Introduction
Cumulus–oocyte complexes (COCs) are dependent on adequate gene expression to initiate and to undergo oocyte maturation (meiotic progression) and embryonic development (Labrecque et al., Reference Labrecque, Vigneault, Blondin and Sirard2013; Li et al., Reference Li, Ray and Ye2013). The mechanisms by which oocytes acquire competence to develop up to the blastocyst stage are still not fully understood. There is evidence that the acquisition of competence is correlated with RNA and protein molecules processed and stored during growth and maturation periods (Ferreira et al., Reference Ferreira, Vireque, Adona, Meirelles, Ferriani and Navarro2009; Caixeta et al., Reference Caixeta, Sutton-McDowall, Gilchrist, Thompson, Price, Machado, Lima and Buratini2013). To enable the storage and the convenient use of the molecules stored in oocytes, several mechanisms should act efficiently (Gandolfi & Gandolfi, Reference Gandolfi and Gandolfi2001; Tomek et al., Reference Tomek, Torner and Kanitz2002). Some transcripts have already been associated with oocyte developmental competence (Caixeta et al., Reference Caixeta, Ripamonte, Franco, Junior and Dode2009; Katz-Jaffe et al., Reference Katz-Jaffe, McCallie, Preis, Filipovits and Gardner2009; Biase et al., Reference Biase, Martelli, Puga, Giuliatti, Santos-Biase, Fonseca Merighe and Meirelles2010; Kanka et al., Reference Kanka, Nemcova, Toralova, Vodickova-Kepkova, Vodicka, Jeseta and Machatkova2012; Bessa et al., Reference Bessa, Nishimura, Franco and Dode2013; Biase et al., Reference Biase, Everts, Oliveira, Santos-Biase, Fonseca Merighe, Smith, Martelli, Lewin and Meirelles2014), and those results support the hypothesis that specific RNAs or proteins produced during oogenesis contribute to oocyte competence (Sirard et al., Reference Sirard, Richard, Blondin and Robert2006).
It is estimated that during embryogenesis about 5000–10,000 genes are simultaneously expressed in oocytes with a high level of control (Niemann et al., Reference Niemann, Carnwath and Kues2007). The transcripts for key transcription factors represent a small number of copies and the ones that encode most of the structural proteins may represent approximately 2% of the mRNA pool (Yu et al., Reference Yu, Hecht and Schultz2002). Approximately 10–20% of total RNA consists of polyadenylated mRNAs (Niemann et al., Reference Niemann, Carnwath and Kues2007), which are associated with oocyte developmental competence (Pocar et al., Reference Pocar, Brevini, Perazzoli, Cillo, Modina and Gandolfi2001; Biase et al., Reference Biase, Fonseca Merighe, Santos-Biase, Martelli and Meirelles2008; Biase et al., Reference Biase, Martelli, Puga, Giuliatti, Santos-Biase, Fonseca Merighe and Meirelles2010).
The molecular mechanisms that govern oocyte competence are mostly still unknown. However, some oocyte-specific genes have been described revealing their importance in promoting embryogenesis (Katz-Jaffe et al., Reference Katz-Jaffe, McCallie, Preis, Filipovits and Gardner2009; Biase et al., Reference Biase, Martelli, Puga, Giuliatti, Santos-Biase, Fonseca Merighe and Meirelles2010; Belli et al., Reference Belli, Cimadomo, Merico, Redi, Garagna and Zuccotti2013; Bessa et al., Reference Bessa, Nishimura, Franco and Dode2013; Biase et al., Reference Biase, Everts, Oliveira, Santos-Biase, Fonseca Merighe, Smith, Martelli, Lewin and Meirelles2014). The profiling of gene expression in oocytes during maturation may help us understand the regulation of oocyte competence to mature and to sustain embryo development during the first two cleavages (Fair et al., Reference Fair, Carter, Park, Evans and Lonergan2007). It also promotes the identification of molecular markers for oocyte developmental potential. Nonetheless, most studies have addressed this subject in taurine subspecies (Bos taurus taurus). Here, we performed microarray-based transcriptome analyses of in vivo- and in vitro-matured bovine oocytes collected from Bos taurus indicus cows in order to enrich our knowledge of genes involved in the acquisition of oocyte competence.
Materials and methods
Estrous synchronization and superovulation protocols
Eight Nelore cows (crossbred) with good body condition and in reproductive age were synchronized on random days of the estrous cycle (D0) by intramuscular (IM) application of 2 mg estradiol benzoate RIC-BE (Tecnopec) and with placement of a bovine intravaginal progesterone device (Schering) for 8 days. On the fourth day (D4) follicle-stimulating hormone (FSH) treatment (Follitropin-V, Vetrepharm) was initiated with decreasing doses (80, 60, 40 or 20 mg FSH – IM) during 4 consecutive days. Simultaneously with the last FSH application (D7), cows received 0.150 mg Prolise (d-cloprostenol, IM), a PGF2α analogue (Tecnopec), and after 36 h (D8), the animals were separated randomly into two groups of four animals for oocyte collection.
One group (n = 4) was designated for collection of immature oocytes at the germinal vesicle stage (GV). The intravaginal progesterone device was removed from the cows at D8 and oocyte collection was performed by ultrasound-guided follicular aspiration ovum pick up (OPU) so that they could be matured in vitro.
The second group (n = 4) was designated for collection of in vivo-matured (MII) oocytes. The intravaginal progesterone device was removed from the cows at D8, 25 mg luteinizing hormone (LH) (Lutropin-V, Vetrepharm, IM) was administered and we performed OPU 22–24 h later for collection of MII oocytes. OPU was performed three times in the same animals at intervals of approximately 80 days between the synchronizations.
Oocytes selection criteria
During the three collections for each group, we selected only those follicles with diameter greater than 8 mm for OPU. Cumulus–oocyte complexes collected at GV stage were used for in vitro maturation if the oocyte presented homogeneous cytoplasm and at least two compact layers of cumulus cells. Cumulus–oocyte complexes collected at MII phase were used for further procedures if the oocyte presented homogeneous cytoplasm and several layers of expanded cumulus cells.
The procedures involving animal handling were approved by the Ethics Committee of the University of São Paulo – School of Animal Sciences and Food Engineering.
In vitro maturation of GV oocytes
Oocytes collected at GV phase were matured in vitro for 22 h in TCM-199 medium (Sigma) supplemented with 10% bovine fetal serum (Sigma), 5.0 μg/ml LH, 0.5 μg/ml FSH, 200 μM pyruvate (Sigma), and 50 μg/ml gentamicin (Sigma). In vitro maturation culture was carried out in 100 μl droplets (20–25 oocytes in each droplet) under mineral oil at 38.5°C and an atmosphere of 5% CO2 in air.
RNA extraction and amplification
For each of the three replicates, we selected 50 in vitro-matured and 50 in vivo-matured oocytes presenting the first polar body after removal of cumulus cells. The oocytes were pooled and stored at −80°C in calcium- and magnesium-free phosphate-buffered saline (PBS) with 0.1% polyvinyl alcohol (PVA) and 100 U/ml RNase inhibitor (Invitrogen). RNA extraction from oocytes was performed using RNeasy Protect Mini Kit (Qiagen) following the manufacturer's recommendations. Total RNA (~10 ng) was used as template for mRNA amplification with the SuperScript RNA Amplification Kit (Invitrogen) following the manufacturer's recommendations and oligo(dT)12–18 as primers. Samples of amplified mRNA (mRNAa) were assayed in a Bioanalyzer 2100 equipment to assess quality and integrity using the RNA 6000 LabChip kit, following the manufacturer's recommendations (Agilent Technologies).
Probe labelling and microarray hybridization
Hybridization probes from the mRNAa were prepared by reverse transcription followed by the incorporation of Cy3 or Cy5 fluorophores according to the recommendations of CyScribe Post-Labelling Kit and CyScribe GFX Purification (GE Healthcare). cDNA labelled with Cy3 or Cy5 was measured in a NanoDrop 2000 spectrophotometer. Hybridizations were performed on microarray slides (BLO Plus (GPL9176)), containing oligonucleotides (70-mer) representing 8400 bovine genes. This long oligo set includes 10 bovine control genes and 10 Stratagene Alien Genes spotted multiple times on the array. Approximately 400 ng of labelled cDNA was hybridized to the microarray, following the dye-swap schema with two technical replicates. Thus, for each of the three biological replicates, we hybridized four slides, composing 12 slides for the experiment. Hybridization was carried in an automated station (Tecan HS400) for 6 h at 42°C, for 6 h at 35°C and for 6 h at 30°C, followed by three washes in 2× sodium chloride and sodium citrate (SSC), 1% sodium dodecyl sulfate (SDS) at 37°C, three washes in 0.1× SSC, 0.1% SDS at 30°C and other three washes in 0.1× SSC at 25°C.
Data collection and analysis
The array images were digitalized by GenePix 4000B (Axon Instruments). The images were compiled using Imagene 5.0 (BioDiscovery), followed by the identification of the points of fluorescence and by background reading.
Raw intensities were normalized by the Lowess local regression using the LIMMA computational package according to procedures recommended for dye-swap labelling (Smyth & Speed Reference Smyth and Speed2003; Smyth Reference Smyth, Gentleman, Carey, Huber, Irizarry and Dudoit2005). The data obtained from the array spots were filtered and processed in order to eliminate poor quality, saturated or low fluorescence intensity spots relative to the background. Following data normalization, spots with intensity two-fold or greater than the background were considered for downstream analysis. Student's t-test was used to assess the statistical significance between the gene expression data generated from two experimental groups. Genes were inferred as differentially expressed between in vivo- and in vitro-matured oocytes if fold change was ≥2 and P-value <0.05.
The list of differentially expressed genes (DEG) was queried for biological processes potentially affected by in vitro maturation of oocytes using DAVID Bioinformatics Resources (v6.7, (Huang et al., Reference Huang, Sherman and Lempicki2009)). The probabilities of significance were adjusted for multiple hypotheses testing using false discovery rate (FDR) (Benjamini & Yekutieli Reference Benjamini and Yekutieli2001), and a Gene Ontology term was assumed enriched if FDR < 0.1. The DEGs were overplayed on the topology of a protein--protein network according to the human and mouse BioGRID (v.3.2) database (Chatr-Aryamontri et al., Reference Chatr-Aryamontri, Breitkreutz, Heinicke, Boucher, Winter, Stark, Nixon, Ramage, Kolas, O’Donnell, Reguly, Breitkreutz, Sellam, Chen, Chang, Rust, Livstone, Oughtred, Dolinski and Tyers2013) The network was built by expanding one protein interaction from each gene. The putative protein–protein network with DEGs in oocytes was visualized in Cytoscape (Shannon et al., Reference Shannon, Markiel, Ozier, Baliga, Wang, Ramage, Amin, Schwikowski and Ideker2003).
Validation of the microarray results
In order to validate the microarray, cDNA was synthesized from the mRNAa used for the preparation of probes. The nine genes with the greatest difference in expression between in vivo- and in vitro-matured oocytes and known to be associated with the physiology of oocyte maturation were chosen for validation. Five of those genes were up-regulated (EZR, EPN1, PSEN2, FST, and IGFBP3), and four genes were down-regulated (RBBP4, STAT3, FDPS, IRS1) in in vivo-matured oocytes. Primers and probes for TaqMan Gene Expression Assays were designed by the manufacturer (EZR (Bt03223252_m1), EPN1 (Bt03233436_g1), PSEN2 (Bt03237484_m1), FST (Bt03259671_m1) and IGFBP3 (Bt03223808_m1), RBBP4 (Bt03230465_g1), STAT3 (Bt03259866_g1), FDPS (Bt03216346_g1), Applied Biosystems). The exception was IRS1 whose primers (GGCAGATCTGGATAATCGGT, AATGGAAGCCACAGAGGACT) and probe (CGGACTCACTCTGCGGGCAC) were made to order.
Reverse transcription was performed with the SuperScript II kit, following the manufacturer's recommendations (Invitrogen) and oligo(dT)12–18 as primers. The real-time PCR reactions were set up according to the TaqMan PCR Master Mix Kit (Applied Biosystems). Real-time PCR data were normalized relative to H2A histone family gene, member Z (H2AFZ, Bower et al., Reference Bower, Moser, Hill and Lehnert2007) and fold changes were calculated according to the 2−ΔΔCT method (Livak & Schmittgen, Reference Livak and Schmittgen2001). We used the in vivo-matured oocytes as calibrator sample. The ΔCTs were used as input for analysis of variance (BioEstats 5.0) (Ayres et al., Reference Ayres, Ayres, Ayres and Santos2007) to assess the significance of differential gene expression between the two groups (Yuan et al., Reference Yuan, Reed, Chen and Stewart2006). Differential gene expression between in vivo- or in vitro-matured oocytes was assumed significant when P < 0.05.
Results
Genes expressed in bovine oocytes matured in vivo and in vitro
Our experiment yielded 1488 genes quantified with two-fold or greater intensity than the background. Among these, 51 genes were up-regulated (≥2-fold) in in vivo-matured oocytes compared with in vitro counterparts (Table 1). By comparison, 56 genes were down-regulated (≥2-fold) in in vivo-matured oocytes (Table 2). Among these 107 DEGs genes, 25 genes were annotated to the enriched Gene Ontology (GO) biological process ‘negative regulation of cellular process’ (FDR < 0.1, Table 3). Noticeably, several of these 25 DEGs are also possibly associated with regulation of metabolic processes (FDR < 0.2, Table 3). Further inspection of these 25 DEGs demonstrated that nine of them are involved in regulation of transcription (CTNNB1, DEDD, FST, HMGB2, HDAC2, MYF6, NR1H2, RBBP7, STAT3).
Genes up-regulated in in vivo-matured oocytes compared with in vitro counterparts
Genes down-regulated in in vivo-matured oocytes compared to in vitro counterparts
Top 10 GO biological processes associated with DEG between in vivo- and in vitro-matured oocytes
Next, we searched for altered gene expression that would affect interacting proteins. Most DEGs were part of a protein–protein network. Forty out of 51 genes up-regulated in in vivo-matured oocytes composed a protein interaction network (Fig. 1 A), four of those genes identified two pairs of direct protein-protein interaction (Fig. 1 A, inset). Two of those genes are associated with the transcription complex, namely: polymerase (RNA) II (DNA directed) polypeptide A and the Mediator Complex Subunit 29. By comparison, 41 of the 56 genes up-regulated in in vitro-matured oocytes composed another protein–protein network (Fig. 1 B), 10 of which identified direct protein interactions (Fig. 1 B, inset). It is noteworthy that we found four genes associated with chromatin remodelling factors that were positively modulated by in vitro culture, namely: H3 histone, family3A, retinoblastoma binding protein 4 and 7, and bromodomain adjacent to zinc finger domain, 1A.
Regulatory protein–protein network of genes affected by the in vitro culture of oocytes. (A) Genes up-regulated in in vivo-matured (MII) oocytes collected in vivo. (B) Genes down-regulated in MII oocytes collected in vivo. Differentially expressed genes (DEGs) are marked in red or blue, and yellow depicts the proteins potentially interacted with DEGs. The insets highlight the direct connection between proteins whose coding genes are differentially regulated by in vitro maturation.
Confirmation of differentially expressed genes by real-time PCR
We used real-time PCR to validate our microarray results. The nine genes tested were differentially expressed in in vivo-matured oocytes (EZR, EPN1, IRS1, FDPS, FST, IGFBP3, PSEN2, RBBP4, STAT3) compared with in vitro-matured oocytes in the two analyses, microarray and RT-PCR (P < 0.05. Fig. 2).
Relative expression of transcripts in in vivo- and in vitro-matured oocytes. Significant differences (P < 0.05) between groups (in vivo versus in vitro) are denoted by an asterisk. The results of three replicates are shown.
Discussion
We determined whether the in vitro maturation process affected the gene expression of oocytes collected from Bos taurus indicus cows. With our investigation, we showed that 107 genes have altered expression due to the in vitro maturation system. Functional annotation of the data suggests that dysfunctional gene expression is not random and mostly affected the metabolism of oocytes. Inspection of the GO annotation of the genes suggests that one of the metabolic processes highly affected is the regulation of RNA synthesis.
Corroborating our results, 21 of the DEGs were previously shown to be associated with oocyte developmental competence, 13 of those genes were up-regulated in in vivo-matured oocytes (DGKA (Beltman et al., Reference Beltman, Forde, Furney, Carter, Roche, Lonergan and Crowe2010), GHR (Caixeta et al., Reference Caixeta, Ripamonte, Franco, Junior and Dode2009), FST (Bonnet et al., Reference Bonnet, Bevilacqua, Benne, Bodin, Cotinot, Liaubet, Sancristobal, Sarry, Terenina and Martin2011), HMGB2 (Corcoran et al., Reference Corcoran, Rizos, Fair, Evans and Lonergan2007), TACC3 (Hao et al., Reference Hao, Stoler, Sen, Shore, Westbrook, Flickinger, Herr and Coonrod2002), IGFBP3 (Sawai Reference Sawai2009), EZR (Heng et al., Reference Heng, Cervero, Simon, Stephens, Li, Zhang, Paule, Rainczuk, Singh and Quinonero2011), KEAP1 (Powell et al., Reference Powell, Manandhar, Spate, Sutovsky, Zimmerman, Sachdev, Hannink, Prather and Sutovsky2010), SMARCC1 (Lisboa et al., Reference Lisboa, Bordignon and Seneda2012), PLK1 (Sun et al., Reference Sun, Liu and Sun2012), NGFRAP1 (Jiang et al., Reference Jiang, Xiong, Cao, Xia, Sirard and Tsang2010), NUMA (Kolano et al., Reference Kolano, Brunet, Silk, Cleveland and Verlhac2012) and EPN1 (Liu & Zheng, Reference Liu and Zheng2009)) and eight of them were up-regulated in in vitro-matured oocytes [IRS-1 (Yamamoto-Honda et al., Reference Yamamoto-Honda, Honda, Ueki, Tobe, Kaburagi, Takahashi, Tamemoto, Suzuki, Itoh and Akanuma1996), STAT3 (Mohammadi-Sangcheshmeh et al., Reference Mohammadi-Sangcheshmeh, Held, Ghanem, Rings, Salilew-Wondim, Tesfaye, Sieme, Schellander and Hoelker2011), CCNB1 (Liu et al., Reference Liu, Yin, Zhao, Yun, Wu, Jones and Lei2012), RBBP4 and RBBP7 (Gasca et al., Reference Gasca, Pellestor, Assou, Loup, Anahory, Dechaud, De Vos and Hamamah2008), ATF2 (Vigneault et al., Reference Vigneault, McGraw and Sirard2009), TPX2 (Brunet et al., Reference Brunet, Dumont, Lee, Kinoshita, Hikal, Gruss, Maro and Verlhac2008) and HDAC2 (Caixeta et al., Reference Caixeta, Sutton-McDowall, Gilchrist, Thompson, Price, Machado, Lima and Buratini2013)]. This observation supports our approach and analysis. Most importantly, we showed 86 new potential biomarkers associated with oocyte competence. Further investigation will be required to conclusively demonstrate that the transcription of these 86 genes are specifically altered in oocytes collected from B. taurus indicus and matured in vitro.
Functional analysis of the DEGs revealed that 37 genes were annotated to ‘regulation of cellular metabolic process’, which was previously shown to be important for the maturation of oocytes (Fair et al., Reference Fair, Carter, Park, Evans and Lonergan2007; Katz-Jaffe et al., Reference Katz-Jaffe, McCallie, Preis, Filipovits and Gardner2009). Interestingly, 26 (of the 37) DEGs were also functionally related to ‘negative regulation of cellular process’, and those genes are potentially important for cytoplasmic maturation (Ferreira et al., Reference Ferreira, Vireque, Adona, Meirelles, Ferriani and Navarro2009), and developmental potential of the oocytes. The dysregulation of a metabolic process such as the synthesis of RNA, due to in vitro maturation, is likely to affect the transcription during cleavage stages of development (Smith et al., Reference Smith, Everts, Sung, Du, Page, Henderson, Rodriguez-Zas, Nedambale, Renard, Lewin, Yang and Tian2009) and alter cleavage kinetics during embryo development (Knijn et al., Reference Knijn, Gjørret, Vos, Hendriksen, van der Weijden, Maddox-Hyttel and Dieleman2003).
Our results of the transcriptome analysis were further supported by protein–protein interactome. The formation of protein–protein networks composed of the majority of genes either up-regulated (Fig. 1 A) or down-regulated (Fig. 1 B) regulated in in vivo-matured oocytes strongly suggests biological co-regulation of such genes in MII oocytes. Interestingly, we observed subsets of DEGs whose protein may form regulatory complexes (Fig. 1, insets). Two examples of gene co-expression and protein–protein interaction are potentially associated with gene regulation. First, the transcripts of MED29 and POLR2E are up-regulated in in vivo-matured oocytes, where this protein complex may function in the elongation phase of transcription (Takahashi et al., Reference Takahashi, Parmely, Sato, Tomomori-Sato, Banks, Kong, Szutorisz, Swanson, Martin-Brown, Washburn, Florens, Seidel, Lin, Smith, Shilatifard, Conaway and Conaway2011). Second, we found the complex formed around the retinoblastoma binding protein 4 (RBBP4) in the genes up-regulated in in vitro-matured oocytes. The abnormal abundance of this complex may contribute to negative regulation of genes important for embryo development (Wolffe et al., Reference Wolffe, Urnov and Guschin2000). These results showed that the in vitro culture system also disturbs the regulation of oocyte's gene expression at the transcriptional level.
In summary, we established the transcript profile of in vivo- and in vitro-matured oocytes of Bos taurus indicus cows using microarray technology. Our experiment allowed us to uncover genes potentially involved in the control of oocyte competence. In light of our results, we suggest that the harmonious function of metabolism and regulation of gene expression is pivotal for the acquisition of oocyte developmental competence. The identification of potential competence markers will be useful for developing better in vitro culture conditions to allow the oocyte to adequately obtain competence.
Acknowledgements
We thank Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP 05/58702–4 and 05/59694-5) and Universidade de São Paulo (USP/FZEA) for financially supporting the project.
