An improved method for specific-target preamplification PCR analysis of single blastocysts useful for embryo sexing and high-throughput gene expression analysis

Abstract

Gene expression analysis in preimplantation embryos has been used for answering fundamental questions related to development, prediction of pregnancy outcome, and other topics. Limited amounts of mRNA in preimplantation embryos hinders progress in studying the preimplantation embryo. Here, a method was developed involving direct synthesis and specific-target preamplification (STA) of cDNA for gene expression analysis in single blastocysts. Effective cell lysis and genomic DNA removal steps were incorporated into the method. In addition, conditions for real-time PCR of cDNA generated from these processes were improved. By using this system, reliable embryo sexing results based on expression of sex-chromosome linked genes was demonstrated. Calibration curve analysis of PCR results using the Fluidigm Biomark microfluidic platform (Fluidigm, South San Francisco, CA) was performed to evaluate 96 STA cDNA from single blastocysts. In total, 93.75% of the genes were validated. Robust amplification was detected even when STA cDNA from a single blastocyst was diluted 1,024-fold. Further analysis showed that within-assay variation increased when cycle threshold values exceeded 18. Overall, STA quantitative real-time PCR analysis was shown to be useful for analysis of gene expression of multiple specific targets in single blastocysts.

INTRODUCTION

Before implantation, the newly formed embryo must undergo several rounds of cleavage, activation of the embryonic genome, epigenetic remodeling, and a series of differentiation events leading to formation of the blastocyst-stage embryo possessing 2 distinct lineages represented by the inner cell mass and trophectoderm (Pfeffer, 2018; Rossant and Tam, 2018; Hansen and Tríbulo, 2019). Although lasting only a few days, the period leading up to blastocyst formation is critical for subsequent survival of the embryo (Pribenszky et al., 2017) and has been linked to postnatal function and health (Greenberg et al., 2017; Duranthon and Chavatte-Palmer, 2018; Hansen and Tríbulo, 2019). Studying gene expression in the embryo is an important tool to understand mechanisms responsible for successful preimplantation development. Additionally, assessments of transcript abundance may aid in understanding embryonic mortality and identification of molecular signatures for predicting pregnancy potential of individual embryos (El-Sayed et al., 2006; Jones et al., 2008; Zolini et al., 2020a; Zolini et al., 2020b). RNA sequencing (RNA-seq) analysis of biopsies from human embryos has also been used to identify dysregulated gene expression networks related to aneuploidy and poor morphology and morphokinetic aspects of development (Groff et al., 2019).

Recent developments in small input RNA-seq has allowed analysis of gene expression at the single embryo and even single blastomere level (Chitwood et al., 2013; Xue et al., 2013; Yan et al., 2013; Lavagi et al., 2018). Such a technique usually requires several rounds of whole transcriptome amplification to meet the cDNA input requirement for next generation sequencing. Concerns have arisen regarding the biased amplification of the transcriptome (van Dijk et al., 2014; Parekh et al., 2016; Kroneis et al., 2017; Luo and Zhang, 2018; Wright et al., 2019). An alternative method for analyzing gene expression at the level of the single embryo or single blastomere is quantitative real-time PCR, particularly if the goal is not to profile global gene expression but to focus on specific genes or sets of genes. In addition to its low cost, quantitative real-time PCR has higher sensitivity and lower variability than RNA-seq (Kroneis et al., 2017; Kolodziejczyk and Lonnberg, 2018). Furthermore, this technique can be used in conjunction with microfluidic technology to analyze expression of up to 96 genes in 96 samples by using the Biomark microfluidic platform developed by Fluidigm (Negron-Perez et al., 2017; Kolodziejczyk and Lonnberg, 2018).

Analysis of gene expression from individual embryos using PCR is complicated by the low amount of RNA available. A strategy of direct cDNA synthesis combined with specific-target preamplification (STA) to overcome the issue of small RNA input and avoid bias generated from whole transcriptome amplification has been used for analysis of individual blastomeres of embryos (Nestorov et al., 2015; Negron-Perez et al., 2017; Wei et al., 2017). Not considered in the earlier experiments were artifacts caused by DNA contamination, as well as possible bias caused by between-embryo variation in the effectiveness of the lysis technique.

The goal of the current study was to modify procedures for (STA) PCR of individual blastocysts by establishing complete lysis of every blastocyst, ensuring removal of DNA, and optimizing conditions for real-time quantitative PCR. As used previously (van der Weijden et al., 2017), the method uses multiplex STA incorporated into the cDNA synthesis step to minimize bias associated with amplification preference. Moreover, the method does not require RNA extraction. Modifications to the earlier method include visual confirmation of embryo lysis, addition of DNase to remove all residual DNA before amplification and decreased cycle number for real-time PCR. Here we show that the method can be used to sex blastocysts based on expression of sex-chromosome linked genes and can be incorporated into the high-throughput, quantitative real-time PCR method using the 96 × 96 integrated fluidic circuit (IFC) in the Biomark microfluidic platform. This method for single-blastocyst gene expression analysis provides an improved and useful option for assessing gene expression of hundreds of genes of interest.

MATERIALS AND METHODS

Embryo Production

Bovine embryos were produced as previously described (Tríbulo et al., 2018, 2019; Zolini et al., 2019) with minor modifications. Briefly, groups of 10 cumulus-oocyte complexes (COC) were collected from ovaries that were harvested from an abattoir. The COC were cultured in 50-μL microdrops of maturation medium (BO-IVM, IVF Bioscience, Falmouth, UK) covered with mineral oil (MOFA Global, Verona, WI) for 22 h at 38.5°C and 5% (vol/vol) CO₂ in a humidified air atmosphere. The COC were rinsed in HEPES-Tyrode’s albumin lactate pyruvate (TALP) medium and fertilized for 16 to 18 h with a pool of Puresperm (Nidacon, Mölndal, Sweden) purified sperm from 3 different bulls in in vitro fertilization-TALP medium. After fertilization, presumptive zygotes were treated with hyaluronidase to remove cumulus cells, rinsed in HEPES-TALP, and cultured in groups of up to 30 in 50-μL microdrops of synthetic oviduct fluid–bovine embryo 2 covered with mineral oil. Embryos were cultured at 38.5°C in an atmosphere of 5% (vol/vol) oxygen, 5% CO ₂ in a humidified atmosphere with the balance being N₂. Blastocyst-stage embryos were harvested for analysis at d 7 after fertilization.

Zona Pellucida Removal and Collection of Blastocysts into Modified Capillary Tubes

Each blastocyst was rinsed in diethyl pyrocarbonate-treated Dulbecco’s PBS that contained 0.2% (wt/vol) polyvinylpyrrolidone (DPBS-PVP) and then incubated in acid Tyrode’s solution (Sigma-Aldrich, St. Louis, MO) for 0.5 to 2 min until the zona pellucida was dissolved (https://dx.doi.org/10.17504/protocols.io.6kwhcxe). The embryo was then rinsed in DPBS-PVP again. For hatched blastocysts only, an additional step was performed consisting of 10-min exposure of the embryo to undiluted 10× TrypLE cell dissociation reagent (Thermo Fisher, Waltham, MA). The additional step was included to ensure complete lysis for blastocysts at this stage only.

Blastocysts were placed individually into 200-μL nuclease-free PCR tubes (Bio-Rad, Hercules, CA) containing 1 μL of the resuspension buffer provided in the CellsDirect One-Step qRT-PCR Kit (Thermo Fisher). To minimize dilution of the buffer, embryos were transferred into the tube in a minimum volume using a modified capillary tube (Drummond Scientific, Broomall, PA; Figure 1B–D). The tip of the capillary tube was modified so that the outer diameter was ~300 μm. This was achieved by placing the tube over a flame, pulling the tube and cutting the end off with a mini grinding disc. After transfer of the embryo, the collection tube was placed in liquid nitrogen for snap freezing and then stored at −80°C.

Figure 1.

Overview of procedures for specific-target preamplification PCR analysis of RNA from single blastocysts. (A) Workflow of the procedure highlighting composition of the reaction mixture at specific steps in the process. ZP = zona pellucida; gDNA = genomic DNA. (B–D) Modification of a capillary pipette for use in handling embryos while minimizing dilution of the cell lysis reagent. Shown in panel B is a Wiretrol (Drummond Scientific, Broomall, PA) plunger (left) and a Wiretrol micropipette modified by heating, pulling and cutting of the end of the tube. A detail of the tip of the modified capillary pipette is shown in panel C, and under the stereoscope adjacent to a blastocyst in panel D.

One-Step cDNA Synthesis and Preamplification

The workflow for synthesis of cDNA from a single embryo is shown in Figure 1A. After thawing the blastocyst, lysis was achieved by addition of 0.5 μL of lysis enhancer provided in the CellsDirect One-Step qRT-PCR Kit and digestion at 70°C for 20 min; tubes were tapped once during digestion. Each embryo was then examined using an Olympus model SZX12 stereomicroscope (Tokyo, Japan) while in the tube. If the embryo was incompletely solubilized, incubation at 70°C proceeded for an additional 5 to 10 min until the embryo was completely solubilized. Subsequently. 0.5 μL of 1 U/μL DNase I and 0.22 μL of DNase buffer (provided in the kit) were added to remove DNA from the sample. Incubation proceeded at 25°C for 15 min. Activity of DNase I was terminated by addition of 0.55 μL of 25 mM EDTA and incubation at 70°C for 10 min.

The STA mix contained 5 μL of CellsDirect 2× reaction mix (from the kit), 0.5 μL of SuperScript III RT/Platinum taq mix (from the kit), 1 μL of a mix of 500 nM of each primer, and 1 μL DNA suspension buffer (Teknova, Hollister, CA). A total of 7.5 μL of the STA mix was added to each lysed embryo, and cDNA synthesis and preamplification was carried out by running a program in a PCR machine (CFX-96, Bio-Rad, Richmond, CA) consisting of 50°C for 20 min; 95°C for 2 min; 18 cycles of 95°C for 15 s, and 60°C for 4 min. The sample was then subjected to Exonuclease I treatment to remove primers by digesting with 0.8 μL of 20 U/μL Exonuclease I, 0.4 μL of buffer (New England BioLabs, Ipswich, MA), and 2.8 μL of nuclease-free water at 37°C for 30 min. The Exonuclease I was then deactivated at 80°C for 15 min. Nuclease-free water or DNA suspension buffer was added to achieve a 5-fold dilution of the cDNA.

Real-time PCR

For PCR, 20 μL of real-time PCR master mix containing 10.4 μL EvaGreen (Bio-Rad), 1.26 μL of 10 μM forward and reverse primer mix, and 7.1 μL of water were mixed with 1.2 μL of cDNA or genomic (g)DNA. Denaturation was performed at 95°C for 30 s, followed by 30 or 40 cycles of 95°C for 5 s and 60°C for 5 s in a real-time PCR machine (CFX-96, Bio-Rad). Melt curve analysis was performed from 65 to 95°C in increments of 0.5°C per 5 s.

Efficacy of DNase I Treatment

Zona-free single embryos were lysed as described above. Total RNA was removed by treatment with RNase A (Qiagen, Germantown, MD) at 37°C for 30 min. Samples were treated with or without 0.5 U DNase I (from the kit) for 5, 10, or 15 min and then, to assess amounts of residual DNA in the sample, 40 cycles of real-time PCR were performed for YJU2 using primers that did not span an intron (Table 1).

Table 1.

Primers that were used for sexing embryos

Gene/region	Genbank ID	Forward (5′–3′)	Reverse (5′–3′)	Size (bp)
Autosomal	NC_037346.1	TGGAAGCAAAGAACCCCGCT	TCGTGAGAAACCGCACCCTG	2171
DDX3Y	NM_001172595.1	GGGCGCTATATACCACCTCA	TCCACCCTGAACTGCCTTTA	76
GAPDH	NM_001034034.2	GGGTCTTCACTACCATGGAGAA	GTTCACGCCCATCACAAACA	110
H2AFZ	NM_174809.2	AGCCATCCTGGAGTACCTCA	GTGACGAGGGGTAATACGCTTTA	91
XIST	XR_001495594.2	GATCAACATGCCTGCAATGCTA	GGCAGACATGGAAGAGGGTAA	83
Y-linked	NC_016145.1	GATCACTATACATACACCACT	AAGGCTATGCTACACAAATTCTG	143
YJU2	NM_001075691.1	TTCGATAAGCACCAAGCATAGTCA	TCCGACATCTTCGCCTCCTA	68

¹Other potential amplicons at 1,623 and 3,029 bp; their yield was limited by controlling the extension duration in PCR cycles.

Embryo Sexing

Twenty-four expanded blastocysts were evenly divided into 2 halves using a microblade described elsewhere attached to a micromanipulator (Narishige, Amityville, NY) as described elsewhere (Zolini et al., 2020a). Blastocysts were bisected so that each half contained inner cell mass and trophectoderm. One half was sexed by PCR using gDNA as template, whereas the other half was sexed based on transcript abundance of the sex-linked genes DDX3Y and XIST. Primers are described in Table 1.

The technique for sexing using gDNA as a template was performed as described previously (Park et al., 2001) with modifications. Briefly, dissected embryos were lysed as mentioned above and then treated with 1 μL of 100 μg/mL RNase (Qiagen) at 37°C for 0.5 h. For PCR, 0.2 μL of GoTaq hot start polymerase (Promega, Madison, WI) was used along with 4 μL of 5× Green GoTap Flexi Buffer (provided with the GoTaq kit), 1 μL of 10 mM dNTP, 4 μL of 25 mM MgCl₂, 0.4 μL of 10 μM primers for PCR, and 6.6 μL of nuclease-free water. The PCR was performed in 2 rounds. For the first round, a Y-linked region of DNA was amplified for 20 cycles of 95°C for 15 s, 58°C for 15 s, and 72°C for 15 s. For the second round, 1 μL of 4 μM autosomal primer was added, and an additional 17 cycles of PCR were performed under the same condition as the first round. Products of PCR (8 μL) were mixed with 2 μL of Diamond Nucleic Acid Dye (Promega) and loaded on 2% (wt/vol) agarose gels for electrophoresis. Sex of the embryo was classified as male if the Y-linked amplicon and autosomal amplicon was identified and as female if the only amplicon present was the autosomal amplicon.

The gene expression-based sexing assay was performed using the STA procedure for individual embryos described above and with primers for DDX3Y and XIST as sex-dependent genes and with GAPDH and H2AFZ as housekeeping genes. Preamplified cDNA from 24 half embryos were analyzed in 4 runs of real-time PCR (3–10 per run). The threshold for calling an embryo a male was a fold-change value for DDX3Y relative to the geometric mean of housekeeping genes of 0.1. The threshold for calling an embryo a female was a fold-change value for XIST of 0.8. When making calls for sex, technicians were blinded as to the call for sex based on PCR of gDNA.

Fluidigm Real-Time Quantitative PCR Using cDNA Generated by STA from Individual Blastocysts

To test whether the cDNA generated from the STA procedure was compatible with a high-throughput real-time PCR platform, primer validation was performed for 96 primer pairs for use in the 96 × 96 Biomark microfluid chip (Fluidigm, South San Francisco, CA; Supplemental Table S1, http://dx.doi.org/10.17632/y7y2c3bnvb.1). A total of 92 primer pairs were designed and manufactured using the D3 Assay Design (Fluidigm) and 4 other pairs (DDX28, FOSB, PGAP3, and WBP1) were designed and manufactured by Integrated DNA Technologies (Coralville, IA) on the 96 × 96 Biomark microfluid chip. Primers were used for STA amplification of cDNA samples from 2 individual blastocysts. Amplified and Exonuclease I-treated cDNA was serially diluted 1:2, 1:4, 1:8, 1:16, 1:32, 1:64, 1:128, 1:256, 1:512, 1:1,024, 1:2,048, 1:4,096, 1:8,192, 1:16,384, 1:32,768, and 1:65,536. Each dilution point was analyzed in triplicate except for the 1:512 dilution sample that was run in duplicate for Embryo 2 due to space requirements on the chip. For Fluidigm real-time PCR, 5 μL of assay mix was loaded into left side frame of 96 × 96 IFC, one assay per well. The sample mix (5 μL) was loaded into the right-side frame of the IFC, with one sample per well. The assay mix contained 3 μL of assay loading reagent, 0.6 μL of 50 μM combined forward and reverse Deltagene primers, and 2.4 μL of DNA suspension buffer (Teknova). The sample mix contained 3 μL of 2× SsoFast EvaGreen Supermix with low ROX, 0.3 μL of 20× DNA binding dye, 1.35 μL of preamplified and Exonucleaase I-treated sample, and 1.35 μL of 1× DNA suspension buffer (Teknova). The quantitative PCR step involved 30 cycles of 95°C for 15 s and 60°C for 60 s, followed by a melting phase from 60 to 95°C. Calibration curves and cycle threshold (C_t) were generated with the Fluidigm Real-Time PCR Analysis software (Fluidigm). The criteria for primer validation was that amplification efficiency was between 90 and 110%, and the R² was greater than 0.98 (Taylor et al., 2010; Broeders et al., 2014). The C_t values were used only when melting analysis of the reaction indicated a single product was formed.

RESULTS AND DISCUSSION

Development of Optimum Conditions for Lysing Blastocyst-Stage Embryos

Biases in differences in gene expression between embryos could be induced by incomplete lysis of the embryo. In initial experiments with 6 single blastocysts, it was noted that simply adding the STA master mix to a single frozen embryo followed by repeated pipetting, as performed previously (van der Weijden et al., 2017), did not achieve lysis (Figure 2A and 2B). In most cases, complete lysis of normal and expanded blastocysts was achieved by addition of 0.5 μL of lysis enhancer provided in the CellsDirect kit followed by incubation at 70°C for 20 min (Figure 2C and 2D). An additional 10-min incubation was required for embryos that failed to be lysed within 20 min. The procedure did not always work for hatched blastocysts older than 7.5 d of development even if digestion continued for 1 h (results not shown). Lysis could be achieved for hatched blastocysts if the embryo was exposed to TrypLE cell dissociation reagent before embryo collection. To ensure that all embryos examined using the STA procedure were completely lysed before subsequent processing, lysis of each embryo was confirmed by visual observation of the embryo through the top of the PCR tube with the lid open and observation through the side of the tube.

Figure 2.

Demonstration of importance of lysis method and visual evaluation by stereomicroscopy to ensure complete lysis of a blastocyst. Zona-free blastocysts were stored in 1 μL of cell resuspension buffer at −80°C. Two procedures were used to lyse embryos (n = 3 blastocysts per procedure). In method 1, based on van der Weijden et al., (2017), 9 μL of specific-target preamplification (STA) mix was added to a single frozen blastocyst (A), immediately followed by vigorous pipetting up and down at least 10 times before examination under a stereomicroscope (B). Note that embryo lysis was not complete. In method 2, 0.5 μL of lysis enhancer provided in the CellsDirect One-Step qRT-PCR Kit (Thermo Fisher, Waltham, MA) was added to the embryo (C) and digestion proceeded at 70°C for 20 min with a tap-to-mix step at 10 min. After digestion, no embryo was observed (D). If embryo was not fully dissolved, as was found occasionally, additional incubation for 10 min at 70°C was required. Method 2 has been used for successfully digesting more than 200 embryos. Arrows indicate the presence of embryos. Bar = 0.5 mm.

Removal of gDNA

Contamination with gDNA leads to spurious values for transcript abundance in reverse-transcription PCR procedures. This problem can be avoided by design of primers so that they span an intron. However, this solution is not always practical, including for genes without introns or with short introns or where primer design is difficult. Accordingly, successful application of the STA procedure for a wide range of genes requires elimination of all gDNA. Initial experiments indicated presence of gDNA after the conversion of mRNA to cDNA. Accordingly, conditions to eliminate all gDNA were established. After RNase A treatment to remove RNA, the reaction mixture from single blastocysts was treated with or without DNase I for 5, 10, or 15 min at 25°C and then subjected to PCR using primers for YJU2. As shown in Figure 3, amplicons were generated for YJU2 after RNase treatment in samples not treated with DNase I. Treatment with DNase I for 5 or 10 min reduced the amount of amplicon generated by PCR but treatment with DNase I for 15 min was required to eliminate all signal from RNase-treated samples. Subsequent analyses incorporated the 15-min DNase treatment in the workflow.

Figure 3.

Importance of incorporation of a DNA removal step to eliminate contaminating DNA. Single blastocysts were lysed and treated with RNase. One set of embryos was subjected to DNase I treatment for 5, 10, or 15 min, and the other was treated by buffer and water (n = 2– 4 blastocysts per treatment). Real-time PCR was conducted using primers targeted a single exon of YJU2. Shown in left and right panels are the amplification curves and the melt curves of each reaction, respectively. Note that it was necessary to treat embryos with DNase for 15 min to ensure there was no signal from DNA. RFU = relative fluorescence units and T = time.

Reduced Real-Time PCR Cycles Improves Melt Curve Characteristics

To test whether the STA procedure could be used for gene expression analysis, we first performed 40 cycles of real-time PCR of ACPP and GAPDH using undiluted cDNA from 3 blastocysts that was preamplified for 18 cycles. The fluorescent signal steadily increased with each PCR cycle for both genes. However, melt curve analysis indicated multiple peaks for amplification of ACPP in 2 of 3 samples. Subsequently, real-time PCR was performed for ACPP, YJU2, HNF1B, and GAPDH using STA cDNA from 12 individual blastocysts. The multipeak issues occurred in melt curves of ACPP, YJU2, and HNF1B for some embryos. Multiple peaks in melt curves are indicative of multiple PCR products (Ririe et al., 1997; Ruiz-Villalba et al., 2017). The problem still existed even with diluted cDNA samples unless PCR cycle number was reduced to 30 (Figure 4). The multiple peaks issue was unlikely due to primer dimerization because such melt curves were not observed in no-template controls and were not seen for every embryo. Contamination of gDNA was probably not the cause because the YUJ2 primer did not span an intron and samples were treated with DNase I. Because the cDNA was subjected to 18 cycles of preamplification, the downstream real-time PCR may not tolerate large numbers of amplification cycles before artifacts are generated. Based on these results, the number of cycles of real-time PCR using STA cDNA was limited to 30 for other experiments and issues of multiple peaks were eliminated.

Figure 4.

Importance of reducing PCR cycles for real-time PCR of pre-amplified cDNA. Melting curve analysis was performed in 4 different genes that were amplified for 40 (left panels) or 30 (right panels) cycles of real-time PCR. The cDNA from 6 blastocysts were tested for 40 cycles of real-time PCR. The same 6 cDNA samples as well as cDNA from 6 additional blastocysts were tested for 30 cycles. All cDNA was diluted 1:10. Note that the existence of multiple peaks was limited to samples amplified for 40 cycles. Not shown are no-template controls which were always included and did not generate signals. RFU = relative fluorescence units and T = time.

Sexing of Single Embryos by Gene Expression Analysis using STA cDNA Is Accurate

Sex-biased gene expression emerges as early as the initiation of embryonic genome activation in male and female preimplantation embryos (Larson et al., 2001; Lowe et al., 2015; Gross et al., 2017). Male and female embryos also respond differently to environmental stimuli such as cell-signaling molecules and stress (Dobbs et al., 2014; Perez-Cerezales et al., 2018; Tríbulo et al., 2018). These sexual dimorphic responses alter early embryonic programming and lead to changes in prenatal and postnatal characteristics (Bermejo-Alvarez et al., 2011; Hansen et al., 2016). Therefore, transcriptional analysis on sexed individual embryos could be an invaluable tool to explore sex-dependent gene regulation of preimplantation development. We chose 2 sex-chromosome-linked genes DDX3Y and XIST for embryo sexing because they are differentially expressed in a sex-dependent manner (Siqueira and Hansen, 2016; Groff et al., 2019). The noncoding RNA XIST is responsible for triggering X chromosome inactivation in females to achieve dosage compensation of sex chromosomal genes (Loda and Heard, 2019), and DDX3Y encodes for a DEAD box RNA helicase enzyme that is involved in differentiation of male germ cells (Kotov et al., 2017). To assess the accuracy of gene expression-based sexing, we split 24 blastocysts into 2 halves (Figure 5A). One half of each blastocyst was sexed via gel electrophoresis of amplified DNA fragments (Figure 5B). For the other half, transcription of DDX3Y and XIST were evaluated by real-time PCR of preamplified cDNA (Figure 5C). Eleven (45.8%) half blastocysts were considered females based on low expression of DDX3Y and high expression of XIST (Figure 5C). The remaining 13 embryos (54.2%) were classified as male as determined by high expression of DDX3Y and low expression of XIST. Based on analysis of amplified DNA fragments and gel electrophoresis, the accuracy of classifying blastocysts based on sex using STA of DDX3Y and XIST was 100% (Figure 5D).

Figure 5.

A comparison of real-time PCR using specific-target preamplification cDNA and PCR using genomic DNA (gDNA) for determination of embryo sex. (A) Bisection of expanded blastocyst to produce 2 halves for determination of sex. (B) Gel electrophoresis of amplified gDNA of half blastocysts. The number above each lane is the embryo ID; NTC = no-template control. Also shown are samples of blood from a male (bull) and female (cow). Identification of the automosal amplicon (217 bp) is indicated by the white arrowhead and the Y-specific amplicon (143 bp) by the white arrow. (C) Transcriptional abundance of DDX3Y (Y-linked) and XIST (X-linked) in the other half blastocyst. Data are expressed as fold-change relative to the geometric mean of housekeeping genes (GAPDH and H2AFZ). (D) Comparison of designation of sex (M/F: male/female) using the 2 methods.

Gene expression-based embryo sexing was reported previously using extracted RNA (Hamilton et al., 2012). The RNA isolation may not be an ideal strategy for transcriptional analysis due to RNA loss or degradation during the process, as shown by RNA-seq analysis (Sultan et al., 2014). The STA procedure does not involve RNA extraction and can provide sufficient cDNA for analysis of multiple genes besides those used for sexing embryos.

Note that there was low expression of Y-linked DDX3Y in female embryos (Figure 5C). Such a phenomenon was observed elsewhere using real-time PCR with RNA extracted from pools of embryos (Siqueira and Hansen, 2016). Expression in females could be due to the cross-reaction between DDX3Y primers and transcripts for DDX3X-like (XM_015461814.1) from the X chromosome. Currently, the record of DDX3X-like is removed from the National Center for Biotechnology Information database due to standard genome annotation processing but nucleotide BLAST between DDX3Y and DDX3X-like showed that part of DDX3X-like transcript was identical to the full length of DDX3Y. Other Y-linked genes could be used for sexing instead of DDX3Y to avoid this minor issue.

Use of XIST expression for sexing should be useful for a wide range of species. It has been shown earlier that XIST transcription was active in both male and female human embryos from 4-cell to blastocyst stage (Daniels et al., 1997; Ray et al., 1997; Okamoto et al., 2011). An important question is how early in development differences in expression of sex-chromosome-linked genes can be used to sex embryos. Expression of XIST occurs before embryonic gene expression in the human (van den Berg et al., 2009) but not in the mouse (Sado and Sakaguchi, 2013). The XIST transcripts were not detected by RNA sequencing in male mouse embryos from the 4-cell stage through d 4.5 in peri-implantation blastocysts (Wang et al., 2016). In an early study, XIST was first detected in 8-cell embryos in the cow (De La Fuente et al., 1999) but XIST expression has also been identified in bovine individual oocytes and 2-cell and 4-cell embryos (Mendonca et al., 2019), indicating maternal and embryonic expression of XIST during bovine preimplantation development.

STA cDNA is Compatible with a High-Throughput Real-Time PCR Platform

The Fluidigm microfluidic dynamic array designed for real-time PCR provides a high-throughput platform for analyzing expression of 96 genes in 96 samples in a cost-effective manner (Spurgeon et al., 2008). To test performance of STA cDNA on this platform, we ran calibration curve analysis of 92 primers that were designed using the Delta Gene assay format (Fluidigm), and 4 were designed by Integrated DNA Technologies. The STA cDNA from 2 individual blastocysts were 2-fold serially diluted for 16 dilutions (from 2 to 65,536-fold) and run in triplicate. As expected, C_t gradually increased as dilution of cDNA increased (Figure 6A). Most reactions generated detectable amplification when cDNA was diluted up to at least 1,024-fold. The exception was for 2 genes with low transcript abundance. As dilution increased above 1,024-fold, more reactions showed negative results, even though amplification was detectable in some cases where samples were diluted 65,536-fold (Figure 6A). No amplification was detected for NEPRO and FOSB, indicating issues with primers of these genes. For the other primer pairs, amplification, melt and calibration curves were characterized as shown in Figure 6B. Calibration curve analysis showed that 93.75% (90/96) of primers were validated in both cDNA samples with 8 to 12 continuous dilution points (Figure 7A). Primers for UNK and FGF4 met the validation criteria for one embryo but not the other. Validated calibration curves of DDX3Y and WBP could not be generated (Figure 7B). Overall, these results suggest that STA cDNA is compatible for use with the Fluidigm real-time PCR platform.

Figure 6.

Validation of 96 primers on Fluidigm real-time PCR platform (Fluidigm Corp., South San Francisco, CA) using cDNA produced from specific-target preamplification method in single blastocysts. (A) Heatmap of cycle threshold (C_t) values of 96 genes tested with cDNA that was 2-fold-serially diluted from 1:2 to 1:65,536. Each dilution point was performed in triplicate except the 1:512 dilution that was run in duplicate for Embryo 2. Cells containing a × indicate amplifications associated with abnormal melting curves, such as curves with multiple peaks. NTC = no-template control. (B) Representative amplification, melting curves, and calibration curves. ΔRN = normalized reporter subtracted by baseline; T = time.

Figure 7.

Analysis of primer validation real-time PCR using Fluidigm real-time PCR platform (Fluidigm Corp., South San Francisco, CA). (A) Categories of primers based on validation results. The numbers within the pie graph represent numbers of genes in the category. The criteria for primer validation was that amplification efficiency was between 90 and 110% and R² was greater than 0.98. (B) Calibration curves of DDX3Y and WBP of Embryo 1. The dashed circles indicate high variability of triplicates. C_t = cycle threshold. (C) Variation of triplicates at different dilution folds. Data are shown as mean ± SD. (D) Dot plot of SD as a function of average C_t for data from Embryos 1 and 2 data at each dilution.

Representative examples of the effects of dilution on between-triplicate variation in C_t values is shown in Figure 7B and analysis of between-assay standard deviation for all triplicates at different dilution points and average Ct values are shown in Figures 7C and 7D, respectively. Between-triplicate variation increased when samples were diluted greater than 1,024-fold (Figure 7C) and the average C_t was greater than 18 (Figure 7D). These results suggested that STA cDNA should be diluted no more than 1,024-fold, and that accuracy of estimates of transcript abundance decline when C_t is higher than 18.

One key feature of the method described here is the multiplex STA step which converts mRNA into cDNA and preamplifies specific targets to generate sufficient template for high-throughput real-time PCR. The method is highly efficient based on the fact that it has been used on over 200 single blastocysts for conventional and high-throughput real-time PCR and no defective samples have been identified. Such a result suggests that total RNA was well preserved before the reverse-transcription. As compared with other systems that use RNA isolated from pool of embryos (Bermejo-Alvarez et al., 2009; Heras et al., 2016; Siqueira and Hansen, 2016; Tríbulo et al., 2018; Jannaman et al., 2020), the advantages of this system are elimination of variation due to differences in extraction efficiency between samples and the use of individual blastocysts instead of pools of blastocysts as the experimental unit in experiments. Depending on the experiment, use of STA can reduce the number of blastocysts required for gene expression analysis. This is an especially helpful characteristic of the system when analyzing embryos in species including goats, sheep, cows, horses; and humans where embryo availability can be limited.

One potential concern with STA is the possibility of amplification bias, when not all targets are amplified to the same extent. This can be a serious problem when global preamplification of mRNA is performed, such as for RNA-seq (van Dijk et al., 2014) and single-cell RNA-seq applications (Luo and Zhang, 2018). When used for quantitative PCR, target-specific preamplification was associated with greater reproducibility and sensitivity than global preamplification (Kroneis et al., 2017). There was no evidence of amplification bias when target-specific preamplification was used for reverse-transcription quantitative PCR with the Biomark platform (Sindelka et al., 2010). Similarly, variation in the reverse-transcription step was responsible for more variability in results than variation induced by preamplification in reverse-transcription quantitative PCR using the Biomark platform (Korenková et al., 2015).

Another key feature of the method described here is incorporation of procedures to ensure complete lysis of the embryo before further analysis. In initial experiments, embryos were found to be surprisingly resistant to lysis and it was found necessary to include a specific lysis enhancer step and visual verification of lysis to ensure complete lysis. Others have used target-specific preamplification of genes for multiple gene expression analysis on single cells or blastomeres (Citri et al., 2012; Moignard et al., 2013; Nestorov et al., 2015; Wei et al., 2017) but without verification of complete lysis. Finally, a third key feature was the necessity of incorporating a gDNA digestion step to avoid contamination with DNA.

A variety of methods have been developed to study gene expression at the single embryo and even single-cell method. Although RNA-seq can allow assessment of the entire transcriptome, it is more prone to amplification bias than quantitative PCR and is less sensitive and more variable than real-time quantitative PCR (van Dijk et al., 2014; Kroneis et al., 2017; Kolodziejczyk and Lonnberg, 2018; Luo and Zhang, 2018). Moreover, it is an expensive technique, so the number of individual blastocysts that can be analyzed in an experiment is often less than optimal. The protocol described here is useful for inexpensive and accurate assessments of quantification of specific transcripts for individual blastocysts. The fact that hundreds of genes can be measured from hundreds of blastocysts at a reasonable cost makes the technique valuable to determine how individual blastocysts respond to regulatory signals, actions of specific genes and groups of genes, and presence of cytotoxic molecules.

We also demonstrated here how the technique can be used to simultaneously determine embryo sex and gene expression without the need for separate extraction of RNA and DNA from the embryo. The importance of sex for function of the preimplantation embryo is well established (Hansen et al., 2016; Engel, 2018) but the characterization of which aspects of embryonic function depend upon sex and the mechanisms by which embryo sex regulates development during the preimplantation period are incompletely understood.

CONCLUSIONS

The assay described here can be used to quantify mRNA for individual bovine blastocysts using conventional real-time PCR and with a high-throughput, real-time quantitative PCR platform. Accurate results could be obtained when STA cDNA from single blastocysts was diluted up to 1,024-fold. The procedure was found to be useful for embryo sexing and can be useful for single-blastocyst gene expression analysis for up to hundreds of targets. A detailed, step-by-step protocol for use of the STA procedure is included in Supplemental File S1 (http://dx.doi.org/10.17632/y7y2c3bnvb.1).