Effect of oocyte quality on the relative abundance of specific gene transcripts in bovine mature oocytes and 16-cell embryos

Abstract

Although the developmental potential of oocytes is related to oocyte quality, whether the expression of specific genes is altered in oocytes of different quality and in resulting embryos is not known. Semi-quantitative reverse transcription-polymerase chain reaction was used to compare the relative abundance of 2 transcripts for housekeeping proteins (β-actin and ribosomal protein L30) and 3 transcripts for growth factor ligand or receptors (platelet derived growth factor receptor α (PDGFRα), basic fibroblast growth factor (bFGF)), in mature bovine oocytes of high versus low developmental potential. The transcripts for L30, PDGFRα, and bFGF in 16-cell embryos originating from these oocytes were also examined. No significant effect of oocyte quality was detected for any of the transcripts examined from oocytes or 16-cell embryos. In conclusion, a lower developmental potential of oocytes with advanced signs of atresia, was not associated with a lower level of abundance of the transcripts examined.

Bovine cumulus-oocyte complexes (COC), that were recovered from ovaries after slaughter, displayed a wide range of quality, as judged by their physical appearance. To a certain extent, the physical appearance is correlated with developmental competence. In that regard, COC with early signs of atresia (slight expansion of the cumulus and/or slight granulations of the cytoplasm), have higher developmental competence than those considered healthy (1,2). Advanced atresia; as visualized by full expansion of the cumulus, with the presence of dark clumps and heavy granulations in the cytoplasm, or the absence of cumulus cells, results in low developmental competence (3).

The developmental competence of oocytes may be a function of the presence or abundance of specific transcripts in their mRNA pools, since the earliest stages of embryogenesis in mammals and other animals are regulated by maternally-inherited RNAs and proteins stored within the oocyte (4). Therefore, oocyte quality could be correlated with the prevalence of transcripts in the oocyte itself, as well as the prevalence of these maternal messages in early embryos. Moreover, since activation of the embryonic genome must be mediated by maternal proteins (5), oocyte quality could have an effect on this event and, therefore, on the prevalence of transcripts synthesized after activation of the embryonic genome.

Previous studies have demonstrated the beneficial effects of growth factors, such as platelet derived growth factor (PDGF), transforming growth factor β (TGFβ), and basic fibroblast growth factor (bFGF) on bovine embryo development in vitro (6,7). Moreover, the presence of the transcripts encoding these growth factors and their receptors, have been demonstrated in mature bovine oocytes and in vitro-produced embryos (8,9,10). However, it is not known how oocyte quality affects the abundance of these transcripts. The first objective of the present work was to compare the relative abundance of transcripts for β-actin, ribosomal protein L30 (2 abundant housekeeping proteins), PDGF receptor α (PDGFRα), bFGF, and the bFGF receptor (bFGFR) in oocytes of high versus low potential for development as previously determined (3). The second objective was to compare the relative abundance of L30, PDGFRα, and bFGF transcripts in 16-cell embryos originating from oocytes of high versus low developmental potential to determine if oocyte quality could affect the prevalence of these transcripts after activation of the embryonic genome.

Ovaries from cows and heifers were collected at an abattoir, about 15 min after slaughter. The ovaries were immediately placed in saline solution (35°C) supplemented with 100 U/mL penicillin G, 100 μg/mL streptomycin sulfate, and 250 ng/mL amphotericin B (Sigma-Aldrich Canada, Oakville, Ontario). Cumulus-oocyte complexes were recovered from the ovaries by aspiration of follicles (> 2 mm) with an 18-gauge needle attached to a 10-mL syringe. The follicular fluid was collected in 50-mL conical tubes, spread in petri dishes, and searched with a stereomicroscope. The COC/oocytes were recovered and washed twice in tissue culture medium 199 (TCM-199) supplemented with 20 mM Hepes (N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid), 0.4% BSA, 0.2 mM sodium pyruvate (all from Sigma-Aldrich Canada), and 100 U/mL penicillin, 100 μg/mL streptomycin (Gibco-BRL, Burlington, Ontario). The COC were washed a third time in maturation medium and transferred to 4-well plates containing 0.5 mL of maturation medium per well. The maturation medium consisted of TCM-199 supplemented with 10% (v/v) fetal bovine serum (Characterized FBS; Hy-Clone Laboratories Inc., Logan, Utah, USA), 0.2 mM sodium pyruvate, 35 μg/mL of porcine FSH (Folltropin-V; Vetrepharm Inc., Belleville, Ontario), and penicillin/streptomycin, as above. The COC were classified according to the appearance of their cumulus and cytoplasm, as previously described (3). Briefly, class I COC had a compact cumulus of more than 5 layers of cells and an oocyte with homogeneous ooplasm; class II COC had either slight expansion of the cumulus (more than 5 layers of cells) or slight granulation in the ooplasm; class III COC had a cumulus of fewer than 5 layers of cells (without or with slight expansion) and an oocyte with homogeneous ooplasm; class IV COC had fewer than 5 layers of cumulus cells with granulation in the oocyte; class V oocytes had no cumulus cells; and class VI COC had full expansion of the cumulus with dark clumps and heavy granulation in the oocyte. Oocyte maturation and subsequent cultures were performed at 39°C in an atmosphere of 5% CO₂ in air and saturated humidity. Mature oocytes (after 22 h of culture) were either frozen for subsequent RNA extraction or fertilized to produce 16-cell embryos.

For the production of embryos, mature oocytes were washed twice in warm Hepes-buffered Tyrode's albumin lactate pyruvate solution (TALP, 11) and washed once in fertilization medium (a bicarbonate-buffered modified TALP, 12), and then placed in 50 μL droplets (10-12 COC/droplet) of fertilization medium containing 10 μg/mL of heparin, 0.5 mM penicillamine, 0.25 mM hypotaurine, and 0.25 mM epinephrine (all from Sigma-Aldrich Canada). Frozen bull semen (Western Breeders Service, Balzac, Alberta) was thawed and prepared by a swim-up procedure (12). Spermatozoa were added to the fertilization drops at a concentration of 2 × 10⁶/mL.

Twenty hours after insemination, cumulus cells were removed from the presumptive zygotes by vortexing for 2 min in 500 μL of 300 μg/mL of hyaluronidase (Sigma-Aldrich Canada) in Hepes-buffered Tyrode's solution (TLH). The presumptive zygotes were washed twice in TLH and once in embryo culture medium. The embryo culture medium consisted of modified synthetic oviductal fluid medium (mSOF) supplemented with 0.5 mM citrate (sodium-citrate trisodium salt; Sigma-Aldrich Canada) and 1 × essential and non-essential amino acids (Gibco-BRL) (13). Twenty to 25 embryos were cultured in 50-μL drops under light mineral oil. The embryos were assessed for cleavage 48 h after the beginning of culture and 16-cell embryos were collected 90 to 114 h after fertilization.

Before freezing, oocytes were vortexed for 2 min in the presence of trypsin (Gibco-BRL, 1 mL of 1 mg/mL in Hank's balanced salt solution) in order to detach the cumulus cells. The denuded oocytes were then washed twice in TLH and transferred to a 500 μL tube. They were left to settle to the bottom of the tube for 1 min, as much medium as possible was removed, and the tube was put in liquid nitrogen for quick freezing. Sixteen-cell embryos were frozen in the same way, except that vortexing was not necessary.

For RNA extraction, oocytes of classes I to III were combined since their developmental potential was similar and higher than for the other classes (13.9, 13.7, and 12.7% blastocysts for classes I, II, and III, respectively) (3). Oocytes of classes IV to VI were also combined (representing oocytes with low developmental potential, with 3.5, 0.3, and 1.9% blastocysts for classes IV, V and VI, respectively) (3). Sixteen-cell embryos originating from classes I to III oocytes and from classes IV to VI oocytes were also combined; therefore, 2 populations of oocytes or embryos were compared. Using the Trizol reagent (Gibco-BRL), RNA was extracted from pools of 100 oocytes or 50–70 embryos. For each pool of oocytes or embryos to be extracted (4 pools were prepared for each stage with material from different collection dates), the following were combined: 300 μL of Trizol, 20 μg of Escherichia coli ribosomal RNA (rRNA, in 5 μL H₂O; Roche Molecular Biochemicals, Laval, Quebec), and rabbit globin mRNA (0.125 pg/embryo; Roche Molecular Biochemicals). The rRNA served as a carrier for RNA precipitation and as a measurable marker to allow the estimation of recovery, while the globin mRNA served as a standard for the estimation of the relative abundance of transcripts. The rest of the procedure was performed as suggested by the manufacturer of Trizol; the RNA was precipitated with 95% ethanol at −20°C overnight or longer, centrifuged, washed, and resuspended in 11 μL of water. An aliquot (0.5 μL) was removed for estimation of recovery by spectrophotometry (on average 80%). The remainder was used for cDNA synthesis in a reaction containing 50 pmoles oligodT primers, 1 × reverse transcription buffer, 10 mM DTT, 1 mM dNTP mix, 20 U RNAse inhibitor, and 15 U of Thermoscript reverse transcriptase (all from Gibco-BRL) for 1 h at 55°C. The reverse transcriptase was inactivated at 85°C for 5 min, 100 U RNAse H were added, and the reaction was incubated at 37°C for 20 min, then heated (70°C) for 10 min.

The procedure for semi-quantitative reverse transcription-polymerase chain reaction (sqRT-PCR) was adapted from Temeles et al (14). For each set of primers (Table I), oviduct cell RNA (in amounts equivalent to the amounts of oocyte and embryo RNA utilized) was used to determine the amount of PCR product generated as a function of cycle number. This, in turn, was used to determine the number of cycles for subsequent reactions, so that the amount of product generated could be visible on agarose gel, but not past the exponential stage of PCR amplification (data not shown). To ensure maximal consistency for each experiment, RNA from each of the classes to be compared was isolated concurrently and the RT-PCR was also performed at the same time; however, the different replicates originate from different pools of RNA. In preparation for PCR, a master mix containing all of the reaction components except the template and primers was prepared. This master mix was then divided into small aliquots; the globin primers were added to one aliquot and each of the other aliquots received a primer pair specific for one of the genes to be studied. These smaller master mixes were also divided into aliquots that received oocyte or embryo cDNA of each class. For α-globin, cDNA from 2.5 embryo-equivalent was used; for β-actin and L30, cDNA from 5 embryo-equivalent was used; and for the other transcripts, cDNA from 10 embryo-equivalent was used. Each reaction contained 1 × PCR buffer (including 1.5 mM magnesium chloride), 200 μM each of the 4 dNTPs, 0.6 unit Takara Taq (Panvera Corporation, Madison, Wisconsin, USA), 20 μM each of the appropriate 3' and 5' primers in 25 μL. The PCR conditions were as follows: an initial denaturation step at 94°C for 4 min; then 1 min of denaturation at 94°C; 1 min of annealing at 55°C; 1 min of extension at 72°C for 30, 35, 25; and 38 cycles for α-globin, β-actin, L30, and for the other transcripts, respectively. Final extension was at 72°C for 10 min. Five μL of gel loading buffer was added and 22 μL of each sample was loaded on 1.8% agarose gels. The gels were stained and photographed under ultraviolet (UV) illumination. Photographs were scanned and the intensity of each band was quantified with software (Quantity One; Bio-Rad Laboratories Ltd., Mississauga, Ontario). For each sample, the ratio of intensity of the band representing the RNA of interest over the intensity of the corresponding globin band (the standard) was calculated and called ratio'; then for each replicate, ratio' I-III/ratio' IV-VI was calculated and this value was called ratio''. Results are expressed as mean ± standard error (SE) of ratio'' for each of the transcripts. For each transcript, ratio'' was expected to be equal to 1, if the relative abundance in class I-III and class IV-VI was equal. For statistical analysis, the log of each ratio' was taken and a paired t-test (16) was performed on the transformed data using statistical analysis software (SAS; SAS Institute Inc., Cary, North Carolina, USA) to compare ratio' I-III versus ratio' IV-VI for each replicate for each transcript. The intra-assay variation was estimated from several simultaneous PCR amplifications of globin cDNA from each of the pools to detect differences caused by the PCR reaction and did not exceed 30% of the mean (data not shown). The inter-assay variation estimated from non-simultaneous amplification of different globin templates ranged from 30 to 70% (data not shown); therefore, only transcript ratios between samples processed simultaneously were compared.

Table I.

For PDGFRα and bFGFR primers, preliminary experiments showed that consistent generation of a visible amplification product was dependent on the presence of cDNA from more than 3 oocytes (data not shown), therefore, it is unlikely that the PCR products obtained with these primers resulted from cross amplification from the globin mRNA or the E. coli rRNA. Similarly, a minimum amount of globin mRNA per embryo in the extraction mixture was necessary for successful amplification, making it unlikely that the PCR product resulted from cross amplification from the E. coli rRNA. The primers for β-actin, PDGFRα, and bFGF were previously used in the presence of large amounts of E. coli rRNA and the PCR products were shown to originate from the respective transcripts (8).

Ethidium bromide-stained agarose gels of products obtained after amplification, with different sets of primers, of cDNA from mature oocytes from classes I-III and IV-VI, and 16-cell embryos originating from oocytes of these classes, are shown in Figures 1 and 2, respectively. The ratio' for each replicate for each transcript for each class are shown in Table II. For each of the transcripts examined, average ratio'', which is the average ratio of the relative abundance in class I-III oocytes (or 16-cell embryos from these oocytes) over the relative abundance in class IV-VI oocytes (or 16-cell embryos from these oocytes), was not significantly different from 1 (Table II). Transcripts for bFGFR were not detected in 16-cell embryos.

Figure 1. Representative reverse transcription polymerase chain reaction (RT-PCR) figures for the comparison of the relative abundance of the studied transcripts in oocytes with higher (I-III) versus lower (IV-VI) developmental potential. Transcript names are on the left and product sizes in base pairs are on the right. The L lane is the 100-base pair DNA marker lane.

L-30 — ribosomal protein L30

bFGF — basic fibroblast growth factor

PDGFRα — platelet derived growth factor receptor α

bFGFR — basic fibroblast growth factor receptor

Figure 2. Representative reverse transcription polymerase chain reaction (RT-PCR) figures for the comparison of the relative abundance of the studied transcripts in 16-cell embryos originating from oocytes with higher (I-III) versus lower (IV-VI) developmental potential. Transcript names are on the left and product sizes in base pairs are on the right. The L lane is the 100-base pair DNA marker lane.

L-30 — ribosomal protein L30

PDGFRα — platelet derived growth factor receptor α

bFGF — basic fibroblast growth factor

Table II.

For this study, we used an established method to compare the relative abundance of specific mRNA transcripts in mature oocytes of high and low developmental potential, and in 16-cell embryos originating from these oocytes. The 16-cell stage was chosen to represent a development stage during which the embryonic genome is active, and since activation of the embryonic genome occurs at the 8- to 16-cell stage in cattle oocytes (17), the transcripts present at the 16-cell stage could be of maternal or embryonic origin. Comparison of the relative abundance of transcripts in blastocysts would have allowed for a better determination of the effects of oocyte quality on embryonic transcription of the genes studied, but it was very difficult to obtain enough blastocysts from class IV-VI oocytes for experiments (3).

Oocyte quality (as assessed by visual appearance) did not have any significant effects on the relative abundance of 2 mRNA transcripts encoding housekeeping proteins (L30 and β-actin) and 3 mRNA transcripts of lower abundance (PDGFRα, bFGF, and bFGFR), suggesting that the lower developmental potential of oocytes of class IV-VI was not due to the absence of these transcripts, dramatically lower level of expression of these genes, or increased degradation of the transcripts. The relative abundance of the mRNA transcripts for L-30, PDGFRα, and bFGF was also similar in 16-cell embryos originating from oocytes of high versus low developmental potential; therefore, oocyte quality did not dramatically affect degradation of maternal transcripts for these proteins or embryonic expression of these genes. De Sousa et al (18) calculated the precision of the relative abundance values for the α1 subunit of the Na⁺ K⁺-ATPase to be sufficient to resolve 2- and 3-fold differences between individual blastocysts and oocytes respectively. Therefore, we can conservatively say that the sensitivity of the method would allow detection of major differences in gene expression.

In a previous study, we observed a trend for lower transcriptional activity (as measured by incorporation of ³H-uridine into RNA) in 16-cell stage embryos originating from oocytes of classes IV and V-VI, compared to embryos originating from classes I-II and III oocytes (3). It is possible that the activation of the embryonic genome was delayed in more of the embryos from class IV-VI oocytes compared to embryos from class I-III oocytes. The 16-cell embryos used in the previous and present studies are the ones who survived embryonic death at the earlier stages. It is, however, expected that fewer of the 16-cell embryos from class IV-VI would have reached the blastocyst stage if allowed to develop (3); therefore, it is highly probable that there would be differences in gene expression between the 2 populations of embryos. Our previous study (3), as well as the present study, indicate that oocyte quality would dramatically affect the level of expression of a small number of specific genes only, as observed for the composition of the maturation medium, which affected the relative abundance of some of the transcripts examined in bovine oocytes (19). Very few studies have examined the effect of oocyte quality on the expression of specific genes in bovine oocytes or embryos. In a recent study on the role of PGE₂ in cumulus expansion, there was no effect of COC quality grade on the expression of cyclooxygenase-2 and prostaglandin E₂ receptor 2 mRNA; however, higher expression of prostaglandin E₂ receptor 3 mRNA was detected in oocytes of higher quality grades (20). Perhaps small differences in relative abundance, qualified as non-significant by this method, are physiologically significant. In the future, real-time RT-PCR, with its increased sensitivity, will be the method of choice for comparing transcript abundance since reaction products for each sample are quantified in every cycle.

Transcripts for PDGFRα were previously shown to be present in mature bovine oocytes and embryos from the 2-cell to the blastocyst stage, while mRNA transcripts for bFGF were detected in mature oocytes and in early stages of preimplantation development (up to the 8-cell stage) but were not detected in 16-cell through to blastocyst-stage embryos (7). In the present study, transcripts for bFGF were detected in 16-cell embryos; this difference could be due to the different culture systems employed or to the improved efficiency of enzymes used for reverse transcription and PCR in the present study. Transcripts for bFGFR were not detected in 16-cell embryos, as previously observed by Yoshida et al (8).

In conclusion, lower developmental competence of oocytes with advanced signs of atresia was not due to dramatically lower levels of the transcripts for β-actin, L30, PDGFRα, bFGF, and bFGFR.