發情周期對香豬卵巢生物鐘相關基因表達的影響（英文）

關鍵詞：晝夜節律；香豬；卵巢；發情周期；可變剪接

doi：10.13304/j.nykjdb.2023.0132

中圖分類號：S828 文獻標志碼：A 文章編號：10080864（2023）12006718

Circadian rhythm is the regular change of manyfunctional activities of the biological organism to adaptto the alternation of day and night in the environment. Itis endogenous and persists even without externalsynchronising factors， for example， in constant darkness.Circadian rhythm is controlled by central biologicalclocks related to hypothalamic suprachiasmatic nucleus（SCN） and peripheral biological clocks distributed invarious tissues[1]. SCN cells modulate the activity ofgonadotropin-releasing hormone （GnRH） neurons bydirect innervation or other hypothalamic nucleus， whichare necessary for the occurrence of ovulation[23].It isknown that the molecular model of mammals circadianclock including human is a multi-feedback loop fortranscription and translation including transcriptionalregulation and post-transcriptional regulation[4]. Itconsists of positive regulatory components （BMAL1 andCLOCK） and negative regulatory components （PER 1/2and CRY 1/2）[5]. The positive and negative regulatoryelements also participate in the circadian regulation ofclock controlled genes （CCG）. The clock-controlledgenes act as specific regulators of rhythmic physiologicalfunctions in cells and tissues[6]. For example， thefunction of ovary is regulated by the biological clock，including the estrus cycle， luteinizing hormone （LH）levels， ovulation and embryo implantation[78].

Ovary is an important reproductive organ offemale mammal. In each estrus cycle， the ovariesundergo a series of morphological， hormonal andbiochemical changes， which directly affect and/ordetermine the litter size of female animals[9]. Therealization of ovarian function is a very complexbiological process， involving the transcriptionalregulation of a large number of genes[10]. Alternativesplicing （AS） is also an important way of geneexpression regulation， which occurs widely invarious tissues and cells of almost all organisms[1112].Many studies demonstrate that the biological clockexisted in the ovary[13-16]. The expression of clockgenes have been described in the ovary of rat[17]，mice[18] and ruminant[19]. It indicates that clock genesare involved in the process of follicle formation，ovulation， oocyte maturation and steroid hormonesynthesis[20]. The lack of circadian clock might berelated with varieties of reproductive disorders. Forinstance， mice with per1 and per2 knockout present asignificant decrease in reproductive rate due to theirregular and non-cyclical estrus cycle[2122].Althoughmice with bmal1 knockout are capable of ovulation，they exhibit delayed puberty， irregular estrus cycles，and small ovaries and uterus[23-25]. Loss of bmal1 alsonegatively affects progesterone level[25]， leading toembryo implantation failure[23]. Mice mutant with 51amino acid deletions in the transcription activationdomain of clock protein show an irregular estruscycle， and the normal LH surges disappear on the dayof estrus[26]. In addition， genes of ovarian granulosacells that are important for follicular developmenthave been shown to be controlled by clock genes，such as LH receptor （lhcgr）， prostaglandin synthase（ptgs2）， steroid-producing enzymes （steroid-producingacute regulatory protein， cyp11a1， aromatase， etc.），and gap junction protein Connexin-43[27-29]. Estrus is asign of reproductive maturity. Moreover， it has beenconfirmed that the central biological clock located inthe SCN was the core of maintaining normalreproductive function and the estrus cycle[30].However， the function of ovarian biological clockduring estrus period is still needed to be clear.

The Xiang pig is a small local breed in China withsome unique biological characteristics， such as smallsize， precocious puberty， low birth rate and insignificantestrous behavior[31]. In order to understand more in-depthviews of the influence of estrus on biological clock andfurther explore the physiological functions of clockgenes in porcine ovary， this paper analyzed theexpression of biological clock-related genes andalternative splicing of transcripts in the ovaries of Xiangpig during estrus and diestrus period based on the wholetranscriptome sequencing analysis.

1 Material and method

1.1 Materials

About 40 sows of Xiang pig were selected for"observation of estrus performance. They were rearedin a standard manner that the pigs drank freely andfed timely under the normal condition. The averagelight intensity was about 200 lx， the ambienttemperature was maintained at 15～18 ℃ ， and thehumidity was in the range of 50%±10%. The estrusdetection was carried out 2～3 times per day.Redness and swelling of the vulva and a positivestanding reflex test of the boar were defined as estrusmanifestations. Ovary samples of 14 Xiang pig sowswere taken， 7 of them were in the diestrus period，and the others were subjected to routine castrationsurgery on the 3th day of estrus （the sows showedstrong redness and swelling of the vulva）. Thesampling day was at 8：00—9：00 am at March 15 to30， 2022， with the same light and temperatureconditions as previously mentioned， and fastingbefore ovariectomy surgery. All samples wereimmediately immersed in liquid nitrogen and storedat -80 ℃ for RNA extraction. The animal procedureswere approved by the guidelines of Guizhou UniversitySubcommittee of Experimental Animal Ethics withNo. of EAE-GZU-2020-P002.

1.2 Total RNA extraction

Total RNA from 14 ovarian samples wasextracted using TRIzol reagent kit （Invitrogen， USA）according to the instructions. The quantity andcompleteness of total RNA were evaluated by theBio-analyzer system （Agilent 2100， AgilentTechnologies， USA）. The total RNA was stored at-80 ℃ for subsequent analysis. The same sampleswere used for sequencing and PCR analysis.

1.3 Library construction and sequencing

Total of 14 cDNA libraries were constructedusing RNA as templates from ovarian tissues. Thelibrary construction was carried out according to thestandard procedures of BGI （Shenzhen Huada GeneCo.， Ltd.）. After DNAse Ⅰ enzyme treatment，magnetic beads with oligo （dT） were used to isolatemRNA. Under appropriate temperature conditions，the purified mRNA was fragmented by addingfragmentation buffer， and then cDNA was synthesizedusing random primers and mRNA fragment templates.After 3’-end adenylation， the DNA fragments wereligated to the linker adaptors. cDNA fragments of100～200 bp in length were selected for PCRamplification to generate a cDNA library. Thelibraries were sequenced on the Illumina HiSeqTM2500 sequencing platform， and paired-end sequenceswith 150 bp in length were generated.

1.4 Sequencing data analysis

The original sequences generated by theIllumina platform were saved in fastq format. Theclean reads were obtained by removing low-qualitysequences， reads （more than 5% unknownnucleotides）， and sequencing adapters. The cleanreads should be evaluated through FastaQC software（Q20lt;20%）. The reference genome sequence （ssc11.1version） and annotation files of pig genomes weredownloaded from the Ensembl database （http：//www.ensembl. org/index. html）. Used HISAT2 software（v2.1.0） to map the clean reads to the pig referencegenome. HISAT2 is an efficient alignment tool forRNA-seq experiments， which uses an indexingframework based on the BWT and Ferragina-Manzini（FM） index algorithms for alignment and datanormalization. The Subread featureCounts software（v2.0.0） was used to count the number of reads mappedto the gene. Subread featureCounts applies the readsummarization tool to analyze the expression level fromRNA-seq， which applies efficient chromosome hashalgorithm and feature block technology for statisticsand data standardization.

1.5 Analysis of differential expression

The expressed level of each gene wascalculated by CPM values （counts per millionmapped reads） in all samples， and added 0.001 toCPM value as the gene expression level. Then weused bioconductor software package Limma andDESeq2 to analyze differentially expressed genes（DEGs） between two groups with biologicalreplicates. The FDR≤0.1 and |log2FC|≥0.5 were usedas the thresholds for determining the significance ofgene expression difference.

1.6 Identification of differential alternativesplicing events

The rMATS software （v4.0.2） was used toanalyze differentially alternative splicing eventsbetween two groups. The rMATS recognizes fivealternative splicing events， which are skipped exon（SE）， alternative 5’ splice site （A5SS）， alternative 3’splice site （A3SS）， mutually exclusive exons （MXE）and retained intron （RI） [32]. Also， FDR≤0.05 wasused as a threshold to detect the differentialsignificance of alternative splicing.

1.7 Validation

The same RNA samples from the ovarian forRNA-seq experiment were used to validate thedifferential expression by qRT-PCR and the ASevents by RT-PCR method. The primers weredesigned using Primer 6.0， which were listed inTable 1. GAPDH and β-actin genes were taken asinternal controls. The PCR efficiency of the primerswas controlled within 100%±10%. All qPCRreactions were performed on the CFX98TM real-timesystem （Bio-rad Co.） according to the manufacturer’sinstructions and recommended cycling conditions. 3replicate reaction tests were performed using avolume of 10 μL containing 5 μL of 2× Super RealPre Mix Plus， 0.3 μmol of each of the forward andreverse primers （10 pmol·μL-1） and 1 μL cDNA.The qPCR conditions were as following： initialdenaturation at 95 ℃ for 3 min， then denaturation at94 ℃ for 10 s in 40 cycles， and annealing andextension at 60 ℃ for 30 s. The relative expressionlevel of target gene utilized the method of 2-ΔΔCt asreported by Livak et al.[33] The different level of geneexpression between two groups was tested bysoftware SPSS （v21.0） taking the Plt;0.05 asthreshold of significant difference. The results werepresented as mean±standard deviation. The 8 ASevents were randomly selected from 8 genes tovalidate by RT-PCR method. The RT-PCR annealingtemperature was 55 ℃ for 30 s， and the extensiontemperature was 72 ℃ for 60 s by 34 cycles.

To verify the rhythmic expression of circadianrhythm-related genes in Xiang pig ovary， 6 genes，per1， cry1， clock， arntl， ppp1cb and ntrk1， wereselected to detected quantitatively by RT-qPCRmethod at 6 time points （08：00， 12：00， 16：00， 20：00，24：00， 04：00）. 3 ovary samples of Xiang pig werecollected at each time point， and RNA was extractedfrom the ovarian samples for qPCR verification ofgenes. The amplification conditions were the sameas the previous RT-qPCR method.

2 Result and analysis

2.1 Summary of sequencing data

The 14 cDNA libraries from ovaries in estrusand diestrus stages were constructed and sequencedrespectively， and the sequencing data were filteredand normalized， which were then mapped to thereference genome with HISAT2. It generated75 002 160～77 100 634 read pairs， and 93.57%～96.26% of reads were aligned to the referencegenome （Table 2）. The only matching reads wereperformed the following statistical analysis.

2.2 Expression of biological clock-related genesin Xiang pig ovary

After mapping and annotating based on the pigreference genome assembly Sscrofa11.1， the RNAseqdata of circadian rhythms and clock genes werepicked up from the genome-scale data sets （Table 3）.Taking threshold value of 0.1 CPM， 90 rhythmsgenes were determined， in which 81 were expressed inthe estrus ovaries and 83 were detected from thediestrus ovaries. The expression levels of those geneswere ranged from 867.2 CPM in the highest abundanttranscripts to the cut off value of 0.1 CPM. Comparedwith the diestrus ovary， there were more genes （55.6%in estrus， 42.2% in diestrus） at the medium expressionlevel （10～100 CPM）， and fewer genes at the high（gt;100 CPM） or low （0.1～10 CPM） expression level inthe estrus ovary （Fig. 1）. The top expressed genes werenono and sfpq with 455.6 and 446.1 CPM in the estrusovary， and sfpq and top2a with 532.6 and 512.1 CPMin the diestrus ovary， respectively.

2.3 Analysis of differential expression genes

R package Limma and DESeq2 were used toanalyze differentially expressed genes （DEGs）between estrus and diestrus ovaries. Also， the Pvalue≤0.01 and absolute value of log2FC≥0.5 wereused as the threshold for determining the significantdifference of gene expression. The results showedthat 33 DEGs in the ovaries between estrus anddiestrus cycles， comprising 13 down-regulated genes（top2a， nrip1， prkdc， setx， ncoa2， pparg， pspc1， per3，sirt1， timeless， ezh2， ppara， and ppp1ca） and 8 upregulatedgenes （atf4， csnk1d， klf10， id2， bhlhe40，id4， ass1， and ntrk1） in at least 2-fold at estrus stage.The fold changes for DEGs were ranged from 0.17 to26. Furthermore， it was found that 5 genes （ppp1cb，id3， per1， klf9， and nampt） were at least 1.5～2.0 foldupregulated and 7 genes （ep300， ube3a， dyrk1a，sin3a， clock， creb1， and magel2） were at least 1.5～2.0 fold down regulated in estrus ovary （Table 4）. Alot of the genes were expressed in ovaries at mediumor high levels （48.9% and 27.4%）， including the coreclock components， arntl， clock， per1/2/3， and cry1/2.

2.4 Detection of differential splicing genes

The rMATS was used to analyze differentiallyalternative splicing genes （DSGs） between estrus anddiestrus ovaries. Overall， 44 differential splicingevents were identified from the transcripts of 34circadian rhythms genes （Plt;0.05） including 3 A5SS，12 RI， 14 SE and 15 MXE events （Table 5）. Therewere multiple genes with complex alternativesplicing patterns. Of those， nono presented the mostabundant differential alternative splicing events，which included four differential splicing types：A5SS， RI， SE and MXE. Also， ppp1cb harbored threedifferential splicing types （RI， SE and MXE）. Then，ENSSSCG00000000107， per1， clock， ube3a andfbxw11 each contained 2 differential alternativesplicing types. It was found that 20 genes comprisedthe core clock components， arntl and cry1 weredifferentially regulated only at AS level. At least 14genes， such as per1 and clock， were differentiallyregulated in both expression and AS level in ovariesbetween estrus and diestrus.

2.5 Validation of DEGs by RT-qPCR method

To validate the results from RNA-seq， RT-qPCRwere used method to further examine thedifferentially expressed genes using the same RNAsamples from the ovaries （Fig. 2）. It confirmed thatppp1cb， csnk1d， per1， klf9， nampt， bhlhe40， id4， andass1 were significantly high expressed in estrussamples， which were in accordance with the RNAseqanalysis. The expressions of atf4 and ntrk1 inestrus and diestrus ovaries were not significantlydifferent. Moreover， 8 AS events， selected randomlyfrom 8 genes， were detected by RT-PCR andconsistent with the RNA-Seq results （Fig. 3）.

2.6 Validation of rhythmic expression of clockrelatedgenes

To verify whether the expression of clock-related genes in Xiang pig ovaries was rhythmic， wefurther examined the expression patterns of per1，cry1， clock， arntl， ppp1cb and ntrk1 at 6 time points（08：00， 12：00， 16：00， 20：00， 24：00， 04：00） using RTqPCR.The samples at each time point were obtainedfrom 3 Xiang pig ovaries collected within 1 h.Among them， the expressions of per1 and ppp1cbpeaked at 24：00 midnight. However， the expressionpatterns of clock and arntl were opposite， in whichthe peaks presented at 8： 00 am and it fell to thelowest at 24：00. The maximum values were appearedat 20：00 and 24：00 pm for gene cry1， and it was at16：00 pm for gene ntrk1. It illustrated that the coreclock genes per1， cry1， clock and arntl and the clockrelatedgenes ppp1cb and ntrk1 were rhythmicexpressed in Xiang pig ovaries （Fig.4）.

3 Discussion

The mammalian biological clock system has amulti-level structure， including the main clocklocated in the suprachiasmatic nucleus and the subclocksof peripheral organs and tissues[1]. The ovarywas a pivotal reproductive organ in female animal，and the expression of the biological clock gene hadbeen found in the ovaries of many organisms[20]. Inthis study， we analyzed the expression profile andalternative splicing of the biological clock gene inXiang pig ovaries in estrous cycle from the RNA-seqdata. The study found out a total of 90 rhythms genesthat were expressed in ovaries. The expressionabundance of these genes was ranging from 0.1 to532.6 CPM. A lot of the genes were expressed inovaries at medium or high levels （48.9% and27.4%）， including the core clock components， arntl，clock， per1/2/3， and cry1/2. Furthermore， 33 geneswere detected to undergo differentially expressionand 34 genes were detected to undergo differentialalternative splicing between estrous and diestrousovaries. The DEGs and DSGs related with rhythmsmight have a connection with the regulation of estrusprocess in Xiang pig.

Transcription and translation of core clockcomponents genes （clock or npas2， arntl/bmal1， orarntl2/bmal2， per1/2/3， and cry1/2） play a criticalrole in rhythm generation process. CLOCK andBMAL1 heterodimerize to activate transcription ofcircadian target genes including per1/2/3 and cry1/2.PER and CRY interact and conversely inhibittranscription of bmal1 and clock genes. These genesand their protein products were organized intointerlocking positive and negative transcriptionaland translational feedback loops， which regulatedcircadian rhythm generation in the brainsuprachiasmatic nucleus （SCN） and peripheralorgans[34]. In this study， we found that 3 core clockgenes were differentially expressed between estrusand diestrus ovaries. clock and per3 were downregulatedand per1 was up-regulated in the estrousovaries. CLOCK played a key role in maintaining thecircadian rhythm and activating downstreamelements. Inhibition of CLOCK could inhibit cellgrowth and increase the rate of apoptosis[35]. In vivoexperiments showed that female mice injected withCLOCK-shRNA had fewer oocytes， fewer litters， anda higher rate of apoptosis. The results indicated thatclock played an important role in fertility， and downregulationof clock leads to cell apoptosis anddecreased reproductive capacity[35]. It was proposedthat the down-regulation of clock in the ovary atestrus stage might be related to the low litter size ofthe Xiang pig. The core clock gene period 1 （per1）might be a prolific gene in Drosophila[19]. Femalemice with the per1 mutation showed a normalnumber of implantation sites but reduced littersize[22]. per1 mRNA located in the secondary oocytesand follicles of ruminants （sheep， cattle） and foundthat there was no relationship between itstranscription level and prolificacy， and this gene didnot map to the known QTL region of ovulation rate incattle[19]. Treatment with Progesterone for 1 h couldinduce the expression of per1 mRNA in MCF-7cells[36]. In fact， the up-regulation of per1 was acommon feature of many tissues in response tocertain types of hormonal stress， for example"luteinising hormone[37-39]. This expression patternindicated that PER1 might be the most sensitiveeffector in the biological clock system[40]. Somestudies have found that estrus leaded to changes inthe expression time and amplitude of the biologicalclock genes[41]. Rising estrogen caused female animalto go into estrous stage[42]. The estrus in Xiang pigmight be speculated to reset the biological clock byup-regulating the per1 gene to response the stimulusof steroid hormone. In this study， per3 was downregulatedin the ovaries of estrus pigs. There waslittle information on the function of per3 gene inreproduction except for brain development[43]. Incontrast， the expression patterns of clock genes inother pig breeds were much different from that inXiang pig. In the ovarian follicles of Large Whitepigs and Mi pig， the core circadian clock genes wereall down-regulated and not differentially expressedbetween estrus and diestrus periods[44]. Comparedwith Large White pig and Mi pig， the characteristicsof Xiang pig were specific， such as low birth rate andinsignificant estrous behavior. It would beinteresting to prove whether the expression profilesof core biological clock genes in ovary was a reasonfor the special reproductive traits such as estrusperformance and litter size in different pig breeds.

The biological clock system consists of an inputpathway， a core oscillator and an output pathway.The post-translational modification of clock proteinswas essential to maintain the accuracy androbustness of the evolutionarily conserved circadianclock[45]. Post-translational modification anddegradation of the clock proteins were key steps indetermining the length of the circadian clockcycle[46]. Our research found that the products ofmost differentially expressed genes were related tothe post-translational modification of the core clockprotein. For example， PPP1CB and PPP1CA couldreduce the phosphorylation of PER2， and affect thenuclear localization of the protein PER2， whichmight at least partially change the cycle and phaseshift characteristics of the biological clock[47]. Also，highlighted DYRK1A was the enzyme responsiblefor the phosphorylation of CRY2 at Ser557 andplayed a key role in regulating the protein level ofCRY2[48]. The casein kinases CSNK1D and CSNK1Ephosphorylated the PER protein and provided amarker for subsequent degradation[46]. The UBE3Abinded and degraded BMAL1 in a ubiquitin ligasedependentmanner[45]. Furthermore， MAGEL2regulated the ubiquitination and stability of CRY1，and changed its nuclear and cytoplasmicdistribution[49]. Lastly， SIRT1 deacetylated PER2 andBMAL1， thereby participated in biological rhythmregulation[50]. The previous reports illustrated that theestrus cycle affected the localization and degradationof the core clock protein by changing their posttranslationalmodification. And transcription andpost-transcriptional regulation were the basis ofclock system component activities， and posttranscriptionalmechanisms accounts for more than halfof the regulatory network[51]. The results of this studyfurther strengthened the view that the post-translationalmechanism participated in the circadian generegulatory network. In addition， we detected somedifferentially expressed genes from Xiang pig ovariesthat might affect the expression of core clock genes atthe transcriptional level. For example， TOP2A bindedto the unique GC-rich open chromatin structure of thebmal1 promoter region， indicating that TOP2 on thebmal1 promoter affected transcription of bmal1[52]. Also，CBP/P300 and tissue-specific cofactors regulatedCLOCK /BMAL1 transcription positively ornegatively[53]. Accordingly， ID （DNA binding inhibitor）was an important transcription repressor. Each IDprotein contained a helix-loop-helix domain throughwhich it could interact with bHLH protein. However，ID lacked the basic domain that allows binding to theE-box element， which leaded to the ID being able tomodify the transactivation of clock genes and CCG byinterfering with the CLOCK-BMAL1 heterodimer，bHLH orange factor and other bHLH factors[54]. Thesedata indicated that the estrus cycle promoted the timechange of the biological clock in reproductive tissues at"the level of transcription and post-translational modifications （Fig.5）.

In addition， we detected several genescontrolled by the circadian rhythm， which weredifferentially expressed in the ovaries of the twoperiods including klf9，star，ptgs2. Other studies hadshown that klf9 was a clock output gene， andCLOCK and BMAL1 complexes could bind andactivate transcription at the 5’ flanking region ofklf9， which was blocked by the co-expression ofPER1. Also， KLF9 might play a role in regulatingthe effects of CLOCK/BMAL1 and in the expressionof DBP， other clock and clock output genes， therebychanged the timing and amplitude of the circadianoscillations of gene transcription[55]. The previousstudies had found that steroidogenic enzymes（STAR）， prostaglandin synthase （PTGS2）， whichwere related to ovarian progesterone synthesis， wereclock-controlled genes （CCGs）. The E-box enhancerexisted in the 5’-flanking region of star， whichbinded to the CLOCK-BMAL1 heterodimer andactivated the transcription of gene star[56]. Coexpressionof the negative regulators PER and CRYattenuated this activation[57]. The promoter of ptgs2had an E-box element and a REV-ERBα/RORαresponse element （RORE）. The secretion of PGF2αcould be balanced by the inhibition or stimulation oftranscriptional regulation on REV-ERBα andBMAL1/CLOCK， respectively[58]. It was thus inferredthat the peripheral circadian oscillator， such asovary， could play an essential role in synchronizinglocal physiology through regulation of the expressionof clock-controlled genes[59]. In this study， star andptgs2 in the ovaries of the estrus pigs were upregulatedby 6.42 and 5.55 times， respectively. Itmight be possible that the core clock genes mightregulate the production of steroid hormones bycontrolling the ovarian-specific CCG， therebyaffecting the estrus cycle of Xiang pig.

Pre-mRNA splicing is a basic biologicalprocess through which introns of nascent RNA areremoved and exons are merged to form mature RNA，which is then translated into protein[60]. Through AS，different proteins are produced， resulting in a widervariety of cellular functions[61]. Except fortranscription or post-translational controls，alternative splicing also plays a key role in circadianrhythms in a cell-type-specific manner[62]. Recentresearch reported several cases， in which AS wasinvolved in the regulation of biological clocks inplants， mice， and fruit flies[63-67]. In this study， arntlonly underwent differential alternative splicing，while clock and per1 experienced both of differentialalternative splicing and differential expression. Thedifferential alternative splicing event occurred inper1 was the retention of introns 14～15. This eventwould result in the deletion of amino acid residuesfrom positions 545 to 1 278 in the encoded protein，and the deletion site was located in the core domainto activate the protein. This intron retention event"occurred more frequently in the ovaries of Xiang pig"in the diestrus period， and the active PER1 proteindecreased. During the estrus period， the occurrenceof the intron retention event decreased， and theactive PER1 protein level increased， which furtherincreased the expression of per1. It indicated that thecore clock component PER1 might respond to thechanges of sterol hormones in the ovary throughtranscriptional regulation and post-transcriptionalregulation. clock underwent the differential exon 17skipping event. When this event occurred， theencoded protein would be deleted amino acidresidues from positions 484 to 513. The deletedamino acids were not located in the core domain andthe influence on the protein activity was stillunclear. These findings suggested that AS mightdirectly change the structure of core clockcomponents to regulate the ovarian biological clock，and the AS regulation of biological clock genesmight be independent in the gene expression.However， more evidence should be accumulatedwhether the AS events in these biological clockgenes would turn on and turn off the circadianrhythm in pig.

（責任編輯：張冬玲）