Monitoring gene expression in a single Xenopus oocyte using multiple cytoplasmic collections and quantitative RT-PCR Alexander A. Tokmakov1, Takanori Hashimoto2, Yushi Hasegawa2, Sho Iguchi2, Tetsushi Iwasaki1,2 and Yasuo Fukami1,2 1 Research Center for Environmental Genomics, Kobe University, Japan 2 Graduate School of Science, Kobe University, Japan
Keywords gene expression variation; mRNA degradation; quantitative RT-PCR; single-cell gene expression analysis; Xenopus oocytes Correspondence A. A. Tokmakov, Research Center for Environmental Genomics, Kobe University, 1-1 Rokko dai, Nada, Kobe 657-8501, Japan Fax: +81 78 803 5951 Tel: +81 78 803 5953 E-mail: [email protected] (Received 28 March 2013, revised 10 October 2013, accepted 21 October 2013) doi:10.1111/febs.12576
Oocytes and eggs of the African clawed frog, Xenopus laevis, are commonly used in gene expression studies. However, monitoring transcript levels in the individual living oocytes remains challenging. To address this challenge, we used a technique based on multiple repeated collections of nanoliter volumes of cytoplasmic material from a single oocyte. Transcript quantification was performed by quantitative RT-PCR. The technique allowed monitoring of heterologous gene expression in a single oocyte without affecting its viability. We also used this approach to profile the expression of endogenous genes in living Xenopus oocytes. Although frog oocytes are traditionally viewed as a homogenous cell population, a significant degree of gene expression variation was observed among the individual oocytes. A lognormal distribution of transcript levels was revealed in the oocyte population. Finally, using this technique, we observed a dramatic decrease in the content of various cytoplasmic mRNAs in aging unfertilized eggs but not in oocytes, suggesting a link between mRNA degradation and egg apoptosis.
Introduction Transcript levels vary substantially among cells in seemingly homogenous populations [1,2], requiring gene expression analysis in individual cells. Recently, several techniques for transcript quantification and imaging in single cells have been developed [3–6]. The methods using fluorescent proteins and hybridization probes, such as MS2-based and molecular beaconbased approaches, hybridization probes and molecular beacons, label non-covalently specific mRNA molecules and allow real-time detection and measurement of transcript levels in single cells. However, these methods target only specific pre-selected mRNAs, and cannot be used for comprehensive transcriptome profiling. Alternatively, single-cell gene expression profiling using microarray and quantitative RT-PCR may be applied for the whole transcriptome. However, these methods cannot measure gene expression in living cells, as they are based on the extraction of total cellular mRNA.
The major difficulty of gene expression profiling in single cells is the need to measure small numbers of mRNA molecules. Microarray technology requires relatively large amounts of RNA for analysis and a preamplification step with broad applicability to many target genes. On the other hand, quantitative RT-PCR, which is currently the most sensitive and reproducible technique for gene expression quantification, has, in principle, the sensitivity to detect a single target mRNA molecule. A practical threshold of approximately 20 copies per reaction has been suggested [7,8]. Pre-amplification may also be required for this technique if a sample is split and multiple transcripts are quantified. However, the maintenance of bona fide gene expression profiles remains a bottleneck of pre-amplification procedures [9,10]. Thus, the existing data for transcriptome profiling have been obtained mostly on large cell populations, representing only mean mRNA levels.
Abbreviations Cq, quantification cycle; GV, germinal vesicle; GVBD, germinal vesicle breakdown; PG, progesterone; T7 RNAP, T7 RNA polymerase; RTPCR, reverse transcriptase PCR.
104
FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
A. A. Tokmakov et al.
Monitoring gene expression in a single oocyte
Oocytes and eggs of the African clawed frog, Xenopus laevis, have been commonly used in expression studies. These cells are a popular model because of their large size and the ease of obtaining them in large numbers. The oocytes are obtained freshly from living frogs and may be treated like a primary cell culture. The technique of oocyte microinjection enables easy delivery of nanoliter volumes of various materials. The oocytes are able to withstand introduction of up to 50 nL of injected volume, and injections may be continued over several days [11]. Such microinjection experiments were described as early as 1971 [12,13]. In addition, nanoliter quantities of the cytoplasm may also be extracted from living frog oocytes and matured eggs. The standard equipment for oocyte microinjection may be used for cytoplasmic sampling with minimal adjustments. Cytoplasmic transfers were originally used by Masui and Markert to reveal maturation-promoting factor in hormone-treated Rana pipiens oocytes [14]. Localized sampling of the Xenopus oocyte cytoplasm has been used to perform capillary electrophoresis [15] and measurements of kinase activation [16] and inositol triphosphate gradients [17]. In the present study, we investigated the feasibility of gene expression monitoring in single Xenopus oocytes and eggs using multiple cytoplasmic collections and quantitative RT-PCR.
Results Multiple cytoplasmic collections from Xenopus oocytes Fully grown Xenopus oocytes at stage VI are large cells of approximately 1.2 mm in diameter arrested in prophase of the first meiotic division. They are visibly A
B
divided into the dark-pigmented animal hemisphere and the light-colored vegetal hemisphere (Fig. 1A,B). The oocyte nucleus, called the germinal vesicle, is located in the animal hemisphere, occupying approximately one quarter of the total oocyte volume. Cytoplasmic samplings for gene expression analysis must be performed from the vegetal near-equatorial area of oocytes in order not to hit the nucleus. The needles for cytoplasmic collection have a wider tip opening than those for microinjection (Fig. 1C). It is important to make the collections from the same area of the cytoplasm to minimize the measurement variation due to sampling. Indeed, samples collected from the equatorial area of oocytes displayed much lower variation in the content of extracted RNA than those collected randomly (Fig. S1; see below also). In our experiments, the volume of a single cytoplasmic collection was set to 10 nL. Given that the estimated cytoplasmic volume of one oocyte is approximately 700 nL, and its total RNA content exceeds 5 lg [18], it may be calculated that a single cytoplasmic collection obtains approximately 70 ng of total RNA, and the total amount of extracted RNA from a single cytoplasmic collection fits this estimate (Fig. S1). Taking into account that at least ten target molecules per reaction are necessary for accurate quantification by quantitative PCR [7], this approach should allow detection of specific transcripts that are present at the level of more than 103 copies per oocyte. Oocyte viability after multiple cytoplasmic collections Cytoplasmic samplings inflict damage to oocytes, therefore, for practical application of the proposed C
Fig. 1. Cytoplasmic collections and microinjections. (A,B) Schematic representation of a Xenopus oocyte (A) and its microscopic image with the inserted sampling capillary during collection of cytoplasmic material (B). The large oocyte nucleus (germinal vesicle, GV) is located in the animal hemisphere. (C) Photograph of microcapillary tips used for cytoplasmic collections (c) and microinjections (i).
FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
105
A. A. Tokmakov et al.
Monitoring gene expression in a single oocyte
technique, it is important to evaluate their damaging effect. We investigated the impact of multiple cytoplasmic collections on Xenopus oocyte viability. It was found that the oocytes are able to withstand well up to eight cytoplasmic collections (Fig. 2A). After four collections, 80% of oocytes remained intact for 36 h, and approximately 50% were still morphologically normal by 48 h. Oocytes deteriorated significantly after 8 cytoplasmic collections. Still, more than half of them survived this procedure by 24 h, and 30% of the oocytes remained intact by 48 h (Fig. 2A). Viability of control oocytes exceeded 90% at that time. Notably, oocyte viability after a single 80 nL collection, i.e. the volume of eight 10 nL cytoplasmic collections, was practically the same as the viability of oocytes after a single 10 nL collection. These data suggest that it is the number of collections rather than the removal of certain cytoplasmic volume (below 80 nL) that causes cell death. In other words, oocytes are able to withstand the loss of approximately 10% of their cytoplasm with little effect on their viability. Generally, two outcomes were observed in damaged oocytes. Most often, they became susceptible to leakage, rapidly lost a significant amount of cytoplasmic material, shrank and became discolored. Alternatively, damaged non-leaky oocytes developed mottling in the animal hemisphere (Fig. 2B). This morphology has previously been associated with oocyte death [19]. In addition, these oocytes were functionally compromised, as they did not respond to the maturation hormone progesterone (Fig. 2B). Based on these results,
A
gene expression in further experiments was monitored mainly over 24 h, the period over which multiple cytoplasmic samplings exerted minimal effect on oocyte viability. Monitoring luciferase expression in a single oocyte Heterologous genes are commonly expressed in Xenopus oocytes to investigate their functions. In order to evaluate the applicability of multiple cytoplasmic collections for monitoring heterologous gene expression in a single oocyte, a T7 promoter-driven luciferase-encoding plasmid was microinjected into the oocyte cytoplasm together with T7 RNA polymerase (T7 RNAP) protein. It was shown previously that efficient coupling of T7 RNAP-mediated transcription with the intrinsic oocyte translation machinery occurs in the oocyte cytoplasm [20–22]. Robust production of luciferase mRNA was detected in the cytoplasm immediately after microinjection (Fig. 3A). Synthesis of luciferase mRNA continued over several hours, producing, according to calibration curve-based calculations, hundreds of mRNA molecules per single molecule of injected DNA. The level of luciferase mRNA decreased to some extent by 24 h (Fig. 3B); however, luciferase protein was still being produced at that time (Fig. 3D). In ovo synthesis of luciferase protein continued for many hours and was still evident at 24 h after DNA/T7 RNAP injection (Fig. 3C,D). These findings are largely consistent with previously published data from bulk-scale experiments
B
Fig. 2. Viability of Xenopus oocytes after multiple cytoplasmic collections. (A) Oocyte viability after zero, one, two, four and eight cytoplasmic collections at the indicated times after the start of sampling (0–48 h). The collections were performed at 1 h intervals. Ten to thirty oocytes were monitored for each number of cytoplasmic collections. The columns marked (1*) refer to oocyte viability after a single 80 nL cytoplasmic collection. (B) Images of normal and damaged immature oocytes, and their response to progesterone treatment, as determined by the occurrence of germinal vesicle breakdown (GVBD).
106
FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
A. A. Tokmakov et al.
Monitoring gene expression in a single oocyte
A
B
C
D
Fig. 3. Monitoring luciferase expression in a single Xenopus oocyte. (A–D) Kinetics of luciferase mRNA (A,B) and protein (C,D) synthesis in Xenopus oocytes after the cytoplasmic coinjection of a T7 promoter-driven luciferase-encoding plasmid and T7 RNAP protein. Four oocytes from the same batch were monitored in the experiment; however, the representative results of only one single-cell analysis are shown. Bars in (C) and (D) represent SDs of three independent measurements of luciferase activity. The SDs of replicate measurements in (A) and (B) were very small and are not visible in the plots.
[22]. It may therefore be concluded that the applied technique of multiple cytoplasmic collections allows effective monitoring of heterologous gene expression in a single Xenopus oocyte. Gene expression profiling in single Xenopus oocytes The filly grown Xenopus oocytes of stage VI contain a large quantity of maternal mRNA that is synthesized and accumulated during oogenesis. We investigated whether it is possible to detect various endogenous transcripts in cytoplasmic samples taken from a single oocyte. The task of gene expression profiling in the nanoliter samples is challenging because of the low abundance of many cellular transcripts. Indeed, expression profiling of multiple genes classified as various functional types, such as housekeeping, regulatory, apoptosis-related and metabolic genes, revealed a great variation in their expression levels (Fig. 4). Although some transcripts were present at high copy numbers, up to 106 copies per sample, FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
only a few copies of other transcripts were present. Gene expression profiling in different oocytes revealed a great variation of transcript levels. As an example, the expression profiles of two oocytes from the same batch are shown in Fig. 4A,B. Most genes in oocyte 1 were expressed at higher levels than in oocyte 2, so these cells may provisionally be described as ‘highexpressing’ and ‘low-expressing’ oocytes. However, the magnitude of difference in the expression levels of different genes was not the same. Some genes were expressed at similar levels in both oocytes, whereas others displayed an approximately 100-fold difference in expression levels. Notably, both ‘high-expressing’ and ‘low-expressing’ oocytes were viable and responded normally to progesterone treatment. These data demonstrate that quantification of lowabundance transcripts in nanoliter samples of the oocyte cytoplasm nears the detection limit of the quantitative RT-PCR technique. As estimated above, this approach allows reliable quantification of specific mRNAs present at the level of more than 103 copies per oocyte. 107
A. A. Tokmakov et al.
Monitoring gene expression in a single oocyte
A
A
B
B
Measurement and sampling variations of the analysis
Fig. 5. Variations of the expression analysis. (A) Estimation of measurement and sampling variations. The dashed line indicates the maximal level of the SD for Cq values determined for the studied transcripts in four replicate measurements. Columns represent SDs of Cq values determined for eight cytoplasmic collections taken from the same oocyte. The samples of extracted RNA were divided into aliquots before reverse transcription. (B) Cell-to-cell variation of gene expression between individual Xenopus oocytes. Columns represent SDs of Cq values determined for 16 oocytes of the same batch.
Experimental variation introduced by the applied technique may obscure gene expression quantification. It is therefore important to estimate its degree. In general, experimental variation of a method includes measurement variation and sampling variation. It has been reported previously that measurement variation of quantitative PCR, i.e. the variation of quantitative PCR measurements for the same template, is typically very low [10]. In our hands, the standard deviation of quantification cycle (Cq) values determined for eight studied transcripts in four replicate measurements was less than 0.25 PCR cycles (Fig. 5A), and it was geneindependent. Given that the above estimation covers both RT and PCR, the low measurement variation of this analysis is directly related to high reproducibility of the first-strand cDNA synthesis and quantitative PCR technique. Under our conditions, the low measurement variation relies on the high efficiency of reverse transcription and PCR due to the low amount of starting material (see Discussion). In a spike control experiment, the isolated oocyte RNA obtained from 10 nL cytoplasmic collection had no inhibitory effect on a control PCR involving a heterologous transcript
(Fig. S2A). As a result, the efficiency of PCR reactions for various analyzed gene products was > 95%. Sampling variation of the performed gene expression analysis includes the variation of cytoplasmic collections, variation of mRNA isolation and variation of first-strand cDNA synthesis. This was significantly higher than measurement variation, with a standard deviation (SD) of 0.5–0.7 PCR cycles, as determined for eight cytoplasmic collections taken from the same oocyte (Fig. 5A). Sampling variation was also geneindependent. These data indicate that sampling variation represents the major source of experimental variation in the proposed approach. To further understand the source of high sampling variation, we evaluated the reproducibility of sample collection and RNA isolation. It was found that the SD of total contents of isolated RNA in the aliquots of a split sample was relatively low at approximately 10% of its mean value, whereas a much higher SD (approximately 40%) was obtained for samples independently collected from the same oocyte (Fig. S1). These data suggest that variability of the collection volume is the main source of sampling variation.
Fig. 4. Expression profiling in Xenopus oocytes. (A,B) Gene expression profiles of two oocytes from the same batch. Gene expression levels were measured in two to four replicates. SDs of replicate measurements were very small and are not visible in the plots.
108
FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
A. A. Tokmakov et al.
Plausible explanations for the observed high sampling variation are discussed below. Variability of gene expression in individual oocytes We next evaluated biological cell-to-cell variation of gene expression between individual Xenopus oocytes. The variability of transcript levels in these cells has not previously been quantitatively estimated and remained undefined at this stage of the study. This kind of variation has important implications for the interpretation of gene expression data. We confirmed the different degree of gene expression variation for different genes by calculating SDs of Cq values for 16 oocytes of the same batch. In contrast to measurement and sampling variations, this parameter was genedependent and varied from 1.5 to approximately 3.0 PCR cycles (Fig. 5B). The lowest variation was observed for the highly expressed constitutive genes, such as actin, histH4 (histone H4) and gpdh (glyceraldehyde-3-phosphate dehydrogenase), validating their housekeeping designation, whereas apoptosis-related and regulatory genes showed much higher expression variations. To further examine gene expression variability, the distributions of Xenopus oocytes according to gene
Monitoring gene expression in a single oocyte
expression levels were plotted in logarithmic and linear scales for the actin and mcl1 genes (Fig. 6). The graphs revealed a lognormal distribution of transcript levels in the investigated oocyte population. In accordance with the data presented in Fig. 5B, the mcl1 gene had much greater expression variation than the actin gene, as indicated by the different widths of the approximated distribution curves (blue and red curves in Fig. 6). mRNA degradation in mature oocytes The technique of multiple cytoplasmic collections was also used for analysis of gene expression changes induced in Xenopus oocytes by the steroid hormone progesterone. The hormone promotes oocyte transition from prophase I to metaphase II, producing fertilization-competent metaphase-arrested eggs. It was found that progesterone treatment had little effect on mRNA levels in the oocytes within the first 12 h following hormone administration (Fig. 7A). However, a significant decrease in the content of all analyzed transcripts was observed after 24 h (Fig. 7B). By 36 h, only the most abundant transcripts were detected in the egg cytoplasm by the applied technique. A moderate decline in mRNA levels was also observed in progesterone-untreated Xenopus oocytes by that time (Fig. 7C). Notably, the timing of mRNA degradation coincided with the hallmark events of Xenopus egg apoptosis, such as cytochrome c release and caspase 3/7 activation (Table S1). By 48 h, no intact mRNA molecules were detected in the egg cytoplasm by the applied technique (data not shown). By that time, prominent egg swelling was observed, indicating disruption of osmotic homeostasis at a final stage of cell death (Table S1).
Discussion
Fig. 6. Variability of gene expression in Xenopus oocytes. Histograms of actin and mcl1 gene expression levels in the individual oocytes of the same batch plotted using logarithmic scale (main graph) and linear scale (inset). The distribution graphs (red and blue curves) were smoothed using the Excel chart smoothing algorithm.
FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
In the present work, we describe a simple and reliable technique for monitoring gene expression in a single living Xenopus oocyte. It is based on multiple collections of nanoliter amounts of the oocyte cytoplasm and quantitative RT-PCR. Our data demonstrate that the oocytes are able to withstand multiple cytoplasmic collections, with a high survival rate over 48 h (Fig. 2). This allows monitoring of gene expression dynamics in various experiments using a single oocyte. As a sample application of this technique, we monitored the expression of firefly luciferase gene in a single oocyte during 24 h (Fig. 3). To initiate the synthesis of both luciferase RNA and protein, a T7 promoter-driven luciferase-encoding plasmid was microinjected into the oocyte cytoplasm together with T7 RNAP protein. 109
A. A. Tokmakov et al.
Monitoring gene expression in a single oocyte
A
B
C
Fig. 7. mRNA degradation in progesterone-treated Xenopus oocytes. Gene expression profiling was performed in a single Xenopus oocyte at 12 h (A), 24 h (B) and 36 h (C) after progesterone addition. The specific transcript levels determined at the indicated times were normalized to those obtained in the same oocyte before hormone administration (dark gray bars in all panels). The light gray bars in (C) refer to the transcript levels in a progesterone-untreated control oocyte. Gene expression levels were measured in replicates. The SDs of two to four replicate measurements were very small and are not visible in the plots. Three progesterone-treated and three progesterone-untreated oocytes from the same batch were monitored in the experiment; however, the results of only one representative single-cell analysis are shown.
Synthesis of both luciferase mRNA and protein was observed in the microinjected oocyte for many hours after DNA/T7 RNAP introduction (Fig. 3). These results are largely consistent with the previously published data from bulk-scale experiments [22], providing a proof of concept for the suggested technique of gene expression monitoring. Next we applied this method to profile multiple endogenous transcripts in a single oocyte. The task of transcript profiling in the nanoliter samples of the oocyte cytoplasm is challenging because of the low abundance of many cellular transcripts. Indeed, the technique approaches the detection limit of quantitative RT-PCR for the low-abundance transcripts (Fig. 4). As a result, the low-copy number mRNAs, such as stat3 and pdk3, showed a higher variation of expression levels than high-copy number mRNAs (Fig. 5B). Notably, the small sampling volume of cytoplasmic collections is directly related to the high quality of PCR data. Previously, yolk platelets localized in the vegetal hemisphere of the oocyte have been reported to inhibit reverse transcription and PCR [23]. However, the spike control experiment did not reveal any inhibitory effects of the samples obtained by multiple cytoplasmic collections on the PCR-based quantification (Fig. S2A). Evidently, it is high dilution of the cytoplasmic inhibitors during RNA isolation from a small-volume sample that allows high PCR efficiency. However, high sampling variation remains a bottleneck of this approach, and appears to be the major source of experimental variation. As the standard deviation of sampling exceeds 0.5 PCR cycles (Fig. 5A), this approach is not applicable for detection of less than twofold changes in transcript levels. Instead, it 110
may be useful for monitoring high-magnitude expression of heterologous genes (Fig. 3). Also, it provides a means to simultaneously monitor significant multifold changes in the expression of endogenous Xenopus genes in a single oocyte. Indeed, using this technique, we observed a dramatic decrease in the content of various cytoplasmic mRNAs in aging unfertilized eggs (Fig. 7). The biological variation of gene expression in Xenopus oocytes varied from 1.5 to approximately 3.0 PCR cycles for various genes (Fig. 5B). It is still much larger than the technical variation of the proposed method, indicating that measurements of gene expression changes in a single oocyte may be more precise than the measurements in the oocyte population. Several factors may contribute to the observed high variation of sampling in this technique, most notably variability of cytoplasmic collections variability of total RNA isolation and variability of first-strand cDNA synthesis. The most prominent of them was found to be variability of the collection volume. The SD of the total content of isolated RNA for samples independently collected from the same oocyte via multiple cytoplasmic collections was 40% of the mean isolated RNA content (Fig. S1). This roughly equals the overall SD of Cq values, which was determined to be ≥ 0.5 PCR cycles for various genes (Fig. 5A), thus strongly suggesting that variability of the collection volume is the main source of sampling variation. Much lower variability was observed for RNA isolation (Fig. S1), reverse transcription and quantitative PCR (Fig. 5A). Presumably, one of the major factors related to the high variability of the collection volume is the high viscosity of the oocyte cytoplasm. This requires needles with a wide opening, prepared by manually breaking the FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
A. A. Tokmakov et al.
needle tip (Fig. 1C). This process is not standardized, and produces a significant variation in tip shapes and diameters, thereby affecting the collection volume. In addition, the depth and angle of needle insertion may be poorly controlled during the collections. Another factor that may significantly increase sampling variation of the applied technique is asymmetry in the spatial distribution of mRNA in Xenopus oocytes. Two classes of maternal mRNAs, one preferentially located towards the animal pole and another towards the vegetal pole, have been identified by previous studies [24–27]. Realtime PCR tomography of Xenopus oocytes demonstrated that this polarization is substantial but not extreme: a significant fraction of mRNA molecules of both distribution patterns were found in the equatorial area [27]. We observed that samples obtained from the same oocyte by serial collections of the cytoplasm from a near-equatorial region generate the most reproducible quantitative estimations of gene expression (Fig. S1). However, it is impossible to completely standardize the sampling procedure given the substantial intracellular mRNA gradient and the practical impossibility of cytoplasmic collections from exactly the same area of the cytoplasm. Our study revealed a significant degree of gene expression variation among the individual Xenopus oocytes (Figs 5 and 6). This variation reflects true biological variability and is not an artifact of our applied technique. AS the expression levels of all analyzed genes, including housekeeping genes, varied from cell to cell, we normalized quantitative PCR data by the amount of extracted RNA at the stage of reverse transcription. Quantification of housekeeping (actin) and apoptotic (mcl1) gene expression in individual oocytes revealed s lognormal distribution of transcript levels in the oocyte population (Fig. 6). Previously, a lognormal transcript distribution has been reported for cells isolated from mammalian tissues [28]. The similar SD values obtained for the actin transcript in the two studies suggest a similar degree of gene expression heterogeneity in the populations of frog oocytes and mammalian cells. The expression variation of the housekeeping genes, such as actin, histone H4 and gapdh, was significantly lower than that of apoptotic genes (mcl1 and casp7) and regulatory genes (stat3 and pdk3) (Fig. 5B). The registered difference in the calculated SDs of Cq values corresponds to an approximately fourfold difference in the variation of housekeeping and apoptotic/regulatory genes. Similarly, the housekeeping mammalian actin gene ActB was found to have a significantly lower standard deviation of expression than the inducible insulin genes Ins1 and Ins2 in mammalian cells [28]. The high expression variation of FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
Monitoring gene expression in a single oocyte
apoptosis-related genes, revealed by the present study, may explain the significant heterogeneity of the apoptotic response observed previously in unfertilized Xenopus eggs [29]. We further applied this technique to monitor gene expression changes elicited in Xenopus oocytes by the maturation-promoting hormone progesterone. In accordance with previous reports, we found that transcript levels remained unchanged in progesterone-treated oocytes during the first 12 h. However, a significant decrease in the content of all analyzed transcripts was observed in the eggs after 24 h of progesterone treatment (Fig. 7B). This change concerned all analyzed transcripts encoding various types of proteins. By 36 h, only the most abundant transcripts were detected in the egg cytoplasm (Fig. 7C). Remarkably, the decrease in mRNA levels coincided with the hallmark events of Xenopus egg apoptosis, such as cytochrome c release and caspase 3/7 activation (Table S1), suggesting a link between mRNA degradation and apoptosis. It has been reported recently that mature aging Xenopus eggs die by apoptosis, whereas immature oocytes are markedly resistant to apoptotic cell death [29,30]. The stability of transcript levels in apoptosis-resistant immature oocytes (Fig. 7C) strongly suggests causality between mRNA degradation and apoptosis. Previously, cytoplasmic mRNAs were found to be specifically degraded in some mammalian cell lines as a part of the early apoptotic response [31]. In addition, specific cleavage of mRNA has been observed during apoptotic cell death in plants [32]. Nevertheless, in contrast to well-characterized changes to proteins and DNA that occur during cell death, the fate of mRNAs in apoptotic cells has not been extensively studied and requires further investigation. Also, the causality between mRNA degradation and apoptosis in the unfertilized frog eggs requires further study. In conclusion, this technique using multiple cytoplasmic collections and quantitative RT-PCR allows reliable detection of various mRNAs expressed at the level of more than 103 copies per oocyte. It may be used for monitoring the expression dynamics of heterologous genes in single living Xenopus oocytes, with a minimal effect on their viability. The technique also allows expression profiling and monitoring of multiple endogenous transcripts in individual oocytes.
Experimental procedures Animals and cells Female frogs of Xenopus laevis were purchased from Hamamatsu Seibutsu Kyozai (Hamamatsu, Japan). Animal han-
111
A. A. Tokmakov et al.
Monitoring gene expression in a single oocyte
dling was performed in accordance with the Kobe University Animal Experimentation Regulations (permission number KEN-12). All experiments with oocytes and eggs were performed at ambient temperature (21–23 °C). To obtain oocytes, the frogs were anesthetized, then the ovaries were surgically removed and placed into OR-2 solution containing 82.5 mM NaCl, 2.5 mM KCl, 1 mM CaCl2, 1 mM MgCl2, 1 mM Na2HPO4, 5 mM HEPES, pH 7.6. The chemicals for buffer preparation were obtained from Wako (Osaka, Japan), Nacalai Tesque (Kyoto, Japan) or Sigma (St Louis, MO, USA). The ovaries were manually dissected into clumps of 50–100 oocytes and extensively washed with ten-fold volume of OR-2 solution. The clumps of oocytes were treated with 0.5 mgmL 1 collagenase (280 Umg 1; Wako) in OR-2 for 3 h by shaking at 60 rpm. Oocytes were extensively washed with 10-fold volume of OR-2 solution and left to stabilize for 4 h. Undamaged defolliculated stage VI oocytes were manually selected and used in experiments. In vitro oocyte maturation was induced by addition of 10 lM progesterone (Sigma). Microscopic observations of oocyte morphology were performed using a stereomicroscope Leica S8APO (Leica Microsystems, Wetzlar, Germany). Cell images were acquired using an EC3 stereomicroscope digital color camera and processed using Leica application suite LAS EZ version 1.8.
Microinjections and cytoplasmic collections The capillaries for microinjections and cytoplasmic collections were prepared by pulling #3–00–203–G/X replacement tubes (Drummond Scientific Company, Broomall, PA, USA) using a dual-stage glass micropipette puller (Narishige, Tokyo, Japan). The capillary tips for microinjection were sharpened using an EG-44 microgrinder (Narishige). However, the needles for cytoplasmic collection differ from those for microinjection. They are thicker and have a wider tip opening that is not grinder-sharpened but is forcepsbroken (Fig. 1C). Using the narrow and sharpened microinjection capillaries for cytoplasmic collections is difficult because they are easily blocked. The capillary was positioned for microinjections and cytoplasmic samplings using a three-axis micromanipulator under microscopic observation. Quantitative injections of a T7 promoter-driven luciferase-encoding plasmid and T7 RNA polymerase (T7 RNAP) protein into oocytes were performed using the pulse-directed Nanoject II nanoliter injector system (Drummond Scientific Company). The buffer-exchanged T7 RNAP protein and luciferase-encoding DNA were premixed and loaded into the injection needles as described previously [22]. Approximately 10 ng of nucleic acid and 10 ng of T7 RNAP protein were injected into an oocyte to initiate luciferase synthesis via coupled cytoplasmic transcription and translation. The same injection equipment albeit with different needles was used for cytoplasmic samplings. The capillary tip was inserted into the oocyte to a
112
depth of 300–500 lm, followed by collection of a 10 nL volume of cytoplasmic material. The tip was withdrawn into 2 lL of water containing 1 UlL 1 of RNasin Plus ribonuclease inhibitor (Promega, Fitchburg, WI, USA). The collected samples were frozen in liquid nitrogen and kept at 80 °C until gene expression analysis.
Quantitative RT-PCR Total RNA was extracted from the collected samples using an RNeasy Mini purification kit (Qiagen, Valencia, CA, USA). RNA quality was assessed using denaturing agarose gel electrophoresis, through detection of sharp 18S and 28S bands of ribosomal RNA and the absence of short degraded RNA fragments. Reverse transcription was performed using an AccuScript High Fidelity 1st strand cDNA synthesis kit (Agilent, Santa Clara, CA, USA), with 50 ng total RNA and random primers. Gene-specific nucleotide sequences were detected using a SYBR Green detection system (Roche, Penzberg, Germany). Quantitative PCR was performed in a final volume of 20 lL using a LightCycler 480 instrument (Roche) under the following cycling conditions: 95 °C for 2 min, 45 cycles at 95 °C for 10 s, 60 °C for 10 s, 72 °C for 10 s. After cycling, a melting curve was recorded between 65 °C and 95 °C with a ramp rate of 0.11 °Cs 1. Detection primers were designed using PRIMER3 software (http://biotools.umassmed.edu/bioapps/ primer3_www.cgi) [33]. Their nucleotide sequences are provided in Table S2. The primers were purchased from Invitrogen (Carlsbad, CA, USA) and used for PCR at a final concentration of 0.5 lM. The primers produced a single peak in the derivative of the melting curve, and did not amplify of non-template controls.
Data analysis and statistics The data for quantitative PCR experiments were analyzed using the LIGHTCYCLER 480 software. Measured Cq values for quantitative PCR technical replicates were averaged. Their standard deviations were rather small and generally cannot be seen in the plots. Absolute quantification of the transcript numbers for various genes was performed using standard curves obtained with the corresponding doublestranded PCR products, on the assumption that the reverse transcription efficiency is 100%. The representative standard curves used for cyclin B2 and luciferase absolute gene quantification are shown in Fig. S2B. For each generated standard curve, the error and efficiency were estimated. The experimentally determined efficiency of PCR reactions for various gene products was > 95%, as calculated by the LIGHTCYCLER 480 analytical software. No PCR efficiency correction was applied to the data. Spike control of quantitative PCR was performed by addition of total Xenopus RNA extracted from one 10 nL cytoplasmic collection to the reaction amplifying the heterologous luciferase gene
FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
A. A. Tokmakov et al.
(Fig. S2A). The distribution graphs were smoothed using the Excel chart smoothing algorithm.
Other methods The total content of isolated RNA in the collected cytoplasmic samples was determined spectrophotometrically using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). Protein content was determined using a Bio-Rad protein assay kit (Hercules, CA, USA), and luciferase protein detection in the oocytes was performed as described previously [22]. Luciferease activity was measured using ImageQuant LAS4000 mini imaging system (GE Healthcare Life Sciences, Uppsala, Sweden) Measurements of Cdk1 activity and activation of mitogen-activated protein kinase, the cytochrome c release assay, the caspase 3/7 activity assay, detection of intracellular ATP and measurements of egg diameter were performed as described previously [29].
Acknowledgements This work was supported by the Research Fund for Foreign Visiting Professors from Kobe University (to A.A.T.), and Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (25440023 to A.A.T., 21770142 to T.I. and 21570225 to Y.F.). We are grateful to Yuji Kageyama (Research Center for Environmental Genomics, Kobe University, Japan) for providing the ImageQuant LAS4000 mini imaging system and to Yumiko Terazawa and Mariko Ikeda (Systems and Structural Biology Center, Yokohama Institute, RIKEN, Japan) for valuable help with the initial set-up of quantitative RT-PCR experiments.
References 1 Levsky JM, Shenoy SM, Pezo RC & Singer RH (2002) Single-cell gene expression analysis. Science 297, 836–840. 2 Cai L, Friedman N & Xie XS (2006) Stochastic protein expression in individual cells at the single molecule level. Nature 440, 358–362. 3 Dirks RW & Tanke HJ (2006) Advances in fluorescent tracking of nucleic acids in living cells. Biotechniques 40, 489–496. 4 Tyagi S (2009) Imaging intracellular RNA distribution and dynamics in living cells. Nat Methods 6, 331–338. 5 Itzkovitz S & van Oudeaarden A (2011) Validating transcripts with probes and imaging technology. Nat Methods 8, S12–S19. 6 Armitage BA (2011) Imaging of RNA in live cells. Curr Opin Chem Biol 15, 806–812.
FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
Monitoring gene expression in a single oocyte
7 Bengtsson M, Hemberg M, Rorsman P & Stahlberg A (2008) Quantification of mRNA in single cells and modeling of RT-qPCR induced noise. BMC Mol Biol 9, 63. 8 Tichopad A, Kitchen R, Riedmaier I, Becker C, Stahlberg A & Kubista M (2009) Design and optimization of reverse-transcription quantitative PCR experiments. Clin Chem 55, 1816–1823. 9 van Haaften RI, Schroen B, Janssen BJ, van Erk A, Debets JJ, Smeets HJ, Smits JF, van den Wijngaard A, Pinto YM & Evelo CT (2006) Biologically relevant effects of mRNA amplification on gene expression profiles. BMC Bioinformatics 7, 200. 10 Croner RS, Lausen B, Schellerer V, Zeittraeger I, Wein A, Schildberg C, Papadopoulos T, Dimmler A, Hahn EG, Hohenberger W et al. (2009) Compatibility of microarray data between amplified and non-amplified RNA in colorectal carcinoma. J Biomed Biotechnol 2009, 837170. 11 Sive HL, Grainger RM & Harland RM (2010) Microinjection of Xenopus oocytes. Cold Spring Harb Protoc. 2010, 12. doi:10.1101/pdb.prot5536. 12 Moar VA, Gurdon JB, Lane CD & Marbaix G (1971) Translational capacity of living frog eggs and oocytes, as judged by messenger RNA injection. J Mol Biol 61, 93–103. 13 Gurdon JB, Lane CD, Woodland HR & Marbaix G (1971) Use of frog eggs and oocytes for the study of messenger RNA and its translation in living cells. Nature 233, 177–182. 14 Masui Y & Markert CL (1971) Cytoplasmic control of nuclear behavior during meiotic maturation of frog oocytes. J Exp Zool 172, 121–145. 15 Luzzi V, Lee CL & Albritton NL (1997) Localized sampling of cytoplasm from Xenopus oocytes for capillary electrophoresis. Anal Chem 69, 4761–4767. 16 Lee CL, Linton J, Soughayer JS, Sims CE & Allbritton NL (1999) Localized measurement of kinase activation in oocytes of Xenopus laevis. Nat Biotechnol 17, 759–762. 17 Wagner J, Fall CP, Hong F, Sims CE, Allbritton NL, Fontanilla RA, Moraru II, Loew LM & Nuccitelli R (2003) A wave of IP3 production accompanies the fertilization Ca2+ wave in the egg of the frog, Xenopus laevis: theoretical and experimental support. Cell Calcium 35, 433–447. 18 Okuyama K, Sagata N, Shiokawa K & Yamana K (1981) Localization of poly(A) sequences that are lost during oocyte maturation of Xenopus laevis. Cell Struct Funct 6, 279–282. 19 Johnson CE, Freel CD & Kornbluth S (2010) Features of programmed cell death in intact Xenopus oocytes and early embryos revealed by near-infrared fluorescence and real-time monitoring. Cell Death Differ 17, 170–179. 20 Geib S, Sandoz G, Carlier E, Cornet V, CheynetSauvion V & De Waard M (2001) A novel Xenopus
113
A. A. Tokmakov et al.
Monitoring gene expression in a single oocyte
21
22
23
24 25
26
27
28
114
oocyte expression system based on cytoplasmic coinjection of T7-driven plasmids and purified T7-RNA polymerase. Receptors Channels 7, 331–343. Altafaj X, Joux N, Ronjat M & De Waard M (2006) Oocyte expression with injection of purified T7 RNA polymerase. Methods Mol Biol 322, 55–67. Tokmakov AA, Matsumoto E, Shirouzu M & Yokoyama S (2006) Coupled transcription-andtranslation – a method of choice for heterologous gene expression in Xenopus oocytes. J Biotechnol 122, 5–15. Sindelka R, Sidova M, Svec D & Kubusta M (2010) Spatial expression patterns in the Xenopus laevis oocytes measured with qPCR tomography. Methods 51, 87–91. Mowry KL & Cote CA (1999) RNA sorting in Xenopus oocytes and embryos. FASEB J 13, 435–445. Zhou Y & King ML (2004) Sending RNAs into the future: RNA localization and germ cell fate. IUBMB Life 56, 19–27. King ML, Messitt TJ & Mowry KL (2005) Putting RNAs in the right place at the right time: RNA localization in the frog oocyte. Biol Cell 97, 19–33. Sindelka R, Jonak J, Hands R, Bustin SA & Kubista M (2008) Intracellular expression profiles measured by real-time PCR tomography in the Xenopus laevis oocyte. Nucleic Acids Res 36, 387–392. Bengtsson M, Rorsman P, Stahlberg A & Kubista M (2005) Gene expression profiling in single cells from the pancreatic islets of Langerhans reveals lognormal distribution of mRNA levels. Genome Res 15, 1388–1392.
29 Tokmakov AA, Iguchi S, Iwasaki T & Fukami Y (2011) Unfertilized frog eggs die by apoptosis following meiotic exit. BMC Cell Biol 12, 56. 30 Pasquier DD, Dupre A & Jessus C (2011) Unfertilized Xenopus eggs die by Bad-dependent apoptosis under the control of Cdk1 and JNK. PLoS ONE 6, e23672. 31 Del Prete MJ, Robles MS, Guao A, Martınez-AC Izquierdo M & Garcia-Sanz JA (2002) Degradation of cellular mRNA is a general early apoptosis-induced event. FASEB J 16, 2003–2005. 32 Hoat TX, Nakayashiki H, Tosa Y & Mayama S (2006) Specific cleavage of ribosomal RNA and mRNA during victorin-induced apoptotic cell death in oat. Plant J 46, 922–933. 33 Rozen S & Skaletsky H (2000) Primer3 on the WWW for general users and biologists programmers. Methods Mol Biol 132, 365–386.
Supporting information Additional supporting information may be found in the online version of this article at the publisher’s web site: Fig. S1. Reproducibility of sample collection and RNA extraction. Fig. S2. Spike control and standard curves of quantitative PCR. Table S1. Intracellular events in progesterone-treated Xenopus oocytes. Table S2. Nucleotide sequences of gene-specific quantitative PCR primers used in the study.
FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
Keywords gene expression variation; mRNA degradation; quantitative RT-PCR; single-cell gene expression analysis; Xenopus oocytes Correspondence A. A. Tokmakov, Research Center for Environmental Genomics, Kobe University, 1-1 Rokko dai, Nada, Kobe 657-8501, Japan Fax: +81 78 803 5951 Tel: +81 78 803 5953 E-mail: [email protected] (Received 28 March 2013, revised 10 October 2013, accepted 21 October 2013) doi:10.1111/febs.12576
Oocytes and eggs of the African clawed frog, Xenopus laevis, are commonly used in gene expression studies. However, monitoring transcript levels in the individual living oocytes remains challenging. To address this challenge, we used a technique based on multiple repeated collections of nanoliter volumes of cytoplasmic material from a single oocyte. Transcript quantification was performed by quantitative RT-PCR. The technique allowed monitoring of heterologous gene expression in a single oocyte without affecting its viability. We also used this approach to profile the expression of endogenous genes in living Xenopus oocytes. Although frog oocytes are traditionally viewed as a homogenous cell population, a significant degree of gene expression variation was observed among the individual oocytes. A lognormal distribution of transcript levels was revealed in the oocyte population. Finally, using this technique, we observed a dramatic decrease in the content of various cytoplasmic mRNAs in aging unfertilized eggs but not in oocytes, suggesting a link between mRNA degradation and egg apoptosis.
Introduction Transcript levels vary substantially among cells in seemingly homogenous populations [1,2], requiring gene expression analysis in individual cells. Recently, several techniques for transcript quantification and imaging in single cells have been developed [3–6]. The methods using fluorescent proteins and hybridization probes, such as MS2-based and molecular beaconbased approaches, hybridization probes and molecular beacons, label non-covalently specific mRNA molecules and allow real-time detection and measurement of transcript levels in single cells. However, these methods target only specific pre-selected mRNAs, and cannot be used for comprehensive transcriptome profiling. Alternatively, single-cell gene expression profiling using microarray and quantitative RT-PCR may be applied for the whole transcriptome. However, these methods cannot measure gene expression in living cells, as they are based on the extraction of total cellular mRNA.
The major difficulty of gene expression profiling in single cells is the need to measure small numbers of mRNA molecules. Microarray technology requires relatively large amounts of RNA for analysis and a preamplification step with broad applicability to many target genes. On the other hand, quantitative RT-PCR, which is currently the most sensitive and reproducible technique for gene expression quantification, has, in principle, the sensitivity to detect a single target mRNA molecule. A practical threshold of approximately 20 copies per reaction has been suggested [7,8]. Pre-amplification may also be required for this technique if a sample is split and multiple transcripts are quantified. However, the maintenance of bona fide gene expression profiles remains a bottleneck of pre-amplification procedures [9,10]. Thus, the existing data for transcriptome profiling have been obtained mostly on large cell populations, representing only mean mRNA levels.
Abbreviations Cq, quantification cycle; GV, germinal vesicle; GVBD, germinal vesicle breakdown; PG, progesterone; T7 RNAP, T7 RNA polymerase; RTPCR, reverse transcriptase PCR.
104
FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
A. A. Tokmakov et al.
Monitoring gene expression in a single oocyte
Oocytes and eggs of the African clawed frog, Xenopus laevis, have been commonly used in expression studies. These cells are a popular model because of their large size and the ease of obtaining them in large numbers. The oocytes are obtained freshly from living frogs and may be treated like a primary cell culture. The technique of oocyte microinjection enables easy delivery of nanoliter volumes of various materials. The oocytes are able to withstand introduction of up to 50 nL of injected volume, and injections may be continued over several days [11]. Such microinjection experiments were described as early as 1971 [12,13]. In addition, nanoliter quantities of the cytoplasm may also be extracted from living frog oocytes and matured eggs. The standard equipment for oocyte microinjection may be used for cytoplasmic sampling with minimal adjustments. Cytoplasmic transfers were originally used by Masui and Markert to reveal maturation-promoting factor in hormone-treated Rana pipiens oocytes [14]. Localized sampling of the Xenopus oocyte cytoplasm has been used to perform capillary electrophoresis [15] and measurements of kinase activation [16] and inositol triphosphate gradients [17]. In the present study, we investigated the feasibility of gene expression monitoring in single Xenopus oocytes and eggs using multiple cytoplasmic collections and quantitative RT-PCR.
Results Multiple cytoplasmic collections from Xenopus oocytes Fully grown Xenopus oocytes at stage VI are large cells of approximately 1.2 mm in diameter arrested in prophase of the first meiotic division. They are visibly A
B
divided into the dark-pigmented animal hemisphere and the light-colored vegetal hemisphere (Fig. 1A,B). The oocyte nucleus, called the germinal vesicle, is located in the animal hemisphere, occupying approximately one quarter of the total oocyte volume. Cytoplasmic samplings for gene expression analysis must be performed from the vegetal near-equatorial area of oocytes in order not to hit the nucleus. The needles for cytoplasmic collection have a wider tip opening than those for microinjection (Fig. 1C). It is important to make the collections from the same area of the cytoplasm to minimize the measurement variation due to sampling. Indeed, samples collected from the equatorial area of oocytes displayed much lower variation in the content of extracted RNA than those collected randomly (Fig. S1; see below also). In our experiments, the volume of a single cytoplasmic collection was set to 10 nL. Given that the estimated cytoplasmic volume of one oocyte is approximately 700 nL, and its total RNA content exceeds 5 lg [18], it may be calculated that a single cytoplasmic collection obtains approximately 70 ng of total RNA, and the total amount of extracted RNA from a single cytoplasmic collection fits this estimate (Fig. S1). Taking into account that at least ten target molecules per reaction are necessary for accurate quantification by quantitative PCR [7], this approach should allow detection of specific transcripts that are present at the level of more than 103 copies per oocyte. Oocyte viability after multiple cytoplasmic collections Cytoplasmic samplings inflict damage to oocytes, therefore, for practical application of the proposed C
Fig. 1. Cytoplasmic collections and microinjections. (A,B) Schematic representation of a Xenopus oocyte (A) and its microscopic image with the inserted sampling capillary during collection of cytoplasmic material (B). The large oocyte nucleus (germinal vesicle, GV) is located in the animal hemisphere. (C) Photograph of microcapillary tips used for cytoplasmic collections (c) and microinjections (i).
FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
105
A. A. Tokmakov et al.
Monitoring gene expression in a single oocyte
technique, it is important to evaluate their damaging effect. We investigated the impact of multiple cytoplasmic collections on Xenopus oocyte viability. It was found that the oocytes are able to withstand well up to eight cytoplasmic collections (Fig. 2A). After four collections, 80% of oocytes remained intact for 36 h, and approximately 50% were still morphologically normal by 48 h. Oocytes deteriorated significantly after 8 cytoplasmic collections. Still, more than half of them survived this procedure by 24 h, and 30% of the oocytes remained intact by 48 h (Fig. 2A). Viability of control oocytes exceeded 90% at that time. Notably, oocyte viability after a single 80 nL collection, i.e. the volume of eight 10 nL cytoplasmic collections, was practically the same as the viability of oocytes after a single 10 nL collection. These data suggest that it is the number of collections rather than the removal of certain cytoplasmic volume (below 80 nL) that causes cell death. In other words, oocytes are able to withstand the loss of approximately 10% of their cytoplasm with little effect on their viability. Generally, two outcomes were observed in damaged oocytes. Most often, they became susceptible to leakage, rapidly lost a significant amount of cytoplasmic material, shrank and became discolored. Alternatively, damaged non-leaky oocytes developed mottling in the animal hemisphere (Fig. 2B). This morphology has previously been associated with oocyte death [19]. In addition, these oocytes were functionally compromised, as they did not respond to the maturation hormone progesterone (Fig. 2B). Based on these results,
A
gene expression in further experiments was monitored mainly over 24 h, the period over which multiple cytoplasmic samplings exerted minimal effect on oocyte viability. Monitoring luciferase expression in a single oocyte Heterologous genes are commonly expressed in Xenopus oocytes to investigate their functions. In order to evaluate the applicability of multiple cytoplasmic collections for monitoring heterologous gene expression in a single oocyte, a T7 promoter-driven luciferase-encoding plasmid was microinjected into the oocyte cytoplasm together with T7 RNA polymerase (T7 RNAP) protein. It was shown previously that efficient coupling of T7 RNAP-mediated transcription with the intrinsic oocyte translation machinery occurs in the oocyte cytoplasm [20–22]. Robust production of luciferase mRNA was detected in the cytoplasm immediately after microinjection (Fig. 3A). Synthesis of luciferase mRNA continued over several hours, producing, according to calibration curve-based calculations, hundreds of mRNA molecules per single molecule of injected DNA. The level of luciferase mRNA decreased to some extent by 24 h (Fig. 3B); however, luciferase protein was still being produced at that time (Fig. 3D). In ovo synthesis of luciferase protein continued for many hours and was still evident at 24 h after DNA/T7 RNAP injection (Fig. 3C,D). These findings are largely consistent with previously published data from bulk-scale experiments
B
Fig. 2. Viability of Xenopus oocytes after multiple cytoplasmic collections. (A) Oocyte viability after zero, one, two, four and eight cytoplasmic collections at the indicated times after the start of sampling (0–48 h). The collections were performed at 1 h intervals. Ten to thirty oocytes were monitored for each number of cytoplasmic collections. The columns marked (1*) refer to oocyte viability after a single 80 nL cytoplasmic collection. (B) Images of normal and damaged immature oocytes, and their response to progesterone treatment, as determined by the occurrence of germinal vesicle breakdown (GVBD).
106
FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
A. A. Tokmakov et al.
Monitoring gene expression in a single oocyte
A
B
C
D
Fig. 3. Monitoring luciferase expression in a single Xenopus oocyte. (A–D) Kinetics of luciferase mRNA (A,B) and protein (C,D) synthesis in Xenopus oocytes after the cytoplasmic coinjection of a T7 promoter-driven luciferase-encoding plasmid and T7 RNAP protein. Four oocytes from the same batch were monitored in the experiment; however, the representative results of only one single-cell analysis are shown. Bars in (C) and (D) represent SDs of three independent measurements of luciferase activity. The SDs of replicate measurements in (A) and (B) were very small and are not visible in the plots.
[22]. It may therefore be concluded that the applied technique of multiple cytoplasmic collections allows effective monitoring of heterologous gene expression in a single Xenopus oocyte. Gene expression profiling in single Xenopus oocytes The filly grown Xenopus oocytes of stage VI contain a large quantity of maternal mRNA that is synthesized and accumulated during oogenesis. We investigated whether it is possible to detect various endogenous transcripts in cytoplasmic samples taken from a single oocyte. The task of gene expression profiling in the nanoliter samples is challenging because of the low abundance of many cellular transcripts. Indeed, expression profiling of multiple genes classified as various functional types, such as housekeeping, regulatory, apoptosis-related and metabolic genes, revealed a great variation in their expression levels (Fig. 4). Although some transcripts were present at high copy numbers, up to 106 copies per sample, FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
only a few copies of other transcripts were present. Gene expression profiling in different oocytes revealed a great variation of transcript levels. As an example, the expression profiles of two oocytes from the same batch are shown in Fig. 4A,B. Most genes in oocyte 1 were expressed at higher levels than in oocyte 2, so these cells may provisionally be described as ‘highexpressing’ and ‘low-expressing’ oocytes. However, the magnitude of difference in the expression levels of different genes was not the same. Some genes were expressed at similar levels in both oocytes, whereas others displayed an approximately 100-fold difference in expression levels. Notably, both ‘high-expressing’ and ‘low-expressing’ oocytes were viable and responded normally to progesterone treatment. These data demonstrate that quantification of lowabundance transcripts in nanoliter samples of the oocyte cytoplasm nears the detection limit of the quantitative RT-PCR technique. As estimated above, this approach allows reliable quantification of specific mRNAs present at the level of more than 103 copies per oocyte. 107
A. A. Tokmakov et al.
Monitoring gene expression in a single oocyte
A
A
B
B
Measurement and sampling variations of the analysis
Fig. 5. Variations of the expression analysis. (A) Estimation of measurement and sampling variations. The dashed line indicates the maximal level of the SD for Cq values determined for the studied transcripts in four replicate measurements. Columns represent SDs of Cq values determined for eight cytoplasmic collections taken from the same oocyte. The samples of extracted RNA were divided into aliquots before reverse transcription. (B) Cell-to-cell variation of gene expression between individual Xenopus oocytes. Columns represent SDs of Cq values determined for 16 oocytes of the same batch.
Experimental variation introduced by the applied technique may obscure gene expression quantification. It is therefore important to estimate its degree. In general, experimental variation of a method includes measurement variation and sampling variation. It has been reported previously that measurement variation of quantitative PCR, i.e. the variation of quantitative PCR measurements for the same template, is typically very low [10]. In our hands, the standard deviation of quantification cycle (Cq) values determined for eight studied transcripts in four replicate measurements was less than 0.25 PCR cycles (Fig. 5A), and it was geneindependent. Given that the above estimation covers both RT and PCR, the low measurement variation of this analysis is directly related to high reproducibility of the first-strand cDNA synthesis and quantitative PCR technique. Under our conditions, the low measurement variation relies on the high efficiency of reverse transcription and PCR due to the low amount of starting material (see Discussion). In a spike control experiment, the isolated oocyte RNA obtained from 10 nL cytoplasmic collection had no inhibitory effect on a control PCR involving a heterologous transcript
(Fig. S2A). As a result, the efficiency of PCR reactions for various analyzed gene products was > 95%. Sampling variation of the performed gene expression analysis includes the variation of cytoplasmic collections, variation of mRNA isolation and variation of first-strand cDNA synthesis. This was significantly higher than measurement variation, with a standard deviation (SD) of 0.5–0.7 PCR cycles, as determined for eight cytoplasmic collections taken from the same oocyte (Fig. 5A). Sampling variation was also geneindependent. These data indicate that sampling variation represents the major source of experimental variation in the proposed approach. To further understand the source of high sampling variation, we evaluated the reproducibility of sample collection and RNA isolation. It was found that the SD of total contents of isolated RNA in the aliquots of a split sample was relatively low at approximately 10% of its mean value, whereas a much higher SD (approximately 40%) was obtained for samples independently collected from the same oocyte (Fig. S1). These data suggest that variability of the collection volume is the main source of sampling variation.
Fig. 4. Expression profiling in Xenopus oocytes. (A,B) Gene expression profiles of two oocytes from the same batch. Gene expression levels were measured in two to four replicates. SDs of replicate measurements were very small and are not visible in the plots.
108
FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
A. A. Tokmakov et al.
Plausible explanations for the observed high sampling variation are discussed below. Variability of gene expression in individual oocytes We next evaluated biological cell-to-cell variation of gene expression between individual Xenopus oocytes. The variability of transcript levels in these cells has not previously been quantitatively estimated and remained undefined at this stage of the study. This kind of variation has important implications for the interpretation of gene expression data. We confirmed the different degree of gene expression variation for different genes by calculating SDs of Cq values for 16 oocytes of the same batch. In contrast to measurement and sampling variations, this parameter was genedependent and varied from 1.5 to approximately 3.0 PCR cycles (Fig. 5B). The lowest variation was observed for the highly expressed constitutive genes, such as actin, histH4 (histone H4) and gpdh (glyceraldehyde-3-phosphate dehydrogenase), validating their housekeeping designation, whereas apoptosis-related and regulatory genes showed much higher expression variations. To further examine gene expression variability, the distributions of Xenopus oocytes according to gene
Monitoring gene expression in a single oocyte
expression levels were plotted in logarithmic and linear scales for the actin and mcl1 genes (Fig. 6). The graphs revealed a lognormal distribution of transcript levels in the investigated oocyte population. In accordance with the data presented in Fig. 5B, the mcl1 gene had much greater expression variation than the actin gene, as indicated by the different widths of the approximated distribution curves (blue and red curves in Fig. 6). mRNA degradation in mature oocytes The technique of multiple cytoplasmic collections was also used for analysis of gene expression changes induced in Xenopus oocytes by the steroid hormone progesterone. The hormone promotes oocyte transition from prophase I to metaphase II, producing fertilization-competent metaphase-arrested eggs. It was found that progesterone treatment had little effect on mRNA levels in the oocytes within the first 12 h following hormone administration (Fig. 7A). However, a significant decrease in the content of all analyzed transcripts was observed after 24 h (Fig. 7B). By 36 h, only the most abundant transcripts were detected in the egg cytoplasm by the applied technique. A moderate decline in mRNA levels was also observed in progesterone-untreated Xenopus oocytes by that time (Fig. 7C). Notably, the timing of mRNA degradation coincided with the hallmark events of Xenopus egg apoptosis, such as cytochrome c release and caspase 3/7 activation (Table S1). By 48 h, no intact mRNA molecules were detected in the egg cytoplasm by the applied technique (data not shown). By that time, prominent egg swelling was observed, indicating disruption of osmotic homeostasis at a final stage of cell death (Table S1).
Discussion
Fig. 6. Variability of gene expression in Xenopus oocytes. Histograms of actin and mcl1 gene expression levels in the individual oocytes of the same batch plotted using logarithmic scale (main graph) and linear scale (inset). The distribution graphs (red and blue curves) were smoothed using the Excel chart smoothing algorithm.
FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
In the present work, we describe a simple and reliable technique for monitoring gene expression in a single living Xenopus oocyte. It is based on multiple collections of nanoliter amounts of the oocyte cytoplasm and quantitative RT-PCR. Our data demonstrate that the oocytes are able to withstand multiple cytoplasmic collections, with a high survival rate over 48 h (Fig. 2). This allows monitoring of gene expression dynamics in various experiments using a single oocyte. As a sample application of this technique, we monitored the expression of firefly luciferase gene in a single oocyte during 24 h (Fig. 3). To initiate the synthesis of both luciferase RNA and protein, a T7 promoter-driven luciferase-encoding plasmid was microinjected into the oocyte cytoplasm together with T7 RNAP protein. 109
A. A. Tokmakov et al.
Monitoring gene expression in a single oocyte
A
B
C
Fig. 7. mRNA degradation in progesterone-treated Xenopus oocytes. Gene expression profiling was performed in a single Xenopus oocyte at 12 h (A), 24 h (B) and 36 h (C) after progesterone addition. The specific transcript levels determined at the indicated times were normalized to those obtained in the same oocyte before hormone administration (dark gray bars in all panels). The light gray bars in (C) refer to the transcript levels in a progesterone-untreated control oocyte. Gene expression levels were measured in replicates. The SDs of two to four replicate measurements were very small and are not visible in the plots. Three progesterone-treated and three progesterone-untreated oocytes from the same batch were monitored in the experiment; however, the results of only one representative single-cell analysis are shown.
Synthesis of both luciferase mRNA and protein was observed in the microinjected oocyte for many hours after DNA/T7 RNAP introduction (Fig. 3). These results are largely consistent with the previously published data from bulk-scale experiments [22], providing a proof of concept for the suggested technique of gene expression monitoring. Next we applied this method to profile multiple endogenous transcripts in a single oocyte. The task of transcript profiling in the nanoliter samples of the oocyte cytoplasm is challenging because of the low abundance of many cellular transcripts. Indeed, the technique approaches the detection limit of quantitative RT-PCR for the low-abundance transcripts (Fig. 4). As a result, the low-copy number mRNAs, such as stat3 and pdk3, showed a higher variation of expression levels than high-copy number mRNAs (Fig. 5B). Notably, the small sampling volume of cytoplasmic collections is directly related to the high quality of PCR data. Previously, yolk platelets localized in the vegetal hemisphere of the oocyte have been reported to inhibit reverse transcription and PCR [23]. However, the spike control experiment did not reveal any inhibitory effects of the samples obtained by multiple cytoplasmic collections on the PCR-based quantification (Fig. S2A). Evidently, it is high dilution of the cytoplasmic inhibitors during RNA isolation from a small-volume sample that allows high PCR efficiency. However, high sampling variation remains a bottleneck of this approach, and appears to be the major source of experimental variation. As the standard deviation of sampling exceeds 0.5 PCR cycles (Fig. 5A), this approach is not applicable for detection of less than twofold changes in transcript levels. Instead, it 110
may be useful for monitoring high-magnitude expression of heterologous genes (Fig. 3). Also, it provides a means to simultaneously monitor significant multifold changes in the expression of endogenous Xenopus genes in a single oocyte. Indeed, using this technique, we observed a dramatic decrease in the content of various cytoplasmic mRNAs in aging unfertilized eggs (Fig. 7). The biological variation of gene expression in Xenopus oocytes varied from 1.5 to approximately 3.0 PCR cycles for various genes (Fig. 5B). It is still much larger than the technical variation of the proposed method, indicating that measurements of gene expression changes in a single oocyte may be more precise than the measurements in the oocyte population. Several factors may contribute to the observed high variation of sampling in this technique, most notably variability of cytoplasmic collections variability of total RNA isolation and variability of first-strand cDNA synthesis. The most prominent of them was found to be variability of the collection volume. The SD of the total content of isolated RNA for samples independently collected from the same oocyte via multiple cytoplasmic collections was 40% of the mean isolated RNA content (Fig. S1). This roughly equals the overall SD of Cq values, which was determined to be ≥ 0.5 PCR cycles for various genes (Fig. 5A), thus strongly suggesting that variability of the collection volume is the main source of sampling variation. Much lower variability was observed for RNA isolation (Fig. S1), reverse transcription and quantitative PCR (Fig. 5A). Presumably, one of the major factors related to the high variability of the collection volume is the high viscosity of the oocyte cytoplasm. This requires needles with a wide opening, prepared by manually breaking the FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
A. A. Tokmakov et al.
needle tip (Fig. 1C). This process is not standardized, and produces a significant variation in tip shapes and diameters, thereby affecting the collection volume. In addition, the depth and angle of needle insertion may be poorly controlled during the collections. Another factor that may significantly increase sampling variation of the applied technique is asymmetry in the spatial distribution of mRNA in Xenopus oocytes. Two classes of maternal mRNAs, one preferentially located towards the animal pole and another towards the vegetal pole, have been identified by previous studies [24–27]. Realtime PCR tomography of Xenopus oocytes demonstrated that this polarization is substantial but not extreme: a significant fraction of mRNA molecules of both distribution patterns were found in the equatorial area [27]. We observed that samples obtained from the same oocyte by serial collections of the cytoplasm from a near-equatorial region generate the most reproducible quantitative estimations of gene expression (Fig. S1). However, it is impossible to completely standardize the sampling procedure given the substantial intracellular mRNA gradient and the practical impossibility of cytoplasmic collections from exactly the same area of the cytoplasm. Our study revealed a significant degree of gene expression variation among the individual Xenopus oocytes (Figs 5 and 6). This variation reflects true biological variability and is not an artifact of our applied technique. AS the expression levels of all analyzed genes, including housekeeping genes, varied from cell to cell, we normalized quantitative PCR data by the amount of extracted RNA at the stage of reverse transcription. Quantification of housekeeping (actin) and apoptotic (mcl1) gene expression in individual oocytes revealed s lognormal distribution of transcript levels in the oocyte population (Fig. 6). Previously, a lognormal transcript distribution has been reported for cells isolated from mammalian tissues [28]. The similar SD values obtained for the actin transcript in the two studies suggest a similar degree of gene expression heterogeneity in the populations of frog oocytes and mammalian cells. The expression variation of the housekeeping genes, such as actin, histone H4 and gapdh, was significantly lower than that of apoptotic genes (mcl1 and casp7) and regulatory genes (stat3 and pdk3) (Fig. 5B). The registered difference in the calculated SDs of Cq values corresponds to an approximately fourfold difference in the variation of housekeeping and apoptotic/regulatory genes. Similarly, the housekeeping mammalian actin gene ActB was found to have a significantly lower standard deviation of expression than the inducible insulin genes Ins1 and Ins2 in mammalian cells [28]. The high expression variation of FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
Monitoring gene expression in a single oocyte
apoptosis-related genes, revealed by the present study, may explain the significant heterogeneity of the apoptotic response observed previously in unfertilized Xenopus eggs [29]. We further applied this technique to monitor gene expression changes elicited in Xenopus oocytes by the maturation-promoting hormone progesterone. In accordance with previous reports, we found that transcript levels remained unchanged in progesterone-treated oocytes during the first 12 h. However, a significant decrease in the content of all analyzed transcripts was observed in the eggs after 24 h of progesterone treatment (Fig. 7B). This change concerned all analyzed transcripts encoding various types of proteins. By 36 h, only the most abundant transcripts were detected in the egg cytoplasm (Fig. 7C). Remarkably, the decrease in mRNA levels coincided with the hallmark events of Xenopus egg apoptosis, such as cytochrome c release and caspase 3/7 activation (Table S1), suggesting a link between mRNA degradation and apoptosis. It has been reported recently that mature aging Xenopus eggs die by apoptosis, whereas immature oocytes are markedly resistant to apoptotic cell death [29,30]. The stability of transcript levels in apoptosis-resistant immature oocytes (Fig. 7C) strongly suggests causality between mRNA degradation and apoptosis. Previously, cytoplasmic mRNAs were found to be specifically degraded in some mammalian cell lines as a part of the early apoptotic response [31]. In addition, specific cleavage of mRNA has been observed during apoptotic cell death in plants [32]. Nevertheless, in contrast to well-characterized changes to proteins and DNA that occur during cell death, the fate of mRNAs in apoptotic cells has not been extensively studied and requires further investigation. Also, the causality between mRNA degradation and apoptosis in the unfertilized frog eggs requires further study. In conclusion, this technique using multiple cytoplasmic collections and quantitative RT-PCR allows reliable detection of various mRNAs expressed at the level of more than 103 copies per oocyte. It may be used for monitoring the expression dynamics of heterologous genes in single living Xenopus oocytes, with a minimal effect on their viability. The technique also allows expression profiling and monitoring of multiple endogenous transcripts in individual oocytes.
Experimental procedures Animals and cells Female frogs of Xenopus laevis were purchased from Hamamatsu Seibutsu Kyozai (Hamamatsu, Japan). Animal han-
111
A. A. Tokmakov et al.
Monitoring gene expression in a single oocyte
dling was performed in accordance with the Kobe University Animal Experimentation Regulations (permission number KEN-12). All experiments with oocytes and eggs were performed at ambient temperature (21–23 °C). To obtain oocytes, the frogs were anesthetized, then the ovaries were surgically removed and placed into OR-2 solution containing 82.5 mM NaCl, 2.5 mM KCl, 1 mM CaCl2, 1 mM MgCl2, 1 mM Na2HPO4, 5 mM HEPES, pH 7.6. The chemicals for buffer preparation were obtained from Wako (Osaka, Japan), Nacalai Tesque (Kyoto, Japan) or Sigma (St Louis, MO, USA). The ovaries were manually dissected into clumps of 50–100 oocytes and extensively washed with ten-fold volume of OR-2 solution. The clumps of oocytes were treated with 0.5 mgmL 1 collagenase (280 Umg 1; Wako) in OR-2 for 3 h by shaking at 60 rpm. Oocytes were extensively washed with 10-fold volume of OR-2 solution and left to stabilize for 4 h. Undamaged defolliculated stage VI oocytes were manually selected and used in experiments. In vitro oocyte maturation was induced by addition of 10 lM progesterone (Sigma). Microscopic observations of oocyte morphology were performed using a stereomicroscope Leica S8APO (Leica Microsystems, Wetzlar, Germany). Cell images were acquired using an EC3 stereomicroscope digital color camera and processed using Leica application suite LAS EZ version 1.8.
Microinjections and cytoplasmic collections The capillaries for microinjections and cytoplasmic collections were prepared by pulling #3–00–203–G/X replacement tubes (Drummond Scientific Company, Broomall, PA, USA) using a dual-stage glass micropipette puller (Narishige, Tokyo, Japan). The capillary tips for microinjection were sharpened using an EG-44 microgrinder (Narishige). However, the needles for cytoplasmic collection differ from those for microinjection. They are thicker and have a wider tip opening that is not grinder-sharpened but is forcepsbroken (Fig. 1C). Using the narrow and sharpened microinjection capillaries for cytoplasmic collections is difficult because they are easily blocked. The capillary was positioned for microinjections and cytoplasmic samplings using a three-axis micromanipulator under microscopic observation. Quantitative injections of a T7 promoter-driven luciferase-encoding plasmid and T7 RNA polymerase (T7 RNAP) protein into oocytes were performed using the pulse-directed Nanoject II nanoliter injector system (Drummond Scientific Company). The buffer-exchanged T7 RNAP protein and luciferase-encoding DNA were premixed and loaded into the injection needles as described previously [22]. Approximately 10 ng of nucleic acid and 10 ng of T7 RNAP protein were injected into an oocyte to initiate luciferase synthesis via coupled cytoplasmic transcription and translation. The same injection equipment albeit with different needles was used for cytoplasmic samplings. The capillary tip was inserted into the oocyte to a
112
depth of 300–500 lm, followed by collection of a 10 nL volume of cytoplasmic material. The tip was withdrawn into 2 lL of water containing 1 UlL 1 of RNasin Plus ribonuclease inhibitor (Promega, Fitchburg, WI, USA). The collected samples were frozen in liquid nitrogen and kept at 80 °C until gene expression analysis.
Quantitative RT-PCR Total RNA was extracted from the collected samples using an RNeasy Mini purification kit (Qiagen, Valencia, CA, USA). RNA quality was assessed using denaturing agarose gel electrophoresis, through detection of sharp 18S and 28S bands of ribosomal RNA and the absence of short degraded RNA fragments. Reverse transcription was performed using an AccuScript High Fidelity 1st strand cDNA synthesis kit (Agilent, Santa Clara, CA, USA), with 50 ng total RNA and random primers. Gene-specific nucleotide sequences were detected using a SYBR Green detection system (Roche, Penzberg, Germany). Quantitative PCR was performed in a final volume of 20 lL using a LightCycler 480 instrument (Roche) under the following cycling conditions: 95 °C for 2 min, 45 cycles at 95 °C for 10 s, 60 °C for 10 s, 72 °C for 10 s. After cycling, a melting curve was recorded between 65 °C and 95 °C with a ramp rate of 0.11 °Cs 1. Detection primers were designed using PRIMER3 software (http://biotools.umassmed.edu/bioapps/ primer3_www.cgi) [33]. Their nucleotide sequences are provided in Table S2. The primers were purchased from Invitrogen (Carlsbad, CA, USA) and used for PCR at a final concentration of 0.5 lM. The primers produced a single peak in the derivative of the melting curve, and did not amplify of non-template controls.
Data analysis and statistics The data for quantitative PCR experiments were analyzed using the LIGHTCYCLER 480 software. Measured Cq values for quantitative PCR technical replicates were averaged. Their standard deviations were rather small and generally cannot be seen in the plots. Absolute quantification of the transcript numbers for various genes was performed using standard curves obtained with the corresponding doublestranded PCR products, on the assumption that the reverse transcription efficiency is 100%. The representative standard curves used for cyclin B2 and luciferase absolute gene quantification are shown in Fig. S2B. For each generated standard curve, the error and efficiency were estimated. The experimentally determined efficiency of PCR reactions for various gene products was > 95%, as calculated by the LIGHTCYCLER 480 analytical software. No PCR efficiency correction was applied to the data. Spike control of quantitative PCR was performed by addition of total Xenopus RNA extracted from one 10 nL cytoplasmic collection to the reaction amplifying the heterologous luciferase gene
FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
A. A. Tokmakov et al.
(Fig. S2A). The distribution graphs were smoothed using the Excel chart smoothing algorithm.
Other methods The total content of isolated RNA in the collected cytoplasmic samples was determined spectrophotometrically using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). Protein content was determined using a Bio-Rad protein assay kit (Hercules, CA, USA), and luciferase protein detection in the oocytes was performed as described previously [22]. Luciferease activity was measured using ImageQuant LAS4000 mini imaging system (GE Healthcare Life Sciences, Uppsala, Sweden) Measurements of Cdk1 activity and activation of mitogen-activated protein kinase, the cytochrome c release assay, the caspase 3/7 activity assay, detection of intracellular ATP and measurements of egg diameter were performed as described previously [29].
Acknowledgements This work was supported by the Research Fund for Foreign Visiting Professors from Kobe University (to A.A.T.), and Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (25440023 to A.A.T., 21770142 to T.I. and 21570225 to Y.F.). We are grateful to Yuji Kageyama (Research Center for Environmental Genomics, Kobe University, Japan) for providing the ImageQuant LAS4000 mini imaging system and to Yumiko Terazawa and Mariko Ikeda (Systems and Structural Biology Center, Yokohama Institute, RIKEN, Japan) for valuable help with the initial set-up of quantitative RT-PCR experiments.
References 1 Levsky JM, Shenoy SM, Pezo RC & Singer RH (2002) Single-cell gene expression analysis. Science 297, 836–840. 2 Cai L, Friedman N & Xie XS (2006) Stochastic protein expression in individual cells at the single molecule level. Nature 440, 358–362. 3 Dirks RW & Tanke HJ (2006) Advances in fluorescent tracking of nucleic acids in living cells. Biotechniques 40, 489–496. 4 Tyagi S (2009) Imaging intracellular RNA distribution and dynamics in living cells. Nat Methods 6, 331–338. 5 Itzkovitz S & van Oudeaarden A (2011) Validating transcripts with probes and imaging technology. Nat Methods 8, S12–S19. 6 Armitage BA (2011) Imaging of RNA in live cells. Curr Opin Chem Biol 15, 806–812.
FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
Monitoring gene expression in a single oocyte
7 Bengtsson M, Hemberg M, Rorsman P & Stahlberg A (2008) Quantification of mRNA in single cells and modeling of RT-qPCR induced noise. BMC Mol Biol 9, 63. 8 Tichopad A, Kitchen R, Riedmaier I, Becker C, Stahlberg A & Kubista M (2009) Design and optimization of reverse-transcription quantitative PCR experiments. Clin Chem 55, 1816–1823. 9 van Haaften RI, Schroen B, Janssen BJ, van Erk A, Debets JJ, Smeets HJ, Smits JF, van den Wijngaard A, Pinto YM & Evelo CT (2006) Biologically relevant effects of mRNA amplification on gene expression profiles. BMC Bioinformatics 7, 200. 10 Croner RS, Lausen B, Schellerer V, Zeittraeger I, Wein A, Schildberg C, Papadopoulos T, Dimmler A, Hahn EG, Hohenberger W et al. (2009) Compatibility of microarray data between amplified and non-amplified RNA in colorectal carcinoma. J Biomed Biotechnol 2009, 837170. 11 Sive HL, Grainger RM & Harland RM (2010) Microinjection of Xenopus oocytes. Cold Spring Harb Protoc. 2010, 12. doi:10.1101/pdb.prot5536. 12 Moar VA, Gurdon JB, Lane CD & Marbaix G (1971) Translational capacity of living frog eggs and oocytes, as judged by messenger RNA injection. J Mol Biol 61, 93–103. 13 Gurdon JB, Lane CD, Woodland HR & Marbaix G (1971) Use of frog eggs and oocytes for the study of messenger RNA and its translation in living cells. Nature 233, 177–182. 14 Masui Y & Markert CL (1971) Cytoplasmic control of nuclear behavior during meiotic maturation of frog oocytes. J Exp Zool 172, 121–145. 15 Luzzi V, Lee CL & Albritton NL (1997) Localized sampling of cytoplasm from Xenopus oocytes for capillary electrophoresis. Anal Chem 69, 4761–4767. 16 Lee CL, Linton J, Soughayer JS, Sims CE & Allbritton NL (1999) Localized measurement of kinase activation in oocytes of Xenopus laevis. Nat Biotechnol 17, 759–762. 17 Wagner J, Fall CP, Hong F, Sims CE, Allbritton NL, Fontanilla RA, Moraru II, Loew LM & Nuccitelli R (2003) A wave of IP3 production accompanies the fertilization Ca2+ wave in the egg of the frog, Xenopus laevis: theoretical and experimental support. Cell Calcium 35, 433–447. 18 Okuyama K, Sagata N, Shiokawa K & Yamana K (1981) Localization of poly(A) sequences that are lost during oocyte maturation of Xenopus laevis. Cell Struct Funct 6, 279–282. 19 Johnson CE, Freel CD & Kornbluth S (2010) Features of programmed cell death in intact Xenopus oocytes and early embryos revealed by near-infrared fluorescence and real-time monitoring. Cell Death Differ 17, 170–179. 20 Geib S, Sandoz G, Carlier E, Cornet V, CheynetSauvion V & De Waard M (2001) A novel Xenopus
113
A. A. Tokmakov et al.
Monitoring gene expression in a single oocyte
21
22
23
24 25
26
27
28
114
oocyte expression system based on cytoplasmic coinjection of T7-driven plasmids and purified T7-RNA polymerase. Receptors Channels 7, 331–343. Altafaj X, Joux N, Ronjat M & De Waard M (2006) Oocyte expression with injection of purified T7 RNA polymerase. Methods Mol Biol 322, 55–67. Tokmakov AA, Matsumoto E, Shirouzu M & Yokoyama S (2006) Coupled transcription-andtranslation – a method of choice for heterologous gene expression in Xenopus oocytes. J Biotechnol 122, 5–15. Sindelka R, Sidova M, Svec D & Kubusta M (2010) Spatial expression patterns in the Xenopus laevis oocytes measured with qPCR tomography. Methods 51, 87–91. Mowry KL & Cote CA (1999) RNA sorting in Xenopus oocytes and embryos. FASEB J 13, 435–445. Zhou Y & King ML (2004) Sending RNAs into the future: RNA localization and germ cell fate. IUBMB Life 56, 19–27. King ML, Messitt TJ & Mowry KL (2005) Putting RNAs in the right place at the right time: RNA localization in the frog oocyte. Biol Cell 97, 19–33. Sindelka R, Jonak J, Hands R, Bustin SA & Kubista M (2008) Intracellular expression profiles measured by real-time PCR tomography in the Xenopus laevis oocyte. Nucleic Acids Res 36, 387–392. Bengtsson M, Rorsman P, Stahlberg A & Kubista M (2005) Gene expression profiling in single cells from the pancreatic islets of Langerhans reveals lognormal distribution of mRNA levels. Genome Res 15, 1388–1392.
29 Tokmakov AA, Iguchi S, Iwasaki T & Fukami Y (2011) Unfertilized frog eggs die by apoptosis following meiotic exit. BMC Cell Biol 12, 56. 30 Pasquier DD, Dupre A & Jessus C (2011) Unfertilized Xenopus eggs die by Bad-dependent apoptosis under the control of Cdk1 and JNK. PLoS ONE 6, e23672. 31 Del Prete MJ, Robles MS, Guao A, Martınez-AC Izquierdo M & Garcia-Sanz JA (2002) Degradation of cellular mRNA is a general early apoptosis-induced event. FASEB J 16, 2003–2005. 32 Hoat TX, Nakayashiki H, Tosa Y & Mayama S (2006) Specific cleavage of ribosomal RNA and mRNA during victorin-induced apoptotic cell death in oat. Plant J 46, 922–933. 33 Rozen S & Skaletsky H (2000) Primer3 on the WWW for general users and biologists programmers. Methods Mol Biol 132, 365–386.
Supporting information Additional supporting information may be found in the online version of this article at the publisher’s web site: Fig. S1. Reproducibility of sample collection and RNA extraction. Fig. S2. Spike control and standard curves of quantitative PCR. Table S1. Intracellular events in progesterone-treated Xenopus oocytes. Table S2. Nucleotide sequences of gene-specific quantitative PCR primers used in the study.
FEBS Journal 281 (2014) 104–114 ª 2013 FEBS
