|Home | About | Journals | Submit | Contact Us | Français|
Parathyroid hormone-related protein (PTHrP) has two different targeting signals: an N-terminal signal peptide for the endoplasmic reticulum (ER) targeting and an internal nuclear localization signal. The protein not only functions as a secretory protein, but is also found in the nucleus and/or nucleolus under certain conditions. PTHrP signal peptide is less hydrophobic than most signal peptides mainly due to its evolutionarily well-conserved region (QQWS). The substitution of four tandem leucine residues for this conserved region resulted in a significant inhibition of the signal peptide cleavage. At the same time, proportion of nuclear and/or nucleolar localization decreased, probably due to tethering of the protein to the ER membrane by the uncleaved mutant signal peptide. Almost complete cleavage of the signal peptide accompanied by a lack of nuclear/nucleolar localization was achieved by combining the hydrophobic h-region and an optimized sequence of the cleavage site. In addition, mutational modifications of the distribution of charged residues in and around the signal peptide affect its cleavage and/or nuclear/nucleolar localization of the protein. These results indicate that the well-conserved region in the signal peptide plays an essential role in the dual localization of PTHrP through ER targeting and/or the membrane translocation.
The prepro-form of parathyroid hormone-related protein (ppPTHrP) has two targeting signals: an N-terminal signal peptide for endoplasmic reticulum (ER) targeting and a nuclear localization signal (NLS) in its mid-region. The protein enters the secretory pathway, and is cleaved to generate the bioactive mature PTHrP (mPTHrP), in which the N-terminal half is homologous to the parathyroid hormone, and other poorly characterized peptides (1). PTHrP is also found in the nucleus and/or nucleolus where it may play a different role from that of the secreted mPTHrP (2). According to the signal recognition particle (SRP)-dependent cotranslational mechanism, the nuclear/nucleolar localization of PTHrP is surprising, since the NLS is located downstream of the ER-targeting signal peptide and is not expected to interact with nuclear transport machineries. In spite of that, several possibilities which could explain nuclear/nucleolar localization of PTHrP have been proposed (3).
The first possibility is reuptake of the once secreted PTHrP through endocytosis in a receptor mediated manner (4). To accomplish this, pro-PTHrP (pPTHrP) must escape from proteolytic processing for maturation during secretion by an unknown regulatory mechanism, since mPTHrP lacks the NLS. Whether some regulation(s) exists for pPTHrP processing between the mPTHrP and NLS by the prohormone convertase is still unclear.
The second possibility is alternative translational initiation from one of the four CUG codons of PTHrP mRNA located downstream of the authentic initiator methionine codon (AUG). The translation product is predicted to have a shorter signal peptide that is non-functional for ER targeting and/or translocation. In support of this possibility, it was shown that expression of a mutant PTHrP cDNA lacking the authentic initiator codon in the cell results in exclusive nuclear localization of the protein, although the translation is much less efficient compared with that from the authentic initiator codon (5, 6). A shunt mechanism for the alternative translation initiation must thus be elucidated.
The third possibility is the regulation of ER targeting and/or translocation, the first possible step for determining the protein localization, through the unique structural features of the PTHrP nascent chain including its signal peptide. In our previous work, we revealed, by examining the effects of a series of deletions of ppPTHrP, that the proportion of nuclear/nucleolar localization of PTHrP is roughly inversely correlated to the proportion of signal peptide cleavage (7). Moreover, the targeting of PTHrP to the nucleus is still observed when a PTHrP cDNA, in which all the CTG codons in the signal peptide has been replaced with TTG codons to avoid alternative translation initiation without changing the encoded amino acids, is expressed in Chinese hamster ovary cells (6). In addition, the precursor forms of PTHrP associated with microsomal vesicles are not completely protected from trypsin digestion, which suggests that a significant portion of the protein is exposed to the cytoplasmic face of the ER membrane for proteasome-mediated degradation (8). These observations suggest that partitioning of the PTHrP nascent chain through ER targeting and/or the membrane translocation is involved in the mechanism(s) for the dual localization of the protein. Although it is known that acute stress of the ER temporarily prevents prion protein from translocating to the ER and redirects it to the cytosolic degradation pathway (9), the underlying mechanism for the translocational regulation of PTHrP, which is insensitive to stress conditions, seems to be different (7).
In this study, we focused on the unique structural features of the N-terminal region of ppPTHrP, including its signal peptide, and assessed their roles in signal peptide cleavage and in the dual localization of the protein in detail by use of mutational analysis. Our results revealed that the well-conserved region in the h-region of the signal peptide and the positively charged residues in the pro-region play important roles in the regulation of ER targeting and/or translocation of the PTHrP nascent chain and have an impact on its dual localization. In addition, we also investigated the structural requirements of the signal peptide for its efficient cleavage and for suppression of its inefficient cleavage by cycloheximide and anisomycin.
Plasmids encoding the mutant proteins listed in Fig. 1B were constructed by site directed mutagenesis of the plasmid encoding the N-terminal 33 residues of rat preproPTHrP fused to a linker sequence (ELGS), the C-terminal 182 residues of Escherichia coli (E. coli) leader peptidase (LP) and an additional nine residues (YPYDVPDYA) of human influenza hemagglutinin (HA) tag (ppPTHrP33-LP) (7) using the Quick Change II site-directed mutagenesis kit (Stratagene) according to the manufacturer’s instructions. Synthesis of the mutagenesis oligonucleotides (Table I) and confirmation of the sequences of the constructed plasmids encoding the mutations were performed by Fasmac Co., Ltd (Kanagawa, Japan).
An expression plasmid encoding full-length ppPTHrP-EGFP fusion protein was constructed by insertion of the HindIII/BamHI fragment of the full-length PTHrP cDNA (7) into HindIII and BamHI-digested pEGFP-N3 (Clonetech). Plasmids encoding the mutant ppPTHrP-EGFP, used in Figs 3 and and4,4, were constructed as described above.
Preparation of an antibody against E.coli LP (anti-LP) is described previously (7). An antibody against the C-terminal 20 residues of rat proPTHrP [anti-PTHrP(122-141)] was prepared as follows. A peptide consisting of the rat proPTHrP(122-141) (CDPQPHTSPTSTSLEPSSRTH) and an extra N-terminal cysteine residue was synthesized and conjugated to keyhole limpet hemocyanin through the sulfhydryl group of the N-terminal cysteine residue using MBS (m-maleimidobenzoyl-N-hydroxysuccinimide ester) (PIERCE, USA). A rabbit (female New Zealand White) was immunized with the conjugated peptide.
COS-1 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal calf serum, penicillin and streptomycin. Cells (1 × 105) were seeded into 24-well plates and transfected with the indicated plasmids using LipofectAMINE 2000 (Life Technologies, Inc.) according to the manufacturer’s instructions. After 20 h incubation at 37°C, 5% CO2, culture medium was replaced with 0.25 ml of DMEM (-Met, -Cys) supplemented with 10% dialysed calf serum, and the cells were further incubated for 4 h. The cells were labelled with 10 µCi of [35S]-Met/Cys for 15 min for the ppPTHrP33-LP fusion proteins or 30 min for the ppPTHrP-EGFP fusion proteins. The labelling reaction was stopped with 0.25 ml of 2 × immunoprecipitation buffer [10 mM Tris-Cl (pH 7.4), 0.2 mM ethylenediaminetetraacetic acid, 0.1% Triton X-100, 0.1% sodium dodecyl sulfate (SDS)] supplemented with protease inhibitors (1 mM phenylmethanesulfonyl fluoride (PMSF), 20 µM each of chymostatin, pepstatin, leupeptin and antipain), and 2 mM each of cold Met and Cys. The cell lysates were treated with DNase I (0.04 mg/ml), 2 mM MgCl2 and RNase A (0.04 mg/ml) for 1 h on ice. Immunoprecipitation by the anti-LP antibody for the ppPTHrP33-LP fusion proteins or the anti-PTHrP(122-141) for the ppPTHrP-EGFP fusion proteins and SDS–PAGE (polyacrylamide gel electrophoresis) were performed as previously described (7). Where indicated, the N-glycan chain of the LP reporter was cleaved by peptide N-glycosidase (PNGase F, New England Biolabs, USA) treatment according to the manufacturer’s instructions. The protein bands were quantified using BAS5000 (Fuji Film) or Typhoon FLA 7000 (GE-Healthcare), and the following values were calculated.
Proportion of glycosylated prepro-form in PNGase F-untreated experiments:
Proportion of signal peptide cleavage in PNGase F-treated experiments:
The proportion of non-glycosylated prepro-form in the PNGase F-untreated experiments was calculated under an assumption that all the glycosylated pro-form is converted to non-glycosylated pro-form by the PNGase F treatment.
Cells transfected with the plasmid encoding the ppPTHrP-EGFP fusion protein were cultured on a ϕ35 mm culture dish. After the medium was changed to phosphate buffered saline (PBS), the transfected cells were observed using an IX-71 inverted microscope equipped with fluorescence apparatus (Olympus).
The signal peptide of PTHrP has several characteristic structural features, which are evolutionarily well-conserved among mammalian species but are unfavourable for its ER targeting and/or translocation (Fig. 1A). To assess the functions of these structural features on the dual localization of PTHrP, we constructed mutants that alter the charge distribution, the hydrophobicity, and/or the cleavage site of the signal peptide (Fig. 1B). Because the C-terminal portion of ppPTHrP including the NLS affects cleavage of the signal peptide (7), the N-terminal region of rat ppPTHrP (ppPTHrP33) that does not contain the NLS was used to assess the intrinsic activity of the signal peptide including the pro-region. Escherichia coli LP was used as a reporter protein. Glycosylation of the reporter domain indicates its translocation across the ER membrane.
The difference of the total sum charges of up to 15 residues in each side flanking the first internal signal-anchor domain is reported to determine its orientation, with the more positive portion facing the cytosol (10). According to this criteria, the net charge difference calculated for the wild-type rat ppPTHrP by subtracting the sum of the charges of the n-region (N) including a positive charge from the N-terminal amino group of the signal peptide from that of the pro-region (C) seems to be unfavourable for translocation of its C-terminal portion across the ER membrane. The difference [Δ(C−N)] for the rat PTHrP signal peptide is +1 (Fig. 1B), suggesting that the signal peptide is preferentially inserted into the ER membrane in an N-terminal exoplasmic/C-terminal cytosolic (Nexo/Ccyto) orientation that might preclude access of the cleavage site to the active site of the signal peptidase located in the lumen of the ER. The well-conserved charged residues in the pro-region (R32 and R33) are functionally unknown, although the C-terminal two residues (R35 and R36) are likely required for cleavage by pro-hormone convertase (1, 11). A C1 mutant was constructed to increase the Δ(C – N) value from +1 to +5. To invert the charge difference in favour of insertion with an Ncyto/Cexo orientation, the positively charged arginines in the c-region (R23 and R25) and in the pro-region (R32 and R33) are replaced with the negatively charged glutamic acids in C2 and C3 mutants, respectively.
The h-region of the PTHrP signal peptide is less hydrophobic than typical h-regions. The predicted apparent free energy (ΔGapp) of insertion of a transmembrane helix into the ER membrane has been presented as a biological hydrophobicity scale (12). We calculated ΔGapp for the N-terminal 20 residues of wild-type and mutant PTHrP signal peptide (Fig. 1B). For example, whereas ΔGapp for the control α2u globulin (GL) signal peptide is −1.221 kcal/mol, ΔGapp for the PTHrP signal peptide is 1.973 kcal/mol. The conserved region (QQWS) significantly contributes to the lower hydrophobicity. Relatively well-conserved hydroxyl residues located on the boundary between the h- and the c-regions also contribute to the lower hydrophobicity.
The conserved region (QQWS) was replaced with four tandem leucine residues (LLLL) in the H1 mutant. The ΔGapp of the H1 mutant signal peptide was predicted to be −3.382 kcal/mol (Fig. 1B). To increase the hydrophobicity of the boundary between the h- and the c-regions, two serine residues were replaced with leucine residues in the H2 mutant. The ΔGapp of the H2 mutant was predicted to be 0.333 kcal/mol, while the predicted ΔGapp was markedly decreased (−5.054 kcal/mol) in the H1/H2 double mutant. An H1/C3 mutant was constructed to assess the influence of the charged residues in the pro-region on the hydrophobic H1 mutant signal peptide function.
The cleavage site of the rat PTHrP signal peptide was predicted to be between G24 and R25 using SignalP 4.0 (13). A de novo designed signal peptide cleavage cassette was originally developed for an E.coli system based on the patterns of conserved amino acids that have been defined by statistical analyses and confirmed by experimental studies (14). It was later found to also work well in a eukaryotic in vitro reconstitution system composed of reticulocyte lysate, purified SRP and dog pancreas microsomes (15). The wild-type cleavage site sequence was replaced with the de novo designed cleavage cassette in a CC mutant. By virtue of this mutation, the cleavage site was moved one residue closer towards the C-terminus. CC/C3, H1/CC and H1/CC/C3 mutants were constructed to assess the influence of the charge inversion in the pro-region and/or in the hydrophobic h-region on the function of the cleavage cassette.
Wild-type and mutant fusion proteins (ppPTHrP33-LP) were expressed in COS-1 cells, and the proportions of glycosylation and of signal peptide cleavage were assessed and calculated as described in Materials and Methods. Although the control signal peptide of the GL-LP fusion protein (7) was completely cleaved, the proportion of wild-type PTHrP signal peptide cleavage was limited to 84% (Fig. 2A). As expected, the signal peptide cleavage decreased to 26% in the C1 mutant signal peptide. Charge inversion of the two positively charged residues on both sides of the cleavage site (R23 and R25) in the C2 mutant had no effect on the signal peptide cleavage. However, the signal peptide cleavage slightly increased to 90% in the C3 mutant in which two tandem glutamic acid residues were substituted for arginine residues in the middle of the pro-region (R32 and R33) to invert the charge difference. The charge difference thus had some effect on cleavage of the PTHrP signal peptide.
The glycosylated prepro-form of the fusion protein was observed in the H1 mutant with a hydrophobic h-region created by replacement of the conserved region (QQWS) with tandem four leucine residues (Fig. 2B). The proportions of both the glycosylated prepro-form and the signal peptide cleavage were 43%, indicating that at least a half of the mutant signal peptide was not cleaved by the signal peptidase after the ER translocation. At least 75% of the prepro-form of the H1 mutant is likely to reside in the ER membrane as a glycosylated type II membrane protein estimated from the ratio of the proportion of the glycosylated prepro-form (PNGase −) to the proportion of the prepro-form (PNGase +). On the other hand, the increased hydrophobicity at the boundary between the h- and the c-regions in the H2 mutant had no effect on the proportion of the signal peptide cleavage. A phenotype similar to that of the H1 mutant was observed for the H1/H2 double mutant. A severe defect of the signal peptide cleavage was observed in the H1/C3 mutant where the proportion of the signal peptide cleavage decreased to 29%, and the proportion of the glycosylated prepro-form concomitantly increased to 52%. Contrary to the case of the lower hydrophobic wild-type h-region, negatively charged residues of the C3 mutation in the pro-region had strong inhibitory effect on signal peptide cleavage when combined with the hydrophobic h-region of the H1 mutation in the H1/C3 mutant.
Replacement of the wild-type cleavage site with the de novo designed signal peptide cleavage cassette, CC, had no effect on the proportion of signal peptide cleavage (Fig. 2C). Similar to the phenotype of the C3 mutant, a little elevation of the signal peptide cleavage was observed in the CC/C3 mutant. On the other hand, almost complete cleavage (97%) of the signal peptide was observed with the H1/CC mutant. Therefore, it can be seen that the combination of the hydrophobic leucine-rich h-region and the de novo designed signal peptide cleavage cassette is important for efficient signal peptide cleavage. Contrary to the case of the H1 mutation, additional introduction of the C3 mutation to the H1/CC mutation (H1/CC/C3) had no inhibitory effect on cleavage of the mutant signal peptide.
To evaluate the effects of those mutations on the final localization of PTHrP, full length wild-type or mutant ppPTHrP fused to EGFP (Fig. 3A) were expressed in COS-1 cells. Similar effects of the mutations on the proportion of signal peptide cleavage to those for the ppPTHrP33-LP fusion proteins were observed for the ppPTHrP-EGFP fusion proteins, except that signal peptide cleavage of the C2 mutant was significantly inhibited (Fig. 3B, compare to Fig. 2).
Subcellular localization of the EGFP fusion proteins was detected by a fluorescence microscopy (Fig. 4). While the reporter EGFP expressed alone was found throughout the entire cell with slightly intense staining in the nucleus, the fusion protein carrying the wild-type signal peptide was almost exclusively detected in the nucleoli. Unexpectedly, such exclusive nucleolar localization was not observed with the C1 mutant in which cleavage of the mutant signal peptide was significantly suppressed. A significant proportion of the C1 mutant fusion protein was found in the nuclear matrix with faint cytoplasmic fluorescence. A similar phenotype was also observed for the C2 mutation. These results suggest that the positively charged residues in the n-region and c-region of the signal peptide are important and/or negatively charged residues in the mutants are inhibitory for the nucleolar translocation of PTHrP from the nuclear matrix. Charge inversion of the positively charged residues in the pro-region in the C3 mutant caused some effects on its final localization, whereby a significant fraction of the mutant protein was found in the perinuclear region, as well as in the nucleus.
Subcellular localization of the PTHrP-EGFP fusion proteins carrying the H1 mutant signal peptide was quite different from that of the wild-type. The substitution of tandem four leucine residues for the conserved QQWS region (H1) resulted in a significant perinuclear localization of the mutant PTHrP-EGFP fusion protein, with a concomitant reduction in the proportion of the nuclear and/or nucleolar localization. More significant perinuclear localization was observed in the H1/C3 and H1/CC mutants, in the former of which the ppPTHrP33-LP was highly glycosylated (Fig. 2B) and in the latter of which the signal peptide was efficiently cleaved both in the ppPTHrP33-LP (Fig. 2C) and ppPTHrP-EGFP fusion proteins (Fig. 3B).
Taken together, the structural features of the signal peptide and the pro-region significantly affect the nuclear and/or nucleolar localization of PTHrP. In particular, the conserved region in the h-region (QQWS) plays an important role in the regulation of ER targeting and/or membrane translocation of PTHrP nascent chain for its dual localization.
COS-1 cells expressing the wild-type or either of the mutant ppPTHrP33-LP fusion proteins were treated with cycloheximide or anisomycin during the labelling reaction to assess the effects of the translation rate and ribosomal peptide elongation cycle on the proportion of the signal peptide cleavage (Fig. 5A). Both cycloheximide and anisomycin treatments had no effect on the proportion of the wild-type signal peptide cleavage even at higher concentrations of the antibiotics (Fig. 6A and B). Similarly, all the mutants in Fig. 5A, except for C1, showed no effect on the signal peptide cleavage when treated with either antibiotic.
For the C1 mutant, the proportion of signal peptide cleavage was increased more than a 2-fold with cycloheximide treatment but not with anisomycin (Fig. 5A). The proportion of C1 mutant signal peptide cleavage increased concomitantly with a decrease in the rate of translation by cycloheximide (Fig. 6A–C, left panels). On the other hand, a moderate elevation of signal peptide cleavage was observed by anisomycin only at concentrations higher than 0.20 µg/ml, a concentration that reduces the rate of translation to as low as 10% that of the untreated control (Fig. 6A–C, right panels). These results suggest that suppression of the defective signal peptide cleavage seen with the C1 mutant depends more on the conformation of the ribosome having high affinity for cycloheximide rather than on the rate of translation.
The proportion of the glycosylated prepro-form decreased by treatment with either cycloheximide or anisomycin with the H1 and H1/C3 mutants (Fig. 5B). Cleavage of the mutant signal peptides increased concomitantly with the reduction of the glycosylated prepro-form (Fig. 7). In contrast with the C1 mutant, suppression of defective signal peptide cleavage in the H1 and H1/C3 mutants seem to be essentially dependent on the rate of translation, irrespective of which antibiotics are used (Figs 6C and and7).7). Taken together, it is clear that the mechanisms underlying the inefficient cleavage of wild-type, C1 and H1 mutant signal peptides are quite different from one another.
The signal peptide and the pro-region of PTHrP have several conserved structural features that are predicted to be unfavourable for ER targeting and for membrane translocation, but are critical for its dual localization. Such evolutionary conservation of the primary structure of a signal peptide and a pro-region is unusual even among mammalian species, where usually their functional properties, rather than the peptide sequence per se, are well conserved (16). Among these structural features, the conserved QQWS region in the h-region of the PTHrP signal peptide seems to be the most important. In addition, the conserved positively charged residues in the pro-region significantly affect mode of insertion and/or cleavage of the signal peptide.
Like secretory proteins in general, most newly synthesized PTHrP enters the secretory pathway (Fig. 8, A→B→C→D, bold arrows). However, a proportion of PTHrP is transported to the nucleus/nucleolus (Fig. 8, G) by partitioning during ER targeting and translocation. Two possible pathways for this dual localization of PTHrP are conceivable.
The first possible pathway is that the nascent chain might fail to be targeted to the ER, due to the weak interaction between the less hydrophobic PTHrP signal peptide and the SRP. In this situation, the NLS of PTHrP could interact with importin β (17), and the protein could then be transported to the nucleus/nucleolus (Fig. 8, A→A′ →G, thin arrows).
Similar to the suppression of inefficient ER translocation in SRP deficient cells by antibiotics (18, 19), the defect of the C1 mutant is efficiently suppressed by cycloheximide treatment, and less so by anisomycin (Figs 5A and and6).6). These results imply that suppression of the C1 mutant phenotype is not merely due to the reduced translation rate with antibiotics treatment, but is also dependent on the conformation of the ribosome. Here, it is possible that the cycloheximide-binding ribosome, which has a higher affinity with the SRP, compensates for the weak interaction between the C1 mutant signal peptide and the SRP. In accordance with this possibility, electrostatic interactions between the positively charged residues of the n-region of signal peptide and the phosphate backbone of SRP RNA are suggested to play a significant role in signal peptide recognition (20). In contrast with the C1 mutant signal peptides, incomplete cleavage of the wild-type PTHrP signal peptide was not suppressed either by cycloheximide or by anisomycin (Figs 5A and and6).6). The mechanism of incomplete cleavage of the wild-type PTHrP signal peptide seems not to depend on conformation of the ribosome.
The second possible pathway is release of the PTHrP nascent chain to the cytosol after the ER targeting (Fig. 8, B→E′→G, thin arrows). The charge distribution of the PTHrP signal peptide including the pro-region might affect the orientation of insertion and hairpin loop formation during the interaction between the protein conducting channel (PCC) and the signal peptide (Fig. 8, B→C). Inefficient formation of the hairpin loop facilitates exposure of the C-terminal portion of the nascent chain to the cytosol and can promote interaction of the NLS with importin β (17) to transport the protein into the nucleus/nucleolus (Fig. 8, E′→G).
It is likely that the positively charged residues in the c-region of the signal peptide or the pro-region of PTHrP might interfere with the hairpin loop formation (Fig. 8, B→C) of the signal peptide (10, 21). At least in some cases, hairpin loop formation occurs by a reorientation of the signal peptide after its head-on insertion (22, 23). Although charge inversion in the c-region (C2) had no effect on the signal peptide cleavage, that in the pro-region (C3) slightly enhanced signal peptide cleavage (Fig. 2A). In addition, the ppPTHrP-EGFP fusion protein carrying the C3 mutation was detected in perinuclear region (Fig.4). These results suggest that the conserved positively charged residues in the pro-region facilitate the release of the PTHrP nascent chain from the ER membrane by inefficient formation of the hairpin loop. In addition to the signal peptide itself, positively charged residues in the pro-region also had some effect on the function of the PTHrP signal peptide for the dual localization. Evolutionary conservation suggests the importance of these positively charged residues (Fig. 1A) in the functional regulation of PTHrP signal peptide. As reported in some cases, the downstream region from the cleavage site of a secretory protein can affect the gating of the PCC, which is one of the critical steps for translocation of the signal peptide (16).
Hydrophobic interaction between the signal peptide and the PCC is important for the opening of the lateral gate (Fig. 8, B→C or B→E) through which the signal peptide gains access to the lipid bilayer (12, 24). It is possible that the low hydrophobicity of the h-region of the PTHrP signal peptide causes a cytosolic release of a significant fraction of the nascent chain through its inefficient lateral gate opening activity (Fig. 8, E′). The relatively high predicted ΔGapp (1.973 kcal/mol, Fig. 1B) of the h-region of the wild-type PTHrP signal peptide seems to be energetically unfavourable for the opening of the lateral gate, but favourable for release of the PTHrP nascent chain from the PCC into the cytosol where it would then be available for nuclear/nucleolar localization (Fig. 8, B→E′→G, thin arrows). Indeed, nuclear/nucleolar localization was significantly precluded when the conserved region was replaced with a tandem of leucine residues in the H1 mutation, for which the predicted ΔGapp was −3.382 kcal/mol, (Fig. 4). In the H1 and H1/C3 mutants, a significant proportion of the signal peptides were not cleaved by the signal peptidase (SPase), resulting in the formation of the glycosylated prepro-form (gpp) and non-glycosylated prepro-form (pp). A major proportion of the uncleaved mutant proteins is integrated into the ER membrane as type II membrane proteins, in which sequestration of the NLS into the ER would prevent it from interacting with cytosolic importin β (Fig. 8, C→C′→D′, dotted arrows). It is also possible that the non-glycosylated prepro-form (pp) of the mutant proteins are integrated into the ER membrane as type I membrane proteins (Fig. 8, B→E→F, dotted arrows). Although the NLS can interact with importin β in this case, tethering of the mutant proteins in the ER membrane by the non-cleavable hydrophobic mutant signal peptides may prevent its nuclear/nucleolar transport.
Cleavage of the H1 mutant signal peptide was significantly inhibited, while its targeting efficiency was apparently unaffected (Fig. 2B). In a study of the model signal peptides consisting of various lengths of leucine residues in the h-region followed by the cleavage cassette (AQAA), signal peptide cleavage was observed in the leucine length from 8 (ΔGapp = 0.833 kcal/mol) to 17 (ΔGapp = −6.189 kcal/mol) residues, but not in the length from 20 (ΔGapp = −8.341 kcal/mol) to 26 (ΔGapp = −12.51 kcal/mol) residues (15). The predicted ΔGapp values of the H1 mutant (−3.382 kcal/mol, Fig. 1B) is within the cleavable hydrophobicity of that study. In spite of this, cleavage of the H1 mutant signal peptide was significantly reduced (Fig. 2B). A similar result was observed with the H1/H2 mutant, for which the predicted ΔGapp value was −5.054 kcal/mol. These results suggest that the combinations of the hydrophobic h-regions and the wild-type cleavage site sequence are unfavourable for cleavage. In contrast, almost complete signal peptide cleavage was observed when combining the hydrophobic h-region and the de novo cleavage cassette (H1/CC), although the combination of the wild-type h-region and the CC mutation (CC) had no effect on signal peptide cleavage (Fig. 2C). This all suggests that both the hydrophobic h-region and the well-adapted cleavage site are simultaneously required for the mutant PTHrP signal peptide to be translocated and cleaved as efficiently as typical signal peptides.
Surprisingly, the combination of the hydrophobic h-region in the H1 mutation and the negatively charged residues in the pro-region of the C3 mutation (H1/C3) caused a more severe inhibition of the signal peptide cleavage (Fig. 2B), even though the C3 mutant slightly enhanced the cleavage in combination with the wild-type h-region (Fig. 2A). Efficient hairpin loop formation and the rapid transfer of the signal peptide to the lipid bilayer, induced by the combination of the negative charge in the pro-region and the hydrophobic h-region may reduce the time window for binding of the wild-type PTHrP cleavage site with the signal peptidase (Fig. 8 C→C′). Alternatively, a conformational change around the cleavage site of the H1/C3 mutant induced by the negatively charged residues in the mutant pro-region may affect its interaction with the active site of the signal peptidase. In the case of the H1/CC/C3 mutant, the cleavage inhibitory effect of the C3 mutation was not observed (Fig. 2C). Here, it is possible that efficient cleavage (Fig. 8, B→C→D, bold lines) by a combination of the H1 and CC mutations could overcome the inhibitory effect of the C3 mutation (Fig. 8, C′).
When the H1 mutation is present, the effects of the two antibiotics on the suppression of the defect of cleavage are similar and mainly dependent on the rate of translation (Figs 5B and and7).7). This implies that due to the slow translation rate caused by the antibiotic treatment, a sufficient time window might be available for binding the cleavage site with the active site of the signal peptidase (Fig. 8, C).
Unlike these mutant signal peptides, incomplete cleavage of the wild-type PTHrP signal peptide was not suppressed either by cycloheximide or by anisomycin (Figs 5A and and6).6). The mechanism for the incomplete cleavage of the PTHrP signal peptide seems to depend on neither conformation of the ribosome nor the rate of translation. The different effects of the two antibiotic treatments observed on the inefficient cleavage of the wild-type, C1, and H1 mutant signal peptides suggest that their underlying mechanism are also different.
To the best of our knowledge, this study provides the first evidence that the unique structural features of the signal peptide of PTHrP play a significant role in its dual localization through targeting and/or translocational regulation at the ER membrane. The function of the nuclear/nucleolar PTHrP has been suggested using mutants lacking the NLS (25, 26). It must be noted that some of the poorly characterized peptides such as osteostatin, generated during secretion (1, 27–29), are deficient in these mutants. The possibility cannot be excluded that at least some aspect of these phenotypes are caused by deficiencies of these peptides. Also, it must be noted that C-terminal truncation of prepro-PTHrP increases the secretion of mature PTHrP (7, 30). To avoid these ambiguities, the signal peptide mutants shown in this report, in which the nuclear/nucleolar translocation is severely impaired, can be useful alternative tools for the functional analysis of nuclear/nucleolar localized PTHrP. Moreover, these mutant signal peptides may be used for probing interactions of the nascent chain with the components of the ribosome, the PCC, and the lipid bilayer during ER translocation to reveal their regulatory functions.
We thank Yoshiaki Takahashi (Niigata University) for encouragement during the course of this work, Haruko Suzuki (Yokohama City University) for antigen peptide synthesis, Hui Lin Chai (Yokohama City University) for antibody preparation, Shigeru Hirayama (Niigata University) for the cell culture experiments, and Takuji Izumikawa (Niigata University) and Jun Goto (Niigata University) for the radio isotope labelling experiments.
This work was supported in part by a grant-in-aid for scientific research (13671934) from the Ministry of Education, Science, Sports and Culture of Japan.