Expression of Variant Ribosomal RNA Genes in Mouse Oocytes and Preimplantation Embryos

Abstract

Ribosomal DNA (rDNA) is not composed of multiple copies of identical transcription units, as commonly believed, but rather of at least seven rDNA variant subtypes that are expressed in somatic cells. This finding raises the possibility that ribosome function may be modulated as proposed by the ribosome filter hypothesis. We report here that mouse oocytes and preimplantation embryos express all the rDNA variants except variant V and that there is no marked developmental change in the qualitative pattern of variant expression. The maternal and embryonic ribosome pools are therefore quite similar, minimizing the likelihood that developmental changes in composition of the ribosome population are critical for preimplantation development.

Mouse oocytes and preimplantation embryos express different rRNA genes, but there is no dramatic change in the variants that are expressed during preimplantation development.

INTRODUCTION

Recent studies suggest that ribosomal RNA (rRNA) synthesis and regulation are not as simple as previously perceived (i.e., serving strictly a housekeeping function and regulated by universal mechanisms) and challenge the prevailing notion of ribosomal DNA (rDNA) array and regulation [1]. The rDNA array is generally assumed to contain multiple copies of identical transcription units. The foundation of this concept was laid down more than three decades ago. The repetitive nature of the rDNA was revealed when it was first isolated from Xenopus genomic DNA, and rRNA-DNA hybridization experiments showed that it contained multiple copies [2], which are in a tandem array [3]. For mammals, cytogenetic and in situ hybridization showed that rDNA arrays are localized to specific chromosomal sites [4–6]. The assumption of homogeneity of rDNA transcription units has likely contributed to the fact that only one mouse and one human rDNA transcription units have been sequenced, and rDNA loci constitute the major gaps in the human and mouse genome sequencing projects (e.g., GenBank Release 154.0, June 2006). This concept has also influenced the study of rDNA transcription and posttranscription regulation. The components of RNA polymerase I transcription complex and their assembly sequence at the rDNA promoter were dissected biochemically in many species, from yeast to human. Although these studies led to many insightful findings regarding the molecular regulation of rRNA synthesis, they focused on one type of rDNA sequence (for each species), thus overlooking potential variation in the rDNA array [7].

Heterogeneity in rDNA array gained support from two recent studies [8, 9], and using restriction fragment length polymorphism as a guide, we previously identified and cloned seven subtypes of rDNA variants (v-rDNA), which are classified by variant-specific single nucleotide differences that occur in the leader region, which is transcribed but excised during processing and therefore not present in mature rRNAs [9]. Although it is not known whether sequence differences occur in the regions of variant rDNA genes that encode the mature rRNAs, the existence of v-rDNA subtypes could indicate functional differences between ribosomes and is consistent with the ribosome filter hypothesis [10]. Although a number of regulatory mechanisms affect translation, the ribosome itself is not generally considered among them. The ribosome filter hypothesis proposes that specific mRNA and rRNA interactions are important for controlling translation. This proposal is mainly based on observations that many mRNAs contain sequences that are homologous to rRNA. Conceivably, these interactions of mRNA and rRNA open possibilities for modulation of ribosome activity. The filter hypothesis also predicts heterogeneity in rRNA, which would influence the efficiency of the ribosome in translating different mRNAs.

The present study attempts to test the ribosome filter hypothesis in the context of a paradox in mammalian oocyte biology. Mouse oocytes increase in volume ∼200- to 250-fold during their growth phase, which takes ∼2.5 wk. During this time, oocytes synthesize and accumulate large amounts of rRNA that is incorporated into ribosomes that will in turn support protein synthesis during early development. In light of the importance of maternal ribosomes in preimplantation development and the extensive amount of energy consumption that must be devoted to ribosomal biosynthesis—more than 60% of the total amount of RNA in oocytes is rRNA—it is intriguing that mouse oocytes destroy a large amount of rRNA and ribosomal protein mRNAs during maturation. Degradation of bulk RNA during meiotic maturation was first noted more than two decades ago [11, 12], and based on nucleic acid hybridization, rRNA degradation constitutes the bulk of the 80 pg reduction of total cytoplasmic RNA during meiotic maturation [13].

The notion of ribosome destruction during oocyte maturation received additional support from a recent microarray study [14] in which it was observed that of the 78 transcripts that encode ribosomal proteins, 68 were significantly degraded during the course of oocyte maturation, strongly suggesting a programmed reduction of ribosomes during oocyte maturation. This reduction is consistent with the maturation-associated decrease in the rate of total protein synthesis [15] but nevertheless is perplexing because it appears wasteful in respect to early embryos, especially considering the huge energy and resource expenditure in ribosome biogenesis. Furthermore, the mouse maternal ribosome storage is evidently not sufficient to support embryonic development beyond the zygotic genome activation because shortly after the 2-cell stage when the zygotic genome activation becomes apparent [15], embryos resume rRNA and ribosomal protein synthesis.

A plausible but unconventional solution to the paradox invokes the ribosome filter hypothesis that assumes heterogeneity in the ribosome population and a differential requirement of subtypes of ribosomes during the development of oocytes and early embryos. Degradation of maternal ribosomes would pave the way for their replacement with zygotic ribosomes of a different composition and function. A key piece of evidence for testing the differential requirement of subtypes of ribosomes would be to demonstrate that a change in expression of differential subtypes of rRNA occurs during early development. As described above, we identified and cloned seven subtypes of v-rDNA, and the expression profile of these v-rDNAs in various tissues suggests that they are independently regulated [9]. This finding provided us the opportunity to examine if different subtypes of rDNAs are transcribed in oocytes and early embryos. We report here that different subtypes of rDNA variants are expressed in oocytes and early embryos, but that there is no marked change in the types of variants that are expressed.

MATERIALS AND METHODS

Oocyte and Embryo Collection and Culture

Collection and culture of meiotically incompetent oocytes, full-grown oocytes, and 2- and 8-cell stages were performed as previously described [16, 17]. Briefly, full-grown germinal vesicle (GV)-intact oocytes were collected from 6- to 8-wk-old CF1 female mice 46–48 h after equine chorionic gonadotropin (eCG) injection (5 IU [international units]). Embryos were obtained from CF1 female mice mated to B6D2F1/J males. Females were superovulated by eCG (5 IU) followed by an injection of 5 IU hCG (human chorionic gonadotropin) 48 h later, and embryos were harvested 48–52 h (2-cell embryos) and 70 h (8-cell embryos) post-hCG injection. Meiotically incompetent oocytes were obtained from 13-day-old female CF1 mice by incubating pieces of ovarian tissue in Ca²⁺- and Mg²⁺-free CZB medium [18] containing 1 mg/ml collagenase and 0.2 mg/ml DNase I at 37°C for up to 120 min. Oocytes and embryos were cultured in CZB or KSOM [19] medium, respectively, in an atmosphere of 5% CO₂/5% O₂/90% N₂. All the animal experiments were approved by the Institutional Animal Use and Care Committee and were consistent with National Institutes of Health guidelines.

Real-Time PCR

Total RNA from oocytes or embryos (2000 or 4000) and liver tissue was isolated using RNAqueous-Micro Kit (Ambion, Inc., Austin, TX) according to the manufacturer's instructions. Complementary DNA was prepared by reverse transcription of the same amount of total RNA with random hexamer primers, T4 gene 32 protein (New England Biolabs, Ipswich, MA), and Superscript III enzyme (Invitrogen, Carlsbad, CA) at 55°C for 120 min. Prepared cDNA was purified and precipitated with linear polyacrylamide (Ambion, Inc.), ammonium acetate, and phenol/chloroform/isoamyl alcohol. Variant-specific primers and PCR conditions (primers and cycling parameters) are as previously described [9]. Because the PCR conditions are different for each variant, the assay measures the relative amounts of each variant.

RESULTS AND DISCUSSION

To examine the expression pattern of v-rDNA, total RNA was isolated from meiotically incompetent oocytes (i.e., growing oocytes that are not capable of resuming meiosis when placed into a suitable culture medium) and 2- and 8-cell embryos. Variant rDNA-specific RT-PCRs were then performed to detect the promoter region of 47S rRNA from the seven subtypes of rDNA. Attempts to amplify the RNA using a random hexamer-T7 primer and second-round amplification system were unsuccessful. Because of the small quantity of the specimen (i.e., oocytes and early embryos), thousands of oocytes/embryos were required to obtain sufficient amounts of total RNA for cDNA synthesis, that is, 360 ng for each sample. Note that although this amount was much lower than the 2 μg previously used [9], v-rDNA signals were clearly detected in oocytes and embryos (Fig. 1).

FIG. 1.

Expression of variant-rDNAs (v-rDNAs) in early stage embryos. Variant-rDNA expression was assessed by RT-PCR in mouse incompetent oocytes, 2-cell embryos, 8-cell embryos, and liver tissue (control). RNA (260 ng total) was used for reverse transcription, and all the variant-specific PCRs were performed with the same cDNA preparation. Primer specificity (I to VII) is indicated on the left of the gel image. For each PCR, a set of controls was included to monitor the specificity (the left seven lanes) and the relative quantity and integrity of the RNA preparation (Actb). A PCR detecting all the v-rDNA transcripts (47S) is also included. The templates are indicated above the gel image: I to VII, cloned v-rDNA controls; INC, incompetent oocytes; 2C, 2-cell embryos; 8C, 8-cell embryos; Liv, liver tissue; M_r, molecular weight ladder marker.

Total liver RNA was a positive control, which showed expression of v-rDNA I/II and IV (Fig. 1), a result consistent with previous observations [9]. As anticipated, no signal was detected when total RNA isolated from full-grown oocytes was used (data not shown). A hallmark of oocyte development in all the species examined to date is that starting around the midgrowth phase, transcription decreases such that full-grown oocytes are transcriptionally quiescent [20]. An outcome of this would be that all the 47S rRNA would already have been processed to mature forms. Thus, the absence of a signal from full-grown oocytes was expected and served as a negative control for the specificity of the assay because the RT-PCR scheme allows detection of only nascent rRNA (i.e., 47S) and not the fully processed species, that is, 18S and 28S. The input control was an RT-PCR of Actb mRNA.

Ribosomal RNA transcription initiates during the 2-cell stage [21, 22] and is quite robust in 8-cell embryos [21]. Although the linearity of product formation was not established for conditions of the RT-PCR assay, the presence of the 47S precursor in 2-cell embryos and the further increase between the 2- and 8-cell stages is consistent with these previous findings. Moreover, the increase in band intensity of the v-RNAs between meiotically incompetent oocytes and 8-cell stage embryos is also consistent with production of zygotic ribosomes by the 8-cell stage. In addition, the results demonstrate that all the v-rDNAs identified so far, with the exception of v-rDNA V, are actively transcribed in incompetent oocytes and early stage embryos. No other tissue to date has been shown to utilize six subtypes of v-rDNA. Previously, the highest v-rDNA subtype usage was found in the brain and testis (mixed cell types), where five subtypes (i.e., I/II, III, IV, VI) are used [9], and in corneal epithelium (a single cell type), where five subtypes (i.e., I/II, IV, VI, VII) are detected [23]. The present study also failed to detect transcription from v-rDNA V, making it the least used v-rDNA so far.

The relative distribution of the different subtypes is very similar in oocytes and 8-cell embryos (Fig. 2); the different distribution observed in 2-cell embryos is likely a consequence of the maternal-to-embryo transition in which maternal ribosomes that are largely degraded during maturation are being replaced by zygotic ribosomes. Thus, the maternal pool of ribosomes is essentially replaced by a zygotic pool of ribosomes of a similar composition by at least the 8-cell stage. It is formally possible that developmental changes in ribosomes associated with polysomes contain different variants, a possibility consistent with the ribosome filter hypothesis. In addition, it is also possible that there are variants yet to be identified. Nevertheless, in the absence of a major developmental change in variant expression, our findings do not provide evidence supporting the ribosome filter hypothesis and also suggest that developmental changes in the composition of the ribosome population are not critical for preimplantation development.

FIG. 2.

Relative distribution of different v-rRNAs in oocytes and preimplantation embryos. The data shown in Figure 1 were quantified using NIH Image J software. The intensity of the signal of variants I/II was set at 1 and the relative abundance of the other variants was expressed relative to that value. Variant V was not included because it was not detected. INC, incompetent oocytes; 2C, 2-cell embryos; 8C, 8-cell embryos. The numbers shown above the bars for INC correspond to the different variants.