Establishment and characterization of Prnp knockdown neuroblastoma cells using dual microRNA-mediated RNA interference

Abstract

Prion diseases are fatal transmissible neurodegenerative disorders. In the pathogenesis of the disease, the cellular prion protein (PrP^C) is required for replication of abnormal prion (PrP^Sc), which results in accumulation of PrP^Sc. Although there have been extensive studies using Prnp knockout systems, the normal function of PrP^C remains ambiguous. Compared with conventional germline knockout technologies and transient naked siRNA-dependent knockdown systems, newly constructed durable chained-miRNA could provide a cell culture model that is closer to the disease status and easier to achieve with no detrimental sequelae. The selective silencing of a target gene by RNA interference (RNAi) is a powerful approach to investigate the unknown function of genes in vitro and in vivo. To reduce PrP^C expression, a novel dual targeting-microRNA (miRdual) was constructed. The miRdual, which targets N- and C-termini of Prnp simultaneously, more effectively suppressed PrP^C expression compared with conventional single site targeting. Furthermore, to investigate the cellular change following PrP^C depletion, gene expression analysis of PrP^C interacting and/or associating genes and several assays including proliferation, viability and apoptosis were performed. The transcripts 670460F02Rik and Plk3, Ppp2r2b and Csnk2a1 increase in abundance and are reported to be involved in cell proliferation and mitochondrial-mediated apoptosis. Dual-targeting RNAi with miRdual against Prnp will be useful for analyzing the physiological function of PrP^C in neuronal cell lines and may provide a potential therapeutic intervention for prion diseases in the future.

Introduction

Prion diseases are group of fatal transmissible neurodegenerative disorders characterized by a long incubation period and broad neuropathology including spongiform encephalopathy, gliosis and neuronal loss.¹ They are classified into sporadic, inherited and acquired forms, which can spread within and between mammalian species.² The central feature of prion diseases is the conversion of a host-encoded prion protein (PrP^C) into a disease-associated isoform (PrP^Sc), which shares same amino acid sequence with PrP^C but is posttranslationally converted to the infectious form, accumulating primarily in the central nerve system.³ PrP^C is implicated in prion pathogenesis as it is required for propagation and development of prion pathology.⁴ Although loss of PrP^C function or deposition of PrP^Sc is not sufficient to cause the disease, the process of conversion into PrP^Sc may alter signal transduction, thereby giving a toxic dominant function.^5–7

There have been extensive efforts to develop prion disease models in vivo, especially Prnp knockout models for defining the physiological role of PrP^C.^8,9 These transgenic animal models are, however, expensive and time consuming.^10,11 To overcome these problems, cell culture models of prion diseases have been used for screening and elucidating the mechanism of action of anti-prion agents and to analyze biological properties of PrP^C at the molecular and cellular levels.^7,11,12 Elucidating the cellular function of PrP^C may help in deciphering mechanisms of prion pathogenesis and in devising therapeutic strategies.^5,13

Mature PrP^C translocates to the outer leaflet of the plasma membrane in proximity to raft-associated signaling molecules. It traffics in and out of lipid rafts and is involved in a diversity of molecular and physiological functions.^14–16 These properties of PrP^C suggest the possibility that several changes in Prnp knockout cell culture models may be caused by an altered expression of interlinked genes with different cellular functions. Prnp knockouts could reflect not only the missing function but also the compensation for that missing function, as gene deletions in individual cells or tissues are often accompanied by compensatory changes in gene expression patterns.^17,18

RNA interference (RNAi) refers to the use of 21–23 nucleotide short interfering RNAs (siRNAs) mediating post-transcriptional degradation or translational repression of homologous gene transcripts. Stable knockdowns can be obtained by constitutive expression of the siRNA from the host chromosome.^19,20 RNAi has been accepted as the most powerful reverse-genetics approach in mammalian cells, however, there are significant technical limitations including identifying efficient methods to design and deliver siRNA.²¹ Unlike other approaches, such as traditional gene targeting by homologous recombination, antisense vectors and catalytic RNA or DNA molecules, RNAi is an endogenous natural pathway and allows cross-species application.¹² RNAi, which acutely decreases target expression, also has the advantage of avoiding compensatory or mechanistic adaptations.¹⁸

The present study describes the development of a prion knockdown cell culture model by introducing artificial microRNA (miR) targeting Prnp into the mouse neuroblastoma cell line (N2a) as well as RK13 cells expressing mouse PrP^C. The established Prnp knock-down N2a cells were characterized by investigating the expression profiles of the genes known as PrP^C interacting and/or associating molecules, proliferation, viability and apoptotic resistance.

Results

Establishment of stable Prnp knockdown cells.

Knockdown of a gene using exogenously introduced siRNA is always transient due to a cell division and/or degradation of siRNA molecules.²² To develop stable Prnp knockdown cells, small hairpin structural artificial miRs were designed (Fig. 1A). The direction and sequence of the artificial miRs were confirmed by sequence analysis, and N2a cells were transfected with the either miR1, miR2, miRdual or miRscr constructs. The transient effects of each miR on prion protein expression were assessed by western blot analysis. The miRdual most effectively knocked down the expression of PrP^C by 75 ± 2% compared with wild-type N2a cells (p < 0.01), whereas PrP^C level in N2a cells transfected with miR1, miR2 and miRscr was decreased by 17 ± 7%, 24 ± 9% (p < 0.05) and no knockdown effect, respectively (Fig. 1B and C). To establish stable Prnp knockdown cells, the N2a cells transfected with miRdual or miRscr were selected with Blasticidin S for a month and are referred to here as N2amiRdual and N2amiRscr, respectively. The single clones were further selected by limiting dilution and clonally expanded cells were analyzed by genomic PCR, qPCR and western blot to confirm stable transformation with miR as well as transcription and expression of Prnp. Association of the miR plasmid with host chromosomal sequences was assessed by PCR of genomic DNA. N2amiRdual tested positive for the 550 bp amplicon, while N2amiRscr yielded 412 bp amplicon. No amplicon occurred with genomic DNA from wild-type N2a cells (Fig. 2A). To quantify the effect of miR on Prnp transcription, qPCR was used and transcription input of each sample was normalized using qPCR of Gapdh. The Prnp transcript was decreased by 87 ± 1% in N2amiRdual (p < 0.01), while it was 50 ± 15% lower in N2amiRscr compared with wild-type N2a (p < 0.05) (Fig. 2B). PrP^C expression was quantified by western blot. Band intensity analysis of western blot indicated that the resultant protein expression of PrP^C was decreased in N2amiRdual by 96 ± 1% and 94 ± 2% in two different cell clones compared with wild-type (p < 0.01). N2amiRscr showed a reduced expression of PrP^C by 27 ± 4% (p < 0.05) (Fig. 2C).

Figure 1.

Schematic features of miR cassettes and knockdown efficacy in N2a cells. (A) Each miR cassette in pcDNA6.2-GW/EmGFP vector. (B) Western blot of N2a cells transfected with each miR expression construct. (C) Quantitation of PrP^C expression. When miRdual was introduced into the cells, PrP^C expression was the most efficiently decreased as compared with wild-type cells (*p < 0.05, **P <0.01).

Figure 2.

Transfection of miR cassettes and Prnp knockdown efficacy at mRNA and protein level. (A) PCR of genomic DNA shows miR cassette-specific amplicons. Lane L, 100 bp DNA ladder; lane 1, expression vector with miRdual cassette; lane 2, expression vector with miRscr; lane 3, wild-type cells; lane 4, N2amiRdual; lane 5, N2amiRscr; lane 6, no template. (B) qPCR analysis of PrP^C transcript abundance showing knockdown efficiency of miRdual in mRNA levels. (C) Western blot analysis of PrP^C protein abundance. Top part is the western blot and bottom panel is densitometric evaluation. dual1 and dual2 are N2amiRdual cell clones (*p < 0.05, **p < 0.01).

Differential gene expression in stably transformed N2amiRdual cells.

We next analyzed gene expression changes in response to Prnp knockdown. These studies were undertaken in N2a cells and N2amiRdual derivative and corresponded to the quantified expression of several target genes thought to interact and/or associate with PrP^C (Table 1). For all targets, transcript inputs were normalized using Gapdh and standard curves were generated using serially diluted PCR products of each specific gene. Relative expression levels were expressed as percent compared to wild-type N2a cells. Of the seven targets selected, 670460F02Rik, Csnk2a1, Plk3 and Ppp2r2b increased in abundance (220 ± 33%, 230 ± 52%, 183 ± 10% and 519 ± 5%, respectively) in N2amiRdual cells compared with wild-type N2a cells (p < 0.01) (Fig. 3). The abundance of these transcripts also increased, but to a less degree, in N2amiRscr (121 ± 6%, 132 ± 11%, 147 ± 16% and 350 ± 21% respectively). The trend in expression patterns for all target genes was negatively correlated with the amount of PrP^C in N2amiRdual and N2amiRscr cells (Fig. 2B and C). For Mpg, the mRNA expression was increased both in N2amiRdual and N2amiRscr by 195 ± 21% and 184 ± 24%, respectively (p < 0.05) (Fig. 3). As expected, the Cdr34 and Gfap transcripts as controls were not detectable until 40 cycles of qPCR reaction in both N2amiRs and wild-type N2a cells.

Table 1.

Target genes of qPCR for characterizing the established N2amiRdual cell line

Gene Symbol	Gene Name	GenBank Accession No.	General Putative Function	Ref.
Prnp	prion protein	NM_011170	The function of PrPC is not known. PrPC is encoded in the host genome and is expressed both in normal and infected cells.
6720460 F02Rik	RIKEN cDNA 6720460F02 gene	AK049710	Unknown	30
Cdr34	Cerebellar degeneration-related antigen 1	NM_001166658	In fission yeast, acts as a mitotic inducer. In G2 it negatively regulates wee1, a mitotic inhibitor. Also has a role in cytokinesis where it required for proper septum formation.	31
Csnk2a1	casein kinase 2, alpha1 polypeptide	AK011501	Casein kinases are operationally defined by their preferential utilization of acidic proteins such as caseins as substrates. The alpha and alpha′ chains contain the catalytic site.	32
Gfap	Glial fibrillary acidic protein	AF332061	GFAP, a class-III intermediate filament, is a cell-specific marker that, during the development of the central nervous system, distinguishes astrocytes from other glial cells.	33
Mpg	N-methylpurine-DNA glycosylase	NM_010822	Hydrolysis of the deoxyribose N-glycosidic bond to excise 3-methyladenine, and 7-methylguanine from the damaged DNA polymer formed by alkylation lesions.	30
Plk3	Polo-like kinase 3 (Drosophila)	NM_013807	Serine/threonine protein kinase involved in regulating M phase functions during the cell cycle. May also be part of the signaling network controlling cellular adhesion.	30
Ppp2r2b	protein phosphatase 2 (formerly 2A), regulatory subunit B (PR 52), beta isoform	NM_028392	The B regulatory subunit might modulate substrate selectivity and catalytic activity, and also might direct the localization of the catalytic enzyme to a particular subcellular compartment.	31

Figure 3.

Quantitative PCR analysis of PrP^C interacting and/or associated molecules. 6720460F02Rik, Plk3, Ppp2r2b, Csnk2a1 transcripts were upregulated in prion knockdown N2amiRdual cells compared to wild-type N2a cells. The increased Mpg transcripts in both N2amiRdual and N2amiRscr seemed to have no correlation with decreased PrP^C (*p < 0.05, **p < 0.01).

Alteration in cellular characteristics following PrP^C knock-down.

After establishment of the prion knockdown cell line, N2amiRdual, proliferation, viability and mitochondrial-mediated apoptosis were examined. In proliferation assays, the N2amiRdual showed decreased proliferation, to 88 ± 4%, in mitochondrial dehydrogenase activity tests compared with wild-type (p < 0.05). No significant difference was observed in lactate dehydrogenase tests for cell viability (Fig. 4A). Based on the observed morphology of the cell body and neuritis, no differences were observed in N2amiRdual and N2amiRscr compared with wild-type. Both cell types expressed the GFP reporter molecules, however, the signal was stronger in miRscr cassette than in miRdual (Fig. 4B). GFP signals derived from miR expression cassette were attenuated in N2amiRdual cells, likely due to increased processing of the miRdual.

Figure 4.

Cell proliferation, viability and GFP expression test. (A) Viability and proliferation of N2a cells stably transformed with miRdual or miRscr (*p < 0.05). (B) Confocal microscopy images of clonal expanded cells. N2amiRdual and N2amiRscr showed green fluorescent signals. Cell nuclei were stained with DAPI (blue). Photos were taken at 20 doubling passages. GFP signals stably expressed over 135 doubling passages. Scale bar, 50 um.

To measure mitochondrial-mediated apoptosis induced by serum withdrawal, the expression level of mitochondria fission related protein, dynamin related protein 1 (Drp1), was quantified. Drp1 expression was increased in N2amiRdual by 112 ± 4% 24 h following apoptosis induction when compared to the level before apoptosis induction (p < 0.05). There was, however, no significant difference in the Drp1 expression level between N2amiRscr and N2a wild-type, regardless of growth condition such as serum deprivation (Fig. 5A). The level of pro-apoptotic Cytochrome C (Cyt c) in cytosol increased 215 ± 2% in N2amiRdual cells during serum deprivation compared to the level before serum deprivation (p < 0.05). Again there was no difference between N2amiRscr and wild-type N2a cells even though cells had undergone apoptotic stimulation (Fig. 5B). During apoptotic stimulation, morphological changes between each cell lines were observed by light microscopy. Neurite retraction was induced and proliferation inhibited in N2amiRdual cell 24 h post serum deprivation, while N2amiRscr and wild-type N2a cells maintained their structural integrity. The caspase 3 activity in N2amiRdual cells was increased from 1.6 pmol pNA liberated/hour at 0 h to 18.8 pmol pNA liberated/hour at 48 h post serum deprivation (p < 0.01), whereas there were no significant differences in N2amiRscr and wild-type N2a cells (Fig. 5C). The proportion of apoptotic cells bound to Annexin V was increased by 212 ± 4% in N2amiRdual at 24 h post apoptosis induction when compared to the percentage of dying cells before apoptosis induction (p < 0.05). Most N2amiRscr and wild-type N2a cells were resistant to apoptotic stimulation regardless of growth environment (Fig. 6).

Figure 5.

Apoptotic resistance to serum deprivation. Cells were cultured under conditions of serum deprivation and collected at indicated time points. (A) Detection of Drp1 expression levels analyzed by densitometric analysis of western blot. (B) Cytochrome C expression in cytosolic fractions detected by immunoassay. (C) Analysis of caspase 3 activity calculated by comparison with the free pNA. dual, N2amiRdual; scr, N2amiRscr; wt, miR non-treated wild-type N2a (*p < 0.05, **p < 0.01).

Figure 6.

Detection of apoptotic cells following serum deprivation using Annexin V-Cy3 assay. Cells were collected at the indicated time points and 5 × 10⁵ cell suspensions were treated with Annexin V. The proportion of cells showing apoptotic change was increased in N2amiRdual whereas most of cells showed the resistance against apoptosis stimulation in N2amiRscr and wild-type N2a cells (*p < 0.05).

Lentiviral transduction of RK13 expressing mouse PrP^C with miRs.

To confirm the reduction of PrP^C levels by the dual knockdown method, a stable clone of rabbit kidney epithelial cells (RK13),²³ expressing both mouse PrP^C (RK13moPrnp) and GFP was transduced with lentivirus containing either the miRdual or miRscr expression cassettes. The cells were harvested at 70% and 100% confluence and the relative abundance of PrP^C analyzed by western blot (Fig. 7A). In these analyses quantities of PrP^C and GFP were normalized to the sample with either the miRdual or miRscr at 70% cell confluency. PrP^C level was decreased in RK13moPrnp cells with miRdual (46 ± 8% and 58 ± 5%), while no significant difference in PrP^C level was observed in RK13moPrnp cells with miRscr (p < 0.05) (Fig. 7B). Besides assessing protein loading with an actin probe, the GFP also encoded within the bigenic plasmid²⁴ was evaluated by use of an anti-GFP antibody. No diminution was seen when using the “dual” lentivirus, speaking to specificity, and in fact GFP protein levels were elevated by ∼30% both in RK13moPrnp-miRdual (136 ± 4% and 139 ± 7%) and in RK13moPrnp-miRscr cells (127 ± 6% and 129 ± 4%). This effect cannot represent alleviation of a transcription interference effect from Prnp cassette within the bigenic pBUD vector (which has tandem expression cassettes), as it is seen with viruses with different effects upon PrP expression.

Figure 7.

Downregulation pattern of PrP^C in RK13-moPrnp cells harvested at different confluences. RK13-moPrnp cells were lentiviral-transduced with miR expression cassettes and the relative abundances of PrP^C and GFP processed were determined by western blot (*p < 0.05, **p < 0.01).

Discussion

The knockdown of PrP^C in N2a cells using a miR-based RNAi system provides a novel cell culture model for studying prion biology. The artificial miRs were designed to target nucleotide residues 36–56 and 668–688 of mouse PrP^C ORF region, respectively. The miRs were constructed to be expressed as a cluster in a long primary transcript driven by RNA polymerase II. This dual targeting strategy enhanced knockdown efficacy more than 3-fold compared to single site targeting method. It was reported that a single miR can bind to and regulates many different mRNA target, conversely, several different miRs can pair with and cooperatively control a single mRNA target.^25,26 Downregulation of Prnp was also observed in N2a transfected with miRscr, an unexpected result as the scrambled sequence does not target any known vertebrate gene. To better define this targeting effect of miRscr to Prnp, we examined RK13moPrnp cells, which were lentiviral-transduced with miRdual or miRscr, respectively. In the viral-tranduced cells with miRdual, the expression levels of PrP^C varied inversely with the amount of GFP. These results suggest that, rather than specific targeting by miRscr, the reduced PrP^C expression in N2amiRscr is likely due to exogenous factors through gene transfection, such as genomic insertion of foreign genes that may disrupt cellular processes.^27–29

An exploration of cellular interactors is one of the way to disclose unknown function of protein of interest.¹⁴ Transcript abundances of PrP^C interacting and/or associating proteins following stable Prnp knockdown can provide valuable information concerned with normal cellular function of PrP^C. Therefore, we selected seven target genes from previously reported microarray and protein microarray data^30–33 and transcript abundance was investigated by qPCR. Of the seven genes examined in this study, the abundance of four PrP^C interactors increased following Prnp downregulation. One of the upregulated genes in N2amiRdual is 6720460F02Rik gene, also known as CATS (computer-annotated as FAM64A), the Homo sapiens family with sequence similarity 64, member A in GenBank (NM_144526). Its transcript is abundantly expressed in proliferative state brain tissues, glioma and primitive neuroectodermal tumors, however, the precise cellular function of the protein is unknown.³⁰ FAM64A was shown to interact with PrP^C through protein microarray and coimmunoprecipitation with recombinant human PrP^C spanning amino acid residues 23–231. In addition, FAM64A is located predominantly in the nucleus, where it colocalizes with the recombinant human PrP^C.³⁰ This suggests that FAM64A/CATS/6720460F02Rik might have a role in control of cell proliferation consistent with the decreased cell proliferation observed in our studies.³⁴ The increased transcription level could be a compensatory reaction to the loss of PrP^C function associated with cell proliferation.

Plk3 is a member of the polo family of serine/threonine kinases and is also as a PrP interacting partner.³⁰ It localizes to the nucleolus and is involved in regulation of the G¹/S phase transition.^35,36 The control of cell proliferation by PrP^C may alter the expression of cell cycle-related genes, such as cyclin D1, Esp8 and CD44, in a cell type-specific way.³⁷ The increased level of Plk3 in N2amiRdual cells may be a compensatory response to a PrP^C-dependent cell proliferation event. Moreover, the overexpression of Plk3 may mediate inhibition of proliferation as a result of the loss of the anti-apoptotic function of PrP^C.^36,37 Further analysis into the mechanism of PrP^C function associated with Plk3 will be required.

Ppp2r2b encodes the neuron-specific regulatory beta subunit of the protein phosphatase 2A (PP2A) holoenzyme and the mutations in Ppp2r2b are responsible for the neurodegenerative disorder, spinocerebellar ataxia 12. Increased expression of Ppp2r2b reported to induce mitochondrial fragmentation through stimulating mitochondrial fission protein such as Drp1, which can lead to neuronal apoptosis.^38,39 Our observation of increased expression of Ppp2r2b following PrP^C knockdown was not consistent with previous study that showed the increased Ppp2r2b in a PrP^C overexpressing HEK293 cell line.³¹ It is conceivable that dysregulation of Ppp2r2b via abnormal expression of PrP^C might be a critical cause of cell death secondary to mitochondrial dysfunction. Mitochondrial fragmentation by overexpression of Ppp2r2b occurs upstream of apoptosis and was sufficient for hippocampal neuron death.³⁹ It is consistent that transfection of PrP knockout cells with a Prnp expression vector prevented typical apoptotic changes.³⁷

Casein kinase 2 (CK2) participates in the regulation of diverse cellular processes. It is a messenger-independent protein serine/threonine kinase and especially abundant in the brain. It is localized to the outer plasma membrane where PrP^C also resides.⁴⁰ The catalytic α subunit of CK2 (CK2α), encoded by the Csnk2a1 gene, is reported to bind to cytosolic non-glycosylated PrP^C.⁴¹ CK2α also binds to PP2A in mitogen-starved cells in vitro and overexpression of CK2α results in suppression of cell growth.⁴² These reports are consistent with the result of decreased proliferation in N2amiRdual cells that may cause compensatory increases of Ppp2r2b and Csnk2a1 following decreased expression level of PrP^C.

The Mpg gene encodes the base-excision repair enzyme, 3-methyladenine DNA glycosylase and has been reported as a PrP interacting protein.³⁰ Mpg plays a vital role in preserving cerebellar development and protecting mature neurons from insults by environmental genotoxins.^43,44 The upregulation of Mpg both in N2amiRdual and N2amiRscr is curious, but may be a response to the expression of double-stranded miR that has internal loops.

There was no difference observed in viability of the N2a lines but proliferation was decreased in N2amiRdual compared with N2amiRscr and wild-type cells. This is likely due to the downregulation of PrP^C rather than other exogenous factors. It has been shown that the expression level of PrP^C was positively correlated with the cell proliferation.^37,45 In agreement with other studies of prion knockout model,^46–48 newly constructed prion knockdown N2amiRdual cells were more vulnerable to apoptotic stimulation than N2amiRscr and wild-type N2a. Inhibition of Drp1 activity during apoptosis in mammalian cells inhibits mitochondrial fragmentation, Cyt c release and the rate of cell death.^39,49,50 Altering the mitochondrial fusion/fission balance towards fission by Drp1 could lead to cell death by promoting mitochondrial release of pro-apoptotic proteins including Cyt c.³⁹ The release of Cyt c from mitochondrial inter-membrane space to the cytosol has reported to activate the caspase family,^51,52 an important event for the downstream activation of apoptosis cascade.^49,53 This phenomenon is supported by Cyt c knockout cell lines study showing reduced caspase 3 activation and resistance to various apoptotic stimuli.⁵⁴

In conclusion, the stable N2amiRdual cells will be a useful tool for elucidating the physiological function of PrP^C and neuropathogenesis of prion disease. Moreover, since miR is known as more suitable approach for achieving RNAi in mouse brain, in terms of toxicity,⁵⁵ it is expected that the dual targeting artificial miR-based RNAi used in this study will provide a promising therapeutic intervention for prion diseases and other neurodegenerative disorders.

Materials and Methods

Cell culture.

N2a and RK13moPrnp cells were maintained in Dulbecco's modified Eagle's Medium (DMEM) with high glucose (4.5 g/l) and 2 mM glutamine (Gibco^®, Invitrogen, #11995), supplemented with 10% fetal bovine serum, penicillin and streptomycin. Cells were cultured at 37°C in a 5% CO₂ incubator.

Designing artificial miRNAs.

miRs were designed to knockdown the endogenous Mus musculus Prnp in N2a cells based on previous report in reference ⁵⁶. The miR design algorithm, provided by the Invitrogen web-site,⁵⁷ was used to select target sequences. The basic local alignment search tool (BLAST) was used with the mouse genome database to identify unique regions in the murine Prnp open reading frame (ORF, NM_011170). Two regions were chosen to target; the N-(miR1) and C-termini (miR2) of the Prnp ORF. miR1 and miR2 bind at 191 to 211 and 823 to 843 of Prnp ORF region, respectively. Each single-stranded DNA (ssDNA) was designed to contain the antisense sequence derived from target sites, followed by 19 nucleotides to form the terminal loop and sense target sequence with 2 nucleotides removed to create an internal loop (Table 2).

Table 2.

Sequences of artificial miRs targeting Mus musculus Prnp ORF

			Sequence
		L	Mature miRNA Sequence	Loop Sequence	Sense Sequence	L
miR1	T	TGC TG	ACA TCA GTC CAC ATA GTC ACA	GTT TTG GCC ACT GAC TGA C	TGT GAC TAT GGA CTG ATG T
	B	C	TGT AGT CAG GTG TAT CAG TGT	CAA AAC CGG TGA CTG ACT G	ACA CTG ATA CCT GAC TAC A	GTC C
miR2	T	TGC TG	ATC TTC TCC CGT CGT AAT AGG	GTT TTG GCC ACT GAC TGA C	CCT ATT ACC GGG AGA AGA T
	B	C	TAG AAG AGG GCA GCA TTA TCC	CAA AAC CGG TGA CTG ACT G	GGA TAA TGG CCC TCT TCT A	GTC C
miRscr	T	TGC TG	AAA TGT ACT GCG CGT GGA GAC	GTT TTG GCC ACT GAC TGA C	GTC TCC ACG CAG TAC ATT T
	B	C	TTT ACA TGA CGC GCA CCT CTG	CAA AAC CGG TGA CTG ACT G	CAG AGG TGC GTC ATG TAA A	GTC C

L, linker; T, top strand; B, bottom strand.

Construction and expression of miRNAs.

The miRs were synthesized commercially (Invitrogen) and annealed for directional cloning. The synthesized top- and bottom-strand DNAs (Table 2) were annealed and ligated into pcDNA6.2-GW/EmGFP-miR (Invitrogen, #K4936-00) expression vector. The ligation mixture was transformed into TOP10 competent E. coli. To confirm the orientation and DNA sequence of the miRs, each construct was sequenced using an automated DNA Sequencer (ABI PRISM 377 L). To express miRs as one primary transcript, miR2 cassette were digested and ligated into miR1 construct. The linked miRs (miRdual) was transformed and analyzed as described above. A miR cassette containing a scramble sequence (miRscr) (Invitrogen, #K4936-00), which can be processed into mature miR but is not predicted to target any known vertebrate gene, was used as a negative control (Fig. 1A).

Transfection of N2a cells with miRNAs.

The wild-type N2a cell line was transfected with the expression construct using the FuGENE6 transfection reagent (Roche, #11814443001). At 24 h post transfection, the cells were lysed with RIPA lysis buffer (1% Triton X-100, 1% sodium deoxycholate, 150 mM NaCl, 50 mM Tris-HCl, pH 7.4, 0.1% SDS, 1 mM EDTA) containing a protease inhibitor cocktail (Amresco, #M222) and the efficiency of downregulation of Prnp was measured by western blot. Through drug selection with Blasticidin S (Invitrogen, #R21001) and limiting dilution, clonal expanded cells expressing miRdual (N2amiRdual) or miRscr (N2amiRscr) were obtained and analyzed by expression of green fluorescent protein (GFP) reporter gene, genomic PCR, quantitative PCR (qPCR) and western blot.

Transfection of RK13 cells with mouse Prnp.

RK13 cells were transfected with a derivative of the bigenic pBud mammalian vector expressing both GFP and a wt mouse Prnp coding region. This procedure used lipofectamine2000 (Invitrogen, #11668). Cells stably expressing mouse PrP^C (RK13moPrnp) were selected with Zeocin (Invitrogen, #R250) and screened by western blot.

Production of lentivirus and infection of RK13moPrnp.

The miR expression cassettes including GFP were cloned into pLenti6/V5-DEST lentiviral vectors (Invitrogen, #K4937) by Gateway site-specific recombination. The constructed lentiviral vectors (3 µg) were co-transfected with a third-generation lentivirus packaging vector (9 µg, Invitrogen, # K4970) into HEK293FT cells using Lipofectamine 2000 reagent (Invitrogen, #11668). The viral supernatant was collected 48 h after co-transfection and concentrated by ultracentrifugation for 2 h at 153,725 g and 4°C. The titer of lentiviruses was determined using HT1080 cells and RK13moPrnp cells were infected with the virus particles at an MOI of 10. After Blasticidin S selection for two weeks, cells were analyzed for knockdown of PrP^C by western blot.

Genomic PCR.

Genomic DNA of clonal expanded cells was isolated using a genomic DNA purification kit (Promega, #A11120). Primer pairs were designed to amplify specific miR cassettes (Table 3). The PCR parameters consisted of an initial denaturation step at 95°C for 5 min followed by 40 cycles of denaturation at 95°C for 30 sec, annealing at 60°C for 30 sec and extension at 72°C for 1 min and a final extension at 72°C for 15 min. PCR products were analyzed by electrophoresis on 1.0% agarose gel.

Table 3.

Primer sequences used in this study

Target Gene	Purpose	Label		Sequence (5′ to 3′)	Size	Product Size
miR cassette	PCR	-	F	TGC TGC TGC CCG ACA ACC ACT ACC TGA GCA	30	412
		-	R	ATC AGC GAA CCG CGG GCC CTC TAG ATC AAC	30
6720460 F02Rik	PCR	-	F	GAT CAC TCC TAA AGG AGG AAC AGC TAG AG	29	643
		-	R	ATA AGG GAG ACG GTC ATG TCA CCA C	25
	qPCR	FAM	F	CGT TAC TCA GCC TGA ACC TCG GTA A	25	68
		-	R	CTG CTC TGG CTC TGG ACT TTC C	22
Cdr34	PCR	-	F	ATG CTG GCA GAT AAC CTA GTG GAG G	25	1,240
		-	R	AAC TGA TGG ACT CTG GGC TTA CAG G	25
	qPCR	FAM	F	CGT GGA GAG CTA CTG AAG AAG TGC CA	26	67
		-	R	AGG TCT GCA CAG CCT TGT GTG	21
Csnk2a1	PCR	-	F	AAG CAG GGC CAG AGT TTA CAC AGA T	25	1,137
		-	R	GCT GGA ACA GGT ATC CCA AGT GAG T	25
	qPCR	FAM	F	CGA AGT TTC TGG ACA AGC TGC TT	23	104
		-	R	AGC CTG GTC CTT CAC AAC AGT G	22
Gfap	PCR	-	F	ACG CTT CTC CTT GTC TCG AAT GAC T	25	1,178
		-	R	GCT CCT GCT TCG AGT CCT TAA TGA C	25
	qPCR	FAM	F	CGG ATA CTT TCT CCA ACC TCC AGA TC	26	77
		-	R	CCT CTT GAG GTG GCC TTC TGA C	22
Mpg	PCR	-	F	GGG CAG AGG ATC CCT AAA ACC GGT G	25	942
		-	R	CAG GCT GTT TGC TGA GGC TGA TCC A	25
	qPCR	FAM	F	CGA CAG CCC TAA AGA GAG ACT CCT GT	26	74
		-	R	GGT CCT CTG GGC TGG AGA AGT	21
Plk3	PCR	-	F	CTA TGA AGC CAC TGA CAC CGA GTC T	25	1,629
		-	R	GAA GCG AGG TAA GTA CAA GCA CTG C	25
	qPCR	-	F	CTT GGT GAG TGG CCT CAT GC	20	75
		FAM	R	CGG AGT GAA ACT ACA GGA GCC TC	23
Prnp	PCR	-	F	CTG CTG GCC CTC TTT GTG ACT ATG T	25	736
		-	R	CGA TCA GGA AGA TGA GGA AGG AGA T	25
	qPCR	FAM	F	CGA AGC AGC AAC CAG AAC AAC TT	23	65
		-	R	ACC GTG TGC TGC TTG ATG GT	20
Ppp2r2b	PCR	-	F	AGG AGG ACA TTG ATA CCC GCA AAA T	25	1,273
		-	R	ATG TTT TCT GAA GGA TGC CAA GCT G	25
	qPCR	FAM	F	CGG TCT TCT TCA GGA TGT TCG AC	23	191
		-	R	AGG ATG CCA AGC TGT ATG CAA G	22

Western blot.

For western blot analysis, 5 µg of protein from each cell lysate was prepared and fractionated on a 12% Tris-glycine gel and transferred to a 0.2 micron nitrocellulose membrane using a dry blotting system (Invitrogen, #IB1001). The membrane was blocked with 5% nonfat dry milk in TritonX-Tris buffered saline (TTBS: 0.1% TritonX-100 in 100 mM Tris-HCl, pH 7.5, 0.9% NaCl) and then probed with anti-PrP antibody, 3F10,⁵⁸ anti-GFP antibody (AbCAM, #ab290) or anti-dynamin-related protein 1 (Drp1) antibody (BD Transduction Laboratories, #611112) at 4°C for overnight. An anti-mouse IgG conjugated to horseradish peroxidase (Santa Cruz Biotechnology, #SC-2005) was used as the secondary antibody. The bound antibodies were visualized by chemiluminescence (Amersham, #RPN2106) and protein levels were measured by scanning and evaluating density (Multi Gauge v2.3 software, Fujifilm). The membrane was stripped in stripping solution (100 mM Tris-HCl, pH 7.5, 0.9% NaCl, 7 µl/ml β-mercaptoethanol, 2% SDS) and then blocked with 5% nonfat dry milk in TTBS and probed with anti-mouse glyceraldehydes-3-phosphate dehydrogenase antibody (GAPDH, Santa Cruz Biotechnology, #SC-32233) or anti-mouse β-actin antibody (AbCAM, #ab20272) as a loading control. Protein levels were visualized as described above.

Quantitative PCR.

To quanify transcript abundance in N2amiRdual cells, genes reported to interact and/or associate with PrP^C were selected^30–33 and their abundance was determined by qPCR (Table 1). Total RNA from clonal expanded cells was extracted using Trizol reagent (Invitrogen, #15596-018) and single stranded cDNA was synthesized using SuperScript™ III First-Strand Synthesis System (Invitrogen, #18080-051) with Oligo dT following the manufacturer's instructions. ORFs corresponding to each target gene were amplified by PCR using specific primers (Table 3) and each PCR product was serially diluted and used to generate standard curves. The primers for qPCR were designed by D-LUX™ Designer (Invitrogen) (Table 3). Thermal cycling for the ABI PRISM 7300 (Applied Biosystems) was performed as follows: 2 min at 50°C for uracil N-glycosylase (UDG) incubation which removes uracil-containing products from previous reactions, UDG inactivation and DNA polymerase activation for 2 min at 95°C, followed by 40 cycles of 15 sec at 95°C for denaturation and 30 sec at 60°C for hybridization and elongation.

Proliferation and viability assay.

The established N2amiRdual, N2amiRscr and wild-type cell lines were cultured in 96-well plates with 3 × 10⁴ cells in 100 µl of DMEM culture medium per each triplicate well. Activity of cellular mitochondrial and lactate dehydrogenase were determined using commercial kits (Biovision, #302-500, #313-500) to measure proliferation and cells viability respectively. After 48 h incubation, the tetrazolium salt (WST) was added to each well and the amount of formazan production quantified using a microtiter plate reader at 450 nm. Non-viable controls were prepared by treating cells with 1% TritonX-100.

Cytochrome C assay.

The cytosolic and mitochondrial fractions from each cell line were separated using a mitochondria isolation kit (Pierce, #89874) following manufacturer's instruction. Total protein concentration of the cytosolic fractions were adjusted using BCA protein assay (Pierce, #23227) and the amount of Cyt c in each fraction were analyzed by sandwich Enzyme Linked-Immuno-Sorbent Assay (Invitrogen, #KHO1051).

Caspase 3 assay.

Caspase 3 activity in cells induced to apoptosis was measured by CaspACE Assay System (Promega, #G7220). Cells were collected at 0 h, 12 h, 24 h, 48 h post serum deprivaton and lysed with RIPA lysis buffer. Each well of a 96 well plate contained 32 µl of caspase assay buffer, 2 µl of DMSO, 10 µl of 100 mM DTT, 20 µl of cell lysates (30 µg), 34 µl of distilled water and 2 µl of the 10 mM DEVD-aminoluciferin labeled with chromophore p-nitroaniline substrate (DEVD-pNA). After 4 h incubation at 37°C, the release of pNA from the substrate by caspase 3 (DEVDase) was detected using microtiter plate reader at 405 nm.

Apoptosis assay.

Cells showing apoptotic change were detected based on the translocation of membrane phosphatidyl-serine (PS) from the inner space to the surface of the cells. When PS is displayed externally, it can be detected by PS-affinitive Annexin V which has a fluorescent conjugate. The proportion of cells undergoing apoptosis in each cell line under both normal growing condition and following serum deprivation was quantified using Annexin V-Cy3 apoptosis detection kit (Biovision, #K102-100). After adding 5 µl of Annexin V-Cy3 into each cell suspension, apoptotic cells bound to Annexin V were detected by flow cytometry (BD FACSCalibur™ Flow Cytometer, BD Bioscience) at 570 nm emission wavelengths. In parallel, control cell suspension without Annexin V-Cy3 were also analyzed.

Statistical analysis.

All experiments were performed at least three times. Statistical analysis was performed by ANOVA and LSD (SPSS Software version 17.0, SPSS Inc.). Data were presented as mean ± standard deviation. Differences were considered to be significant if probability values of p < 0.05 or p < 0.01 were obtained.