J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X
Q4
Ava i l ab l e on l i ne a t www.sc i enced i r ec t . com
ScienceDirectwww.e l sev i e r . com/ loca te / j p ro t
JPROT-01805; No of Pages 10
Regulation of PPAR-alpha pathway by Dicer revealedthrough proteomic analysis
PRO
OF
Nandini A. Sahasrabuddhea,b,1, Tai-Chung Huangc,1, Sartaj Ahmada, Min-Sik Kimc,d,Yi Yangc, Bidyut Ghoshg, Steven D. Leachc,g, Harsha Gowdaa, B.L. Somania,Raghothama Chaerkadya,b,c, Akhilesh Pandeya,c,d,e,f,⁎aInstitute of Bioinformatics, International Technology Park, Bangalore 560066, IndiabManipal University, Madhav Nagar, Manipal 576104, IndiacMcKusick-Nathans Institute of Genetic Medicine, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USAdDepartment of Biological Chemistry, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USAeDepartment of Pathology, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USAfDepartment of Oncology, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USAgDepartment of Surgery, Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA
Please cite this article as: Sahasrabuddhanalysis, J Prot (2014), http://dx.doi.org/10
A B S T R A C T
Article history:Received 28 November 2013Accepted 13 April 2014
RECTE
Dicer is a crucial RNase III enzyme inmiRNA biogenesis pathway. Although numerous studieshave been carried out to investigate the role ofmiRNAsandDicer in the regulation of biologicalprocesses, few studies have examined proteomic alterations upon knockout of Dicer. Weemployed a Cre–loxP-based inducible knockout mouse system to investigate the proteomeregulated by Dicer-dependent miRNAs. We utilized spiked liver lysates from metabolicallylabeled mice to quantify the subtle changes in the liver proteome upon deletion of Dicer. Weidentified 2137 proteins using high resolution tandem mass spectrometry analysis. Theupregulated proteins included several enzymes involved in peroxisomal β-oxidation of fattyacids and a large majority of the upregulated proteins involved in lipid metabolism wereknown PPARα targets. MRM-based assays were carried out to confirm the upregulation ofenzymes including peroxisomal bifunctional enzyme, phosphoenolpyruvate carboxykinase 1,cytochromeP450 3A13, cytochromeP450 3A41andmyristoylatedalanine-richprotein kinaseCsubstrate. Further, miRNA-124 which is predicted to regulate expression of peroxisomalbifunctional enzyme was confirmed to be downregulated in the Dicer knockout mice. Ourstudy demonstrates the strength of coupling knockout mouse models and quantitativeproteomic strategies to investigate functions of individual proteins in vivo.
Biological significanceDicer dependent miRNA biogenesis is themajor pathway for generation of mature miRNAs. Wedeveloped SILACmouse-based proteomics screen to identify protein targets of Dicer-dependent
2 J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X
Please cite this article as: Sahasrabuddheanalysis, J Prot (2014), http://dx.doi.org/10
miRNAs in liver of Dicer knockoutmice.We spiked liver lysates of induced and uninduced Dicerknockout mice with liver lysate of SILAC labeled mice for identification of dysregulatedproteome. We quantitated 1217 proteins of which 257 were upregulated in induced Dicerknockout mice. We observed enrichment of PPAR-α targets and proteins involved in lipidmetabolism among upregulated proteins. We further carried out MRM-based validation ofperoxisomal bifunctional enzyme, phosphoenolpyruvate carboxykinase 1, Cyp3A13, Cyp3A41andmyristoylated alanine-rich protein kinase C substrate.We further validated upregulation ofperoxisomal bifunctional enzyme using Western blot analysis and downregulation of itspredictedupstreammiRNA,miR-124usingqRT-PCR.Our studydemonstrates that uponablationofDicer, certainDicer-dependentmiRNAsaredysregulatedwhich result indysregulationof theirtarget proteins suchasproteins associatedwith lipidmetabolism.Our study illustrates theuseofSILAC strategy for quantitative proteomic investigations of animal model systems.
MicroRNAs (miRNAs) are known to be one of the keymediatorsof post-transcriptional gene regulation [1]. Regulation of genesthrough miRNAs has been implicated in a variety of biologicalprocesses. MicroRNAs exert post-transcriptional regulationthrough binding to 3′ UTRs of target mRNAs and leading tohalted translation or directing them for degradation [1]. Dicer isa ubiquitously expressed and evolutionarily conserved ribonu-clease in the si/miRNAbiogenesis pathway. It contains RNase IIIdomains, whose role is to excise precursor miRNAs to generatemature miRNAs.
The role of Dicer has been investigated in several biologicalsystems. Various in vitro and in vivo systems have beendeveloped to explore the role of Dicer [2–6]. Dicer-deficientmouse embryonic stem cells exhibit defective proliferation,differentiation and centromeric silencing [7,8]. Depletion ofDicer is lethal at embryonic stage. Thus, inducible knockoutsystems were utilized to study the function of Dicer in specificcell types of adult mice including hepatocytes, CD8 cells andspermatogonia [9–12]. Deletion of Dicer results in dysregulationof enzymes involved in lipid metabolism in hepatocytes andsmall intestine [9,13]. Ablation of Dicer in hepatocytes results inthe development of hepatic tumors in mice, indicating crucialregulatory role of Dicer andDicer-dependentmiRNAs [9]. In thisstudy, we aimed to investigate the proteomic alterations thatare induced by Dicer-dependent miRNAs in the liver.
We employed an inducible Cre–loxP knockout system forthe deletion of Dicer. For quantitative profiling of the liverproteome, we used SILAC labeled mouse liver lysates forspiking. Previously, various in vitro labelingmethods includingDIGE, 18O and iTRAQ labeling have been adopted for quanti-tative proteomic profiling of mouse systems which havevarious limitations [14–19]. Gel-based quantitative proteomicsapproaches are known for low reproducibility and separationdue to the limited pI range. Back exchange of 18O isotope,compression of iTRAQ ratios and probability of introducingmanual errors in labeling are other limitations of in vitroquantitation approaches. Label free quantitation has beenused as an alternative quantitative proteomic approach.However, it requires highly reproducible LC–MS conditions,which are often difficult to achieve, especially across multiplesample runs [20]. Therefore, in vivo labeling strategies suchas 15N labeling and stable isotope labeling by amino acidsin cell culture (SILAC) are generally preferred for quantitative
NA., et al, Regulation.1016/j.jprot.2014.04.027
ED P
RO
Oproteomics [21,22]. Some of the imitations of the 15N labelingapproach include challenges in achieving complete labelingand complexity of the quantitation data. As a result, SILAChas been the method of choice for in vivo labeling. Untilrecently, the use of SILAC was limited to cell lines, but withthe development of 13C6-lysine enriched chow, mice can alsobe labeled in vivo, extending the application of SILAC toanimal systems [13,22,23]. This quantitation approach isrelatively free of technical errors as proteins are labeled invivo.
We developed a SILAC mouse-based quantitative proteo-mics assay to identify differentially expressed proteins upondepletion of Dicer in liver. We carried out high resolutionmass spectrometry analysis and identified 2137 proteins. Ofthe 257 proteins that were upregulated in the liver of Dicerknockout mice, we observed enrichment of proteins involvedin fatty acid metabolism. We further carried out MRM assaysto validate candidate proteins, which include peroxisomalbifunctional enzyme, phosphoenolpyruvate carboxykinase 1,Cyp3A13, Cyp3A41 and myristoylated alanine-rich proteinkinase C substrate. We also carried out Western blot analysisto validate upregulation of peroxisomal bifunctional enzyme.We further validated downregulation of miR-124 which ispredicted to regulate expression of peroxisomal bifunctionalenzyme using qRT-PCR assay. Our findings highlight crucialroles of Dicer and Dicer-dependent miRNAs in the regulationof proteins involved in lipid transport and metabolism inmouse liver. Our study also demonstrates the utility of SILACmouse-based proteomics and MRM assays as robust massspectrometry-based approaches for the development of in vivoquantitative proteomics strategies.
2. Materials and methods
2.1. Generation of inducible Dicer knockout mouse
We used Cre–loxP system to generate inducible knockoutmice. ROSA26-CreERT2mice andmice with floxed Dicer exons21 and 22 were crossed. The progeny was responsive totamoxifen, resulting in the deletion of floxed Dicer exons 21and 22. ROSA26-CreERT2 mice without the treatment oftamoxifen were used as control. Mice were monitored dailyfor any obvious pathology. On day 8 post-induction, micewere starved for 3 h prior to euthanasia and necropsy.
of PPAR-alpha pathway by Dicer revealed through proteomic
3J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X
UNCO
RREC
2.2. Generation of SILAC mice
As described previously, SILACmice were generated by feedingstable isotope-labeled mouse food obtained from CambridgeIsotope Laboratories (Mouse Feed Labeling Kit, catalog number:MLK-LYS-C) [13,24]. Briefly, a twenty-eight day old femalemouse (F0) was fed diet containing 13C6-lysine (heavy diet) andmated 14 days later. The heavy diet was maintained until F1pups were weaned. One F1 female mouse was maintained onheavy diet and propagated to F2. Littermate female mice (F0)were fed light diet to generate unlabeled control mice. Thelabeling efficiency was monitored in blood, liver and lungspecimens collected at 10 weeks of age from F0 and 4 weeks ofage from the F1 generation.
2.3. Sample preparation and basic RPLC fractionation
Liver tissues excised from five uninduced Dicer knockout(control) and induced Dicer knockout (Dicer knockout) micewere lysed in 9 M urea in 20 mM HEPES supplemented withcomplete protease inhibitor cocktail (Roche Diagnostic Sys-tems). Protein concentration was measured using BCA assays.Pooled control and pooled Dicer knockout liver lysates werespiked with liver lysate from SILAC mouse at 2:1 (w/w) ratio.The spiked lysates were reduced using 4.5 mM dithiothreitolfor 30 min at 60 °C followed by alkylation in the dark using10 mM iodoacetamide for 10 min. Lysates were digested insolution using Lys-C protease. Lys-C was added at 1:50 (w/w)(Lysyl Endopeptidase® Mass Spectrometry grade, WakoChemical USA, Richmond, VA) to the lysates and incubated at37 °C for 4 h. Additional Lys-C was added to the pre-digestedlysate at 1:50 (w/w) ratio and incubated for 12 h.
Basic reversed phase liquid chromatography (BRPLC) wascarriedout asdescribedpreviously [13]. Briefly, Lys-Cdigestswerereconstituted in BRPLC solvent A (10 mM triethylammoniumbicarbonate, pH 9.5) and were separated on XBridge BEH C18Column (Waters, UK)with a linear increase in gradient from5 to100% of 10 mM triethylammonium bicarbonate (TEABC) with90% acetonitrile (pH 9.5) over 30 min. and persisting for 10 min.For each condition, 24 fractions were collected and dried beforeLC–MS/MS analysis.
2.4. LC–MS/MS analysis
LC–MS/MS analysis of 48 bRPLC fractions was carried outusing Eksigent nanoLC interfaced with the LTQ-Orbitrap XLETDmass spectrometer (Thermo, San Jose, CA). Peptides wereloaded on a trap column (75 μm × 2 cm) packed with C18
material (5 μm Magic C18 AQ) at a flow rate of 5 μl/min of 97%solvent A (3% acetonitrile and 0.1% formic acid) and separatedon an analytical column (75 μm × 12 cm) packed with thesame material using linear gradient of solvent B (0.1% formicacid in 90% acetonitrile) from 10% to 60% solvent B for 60 min,to 97% solvent B from 74 to 90 min. All the MS spectra wereacquired on an Orbitrap analyzer at a resolving power of60,000 at 400 m/z while MS/MS spectra were acquired on theLTQ analyzer. Ten most intense precursor ions from a surveyscan within m/z range from 350 to 1800 above intensity of 1000counts were isolated with a 2 Da window and fragmented byCID with 30% normalized collision energy. The precursors were
Please cite this article as: Sahasrabuddhe NA., et al, Regulation oanalysis, J Prot (2014), http://dx.doi.org/10.1016/j.jprot.2014.04.027
ED P
RO
OF
excluded, after fragmentation, for 30 s with a 7 ppm window.Maximum ion injection times were set to 500 msec for MS and200 msec for MS/MS. The automatic gain control targets wereset to 5 × 105 for MS in the Orbitrap, 1 × 104 for MSn in the LTQ.
2.5. LC–MS/MS data availability
Raw LC–MS/MS data was uploaded on Tranche to sharewith the community. The data may be downloaded fromProteomeCommons.org Tranche using the following hash:
Mass spectrometry data analysis was carried out using Prote-ome Discoverer 1.2 suite (Thermo Fisher Scientific, Bremen,Germany). Precursormass range of 300 to 5000 Da and signal tonoise ratio of 1.5were used as filtration criteria for generation ofpeak lists. NCBI RefSeq 42 containing mouse proteins withknown contaminants (29,109 entries) was used as a referencedatabase. SEQUEST and Mascot algorithms were used to carryout database searches. The parameters used for databasesearches include Lys-C as a protease with allowed one missedcleavage, carbamidomethyl cysteine as a fixed modificationand 13C6-lysine, oxidation of methionine as variable modifica-tions. MS error window of 10 ppm and MS/MS error window of0.8 Da were allowed. As described earlier, LC–MS/MS data wassearched against a reversed database to calculate 1% falsediscovery rate (FDR) score cut-off at the peptide level [25]. Peakarea was computed using Proteome Discoverer 1.2. Light/heavyratios of peak areas were calculated for every peptide andprotein for control and Dicer knockout samples. To determinedifferentially expressed proteins in Dicer knockout liver, theratios from control and Dicer knockout liver samples werecompared to obtain ‘ratio-of-ratios’ [23]. Proteins for whichlight/heavy ratio fromDicer knockout liver sampleswas ≥2-foldcompared to the ratio from control liver samples, wereconsidered to be differentially expressed.
Upregulated proteins in the liver of Dicer knockout micewere considered analysis using DAVID [26]. Enriched KEGGpathways obtained through DAVID with p-value ≤ 2E-2 wereconsidered for further analysis.
2.7. MRM assays
MRM assays were designed to validate candidate upregulatedproteins in Dicer knockout mice. Five differentially expressedproteins were selected as candidates for MRM analysis.Skyline v1.2 was used to create a transition list of proteotypicpeptides from the selected proteins [27]. Preference was givento proteotypic peptides with precursor charge of +2 that didnot contain cysteine and methionine.
In solution digestion of equal amounts of four control andDicer knockout mouse liver samples was carried out and thepeptides were acidified and stored at −80 °C. All samples wereanalyzed in triplicate on a TSQ Quantum Ultra mass spec-trometer (Thermo, San Jose, CA) interfaced with Agilent's 1100series LC. Peptides were enriched on a trap column (5 μm,
f PPAR-alpha pathway by Dicer revealed through proteomic
4 J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X
UNCO
RREC
75 μm × 2 cm.) and separated using analytical column (3 μm,75 μm × 10 cm) with a linear gradient of 90% ACN 0.1% formicacid from 5% to 40% for 80 min followed by holding at 90% for5 min at a constant flow rate of 350 nl/min. Both columnswere packed in-house using Magic C18 AQ (MichromBioresources). Spray voltage of 2.5 kV was applied and iontransfer tube was maintained at 275 °C. The collision energyfor each transition was optimized based on the formula C.E. =0.034 × (m/z) + 3.314. Data was acquired with Q1 and Q3 set at0.4 and 0.7 unit mass resolutions, respectively.
2.8. Bioinformatics analysis for enrichment of upstreammiRNAs
Proteins upregulated in the liver of Dicer knockout mice weretaken as queries to search for the potential upstream miRNAs.From TargetScan (Release 6.2), ‘Summary Counts’ were consid-ered to identify seed sites in 3′ UTR of mRNAs and therepresentative target miRNAs [28]. RefSeq was used as thereference database for mouse mRNAs. Only miRNAs belongingto the speciesMus musculus (mmu-miR) were used for analysis.Gene symbols were used to map identifiers of proteins and thepredicted regulating miRNAs.
2.9. Quantitative RT-PCR (qRT-PCR) and Western blotting
qRT-PCRwas carried out to confirm the deletion of Dicer1 exons21 and 22 and to examine the altering abundance of miRNA-124after Dicer knockout. Total RNA was purified using RNeasy MiniKit (Qiagen, Valencia, CA) after the integrity of RNAwas checked(2100 Bioanalyzer, Agilent Technology, Santa Clara, CA). Toquantify Dicer transcripts, total RNA was reverse-transcribedusing SuperScript III Reverse Transcriptase with oligo (dT) 20primers (Invitrogen, Carlsbad, CA) for confirmingDicer1deletion.Primer sequences to span theDicer exon 21 and 22 junctions are:5′-CTGTTTTGCACGTACCCTGA-3′ and 5′-GAAGCCAATTCACAGGAGGA-3′. Biological triplicate samples were analyzed using iQ™SYBR Green Supermix Assay (Bio-Rad, Hercules, CA). Fordetermining mouse miR-124 level, MultiScribe™ reverse tran-scriptase was used and followed by Taqman® small RNA assayfor mmu-miR-124a (Life Technologies, Grand Island, NY).Biological triplicate samples from 3 control and 3 Dicer knockoutmouse liver were selected. The relative abundance of transcriptsin biological triplicates from 3 control and 3 Dicer knockoutmicewas derived by comparing to the standard curve composed bythe no-tamoxifen control mouse liver.
Western blotting for peroxisomal bifunctional enzyme wascarried out essentially as described previously [29]. Biologicaltriplicate samples from 3 control and 3 Dicer knockout mouseliver were used for the assays. Anti-EHHADH (K-16) rabbitpolyclonal antibody (Santa Cruz, sc-99388), and anti-β-actin(13E5) rabbit monoclonal antibody (Cell Signaling Technology,#4970) were used for the loading control.
365
366
367
368
369
370
371
372
3. Results and discussion
3.1. Dicer is essential for survival of adult mice
To assess the role of Dicer in adult mice and to circumventembryonic lethality, we adopted an inducible Cre–loxP knockout
Please cite this article as: Sahasrabuddhe NA., et al, Regulationanalysis, J Prot (2014), http://dx.doi.org/10.1016/j.jprot.2014.04.027
ED P
RO
OF
system. Ubiquitous expression of tamoxifen responsive Crerecombinase (Cre-ERT2) was driven by the ROSA26 locus. TheCre-ERT2micewere crossed tomicewith floxed exons 21 and 22of the Dicer gene. Upon administration of tamoxifen, the Crerecombinase translocates to the nucleus deleting floxed exons21 and 22 of Dicer. Henceforth, the floxedDicermicewithout thetreatment of tamoxifen are referred to as control mice andinduced Dicer knockout mice are referred as Dicer knockoutmice. RT-PCR analysis targeting the junctional region of exons 20and 21 demonstrated that 80% of the mice had disrupted Dicer1.Representative RT-PCR gel image is depicted in SupplementaryFig. 1A. The mice were monitored daily for any obviousabnormalities. Dicer ablation proved to be lethal in most of thecases as 80% of the mice died ~10 days after ablation of Dicerwas induced by tamoxifen administration [13].
The major histopathological changes observed by day 8post-induction of Dicer knockout included inflamed smallintestine and bonemarrow depletion with the reduced numberof myeloid lineage cells. Dysregulation of lipid metabolism andproteomic changes associated with ablation of Dicer in smallintestine have been reported earlier [13]. Liver performs crucialfunctions including metabolism of lipids, glucose and detoxifi-cation. When Dicer knockout mice were euthanized, we did notobserve any gross abnormality of the liver during necropsy. Thehistological studies of the same specimens did not reveal anyobvious pathology. Immunohistochemical staining of liver fromcontrol andDicer knockoutmice is illustrated in SupplementaryFig. 1B. Administration of 1 mg and 4 mg of tamoxifen showedsimilar downregulation ofDicer; thus,weused1 mg/day dose oftamoxifen for subsequent experiments. We hypothesize thatcertain tissues developed an observable pathology in this timeframe because of their rapidly dividing nature.
3.2. Development of a quantitative proteomic screen to identifythe targets of Dicer in liver
The aim of designing a quantitative proteomic assay was todevelop a suitable labeled internal standard that would allowsystematic quantitation of alterations in the liver proteomeupon Dicer knockout. SILAC mice have been previously usedto explore proteomic changes associated with aging in mice[23]. We developed a SILAC mice-based screen as depicted inFig. 1. In order to minimize inter-individual variation, wepooled liver lysates from five mice. As described in themethods, the SILAC mice were metabolically labeled byfeeding 13C6-lysine labeled diet. As a consequence, all proteinswere labeled in vivo with 13C6-lysine, making them robustlabeled internal standards. Control and Dicer knockout liverlysates were spiked with lysates of liver from SILAC mice at a2:1 ratio. Because generation and maintenance of SILAC miceare costly, this approach of spiking SILAC mouse tissuesprovides an economical alternative for in vivo quantitativeproteomics.
Proteolysis was carried out in-solution using Lys-C prote-ase which specifically hydrolyzes C-terminal of lysine toenable quantitation for the maximum number of peptideswith paired 13C6-lysine containing peptides from SILAC mice.Reversed phase liquid chromatography operating at themacroflow rate in the high pH condition (BRPLC) was used toobtain 24 fractions per condition. The LC–MS/MS analysis was
of PPAR-alpha pathway by Dicer revealed through proteomic
validation using MRM assays, Western blot and qRT-PCR
2:1 (w/w) 2:1 (w/w)
Fig. 1 – SILACmice based proteomic screen to investigate therole of Dicer. Liver lysates of five control and five Dicerknockout mice were spiked at 2:1 ratio with liver lysates oftwo SILAC mice. Proteolysis was carried out using Lys-Cprotease. Digests were separated on BRPLC to obtain 24fractions each for control and Dicer knockout spiked liverlysates. LC–MS/MS analysis of the BRPLC fractions wascarried out on an Orbitrap-XL ETD mass spectrometer.Database searches and quantitation were carried out usingProteome Discoverer (v 1.2) and ‘ratio-of-ratios’ wascalculated to identify differentially regulated proteins in theliver of Dicer knockout mice. Candidate upregulated proteinsin Dicer knockout were validated using MRM assays,Western blotting and qRT-PCR.
Table 1 t1:1– A partial list of proteins involved in lipidt1:2metabolism.t1:3t1:4Protein Gene
symbolDicer
knockout/control
t1:51. Sodium/bile acid cotransporter Slc10a1 15.5t1:62 ATP-binding cassette sub-family D
5J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X
UNCOcarried out on nanoflow reversed phase liquid chromatography
coupled with LTQ-Orbitrap XL ETDmass spectrometer. MS wasacquired at a high resolution (60,000 at 400 m/z) to obtain highaccuracy quantitation information. In total, we analyzed 48 LC–MS/MS fractions. At better than 1% FDR, we identified 58,684peptides corresponding to 2137 proteins.
3.3. Differentially regulated proteins in liver of Dicer knockoutmice
We identified 1414 proteins from liver of control mice and1992 from Dicer knockout mice. As listed in SupplementaryTable 1, we selected 1217 proteins which were quantified inboth conditions and calculated ratio of SILAC ratios from eachcondition to obtain relative quantitation of proteins. Peptidesidentified in the control mice and in Dicer knockout mice liverare listed in Supplementary Tables 2 and 3, respectively. Weidentified 319 proteins that were differentially regulated(≥2-fold) in Dicer knockout mice. Among them, 257 proteinswere upregulated andmight be the targets of Dicer-dependent
Please cite this article as: Sahasrabuddhe NA., et al, Regulation oanalysis, J Prot (2014), http://dx.doi.org/10.1016/j.jprot.2014.04.027
ED P
RO
OF
miRNAs. MS and MS/MS spectra for representative peptidesfrom upregulated proteins, peroxisomal bifunctional enzymeand phosphoenolpyruvate carboxykinase 1 are illustrated inFig. 2.
The upregulated proteins include aldehyde oxidase (4.5-fold),myristoylated alanine-rich protein kinase C substrate (4.3-fold),sorbin, SH3 domain-containing protein 1 (3.4-fold), filamin A(3-fold) and phosphoenolpyruvate carboxykinase 1 (2.7-fold).Representative upregulated proteins are listed in Table 1. Ourdataset represents enrichment of cytochrome P450 familymembers which were upregulated in Dicer knockout liver suchas cytochrome P450 2d22 (15.4-fold), cytochrome P450 3a41b(9.4-fold), cytochrome P450 4v3 (5.8-fold), cytochrome P450 3a13(4.3-fold), cytochrome P450 2c70 (2.8-fold) and cytochrome P4502a12 (2-fold). Cytochrome P450 is a family of heme proteinsthat function as oxygenases. These perform diverse functionsincluding xenobiotic metabolism, fatty acid metabolism andsteroid hormone synthesis. Upregulation of cytochrome P450family of proteins can indicate elevation in the reactive oxygenspecies (ROS). Association of the polymorphisms in cytochromeP450 family members and hepatocellular carcinoma has beenestablished in previous studies [30,31].
The upregulated proteins potentially are targets of Dicer-dependent miRNAs, biogenesis of which is affected owing toablation of Dicer. The unchanged and downregulated proteinscan be potentially regulated by Dicer-independent miRNAs orbe regulated by the protein targets of Dicer-dependentmiRNAs.The downregulated proteins include 15-hydroxyprostaglandindehydrogenase (6-fold), cytochrome P450 2C29 (5.5-fold) andalcohol dehydrogenase 4 (4.5-fold).
3.4. PPARα targets are upregulated upon depletion of Dicer inliver
To identify enriched classes of molecules, we analyzed thedifferentially regulated proteins using DAVID. Our datasetreflected enrichment of proteins involved in PeroxisomeProliferator-Activated Receptor (PPAR) signaling. Other enrichedpathways included focal adhesion and regulation of actincytoskeleton. PPARs are steroid hormone nuclear receptors.There are multiple isoforms of PPAR with varying tissueexpression. PPARα is the predominant isoform expressed inliver [32]. It is known to target genes involved in fatty acidcatabolism, gluconeogenesis and ketone body synthesis andlipoprotein assembly [33,34].
f PPAR-alpha pathway by Dicer revealed through proteomic
A Peroxisomal bifunctional enzyme B Phosphoenolpyruvate carboxykinase 1
Fig. 2 – MS and MS/MS spectra of the representative upregulated proteins in Dicer knockout mice. A)Q2 MS and MS/MS ofLGILDVVVK from upregulated peroxisomal bifunctional enzyme in liver of Dicer knockout mice. B) MS and MS/MS ofGLGGVNVEELFGISK from upregulated phosphoenolpyruvate carboxykinase 1 in liver of Dicer knockout mice.
6 J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X
UNCO
RR
As depicted in Fig. 3, some of the major enzymatic stepsinvolved in peroxisomal β-oxidation of fatty acids includeoxidation, hydration, dehydrogenation and thiolytic cleavageof fatty acids. Acyl-CoA oxidase was upregulated (2-fold) inDicer knockout liver which acts upon bile acid intermediatesand branched chain fatty acids. The second and third stepsof β-oxidation involve hydratase and 3-hydroxyacyl-CoaAdehydrogenase, for which peroxisome harbors hydroxyste-roid (17-beta) dehydrogenase 4 or peroxisomal bifunctionalenzyme type 2 and peroxisomal bifunctional enzyme whichshow overlapping functions. Hydroxysteroid (17-beta) dehy-drogenase 4 was 1.4-fold upregulated while peroxisomalbifunctional enzyme was 3.7-fold upregulated in Dicer knock-out liver. Peroxisomal bifunctional enzyme is involved inproduction of medium chain dicarboxylic acids. Acaa1a is aperoxisomal 3-ketoacyl-CoA thiolase A which performs thelast step of β-oxidation. It was 4.1-fold upregulated in Dicerknockout liver.
The substrates for peroxisomal β-oxidation are end prod-ucts of ω-oxidation of fatty acids which is initiated bymicrosomal Cyp4 family of cytochromes. Cyp4V3 of mouse isa homologue of human Cyp4V2 which functions as fatty acidω-hydroxylase. Cyp4V3 was 5.8-fold upregulated in the Dicer
Please cite this article as: Sahasrabuddhe NA., et al, Regulationanalysis, J Prot (2014), http://dx.doi.org/10.1016/j.jprot.2014.04.027
knockout liver. Also, enzymes involved in α-oxidation of fattyacids such as phytanoyl-CoA dioxygenase (3.6-fold), verylong-chain acyl-CoA synthetase (2.3-fold) were upregulated.Though these enzymes are not reported to be a direct target ofPPARα, their upregulation along with the enzymes involved inperoxisomal β-oxidation, indicates over activation of oxida-tion of fatty acids. These events might have led to theobserved loss of weight in the Dicer knockout mice. Otherupregulated PPARα targets in Dicer knockout liver includeproteins involved in lipid transport such as ATP-bindingcassette sub-family D member 3 (11-fold), apolipoproteinA-IV (3.9-fold), perilipin 2 (3.4-fold) and solute carrier family27 (fatty acid transporter), member 2 (2.3-fold).
As described earlier, down-regulation of Dicer is reportedto be associated with hepatocellular carcinoma [9,35]. PPARαpathway is also observed to be associated with hepatocellularcarcinoma [36]. In our study, PPARα targets were upregulatedupon depletion of Dicer. We also observed overexpression ofcytochrome P450 family of proteins which can be attributed toelevation of ROS stress due to over activation of microsomaland peroxisomal fatty acid oxidation. The chronic effect ofDicer depletion on the PPAR signaling and the development ofhepatocellular carcinoma warrant further investigation.
of PPAR-alpha pathway by Dicer revealed through proteomic
Fig. 3 – Enrichment of PPARα targets in liver of Dicer knockout mice. Upregulated enzymes involved in ω-oxidation of fattyacids and peroxisomal β-oxidation of fatty acids and α-oxidation of fatty acids in liver of Dicer knockout mice. Most of theenzymes involved in β-oxidation have been reported to be PPARα targets. Key enzymes involved in the steps of peroxisomalβ-fatty acid oxidation included acyl-coA oxidase 2 (Acox2), beta-hydroxyacyl dehydrogenase (Hsd17b4), peroxisomalbifunctional enzyme (Ehhadh) and beta-ketothiolase A (Acaa1a). PPARα targets involved in lipid transport proteins were alsoupregulated in the liver of Dicer knockout mice.
7J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X
UNCO
RR3.5. Bioinformatics analysis of the enriched upstream miRNAs
The majority of precursor miRNAs depend on Dicer formaturation. We hypothesized that upon depletion of Dicer;miRNA maturation is severely affected, resulting in dysregu-lation of downstream miRNA targets. To identify the up-stream miRNAs of the upregulated proteins, we carried out abioinformatics analysis using TargetScan algorithm.
We identified 598 unique mouse miRNAs (mmu-miRs)predicted to be regulators of the upregulated proteins in liverof Dicer knockout mice. As listed in Supplementary Table 4,some of the predicted upstream miRNAs targeting multipleupregulated proteins included mmu-miR-124 and mmu-miR-143. miR-124 has been described as a tumor suppressorgene in recent studies carried out on hepatocellular carcinoma[37,38]. Vimentin (2-fold upregulated), a marker for epithelial–mesenchymal transition has been demonstrated previously asdirect target of miR-124 [38]. Apart from vimentin, multipleupregulated proteins associated with actin binding and mem-brane trafficking are predicted targets of miR-124. It is alsopredicted to regulate expression of peroxisomal bifunctionalenzyme. miR-143 was identified as upstream miRNA ofmultiple proteins involved in lipid metabolism or transportwhich included peroxisomal bifunctional enzyme, Abcd3 and
Please cite this article as: Sahasrabuddhe NA., et al, Regulation oanalysis, J Prot (2014), http://dx.doi.org/10.1016/j.jprot.2014.04.027
perilipin 2. Marcks was also predicted as a target of miR-143.miR-143 is known to be involved in regulation of lipidmetabolism [39,40]. It induces adipocyte differentiation withaccumulation of triglycerides. Along with downregulation ofmiR-143, downregulation of miR-124 probably results in upreg-ulation of the target genes involved in fatty acid oxidation.Further studies are required to investigate functions of thesecandidate miRNAs and their targets in liver. Also, studies arewarranted to advance our understanding of the correlation ofdownregulationofDicer,miRNAs, and PPAR-alpha downstreamtargets possibly with carcinogenesis.
3.6. Validation of candidate targets of Dicer using MRM assays
To validate upregulated proteins in liver of Dicer knockoutmice using a complementary mass spectrometry-basedmethod, we designed MRM assays. Candidate proteins wereselected based on the functional role of the proteins as well asthe predicted upstream miRNAs. Peroxisomal bifunctionalenzyme (Pbe) is predicted to be regulated by miRNA-124 andmiRNA-143, while myristoylated alanine-rich C-kinase sub-strate (Marcks) is predicted to be regulated by miRNA-143.Among the cytochrome P450 family of proteins, Cyp3a41 andCyp3a13 were found to be most dysregulated. Thus, MRM
f PPAR-alpha pathway by Dicer revealed through proteomic
Fig. 4 – Validation of Dicer targets. A) Four transitions were monitored for LGILDVVVK from peroxisomal bifunctional enzymein Dicer knockout and control mice. Box and whiskers plot depict relative abundance in liver of the Dicer knockout and controlmice. B) Western blot validation of peroxisomal bifunctional enzyme in liver of Dicer knockout and control mice. C) qRT-PCRanalysis to assay levels of miR-124 in liver of Dicer knockout and control mice.
t2:1t2:2t2:3t2:4
t2:5
t2:6
t2:7
t2:8
t2:9
t2:10
8 J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X
UNCO
Rassays were developed for Pbe, Pck1, Cyp3a41, Cyp3a13 andMarcks. We selected 4 control and 4 Dicer knockout mice. Atleast four transitions were monitored for a minimum of oneunique peptide for each protein. For normalization of data,housekeeping protein, isocitrate dehydrogenase (Idh2) wasselected as an unchanged internal standard from LC–MS/MSdata (Supplementary Fig. 1C). We monitored LGILDVVVK (z =+2, m/z = 478.3) and GWYQYDKPLGR (z = +2, m/z = 691.8) forvalidation of Pbe, a PPARα target which is involved inperoxisomal β-oxidation of fatty acids. As depicted inFig. 4A, LGILDVVVK was 5.2-fold upregulated in liver ofDicer knockout mice corroborating our findings. Similarly as
Table 2 – List of validated proteins upregulated in Dicer knocko
Please cite this article as: Sahasrabuddhe NA., et al, Regulationanalysis, J Prot (2014), http://dx.doi.org/10.1016/j.jprot.2014.04.027
listed in Table 2, we also monitored LTPIGYIPK (z = +2, m/z =501.3), LQEEIDETLPNK (z = +2, m/z 714.8), LQDEIDAALPNK(z = +2, m/z = 663.8), and VNGDASPAAAEPGAK (z = +2, m/z:677.8) corresponding to Pck1, Cyp3a41, Cyp3a13 and Marcks,respectively. Quantitation obtained using MRM-based assayswas in agreement with the quantitation obtained from theLC–MS/MS data. Monitored proteins, corresponding peptideswith transitions and applied collision energy are listed inSupplementary Table 5. We observed that MRM assaysprovided as a robust and antibody-independent approachfor validation of differentially expressed proteins in ourstudy.
9J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X
3.7. miR-124 — a predicted upstream miRNA of peroxisomalbifunctional enzyme is downregulated upon ablation of Dicerin liver
In order to investigate the correlation of miRNA and its target,we carried out qRT-PCR andWestern blot-based validation. Asdiscussed above, miRNA-124 was predicted to regulate severalupregulated proteins upon depletion of Dicer. One of thepredicted targets of miRNA-124 is peroxisomal bifunctionalenzyme. Upregulation of Pbe was also confirmed using MRMassays as discussed above. We further carried out Westernblotting for Pbe in 3 control and 3 Dicer knockout mice. Pbewas observed to be significantly upregulated in the Dicerknockout mice as shown in Fig. 4C. In order to assessexpression levels of miR-124, we carried out qRT-PCR exper-iments. We observed 4.5-fold downregulation of miR-124 inliver of Dicer knockout mice as depicted in Fig. 4D. Theseresults corroborated our hypothesis that certain miRNAsincluding miRNA-124 are likely Dicer-dependent miRNAsand their potential targets like peroxisomal bifunctionalenzyme are dysregulated upon ablation of Dicer.
T626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
CO
RREC
4. Conclusions
We investigated the global proteomic changes in mouse liverupon deletion of Dicer. We coupled in vivo metaboliclabeling-based proteomic screening with a mouse model. Wereport a number of proteins involved in the lipid metabolismthat were upregulated upon ablation of Dicer. We observedenrichment of PPARα targets in the upregulated proteins. Wevalidated upregulation of candidate protein targets ofDicer-dependent miRNAs including peroxisomal bifunctionalenzyme upon ablation of Dicer through MRM assays andWestern blotting. We also confirmed downregulation ofpredicted upstream miRNA of peroxisomal bifunctional en-zyme, miR-124. This indicates that certain Dicer-dependentmiRNAs regulate proteins involved in lipid metabolism whichare dysregulated upon ablation of Dicer. Further explorationof the candidate miRNAs can shed light on the association ofDicer with the regulatory network of lipid metabolism. Ourstudy illustrates an application of SILAC mouse as a powerfultool for quantitative proteomic characterization in vivo.
Supplementary data to this article can be found online athttp://dx.doi.org/10.1016/j.jprot.2014.04.027.
N 654
655
656
657
658
659
UConflict of interest
The authors declare no conflict of interest.
660
661
662
663
664
665
666
667
668
669
670
Acknowledgments
We thank the Department of Biotechnology (DBT), Govern-ment of India for research support to the Institute ofBioinformatics. Nandini A. Sahasrabuddhe is a recipient ofSenior Research Fellowship from the Council for Scientific andIndustrial Research (CSIR), India. Sartaj Ahmad is a recipientof Junior Research Fellowship from the University Grants
Please cite this article as: Sahasrabuddhe NA., et al, Regulation oanalysis, J Prot (2014), http://dx.doi.org/10.1016/j.jprot.2014.04.027
Commission (UGC), India. H.C. Harsha is a Wellcome Trust/DBT India Alliance Early Career Fellow. We thank ThermoScientific, India for access to instrumentation.
ED P
RO
OF
R E F E R E N C E S
[1] He L, Hannon GJ. MicroRNAs: small RNAs with a big role ingene regulation. Nat Rev Genet 2004;5:522–31.
[2] Iida A, Shinoe T, Baba Y, Mano H, Watanabe S. Dicer playsessential roles for retinal development by regulation ofsurvival and differentiation. Invest Ophthalmol Vis Sci2011;52:3008–17.
[3] Mudhasani R, Puri V, Hoover K, Czech MP, Imbalzano AN,Jones SN. Dicer is required for the formation of white but notbrown adipose tissue. J Cell Physiol 2011;226:1399–406.
[4] Nagalakshmi VK, Ren Q, Pugh MM, Valerius MT, McMahonAP, Yu J. Dicer regulates the development of nephrogenic andureteric compartments in the mammalian kidney. Kidney Int2011;79:317–30.
[5] Zhang L, Zhang B, Valdez JM, Wang F, Ittmann M, Xin L. Dicerablation impairs prostate stem cell activity and causesprostate atrophy. Stem Cells 2010;28:1260–9.
[6] Zehir A, Hua LL, Maska EL, Morikawa Y, Cserjesi P. Dicer isrequired for survival of differentiating neural crest cells. DevBiol 2010;340:459–67.
[7] Murchison EP, Partridge JF, Tam OH, Cheloufi S, Hannon GJ.Characterization of Dicer-deficient murine embryonic stemcells. Proc Natl Acad Sci U S A 2005;102:12135–40.
[8] Kanellopoulou C, Muljo SA, Kung AL, Ganesan S, Drapkin R,Jenuwein T, et al. Dicer-deficient mouse embryonic stem cellsare defective in differentiation and centromeric silencing.Genes Dev 2005;19:489–501.
[9] Sekine S, Ogawa R, Ito R, Hiraoka N, McManus MT, Kanai Y,et al. Disruption of Dicer1 induces dysregulated fetal geneexpression and promotes hepatocarcinogenesis.Gastroenterology 2009:e1–4 (136:2304-15).
[10] Zhang N, Bevan MJ. Dicer controls CD8+ T-cell activation,migration, and survival. Proc Natl Acad Sci U S A2010;107:21629–34.
[11] Korhonen HM, Meikar O, Yadav RP, Papaioannou MD, RomeroY, Da Ros M, et al. Dicer is required for haploid male germ celldifferentiation in mice. PLoS One 2011;6:e24821.
[12] Bernstein E, Kim SY, Carmell MA, Murchison EP, Alcorn H, LiMZ, et al. Dicer is essential for mouse development. NatGenet 2003;35:215–7.
[13] Huang TC, Sahasrabuddhe NA, Kim MS, Getnet D, Yang Y,Peterson JM, et al. Regulation of lipid metabolism by dicerrevealed through SILAC mice. J Proteome Res2012;11:2193–205.
[14] Tewari AK, Popova-Butler A, El-Mahdy MA, Zweier JL.Identification of differentially expressed proteins in bloodplasma of control and cigarette smoke-exposed mice by 2-DDIGE/MS. Proteomics 2011;11:2051–62.
[15] Golinska M, Troy H, Chung YL, McSheehy PM, Mayr M, Yin X,et al. Adaptation to HIF-1 deficiency by upregulation of theAMP/ATP ratio and phosphofructokinase activation inhepatomas. BMC Cancer 2011;11:198.
[16] Lane CS, Wang Y, Betts R, Griffiths WJ, Patterson LH.Comparative cytochrome P450 proteomics in the livers ofimmunodeficient mice using 18O stable isotope labeling. MolCell Proteomics 2007;6:953–62.
[17] Hood BL, Lucas DA, Kim G, Chan KC, Blonder J, Issaq HJ, et al.Quantitative analysis of the low molecular weight serumproteome using 18O stable isotope labeling in a lung tumorxenograft mouse model. J Am Soc Mass Spectrom2005;16:1221–30.
f PPAR-alpha pathway by Dicer revealed through proteomic
10 J O U R N A L O F P R O T E O M I C S X X ( 2 0 1 4 ) X X X – X X X
REC
[18] Iff J, Wang W, Sajic T, Oudry N, Gueneau E, Hopfgartner G,et al. Differential proteomic analysis of STAT6 knockout micereveals new regulatory function in liver lipid homeostasis. JProteome Res 2009;8:4511–24.
[19] Zhang Y, Askenazi M, Jiang J, Luckey CJ, Griffin JD, Marto JA. Arobust error model for iTRAQ quantification reveals divergentsignaling between oncogenic FLT3 mutants in acute myeloidleukemia. Mol Cell Proteomics 2010;9:780–90.
[20] Chaerkady R, Pandey A. Quantitative proteomics foridentification of cancer biomarkers. Proteomics Clin Appl2007;1:1080–9.
[21] Wu CC, MacCoss MJ, Howell KE, Matthews DE, Yates 3rd JR.Metabolic labeling of mammalian organisms with stableisotopes for quantitative proteomic analysis. Anal Chem2004;76:4951–9.
[22] Kruger M, Moser M, Ussar S, Thievessen I, Luber CA, Forner F,et al. SILAC mouse for quantitative proteomics uncoverskindlin-3 as an essential factor for red blood cell function.Cell 2008;134:353–64.
[23] Walther DM, Mann M. Accurate quantification of more than4000 mouse tissue proteins reveals minimal proteomechanges during aging. Mol Cell Proteomics 2011 (10:M110004523).
Singh TD, Nanjappa V, et al. Comprehensive proteomicanalysis of human bile. Proteomics 2011;11:4443–53.
[26] Huang da W, Sherman BT, Lempicki RA. Bioinformaticsenrichment tools: paths toward the comprehensivefunctional analysis of large gene lists. Nucleic Acids Res2009;37:1–13.
[27] MacLean B, Tomazela DM, Shulman N, Chambers M, FinneyGL, Frewen B, et al. Skyline: an open source document editorfor creating and analyzing targeted proteomics experiments.Bioinformatics 2010;26:966–8.
[29] Kashyap MK, Harsha HC, Renuse S, Pawar H, SahasrabuddheNA, Kim MS, et al. SILAC-based quantitative proteomicapproach to identify potential biomarkers from theesophageal squamous cell carcinoma secretome. Cancer BiolTher 2010;10:796–810.
UNCO
R
Please cite this article as: Sahasrabuddhe NA., et al, Regulationanalysis, J Prot (2014), http://dx.doi.org/10.1016/j.jprot.2014.04.027
ED P
RO
OF
[30] Mochizuki J, Murakami S, Sanjo A, Takagi I, Akizuki S,Ohnishi A. Genetic polymorphisms of cytochrome P450 inpatients with hepatitis C virus-associated hepatocellularcarcinoma. J Gastroenterol Hepatol 2005;20:1191–7.
[31] Li R, Shugart YY, Zhou W, An Y, Yang Y, Zhou Y, et al.Common genetic variations of the cytochrome P450 1A1 geneand risk of hepatocellular carcinoma in a Chinese population.Eur J Cancer 2009;45:1239–47.
[32] Auboeuf D, Rieusset J, Fajas L, Vallier P, Frering V, Riou JP,et al. Tissue distribution and quantification of the expressionof mRNAs of peroxisome proliferator-activated receptors andliver X receptor-alpha in humans: no alteration in adiposetissue of obese and NIDDM patients. Diabetes1997;46:1319–27.
[33] Mandard S, Muller M, Kersten S. Peroxisomeproliferator-activated receptor alpha target genes. Cell MolLife Sci 2004;61:393–416.
[34] Tachibana K, Kobayashi Y, Tanaka T, Tagami M, Sugiyama A,Katayama T, et al. Gene expression profiling of potentialperoxisome proliferator-activated receptor (PPAR) targetgenes in human hepatoblastoma cell lines induciblyexpressing different PPAR isoforms. Nucl Recept 2005;3:3.
[35] Wu JF, Shen W, Liu NZ, Zeng GL, Yang M, Zuo GQ, et al.Down-regulation of Dicer in hepatocellular carcinoma. MedOncol 2011;28:804–9.
[36] Yeldandi AV, Rao MS, Reddy JK. Hydrogen peroxidegeneration in peroxisome proliferator-induced oncogenesis.Mutat Res 2000;448:159–77.
[37] Hatziapostolou M, Polytarchou C, Aggelidou E, Drakaki A,Poultsides GA, Jaeger SA, et al. An HNF4alpha-miRNAinflammatory feedback circuit regulates hepatocellularoncogenesis. Cell 2011;147:1233–47.
[38] Furuta M, Kozaki KI, Tanaka S, Arii S, Imoto I, Inazawa J.miR-124 and miR-203 are epigenetically silencedtumor-suppressive microRNAs in hepatocellular carcinoma.Carcinogenesis 2010;31:766–76.
[39] Li G, Li Y, Li X, Ning X, Li M, Yang G. MicroRNA identity andabundance in developing swine adipose tissue as determinedby Solexa sequencing. J Cell Biochem 2011;112:1318–28.
[40] Esau C, Kang X, Peralta E, Hanson E, Marcusson EG,Ravichandran LV, et al. MicroRNA-143 regulates adipocytedifferentiation. J Biol Chem 2004;279:52361–5.
of PPAR-alpha pathway by Dicer revealed through proteomic
