Misalignment of PLP/DM20 transmembrane domains determines protein misfolding in Pelizaeus-Merzbacher disease

Abstract

A large number of genetic diseases have been associated with truncated or misfolded membrane proteins trapped in the endoplasmic reticulum (ER). In the ER, they activate the unfolded protein response, which can trigger cell death. Hence, a better understanding of protein misfolding features might help in developing novel therapies. Here, we have studied the molecular basis of Pelizaeus-Merzbacher disease, a leukodystrophy defined by mutations of the PLP1 gene and ER retention of two encoded tetraspan myelin proteins, PLP and DM20. In mouse oligodendroglial cells, mutant isoforms of PLP/DM20 with fewer than all four transmembrane (TM) domains are fully ER retained. Surprisingly, a truncated PLP with only two N-terminal TM domains shows normal cell-surface expression when coexpressed with a second truncated PLP harboring the two C-terminal TM domains. This striking ability to properly self-align the TM domains is disease relevant, as shown for the smaller splice isoform DM20. Here, the increased length of TM domain 3 allows for compensation of the effect of several PLP1 point mutations that impose a conformational constraint onto the adjacent extracellular loop region. We conclude that an important determinant in the quality control of polytopic membrane proteins is the free alignment of their TM domains.

Figure 1.

Structure of PLP/DM20 and mutations associated with Pelizaeus–Merzbacher disease. A, Two-dimensional model of PLP (276 residues as black beads) and its splice isoform DM20, lacking 35 residues (marked in gray) from an intracellular loop. The orientation of TM1–TM4 positions of both the N and C termini into the cytoplasm. Disulfide bonds in EC2 are joined by black lines and are critical for PLP folding (Dhaunchak and Nave, 2007). Positions of amino acids in EC2 that are substituted in patients with PMD are marked in yellow. Those that have been studied in detail here carry the single letter code of the wild-type sequence, labeled in red. Also indicated are C-terminal epitope tags used in this study (EGFP or myc) and the approximate positions of extracellular (3F4) and intracellular (A431) antibody binding sites common to PLP and DM20. B, Schematic view of the intracellular loop of PLP/DM20, flanked by TM2 and TM3, including a PLP-specific sequence (open gray circles) and 5 aa (open red circles) that extend the hydrophobicity of DM20 (see Fig. 6 for hydropathy plot). The positions of four segments (S1–S4) are indicated, each of which has been replaced individually with the myc-epitope tag, giving rise to different PLP “chimeras” (S1–S4). These serial replacement chimeras were used to test whether PLP harbors an essential ER retention signal (Fig. 5). C, A two-dimensional model of PLP, DM20, and various truncations used in the study is shown. The four TMs (block rectangles), the PLP-specific region (gray rectangle), the previously proposed retention signals (stars), and the signal peptide for proper insertion of TM4 (blue rectangle) are highlighted.

Figure 2.

Truncated PLPs with fewer than four TMs are retained in the ER. A, When expressed, PLP^wt readily traffics to the cell surface as demonstrated by green fluorescent microspikes at the tips of processes. COS-7 and glial cells (Fig. 4A) protrude numerous filopodial processes, and PLP^wt accumulates in the endosomal/lysosomal compartment. Magnifications (Aa, Ab) of the boxed area in A highlight PLP-positive filopodial protrusions and endosomes/lysosomes. Scale bar, 10 μm. B, In contrast, mutant protein derived from jimpy-msd mice (PLP^msd; A242V substitution) fails to reach the cell surface. Cells lack any labeled microspikes (magnified in insets). There is a reticular distribution of EGFP fluorescence with no accumulation in the endosomal/lysosomal compartment. Magnifications (Ba, Bb) of boxed area in B are magnifications of areas in a close proximity to the cell surface. Note that there is complete absence of fine microspikes. Scale bar, 10 μm. C–F, Fixed cells permeabilized and stained for the C- or N-terminal myc epitope are shown in the left column and EGFP fluorescence is shown in the right column. Cells expressing truncated PLPs, regardless of whether they contain two or three TMs, are strongly retained in the ER (C, E). Interestingly, the second half of the protein whether consisting of TM3,4 or TM4 alone is also retained in the ER (D, F, G), exactly like PLP^msd. To facilitate proper insertion of TM4, with N terminus luminal and C terminus cytosolic, we added an MHC class I signal peptide to the N terminus (Kleijnen et al., 1997; Swanton et al., 2003) and stained methanol-fixed and digitonin-permeabilized cells (G).

Figure 3.

Self-assembly of TM domains. TM1,2 when coexpressed in oligodendroglial (A, B) or COS-7 cells (C, D) with EGFP-tagged TM3,4 derived from PLP^wt but not from PLP^msd results in surface expression of both halves of the protein (only TM3,4 is shown). In B, the endosomal/lysosomal accumulation of TM3,4 derived from PLP^wt was verified by colabeling with lamp-1 (red). Also, note that, in addition to endosomal/lysosomal accumulation, glial and COS-7 cells protrude microspikes, depicted in a higher magnification of the boxed area in B (Ba) that are absent in D. Scale bar, 10 μm.

Figure 4.

The function of extracellular disulfide bridges in DM20 folding. Of the two disulfide bridges in EC2, the “outer” one (and C¹⁶⁵–C¹⁸⁴) is dispensable for folding and cell-surface expression of PLP. To test the function of each cysteine bridge (see Fig. 1A) in DM20 folding and surface appearance, single and double cysteine-to-serine substitutions were engineered for each disulfide bridge. A–C, Replacing one or both cysteines of the membrane proximal bridge C¹⁴⁸–C¹⁹² led to severe misfolding, as visualized by ER retention with no surface labeling of DM20 (in red; stained live using 3F4 antibody; left panels). Surprisingly, replacing one or both cysteines of the outer bridge and C¹⁶⁵–C¹⁸⁴ did not interfere with cell-surface labeling of DM20 (in red). Thus, only the membrane proximal disulfide bridge is essential for normal folding of DM20. Only merged images are shown. The percentages of cells stained with 3F4 antibody (B) or showing endosomal/lysosomal accumulation (C) are depicted in histograms on the right. As expected, DM20 lacking the membrane proximal bridge shows no accumulation in the endosomal/lysosomal compartment. D, Wild-type (W) PLP, DM20, and PLP/DM20 lacking outer (O) or inner (I) cysteines were analyzed either under nonreducing or reducing (150 mm mercaptoethanol) conditions. M and D highlight PLP/DM20 monomers and dimers, respectively. As expected, in MO3.13 cells, the wild-type, and PLP and DM20 lacking the outer cysteines do not form cysteine-mediated cross-links. The short and long exposures show that both PLP and DM20 lacking the inner bridge form cysteine-mediated cross-links (highlighted with dashed red boxes). The faint SDS and mercaptoethanol-resistant dimer seen on reducing gels constitutes between 7 and 13% (quantified from two independent transfection experiments) of the total dimer seen on nonreducing gels. Such dimers are also seen with wild-type protein but only during longer exposures and were also reported previously (Swanton et al., 2005).

Figure 5.

ER retention in oligodendroglial cells distinguishes PMD-associated isoforms of PLP, DM20, and experimental chimeras. A, Wild-type PLP and DM20, here fused to a C-terminal EGFP, exit the ER and reach the cell membrane (a, d), as shown by confocal imaging revealing distal transport vesicles in glial processes and membrane-associated fluorescent microspikes (magnified in inset). Specific PMD substitutions that map into EC2 (D202N, P215S; amino acid positions refer to the sequence of PLP) cause ER retention of mutant PLP (b, c) but not of mutant DM20 (e, f). In addition to labeling the membrane of live glial cells with an antibody raised against the cell-surface protein NG2 (blue), the surface appearance of PLP was confirmed by live staining (in red) with monoclonal antibody 3F4 (b, c, e, f). The intracellular EGFP signal marks either a late endosomal compartment (a, d–f) or the reticular ER when PLP is retained (b, c). Scale bar: A–C, 10 μm. B, In searching for a PLP-specific ER retention signal, PLP–myc chimeras S1–S4 were expressed that lacked different segments (S1–S4 in Fig. 1B) of the PLP-specific intracellular loop (schematically indicated at the top). In the absence of any additional modification, chimeras S1–S3 were able to exit the ER of oli-neu cells, reaching a late endosomal compartment (a–c). Chimera S4 was ER retained, suggesting that no retention signal had been removed. C, Importantly, in combination with PMD mutations mapping into EC2 [D202N (not shown) and P215S], chimera S2 (a) and chimera S3 (c) were also ER retained, with chimera S2 showing a partial rescue of mutant PLP highlighted by lamp-1 colabeling (red). This suggests the absence of an essential retention signal in the PLP-specific cytoplasmic loop that could completely explain the differential ER retention of mutant PLP and DM20 (A).

Figure 6.

The length and position of TM3 determine ER retention or release of mutant DM20. A, Kyte–Doolitle hydropathy plot of PLP, DM20, and DM20^LSAT–HPDK chimera. Scale bar: same for all as in PLP. Hydrophilicity plotted with a window of 11 residues, using DNA star. Negative hydrophilicity implies a highly hydrophobic stretch. Both PLP and DM20 share four highly hydrophobic TM stretches. The motif itself has been conserved among proteolipids and has been proposed to evolve from an ancestral gene encoding pore-forming proteins (Kitagawa et al., 1993). The PLP-specific region imparts a highly hydrophilic nature to intracellular loop (IC). In contrast, DM20 bears an extended hydrophobic stretch, which might allow TM3 to glide a single α helical turn up or down to reorient EC2 during local misfolding. This gliding phenomenon can be completely reversed by simply reversing the hydrophobicity of DM20-specific region to a positive hydrophilicity by replacing amino acid residues LSAT to HPDK. This engineered DM20^HPDK displays hydrophilic characteristics of PLP and also traffics like PLP^wt in transfected cells. The third TM domain of DM20 is potentially longer than in PLP as a result of a stretch of four uncharged residues in DM20 (positions 112–115) immediately preceding TM3. Replacing the juxtamembrane threonine by a lysine in DM20 (T115K) did not prevent cell-surface expression (B) but is likely to reduce transmembrane domain sliding of TM3. Importantly, T115K partially impaired transport of two PMD mutant isoforms (C, D). The substitution of four consecutive residues had an even greater effect on DM20 retention [HPDK¹⁵⁰: the predicted TM3 “stop transfer” signal in PLP at the equivalent position in DM20 (LSAT¹¹⁵)]. Also this modification (depicted on the right) allowed DM20 to traffic normally (E) but caused complete ER retention when combined with mutations D202N (F) or P215 (G). Together, this strongly suggests that a subtle TM domain sliding of TM3 allows DM20 (but not PLP) to properly fold the globular EC2 domain in the ER lumen, despite its PMD-causing substitution. Scale bar, 10 μm.

Figure 7.

Rescue of PMD mutation bearing chimeric PLP and DM20s from ER retention suggests coordination of QC of TM domain assembly and EC2 folding. oli-neu cells transfected with EGFP fusion constructs. Surprisingly, PMD mutation bearing chimeras S2, S3, DM20^T115K, and DM20^LSAT–HPDK with removal of the outer cysteine pair are not retained in the ER. Quadruple mutants are also readily detectable in the lamp-1-positive endosomal/lysosomal compartment (red) as depicted in the corresponding magnified panels (right). This suggests coordination of QC of TM domain assembly and EC2 folding in PLP and DM20.

Figure 8.

QC of EC2 folding and TM domain assembly of PLP/DM20 is not glial cell specific. A, When resolved under nonreducing conditions, the cell-surface DM20 proteins bearing PMD-causing mutations (D202N, L209H, P215S) form fewer cross-links, whereas ER-retained DM20 mutants (L223P and A242V) form cross-links to the same extent as the corresponding PLP mutants in a human oligodendroglial cell line (MO3.13). As reported previously (Dhaunchak and Nave, 2007), when the same samples were resolved under reducing conditions, no cross-links were observed at similar exposure times (data not shown). B, PMD causing PLP^P215S forms twofold to threefold more cross-links when compared with the corresponding DM20^P215S or wild-type PLP. C, The extent of cross-linking of PLP^P215S is greater than that of the corresponding DM20^P215S or wild-type PLP when analyzed in three different cell lines (COS-7, MO3.13, and HEK293). D, Removal of the outer cysteines (R) rescues the PMD-causing (P215S) mutant chimeras (M) S2, S3, DM20^T115K, and DM20^HPDK from cross-linking in four different cell lines. The percentage of cross-linking in MO3.13 cells quantified from two independent transfection experiments is shown on top. E, F, Removal of outer cysteines (R) releases the PMD-causing (P215S) mutant chimeras (M) S2, S3, DM20^T115K, and DM20^HPDK from ER retention in COS-7 cells and oli-neu cells (Fig. 7). The percentage of cell-surface PLP in COS-7 cells quantified from one of the three experiments (each performed in triplicate) is shown in F.

Figure 9.

Experimental elongation of TM3 in PLP rescues mutant PLP from ER retention. A, Incorporating the DM20 juxtamembrane pentapeptide (GLSAT) after lysine 151 in PLP does not perturb the cell-surface expression or lysosomal accumulation of chimera S5^wt but is likely to introduce transmembrane domain sliding of TM3 (Fig. 6). A representative phenotype of S5, PLP, DM20-bearing PMD mutation (P215S) is shown. Note that S5^P215 accumulates in lamp-1-positive compartment similar to wild-type S5 or DM20^P215S. PLP^P215S, however, is present in calnexin-positive reticular compartment and is absent from lamp-1-positive compartment. The merged and individual magnification panels of boxed area are shown on right. Scale bar, 10 μm. B, When compared with corresponding PLP mutants, using unpaired t test, the S5^D202N and S5^P215S show significant rescue from ER retention. *p < 0.05; **p < 0.01. The quantification from two independent experiments (each done in triplicate) shows that S5^D202N (54.57 ± 4.56 SEM) and S5^P215S (68.52 ± 2.66 SEM) are detected at the cell surface.