FEBS Letters 585 (2011) 2493–2500

journal homepage: www.FEBSLetters .org

Cyclin T2: A novel miR-15a target gene involved in early spermatogenesis

Yu Teng, Yong Wang, Jun Fu, Xiaowen Cheng, Shiying Miao, Linfang Wang ⇑State Key Laboratory of Medical Molecular Biology, Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences, Peking Union Medical College,Tsinghua University, 5 Dong Dan San Tiao, Beijing 100005, China

a r t i c l e i n f o

Article history:Received 26 April 2011Revised 18 June 2011Accepted 27 June 2011Available online 2 July 2011

Edited by Tamas Dalmay

Keywords:miR-15aCyclin T2Cell differentiationSpermatogenesisMicroRNA microarray

0014-5793/$36.00 � 2011 Federation of European Biodoi:10.1016/j.febslet.2011.06.031

Abbreviations: miRNA, microRNA; Ccnt2, cyclin Tdpp, day post-partum; DEM, differentially expremedium; DM, differentiation medium; RT-qPCR, revetative PCR; HRP, horseradish peroxidase; snRNA,positive transcription elongation factor b; MHC, myogeCTD, carboxyl-terminal domain⇑ Corresponding author. Fax: +86 10 6524 0529.

E-mail addresses: lfwangz@yahoo.com, wang.linfan

a b s t r a c t

MicroRNAs (miRNAs) are posttranscriptional modulators of gene expression that play importantroles in various biological processes. Spermatogenesis is a highly regulated process in which diploidspermatogonia eventually differentiate into haploid spermatozoa. In this study, we identified fourdifferentially expressed miRNAs between two premeiotic male germ cells, made predictions abouttheir putative targets, and confirmed cyclin T2 (Ccnt2) as a direct target of miR-15a. We also reportthat miR-15a inhibited muscle differentiation at least in part by targeting Ccnt2, which represents anovel interaction. Subsequently, miR-15a and Ccnt2 were profiled in developing mice testes toobserve their inverse correlations in the postnatal 3-week period to understand their roles inspermatogenesis.� 2011 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

1. Introduction

MicroRNAs (miRNAs) are a class of widely distributed andhighly conserved non-coding RNAs about 22 nucleotides (nt) long.They are thought to play negative regulatory roles posttranscrip-tionally, causing mRNA cleavage or translation repression, by tar-geting the imperfect complementary sequences, usually locatedin the 30-untranslated regions (UTRs) of mRNAs [1]. It has been re-ported that miRNAs are involved in various biological processes,including the cell cycle, differentiation and tumorigenesis [2–4].Additionally, the temporal and spatial specificity that is frequentlyassociated with development and differentiation is another charac-teristic of miRNA expression [5,6].

As a multilevel differentiation process, spermatogenesis func-tions to perform cyclic gamete production in which diploid sper-matogonia eventually differentiate into haploid spermatozoa.Spermatogenesis can be theoretically separated into three phases:the mitotic, meiotic and spermiogenesis phases. For mice, the firstwave of spermatogenesis takes about 35 days, during which the

chemical Societies. Published by E

2; UTR, untranslated region;ssed miRNA; GM, growthrse transcription and quanti-small nuclear RNA; P-TEFb,nin and myosin heavy chain;

g@imicams.ac.cn (L. Wang).

first generation of spermatogonia emerge at the 6th day post-partum (dpp), and the type B spermatogonia first appear at the8th dpp, whereas the pre-leptotene spermatocytes emerge aroundthe 10th dpp and the round spermatids at the 20th dpp [7].

In recent years, studies on miRNA regulation in spermatogene-sis have rapidly become one of the focal points in studies on malereproduction. For instance, miR-122a was thought to be involvedin the direct regulation of Tnp2, a testis-specific, crucial protein in-volved in chromatin remodelling in postmeiotic germ cells [8].miR-34c was recently suggested to enhance the late stages of sper-matogenesis by reinforcing the germ cell phenotype [9]. Dead end1 (Dnd1), an evolutionarily conserved RNA binding protein, wasshown to be essential for primordial germ cell development andspermatogenesis in zebrafish and mouse by protecting certainmRNAs from miRNA-mediated repression [10–13], and was newlyreported to regulate mitotic arrest and differentiation in male germcells [14]. Additionally, Dicer, a requisite enzyme for miRNA pro-cessing, was proved to be crucial for both the male germline andSertoli cells in spermatogenesis [15–18].

In this work, a miRNA microarray analysis was performed usingtwo lines of mouse germ cells, GC-1 spg and GC-2 spd, which weretype B spermatogonia and premeiotic spermatocytes, respectively,to determine the profiles of differentially expressed miRNAs(DEMs). During spermatogenesis, these cells are representative oftwo successive differentiation stages prior to the onset of meiosis,or rather as two transitional stages in the shift from mitosis to mei-osis. The profiling of these stages may allow for some insight into

lsevier B.V. All rights reserved.

2494 Y. Teng et al. / FEBS Letters 585 (2011) 2493–2500

the nascent period of meiosis. We identified cyclin T2 (Ccnt2) as anovel target of miR-15a and report that by partly targeting Ccnt2 atleast, miR-15a had a negative effect on muscle differentiation, aswell as in the early steps of spermatogenesis.

2. Materials and methods

2.1. miRNA microarray

Total RNA was extracted using TRIzol Reagent (Invitrogen)according to the manufacturer’s instructions. Total RNA (5 lg)was labelled with Hy3 using the miRCURY™ LNA microRNA Arraylabelling kit (Exiqon A/S), and the labelled samples were concen-trated using the RNeasy Mini Kit (Qiagen) as recommended. miR-NA expression profiles between GC-1 and GC-2 cells wereproduced by the hybridisation of the labelled RNA and the LNA-based capture probes on the microarray slides (miRCURY LNA™microRNA Array V.8.0, Exiqon A/S). The spike-in miRNA controls(miRCURY LNA™ Array Spike-in miRNA Kit, Exiqon A/S) were ap-plied as on-chip normalisers. Hybridisation and washing stepswere performed as instructed. Hybridisation signals were detectedusing the Genepix 4000B Array Scanner (Axon Instruments) at635 nm. Images were analysed in Genepix Pro 6.0 (Axon) andsaved as Excel files. Hybridisations were performed in duplicate,and the mean values of miRNA expression and the fold changes be-tween GC-1 and GC-2 cells were recorded.

2.2. Cell culture and differentiation

GC-1 spg and GC-2 spd cells were purchased from AmericanType Culture Collection (ATCC) and cultured in DMEM containing10% foetal bovine serum (FBS, Invitrogen). HEK 293T, NIH/3T3,C2C12, TM3, TM4, P19 and CT26.WT cells were obtained fromthe Cell Resource Centre, Institute of Basic Medical Sciences, Chi-nese Academy of Medical Sciences (CAMS). NIH/3T3 cells were cul-tured in DMEM with 10% bovine calf serum (BCS, Invitrogen); TM3and TM4 cells in a 1:1 mixture of Ham’s F12 medium (Invitrogen),DMEM with 1.2 g/l sodium bicarbonate, 15 mM HEPES supple-mented with 5% horse serum (HS, Invitrogen) and 2.5% FBS. P19cells were maintained in Alpha Minimum Essential Medium withribonucleosides, deoxyribonucleosides (Invitrogen), 7.5% BCS and2.5% FBS, and CT26.WT cells were cultured in RPMI-1640 mediumwith 10% FBS. Other cells were cultured in DMEM with 10% FBS. Toinduce myodifferentiation [19], growth medium (GM) maintaininglogarithmically-grown C2C12 cells was substituted with DMEMcontaining 2% HS (differentiation medium, DM) at a 70%confluence.

2.3. Reverse transcription and quantitative PCR (RT-qPCR)

miRNAs in 1 lg total RNA were reverse transcribed using spe-cific stem-loop RT primers and the ReverTra Ace-a-First strandcDNA synthesis kit (Toyobo) as instructed by the manufacturer.qPCR was performed following Chen’s methods [20] with PowerSYBR Green PCR Master Mix (Applied Biosystems), the RT product,a miRNA-specific forward primer and a universal reverse primerusing the StepOnePlus Real-Time PCR System (Applied Biosys-tems), according to the company’s instructions. The reactions wereincubated at 95 �C for 10 min, followed by 40 cycles at 95 �C for15 s and 60 �C for 45 s. Melt curves were analysed using defaultmethods. U6 small nuclear RNA (snRNA) was utilised as an internalnormaliser. As for Ccnt2, 1 lg total RNA was reverse transcribed asinstructed, using b-actin as an internal reference gene (all primersare listed in Supplementary materials). The method of the compar-ative CT (DDCT) quantitation was applied as previously described

[21]. Three independent experiments were completed, and eachPCR reaction was performed in triplicate.

2.4. Target prediction

Putative target genes of the DEMs were generated from threepopular on-line algorithms: miRanda, PicTar and TargetScan. Com-mon targets were chosen to undergo subsequent luciferase assays.

2.5. Constructs, transfection and Dual-Luciferase Reporter Assay

The 30-UTR sequences containing the putative targets of miRNAcandidates were cloned immediately downstream of the fireflyluciferase coding region in the pGL3-Promoter Vector (Promega),between the XbaI and FseI restriction sites and later betweenEcoRV and PstI sites added to the vector. Specific mutations corre-sponding to the potential target site were set up by PCR-drivenoverlap extension as Heckman described [22] and inserted intothe pGL3 vector (primers listed in the Supplementary materials).

Lipofectamin 2000 (Invitrogen) and FuGENE HD (Roche) wereused for GC-1 and C2C12 cell transfection, respectively, as in-structed. Synthetic miRNA mimic/mimic negative control (NC)and/or inhibitor/inhibitor NC (20 nM each, GenePharma), pGL3plasmid (0.4 lg) and the Renilla luciferase expression vector,pRL-TK (0.01 lg), were cotransfected into HEK 293T cells in 12-well plates using FuGENE HD for 24 h. Firefly and Renilla luciferaseactivities were detected using the Dual-Luciferase Reporter (DLR)Assay System (Promega) as instructed, with the latter serving asa normaliser. Each assay was performed independently in tripli-cate, at least two times.

2.6. Animals, tissue preparation and isolation of spermatogenic cells

Male BALB/c mice were obtained from the Laboratory AnimalCentre, CAMS. All animal experiments were approved by the Ani-mal Care and Use Committee at the Institute of Basic Medical Sci-ences, CAMS. Testes from 1- to 8-week-old male BALB/c mice werecollected and divided into two parts. One part was quickly frozenin liquid nitrogen and homogenised in TRIzol Reagent for RNAextraction, while the other half was homogenised in SDS lysis buf-fer for protein detection.

Developing mice mentioned above were collected for isolationof spermatogenic cells according to a procedure described previ-ously [23]. Briefly, decapsulated testes were incubated in 0.5 mg/ml collagenase (Sigma) at room temperature for 15 min with gen-tle oscillation to release interstitial cells. After passing through100 lm copper meshes, the interstitial cells were discarded. Sem-iniferous tubules were resuspended in 1 mg/ml collagenase atroom temperature for 20 min to remove myoid cells. The tubuleswere then incubated in 0.5 mg/ml hyaluronidase (Sigma) for15 min with oscillation and pipetting. The dispersed cells were cul-tured in F12/DMEM with 10% FCS at 32 �C for 24 h. Non-adherentspermatogenic cells were collected.

2.7. Western blotting

Total protein from mouse testes (from 1- to 8-week-old mice) orvarious cell types was loaded on a SDS–polyacrylamide gel andelectrically transferred to a PVDF membrane (Millipore). The mem-brane was blocked in 5% skim milk for 1 h at room temperature andincubated with the appropriately diluted primary antibodies at 4 �Covernight. After three washes with 0.1% TBS-T, the membrane wasincubated with horseradish peroxidase (HRP)-conjugated second-ary antibodies for 1 h at room temperature. The membrane waswashed another three times, and the image was visualised bychemiluminescence using the ECL Reagent (Engreen). Primary

Y. Teng et al. / FEBS Letters 585 (2011) 2493–2500 2495

antibodies included anti-cyclin T2a/b (Santa Cruz), anti-myogenin(Abcam) and anti-myosin heavy chain (Abcam). As references,anti-GAPDH (Santa Cruz) or anti-b-tubulin (Abcam) were applied.

2.8. Northern blotting

Total RNA (20 lg per lane) was fractionated on a 15% denatur-ing polyacrylamide gel containing 8 M urea, transferred to a Hy-bond-N+ nylon membrane (Amersham Biosciences) and fixed byUV cross-linking (GS Gene linker UV Chamber, Bio-Rad) and baking(80 �C, 15 min). Membranes were hybridised with a c-32P-labelledLNA-modified probe (Exiqon A/S), complementary to miR-15a.Hybridisations were carried out overnight at 37 �C in PerfectHyb™Hybridization Solution (Toyobo). The membranes were washedtwice for 5 min in 2 � Standard Saline Citrate (SSC), 0.1% SDS atroom temperature. Autoradiography was performed to visualisethe images. For reference, the membranes were stripped for rehy-bridisation with U6 snRNA with a probe sequence of 50-GGGCCATGCTAATCTTCTCTGT-30.

2.9. Statistical analysis

Data are presented as the means ± S.D. Student’s t-test wasadopted to analyse the differences between groups. P values<0.05 were considered statistically significant.

3. Results

3.1. Screening and verification of the DEMs

miRNA microarrays of GC-1 and GC-2 cells were performed(Fig. 1A). Following the output analysis, 28 miRNAs stood outaccording to the inclusion criteria, i.e., GC-1/GC-2-fold change >2or <0.5, among which 20 miRNAs were higher in GC-1 cells, whilethe other 8 were higher in GC-2 cells (Table 1).

We were primarily interested in which genes were upregulatedduring the transitional period from mitosis to meiosis. Given the re-ciprocal relationship between a miRNA and its target, the 20 miR-NAs that were highly expressed in GC-1 cells were the first batchto be verified. Four miRNAs—miR-15a, miR-125a-5p, miR184 andmiR-468—were selected randomly to determine whether they wereactually expressed at higher levels in GC-1 cells (Fig. 1B), whichwould be consistent with the outputs of the microarrays.

3.2. Ccnt2 is a direct target of miR-15a

Target prediction was subsequently performed with the widelyused algorithms: miRanda, PicTar and TargetScan. DLR assays wereperformed using the constructs containing the predicted sites toevaluate the authenticity of the putative targets (SupplementaryFig. 1), among which only the 30-UTR of Ccnt2, one of the putativetargets of miR-15a, caused a marked reduction in luciferase activ-ity (Fig. 2A).

Ccnt2, which is composed of two splice variants (T2a and T2b),is a member of the T-type cyclins, which have been identified onlyin the past decade [24] and are widely expressed among tissues[25]. Coupled with CDK9, Ccnt2 forms a complex known as ‘posi-tive transcription elongation factor b’ (P-TEFb), which is generallyconsidered to function in a cell cycle-independent manner [26]and is involved in muscle differentiation [19,27]. Eight mouse celllines were collected for the detection of Ccnt2 and miR-15a. It wasobserved that an inverse relationship existed between Ccnt2 andmiR-15a (Fig. 2B and C), which suggested a regulatory relationshipbetween them. Alignments of the potential target site indicatedthat the sequence was highly conserved across species (Fig. 2D).

We noticed perfect base pair matches in the seed region (2–8 nt,at the 50 end of the miRNA). Mutation assays were performed asindicated (Fig. 2E and F). In contrast to the negative control, theluciferase activity of the wild-type group decreased significantly(P < 0.01) in the presence of a miR-15a mimic, whereas the mut-agenised groups restored luciferase activity to varying degrees.Especially in mutant-2, which contained a mutation that coveredthe entire base-pairing region at the 50 end, the restored activitylevel was comparable to that observed in the control, suggestingthe importance of base pairing beyond the seed region. Conditionsin mutant-1 and mutant-3 revealed the intermediate states(P < 0.01), supporting the idea that the seed region of miRNAs isnot always dominant enough to target a gene. Conversely, theaddition of a miR-15a inhibitor restored luciferase activity to thewild-type, mutant-1 and mutant-3 groups (P < 0.01) but not inthe null and mutant-2 groups. Furthermore, transfection of amiR-15a mimic or inhibitor (50 nM each) into GC-1 cells resultedin a marked reduction or increase of Ccnt2 protein levels, respec-tively (Fig. 2G, Supplementary Fig. 2).

3.3. miR-15a inhibits muscle differentiation by targeting Ccnt2

Based on the involvement of the CDK9/Ccnt2 complex in muscledifferentiation [19,27], we hypothesised that miR-15a may have aneffect on this process. We observed this process in a differentia-tion-inducible myoblast, C2C12. A miR-15a mimic or siRNA fromCcnt2 (50 nM each) were transfected into C2C12 cells in advance.After a 48-h incubation, miR-15a was exclusively overexpressedin corresponding cells (Fig. 3A), while Ccnt2 mRNA was downreg-ulated in both siRNA and mimic treated cells compared with con-trols (Fig. 3B), suggesting that a cleavage of Ccnt2 mRNA wasinduced by miR-15a. Similar results were obtained at the proteinlevel (Fig. 3C). Myodifferentiation was induced after transfection,with early and late differentiation markers, myogenin and myosinheavy chain (MHC), respectively, being detected at the appropriatetimes. Following the progress of differentiation, the levels of myog-enin and MHC increased continuously, whereas that of Ccnt2 de-creased gradually. Compared to the negative control, miR-15a ledto a truncated build-up of myogenin and MHC (Fig. 3D and E). Mor-phologically, the formation of myotubes was delayed by miR-15a(Fig. 3F). Briefly, miR-15a negatively affected muscle differentia-tion by partly targeting Ccnt2 at least.

3.4. Ccnt2 is important for early spermatogenesis

Do Ccnt2 and miR-15a make a difference during spermatogen-esis? Ccnt2 was detected in developing mice testes at different lev-els. Ccnt2 mRNA rose continually in the postnatal 3-week periodand suddenly dropped off from the 4th week (Fig. 4A), suggestingits importance in early spermatogenesis. As for the Ccnt2 protein,the larger form (T2a in mice) increased markedly from the 3rdweek, while the smaller form (T2b in mice) decreased simulta-neously (Fig. 4B), implying that they affect different stages of sper-matogenesis. Parallel detections of miR-15a showed that itdropped continuously after birth (Fig. 4C and D), especially withinthe postnatal 3-week period, the inverse of what was observedwith Ccnt2. Quantitative detections of Ccnt2 mRNA and miR-15ain developing mice spermatogenic cells displayed similar resultswith that in testes (Fig. 4E and F). Based on these results, Ccnt2and miR-15a might be crucial in early spermatogenesis.

4. Discussion

In this study, we identified 4 DEMs between GC-1 and GC-2cells and demonstrated that Ccnt2 is a novel target of miR-15a.

Fig. 1. Screening and verification of the DEMs between GC-1 and GC-2 cells. (A) Screening for DEMs via miRNA microarrays. (B) Verification of miRNA expression by RT-qPCR.Fold changes of miRNA expression between the two cell types, with the expression values in GC-2 cells being set to 1.

Table 1Differentially expressed miRNAs between GC-1 and GC-2 cells, showing as the foldchanges of GC-1/GC-2 expression values.

MicroRNA Fold changea MicroRNA Fold changea

mmu-miR-138 4.613240095 mmu-miR-369-5p 0.150993389mmu-miR-424 4.401021296 mmu-miR-26b 0.430883272mmu-miR-17-3p 4.371026523 mmu-miR-128b 0.433985867mmu-miR-184 3.672312341 mmu-miR-369-3p 0.434825246mmu-miR-468 3.631771003 mmu-miR-144 0.44893017mmu-miR-15a 3.395488654 mmu-miR-302d 0.467744243mmu-miR-382 3.104128219 mmu-miR-33 0.476561636mmu-miR-200b 3.00372182 mmu-miR-297 0.497366955mmu-miR-130b 2.891435729mmu-miR-325 2.650311185mmu-miR-134 2.601290953mmu-miR-467a 2.524951893mmu-miR-195 2.477003861mmu-miR-381 2.39370409mmu-miR-195 2.320682056mmu-miR-370 2.320139068mmu-miR-125a-5p 2.301493625mmu-miR-466 2.234111036mmu-miR-490 2.227547943mmu-miR-15b 2.000839317

a Fold change >2 or <0.5 is considered as an inclusion criteria.

2496 Y. Teng et al. / FEBS Letters 585 (2011) 2493–2500

Notably, our results indicated that the base pairing at the 30 end ofmiR-15a was more essential for Ccnt2 recognition, which is in dis-agreement with the popular theory that the 50 seed region of miR-NAs acts as a key determinant in recognition [28,29]. In fact,miRNAs tend to be conserved over the full length, therefore the30 end of miRNAs probably make sense too. Generally, miRNA tar-get sites are classified into three categories [30,31], 50-dominantcanonical, 50-dominant seed and 30-compensatory sites. The canon-ical sites pair well at both 50 and 30 ends of the miRNA, the seedsites have perfect base pairing at the 50 end and require limitedor no 30 pairing support, while the 30-compensatory sites haveweak 50 base pairing and depend on strong compensatory pairingto the 30 end of the miRNA. It has been also reported that tran-scripts with 30-compensatory sites can discriminate between miR-NA family members [30]. None of the above types matched well tothe site we identified. An intermediate form between the 50-dom-inant seed and 30-compensatory sites might be more indicative ofwhat we observed.

As a member of the miR-15a/miR-16-1 cluster, miR-15a haslong been considered an important tumour suppressor associatedwith the inhibition of cell proliferation, the promotion of apoptosisand the suppression of tumorigenesis by specifically targeting

Fig. 2. miR-15a directly targets the 30-UTR of Ccnt2. (A) Normalised luciferase activity from the Ccnt2 30-UTR in the miR-15a mimic or in the negative control (NC)-transfectedHEK 293T cells. The ratio in the NC was set to 100%. (B, C) Detection of Ccnt2 protein and miR-15a among types of mouse cell lines. GAPDH and U6 snRNA were used asreferences, respectively. In (C), the expression value from GC-1 cells was set to 1. (D) Imperfect complementation between miR-15a and the potential target site in the Ccnt230-UTR, and the alignments of the target site with various species. The red letters represent the sites complementary to the seed region of miR-15a; the blue letters indicatethe base differences between species. The evolutionarily conserved nucleotides are listed at the bottom. (E) Mutations corresponding to the target sites. Different mutagenicsites are indicated in corresponding colours. (F) Normalised luciferase activities of wild-type (WT) construct and a series of mutants cotransfected with small RNAs in HEK293T cells. The empty pGL3-promoter (Null) served as the intergroup reference. The ratios of the NC of each group were set to 100%. Mi-NC, mimic negative control; In-NC,inhibitor negative control; 15a Mi, miR-15a mimic; 15a In, miR-15a inhibitor. ⁄⁄ and DD represent P < 0.01. (G) Detection of Ccnt2 in small RNA transfected GC-1 cells. GAPDHserved as the loading control. LO, Lipofectamin Only.

Y. Teng et al. / FEBS Letters 585 (2011) 2493–2500 2497

Fig. 3. Effect of miR-15a on muscle differentiation. (A–C) Detection of miR-15a and Ccnt2 in small RNA transfected C2C12 cells. In (A, B), the expression value in the FO groupwas set to 1. In (C), tubulin served as the loading control. FO, FuGENE HD Only; C-Si, Ccnt2 siRNA. (D–F) Differentiation of C2C12 cells in different time courses after 48 h oftransfection. In (D, E), levels of Ccnt2, myogenin and MHC were determined by Western blotting, and tubulin was detected as a loading control. In (F), images under the lightmicroscope were captured. The scale bars indicate 500 lm.

2498 Y. Teng et al. / FEBS Letters 585 (2011) 2493–2500

numerous carcinogenic genes since it was identified [32]. miR-15adownregulation has been associated with various cancers [32,33].Few studies have suggested that miR-15a is involved in cell differ-entiation, although it has been thought to be associated withdevelopment and differentiation [34,35]. Our findings demon-strated that miR-15a decreased continuously before meiosis andprobably contributed to the on-going process of spermatogenesis,which is similar to the effect observed with macrophage differen-tiation [36].

P-TEFb is thought to facilitate the transition from abortive toproductive elongation catalytically by phosphorylating the car-boxyl-terminal domain (CTD) of RNA polymerase II [37]. Geneticanalysis has implicated a broad requirement for the CDK9/cyclinT complex in Caenorhabditis elegans during early embryonic gene

expression [38]. As a regulatory subunit of P-TEFb, cyclin T, includ-ing T1, T2a and T2b, is also required for most genes transcribed byRNA polymerase II [39]. However, the genes affected by thesecyclins are different, and among them, the pathways importantfor early embryogenesis and development are specifically underthe control of Ccnt2 [25]. Additionally, as reported previously, cy-clin T functions in various types of cell differentiation [19,40]. Inthis work, we confirmed that Ccnt2 was essential for muscledifferentiation. The continuous reduction of Ccnt2 during myodif-ferentiation, described previously [41], suggested that Ccnt2 wasimportant for the initiation of cellular differentiation.

Until now, little was known about how Ccnt2 and miR-15afunctioned in spermatogenesis. Our work shows an inverse rela-tionship between them in the developing mice testes during the

Fig. 4. Expression of Ccnt2 and miR-15a in developing mice testes and spermatogenic cells. Ccnt2 in developing mice testes up to 8 weeks old was detected at both RNA (A)and protein (B) levels. miR-15a concentration in developing mice testes was determined by RT-qPCR (C) and Northern blotting (D). In (A, C), the expression value from 1-week-old testis was set to 1. In (B, D), GAPDH and U6 snRNA served as loading controls, respectively. The expression levels of Ccnt2 and miR-15a were quantified by densityanalysis, then were normalised to their loading controls, shown as the numbers between the two panels of blots. The normalised density value from 1-week-old testis was setto 1. Ccnt2 mRNA (E) and miR-15a (F) in developing mice spermatogenic cells up to 8 weeks old were detected by RT-qPCR. The expression value from 1-week-oldspermatogenic cells was set to 1.

Y. Teng et al. / FEBS Letters 585 (2011) 2493–2500 2499

postnatal 3-week period in which spermatogenic cells exist asspermatogonia and spermatocytes. Following the production of alarge amount of spermatids, the expression of Ccnt2 and miR-15ain the testis decreases to a relatively low level, suggesting thatCcnt2 and miR-15a are crucial in early spermatogenesis, includingmeiosis, but not in the postmeiotic steps. Consequently, based onits known function as well as our findings, we rationally speculatethat Ccnt2 promotes germ cell differentiation in the early orpremeiotic steps of spermatogenesis, whereas miR-15a negativelyregulates Ccnt2 by targeting it during this time.

Due to the lack of a spermatogenic differentiation model invitro, we cannot observe the effects of Ccnt2 and miR-15a on germcell differentiation directly. Because Ccnt2�/� mice have an earlyembryonic lethal phenotype [25], the generation of miR-15a genet-ically inactivated mice might help clarify their effects on this pro-cess. Additionally, in light of our results and its tumour suppressorcharacteristics, miR-15a is quite promising and may be applied inthe treatment of male sterility or testicular cancer in the future.

Acknowledgments

We are grateful to Prof. Daishu Han, Jia Yu, Dr. Fang Wang andour colleagues for their kind assistance. This work was supported

by grants from the National Important Research Plan of China(2011CB944302) from the Ministry of Science and Technology ofChina and the State Key Laboratory Special fund (2060204). Theexperimental study was performed in compliance with the lawsof the People’s Republic of China.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.febslet.2011.06.031.

References

[1] Lai, E.C. (2002) MicroRNAs are complementary to 30-UTR sequence motifs thatmediate negative post-transcriptional regulation. Nat. Genet. 30, 363–364.

[2] Carleton, M., Cleary, M.A. and Linsley, P.S. (2007) MicroRNAs and cell cycleregulation. Cell Cycle 6, 2127–2132.

[3] Sarkar, S., Dey, B.K. and Dutta, A. (2010) MiR-322/424 and -503 areinduced during muscle differentiation and promote cell cycle quiescenceand differentiation by down-regulation of Cdc25A. Mol. Biol. Cell 21, 2138–2149.

[4] Lee, Y.S. and Dutta, A. (2009) MicroRNAs in cancer. Annu. Rev. Pathol. 4, 199–227.

[5] Sempere, L.F., Sokol, N.S., Dubrovsky, E.B., Berger, E.M. and Ambros, V. (2003)Temporal regulation of microRNA expression in Drosophila melanogaster

2500 Y. Teng et al. / FEBS Letters 585 (2011) 2493–2500

mediated by hormonal signals and broad-Complex gene activity. Dev. Biol.259, 9–18.

[6] Wienholds, E. et al. (2005) MicroRNA expression in zebrafish embryonicdevelopment. Science 309, 310–311.

[7] Itman, C., Mendis, S., Barakat, B. and Loveland, K.L. (2006) All in the family:TGF-beta family action in testis development. Reproduction 132, 233–246.

[8] Yu, Z., Raabe, T. and Hecht, N.B. (2005) MicroRNA Mirn122a reduces expressionof the posttranscriptionally regulated germ cell transition protein 2 (Tnp2)messenger RNA (mRNA) by mRNA cleavage. Biol. Reprod. 73, 427–433.

[9] Bouhallier, F. et al. (2010) Role of miR-34c microRNA in the late steps ofspermatogenesis. RNA 16, 720–731.

[10] Koprunner, M., Thisse, C., Thisse, B. and Raz, E. (2001) A zebrafish nanos-related gene is essential for the development of primordial germ cells. GenesDev. 15, 2877–2885.

[11] Weidinger, G. et al. (2003) Dead end, a novel vertebrate germ plasmcomponent, is required for zebrafish primordial germ cell migration andsurvival. Curr. Biol. 13, 1429–1434.

[12] Youngren, K.K. et al. (2005) The Ter mutation in the dead end gene causesgerm cell loss and testicular germ cell tumours. Nature 435, 360–364.

[13] Kedde, M. et al. (2007) RNA-binding protein Dnd1 inhibits microRNA access totarget mRNA. Cell 131, 1273–1286.

[14] Cook, M.S., Munger, S.C., Nadeau, J.H. and Capel, B. (2011) Regulation of malegerm cell cycle arrest and differentiation by DND1 is modulated by geneticbackground. Development 138, 23–32.

[15] Maatouk, D.M., Loveland, K.L., McManus, M.T., Moore, K. and Harfe, B.D. (2008)Dicer1 is required for differentiation of the mouse male germline. Biol. Reprod.79, 696–703.

[16] Hayashi, K. et al. (2008) MicroRNA biogenesis is required for mouse primordialgerm cell development and spermatogenesis. PLoS ONE 3, e1738.

[17] Papaioannou, M.D. et al. (2009) Sertoli cell Dicer is essential forspermatogenesis in mice. Dev. Biol. 326, 250–259.

[18] Kim, G.J., Georg, I., Scherthan, H., Merkenschlager, M., Guillou, F., Scherer, G.and Barrionuevo, F. (2010) Dicer is required for Sertoli cell function andsurvival. Int. J. Dev. Biol. 54, 867–875.

[19] Simone, C. et al. (2002) Activation of MyoD-dependent transcription by cdk9/cyclin T2. Oncogene 21, 4137–4148.

[20] Chen, C. et al. (2005) Real-time quantification of microRNAs by stem-loop RT-PCR. Nucleic Acids Res. 33, e179.

[21] Livak, K.J. and Schmittgen, T.D. (2001) Analysis of relative gene expressiondata using real-time quantitative PCR and the 2(�Delta Delta C(T)) Method.Methods 25, 402–408.

[22] Heckman, K.L. and Pease, L.R. (2007) Gene splicing and mutagenesis by PCR-driven overlap extension. Nat. Protoc. 2, 924–932.

[23] Wang, H., Xiong, W., Chen, Y., Ma, Q., Ma, J., Ge, Y. and Han, D. (2006)Evaluation on the phagocytosis of apoptotic spermatogenic cells by Sertolicells in vitro through detecting lipid droplet formation by Oil Red O staining.Reproduction 132, 485–492.

[24] Peng, J., Zhu, Y., Milton, J.T. and Price, D.H. (1998) Identification of multiplecyclin subunits of human P-TEFb. Genes Dev. 12, 755–762.

[25] Kohoutek, J., Li, Q., Blazek, D., Luo, Z., Jiang, H. and Peterlin, B.M. (2009)Cyclin T2 is essential for mouse embryogenesis. Mol. Cell. Biol. 29, 3280–3285.

[26] Wang, S. and Fischer, P.M. (2008) Cyclin-dependent kinase 9: a keytranscriptional regulator and potential drug target in oncology, virology andcardiology. Trends Pharmacol. Sci. 29, 302–313.

[27] Giacinti, C., Bagella, L., Puri, P.L., Giordano, A. and Simone, C. (2006) MyoDrecruits the cdk9/cyclin T2 complex on myogenic-genes regulatory regions. J.Cell. Physiol. 206, 807–813.

[28] Lim, L.P. et al. (2005) Microarray analysis shows that some microRNAsdownregulate large numbers of target mRNAs. Nature 433, 769–773.

[29] Lambert, N.J., Gu, S.G. and Zahler, A.M. (2011) The conformation of microRNAseed regions in native microRNPs is prearranged for presentation to mRNAtargets. Nucleic Acids Res. 39, 4827–4835.

[30] Brennecke, J., Stark, A., Russell, R.B. and Cohen, S.M. (2005) Principles ofmicroRNA-target recognition. PLoS Biol. 3, e85.

[31] Sethupathy, P., Megraw, M. and Hatzigeorgiou, A.G. (2006) A guide throughpresent computational approaches for the identification of mammalianmicroRNA targets. Nat. Methods 3, 881–886.

[32] Calin, G.A. et al. (2002) Frequent deletions and down-regulation of micro-RNAgenes miR15 and miR16 at 13q14 in chronic lymphocytic leukemia. Proc. Natl.Acad. Sci. USA 99, 15524–15529.

[33] Bonci, D. et al. (2008) The miR-15a–miR-16-1 cluster controls prostatecancer by targeting multiple oncogenic activities. Nat. Med. 14, 1271–1277.

[34] Friedman, L.M. et al. (2009) MicroRNAs are essential for development andfunction of inner ear hair cells in vertebrates. Proc. Natl. Acad. Sci. USA 106,7915–7920.

[35] Garzon, R. et al. (2007) MicroRNA gene expression during retinoic acid-induced differentiation of human acute promyelocytic leukemia. Oncogene26, 4148–4157.

[36] Li, T., Morgan, M.J., Choksi, S., Zhang, Y., Kim, Y.S. and Liu, Z.G. (2010)MicroRNAs modulate the noncanonical transcription factor NF-kappaBpathway by regulating expression of the kinase IKKalpha duringmacrophage differentiation. Nat. Immunol. 11, 799–805.

[37] Marshall, N.F., Peng, J., Xie, Z. and Price, D.H. (1996) Control of RNApolymerase II elongation potential by a novel carboxyl-terminal domainkinase. J. Biol. Chem. 271, 27176–27183.

[38] Shim, E.Y., Walker, A.K., Shi, Y. and Blackwell, T.K. (2002) CDK-9/cyclin T (P-TEFb) is required in two postinitiation pathways for transcription in the C.elegans embryo. Genes Dev. 16, 2135–2146.

[39] Chao, S.H. and Price, D.H. (2001) Flavopiridol inactivates P-TEFb and blocksmost RNA polymerase II transcription in vivo. J. Biol. Chem. 276, 31793–31799.

[40] Leucci, E. et al. (2007) The role of the Cdk9/Cyclin T1 complex in T celldifferentiation. J. Cell. Physiol. 212, 411–415.

[41] Cottone, G., Baldi, A., Palescandolo, E., Manente, L., Penta, R., Paggi, M.G. and DeLuca, A. (2006) Pkn is a novel partner of cyclin T2a in muscle differentiation. J.Cell. Physiol. 207, 232–237.