Correction: A Novel Artificial MicroRNA Expressing AAV Vector for Phospholamban Silencing in Cardiomyocytes Improves Ca2+ Uptake into the Sarcoplasmic Reticulum

In failing rat hearts, post-transcriptonal inhibition of phospholamban (PLB) expression by AAV9 vector-mediated cardiac delivery of short hairpin RNAs directed against PLB (shPLBr) improves both impaired SERCA2a controlled Ca cycling and contractile dysfunction. Cardiac delivery of shPLB, however, was reported to cause cardiac toxicity in canines. Thus we developed a new AAV vector, scAAV6-amiR155-PLBr, expressing a novel engineered artificial microRNA (amiR155-PLBr) directed against PLB under control of a heart-specific hybrid promoter. Its PLB silencing efficiency and safety were compared with those of an AAV vector expressing shPLBr (scAAV6-shPLBr) from an ubiquitously active U6 promoter. Investigations were carried out in cultured neonatal rat cardiomyocytes (CM) over a period of 14 days. Compared to shPLBr, amiR155-PLBr was expressed at a significantly lower level, resulting in delayed and less pronounced PLB silencing. Despite decreased knockdown efficiency of scAAV6-amiR155-PLBr, a similar increase of the SERCA2a-catalyzed Ca uptake into sarcoplasmic reticulum (SR) vesicles was observed for both the shPLBr and amiR155-PLBr vectors. Proteomic analysis confirmed PLB silencing of both therapeutic vectors and revealed that shPLBr, but not the amiR155-PLBr vector, increased the proinflammatory proteins STAT3, STAT1 and activated STAT1 phosphorylation at the key amino acid residue Tyr701. Quantitative RT-PCR analysis detected alterations in the expression of several cardiac microRNAs after treatment of CM with scAAV6-shPLBr and scAAV6-amiR155-PLBr, as well as after treatment with its related amiR155and shRNAs-expressing control AAV vectors. The results demonstrate that scAAV6-amiR155-PLBr is capable of enhancing the Ca transport function of the cardiac SR PLB/SERCA2a system as efficiently as scAAV6-shPLBr while offering a superior safety profile. Citation: Gröbl T, Hammer E, Bien-Möller S, Geisler A, Pinkert S, et al. (2014) A Novel Artificial MicroRNA Expressing AAV Vector for Phospholamban Silencing in Cardiomyocytes Improves Ca Uptake into the Sarcoplasmic Reticulum. PLoS ONE 9(3): e92188. doi:10.1371/journal.pone.0092188 Editor: R. Jude Samulski, University of North Carolina at Chapel Hill, United States of America Received September 12, 2013; Accepted February 19, 2014; Published March 26, 2014 Copyright: 2014 Gröbl et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This work was supported by the Deutsche Forschungsgemeinschaft through SFB Transregio 19 project grants to HF, UV and RV. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: HF and JK submitted a patent application entitled ‘‘Polynucleotide sequence for the inhibition of phospholamban synthesis’’ (Patent application number: 20100310521, Publication date: 2010-12-09). HF and WP submitted a patent application entitled ‘‘RNA Interference for the Treatment of Heart Failure’’ (Patent application number: 20110098338, Publication date: 2011-04-28). This does not alter adherence to all the PLOS ONE policies on sharing data and materials, as detailed online in the guide for authors of PLOS ONE. * E-mail: henry.fechner@tu-berlin.de . These authors contributed equally to this work.


Introduction
Heart failure is the leading cause of mortality and morbidity in Western countries and a common endpoint of cardiac disorders, including atherosclerosis, ischemic cardiomyopathies, familiar cardiomyopathies, valvular-induced myocardial pathologies and arterial hypertension [1,2]. A common feature of heart failure is an altered Ca 2+ homeostasis, which is a consequence of dysregulation of Ca 2+ cycling proteins. The sarcoplasmatic reticulum (SR) Ca 2+ -ATPase (SERCA2a) plays a key role in regulating normal Ca 2+ levels in cardiomyocytes. It is the major contributor to lowering cytosolic Ca 2+ levels and reloading the SR with Ca 2+ during each contraction-relaxation cycle [3]. Reduced SERCA2a expression and/or activity are at least partly responsible for dysregulation of cellular Ca 2+ homeostasis in heart failure [4], and enhancement of Ca 2+ re-uptake activity of the cardiac SR by gene transfer of SERCA2a has been shown to improve contractile dysfunction in rodent and large animal heart failure models [5,6]. Furthermore, recent phase I and II clinical trials have confirmed the safety and clinical benefit of AAV1 vector-mediated SERCA2a gene therapy in patients with terminal heart failure [1].
The Ca 2+ affinity and transport activity of SERCA2a are regulated by the SR protein phospholamban (PLB). Nonphosphorylated PLB keeps the Ca 2+ affinity of SERCA2a low and reduces Ca 2+ re-uptake into the SR, whereas PLB phosphorylation in response to b-adrenergic stimulation reverses this inhibition. The well-known regulatory role of PLB for the SERCA2a-catalyzed Ca 2+ -transport of the cardiac SR has qualified this small membrane protein as a promising molecular target for heart failure therapy. In this context, gene therapeutic approaches for inhibiting PLB expression by employment of dominant-negative PLB mutants [7], intracellular inhibitory antibodies targeting PLB [8] or engineered zinc-finger protein transcription factors (ZFP TFs) to the endogenous PLB gene [9] have been shown to improve the SERCA2a-catalyzed Ca 2+ transport activity in cardiomyocytes and the contractile function of the heart in animals models of heart failure. An alternative strategy to reduce expression of PLB comprises using RNA interference (RNAi) technology [10][11][12][13][14]. RNAi represents a mechanism of post-transcriptional gene silencing induced by short doublestranded (ds) RNAs with length of 21-22 nucleotides [15]. The short dsRNAs are incorporated into the RNA-induced silencing complex (RISC) in the cytoplasm. One strand, the guide strand, binds to the complementary target sequence in the transcript leading to its degradation [16]. For therapeutic use, short dsRNAs can be applied as synthetic small interfering (si)RNAs, short hairpin (sh)RNAs or artificial microRNAs (amiR) [17]. After transfection, siRNAs directly target the corresponding mRNA in the cytoplasm [18]. While siRNAs are commonly used for in vitro investigations, their delivery for in vivo application is challenging with regard to target organ-specificity, transfection efficiency and long-term efficacy. An alternative is the expression of shRNAs, which contain both a sense and an antisense sequence that are connected by a loop of unpaired nucleotides. They are expressed from vectors with strong, constitutively active polymerase III promoters (U6, H1, and 7SK). After transcription, shRNAs are exported from the nucleus into the cytoplasm via the nuclear karyopherin exportin-5 and are cleaved subsequently by the RNase Dicer into functional active siRNAs [19]. The high efficiency of shRNAs in knocking down genes in vivo has been widely demonstrated [12,20,21]. However, shRNAs were found to induce cellular innate immune responses resulting in enhanced production of interferons (IFNs). The released IFNs can subsequently activate IFN-stimulated genes, resulting in cellular gene expression profile changes that may contribute to adverse sideeffects such as global degradation of mRNA, inhibition of general protein translation or cell death [22][23][24][25]. Moreover, long-term and high-level shRNA expression can result in oversaturation of the endogenous cellular microRNA (miR) pathways leading to cytotoxicity and eventually fatality, as reported previously [26][27][28][29].
The amiRNAs represent a type of shRNAs in which a siRNA sequence is embedded into a native microRNA scaffold [30], most commonly into that of miR-30 [30,31] or miR-155 [32,33]. In contrast to shRNAs, amiRs are expressed from polymerase II promoters and are processed consecutively into functionally active siRNAs via the nuclear class 2 RNase III enzyme Drosha and the cytoplasmic endoribonuclease Dicer [34]. Several studies have compared shRNAs and amiRs regarding their efficiency and safety. The general trend of these studies is that amiRs substantially reduce the toxicity observed for the shRNAs of the same constructs [27][28][29]. The efficiency of gene silencing mediated by amiRNAs has been reported to be higher [27,35,36], comparable [28,37] or lower [38] in comparison to shRNAs.
We demonstrated the first successful treatment of experimental heart failure by RNAi in vivo [12]. An AAV9 vector and adenoviral vector-mediated cardiac expression of rat PLB-specific shRNA (shPLBr) resulted in strong cardiac down-regulation of PLB. This led to restoration of the compromised left ventricular contractile function. It also significantly reduced cardiac dilatation, hypertrophy, cardiomyocyte diameter and cardiac fibrosis. Importantly, no adverse side effects were observed using this shPLBr gene therapeutic approach in rats. More recently, however, Bish et al. [14] evaluated the efficiency and safety of an adapted shPLB in healthy dogs. They found that AAV6 vector-mediated cardiac shPLB delivery to these animals resulted in the strong knockdown of PLB, accompanied by severe cardiac toxicity.
In our present study we compared the PLB silencing efficiency and the safety profile of a newly developed AAV vector (scAAV6-amiR155-PLBr) expressing a novel engineered artificial micro-RNA (amiR155-PLBr) directed against PLB with those of a previously described AAV vector (scAAV6-shPLBr [12]) expressing shPLBr in neonatal rat cardiomyocytes (CM). Compared to shPLBr, amiR155-PLBr expression was delayed and lower. A comparable increase of the SERCA2a-catalyzed Ca 2+ transport activity, however, was observed after long term treatment of CM with the respective AAV vectors. MicroRNA and proteome analysis detected similarities but also differences between amiR155-PLBr and shPLBr vectors regarding undesirable offtarget effects. scAAV6-shPLBr, but not scAAV6-amiR155-PLBr, caused elevation of proinflammatory proteins such as STAT3, STAT1 and phosphorylated STAT1. Both vectors, as well as control RNAi vectors, induced alterations of microRNA expression levels in cardiac cells. Thus, scAAV6-shPLBr and scAAV6-amiR155-PLBr display comparable functional efficiency, but the safety profile of the amiR155-PLBr vector is superior to that of shPLBr vector.

Cell culture
Human embryonal kidney (HEK) 293 cells were cultured in Dulbecco's Modified Eagle's Medium (DMEM) (Life Technologies, Gibco, Darmstadt, Germany) supplemented with 10% fetal calf serum (FCS) and 1% of each penicillin/streptomycin (Sigma Aldrich Chemie GmbH, München, Germany). Primary neonatal rat CM were isolated from ventricular tissue of 1-3-day-old Wistar rat pups (Charles River, Sulzfeld, Germany) and cultured in CMRL 1415 medium (Biochrom AG, Berlin, Germany) supplemented with, 5.4 mM KCl, 1.26 mM CaCl 2 , 10 mM Hepes, 2 mM 59-Fluoro-29-deoxyuridine (Sigma-Aldrich), 10% FCS and 2 mg/ml of gentamycin (Biochrom AG) as described previously [11,39]. The seeding cell density using 6-well Nunc TM Cell-Culture dishes with surface coating Nonclon Delta (Cat-No. 140685, Thermo Fisher Scientific; Rockford, IL, USA) was 1.25610 5 cells/cm 2 . Hearts were removed from decapitated rat pups and used for isolation of CM in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). Decapitation and removal of hearts was approved by the Landesamtes für Gesundheit und Soziales (LAGeSo) in Berlin, Germany under the Permit Number: O 0139/05.

Plasmids
Generation of amiR155 expression cassettes was performed with the Block-iT Pol II miR RNAi Expression Vector Kit according to the recommendations of the supplier (Life Technologies, Invitrogen, Darmstadt, Germany). Briefly, a double stranded (ds) oligonucleotide encoding murine miR155 pre-miRNA containing the siPLB-17 sequence [11] was generated by annealing of the primers 59-TGCTGTAGCCGAGCGAGTAAGGTATTGTTTTGGCCA-CTGACTGACAATACCTTTCGCTCGGCTA-39and 59-CTGTAGCCGAGCGAAAGGTATTGTCAGTCAGTG-GCCAAAACAATACCTTACTCGCTCGGCTAC-39.
The reporter plasmid necessary to prove silencing activity of shPLBr and amiR155-PLBr was generated as follows. Total RNA was isolated from CM with TRIzol Reagent (Life Technologies, Invitrogen) according to the recommendation of the supplier and reverse transcribed using the High Capacity cDNA Reverse Transcription Kit (Life Technologies, Applied Biosystems Inc., Darmstadt, Germany). The cDNA of rat PLB was amplified by PCR using the primer pair NruI-174s 59-CGACTCGCGAATG-GAAAAAGTCCAATACCT-39 and XbaI-332as 59-CAGCTC-TAGAGTCACAGAAGCATCACAATGA-39. The resulting PCR fragment containing an NruI site at the 59end and a XbaI site at the 39end was inserted into the pGEM-T vector with the pGEM-T Easy Vector System II (Promega GmbH, Mannheim, Germany). The vector was digested with NruI/XbaI and the resulting fragment was inserted into NruI/XbaI-digested luciferase reporter plasmid pUF-CMV enh /MLC0.26-Luc [40]. The reporter plasmid was termed pLuc-PLBr_cDNA. The correctness of the plasmids was verified by sequence analyses using an ABI 310 Genetic Analyzer (Life Technologies, Applied Biosystems Inc.).

AAV vectors
For the production of pseudotyped self-complementary (sc) AAV6 vectors HEK293 cells were co-transfected with the respective AAV shuttle plasmids pscAAV-shPLBr, pscAAV-shCon, pscAAV-amiR155-PLBr, and pscAAV-amiR155-Con and the AAV packaging plasmid pDP6 (Plasmid Factory GmbH & Co. KG, Bielefeld, Germany) providing the AAV2 rep genes, AAV6 cap genes and adenoviral helper functions. The procedure for vector generation was carried out as described previously [43]. Purification of the AAV vectors was carried out by an iodixanol step gradient centrifugation [44]. Titers of shPLBr and shCon containing AAV constructs were determined by real-time PCR using the U6 promoter specific primers 59-AAAACTGCAAAC-TACCCAAGAAA-39 and 59-AAGGTCGGGCAGGAAGAG-39 and a 6-FAM-TGCAAATATGAAGGAATCATGGGAAA-BHQ-1 labeled probe. Titers of amiR155-PLBr and amiR155-Con containing constructs were determined by real-time PCR with the primers 59-TGCCCAGTACATGACCTTATGG-39 and 59-GAAATCCCCGTGAGTCAAACC-39 and a 6-FAM-AGT-CATCGCTATTACCATGG-BHQ-1 labeled probe (Metabion GmbH, Martinsried, Germany) which were directed against a target sequence located in the CMV promoter enhancer present in the amiR155-PLBr and amiR155-Con bearing AAV vector genomes. Real Time PCR quantification was carried out with the StepOnePlus Realtime PCR System (Life Technologies, Applied Biosystems Inc.).

Plasmid transfection
HEK293 cells were seeded in 24-well plates at a density of 2610 5 cells/well. After 24 h cells were transfected with 50 ng of the plasmid expressing Firefly luciferase, 75 ng of the plasmids expressing shPLBr, amiR155-PLBr, shCon, amiR155-Con and amiR155-Luc, respectively, 5 ng of a plasmid encoding Renilla luciferase for standardisation and 800 ng of an unrelated carrier plasmid containing a green fluorescent protein using polyethylenimine solution (Sigma-Aldrich) as transfection reagent. Analysis of reporter gene expression was carried out 72 h after transfection.

Luciferase assay
Firefly and Renilla luciferase activity were measured using the Dual-Luciferase Reporter Assay (Promega GmbH) according to manufacturer's instructions. Luciferase activity was measured in a Lumat LB 9507 luminometer (Berthold Technologies, Bad Wildbad, Germany).

Transduction of cardiomyocytes with AAVs
CM were seeded in 6-well-(1.2610 6 cells per well) or 24-well-(3610 5 cells per well) cell culture plates. After two days, the medium was replaced by fresh medium and the cells were transduced with the AAV6 vectors. At this time point cells had a confluency of about 70-80%. Twenty four hours later the cell culture medium was changed and the cells were washed once with PBS and further incubated with fresh medium until the cells were harvested.

qRT-PCR
Quantification of processed shPLBr and amiR155-PLBr was carried out using two-step RT-PCR as described previously [45] with some modifications. Total RNA was isolated from CM as described and 1 mg of RNA was digested with peqGOLD DNaseI. 250 ng of DNaseI-digested RNA was reverse transcribed using the stem-loop RT primer 59-GTCGTATCCAGTGCAGGGTCC-GAGGTATTCGCACTGGAT ACGACAATACC-39 and the High-Capacity cDNA Reverse Transcription Kit (Life Technologies, Applied Biosystems Inc.). Mature shPLBr and amiR155-PLBr were amplified using the primers 59-GTGCAGGGTCC-GAGGT-39 and 59-GCGGCTAGCCGAGCGAGTAA-39. For normalization of equal RNA load the expression of the 18 S rRNA was determined by quantitative RT-PCR using the primer pair 59-CGCGGTTCTATTTTGTTGGT-39 and 59-AGTCGG-CATCGTTTATGGTC-39. Real-time PCR reactions were carried out with the SYBR Green PCR Master Mix (Life Technologies, Applied Biosystems Inc.) and StepOnePlus Real-Time PCR System (Life Technologies, Applied Biosystems Inc.). Each sample was tested in triplicate.
To analyse PLBr gene expression 1 mg of total RNA was isolated from CM and digested with peqGOLD DNaseI. 250 ng of DNaseI-digested RNA were reverse transcribed using the High Capacity cDNA Reverse Transcription Kit (Life Technologies, Applied Biosystems Inc.). Expression levels were determined by real-time PCR (TaqMan Gene Expression AssayRn01434045_ m1, Life Technologies, Applied Biosystems Inc.). Rat GAPDH gene expression was used for normalization (TaqMan Gene Expression Assay Rn99999616_m1, Life Technologies, Applied Biosystems Inc.).
PCR reactions for miRs and gene expression analyses were carried out in triplicate with TaqMan Universal PCR Master Mix (Life Technologies, Applied Biosystems Inc.) and the StepOnePlus Real-Time PCR System (Life Technologies, Applied Biosystems Inc.).

Northern blot analysis
Northern blot hybridization was carried out as described previously [11]. Hybridized filters were exposed to Kodak BioMax MR film (Sigma Aldrich).
For detection of total and phosphorylated STAT1 and STAT3, 50 mg of protein lysates of three biological replicates were boiled for 3 min at 95uC, separated on a 7.5% SDS polyacrylamide gel and subsequently transferred onto a nitrocellulose transfer membrane (Schleicher&Schüll, Dassel, Germany). After blocking with 5% non-fat milk powder dissolved in TBS-T, the membranes were incubated with the following primary rabbit monoclonal antibodies at 4uC for 16 h: anti-total STAT1 (1:1,000), anti-total STAT3 (1:1,000) anti-phosphorylated STAT1-Tyr701 (1:500), anti-phosphorylated STAT3-Tyr705 (1:1,000) and anti-phosphorylated STAT3-Ser727 (1:1,000). All primary STAT antibodies were purchased from Cell Signaling Technology (Danvers, MA, USA). For normalization, we also analysed the expression of GAPDH as loading control using anti-GAPDH (1:3,000, 1 h at room temperature, mouse monoclonal, Biodesign Int., Meridian Life Science, Memphis, TN, USA). After incubation with primary antibodies blots were washed extensively with TBS-T followed by incubation with the secondary peroxidase-conjugated sheep antirabbit or sheep anti-mouse antibodies (both from Dako Deutschland GmbH, Hamburg, Germany) for 1 h at room temperature. Signals were visualized using ECL Western Blotting Detection Reagent (GE Healthcare Europe GmbH, Freiburg, Germany) and densitometric analysis was performed using Quantity One (Bio-Rad Laboratories GmbH, Munich, Germany). The relative optic densities of the specific bands were calculated and normalized to GAPDH.

Preparation of cardiomyocyte homogenates
At different time points after transduction of the CM, medium was removed from 6-well culture plates, the cell monolayer was washed twice with 2 ml ice-cold PBS and then removed from the bottom of each well in the presence of 0.6 ml ice-cold PBS using a plastic cell scraper. The content of three wells was transferred to a 2 ml Eppendorf tube and then sedimented by centrifugation (400 6 g, 10 min, 4uC). Cell pellets were shock-frozen in liquid nitrogen and stored at -80uC until use. The pellets containing cells of three wells were homogenized in 150 ml ice-cold homogenization buffer containing 250 mM sucrose, 50 mM NaH 2 PO 4 , 10 mM NaF, 1 mM EDTA and 10 mM histidine, pH 7.4, 0.3 mM phenylmethylsulfonyl fluoride (PMSF) using a mini glass Teflon homogenizer on ice by 10 strokes. The three 50 ml aliquots of the homogenates obtained per pellet were immediately shockfrozen in liquid nitrogen and stored at -80uC until use. Protein was determined using the BCA TM Protein Assay (Pierce, Rockford, IL, USA) with bovine serum albumin as a standard.
Measurement of sarcoplasmic reticulum oxalatesupported Ca 2+ uptake Initial rates of oxalate-supported Ca 2+ uptake into SR vesicles were estimated in homogenates of cultured neonatal rat CM by a standard procedure [39]. The reaction medium contained 40 mM imidazole (pH 7.0), 100 mM KCl, 5 mM MgCl 2 , 5 mM Tris (hydroxymethyl)-aminomethane(Tris)-ATP, 6 mM phosphocreatine, 10 mM K-oxalate, 0.2 mM ethylene glycol-bis-(b-aminoethylether)-N,N,N9,N9-tetraacetic acid (EGTA), 10 mM NaN 3 , 0.025 to 0.25 mM 45 CaCl 2 (3 to 8610 11 Bq/mol), and 15-20 mg of homogenate protein per 0.2 ml. The resulting submicromolar and saturating free Ca 2+ concentrations, amounting to 0.34 mM and 3.68 mM, were calculated using Fabiato's computer program as described elsewhere [46]. After 2 min of preincubation at 37uC, the duplicate measurement was started by addition of protein. After 1, 2 and 3 min, a 50 ml sample was filtered through the 0.45 mm membrane filter ME25 (Schleicher &Schüll) using a vacuum pump. Filters were then washed twice with 3 ml ice-cold solution containing 100 mM KCl, 2 mM EGTA and 40 mM imidazole (pH 7.0). Radioactivity associated with the dry filters was determined by liquid scintillation counting. Solutions for Ca 2+ transport measurements were made with deionizised water, p.a. (Merck, Darmstadt, Germany); contaminant Ca 2+ did not exceed 0.005 mg/l. The 45 CaCl 2 was obtained from PerkinElmer (Waltham, MA, USA). Rates of transport of Ca 2+ into SR vesicles were calculated by linear regression of data points at 1, 2, and 3 min as described earlier [39] and measured in duplicate or triplicate. Specific Ca 2+ uptake activity expressed in nmol Ca 2+ / min/mg protein is defined as the rate of oxalate-supported Ca 2+ uptake related to milligram of cell homogenate protein. Rate values obtained at the submicromolar free Ca 2+ concentration were normalized to the respective rate values obtained at saturating free Ca 2+ concentration as reported earlier [11]. This approach relies on the fact both the Ca 2+ affinity and the activity of SERCA2a -at the saturating Ca 2+ concentration -are not altered by the abundance nor the phosphorylation status of phospholamban.

Sample preparation for proteomics
At appropriate time points, cells were washed twice with 2 ml PBS and subsequently harvested with 0.2 ml of a solution containing 8 mol/l urea, 2 mol/l thiourea. The cell lysates were rapidly frozen by immersion in liquid nitrogen. Proteins were extracted by five cycles of thawing and freezing, in which the cell lysates were incubated for 10 min at 30uC with vigorous shaking at 1.400 rpm, mixed by vortexing and then frozen in liquid nitrogen again. Afterwards, high molecular nucleic acids were fragmented by sonification 3 times for 3 sec at 50% power (Sonoplus, Bandelin, Berlin, Germany). The protein extract was obtained by centrifugation for 1 h at 4uC and 16.000xg. The protein concentration of the supernatant was determined with a Bradford assay kit (Thermo Fisher Scientific Inc.). Sample aliquots were stored at -80uC.
Sample aliquots of 4 mg were diluted with 20 mM ammonium bicarbonate (ABC) until a final concentration of less than 1 M urea was reached. Following reduction by 2.5 mM DTT (1 h at 60uC) and alkylation by 10 mM iodoacetamide (30 min at 37uC), sequencing grade modified trypsin (Promega GmbH) at a ratio of 1:25 enzyme to protein was added and digestion proceeded for 18 h at 37uC. The reaction was stopped with a final concentration of 1% acetic acid and peptide mixtures were desalted on C-18 reverse phase material (ZipTip m-C18, Millipore Corporation, Billerica, MA, USA). Peptides were eluted in 50% and 80% acetonitrile (ACN) each in 1% acetic acid. Pooled eluates were concentrated to 2 ml in a vacuum concentrator (5303, Eppendorf, Wesseling, Germany) and resuspended in 2% ACN in 0.1% acetic acid.

Global protein profiling by mass spectrometry and data analysis
Peptide separation was performed using a non-linear 86 min gradient of 5-40% ACN in 0.1% acetic acid at a constant flow rate of 300 nl/min on a reverse phase column (PepMap C18, 75 mm id x 15 cm, LC Packings, Idstein, Germany) operated on an Easy-nLC (Proxeon, Thermo Scientific, Dreieich, Germany). The eluate was directed into a LTQ-OrbitrapVelos mass spectrometer (ThermoElectron, Bremen, Germany) equipped with a nanoelectrospray ion source. After a survey scan in the Orbitrap analyzer, 13 high resolution MS/MS spectra were recorded in a data dependent mode with exclusion times of 60 sec.
Proteins were identified via an Elucidator web interface by an automated database search in a NCBI database (rel. 2009, limited to rat entries) using the Sequest algorithm rel. 2.7 (Sorcererbuilt 4.04, Sage-N Research Inc., Milpitas, CA, USA). Parent mass tolerance (MS) was set to 10 ppm and fragment mass tolerance 1 Da. Carbamido methylation of cysteine was a fixed modification and methionine oxidation was considered as optional modification. Peptides were annotated at a false positive rate of 1% calculated in Elucidator (Ceiba Solutions, Boston, MA, USA) based on Peptide Teller (Peptide Prophet [47]).
The obtained raw data were analyzed with the Rosetta Elucidator software version 3.3 (Ceiba Solutions) employing a workflow described earlier [48]. Briefly the following steps were carried out and realized with an in-house-created visual script (step iii to viii): (i) feature detection, (ii) feature alignment across all MS runs, (iii) filtering for features p,0.05, (iv) feature annotation, (v) median normalization by a feature set containing search results and therefore discriminating features arising from single-charged contaminants, (vi) combination of signal intensities of biological replicates, (vii) filtering for unique peptides, (viii) statistical analyses.
For functional classification of proteins Ingenuity Pathway Analysis, Version 8.7 (Ingenuity Systems, Redwood City, CA, USA) was used. Furthermore, this software tool allowed the identification of biological processes particularly influenced by interfering RNA approaches and reduced PLB levels.

Statistical analysis
Data are presented as means 6 S.E.. For statistical comparison of Ca 2+ transport data the Mann-Whitney Rank Sum test was used. All other tests were done using the unpaired Student's t-test. Statistical significance was assumed at p,0.05 or p,0.01 (proteome analyses).

Silencing efficiency of amiR155-PLBr
To develop an improved vector system for modulation of cardiac Ca 2+ homeostasis by silencing PLB, an siRNA sequence directed against the rat PLB [11] was embedded into the native environment of the miR-155 stem loop structure (amiR155-PLBr) and placed under control of the CMV-MLC0.26 promoter, which is specifically active in cardiac cells [43,49]. To assess its functionality, a second construct containing a previously described shPLBr under control of an ubiquitously active U6 promoter was used ( Figure 1A). Co-transfection of HEK293 cells with either amiR155-PLBr or shPLBr expression plasmids and plasmids expressing a luciferase reporter containing the rat PLB cDNA within its 39UTR was used to determine the PLB silencing activity of these constructs ( Figure 1A). The control vectors expressing amiR155-Con and shCon were used to test the target specificity of the amiR155-PLBr and the shPLBr, respectively. Luciferase activity assays showed that both amiR155-PLBr and shPLBr recognized the target sequence in the PLB cDNA and efficiently silenced the luciferase reporter ( Figure 1B).

Quantification of amiR155-PLBr and shPLBr expression
AAV vectors represent the most efficient vehicles to deliver transgenes or shRNAs to the heart. Therefore, we generated AAV vectors expressing either amiR155-PLBr (scAAV6-amiR155-PLBr), shPLBr (scAAV6-shPLBr) or the respective short regulatory control RNAs (scAAV6-shCon, scAAV6-amiR155-Con). The AAV6 serotype was chosen because it has been shown to efficiently transduce cardiac cells in vitro [20,44]. Indeed, transduction of CM with 25610 3 viral vector equivalents per cell (vg/c) of a GFP reporter vector resulted in GFP expression in approx. 90% of CM at post transduction day 14 (data not shown). To study possible functional differences between the used PLB silencing vectors, we first determined the expression levels of mature shPLBr and amiR155-PLBr. For this purpose, CM were transduced with scAAV6-shPLBr or scAAV6-amiR155-PLBr at doses of 0.25610 3 to 25610 3 vg/c. The levels of shPLBr and amiR155-PLBr were then determined by qRT-PCR at days 4 and 14 post transduction. As shown in Figure 2A, both shPLBr and amiR155-PLBr levels increased with rising vector doses at day 4 as well as at day 14 post transduction. The measured levels of shPLBr, however, markedly exceeded those of amiR155-PLBr, i.e. 3-to 13-fold at day 4 and 4-to 15-fold at day 14 within the tested vector dose range. These differences were inversely related to the vector-induced reduction of respective PLB mRNA levels shown in Figure 2B.

Dose-and time-dependent PLB silencing by amiR155-PLBr
For further detailed analysis of the influence of the investigated vectors on PLB expression at the transcriptional and translational levels, CM were transduced with either scAAV6-amiR155-PLBr, scAAV6-shPLBr or the control vector scAAV6-amiR155-Con at doses between 2.5610 3 and 50610 3 vg/cell. Northern blot analysis performed at day 7 post transduction revealed that a low dose of 5610 3 vg/c scAAV6-shPLBr was sufficient to reduce PLB mRNA nearly completely ( Figure 3A). By contrast, a much higher dose of scAAV6-amiR155-PLBr was needed to achieve the same goal ( Figure 3A). These findings are in line with the qRT-PCR data shown in Figure 2B. In addition, a different vector dose-dependent decrease of immunoreactive PLB abundance was observed in CM after transduction with either scAAV6-shPLBr or The shPLBr expression is driven by an U6 polymerase III promoter and terminated by five consecutive thymidines. The amiR155-PLBr expression is driven by a CMV-enhanced 0.26 kb rat MLC hybrid polymerase II promoter (CMV enh -MLC0.26) and terminated by a SV40 poly A signal (pA). Note: The 39 ITR of the AAV genome has a mutation in the terminal resolution site resulting in self complementary AAV vector genomes after packaging into AAV6 capsids. Lower panel: Schematic illustration of plasmids containing the Firefly luciferase reporter gene and the PLB cDNA (pLuc-PLBr_cDNA) in its 39UTR. (B) Silencing activity of shPLBr and amiR155-PLBr using reporter assays. HEK293 cells were transfected with pLuc-PLBr_cDNA and Renilla luciferase expression plasmid and cotransfected with either amiR155-PLBr or shPLBr expressing plasmids. In addition, cotransfection was also done using the respective control plasmids amiR155-Con, shCon and amiR155-Luc. Cells were harvested 72 h after transfection and luciferase activity was determined. Firefly luciferase expression signals of the pLuc-PLB_cDNA reporter were normalized to Renilla luciferase and related to the respective signals in exclusively with Firefly luciferase expression vector transfected cells (Mock). *p,0.01 vs. respective control. doi:10.1371/journal.pone.0092188.g001 scAAV6-amiR155-PLBr ( Figure 3B, upper panel). In fact, semi-quantitative Western blot analysis revealed a reduction of the PLB abundance to 11% of the control after transduction with 50610 3 vg/cell of scAAV6-shPLBr, while the same dose of scAAV6-amiR155-PLBr reduced the PLB abundance to 44% of the control only ( Figure 3B, lower panel).
We then compared the amiR155-PLBr-and shPLBr-mediated PLB silencing as a function of time after transduction. For these analyses, CM were transduced with either scAAV6-amiR155-PLBr, scAAV6-shPLBr or the control vector scAAV6-amiR155-Con using a dose of 25610 3 vg/c of each vector. The PLB mRNA and protein levels were determined between days 1 to 14 post transduction using qRT-PCR and semiquantitative Western blot analysis ( Figure 3C-E). At day 1 after transduction, PLB mRNA levels of all vector-treated CM did not differ from the untreated cells. After another three days, however, a strong decline of PLB mRNA to 9% of control was found in scAAV6-shPLBr treated cells. At day 14 post transduction, this value amounted to 4% of control. By contrast, a decline of the PLB mRNA levels following scAAV6-amiR155-PLBr treatment was first visible at day 7 (33% of control), and this value reached 20% of the control value at day 14 post transduction ( Figure 3C). The quantitation of PLB protein abundance shown in Figure 3D, E confirms the distinct dose-dependent silencing of PLB for the two types of vectors. Thus, PLB protein levels in scAAV6-shPLBr-treated CM amounted to 14% and 2% of control at post transduction days 7 and 14, respectively. In scAAV6-amiR155-PLBr-treated cells, the respective PLB abundance was reduced to 53% and 29% of control only ( Figure 3E). Importantly, abundance of other proteins involved in Ca 2+ metabolism such as SERCA2a ( Figure 3D, E), NCX1 and CSQ2 (not shown) remained unaffected in the experimental groups.
These data demonstrate that treatment of CM with scAAV6-amiR155-PLBr results in delayed and significantly reduced PLB silencing compared to treatment of the cells with its shPLBexpressing AAV6 counterpart. To examine functional consequences of scAAV6-amiR155-PLBr-and scAAV6-shPLBr-induced silencing of PLB on the SERCA2a-catalyzed Ca 2+ uptake in CM, oxalate-supported SR Ca 2+ uptake was measured in cell homogenates. In Figure 4A, the Ca 2+ -dependence of the Ca 2+ uptake rate is shown. At day 14 post transduction, there was a clear leftward shift of the sigmoidal Ca 2+ -dependence curve due to PLB silencing by scAAV6-amiR155-PLBr. The respective EC 50 (Ca 2+ ) value was reduced to 63% of the control value found for scAAV6-amiR155-Contreated CM. This effect indicates increased Ca 2+ sensitivity of the SERCA2a-mediated Ca 2+ transport due to PLB silencing.
The stimulating effect of PLB silencing by scAAV6-amiR155-PLBr on SR Ca 2+ uptake is clearly dependent on the time after transduction, as shown in Figure 4B. For the time course, SR and SERCA2a (lower panel) abundance found in scAAV6-shPLBr (shPLBr) and scAAV6-amiR155-PLBr (amiR155-PLBr) vector transduced cells as compared to respective control vector transduced CM exhibiting PLB and SERCA2a abundance of 100%, respectively. *p,0.05 shPLBr vs. amiR155-PLBr and amiR155-PLBr vs amiR155-Con; # p,0.05 vs. cells transduced with the respective control vector scAAV6-amiR155-Con or scAAV6-shCon. doi:10.1371/journal.pone.0092188.g003 Ca 2+ uptake activity was measured at a submicromolar free Ca 2+ concentration of 0.34 mM, at which the Ca 2+ affinity of the Ca 2+ transporter is known to be dependent on PLB. A significant increase in Ca 2+ uptake activity was observed 7 and 14 days after treatment of CM with scAAV6-amiR155-PLBr. At day 14 after transduction, the normalized Ca 2+ transport activity reached 77% of the maximal value measured at saturating Ca 2+ . A comparable value of 85% was measured for PLB silencing with the scAAV6-shPLBr vector ( Figure 4C). By contrast, respective relative Ca 2+ uptake rate values around 50 to 55% of the maximum were found for CM that were treated with the two control vectors scAAV6-amiR155-Con ( Figure 4B) and scAAV6-shCon ( Figure 4C) as well as for non-transduced cells (Figure 4c) at all investigated time points. Interestingly, a detectable functional effect of PLB silencing by the shPLB-producing vector was that it exerted its functional effect on SR Ca 2+ uptake at an earlier time point than the amiR155-PLBr producing vector ( Figure 4B, C).

ScAAV6-shPLBr, but not scAAV6-amiR155-PLBr, induces an inflammatory response
To investigate possible side effects of both amiR155-PLBr and shPLBr-encoding vectors, the protein expression profile of CM was analyzed 14 days after transduction with 25x10 3 vg/c of either scAAV6-amiR155-PLBr, scAAV6-shPLB, scAAV6-shCon or scAAV6-amiR155-Con. At this time point, no differences in morphology, cell growth and functionality (beating frequency) were obvious between the groups. Analysis of the 1,241 proteins covered by shotgun proteome analysis revealed that a total of 60 to 80 proteins were dysregulated through CM transduction with the AAV vectors ( Figure S1). No major intergroup differences of the overall protein pattern were detected, except for a small subset of proteins which exhibited altered abundance levels (.1.5-fold, p,0.05) after transduction with scAAV6-amiR155-PLBr and scAAV6-shPLBr as compared to the respective control vectors ( Figure 5A) or to non-transduced CM ( Figure S2). Notably, the number of affected proteins was lower in the scAAV6-amiR155-PLBr treatment group (vs. scAAV6-amiR155-Con: two up-and 14 down-regulated) compared to CM transduced with scAAV6-shPLBr (vs. scAAV6-shCon: 12 and 18 proteins were up-and down-regulated, respectively). This analysis also confirmed efficient PLB silencing down to less than 10% of the respective control for both amiR155-PLBr and shPLBr vectors (data not shown). Furthermore, a slight increase of troponin I levels (amiR155-PLBr: FC 1.7, shPLBr: FC 1.5) was observed. Interestingly, a decreased level of the regulatory subunit of cAMP-dependent protein kinase type I (FC 1.7) was found after transduction with scAAV6-amiR155-PLBr, while the level of this protein was elevated in scAAV6-shPLBr-transduced CM. None of the other identified calcium cycling or various signaling proteins displayed altered levels after transduction with scAAV6-amiR155-PLBr. A different observation was made for CM transduced with scAAV6-shPLBr. Several proteins were altered which are involved in acute phase response signaling and inflammatory processes, such as the b-isoform of the signal transducer and activator of transcription 1 (STAT1) shown in Figure 5B, a-2-HS-glycoprotein, serpine peptidase inhibitor clade I, b-2-microglobulin, calnexin and alpha-1-antitrypsin. In addition to these proteins, two IFN-regulated proteins, the galectin-3-binding protein (Gal3BP) and the major prion protein (PrP), were dysregulated by shPLBr but remained unchanged after treatment with amiR-155-PLBr. To further analyze the CM inflammatory response in more detail, Western blot analyses were performed for selected members of the JAK/STAT pathway. As shown in Figure 5B, C, the upregulation of STAT1 after transduction of CM with scAAV6-shPLBr that was found by proteomics was verified by Western blot analysis ( Figure 5C). Semiquantitative evaluation of the latter revealed an estimated 1.4-fold increase in total STAT1 abundance and a 3.3-fold rise of the level of STAT1 phosphorylated at the amino acid residue Tyr 701 (pSTAT1Tyr701) in scAAV6-shPLBrtransduced cells as compared to CM transduced with scAAV6-shCon. In addition to STAT1, a significant 1.7-fold up-regulation of total STAT3 was seen after transduction of CM with scAAV6-shPLBr compared to scAAV6-shCon. In contrast to STAT1, however, the phosphorylation status of STAT3 was not altered at the time point measured. Most notably, neither the total nor the phosphorylated levels of STAT1 or STAT3 were altered after transduction of CM with the amiR155-PLBr encoding vector ( Figure 5B, C).

Alteration of cardiac miR levels by scAAV6-amiR155-PLBr or scAAV6-shPLBr
Bish et al. [14] reported that strong cardiac shPLB expression in canines altered the expression of several miRs and induced toxic side effects such as transiently increased serum levels of troponin I. To investigate whether scAAV6-shPLBr and scAAV6-amiR155-PLBr treatment might affect the expression levels of selected cardiac miRs that had been analyzed in the in vivo study of Bish et al. [14], the levels of miR-1, miR-21, miR-124, miR-195 and miR-199a were determined in CM at day 14 after transduction with 25610 3 vg/c of scAAV6-shPLBr, scAAV6-amiR155-PLBr, scAAV6-shCon or scAAV6-amiR155-Con. We detected a common decrease of the miRs expression levels between 20-50% for all vectors compared to non-transduced control CM. Interestingly, among the investigated miRs, the miR-21 expression levels were significantly increased by scAAV6-shPLBr and scAAV6-amiR155-PLBr compared to its respective shCon and amiR155-Con control vectors. No such changes were found for any of the other quantified miR ( Figure 6).

Discussion
Here, we show that scAAV6-amiR155-PLBr, which expresses a newly engineered small regulatory RNA directed against the negative SERCA2a modulatory protein PLB, improves the SERCA2a-catalyzed Ca 2+ transport activity of the SR in CM. The efficiency of scAAV6-amiR155-PLBr was as high as that of an AAV6 vector expressing the conventional shPLBr that was previously used for PLB silencing [11,12]. Importantly, our data reveal that scAAV6-amiR155-PLBr exhibits an improved cardiac safety compared to the shPLBr-expressing vector.
In a previous study, we demonstrated the therapeutic efficacy of shPLBr in a rat heart failure model using AAV9 vectors for cardiac delivery. No negative side effects were detected in this study [12]. Severe cardiac toxicity, however, was observed in healthy canines after AAV6 vector-mediated delivery of shPLB [14]. Reducing the expression of the shRNA by decreasing the vector dose or the use of weaker promoters are potential ways to prevent shRNAinduced side effects [26,50]. This procedure, however, is not always successful and shRNAs can remain toxic, even when shRNA levels are reduced [28]. Moreover, shRNA expression levels below a critical value can lead to markedly reduced silencing effect [28]. Thus, reduction of shRNA toxicity might come at the price of loss of therapeutic efficiency.
A new class of engineered regulatory RNAs, amiRs, appears to be safer than shRNAs [27][28][29]. For this reason we have used and characterized amiR155-PLBr for silencing of PLB in CM. In accordance with previous reports [27,28,51] comparing shRNAs versus amiRs approaches, we found that amiR155-PLBr was expressed at a lower level (about 10-fold) than shPLBr. The lower activity of the heart-specific CMV-MLC0.26 promoter [43] driving the expression of miR155-PLBr compared to the highly active U6 promoter used for shPLBr delivery is the most reasonable explanation for the observed differences between the two expression systems. We cannot rule out, however, that other factors, for example individual structural characteristics, cause the differential expression of amiR155-PLBr and shPLBr. The lower expression of amiR155-PLBr was linked to distinctly (about 10fold) less pronounced PLB silencing. Interestingly, when both amiR155-PLBr and shPLBr were expressed at similar levels, achieved by about 10-fold increase of scAAV6-amiR155-PLBr vector dose, PLB silencing was similar ( Figure. 2A and B). Thus, our data suggest, that amiR155-PLBr and shPLBr have per se similar silencing efficiency. This, however, is somewhat contradictory to previous reports. McBride et al. [28] and Mazcuga et al. [51] found comparable silencing efficiency of amiRs and shRNAs despite the lower expression they found for amiRs. Other studies confirmed these data and reported higher efficacy of amiRs as compared to shRNAs [27,35,36,51]. So far, the mechanisms underlying the differing activities of amiRs and shRNA are not well understood. Differences in stability [28] and processing to generate mature siRNA molecules [37] may be important factors. In particular, the specific structure of amiRs mimicking the naturally occurring cellular miR structure has been assumed to be a determinant that makes amiRs more efficient than shRNAs [28]. In addition, the use of different termination signals and the distance between the promoter, the shRNA/amiR sequence and the termination signal may affect the respective silencing efficiency [37,51]. Although no in-depth analysis has been carried out yet, sequence-specific determinants may play an important role as well. This would explain why, in some cases, amiRs are more efficient than shRNAs, while in other cases, amiRs and shRNAs have similar activity or amiRs may even be less efficient. We confirmed that the silencing efficiency of amiR155-PLBr used here was comparable to that using a miR-30 scaffold (data not shown). Hence, the miR-155 scaffold does not affect the inhibitory activity of amiRs.
The repression of the PLB protein in CM by amiR155-PLBr by 50% and 70% at days 7 and 14 post transduction, respectively, was considerably lower than that induced by shPLBr (decline of 85% at day 7 and . 95% at day 14). Nevertheless, the amiR155-PLBr-mediated PLB silencing was functionally as efficient as that of shPLBr, i.e. the increase of SERCA2a-mediated SR Ca 2+ transport was similar for both small regulatory RNAs. Hence, the diminished PLB silencing by amiR155-PLBr appears to be functionally irrelevant in the case studied here. The effect observed here may be explained by a functional threshold for PLB-induced inhibition of SERCA2a. Both silencing strategies (amiR155-PLBr and shPLBr) diminished PLB to a level below this value. Stronger PLB silencing will thus not result in further enhancement of SERCA2a-mediated SR Ca 2+ transport.
Avoiding undesirable side effects is critical for therapeutic applications since they restrict application of the therapeutic agent in humans. Thus, we have carried out a proteomic analysis and validated some of our findings by semiquantitative Western blot analysis. Importantly, shPLBr, but not amiR155-PLBr, induced the expression of the IFN-regulated proteins PrP and Gal3BP and the proinflammatory genes STAT1 and STAT3, accompanied by STAT1 activation via phosphorylation at the Tyr 701 residue. Both STAT proteins are part of the Janus kinase (JAK)/STAT pathway, which mediates transduction of stress signals from the plasma membrane to the nucleus. IFNs are important triggers of the JAK/STAT pathway and its activation has been described in association with shRNA treatment [22,24]. Hence, our results demonstrate that long-term and high level shPLBr expression can induce undesirable induction of stress response genes in CM, probably via IFN-dependent pathways. This activation may be avoided by the use of amiR155-PLB. We also found dysregulation of several other proteins after shPLBr and amiR155-PLBr vector treatment. In most cases this may be caused by adaptation of the cellular protein network to the changed cellular metabolism as a result of PLB downregulation. Whether these findings will have relevance for the in vivo application of the vectors is currently not clear. It should, however, be considered in follow up studies.
We also detected alterations in the expression of cardiac miRs after transduction of CM with the shRNA and amiR155 expressing AAV vectors used in this study. Compared to untreated CM, the miRs levels were distinctly reduced, indicating that RNAi vectors have affected the miR biosynthesis. Importantly, miR downregulation occurred through treatment with PLB-silencing as well as its respective shRNA and amiR155 expressing control vectors. Hence, the suppressing effect was not restricted to shPLBr and amiR155-PLBr but represented a common outcome of AAV vector-mediated transduction and expression of interfering RNAs. High shRNA expression can interfere with the cellular miR expression [14,26] and thus high vector doses used here, leading to strong shRNAs and amiR155 expression, may be the reason for our observations. Vector transduction effects, however, have also been taken into account. In a recent study, Bish et al. [14] detected dysregulation of several cardiac miR after AAV6-shPLB application to dogs. In contrast to our observations, they detected downand up-regulation of miRs. The state of cellular differentiation (neonatal cardiomyocytes vs. adult cardiomyocytes), speciesspecific differences as well as differences between the in vitro and in vivo situation may be possible explanations for these differences. In this regard it should also be noted that very strong downregulation of PLB as triggered by RNAi-inducing vectors may be rather undesirable for treatment of heart failure in humans, because in humans PLB is essential for cardiac health and its absence results in lethal heart failure [52]. Thus it is obvious that further improvement of the safety of the RNAi vectors is desirable. An increase of safety might be accomplished by the use of gene expression systems subject to pharmacological regulation. For example, the tetracycline-dependent gene expression system that has already been used successfully for the expression of amiR [53] may be easily adaptable for expression of amiR155-PLBr. The system allows adjustment amiR expression within a therapeutic window, as well as outright on-off switching.
Another interesting finding of the miR analysis was a significantly higher expression of miR-21 in CM after treatment with PLB silencing vectors as compared to the respective control vectors. MiR-21 is overexpressed in many tumors and considered as key regulator of oncogenic processes [54]. It is also highly expressed in CM [55] and cardiac fibroblasts [56] and apparently plays an important role in cardiac metabolism. Both miR-21 regulation as well as PLB-SERCA2a interaction are known to be regulated via b-adrenergic receptor signaling [57,58]. Thus our data suggest that PLB-SERCA2a system could be an upstream regulator of miR-21 in the b-adrenergic signaling cascade.
In conclusion, our data suggest that the amiR155-PLB AAV vector could be a superior new RNAi vector system for improving the compromised Ca 2+ handling function of the cardiac SR. The efficiency and the safety of this novel tool will have to be tested in experimental heart failure animal models in further studies.