Identification of Sucrose Non-Fermenting–Related - Diabetes

14 Pages·2013·1.66 MB·English

ORIGINALARTICLE
–
Identification of Sucrose Non-Fermenting Related
Kinase (SNRK) as a Suppressor of
Adipocyte Inflammation
Yujie Li,1,2 Yaohui Nie,1,3 Ynes Helou,4 Guoxian Ding,2 Bin Feng,1 Gang Xu,3 Arthur Salomon,5,6
and Haiyan Xu1,7
In this study, the role of sucrose non-fermenting–related kinase inflammatory cytokines, contributes to systemic insulin re-
(SNRK) in white adipocyte biology was investigated. SNRK is sistance through elevated circulating free fatty acid (FFA)
abundantlyexpressedinadiposetissue,andtheexpressionlevel levels. Multiple types of immune cells have been identified
is decreased in obese mice. SNRK expression is repressed by toregulateinflammatorypathwaysinobeseadiposetissue,
inflammatory signals but increased by insulin sensitizer in cul- such as macrophages, neutrophils, T cells, and mast cells
tured adipocytes. In vivo, adipose tissue SNRK expression can (6,7).Potentanti-inflammatoryeffects,suchassuppression
be decreased by lipid injection but enhanced by macrophage ofadiposemacrophagegeneexpressioninvitroandinvivo
ablation. Knocking down SNRK in cultured adipocytes activates and inhibition of proinflammatory mononuclear cells, have
both JNK and IKKb pathways as well as promotes lipolysis.
been reported for the only class of insulin sensitizer, thia-
Insulin-stimulated Akt phosphorylation and glucose uptake are
zolidinediones (8–11). Curbing inflammation of obese pa-
impaired in SNRK knockdown adipocytes. Phosphoproteomic
analysiswithSNRKknockdownadipocytesrevealedsignificantly tients with salsalate, a prodrug of salicylate for treating
decreasedphosphorylationof49proteinsby25%ormore,which arthritis, has been reported to improve glycemic control
areinvolvedinvariousaspectsofadipocytefunctionwithaclear (12). These results indicate that obesity-related adipose in-
indication of attenuated mTORC1 signaling. Phosphorylation of flammation plays an important role in the development of
43proteinsissignificantlyincreasedbyonefoldorhigher,among insulin resistance. However, the molecular pathways in-
whichseveralproteinsareknowntobeinvolvedininflammatory volved in the development of adipose inflammation in re-
pathways. The inflammatory responses in SNRK knockdown adi-
sponse to overnutrition are not fully understood.
pocytes can be partially attributable to defective mTORC1 signal-
Beingthemostextensivelystudiedmemberofthefamily,
ing,sincerapamycintreatmentactivatesIKKbandinduceslipolysis
AMPkinase(AMPK)hasbeendescribedasamasterenergy-
inadipocytes.Insummary,SNRKmayactasasuppressorofadi-
pocyteinflammationanditspresenceisnecessaryformaintaining sensing enzyme activated by increased AMP-to-ATP ratio.
normaladipocytefunction.Diabetes62:2396–2409,2013 Activation of AMPK turns on catabolic pathways that
generateATPwhileswitchingoffATP-consuminganabolic
pathways.Recentstudiesalsoindicateananti-inflammatory
S role for AMPKa1 in macrophages through diminishing
edentary lifestyle and excessive energy intake inflammasomeformationandactivationofsirtuin1(SIRT1)
havecausedobesityepidemics.Extensivestudies (13,14).Extensiveresearcheffortshaveestablishedaclear
have demonstrated that obesity-related insulin role for AMPK in energy metabolism in muscle, liver, and
resistanceandtype2diabetesareassociatedwith macrophages. In contrast, limited and contradictory in-
a low degree of inflammation in adipose tissue (1). Obese formation is available on the role of AMPK in adipocytes.
adiposetissuesecretesavarietyofinflammatorymarkers, AMPK phosphorylates Ser565 of hormone-sensitive lipase,
cytokines, and chemokines at elevated levels. Some of which was proposed to exert an antilipolytic effect
these factors, such as tumor necrosis factor (TNF)-a, in- through preventing phosphorylation of Ser563 by protein
terleukin (IL)-6, IL-1b, and MCP-1, have been reported kinase A (PKA) (15). However, the significance of this
to impair insulin signaling (2–5). Dysregulation of adipo- mechanism has been questioned because of the recent
cyte lipolysis, induced by increased expression of adipose finding that Ser563 is not essential for hormone-sensitive
lipaseactivation(16).Inaddition,bothpro-andantilipolytic
effects have been described for AMPK in adipocytes
Fromthe1HallettCenterforDiabetesandEndocrinology,RhodeIslandHos- (17,18). Although the AMPK activator AICAR stimulates
pital, Warren Alpert Medical School of Brown University, Providence,
Rhode Island; the 2Department of Geriatric Endocrinology, Jiangsu Prov- glucoseuptakeinbothmusclecellsandadipocytes,AMPK
inceHospital,NanjingMedicalUniversity,Nanjing,China;the3Department onlyappearstomediateAICAR-inducedglucoseuptakein
ofMedicineandTherapeutics,ThePrinceofWalesHospital,TheChinese muscle cells—not in adipocytes (19). AMPKa2 knockout
UniversityofHongKong,HongKong,China;the4DepartmentofMolecular
PharmacologyandPhysiology,BrownUniversity,Providence,RhodeIsland; mice have unchanged fat mass despite impaired glucose
the 5Department of Molecular Biology, Cell Biology and Biochemistry, tolerance when fed a normal chow diet, suggesting that
BrownUniversity,Providence,RhodeIsland;the6DepartmentofChemistry, AMPKa2 is not important for energy metabolism in adi-
Brown University, Providence, Rhode Island; and the 7Pathobiology Pro-
pose tissue of normal mice (20). Interestingly, AMPKa2
gram,BrownUniversity,Providence,RhodeIsland.
Correspondingauthor:HaiyanXu,[email protected]. knockout mice develop adipocyte hypertrophy when fed
Received10August2012andaccepted13March2013. a high-fat diet, indicating that AMPKa2 may be able to
DOI:10.2337/db12-1081
prevent excess lipogenesis in states of nutrition surplus
This article contains Supplementary Data online at http://diabetes
.diabetesjournals.org/lookup/suppl/doi:10.2337/db12-1081/-/DC1. (21). However, it is unclear whether this is because of the
(cid:1)2013bytheAmericanDiabetesAssociation.Readersmayusethisarticleas deficiency of AMPKa2 in adipocyte or because of the
longastheworkisproperlycited,theuseiseducationalandnotforprofit, secondary effects of AMPKa2 deficiency in liver or mus-
and the work is not altered. See http://creativecommons.org/licenses/by
-nc-nd/3.0/fordetails. cle because AMPKa2 is expressed at a low level in fat.
2396 DIABETES,VOL.62,JULY2013 diabetes.diabetesjournals.org
Y.LIANDASSOCIATES
AMPKa1 has been reported to be the major isoform of complementary strands were designed to target SNRK and ligated into the
AMPK in adipose tissue; yet, global AMPKa1 knockout Gateway-based pENTR/U6 vector (Invitrogen) and recombined into the
mice derived from the same founders were reported to BLOCK-iT RNAi system after sequence confirmation. As a negative control,
adenovirus expressing short hairpin (sh)GFP was used as previously de-
increase adiposity from one laboratory and decrease adi-
scribed (33). Amplification of recombinant adenovirus was performed
posity from another (18,22). These results raise questions accordingtothemanufacturer’sinstructionsusinghumanembryonickidney
regarding the importance of AMPKa1 in adipocyte func- 293Acells.
tion. RNAisolationandreal-timePCRanalysis.RNAsampleswereextracted
The biological functions of many AMPK-related family usingtheTRIZOLreagent.Geneexpressionwasmeasuredusingreal-timePCR
members have not been clearly defined. In this study, we analysis.Randomhexamerswereusedforreversetranscription.Real-timePCR
analysis was performed in a 15-mL reaction on 96-well clear plate using
describe the potential role of an AMPK-related kinase, su-
aPowerSYBRGreenRT-PCRReagentskitonABIPrismthermalcyclermodel
crosenon-fermenting–relatedkinase(SNRK),asapotential
StepOnePlus. The relative mRNA expression levels were normalized to ex-
suppressor of inflammation in adipocytes. As a family pression of b-actin or 28S rRNA. The sequences of primers are shown in
member of AMPK-related kinases, SNRK and its function SupplementaryTable1.
have not previously been studied in adipose tissue. SNRK ImmunoprecipitationandWesternblotanalysis.Forimmunoprecipitation
was initially cloned from fat cell cDNA library (23). The ofendogenousSNRKfromadiposetissue,thirtymicrolitersofA/Gplusmatrix
(SantaCruzBiotechnology)slurrywereusedtopreclear1mgproteinlysatesat
NH -terminal catalytic domain has a low homology to
2 4°Cfor30min.Then,3mgSNRKantibodywasaddedfor1hfollowedby40mL
AMPKa, but the noncatalytic domain is unique (23). SNRK
A/Gplusmatrixslurryforovernightincubationat4°C.Forimmunoblotanal-
can be activated by liver kinase B (LKB)1, the same up- ysis,immunoprecipitatedproteinor100mgproteinlysatesfromeachsample
stream kinase that activates AMPK (24). Gene knockdown was used. After PAGE on 4–12% gel (Bio-Rad Laboratories), the resolved
study in zebrafish indicates that SNRK may play a role in proteinsweretransferredontopolyvinylidenefluoridemembranes.Membranes
angioblast development (25,26). A recent study reported were blocked in 1% BSA/13 Tris-buffered saline with Tween (TBST) or 5%
milk/13TBSTfor1handthenincubatedwiththeappropriateprimaryanti-
thatSNRKinhibitscoloncancercellproliferation(27).Itis
bodiesinthepresenceof1%BSA/13TBSTor5%milk/13TBSTfollowedby
worthy to note thatSNRK is a completely different protein incubation with appropriate horseradish peroxidase–linked secondary anti-
fromsucrosenonfermentingAMPK-relatedkinase(SNARK), bodiesfor1hin5%milk/13TBST.Proteinbandsweredetectedbyenhanced
which is also a member of the AMPK/SNF1 family and can chemiluminescence(ECL)Westernblottingdetectionreagent(PerkinElmer).
be activated by LKB1 (28,29). Phosphoproteomicstudy.3T3-L1CARpreadipocyteswereculturedinSILAC
mediumforeightdoublingsandthenseededonsixwellplatesfordifferentiation
inSILACmedium.One-halfofthecellswerelabeledwithlightArgandLysand
infectedwithadenovirusexpressingshSNRK.Theotherhalfwerelabeledwith
RESEARCHDESIGNANDMETHODS heavyArgandLysandinfectedwithadenovirusexpressingshGFP.Foreach
Cells,reagents,andtreatments.3T3-L1cellswereobtainedfromAmerican sample,2.5mglightArg-andLys-labeledproteinlysateswasmixedwith2.5mg
Type Culture Collection. 3T3-L1 coxsackie virus and adenovirus receptor heavyArg-andLys-labeledproteinlysates.Sampleswerereducedwith45mmol/L
(CAR)cells,a3T3-L1sublinestablyexpressingthetruncatedadenovirusre- dithiothreitolat60°Cfor20min,alkylatedwith100mmol/Liodoacetimidefor
ceptor, were provided by Dr. David Orlicky (University of Colorado Health 15mininthedark,anddigestedovernightwithL-(tosylamido-2-phenyl)ethyl
Sciences Center) (30). Preadipocytes were differentiated as previously de- chloromethyl ketone (TPCK)-treated trypsin. Peptides were then desalted
scribed(31).FFAmixture,dexamethasone,insulin,andisobutylmethylxanthine using Sep-Pak C18 columns and enriched for phosphopeptides using Titan-
werepurchasedfromSigma.LiposynIIwaspurchased fromWebster Veteri- spherePhos-TiokitbeforeloadingtoOrbitrapVelosETDmassspectrometer.
nary.JunNH-terminalkinase(JNK),acetylCoAcarboxylase(ACC),andIkB Resultsfromfourbiologicalreplicateswereusedfordataanalysis.Qvalues
2
kinase (IKK)b antibodies were purchased from Santa Cruz Biotechnology. for multiple hypothesis tests were also calculated based on the determined
Phosphorylated(phospho-)JNKandphospho-IKKbantibodieswerepurchased PvaluesusingtheRpackageQVALUEaspreviouslydescribed(34).
fromCellSignalingTechnology.SNRKantibodywaspurchasedfromUniversity Statisticalanalysis.Studentttestwasusedtoanalyzeexperimentsinvolving
ofDundee(24).TubulinantibodywaspurchasedfromAbcam.Phospho-ACC twogroups.Two-wayANOVAandBonferroniposttestswereusedtoanalyze
antibody was purchased from Millipore. LC3 antibody was purchased from multipleexperimentalgroups.Errorbarsrepresentmeans6SE.
NovusBiologicals.Raptorexpressionconstructsandantibodieswereprovided
byDr.DianeFinger(UniversityofMichiganMedicalSchool).Clodronatelip-
osomes were purchased from clodronateliposomes.org (Vrije Universiteit, RESULTS
Amsterdam,the Netherlands)aspreviouslydescribed(32).Stableisotopela- Expression and regulation patterns of SNRK in
belingwithaminoacidsincellculture(SILAC)mediumwaspurchasedfrom adipocyte and adipose tissue. Through study of tissue
ThermoScientific.RapamycinwaspurchasedfromCalbiochem.Adenoviruses distribution patterns of .300 human kinases including all
expressinginduciblemyr-Akt1andthetetracyclinetransactivator(tTA)were
familymembersofAMPK/SNF1-relatedkinases,SNRKgene
purchasedfromCellBiolabs.
Mousemodels.Maleob/obmiceandlittermatecontrolswerepurchasedfrom was found to be abundantly and predominantly expressed
TheJacksonLaboratory.Thesemicewerefedstandardchowandkilledat9 in white adipose tissue in human and in both white and
weeks ofage fortissue collection. ForDIO mice,C57BL/6Jmicewere pur- brown adipose tissue in mouse (Fig. 1A and B). SNRK
chasedfromTheJacksonLaboratoryat3weeksofage,acclimatedforaweek, protein expression pattern confirmed its predominant
andfedoneitherachowdiet(5%kcalfromfat)orahigh-fatdiet(60%kcal
presence in adipose tissue (Fig. 1C). AMPKa1 has been
fromfat,D12492;ResearchDiets)forupto20weeks.Formacrophageabla-
shown to be the major isoform of AMPK in adipose tissue
tionexperiment,micewereinjectedwithclodronateliposomesatthedoseof
110mg/kg.Attheendofthestudy,micewerekilledbyCO inhalationand (18).However,therelativemRNAexpressionlevelofSNRK
2
epididymalfatpadswerecollected.Allanimalexperimentswereapprovedby is74%higherthanthatofAMPKa1inhumanwhiteadipose
theInstitutionalAnimalCareandUseCommitteeofRhodeIslandHospital. tissue and 2.1-fold higher than that of AMPKa1 in mouse
Isolationofprimaryadipocytes.EpididymalwhitefatpadsfromDIOmice adipose tissue. Expression levels of AMPKa2 are very low
wereexcised,weighed,andrinsedinisolationbuffer.Fatpadswerethencutinto
in both human and mouse adipose tissue, supporting the
smallpiecesinisolationbuffersupplementedwith1mg/mLtypeIcollagenase
hypothesisthatAMPKa2maynotplayanimportantrolein
and digested at 37°C in shaking-water bath at 100 rpm for 45 min. Digested
tissues were filtered through400 mmol/L meshto get single cell suspension. adipocyte energy metabolism. SNRK has a nuclear locali-
Aftercentrifugation,floatingadipocyteswererinsedtwicewithisolationbuffer. zationsignalandhasbeenshowntobelocalizedinnucleus
Adenovirusconstruction.Forconstructionofadenoviralvectorformouse of cultured rat cerebellar granule neurons by immunohis-
SNRK,thecodingsequencewasamplifiedbyPCR,clonedintotheentryvector, tochemistry (35). For localization of SNRK in adipocyte,
andsequenceconfirmed.Thecodingsequencewasthenrecombinedintothe
SNRK-GFP fusion protein was expressed via adenoviral-
Gateway-based pAd-CMV DEST vector (Invitrogen) according to the manu-
facturer’s instructions. For construction of adenoviral vector for short hair- mediated gene transfer and revealed a large overlapping
pin interfering RNA against SNRK, six short hairpin oligonucleotides and withlysosomaltracker(Fig.1D),indicatingthatSNRKmay
diabetes.diabetesjournals.org DIABETES,VOL.62,JULY2013 2397
SNRKANDADIPOCYTEINFLAMMATION
FIG.1.TissuedistributionofSNRKandAMPKa.A:RelativemRNAexpressionlevelsofSNRK,AMPKa1,andAMPKa2inhumantissues.RNA
samplesfrompooledhumantissueswerepurchasedfromClonetech(stomachcat.no.636126,smallintestine636125,pancreas636119,white
adiposetissue636162,lung636105,heart636113,testis636115,liver636101,kidney636118,brain636102,spleen636121,andskeletalmuscle
636120).B:RelativemRNAexpressionlevelsofSNRK,AMPKa1,andAMPKa2inmousetissues.TissuesfromfourmaleC57BL/6micewerepooled
forRNApreparation.C:SNRKproteinlevelsinmousetissues.TissuesfromfourmaleC57BL/6micewerepooledforproteinpreparation.SNRK
wasimmunoprecipitatedfrom1mgproteinlysates.D:SNRKlocalization.TheSNRK-GFPfusionproteinwasexpressedin3T3-L1CARadipocytes
viaadenovirus-mediatedgenetransfer.Forty-eighthoursafterinfection,adipocyteswerestainedwithLysotracker(10,0003,cat.no.L-7528;
Invitrogen)(red)forpresenceoflysosomesandDAPI(finalconcentrationat1mg/mL)(blue)forpresenceofnucleusduringa2-hincubation.The
imageswereoverlaid,andtheorangecolorindicatesoverlappingofSNRK-GFPandLysotracker.BAT,brownadiposetissue;S.,small;WAT,white
adiposetissue.
2398 DIABETES,VOL.62,JULY2013 diabetes.diabetesjournals.org
Y.LIANDASSOCIATES
FIG.2.RegulationofSNRKexpressioninadiposetissue.A:RegulationofSNRKgeneexpressioninadiposetissueofob/obmice(toppanel)(n=5
pergroup)andhigh-fatdietDIOmice(bottompanel)(n=4pergroup).B:RegulationofSNRKproteinexpressioninadiposetissueofob/obmice
(n=3pergroup).C:RegulationofSNRKproteinexpressioninadiposetissueofDIOmice(n=4pergroup).D:RegulationofSNRKgeneex-
pressioninadipocytesandstromalvascularcellsisolatedfromchow-fedandDIOmice(fatpadswerepooledfrom6–8miceineachgroup).Adi,
adipocytes;HF,highfat;SV,stromalvascularcells.*P<0.05.Errorbarsstandformean6SE.
have different cellular localization patterns in different cell and protein levels (Fig. 2A–C). Further analysis indicates
types. SNRK expression in white adipose tissue is signifi- that the decrease of SNRK gene expression in adipose tis-
cantly reduced in both ob/ob and DIO mice at both mRNA sue of DIO mice occurs in adipocyte, not in stromal
diabetes.diabetesjournals.org DIABETES,VOL.62,JULY2013 2399
SNRKANDADIPOCYTEINFLAMMATION
vascular, cells (Fig. 2D). Interestingly, SNRK mRNA and FFAandTNF-a,indicatingthatactivationofinflammatory
proteinlevelscanbeinducedbyfastinginadiposetissueof pathwayssuppressesSNRKexpression.Theconstitutively
lean mice, but this regulation is lost in DIO mice (Fig. 2A active form of IkB kinase b (IKKb SE), a master kinase
andC),suggestingthatSNRKmaybeinvolvedinregulating that turns on transcription of many inflammatory genes,
metabolism in response to fasting in lean mice. was also overexpressed in 3T3-L1 CAR adipocytes by
ItiswellestablishedthatFFAreleaseandproductionof adenovirus-mediated gene transfer. IKKb SE significantly
inflammatory factors are increased in adipose tissue of repressed SNRK gene expression (Fig. 3C). In addition,
obese state and contribute to the development of insulin significantlydecreasedSNRKgeneexpressionisobserved
resistance and hyperglycemia. For exploration of whether in isolated adipocytes prepared from lean mice treated
these factors are potentially responsible for down- with intravenous injection of liposyn II (Fig. 3D), sug-
regulating SNRK expression in adipose tissue, FFAs and gesting that the decreased SNRK expression in obese
proinflammatory factor TNF-awereusedtotreatcultured adipose tissue is likely due to the occurrence of obesity-
adipocytes. As shown in Fig. 3A and B, SNRK gene ex- induced adipose inflammation and/or lipid toxicity. In
pression was significantly reduced by treatmentwith both contrast,treatmentwithrosiglitazone,aninsulinsensitizer
FIG.3.EffectofproinflammatorysignalsonSNRKexpression.A:EffectofFFAonSNRKgeneexpressioninculturedadipocytes.FFAmixture
(cat.no.F7050;Sigma)wasusedatafinalconcentrationof0.25mmol/Lfor6h.B:EffectofTNF-aonSNRKgeneexpressioninculturedadi-
pocytes.TNF-awasusedatafinalconcentrationof25ng/mLfor6h.C:EffectofoverexpressingtheconstitutivelyactiveIKKb(IKKbSE)on
SNRKgeneexpressioninL1-CARadipocytes.D:EffectofintravenouslipidinjectioninleanmiceonSNRKgeneexpressioninisolatedadipocytes.
Epididymalfatpadswerepooledfromthreemiceineachgroupforadipocyteisolation30minafterinjectionwithliposynIIat3mL/kg/h.E:Effect
ofrosiglitazone(Rosi)treatmentonSNRKgeneexpressioninculturedadipocytes.Rosiglitazonewasusedatafinalconcentrationof10mmol/L
for6h.F:EffectofadiposetissuemacrophagedepletiononSNRKgeneexpressioninleanandDIOmice(n=4ineachgroup).Micewere23weeks
oldandonahigh-fatdietfor19weeks.ClodronateorPBSliposomeswereinjectedintoperitonealcavityatthedoseof110mg/kg.G:EffectofFFA
onSNRKproteinexpressioninculturedadipocytes.H:EffectofTNF-aonSNRKproteinexpressioninculturedadipocytes.I:Effectofover-
expressing IKKb SE on SNRK protein expression in L1-CAR adipocytes. J: Effect of rosiglitazone treatment on SNRK protein expression in
culturedadipocytes.*P<0.05.Forcell-basedexperiments,triplicatesampleswereusedingeneexpressionandduplicatesampleswereusedin
proteinexpression.Resultsshownarerepresentativefromthreeindependentexperiments.Errorbarsstandformean6SE.Ab,antibody;Clod,
clodronate liposomes; F, FFA; Lipid, liposyn II; PBS, PBS liposomes; R, rosiglitazone; T, TNF-a; V, vehicle; Veh, vehicle. (A high-quality color
representationofthisfigureisavailableintheonlineissue.)
2400 DIABETES,VOL.62,JULY2013 diabetes.diabetesjournals.org
Y.LIANDASSOCIATES
withpotentanti-inflammationeffect,increasedSNRKgene Effect of SNRK knockdown in cultured adipocytes.
expression in cultured adipocytes (Fig. 3E). Ablation of For understanding of the function of SNRK, adenovirus-
adipose macrophages, an important source of inflamma- mediatedshorthairpininterferingRNAwasusedtoknock
tory factors in obesity, by clodronate liposomes signifi- down SNRK expression in 3T3-L1 CAR adipocytes. Ex-
cantly increased SNRK gene expression in adipose tissue pression levels of SNRK mRNA and protein were signifi-
ofDIOmice(Fig.3F).Furtherexperimentsconfirmedthat cantlyreducedbyinterferingRNAagainstSNRKcompared
SNRK protein expression levels are also decreased by in- withcontroladipocytesexpressinginterferingRNA against
flammatory signals and increased by rosiglitazone in cul- GFP (Fig. 4A). Reduction of SNRK expression in 3T3-L1
tured adipocytes (Fig. 3G–J). CAR adipocytes activated IKKb and JNK pathways, as
FIG.4.EffectofSNRKknockdownonadipocytebiology.A:SNRKknockdownbyadenovirus-mediatedshRNAonmRNA(toppanel)andprotein
(bottompanel)levels.B:EffectofSNRKknockdownonphosphorylationofIKKbandJNK.C:EffectofSNRKknockdownonlipolysis.D:Effectof
SNRK knockdown on expression of TNF-a, IL-6, MCP-1, MCP-2, MCP-3, and chop genes. E: Effect of SNRK knockdown on TNF-a secretion.
F:EffectofSNRKknockdownonIL-6secretion.G:EffectofSNRKknockdownonexpressionofGlut4,Adipo,VLACAD,MACAD,andAtg12genes.
H:EffectofSNRKknockdownoninsulin-stimulatedAktphosphorylation.I:EffectofSNRKknockdownoninsulin-stimulatedglucoseuptake.*P<
0.05.Triplicatesampleswereusedineachexperiment.Resultsshownwererepresentativefromthreeindependentexperiments.Errorbarsstand
formean6SE.S,Ser;Veh,vehicle.
diabetes.diabetesjournals.org DIABETES,VOL.62,JULY2013 2401
SNRKANDADIPOCYTEINFLAMMATION
TABLE1
Phosphoproteins decreasedby25%ormoreinSNRKknockdown adipocytes
KD-to-
Phosphorylation control
Symbol Name sites ratio QforSILAC
Insulinsignaling/protein
synthesis/aminoacid
metabolism
RPTOR Regulatory-associatedproteinofmTOR S863 0.25 0.031
PDCD4 Programmedcelldeathprotein4 T93,S94 0.26 0.010
IRS1 Insulinreceptorsubstrate1 S1097 0.29 0.030
PI4K2A Phosphatidylinositol 4-kinasetype2-a S47 0.45 0.025
PEA15 AstrocyticphosphoproteinPEA-15 S116 0.51 0.025
BCKDK 3-methyl-2-oxobutanoate S33 0.67 0.029
dehydrogenasekinase,mitochondrial
AKT1S1 PRAS40 T247 0.75 0.025
Transcription
HDAC1 Histone deacetylase1 S393 0.18 0.023
TCOF1 Treacleprotein S1191 0.23 0.006
NF1 Nuclearfactor1 S265 0.66 0.040
PTRF PolymeraseI andtranscript T198,S218 0.74 0.046
releasefactor
mRNAprocessing
SNRNP70 Snrnp70protein S81 0.11 0.026
SF1 Splicingfactor1 S80,S82 0.39 0.029
SCAF1 Splicingfactor,Arg/Ser-rich 19 S676,S682 0.41 0.044
Smg9 ProteinSMG9 S32 0.47 0.026
SRRM1 Ser/Argrepetitivematrix 1 S387,S391 0.47 0.011
CSTF3 Cleavagestimulation factorsubunit 3 S691 0.56 0.039
G3BP2 RasGTPase-activatingprotein– T227 0.61 0.027
bindingprotein2
FIP1L1 Pre-mRNA39-end-processing factorFIP1 S479 0.66 0.031
Translation
EEF1D Eef1dprotein S128 0.38 0.022
Glucoseandlipid
metabolism
Pck2 Phosphoenolpyruvate S2 0.11 0.010
carboxykinase(GTP),mitochondrial
PC Pyruvatecarboxylase,mitochondrial S19,S20 0.42 0.007
ACC1 Acetyl-CoA carboxylase1 S29 0.52 0.025
ATGL Patatin-like phospholipase S406 0.67 0.026
domain–containingprotein2
ATGL Patatin-like phospholipase S409 0.70 0.037
domain–containingprotein2
OSBP Oxysterol-bindingprotein S188,S191 0.70 0.026
ACLY ATPcitratelyase S455 0.71 0.023
OSBPL11 Oxysterol-bindingprotein– S186 0.72 0.025
related protein11
Cytoskeleton
VCL Vinculin S290 0.00035 0.009
Sdccag8 Serologically definedcolon T430,S438 0.0013 0.023
cancerantigen8homolog
Mylk Myosin,light polypeptidekinase S611 0.013 0.023
Ppfia1 Ppfia1protein T770 0.16 0.025
RABGAP1 RabGTPase-activatingprotein1 S42 0.44 0.023
PXN Paxillin S321,S328 0.63 0.022
SH3PXD2A SH3andPXdomain–containing S420 0.67 0.025
protein2A
Tpr Nuclearporecomplex–associated S2067 0.74 0.031
intranuclearcoiled-coilproteinTPR
Membranetrafficking
PLVAP Plasmalemmavesicle–associated protein S4 0.40 0.040
Tgoln1 trans-Golgi networkintegral S266,S267 0.56 0.035
membraneprotein1
Tgoln2 trans-Golgi networkintegral S277 0.63 0.016
membraneprotein2
Continuedonfacingpage
2402 DIABETES,VOL.62,JULY2013 diabetes.diabetesjournals.org
Y.LIANDASSOCIATES
TABLE1
Continued
KD-to-
Phosphorylation control
Symbol Name sites ratio QforSILAC
TRAM1 Translocatingchain–associated S365 0.73 0.028
membraneprotein1
CANX Calnexin S563 0.75 0.037
Miscellaneous
SSFA2 Sperm-specificantigen2 S90 0.11 0.022
SASH1 SAMandSH3domain–containingprotein1 S805 0.13 0.009
CLASP2 Clasp2protein S376 0.45 0.042
SDPR Serumdeprivation–responseprotein S363 0.49 0.029
THUMPD1 THUMPdomain–containingprotein1 S86,S88 0.50 0.030
HMOX1 Hemeoxygenase1 S174 0.57 0.047
FAM73B ProteinFAM73B T208 0.68 0.011
BAG3 BAGfamilymolecularchaperoneregulator3 S270,S274 0.69 0.034
LUC7I3 Luc7-likeprotein3 S425,S431 0.71 0.044
Experimentsweredonewithfourindependentreplicates,anddatawereanalyzedwithratioofmean.KD,knockdown.
shown by increased phosphorylation (Fig. 4B), which is significantly reduced signals of 55 phosphopeptides by 25%
accompanied by increased basal lipolysis as reflected by or more, which are derived from 49 proteins involved in
elevated glycerol release (Fig. 4C). Consistent with acti- various aspects of cell function (Table 1), including tran-
vation of inflammatory pathways, markedly increased scription, mRNA processing, translation, glucose and lipid
expression levels of proinflammatory cytokines, includ- metabolism, and cytoskeleton. Most of the phosphosites
ing TNF-a and IL-6, were observed in SNRK knockdown have not been functionally annotated, but reduced phos-
3T3-L1 CAR adipocytes compared with control cells phorylation occurs in factors normally required for tran-
(Fig.4D–F).TNF-aisapotentinduceroflipolysis,which scription initiation, RNA processing, and protein synthesis.
may contribute to increased lipolysis. It has been pre- Interestingly, several components of mTORC1-signaling
viously reported that multiple monocyte chemotactic pathwayhavereducedphosphorylation,includingregulatory-
factors are upregulated in adipose tissue and adipocytes associated protein of mTOR (raptor) on Ser863, PDCD4 on
of DIO mice (31). Reduction of SNRK expression in cul- Thr93andSer94,insulinreceptorsubstrate(IRS)-1onSer1097,
turedadipocytesalsoledtoincreasedgeneexpressionof and proline-rich Akt1 substrate 1 (PRAS40) on Thr247.
three monocyte chemotactic factors, MCP-1, MCP-2, and Phosphorylation on Ser863 of raptor activates mTORC1 sig-
MCP-3, indicating that decreased SNRK expression in naling(37).Ser1097ofIRS-1isreportedtobeasiteforS6K
adipocytes may potentially enhance macrophage infiltra- phosphorylation(38).PRAS40inhibitsmTORsignaling,and
tioninobesity.ReducedSNRKexpressionalsoincreased Akt phosphorylates PRAS40 on Thr247 to relieve the inhi-
expression of chop gene. In contrast, mRNA expression bition. Decreased phosphorylation on raptor Ser863, IRS-1
levels of insulin-responsive glut4 and insulin-sensitizing Ser1097, and PRAS40 Thr247 indicates attenuated mTOR
adiponectin are significantly decreased upon SNRK knock- signaling. Phosphorylation of ACC1, a well-characterized
down (Fig. 4G). Insulin-stimulated Akt phosphorylation on AMPKsubstrate,isdecreasedonSer29,whichisanunchar-
Ser473 and glucose uptake are significantly reduced (Fig. acterizedsite.Inaddition,SNRKknockdownalsodecreased
4H–I). Expression levels of very-long-chain acyl-coA de- phosphorylationofseveralenzymesinvolvedinglucoseand
hydrogenase and medium-chain acyl-coA dehydrogenase lipid metabolism, including mitochondrial phosphoenolpyr-
are also significantly reduced in SNRK knockdown adipo- uvate carboxykinase, mitochondrial pyruvate carboxylase,
cytes, suggesting impaired fatty acid oxidation. Further- patatin-like phospholipase domain–containing protein 2,
more, Atg12 gene was significantly decreased, suggesting ATP citrate lyase, oxysterol-binding protein, and oxysterol-
deficiency in autophagy. binding protein–related protein 11.
EffectofSNRKknockdownonproteinphosphorylation Reduction of SNRK expression in adipocytes also sig-
in cultured adipocytes. For a broad understanding re- nificantly increased signals of 55 phosphopeptides by
garding the function of SNRK in adipocytes and the down- onefold or more, which are derived from 43 proteins (Ta-
stream signaling events, a quantitative phosphoproteomic ble 2). These proteins also affect various aspects of adi-
approach was used to evaluate changes of global phos- pocyte function. Distinct from the dataset of decreased
phorylationinSNRKknockdownadipocytescomparedwith phosphorylation, several proteins involved in inflamma-
control adipocytes expressing shGFP. The SILAC method tory pathways have increased phosphorylation. Among
wasusedtoensureaccuratecomparison.SNRKknockdown these proteins, caspase recruitment domain family mem-
adipocyteswerelabeledwithlight(12Cor14N)ArgandLys, ber 6 (Card6), IFIH1, ZBP1, and receptor-interacting Ser/
whereascontroladipocyteswerelabeledwithheavy(13Cor Thr–proteinkinase2areknowntoactivatenuclearfactor-
15N) Arg and Lys. All peptide sequence assignments were kBtranscriptionfactors,thesubstrateofIKKb.Interferon-
filtereddownto1%falsediscoveryratebyalogisticspectral induced double-stranded RNA-activated protein kinase
score (36). Results shown in Supplementary Tables2 and 3 (PRKRA) has been characterized as a critical component
arethestatisticallysignificant(qvalue,0.05)averagepeak of an inflammatory complex that activates JNK and IKKb
ratiosoflightversusheavyaminoacidsfromfourbiological pathways and induces insulin resistance (39). Phosphory-
replicates. Reduction of SNRK expression in adipocytes lation levels of multiple sites of protein PML, a potent
diabetes.diabetesjournals.org DIABETES,VOL.62,JULY2013 2403
SNRKANDADIPOCYTEINFLAMMATION
TABLE2
Phosphoproteins increasedbyonefoldormoreinSNRKknockdownadipocytes
Symbol Name Phosphorylationsites KD-to-control ratio QforSILAC
Inflammatory
pathways/apoptosis
Card6 Caspaserecruitmentdomain S870, T873 687.5 0.0065
family, member6
IFIH1 Interferon-inducedhelicaseC S289 11.2 0.0397
domain–containingprotein1
ZBP1 Z-DNA–bindingprotein1 S384 7.5 0.0334
PML ProteinPML S515, T527,S528 4.8 0.0270
PML ProteinPML S514 4.5 0.0440
RIPK2 Receptor-interacting Ser/Thr– S414 3.7 0.0372
proteinkinase2
SAMHD1 Samdomain andHDdomain– T21,S24,T29 2.6 0.0221
containingprotein1
H2-L H-2classIhistocompatibility S353, S356 2.3 0.0293
antigen,L-Dachain
SAMHD1 Samdomain andHDdomain– T25 2.2 0.0397
containingprotein1
Insulinsignaling/protein
synthesis
Sik2 Ser/Thr–proteinkinaseSIK2 S358 16.2 0.0095
PI4KB Phosphatidylinositol 4-kinaseb S511 7.2 0.0230
PRKRA Interferon-induceddouble- S32 4.7 0.0264
strandedRNA-activated
proteinkinase
Transcription/messenger
RNAprocessing/
translation
RBM15 RNAbindingmotif protein15 S655 6.8 0.0497
Trim28 Transcriptionintermediary S473 3.5 0.0308
factor1-b
EIF4G3 Eukaryotictranslationinitiation S267 2.9 0.0095
factor4g3
EIF4G1 Eukaryotictranslationinitiation S1189 2.7 0.0434
factor4g1
RBM8A RNA-bindingprotein8A S56 2.7 0.0317
Ifi204 Interferon-activableprotein204 T106 2.3 0.0331
INTS1 Integratorcomplexsubunit1 S1320,T1321,S1328,S1329 2.3 0.0246
Srrm1 Ser/Argrepetitive matrixprotein1 S610, S612 2.3 0.0313
Patl1 ProteinPAT1homolog1 S179 2.2 0.0120
NFIB Nuclearfactor1B-type S328 2.0 0.0461
Glucoseandlipid
metabolism
PLIN1 Perilipin1 S81 3.0 0.0366
PLA2G4B CytosolicphospholipaseA2 S437 2.6 0.0230
Cytoskeleton/scaffolding
protein/endocytosis/
exocytosis
AaK1 AP-2–associatedproteinkinase1 T618,S621 80.6 0.0065
AKAP13 Akap13Akinaseanchorprotein13 S1910 4.1 0.0340
MAPT Microtubule-associatedprotein S761 3.5 0.0289
CLASP1 CLIP-associatingprotein1 S600 3.5 0.0246
Sept2 Septin-2 S218 2.8 0.0308
ACTB Actin,cytoplasmic1 S239 2.8 0.0441
Ctnnb1 Cateninb-1 T556 2.4 0.0441
AP3D1 AP-3complexsubunitD-1 S755, T758,S760 2.1 0.0230
Cell growth
MEK5 Mitogen-activatedproteinkinase S433 3.5 0.0321
kinasekinasekinase5
HDGF Hepatoma-derivedgrowthfactor S165 3.5 0.0332
Miscellaneous
Likelyortholog ofHomosapiens T311 1030.4 0.0065
chromosome9ORF138
TNKS1BP1 182kDatankyrase-1–bindingprotein S429 5.5 0.0387
Continuedonfacingpage
2404 DIABETES,VOL.62,JULY2013 diabetes.diabetesjournals.org
Y.LIANDASSOCIATES
TABLE2
Continued
Symbol Name Phosphorylationsites KD-to-control ratio QforSILAC
TMCC2 Transmembraneandcoiled-coil S435 5.5 0.0366
domainsprotein2
Testisexpressedgene2 S195 5.5 0.0276
Arhgef17 rguaninenucleotide S458 3.2 0.0380
exchangefactor17
Myl12a MCG5400 T19,S20 3.0 0.0343
DCLK1 Dclk1protein S330, S332,T336 2.9 0.0246
Tjp1 Tjp1 S617 2.9 0.0481
Rplp2-ps1 MCG10050 S105 2.5 0.0246
Marcksl1 MARCKS-relatedprotein S104 2.2 0.0492
Nop58 NUcleolarprotein58 S509, S521 2.2 0.0366
Experimentsweredonewithfourindependentreplicates,anddatawereanalyzedwithratioofmean.
growth suppressor and proapoptotic factor, are signifi- knockdown adipocytes, both LC3I and LC3II were signifi-
cantly increased. In addition, increased phosphorylation cantly reduced after normalization to tubulin control, but
occurs in proteins normally involved in repression of tran- LC3II-to-LC3I ratio was not affected (Fig. 6A). Reduction
scription(Trim28,Ifi204,andHDGF),promotionofmRNA of SNRK also significantly reduced phosphorylation levels
degradation (IFIH1 and Patl1), and inhibition of protein of raptor S863 and ACC S79. Defective autophagy has
synthesis (PRKRA). Phosphorylation also increases in pro- been linked to induction of inflammation and may be
teins involved in hydrolysis of glycerophospholipids (phos- partially responsible for the inflammatory phenotype in
pholipaseA2onSer437)andtriglycerides(perilipinonSer81, adipocytes with reduced SNRK expression. In addition,
a PKA phosphorylation site), consistent with observed phosphoproteomic study revealed many potential down-
increase of basal lipolysis. Phosphorylation on Ser358 of stream targets of SNRK. The mTORC1 pathway stood
SIK2 increased 15-fold, a site that has been reported to out owing to decreased phosphorylation of several com-
be phosphorylated by PKA in response to forskolin and ponents. To explore whether defective mTOR signaling
CL316243 in adipocytes (40). pathway contributes to adipocyte inflammation, 3T3-L1
Effect of SNRK overexpression in hepatocytes. De- adipocytesweretreatedwithrapamycinandexaminedfor
spite successful SNRK knockdown in 3T3-L1 CAR adipo- activations of JNK and IKKb pathways, lipolysis, and ex-
cytes, SNRK protein cannot be overexpressed at a high pression of inflammatory cytokines. Treatment with rapa-
level in this cell type in spite of hugely overexpressed mycin increased phosphorylation of IKKb but not JNK
SNRK mRNA, but SNRK-GFP fusion protein could be and stimulated lipolysis (Fig. 6B and C). Gene expression
expressed at a low level that is observable under fluores- levelsofproinflammatoryfactorsIL-6,MCP-1,andMCP-3
cence microscope. One possibility is that the COOH- are also significantly upregulated (Fig. 6D). Consistent
terminal of SNRK contains two regions of PEST-rich with these results, activation of mTOR signaling by over-
sequences, which are motifs indicative of rapid protein expressingtheconstitutivelyactiveAkt1(myr-Akt)repressed
turnover (23). In addition, there may be a mechanism in lipolysis and decreased expression of proinflammatory fac-
adipocytes to limit SNRK overload due to abundantly tors (Fig. 6E–F). These data indicate that defective mTOR
expressed endogenous SNRK protein in this cell type. We signaling in SNRK knockdown adipocytes may be partially
therefore used a liver cell line to examine whether SNRK responsible for the inflammatory phenotype and that other
overexpression can increase phosphorylation of raptor pathways are likely involved, since JNK is not activated by
and ACC. As shown in Fig. 5A, SNRK protein was over- rapamycin treatment.
expressedinFaohepatomacellsandsignificantlyincreased
phosphorylation of ACC on Ser79 (an AMPK phosphoryla- DISCUSSION
tion site) and raptor on Ser863. SNRK kinase assay con- The underlying molecular mechanism responsible for orig-
firmed that overexpressed SNRK is active in hepatoma ination of inflammation in adipocyte upon overnutrition
cells (Fig. 5B). For determination of whether raptor and is still not fully understood. In this study, we identified
SNRKexistinthesameproteincomplex,raptorandSNRK SNRKasapotentialsuppressorofadipocyteinflammation
were coexpressed in 293A cells. Immunoprecipitation of and hypothesize that decreased SNRK expression in adi-
SNRK brought down raptor as evidenced by immunoblot pose tissue may contribute to the development of adipose
analysis (Fig. 5C). SNRK overexpression also suppressed inflammation commonly observed in obesity. AMPKa has
chop gene expression (Fig. 5D), consistent with increased been reported to repress inflammation in macrophages
chop gene expression in SNRK knockdown adipocytes. and polarize macrophages to an anti-inflammatory phe-
Interestingly,SNRKoverexpressionsignificantlyincreased notype (13,14,41). AMPKa-deficient bone marrow macro-
LC3II-to-LC3I ratio under the fed condition when treated phages are proinflammatory and are sufficient to induce
with chloroquine and rapamycin and under the fasted insulin resistance when transplanted to irradiated wild-
condition when treated with rapamycin, indicating acti- type mice upon high-fat-diet feeding (22). In addition,
vation of autophagy (Fig. 5E). AMPKa activity is decreased in adipose tissue of obese
mTOR pathway and adipocyte inflammation. The and insulin-resistant humans (42), suggesting that AMPKa
effects of SNRK on autophagy and ACC/raptor phosphor- is potentially another candidate for suppressing obesity-
ylation were further confirmed in adipocytes. In SNRK induced adipose tissue inflammation. However, further
diabetes.diabetesjournals.org DIABETES,VOL.62,JULY2013 2405