Sequential analysis of global gene expression profiles in immature and in vitro matured bovine oocytes: potential molecular markers of oocyte maturation

Transition from the maternal to -embryonic genome control of development occurs relatively late in cattle, during the fourth cell cycle [1]. Thus, the oocyte is the main driver of early embryo development, drawing on maternal mRNAs and proteins accumulated during the oocyte growth phase [2, 3]. In order to achieve a developmentally competent status, the oocyte has to gradually undergo a number of physiological changes that include physical and molecular remodeling [4]. During fetal life, mammalian oocytes initiate meiosis and become arrested at the diplotene stage of prophase I (dictyate stage). The ability of these oocytes to resume meiosis and to complete the first meiotic division is acquired sequentially during their growth phase [5]. In fully grown oocytes, meiotic resumption and nuclear maturation, in response to the preovulatory gonadotrophin surge in vivo or release from the follicle in vitro, is characterized by germinal vesicle breakdown (GVBD), chromosomal condensation, cumulus cell expansion, hyaluronic acid and cyclic AMP production [5, 6, 7, 8], and progression through metaphase I to anaphase and telophase, with extrusion of the first polar body and arrest at metaphase II (MII) until reactivation at fertilization. The basic molecular machinery governing these developmental processes is relatively well conserved across mammalian species [9, 10]; however, there is a clear difference in the timing of these processes between species (reviewed in [11, 12]).

The origin of the oocyte and more specifically, the environment in which oocyte growth and maturation occur [13, 14, 15, 16, 17] has been implicated as an important determinant of the subsequent developmental competence of the oocyte. Transcriptional profiling of in vivo and in vitro matured (IVM) oocytes in cattle [18, 19, 20], humans [21], and rhesus monkeys [22] have shown variations in a number of genes and distinct pathways, which may have consequential effects during the post fertilization development. For example during the process of in vitro embryo development, while maturation and fertilization proceed apparently normally (based on first polar body extrusion and mitotic cleavage, respectively), the proportion of embryos reaching the transferable (blastocyst) stage rarely exceeds 40 to 50% and those that do reach this stage are often compromised in quality and further developmental competence [23], an effect partly attributed to inadequate oocyte cytoplasmic maturation [24, 25]. In contrast, fertilization and culture in vitro of oocytes matured in vivo results in high rates of blastocyst development [15, 26] providing further evidence of the importance of oocyte quality in determining developmental competence.

To date, efforts to characterize developmentally competent oocytes have been hampered partly by the difficulty in assessing cytoplasmic maturation [27], and the small volume of material available for analysis. Currently, while nuclear maturation can be assessed by the extrusion of the first polar body and the formation of the second metaphase plate, there is no reliable assay for the assessment of oocyte cytoplasmic maturation, other than the development of the fertilized oocyte to a live offspring [24, 27, 28]. This lack of oocyte quality markers has led some researchers to use embryo morphology or blastocyst rate as alternative early quality predictors of developmental competence [29, 30, 31]. Until the mechanisms involved in establishing oocyte quality are elucidated, any effort to use assisted reproductive technologies in the treatment of human infertility or in animal production will be inefficient [28]. Analysis of the oocyte transcriptome during maturation using global mRNA analysis techniques provides a resource that can be mined and expanded continuously with new software and sequencing techniques to identify genes involved in the processes of meiotic and cytoplasmic oocyte maturation and the specific checkpoints regulating acquisition of full competence [32].

It is well understood that mRNAs stored during the oocyte growth phase are systematically, and in a step-wise manner, degraded or translated to form proteins that regulate subsequent developmental processes [33, 34]. As a result, down regulation of stored transcripts during oocyte maturation is believed to occur. Although GV intact oocytes show some transcriptional activity [4, 35], there are conflicting reports on the occurrence of transcription following the resumption of oocyte meiotic maturation. Comparison of immature (germinal vesicle stage, GV) and in vivo matured oocytes in mice [36], and in humans [32] indicated up regulation of a number of transcripts during oocyte maturation. However, other studies comparing immature and in vitro matured (IVM) oocytes in cattle [37] and in mice [38] reported the absence of over-expressed transcripts during oocyte maturation. Given these apparently conflicting observations, and the opportunity to benefit from the increased coverage (more than 92%) of the bovine genome sequence [39], we examined transcriptional activities during bovine oocyte maturation and present a dataset that provides a unique reference resource for studies concerned with the molecular mechanisms controlling oocyte meiotic maturation in cattle. These data have been superficially referred to in a previous review paper [2] but a thorough analysis of the gene lists and associated ontologies has not yet been published.

The issue of transcript regulation during oocyte maturation is a controversial topic in developmental biology. Various previous efforts [32, 36, 37, 38] to decipher the transcriptional changes during oocyte maturation have been masked by contradictory outcomes. In the current study, we established the global changes in transcriptomic profile during meiotic maturation in bovine oocytes. Moreover, in an effort to examine the possible sources of conflicting reports, we further investigated transcription in the presence of transcription inhibitor, effects of primers, amplified product sizes, reference genes and sample types (in vitro and in vitro) on transcript abundance. Here we (1) report the existence of transcription activities during bovine oocyte maturation, (2) show suppression of over-expressed genes when oocytes were matured in the presence of an inhibitor of transcription suggesting that the observed transcript changes were newly synthesized, (3) report that the new transcription was not sample type (in vitro or in vivo) -specific, (4) report primers used during cDNA synthesis and reference genes used for normalization have an impact on the interpretation of the gene expression data, but not the amplified product sizes.

The genome-wide interrogation of immature and in vitro matured bovine oocytes on the Affymetrix GeneChip bovine array identified a number of differentially expressed genes, the vast majority (~75%) of which were over-expressed in immature oocytes (Table 1). It is well accepted that mammalian oocytes have already accumulated the majority of their transcripts at the fully grown immature (GV) stage that will drive subsequent development through degradation, translation and post-transcriptional modifications [33, 34] up to embryonic genome activation. Therefore, the observation of a massive reduction of the initial transcript stock in our study was in line with the current understanding of events during oocyte maturation. On the other hand, the detection and confirmation of certain over-expressed transcripts during bovine oocyte maturation suggests the existence of transcription, perhaps to complement the depleting transcriptional stock.

Despite the multiple roles of a particular gene [43], Ingenuity Pathway Analysis classified the differentially regulated genes into various associated functional network groups (Table 2). These include cellular assembly, molecular transport, post-translational modification and cell to cell signaling, all of which occur during oocyte maturation. For example, cell-to-cell signaling between oocytes and their surrounding somatic cells is important for oocyte cytoplasmic maturation and the acquisition of developmental competence. This is a bidirectional communication mediated through the transport of various growth factors, such as GDF9 and BMP15, from oocytes to their surrounding cumulus cells [44, 45, 46], and cyclic adenosine monophosphate (cAMP) from somatic cells to the oocyte [47] via gap junctions.

Generally, the sequence of events leading to the GVBD (Germinal Vesicle Breakdown) and the requirements for transcription and/or protein synthesis differs markedly between species [12, 48]. For example, in frog, mouse, rat and fish oocytes, high levels of cAMP prevent oocyte maturation in vitro, while a decrease in oocyte cAMP is associated with the resumption of meiosis [49, 50]. In contrast, maturing oocytes from pig, sheep, cattle and rabbit exhibit a transient increase rather than a decrease in cAMP levels, and treatments that increase cAMP levels can induce oocyte maturation [49, 51, 52, 53]. Similarly, an earlier study [12] confirmed the requirements of transcription and protein synthesis as requirements for GVBD in domestic animals (sheep, cattle and pigs) while neither event is required for the initiation of maturation in mouse oocytes. Generally our finding is in line with earlier studies in bovine [54, 55, 56] that observed various over-expressed transcripts during oocyte maturation.

In order to verify if the over-expressed transcripts were transcribed following submission of oocytes to IVM, oocytes were matured in the presence of the RNA polymerase II inhibitor, α-amanitin for 24 h. The resulting transcript profile was similar to that of the immature oocytes, which is consistent with the notion that 24 h exposure to α-amanitin prevents meiotic resumption in most oocytes [57]. This study [57] also reported that addition of α-amanitin after 3 h of culture had no effect on meiotic maturation. However, in the current study exposure of the oocytes to the α-amanitin treatment following an initial 3 h culture in α-amanitin-free medium resulted in a similar level of expression to that observed when the inhibitor was present throughout. These findings suggest that some de novo transcription is occurring in bovine oocytes following the resumption of meiosis.

In order to examine the contributions of some downstream analysis procedures on the final transcript data, we further examined the implications of primer choices during cDNA synthesis. Although not significant, there was a tendency for higher expression fold change (ratio) for oligo (dT)-based cDNA preparations compared to random-based cDNAs. This suggests the preferential amplification of oligo (dT)-based primers, and the finding is in agreement with most other previous observations [54, 58, 59, 60, 61], although another study [19] reported an identical results irrespective of the primers used. The fact that anchored oligo (dT) was used in our study may have narrowed the difference. It has been shown previously that anchored oligo (dT) primers are better than the conventional oligo (dT) primers in maintaining the fidelity of the probes, as the latter generates a high frequency of truncated cDNA through internal poly (A) priming [62]. This observation further signifies the contribution of primers to the final conclusions, and the need to select appropriate primers commensurate with the sample type for analysis. Therefore, it is possible to speculate that this may have also contributed to the earlier contradictory reports on the occurrence of transcription during meiotic maturation [32, 36, 37, 38].

Increasing stringency by controlling primers for cDNA synthesis, designing intron-spanning primers for amplification and the use of alternative validated reference genes during the analysis appears to reduce the number of significant genes at 24 h. Interestingly, when we analysed the kinetics of transcript expression during maturation the abundance of several transcripts was significantly higher at 12 h compared to 0 h. The kinetics of bovine oocyte maturation has been well described [63, 64, 65, 66]. In most oocytes GVBD occurs between 4 and 8 h after initiation of maturation, and has occurred in the majority of oocytes by 8 h. By 12 h the majority of oocytes have reached metaphase I. Progression from GVBD through the subsequent stages of meiosis is under the control of the anaphase promoting complex (APC) which is mainly regulated through sequential polyadenylation and deadenylation of transcripts [67] and the increased abundance of these transcripts at this time may reflect their association with APC. These genes are implicated in various developmental activities including cell signaling, apoptosis and membrane trafficking (ANXA1) [68], and cumulus cell expansion (PLAU) (reviewed in [55]).

Irrespective of the introduction of increased stringency measures, the abundance of STC1 and LUM was significantly higher in both in vivo and in vitro matured oocytes compared to 0 h (GV stage oocytes) (Figure 5 and Figure 6). Furthermore, transcript abundances were maintained at 0 h levels when oocytes were matured in the presence of α-amanitin. Taken together, these findings strongly suggest de novo transcription of STC1 and LUM following the resumption of meiosis. Analysis of various studies suggests STC1 has effects on metabolism, reproduction, and developmental processes in addition to affecting mineral homeostasis (reviewed in [69]. STC1 expression was highest in mouse ovary, with lower but detectable levels in most other tissues [70]. Based on this result and the initial implication of the gene in mineral metabolism, it was suggested that STC1 may have acquired an important function in reproduction during its evolution in mammals [70]. Similarly, increasing evidence suggests that LUM may also serve as a regulatory molecule of several cellular functions [71, 72]. Previous study in mice using Northern and In situ hybridization indicated that, in early stages of embryonic development before day 7 post-coitus, the embryo does not express LUM or expresses only very low amounts [73]. This is the first study to examine and reveal the expression of these two genes (LUM and STC1) during oocyte maturation. Based on the consistent expression pattern in repeated experiments of in vitro and in vivo derived oocytes, it is plausible to speculate that these two genes (LUM and STC1) may be potential molecular markers of oocyte maturation and may contribute to the early events of embryo development.

We have used global microarray analysis to establish the molecular transcriptome blueprint of immature and matured oocytes and to identify and validate genes that are unique to and predominantly expressed in bovine immature and IVM oocytes. The genes identified will be invaluable in further studies examining the processes of oocyte maturation in addition to addressing the existing conflicting issue of transcription during meiotic maturation. Moreover, it will enable the comparisons across many species and contributes to the goal of improving assisted reproductive technology.

All chemicals, unless stated otherwise, were purchased from Sigma-Aldrich Chemical Inc. (St. Louis, MO, USA). All experimental procedures involving animals were licensed by the Department of Health and Children, Ireland, in accordance with the Cruelty to Animals Act (Ireland 1897) and the European Community Directive 86/609/EC and were sanctioned by the Animal Research Ethics Committee of University College Dublin. Four experiments each with different objectives and methodologies were carried out, as described separately in the result sections.