supernatants upon long exposures—their detection in the presence of ADPNP and/or recombinant NDPEX14 indicates that they are not engaged in the peroxisomal protein import pathway; numbers to the left indicate the molecular weight markers in kDa; autoradiographs (“Autorad.”) and Coomassie-stained dried gels (“Coomassie”) are shown. DTM, docking/translocation module; NEM, N-ethylmaleimide; REM, receptor export module.

Fig 2. DTM-embedded PEX5(C11K) is polyubiquitinated at residue 11 yielding export-incompetent species.

(A) ³⁵S-His-PEX5(C11K) and ³⁵S-His-PEX5(C11A) were subjected to E1/E2-fortified single-step in vitro assays in the presence of ADPNP. Organelles (lanes “P”) were isolated by centrifugation and analyzed by reducing SDS-PAGE/autoradiography. (B) PEX5(C11K) was subjected to an E1/E2-fortified two-step in vitro assay. After the first step, organelles (lane “P”) were isolated by centrifugation, resuspended in import buffer, and incubated (second step) in the presence of either ATP or ADPNP for 10 min. Organelle and soluble fractions (lanes “PE” and “SE,” respectively) were isolated and analyzed by reducing SDS-PAGE/autoradiography. (C) Export kinetics of ubiquitinated PEX5(C11K) species. PEX5(C11K)-containing organelles isolated from an E1/E2-fortified assay were resuspended in import buffer (lane “P”) and incubated at 37°C in the presence of ATP. Aliquots were withdrawn at the indicated time points and centrifuged to separate organelle (lanes “PE”) from soluble proteins (lanes “SE”). Samples were analyzed by reducing SDS-PAGE/autoradiography. Lanes I, radiolabeled PEX5 proteins used in the assays; H₆P5(C11K) and H₆P5(C11A) indicate unmodified His-tagged PEX5(C11K) and His-tagged PEX5(C11A), respectively; Ub- and oligo/polyUb-H₆P5(C11K) indicates mono- and oligo/polyubiquitinated His-tagged PEX5(C11K); P5(C11K), Ub-P5(C11K), oligoUb-P5(C11K), and polyUb-P5(C11K) indicate unmodified, mono-, oligo-, and polyubiquitinated PEX5(C11K), respectively; numbers to the left indicate the molecular weight markers in kDa; autoradiographs (“Autorad.”) and Coomassie-stained dried gels (“Coomassie”) are shown. DTM, docking/translocation module.

Fig 3. Wild-type DTM-embedded PEX5 is not polyubiquitinated when the REM is blocked.

Radiolabeled PEX5 and PEX5(C11K) were incubated with a PNS under standard conditions in the presence of ADPNP for 20 min. After adding recombinant NDPEX14 and incubating for 2 min (to stop further insertion of PEX5 proteins into the DTM), pre-charged E1/E2 were then added to the PNSs and incubation proceeded at 37°C. Aliquots were removed at the indicated time points, treated with NEM, and organelles were isolated by centrifugation. Samples were analyzed by nonreducing SDS-PAGE/autoradiography. Lanes I, ³⁵S-PEX5 proteins used in the assays; P5 and Ub-P5, and P5(C11K), Ub-P5(C11K), oligoUb-P5(C11K), and polyUb-P5(C11K) indicate unmodified and monoubiquitinated PEX5, and unmodified, mono-, oligo-, and polyubiquitinated PEX5(C11K), respectively; numbers to the left indicate the molecular weight markers in kDa; long- and short-exposure autoradiographs (“Autorad.”) and the Coomassie-stained dried gel (“Coomassie”) are shown. Lower panel, the graph shows the amount of monoubiquitinated PEX5(C11K) (percentage of Ub-PEX5(C11K)) in organelles over time; a one phase exponential decay model was fitted to the data (averages and standard deviations of 4 experiments performed on different days with 2 independent PNSs are presented; t_1/2, half-life). The data underlying the graph shown in the figure can be found in S1 Data. DTM, docking/translocation module; NEM, N-ethylmaleimide; PNS, post-nuclear supernatant; REM, receptor export module.

Fig 4. The side-chain length of residue 11 of PEX5 is not important for its polyubiquitination.

(A) Schematic representation of the ubiquitin acceptor chemical group (shown in red) in PEX5(C11K) (P5(C11K)), wild-type PEX5 (P5), and PEX5(11–324;C11A,K0) (P5(11–324;C11A,K0)) proteins. N-ter and C-ter, N- and C-termini, respectively. (B) Radiolabeled PEX5(11–324;C11A,K0) was incubated with a PNS in the presence of ADPNP for 20 min at 37°C, either in the absence or presence of NDPEX14 (lanes “-” and “+NDPEX14,” respectively). After this incubation, an additional amount of NDPEX14 was added to both reactions, and aliquots were removed for analysis (lanes “0’”). Pre-charged recombinant E1/E2 were then added to the assays and incubation at 37°C proceeded for 5 min (lanes “5’”). Organelles were isolated by centrifugation and analyzed by SDS-PAGE/autoradiography. Lane I, ³⁵S-PEX5 protein used in the assays; P5(11–324;C11A,K0), Ub-P5(11–324;C11A,K0), oligoUb-P5(11–324;C11A,K0), and polyUb-P5(11–324;C11A,K0), indicate unmodified, mono-, oligo-, and polyubiquitinated PEX5(11–324;C11A,K0), respectively; numbers to the left indicate the molecular weight markers in kDa; autoradiographs (“Autorad.”) and the Ponceau S-stained membrane are shown (“Ponc. S”). PNS, post-nuclear supernatant.

Fig 5. Ubiquitination of PEX5 at the conserved cysteine residue is a reversible reaction.

(A) ³⁵S-labeled His-PEX5(1–324) was subjected to an in vitro ubiquitin exchange assay comprising 2 steps. In the first step, the radiolabeled protein was incubated with a PNS for 15 min at 37°C in the presence of ADPNP and 2 μM of bovine ubiquitin. After inhibiting further insertion of the PEX5 protein with recombinant PEX5(1–324) (rDC1), an aliquot was withdrawn (lane 2, “0’”). In the second step, one fraction of the reaction was diluted 10-fold in buffer containing recombinant E1/E2 pre-charged with His-Ub and aliquots were withdrawn at the indicated time points (lanes 3–6, 9). Another fraction received instead E1/E2(C85A) pre-incubated with His-Ub (lane 7). An organelle pellet containing radiolabeled His-PEX5(1–324) monoubiquitinated with His-Ub instead of Ub was also loaded (lane 8, “P_H6Ub”) to show the electrophoretic migration of the corresponding conjugate. Organelle (“P”) and soluble (“S”) fractions (lanes 2–8 and lane 9, respectively) were analyzed. (B) Left panel, E2-mediated deubiquitination of DTM-embedded Ub-PEX5. Radiolabeled PEX5 was incubated with a PNS in the presence of ADPNP. After stopping further insertion/ubiquitination of PEX5 with recombinant PEX5(1–324) (rDC1) and EDTA, the PNS was diluted 10-fold in buffer containing EDTA plus 2 μM of either E2 or E2(C85A). Aliquots were withdrawn at the indicated time points and centrifuged to obtain organelle (“P”; lanes 2–6 and 9–13) and soluble (S_10’) fractions. Right panel, the graph shows the amount of monoubiquitinated PEX5 (percentage of Ub-PEX5) in organelles over time; a one phase exponential decay model was fitted to the data (averages and standard deviations of 3 experiments performed on different days with the same PNS are presented; t_1/2, half-life). The data underlying the graph shown in the figure can be found in S1 Data. (C) E2-mediated deubiquitination of PEX5 modified with a di-ubiquitin(K48) chain. As in (B), with the exception that 5 μM of di-Ub(K48) instead of Ub was used in the first step of the assay. Lanes I, ³⁵S-PEX5 protein used in the assays; H₆P5(1–324), Ub-H₆P5(1–324), and H₆Ub-H₆P5(1–324), indicate unmodified his-tagged PEX5(1–324) and his-tagged PEX5(1–324) modified with ubiquitin and His-Ub, respectively; P5, Ub-P5, and diUb_(K48)-P5, indicate unmodified PEX5 and PEX5 modified with Ub or di-Ub(K48), respectively; samples were analyzed by nonreducing SDS-PAGE/autoradiography; numbers to the left indicate the molecular weight markers in kDa; autoradiographs (“Autorad.”), the Coomassie-stained dried gels (“Coomassie”), and the Ponceau S-stained membrane (“Ponc. S”) are shown. PNS, post-nuclear supernatant.

Fig 6. Reversibility of PEX5 ubiquitination avoids overubiquitination of PEX5 at the DTM.

Left panel—Cytosolic cargo-loaded PEX5 (PEX5_cyt) gets inserted into the DTM thus delivering its cargo into the organelle matrix. After cargo translocation, DTM-embedded PEX5 is monoubiquitinated at a conserved cysteine residue (Ub-PEX5_DTM) and is then extracted from the DTM by the REM. In the cytosol, Ub-PEX5_cyt is deubiquitinated either enzymatically by a deubiquitinase (DUB) or via a thiol-thioester exchange with glutathione (GSH) [34], and a new protein transport cycle begins. When Ub-PEX5_DTM is not promptly extracted, it may slowly undergo additional ubiquitination yielding oligo/polyubiquitinated species (Ub_n-PEX5_DTM). However, the oligo/polyubiquitin chains built on cysteine 11 of PEX5 are rapidly transferred to uncharged E2, thus generating non-ubiquitinated PEX5_DTM. According to this model, the steady-state amounts of these oligo/polyubiquitinated PEX5_DTM species (light gray forms) are extremely low. Right panel—PEX5(C11K) behaves similarly to wild-type PEX5 with the exception that ubiquitination at its lysine 11 residue is an irreversible reaction. Thus, any oligo/polyubiquitin adventitiously built on lysine 11 cannot be transferred to an uncharged E2. Oligo/polyubiquitinated PEX5(C11K) species are poor REM substrates leading to DTM clogging. Also, oligo/polyubiquitinated PEX5(C11K) may undergo proteasomal degradation. Cyt, cytosol; DUB, deubiquitinase; DTM, docking/translocation module; E2, ubiquitin-conjugating enzyme; GSH, glutathione; mat, matrix; PEX5/PEX5(C11K)_cyt, cytosolic PEX5 or PEX5(C11K), respectively; PEX5/PEX5(C11K)_DTM, DTM-embedded PEX5 or PEX5(C11K), respectively; REM, receptor export module; Ub, ubiquitin; Ub-/Ub₂-/Ub_n-PEX5/PEX5(C11K)_DTM, DTM-embedded mono-/oligo-/polyubiquitinated PEX5 or PEX5(C11K), respectively; Ub-PEX5/PEX5(C11K)_cyt, cytosolic monoubiquitinated PEX5 or PEX5(C11K), respectively.