|Home | About | Journals | Submit | Contact Us | Français|
The cellular prion protein (PrP) is essential for transmissible prion diseases, but its exact physiological function remains unclear. Better understanding the regulation of the human prion protein gene (PRNP) expression can provide insight into this elusive function. Spliced XBP1 (sXBP1) was recently shown to mediate endoplasmic reticulum (ER) stress-induced PRNP expression. In this manuscript, we identify Luman, a ubiquitous, non-canonical unfolded protein response (UPR), as a novel regulator of ER stress-induced PRNP expression. Luman activity was transcriptionally and proteolytically activated by the ER stressing drug brefeldin A (BFA) in human neurons, astrocytes, and breast cancer MCF-7 cells. Over-expression of active cleaved Luman (ΔLuman) increased PrP levels, while siRNA-mediated Luman silencing decreased BFA-induced PRNP expression. Site-directed mutagenesis and chromatin immunoprecipitation demonstrated that ΔLuman regulates PRNP expression by interacting with the ER stress response element 26 (ERSE26). Co-over-expression and siRNA-mediated silencing experiments showed that sXBP1 and ΔLuman both up-regulate ER stress-induced PRNP expression. Attempts to understand the function of PRNP up-regulation by Luman excluded a role in atorvastatin-induced neuritogenesis, ER-associated degradation, or proteasomal inhibition-induced cell death. Overall, these results refine our understanding of ER stress-induced PRNP expression and function.
Cellular prion protein (PrP) plays a fundamental role in the development of prion diseases. PrP is necessary for prion infection and its levels influence the progression of prion disease1,2,3. In non-infectious conditions, PrP has beneficial effects. PrP is involved in synaptic transmission4, cell signaling5, cell adhesion6, white matter maintenance7, hematopoietic differentiation8, and protection against oxidative stress9, endoplasmic reticulum (ER) stress10, and Bax-mediated cell death11,12,13,14,15. Furthermore, PrP has been closely linked to cancer resistance, tumorigenesis, and proliferation (reviewed in ref. 16). Despite these important roles of PrP in maintaining tissue homeostasis, the underlying molecular mechanisms regulating prion protein gene (PRNP) expression are not well defined. A better understanding of the regulation of PRNP expression will help clarify the physiological purpose of PrP, and is necessary to harness the roles of PrP in disease and tissue homeostasis.
The human PRNP is composed of a large intron flanked by two exons17. The PRNP promoter region is devoid of a TATA box, but contains a CpG island characteristic to housekeeping genes. Consistent with this feature, the PRNP is broadly expressed in the human body18. The expression of the PRNP is regulated by p5319, oxygen levels20,21,22, and copper exposure23. In addition, nerve growth factor increases PRNP promoter activity and PRNP mRNA levels in the developing brain24,25. The PRNP promoter contains several elements, including the heat shock (HSE), nuclear factor IL-6 (NF-IL6), specificity protein 1 (SP1), and muscle-specific factor (MyoD) elements26. Recently, four functional endoplasmic reticulum stress response elements (ERSE) were identified in the PRNP promoter region and PRNP expression was shown to be up-regulated by ER stress10.
ER stress triggers the activation of the unfolded protein response (UPR), a signaling cascade that attenuates overall translation, up-regulates the expression of genes necessary to restore adequate protein folding, promote ER-associated degradation (ERAD) of misfolded proteins, or trigger the apoptosis of cells under unresolvable ER stress. The UPR can be activated via three canonical pathways: the ER transmembrane sensors protein kinase RNA-like endoplasmic reticulum kinase (PERK), inositol-requiring enzyme 1α (IRE1α), and activating transcription factor 6α (ATF6α). PERK activation leads to eIF2α phosphorylation, an event that attenuates overall translation, but promotes the translation of the activating transcription factor 4 (ATF4)27. Activation of IRE1α enables the splicing of X-box binding protein 1 (XBP1) mRNA, causing a frame shift necessary to the translation of the functional spliced XBP1 (sXBP1) transcription factor28. Lastly, ATF6α is expressed as an ER-resident transmembrane protein that, upon ER stress, progresses to the Golgi apparatus, where it undergoes a proteolytic cleavage that releases its N-terminal cytosolic region, the active cleaved ATF6α (ΔATF6α) transcription factor29. Of these three factors, sXBP1 and ΔATF6α, but not ATF4, are linked to PRNP expression during ER stress10. Indeed, sXBP1 and ΔATF6 over-expression increases PRNP promoter activity, and both factors bind the PRNP promoter in ER stressed cells10. However, the siRNA-mediated silencing of ATF6α does not influence ER stress-induced PrP levels, and XBP1 silencing attenuates, but does not abolish, ER stress-induced PRNP expression in MCF-7 cells10,30. This indicates that neither factor is fully sufficient for ER stress-induced PRNP expression, and suggests the participation of additional alternative transcriptional UPR mediators.
The OASIS family of transcription factors is emerging as a group of novel, specialized, tissue-specific UPR regulators (reviewed refs 31 and 32). The OASIS family is constituted of OASIS/CREB3L1, BBF2H7/CREB3L2, CREBH/CREB3L3, AIbZIP/CREB3L4/CREB4 and Luman/LZIP/CREB3 family members. All members share bZIP and ER transmembrane domains. However, OASIS family members are differentially expressed, activated by distinct stimuli, and bind to different response elements31. In addition, most OASIS family members show high tissue specificity, with the exception of Luman, which is ubiquitously transcribed33. Like the other OASIS family members and ATF6α, Luman is an ER localized transmembrane protein. During ER stress, Luman undergoes regulated intramembrane proteolysis34,35, a process mediated by Golgi-resident proteases that release the cytosolic N-terminal portion of the protein. Active cleaved Luman (ΔLuman) then translocates to the nucleus, where it interacts with cis-acting promoter elements. ΔLuman binds cAMP- response element (CRE), CCAAT/enhancer binding protein (C/EBP) element33, endoplasmic reticulum stress response element II (ERSEII)36, and unfolded protein response element (UPRE)37. Ultimately, Luman promotes the expression of ERAD-associated genes, such as EDEM37, HERPUD136, Canx and Ubxn438, and of cholesterol metabolism regulators Insig1, Insig2 and Srebp138.
The objective of this study was to investigate the contribution of the OASIS family members and, more specifically, the Luman transcription factor to the regulation of PRNP expression by ER stress. Luman was transcriptionally and proteolytically activated by brefeldin A (BFA) in primary human central nervous system (CNS) neurons and astrocytes, and in MCF-7 breast carcinoma cells. Over-expression of ΔLuman increased PRNP mRNA and promoter activity, and PrP levels, and siRNA-mediated silencing of Luman reduced BFA-induced PRNP expression. Mutation of the ERSE26 element attenuated ΔLuman-mediated increase in PRNP promoter activity, and ΔLuman binding to the PRNP promoter ERSE26 region was confirmed by chromatin immunoprecipitation. Functionally, we exclude a putative role of Luman-mediated PRNP expression in (1) ERAD of misfolded proteins, (2) protecting against proteasomal inhibition-induced apoptosis or (3) atorvastatin-induced neuritogenesis.
Collectively, these results indicate that Luman contributes to BFA-induced PRNP expression by interacting with the ERSE26 element.
To identify which members of the OASIS family could contribute to ER stress-induced PRNP expression, OASIS, BBF2H7, CREBH, AIbZIP, and LUMAN transcript levels were assessed by RT-PCR in breast carcinoma MCF-7 cells, human primary neurons and astrocytes treated with the ER stressing drugs BFA, Th or TM (Fig. 1a). As previously observed, all three ER stressors increased PRNP mRNA levels. OASIS mRNA levels were only increased by BFA in astrocytes. Levels of BBF2H7 and CREBH were undetectable, or very low, in the three cell types and seemed unaffected by ER stress, with the exception of a salient CREBH increase in astrocytes treated with BFA. AIbZIP was very weakly detected in neuronal preparations. TM increased AIbZIP levels in astrocytes, but reduced them in MCF-7 cells. LUMAN transcripts were detected in the three cell types, and BFA treatment clearly increased LUMAN mRNA levels in MCF-7 cells, neurons and astrocytes. However, Th and TM treatments only caused modest and inconsistent increases in LUMAN mRNA levels. To clarify these results, the induction of LUMAN mRNA by ER stress was assessed by quantitative PCR, and showed a significant LUMAN mRNA increase in MCF-7, neuronal and astrocytic cultures treated with BFA, but not with Th and TM (Fig. 1b). Amplification of BiP mRNA (HSPA5) by RT-PCR (Fig. 1a) and quantitative PCR (Fig. 1c) controlled for induction of ER stress by BFA, Th and TM treatments, and the housekeeping gene HPRT1 was used as a control for overall mRNA levels. To confirm that the induction of Luman mRNA levels by BFA was due to an increase in LUMAN transcription, MCF-7 cells were treated with BFA in the presence of the transcription inhibitor actinomycin D or of the translation inhibitor cycloheximide. BFA treatment increased LUMAN mRNA levels, and actinomycin D co-treatment attenuated LUMAN mRNA levels of BFA- and DMSO-treated cells (Fig. 1d). Cycloheximide treatment did not significantly influence BFA-induced LUMAN mRNA, but caused a small LUMAN mRNA increase in the DMSO control condition, as observed previously for other genes39. The housekeeping gene HPRT1 controlled for overall mRNA levels.
Contrary to human primary neurons and astrocytes, MCF-7 cells are readily available, transfectable, and have previously been used as a model of PRNP regulation by ER stress. For these reasons, we focused our attention on MCF-7 cells. To determine whether ER stress led to Luman proteolytic cleavage of the N-terminal region, western blot was performed on MCF-7, neurons and astrocytes treated with the ER stressing drugs BFA, Th or TM. The ΔLuman N-terminal region was increased only in cells treated with BFA. A parallel assessment of the chaperone BiP levels confirmed the induction of ER stress by BFA, Th and TM. β-Actin levels were unchanged by ER stress treatments and controlled for equal protein loading (Fig. 1e). The purity of neuronal and astrocytic cultures was assessed by RT-PCR amplification of microtubule-associated protein 2 (MAP2) and glial fibrillary acidic protein (GFAP) mRNA (Fig. 1f). MAP2 mRNA was abundant in neuronal, but not astrocytic cultures. Conversely, GFAP levels were substantially superior in the astrocytic, than in the neuronal cultures. Neither MAP2 nor GFAP was detected in MCF-7 cells. HPRT1 controlled for overall mRNA levels. Luman cleavage was further investigated using a Luman construct bearing an N-terminal HA-tag. Once again, BFA treatment, but not Th or TM, led to the cleavage of Luman, as expected34 (Fig. 1g), starting as early as thirty minutes after treatment with BFA (Fig. 1h). Taken together, these results show that the ER stressing drug BFA (1) up-regulates LUMAN mRNA levels, and (2) promotes Luman activation by cleavage of its N-terminal cytosolic region.
To investigate the role of ΔLuman on PRNP expression, we transiently transfected ΔLuman (amino acids 1–215) in MCF-7 cells. Over-expression of ΔLuman led to a 1.5-fold increase in PRNP mRNA levels relative to those of empty vector-transfected cells (Fig. 2a). To assess whether this increase in PRNP mRNA translated into higher protein levels, PrP levels were assessed by western blot. There was a 1.8-fold increase in PrP levels in ΔLuman-transfected MCF-7 (Fig. 2b). ΔLuman over-expression was confirmed by western blot and β-Actin served as loading control. The increase in PrP-coding mRNA and protein levels suggest that ΔLuman contributes to the expression of the PRNP during ER stress. LUMAN expression was silenced prior to inducing ER stress to assess its contribution to ER stress-induced PRNP expression. In MCF-7 cells, BFA, but not Th or TM, treatment significantly increased PRNP mRNA levels when compared to DMSO-treated condition, and Luman silencing attenuated the induction of PRNP mRNA by BFA (Fig. 2c). Luman silencing also attenuated BFA-induced PRNP mRNA levels in primary human neurons (Fig. 2d). At the protein level, BFA, Th and TM treatments increased PrP levels, compared to DMSO control, in both scrambled and Luman-targeting siRNA transfected cells (Fig. 2e). BFA-induced PrP was immature glycosylated. Th increased unglycosylated PrP and maintained immature and mature glycosylated PrP levels, while PrP was entirely unglycosylated following TM treatment, as expected since TM inhibits N-linked glycosylation of newly synthesized proteins. However, silencing of Luman attenuated BFA-, but not Th- or TM-, induced PrP levels, thereby confirming the contribution of Luman to the regulation of PRNP expression during BFA-induced ER stress.
To determine if ΔLuman up-regulates PRNP promoter activity, ΔLuman was co-transfected with the secreted luciferase PRNP promoter reporter construct pML2-PRNP538 in HEK293T cells to maximize plasmid transfection efficiency. The over-expression of ΔLuman significantly increased PRNP promoter activity compared to control (Fig. 3a). Empty vector- and mock-transfected cells showed low luciferase activity. To identify the binding site of ΔLuman to the PRNP promoter, we assessed the impact of mutating the ERSE elements of the PRNP promoter on the ability of ΔLuman to increase PRNP promoter activity (Fig. 3b). Site-directed mutagenesis of the ERSE26 site, but not of the other ERSE elements, reduced by forty percent ΔLuman-induced increase in PRNP538 promoter activity (Fig. 3c). Furthermore, cumulative mutations of all the other ERSE elements of the PRNP promoter did not further decrease ΔLuman-induced promoter activity compared to the ERSE26 mutant (Fig. 3d). Levels of Luman were verified by western blot analyses in all transfected cells (Fig. 3a,c,d). The results show that, of the four sites investigated, ERSE26 is the only one required for full induction of PRNP promoter activity by ΔLuman. Chromatin immunoprecipitation was performed to determine whether ΔLuman interacted with the ERSE26 region of the PRNP promoter. To circumvent the lack of highly specific, commercially available antibody against ΔLuman, a Myc-tagged ΔLuman was transiently transfected and immunoprecipitated using an anti-Myc tag antibody. The positive control HERPUD1, a known target of Luman, and the PRNP ERSE26 promoter region were both amplified from the chromatin of ΔLuman-Myc-transfected cells immunoprecipitated with the anti-Myc antibody, but not from the chromatin of empty vector-transfected cells or samples immunoprecipitated without antibody or with a non-specific IgG. The ACTB promoter was not amplified and served as a negative control for ΔLuman–Myc binding (Fig. 3e). Furthermore, PRNP ERSE26, but not HERPUD1, amplification was lost in genetically altered HEK293T cells lacking the base pairs −202 to −191 of the PRNP promoter region (ΔERSE26: 5′-AGCCACGTCAGG-3′), a region that spans the 3′ conserved arm of the ERSE26 element (underlined) (Fig. 3f).
To assess whether silencing both Luman and XBP1 would be sufficient to fully suppress ER stress-induced PRNP expression, one or both transcription factors were silenced prior to treating MCF-7 cells with ER stressing drugs. BFA treatment led to an increase in both PRNP mRNA and protein levels, and silencing of either Luman or XBP1 attenuated by thirty-five percent BFA-induced increase in PRNP expression. Furthermore, silencing of both Luman and XBP1 further reduced ER stress-induced PRNP mRNA and protein levels (Fig. 4a,b). Induction and silencing of ΔLuman and sXBP1 protein levels was confirmed by RT-PCR or western blot. To determine whether ΔLuman and sXBP1 synergistically regulate PRNP promoter activity, Luman and/or sXBP1 were co-transfected with the pML2-PRNP538 reporter plasmid. Both ΔLuman and sXBP1 increased PRNP promoter activity when compared to empty-vector control (Fig. 4c). However, over-expression of both ΔLuman and sXBP1 did not further increase PRNP promoter activity. Over-expression of ΔLuman and sXBP1 was confirmed by western blot.
To better understand the physiological relevance of the regulation of PRNP expression by Luman and XBP1 during ER stress, and because both XBP1 and Luman regulate several genes involved in ERAD, the impact of PrP over-expression on the degradation rate of the ERAD substrate TTRD18G-GFP was assessed. The results show that PrP over-expression in the low PrP-expressing MCF-7 cell line did not facilitate TTRD18G-GFP degradation during a cycloheximide chase (Fig. 5a,b). Conversely, the degradation rate of the TTRD18G-GFP substrate was assessed in high PrP-expressing CR7 glioblastoma cells that underwent CRISPR/Cas9-mediated PrP knockout (KO) (Fig. 5c). No difference was observed between the degradation rate of TTRD18G-GFP in WT and PrP KO CR7 cells (Fig. 5c,d). In summary, these findings argue against a role of PrP in modulating ERAD.
Because (1) proteasomal inhibition causes ER stress and apoptosis, (2) Luman protects against ER stress-induced Caspase-3 activity36, and (3) PrP protects against ER stress-induced cell death10, the possibility that Luman–mediated PRNP expression protects against proteasomal inhibition-induced cell death was investigated. The impact of PrP on proteasomal inhibition-induced apoptosis was assessed by measuring Caspase-3/7 DEVDase activity in the WT or the CRISPR/Cas9-generated PrP KO CR7 cell lines (KO#1, KO#2) treated with epoxomicin. Cells treated with staurosporine served as positive control. Epoxomicin and staurosporine caused a twenty-five- to thirty-fold increase in DEVDase activity. Strangely, epoxomicin- and staurosporine-induced DEVDase activity was exacerbated in the PrP KO#1 cell line, but not in the PrP KO#2 cell line (Fig. 6a). To determine which effect is due to the loss of PrP, siRNA-mediated knockdown of PrP was achieved in CR7. Silencing of PrP expression did not significantly influence epoxomicin- or staurosporine-induced DEVDase activity (Fig. 6b). These results suggest that CRISPR-targeted knockdown has non-target effects in the CR7 KO#1 cell line, and that PrP does not protect CR7 cells against proteasomal inhibition-induced apoptosis.
Given that (1) Luman was recently shown to promote axonal growth40, (2) Luman regulates several genes involved in cholesterol metabolism38, and (3) the cholesterol synthesis inhibitor atorvastatin was reported to promote neuritogenesis in N2a cells by increasing PrP levels41, the contribution of Luman to atorvastatin-induced neuritogenesis was investigated. As expected, a 20μM atorvastatin treatment increased neuritogenesis in N2a and, to a lesser extent, the human neuroblastoma SK-N-SH cell line after 24h (Fig. 7a). Atorvastatin increased PrP in a dose- and time-dependent manner in N2a cells (Fig. 7b), but had the opposite effect in the SK-N-SH cell line (Fig. 7c). Quantification of neurite-bearing N2a cells confirmed neurite formation following atorvastatin treatment (Fig. 7d). However, despite a trend towards lower amounts of atorvastatin-induced neurite-bearing (Fig. 7e), and mean neuritic length (Fig. 7f) in N2a cells, no statistically significant effect was observed by siRNA-targeted knock down of Luman. Furthermore, while Luman proteolytic activation was clearly increased by BFA treatment, the levels of Luman were similar in atorvastatin- and DMSO- treated N2a (Fig. 7g) and HEK293T (Fig. 7h) cells, thus indicating that atorvastatin does not potently induce Luman activation. Blotting for β-Actin and eGFP controlled for equal loading and transfection efficiency, respectively. Overall, these data confirm previous reports that atorvastatin increases neuritogenesis and PrP levels in N2a cells. In addition, the results indicate that Luman is not necessary for atorvastatin-induced neuritogenesis, and exclude atorvastatin as a regulator of Luman proteolytic activation in murine and human cells. Contrary to N2a cells, atorvastatin reduces PrP levels in human SK-N-SH neuroblastoma cells.
The results of this study show that the ER stressing drug, BFA, up-regulates Luman activity in several human cell types, and that Luman contributes to BFA-induced PRNP expression by interacting with the ERSE26 of the PRNP promoter. Attempts to understand the function of Luman-induced PRNP expression excluded a role of PrP in promoting ERAD, protecting against epoxomicin-induced apoptosis, and atorvastatin-induced neuritogenesis.
Luman activation was observed both by increased mRNA levels and proteolysis into Luman by BFA-, but not TM- or Th-, treatment. The ability of BFA to increase LUMAN mRNA levels by five- to ten-fold in human CNS neurons and astrocytes and breast carcinoma MCF-7 cells, has not been previously reported. In contrast, Th, TM and the proteasomal inhibitor, MG132, up-regulates mRNA levels by three- to six-fold LUMAN in HEK293 and C6 glial cells36,42. Our results are consistent with previous reports that BFA, but not Th and TM, triggers Luman proteolytic activation in HEK293, Vero, RAW264 and dendritic cells34,37,43,44. Together, these results identify BFA as a most potent ER stressing inducer of LUMAN transcription in several cell types, and suggest that the induction of LUMAN transcription by Th and TM may be cell type specific.
The presence of LUMAN mRNA in MCF-7, neurons and astrocytes, and its responsiveness to BFA made Luman the most relevant OASIS family member to study as a mediator ER stress-induced PRNP expression. Our results identify Luman as a new regulator of PRNP expression. Over-expression of ΔLuman led to a 1.5-fold increase in PRNP mRNA and a 2.0- to 2.5-fold increase in PRNP promoter activity. This is comparable to the regulation of PRNP expression by sXBP1, which leads to a 2.0-fold increase in PRNP mRNA when over-expressed in MCF-7 cells30, indicating that Luman is as potent as the canonical UPR mediator sXBP1 at promoting PRNP expression. The level of PRNP induction was similar to the 2.0-fold induction in EDEM mRNA caused by ΔLuman in HEK293 cells37, but was much lower than the 9.0-fold increase of HERPUD1 mRNA36 and 4.5-fold induction of HERPUD1 promoter activity36. Moreover, silencing of Luman substantially attenuated BFA-induced PRNP expression, an effect previously observed in XBP1-silenced MCF-7 cells30, and co-silencing Luman and XBP1 led to a greater reduction in BFA-induced PRNP expression than silencing each individually. However, the inability of Luman and XBP1 co-silencing to completely abrogate BFA-induced PRNP implies that other factors are implicated in this regulation.
The discovery that Luman regulates PRNP expression brings us closer to achieving modulation of ER stress-induced PrP levels, by providing a novel target for pharmacological intervention. Because (1) PrP levels influence the progression of prion diseases1,2,3, (2) protects against apoptosis in a broad range of cancers45,46,47, and (3) ER stress is detected in both prion disease48 and solid tumors49,50, pharmacological attenuation of ER stress-induced PRNP expression could constitute a promising strategy in the treatment of these disorders. The recent discovery that Ceapins can specifically inhibit ATF6α proteolytic activation, a target previously perceived as “undruggable”, suggests that similar inhibitors could also be developed against Luman51.
The identification of Luman as a regulator of PRNP expression also helps understand the physiological purpose of PrP. In this study, we excluded three potential functional implications of Luman-regulated PRNP expression: promoting ERAD, protecting against proteasomal inhibition-induced cell death, and mediating atorvastatin-induced neuritogenesis. Although Luman regulates the expression of ERAD-related EDEM37, HERPUD136, Canx and Ubxn438, neither CRISPR/Cas9-mediated disruption, nor over-expression of PrP influenced the degradation rate of the ERAD substrate TTRD18G. Our data also show that neither CRISPR/Cas9-mediated disruption, nor siRNA-silencing of PRNP expression significantly altered epoxomicin- or staurosporine-induced Caspase 3/7 DEVDase activity. This contrasts with the ability of PrP to protect against ER stress-induced cell death10, but confirms previous results from our lab showing that PrP does not attenuate staurosporine-induced cell death15, and signifies that the ability of PRNP to protect against cell death is not universal, and may be influenced by the nature of the cell death stimulus. From a therapeutic standpoint, the inability of PrP to protect against proteasomal inhibition-induced Caspase 3/7 DEVDase activity implies that up-regulation of PRNP expression would not contribute to chemoresistance in cancers treated with proteasomal inhibitor bortezomib, such as multiple myeloma or mantle cell lymphoma. Lastly, this study investigated whether atorvastatin stimulates neuritogenesis through Luman-mediated PRNP expression. Our data confirm that atorvastatin induces neuritogenesis, and increases PrP levels in N2a neuroblastoma41. However, siRNA-mediated silencing of Luman does not significantly reduce atorvastatin-induced neuritogenesis, nor does atorvastatin trigger Luman activation, thereby excluding Luman induced PRNP expression as a mediator of atorvastatin-induced neuritogenesis. The exact function of Luman-mediated up-regulation of PRNP gene expression thus remains to be discovered remains a difficult task in the absence of a clear function for Luman.
Overall, this study describes an unprecedented induction of LUMAN transcription by BFA that reflects a role in long-term cell adaptation, and refines our understanding of UPR-mediated PRNP expression by identifying Luman as a novel mediator of BFA-induced PRNP expression. This regulation offers a new pharmacological target to attenuate PRNP expression, and designates PrP as an effector of Luman function. Although, this function remains unclear, current results exclude the regulation of PRNP expression by Luman as a mean to facilitate ERAD, protect against proteasomal inhibition-induced apoptosis or promote atorvastatin-induced neuritogenesis.
HEK293T, N2a, and SK-N-SH cells (from ATCC, Manassas VA) were cultured in DMEM, MCF-7 in RPMI1640. CR7 cells, which express high endogenous levels of PrP, were derived from a human glioblastoma, cultured in OptiMEM (Gibco Life Technologies, NY), and obtained from Dr Melinda Estes (Cleveland Clinic, Cleveland, OH)52. All culture media were supplemented with 10% fetal bovine serum. Neurons and astrocytes were prepared as previously described53, with the ethical approval of the McGill University Institutional Review Board. Pharmacological induction of ER stress was achieved using BFA, thapsigargin (Th) or tunicamycin (TM) at a final concentration of 5μg/mL for all three drugs, doses that maximize ER stress-induced PrP gene expression without significant toxicity to the cells, as described previously15. Unless specified, atorvastatin, epoxomicin, staurosporine and cycloheximide were used at a final concentration of 20μM, 0.5μM, 0.25μM and 75μg/mL, respectively. All drugs were dissolved in dimethyl sulfoxide (DMSO). For all experiments, the final DMSO concentration did not exceed 0.1%.
Total RNA was purified using Trizol (Invitrogen, Carlsbad CA) and reverse transcription was performed on 1μg of total RNA using AMV-RT, poly-dT, RNAase inhibitor and the following protocol: 10min at 25°C, 60min at 42°C, 5min at 99°C, and 5min at 4°C. The resulting cDNA was used as template for PCR amplification. The sequences of the primers used for PCR amplification are listed in Table 1. The PCR protocol for all OASIS family members and mouse Hprt1 was 1 cycle of 5min at 95°C, 35 cycles of 30sec at 95°C, 1min at 58°C, and 30sec at 68°C, followed by 1 cycle of 1min at 68°C. Identical conditions were used for human HPRT1 amplification, except that an annealing temperature of 62°C was used instead of 58°C. PRNP was amplified using 1 cycle of 5min at 95°C, 35 cycles of 30sec at 95°C, 1min at 60°C, and 2min at 72°C, followed by 1 cycle of 4min at 72°C. Amplification of MAP2 was achieved using 1 cycle of 5min at 95°C, 40 cycles of 30sec at 95°C, 1min at 61°C, and 1min at 68°C, followed by 1 cycle of 2min at 68°C. A similar protocol was used for GFAP, but required 25 cycles and an annealing temperature of 66.1°C. A protocol of 1 cycle of 5min at 95°C, 30 cycles of 30sec at 95°C, 1min at 60°C, and 30sec at 68°C, followed by 1 cycle of 1min at 68°C was used to amplify sXBP1.
qPCR was performed using SYBR Green Taq Mastermix (Quanta Biosciences, Gaithersburg MD) on an Applied Biosystems 7500Fast qPCR apparatus (Invitrogen, Carlsbad CA). Output data was expressed as fold-induction over control condition following normalization to HPRT1, using Pfaffl’s method54. The qPCR primer sequences used for PRNP, LUMAN, HSPA5, and HPRT1 are listed in Table 1.
Transfection of MCF-7 or CR7 cells with either plasmid DNA or siRNA was achieved by nucleofection (Nucleofection kit V, VCA-1003, Lonza, Basel Switzerland). Briefly, 4 million cells were resuspended in nucleofection buffer with either 2μg or 4μg plasmid DNA or 300nM siRNA before nucleofecting (protocol P-020). Cells were then plated at 1 million cells per well and incubated 24h before use. Lipofectamine RNAiMAX (Invitrogen, Carlsbad CA) was used for siRNA transfection in N2a cells (37.5pmol/50,000 cells) 24h before treatment. Single scrambled (sc-37007) sequence and pools of three siRNAs targeting PrP (sc-36318), Luman (sc-37702), murine Luman (sc-37703) or XBP1 (sc-38627) were purchase from Santa Cruz Biotechnologies (Dallas, TX). Silencing of Luman in primary human neurons was performed using Accell SMARTpool scrambled (D-001910-10-20) or Luman-targeting (E-017471-00) pool of four siRNAs (Dharmacon, Lafayette, CO).
Cells were lysed in lysis buffer [50mM Tris-HCl pH 8.0, 150mM NaCl, 1% NP-40 and 0.5mM ethylenediaminetetraacetic acid (EDTA) pH 8.0, 38μg/mL 4-(2-aminoethyl) benzenesulfonyl fluoride (AEBSF), 0.5μg/mL leupeptin, 0.1μg/mL pepstatin, 0.1μg/mL N-α-p-tosyl-L-lysinechloromethyl ketone hydrochloride (TLCK)]. Samples were incubated on ice before centrifugation at 15,000g for 10min. Supernatants were collected and quantified using the bicinchoninic acid method, following the manufacturer’s guidelines (Thermo Scientific, Waltham MA). Equal amounts of protein were diluted in loading buffer (2% SDS, 5% β-mercaptoethanol, 10% glycerol, 0.01% bromophenol blue, 62.5mM Tris-HCl, pH 6.8) and boiled 5min before loading on 10 or 15% poly-acrylamide gels. Following electrophoresis, proteins were transferred to polyvinylidene fluoride membranes using a Turbo blot apparatus (BioRad, Hercules CA). Membranes were blocked in 5% milk for one hour before probing with anti-Luman (1:200, M13, kindly provided by Dr. Vikram Misra, University of Saskatchewan, Saskatoon, SK), anti-PrP (1:10,000, 3F4, PrP109-112 or 1:500 ab52604, PrP214-230, Abcam, Toronto, ON), anti-BiP (C50B12, 1:5,000, Cell Signaling, Beverly MA or H-129 (1:2,500, Santa-Cruz Biotechnology, Dallas, TX), anti-HA (1:2,500, 16B12, Covance, Princeton NJ), anti-XBP1 (M-186 1:200, Santa Cruz Biotechnology, Dallas, TX), anti-GFP (B-2 1:2,500 Santa Cruz Biotechnology, Dallas, TX), anti-Bim (Y-36 1:5,000, Epitomics, Burlingame, CA, USA) or anti-β-Actin (1:5,000, AC15, Sigma, St-Louis MO) antibodies. Detection of primary antibodies was achieved using an HRP-coupled anti-mouse (1:5,000, NA9310, GE Healthcare, Baie-D’Urfe QC) or anti-rabbit (1:5,000, P0217, Dako, Burlington ON) secondary antibody, chemiluminescent substrate (RPN2232, GE Healthcare, Baie-D’Urfe QC) and Kodak BioFilms (Kodak, San Diego CA). β-Actin was revealed using an alkaline phosphatase-coupled anti-mouse antibody (1:5,000, 115-055-003, Jackson ImmunoResearch, West Grove, PA) and the nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl-phosphate substrate.
The DNA sequence coding for HA-Luman was amplified from MCF-7 cDNA with high fidelity Pfu polymerase (Agilent Technologies, Santa Clara CA) using the forward 5′-CCG CTA AAG CTT ACC ATG GCA TAC CCA TAC GAC GTC CCA GAC TAC GCT GAG CTG GAA TTG GAT GCT GGT GA-3′ and the reverse 5′-ACG CGA GTC GAC TAG CCT GAG TAT CTG TCC TGC-3′ primers, while amplification of the sequence coding for Luman amino acid 1 to 215 was accomplished using the forward 5′-CCG CTA AAG CTT AGC ATG GAG CTG GAA TTG GAT GC-3′ and reverse 5′-ACG CGA GTC GAC TAG GCC TGG AGT TTC CTC AGT TG-3′ primers. Both pairs introduced flanking HindIII and SalI sites used for cloning into the pBud-eGFP vector55. pBud-eGFP-ΔLuman-Myc was generated by mutating the stop codon by QuikChange Site-Directed Mutagenesis (Stratagene, LaJolla, CA), using the forward primer 5′-CTG AGG AAA CTC CAG GCC TCG TCG ACA TCG ATC TTA AGC-3′ and its reverse complement. This allowed the translation of a C-terminal Myc-tag already present in the pBud-eGFP vector.
The PRNP promoter regions −538: +125 was obtained from the pGL2-PRNP538 plasmids, kindly provided by Dr. John Collinge (MCR Prion Unit, London). Briefly, the promoter region was excised using BglII and ligated into the secreted luciferase reporter pML2 plasmid (Clontech, Mountain View CA) between the BglII and HindIII sites. Site-directed mutagenesis of the pML2-PRNP538 plasmid was accomplished by QuikChange Site-Directed Mutagenesis (Stratagene, LaJolla, CA) using the following forward primers and their reverse complements: ERSE-like 5′-AAG ATG ATT TTT ACA GTC AAT GAG ATC TAG AAG GGA GCG ATG GCA CCC GCA GG-3′, in ERSEa 5′-CGG CCC TGC TTG GCA GCG CGA TCG ACT TTA ACT TAA ACC TCG GC-3′, in ERSE-II 5′-GCG CGG CAA TTG GTC ATA TGG CCG ACC TCC GCC CGC G- 3′, and in ERSEb 5′-GCG GCA ATT GGT CCC CGC ATA TGT CTC CGC CCG CGA GCG CCG-3′. To mutate the ERSEb site next to the already mutated ERSEII site the following forward primer and its reverse complement were used: 5′-GCG GCA ATT GGT CAT ATG ATA TGT CTC CGC CCG CGA GCG CCG-3′.
HEK293T were used because they are highly transfectable. Cells were seeded at 300,000 cells/well in 6-well plates and incubated overnight. Transfection was performed using the polyethylenimine (PEI, Polyscience Inc., Warrington PA) method56. Briefly, 2μg of pML2-Luciferase reporter DNA and 2μg of pBud-eGFP or pBud-eGFP-ΔLuman were co-diluted in OptiMEM (Invitrogen, Carlsbad CA) and combined to 20μg of PEI. After 20min at room temperature, the complex was added to each well in a drop-like manner. Culture media was collected after 24h, and secreted luciferase activity was assessed in duplicate for each sample using the Ready-to-glow substrate, according to manufacturer’s protocol (Clontech, Mountain View CA), and a H4 plate reader (Biotek, Winooski VT).
Guide RNAs targeting the protospacer adjacent motifs (PAM) found at the PRNP Tyr38 codon or the ERSE26 promoter region were generated by inserting the oligos 5′-ACT GGG GGC AGC CGA TAC C-3′ and 5′-ACA GTC AAT GAG CCA CGT C-3′ in the pX330-U6-CBh-hSpCas9 plasmid, respectively. pX330-U6-Chimeric_BB-CBh-hSpCas9 was a gift from Dr. Feng Zhang (Addgene plasmid # 42230). The resulting plasmids were co-transfected with a GFP-expressing plasmid, and GFP-positive cells were individually plated in 96-well plates using a BD FACS Aria Fusion cell sorter. Screening of clones in which the PRNP open reading frame was disrupted was achieved by western blot using the anti-PrP 3F4 antibody. The ERSE26 mutants were screened by restriction digest. Briefly, genomic DNA was purified as previously described57, and used as a template for RT-PCR amplification of the PRNP −536:-137 promoter region, using the forward 5′-CGG AGC GCA TTT TTC TCA TTT G-3′ and reverse 5′-GAG ATT CGC TTG AAC ACT TG-3′ primers. Amplicons were then digested using the restriction enzyme BmgBI (5′-CAC/GTC-3′). Only wild type amplicons were susceptible to BmgBI digestion. The mutation of selected clones was characterized by Sanger sequencing.
Wild type or ΔERSE26 mutant HEK293T cells were grown to seventy percent confluence in T75 flasks, and transfected with pBud-eGFP or pBud-eGFP-ΔLuman-Myc using the PEI method56. After 24h, cells were recuperated by trypsinization, rinsed with phosphate-buffered saline, and cross-linked for 10min at room temperature using 1% formaldehyde. The cross-linking reaction was stopped by bringing the solution to 0.125M glycine, before rinsing, and lysing the cells in swelling buffer (10mM Tris-HCl pH 8.0, 0.25M sucrose, 0.5% NP-40, 2mM dithiothreitol (DTT), 38μg/mL AEBSF, 0.5μg/mL leupeptin, 0.1μg/mL pepstatin and 0.1μg/mL TLCK) for 10min on ice. The nuclei were then pelleted and sonicated on ice for 20min in sonication buffer (50mM 4-(2-hydroxyethyl)-1-piperazine-ethane-sulfonic acid (HEPES) pH 7.4, 140mM NaCl, 1mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, 1% SDS, 38μg/mL AEBSF, 0.5μg/mL leupeptin, 0.1μg/mL pepstatin and 0.1μg/mL TLCK) to shear the DNA. The nuclear lysate was pre-cleared with protein G-sepharose beads coated with 1μg/mL sonicated salmon sperm nuclei (S3126, Sigma, St-Louis MO) and 1mg/mL bovine serum albumin. Immunoprecipitation was performed overnight with no antibody, normal serum IgG or anti-Myc tag antibody (1:1,000, 9B11, Cell Signaling, Beverly MA). The beads were washed twice with sonication buffer, wash buffer A (50mM HEPES pH 7.4, 500mM NaCl, 1mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, 38μg/mL AEBSF, 0.5μg/mL leupeptin, 0.1μg/mL pepstatin and 0.1μg/mL TLCK), wash buffer B (20mM Tris pH 8.0, 1mM EDTA, 250mM LiCl, 0.5% NP-40, 0.5% sodium deoxycholate, 38μg/mL AEBSF, 0.5μg/mL leupeptin, 0.1μg/mL pepstatin and 0.1μg/mL TLCK), Tris-EDTA solution (10mM Tris-HCl, 1mM EDTA pH 8.0) and then eluted in elution buffer (50mM NaHCO3, 1% SDS, 1mM EDTA and 50mM Tris-HCl pH 8.0). Cross-linking was reversed by bringing the solution to 238mM NaCl and incubating at 65°C overnight. Immunoprecipitated DNA fragments were purified using the phenol:chloroform method and used as template for PCR. The primers pairs used to amplify the PRNP ERSE26 region, HERPUD1 and ACTB promoter are listed in Table 1.
Cells were lysed in lysis buffer (50mM 4-(2-hydroxyethyl)-1-peperazineethanesulfonic acid, 0.1% 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS), 0.1mM EDTA) and the lysate protein concentration was quantified by the Bradford method58. Cell lysates (20–30μg protein) were combined with Stennicke’s buffer (20mM piperazine-N,N’-bis (2-ethanesulfonic acid) (PIPES), 30mM NaCl, 1mM EDTA, 0.1% CHAPS, 10% sucrose) containing 10μM N-Acetyl-Asp-Glu-Val-Asp-7-amido-4-trifluoromethylcoumarin (Ac-DEVD-AFC) and 10mM DTT59. The fluorogenic reaction took place at 37°C, and the fluorescence level (Excitation 380nm: Emission 505nm) was acquired every minute for an hour in a black, clear bottomed, 96-well plate, using a H4 plate reader (Biotek, Winooski VT).
The statistical significance between LUMAN or HSPA5 mRNA levels was determined using a one-way ANOVA and a Dunnett post-hoc test (compared to DMSO control). For the PRNP mRNA levels following ΔLuman over-expression or the induction of neurite formation and length by atorvastatin, significance was assessed using a unilateral student t-test assuming equal variance. In Luman-silenced cells, the induction of PRNP mRNA levels by ER stress and the induction of neuritogenesis by atorvastatin were analysed using a two-way ANOVA and Bonferroni post-hoc tests. Caspase3/7 DEVDase activity of PrP-disrupted and PrP-silenced cells was analysed using a two-way ANOVA and a Dunnett post-hoc test, when applicable. All luciferase activity experiments were analyzed using one-way ANOVAs and Dunnett post-hoc tests. For all experiments, a p-value of less than 0.05 was considered significant.
How to cite this article: Déry, M.-A. and LeBlanc, A. C. Luman contributes to brefeldin A-induced prion protein gene expression by interacting with the ERSE26 element. Sci. Rep. 7, 42285; doi: 10.1038/srep42285 (2017).
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The authors gratefully acknowledge Dr. John Collinge for the PRNP promoter construct, Dr. Vikram Misra for the M13 antibody against Luman, Dr. Ron Kopito for the pcDNA3.1(-)-TTRD18G-GFP plasmid, Dr. Randal J. Kaufman for the sXBP1-coding plasmid, and Dr. Feng Zhang for the pX330-U6- Chimeric_BB-CBh-hSpCas9 plasmid. We are grateful to the Birth Defects Research Laboratory (University of Washington, Seattle, WA, USA) for providing conceptal tissue for the culture of neurons and astrocytes, and to Dr Melinda Estes (Cleveland Clinic, Cleveland, OH) for the CR7 cell line. M.-A.D. is a recipient of Canadian Institutes of Health Research Frederick Banting and Charles Best doctoral scholarship. This work was supported by Canadian Institutes of Health Research operating grants MOP89376 and MOP102738 to A.L.B.
The authors declare no competing financial interests.
Author Contributions M.-A.D. performed and analyzed all experiments presented in this paper and wrote the manuscript. A.L.B. directed the study and wrote the manuscript.