Gene regulatory networks (GRNs) that control animal development have been studied in several model organisms (Oliveri and Davidson,2004; Koide et al.,2005; Imai et al.,2006). In these GRNs, many network genes have been reported to be regulated both temporally and spatially (Fuchikami et al.,2002; Amore et al.,2003). To better understand these gene regulatory mechanisms, reporter genes driven by specific cis-regulatory sequences can be introduced into embryos, followed by an analysis of their spatiotemporal patterns of expression (Kassis,1990; Oda-Ishii and Saiga,2003).
In the sea urchins, some of the regulatory mechanisms underlying the specific expression of network genes involved in endomesoderm specification have been analyzed by means of the microinjection of reporter gene constructs (Makabe et al.,1995; Revilla-i-Domingo et al.,2004; Ransick and Davidson,2006). Exogenous reporter genes introduced into fertilized sea urchin eggs are stably integrated into chromosomal DNA (McMahon et al.,1985) and expressed during sea urchin early development (Hough-Evans et al.,1988; Arnone et al.,1997). However, the incorporation of transgenes into embryonic cells is not ubiquitous in this system and transgene expression is restricted to cells in part of the embryonic region, which results in mosaic patterns of transgene expression. Hence, improving the incorporation efficiency is necessary for high throughput analysis of cis-regulatory sequences in the sea urchin. In addition, to ensure the precise promoter-dependent expression of transgenes, it is crucial to protect them from the effects of the adjacent chromatin environment.
For this purpose, insulator sequences that serve as a boundary element between differentially regulated genes are often used (Chung et al.,1993; Gdula et al.,1996). Insulators have two conserved properties: an enhancer-blocking activity and protection from position effects (Sakamoto et al.,2006). We previously showed that a 573-bp fragment (Ars insulator) located in the upstream region of the sea urchin (Hemicentrotus pulcherrimus) arylsulfatase gene has the typical features of an insulator (Akasaka et al.,1999). In fact, the Ars insulator has been shown to protect a transgene against silencing by position effects in both sea urchin larva (Yajima et al.,2007) and mammalian cells (Hino et al.,2004; Watanabe et al.,2006).
Recently, efficient transgenesis methods using I-SceI meganuclease have been reported in fish (Thermes et al.,2002; Grabher et al.,2004) and applied to amphibia (Ogino et al.,2006; Pan et al.,2006) and ascidians (Deschet et al.,2003). I-SceI meganuclease is an intron homing enzyme isolated from Saccharomyces cerevisiae mitochondria and is suggested to facilitate the early integration of transgenes into the host genome (Grabher and Wittbrodt,2007). In our current study, we used I-SceI meganuclease-mediated transgenesis in the sea urchin embryos and found an elevation in the efficiency of transgene expression compared with conventional methods. Furthermore, we demonstrated that Ars insulator increases the efficiency of I-SceI–mediated transgenesis by increasing the copy number of the integrated transgenes.
RESULTS AND DISCUSSION
I-SceI Meganuclease-Mediated Transgenesis in the Sea Urchin
To determine the efficiency of I-SceI meganuclease-mediated transgenesis in the sea urchin, we generated pI-CMV-GFP by introducing two I-SceI recognition sequences into both ends of the CMV-GFP construct, in which GFP is driven by a CMV promoter, to facilitate ubiquitous expression in the entire region of embryos (Fig. 1A). After digestion with I-SceI, these constructs were co-injected with I-SceI into fertilized sea urchin eggs. Transgene expression was first detected in the unhatched blastulae and continued to the pluteus larva stage. In the sea urchin, it has been reported that stable incorporation of transgenes into chromosomal DNA occurs mainly at the third or fourth cleavage stages, when the linearized constructs are microinjected into fertilized sea urchin eggs (Hough-Evans et al.,1988).
Mosaic transgene expression was observed with the CMV promoter using the conventional methods (Fig. 1B–D). Compared with this mosaic expression, more prevalent transgene expression was detected in embryos co-injected with pI-CMV-GFP and I-SceI; Some embryos exhibited GFP expression in 50–100% of their cells (Fig. 1E–G; Table 1A). This result suggested that the integration event had occurred before the third cleavage stage as the number of embryos exhibiting these prevalent GFP expressions was very low when the transgene had been introduced using a conventional method.
Table 1A. Transgenesis Efficiency for Different Promoter Constructs
A. CMV promoter
GFP expressing embryos (%)
Embryos with cells expressing GFP (%)
Represents the amount of linearized DNA construct injected per embryo. The injected solutions contain not only DNA constructs but also restricted genomic DNA. The expression levels were scored at 24 hpf. GFP, green fluorescent protein.
Represents the amount of linearized DNA construct injected per embryo. The injected solutions contain not only DNA constructs but also restricted genomic DNA. The expression levels were scored at 20 hpf. PMC, primary mesenchyme cell; EC, ectoderm; EM, endomesoderm; GFP, green fluorescent protein.
Represents the amount of linearized DNA construct injected per embryo. The injected solutions contain not only DNA constructs but also restricted genomic DNA. The expression levels were scored at 40 hpf. AOE, aboral ectoderm; OE, oral ectoderm; EN, endoderm; SMC, secondary mesenchyme cell; PMC, primary mesenchyme cell.
To optimize the experimental conditions for the I-SceI system, different amounts of pI-CMV-GFP were co-injected with different amounts of I-SceI meganuclease. As shown in Table 1A, the optimum amount of injected DNA was determined to be 20 fg per sea urchin embryo, i.e., the most effective level at reducing mosaic expression although the frequency of transgene expression is comparable to conventional method (Table 1A). Injection of high amounts of plasmid caused abnormal development. Conversely, injection of pI-CMV-GFP without I-SceI exhibited low levels of efficiency, indicating that transgene incorporation is dependent upon the presence of the I-SceI enzyme (Table 1A). The optimum amounts of I-SceI meganuclease were found to be 0.5 μU per sea urchin embryo (Table 1A), and injection of high amounts of I-SceI meganuclease resulted in abnormal development.
We next tested different promoter constructs that drive the cell lineage-specific expression of GFP. This included pI-HpTb-GFP (Fig. 2A), which contains the T-brain promoter and its cis-regulatory regions (Ochiai et al.,2008), and pI-HpSM50-GFP (Fig. 2A), which contains the HpSM50 promoter (Yajima et al.,2007). Both promoters drive transgene expression specifically in the micromere descendants and primary mesenchyme cells (PMC) (HpTb-GFP in Fig. 2B–D, HpSM50 data not shown). We also examined pI-HpArs-GFP (Fig. 2A), which contains the HpArs promoter and its cis-regulatory regions (Kurita et al.,2003), and is specifically expressed in the aboral ectoderm (AOE) (Fig. 2H–J). Using the I-SceI system, transgene expression in embryos injected with these three constructs occurred in a promoter-specific manner (pI-HpTb-GFP in Fig. 2E–G, pI-HpArs-GFP in Fig. 2K–M, pI-HpSM50-GFP data not shown) although some low levels of ectopic expression were also detected (Table 1B,C).
In contrast to the experiment using pI-CMV-GFP, an elevated frequency of transgene expression was observed in each case compared with conventional methods (Table 1B,C). The number of embryos expressing GFP increased to 42%, 39%, and 46% using pI-HpTb-GFP, pI-HpSM50-GFP, and pI-Ars-GFP, respectively. We also observed an increase in the population of GFP-expressing cells in each cell lineage using these different promoter constructs, although their expression was not detected in all cells of each lineage (Fig. 2E–G,K–M). These may be the results of an earlier integration event in the I-SceI system which enabled the incorporation of the transgene into a specific cell lineage and facilitated expression. On the other hand, because transgene expression driven by the CMV promoter was detected in all cell lineages that incorporated the constructs, although the level of mosaicism differed depending on the timing of integration, the frequency seemed not to be elevated using CMV-promoter constructs. From these results, it is indicated that the I-SceI-mediated method is extremely useful for transgenesis in the sea urchin as it not only increases the frequency of cell lineage-specific transgene expression but also reduces the level of mosaicism.
Effects of the Ars Insulator in I-SceI Meganuclease-Mediated Transgenesis
We next examined the effects of an insulator that serves as a boundary element between differentially regulated genes in this I-SceI-meditated system. For this experiment, we used the Ars insulator that was identified originally from the sea urchin arylsulfatase gene and has both enhancer-blocking activity and a barrier function that protects transgenes from silencing by surrounding chromatin (Akasaka et al.,1999; Yajima et al.,2007). A pIA-promoter-GFP series (pIA-CMV-GFP, pIA-HpTb-GFP, pIA-HpSM50-GFP, and pIA-HpArs-GFP) was thus generated in which promoter-GFP constructs are flanked by two Ars insulators and two I-SceI recognition sequences (Figs. 1A, 2A). Each construct was subsequently co-injected with I-SceI into fertilized eggs.
In the sea urchin embryos injected with pIA-CMV-GFP, the frequency of transgene expression was elevated 1.3-fold and the level of mosaicism was slightly reduced compared with reporter constructs without the Ars insulator (Table 1A). Using cell lineage-specific promoter constructs, different effects of the Ars insulators could be observed using the I-SceI system. In embryos injected with pIA-HpTb-GFP, the frequency of transgene expression increased approximately 1.5-fold and the ectopic expression levels were similar to those without Ars insulator (Table 1B). In contrast, no increase in transgene expression frequency was observed for HpSM50-GFP, and ectopic expression was strongly reduced by the Ars insulator in this case (Table 1B). In embryos injected with pIA-HpArs-GFP, we observed both an increase in transgenesis efficiency and a reduction in the ectopic expression levels (Table 1C). These results indicate that the Ars insulator cannot only increase the efficiency of transgenes expression but also reduce their ectopic expression. We further observed that a reduction in ectopic expression by the Ars insulator is dependent on the transgene constructs. This is likely due to the functional differences between cis-regulatory elements in these constructs. The promoter region of the SM50 gene contains positive cis-regulatory elements that can activate the expression specifically in the PMC, but no negative cis-regulatory elements that would repress any ectopic expression (Makabe et al.,1995). Hence, a HpSM50 gene construct without an Ars insulator element could well be influenced by enhancers around the sites of integration, and the Ars insulator may block any ectopic activation to ensure transgene expression in the PMC. On the other hand, because HpTb and HpArs gene constructs possess not only positive but negative elements, ectopic activation from surrounding enhancers is likely to be repressed by these negative elements. Hence, the effects of the Ars insulator would be expected to be less efficient in HpTb and HpArs gene transgenesis compared with the HpSM50 gene.
We next examined the relative copy number of the incorporated GFP transgenes as well as the levels of the GFP mRNA by QPCR analysis of embryos injected with pI-CMV-GFP or pIA-CMV-GFP (Fig. 3). The extent of incorporated GFP transgene and its corresponding mRNA levels in embryos injected with constructs flanked by the Ars insulator were higher than those without this element (Fig. 3A,B). In addition, the relative quantity of transcripts per transgene was found to be almost equivalent for the constructs with or without the Ars insulator (Fig. 3C). These results indicate that the transgene expression in the sea urchin embryos driven by CMV promoter is not silenced by surrounding chromatin during early embryonic development. Furthermore, our data suggest that the Ars insulator increases the efficiency of transgene expression by promoting the integration of a great number of transgenes into the genome using the I-SceI systems. This suggestion is further supported by our finding that the promotion of transgene integration by the Ars insulator was detected also for HpTb promoter constructs (data not shown).
It is hypothesized that the incorporation of transgenes by means of I-SceI-mediated genomic integration is dependent upon the formation of double strand breaks (DSBs) in the injected DNA and host genome, and their eventual repair (Grabher and Wittbrodt,2007). Recently, Haviv-Chesner et al. (2007) have shown in yeast that linear fragments are incorporated by nonhomologous end joining when DSBs induced by HO endonuclease are repaired. However, it has also been reported that Su(Hw), a Drosophila insulator binding protein, alters the DSB repair frequency (Lankenau et al.,2000). Taken together, our current data and these previous findings may suggest that Ars insulator enhances the integration efficiency in the I-SceI system by affecting the DSB repair process. However, further analysis is required for the elucidation of the precise enhancement mechanism.
In summary, we show from our present data that I-SceI–mediated transgenic methodologies elevate the frequency of transgene expression and reduce the level of mosaicism in the sea urchin embryo. Furthermore, we demonstrate that I-SceI-mediated methods combined with the use of the Ars insulator sequence increase the frequency of transgene expression and reduce their ectopic expression levels. Moreover, as the Ars insulator has been reported to function in Drosophila (Akasaka et al.,1999), tobacco cells (Nagaya et al.,2001), and mammalian cells (Hino et al.,2004; Watanabe et al.,2006), the use of this element may improve transgenesis in a wide variety of species.
Sea urchins (H. pulcherrimus) were harvested from the Seto inland sea or from Tateyama Bay and their gametes were obtained by coelomic injection of 1 mM acetylcholine chloride. Fertilized eggs were cultured in filtered sea water at 16°C.
Preparation of Reporter Constructs
The pI vector, in which two I-SceI recognition sequences are inserted at both ends of the multicloning sites (MCS) of pBluescript II SK(+) (pBSK) (Stratagene, La Jolla, CA), was generated by polymerase chain reaction (PCR). Using pBSK as the template, a first round of PCR amplification and subsequent self-ligation of the PCR products was performed to generate I-SceI-pBSK using the primers 5′-ATTACCCTGTTATCCCTACACTGGCC-GTCGTTTTACAACG-3′ and 5′-TAATACGACTCACTATAGGG-3′. Using I-SceI-pBSK as a template, a second round of PCR amplification and self-ligation of the PCR product was then performed to generate the pI vector using the primers 5′-ATTACCCTGTTATCCCTATCATGGTCATAGCTGT-TTCCTG-3′ and 5′-ATTAACCCTCACTAAAGGGA-3′. It has been reported previously that a transgene flanked by two Ars insulators in opposite orientations can be stably expressed in mammalian cells (Watanabe et al.,2006). Hence, the pIA vector was prepared by insertion of two Ars insulators in opposite orientations into the XhoI and SacI sites of the pI vector. Cytomegalovirus (CMV)-GFP was prepared by insertion of the GFP gene and SV40 late poly(A) sequence from pGreenLantern-1 vector (GIBCO BRL, Gaithersburg, MD), along with CMV enhancer and immediate early promoter from pRL-CMV (Promega, Madison, WI) into the EcoRI and PstI sites of pBSK, respectively. pI-CMV-GFP and pIA-CMV-GFP were prepared by insertion of the CMV-GFP-SV40 fragment into the MCS of the pI and pIA vector, respectively. The HpSM50 promoter fragment (EU523700) was amplified from the genomic DNA of H. pulcherrimus using the primer set 5′-CCTCGAGATCTGGAAAATAA-3′ and 5′-CCATGGTTGCTTCTGCAATG-3′. HpSM50-GFP was prepared by insertion of GFP-SV40 and the HpSM50 gene fragment corresponding to −441 to +112 into pBSK. pI-HpSM50-GFP and pIA-HpSM50-GFP were prepared by inserting the HpSM50-GFP-SV40 fragment into the MCS of pI and pIA, respectively. We also used HpTb-GFP in which the fragment of HpTb gene extending from −406 to +6703 was fused to the GFP gene in frame as described (Ochiai et al.,2008), and HpArs-GFP in which the HpArs fragment extending from −727 to +4690 was fused to GFP gene in frame, also as previously described (Kurita et al.,2003). The pI-HpTb-GFP, pIA-HpTb-GFP, pI-HpArs-GFP, and pIA-HpArs-GFP constructs were prepared by insertion of the HpTb-GFP or HpArs-GFP fragment into pI and pIA vectors.
Introduction of DNA Constructs Into Fertilized Eggs by Microinjection
Microinjection was carried out as described previously by Rast (2000) with some modifications. Using conventional transgenesis, CMV-GFP, HpSM50-GFP, HpTb-GFP, and HpArs-GFP were linearized by XhoI, SacI, or XbaI. The linearized reporter constructs were then dissolved at the appropriate concentration in a 120 mM KCl and 10% glycerol solution containing a fivefold mass excess of the sea urchin genomic DNA with compatible ends. Two pl of the linearized reporter construct solutions was subsequently injected into fertilized sea urchin eggs. For I-SceI meganuclease-mediated transgenesis, plasmid DNA was digested with I-SceI meganuclease (NEB, Ipswich, MA) before injection (5–20 ng/μl plasmid DNA; 1× commercial meganuclease buffer [NEB]; 0.1% bovine serum albumin; 0.125–1 U/μl I-SceI meganuclease). The reaction mixture was then incubated for 40 min at 37°C, and 2 pl of the reaction mixture were injected into fertilized sea urchin eggs. To observe the spatial patterns of GFP fluorescence, embryos were fixed in sea water containing 0.1% formaldehyde and visualized on an IX-81 fluorescence microscope (Olympus, Tokyo, Japan). Images were acquired using MetaMorph software (Universal Imaging, West Chester, PA).
Quantification of GFP Reporter DNA and mRNA in Microinjected Embryos
To quantify the levels of transcription from a single copy of an injected reporter construct, genomic DNA and total RNA were extracted from 100 to 150 microinjected embryos by ISOGEN (NIPPON GENE, Tokyo, Japan) as described by the manufacturer. Total RNA was treated with DNaseI and reverse-transcribed with ThermoScrip RT-PCR System (Invitrogen, Carlsbad, CA) as described by the manufacturer. Quantities of genomic DNA and cDNA were determined by Quantitative PCR (QPCR) according to an experimental procedure described by Revilla-i-Domingo et al. (2004) with some modifications. Briefly, to determine the copy number of reporter constructs integrated into the genome, a QPCR was carried out using primers designed to amplify a 150- to 200-bp fragment of the coding sequence of GFP (GFP primer set), the coding sequence of the Ars gene (Ars primer set), and a genomic DNA template. PCR amplification of each gene fragment was analyzed using a Mx3000P Real-Time PCR System (Stratagene) using FullVelocity SYBR Green QPCR Master Mix (Stratagene). Reactions were done in triplicate with samples from five embryos. Thermal cycling parameters were 40 cycles at 95°C for 10 sec followed by 60°C for 30 sec. The copy number of GFP reporter construct integrated into a single genome was estimated by dividing the quantity of integrated GFP construct by that of Ars gene, which is known to be a single copy gene in the H. pulcherrimus haploid genome (unpublished data). To determine the quantity of reporter mRNA, a QPCR was conducted using GFP and MtCOI primer sets with a cDNA template. To obtain the relative amounts of GFP mRNA per embryo, the quantity of GFP reporter mRNA was divided by the quantity of MtCOI mRNA, which is known to be expressed constitutively during the developmental process (Okabayashi and Nakano,1983). Ultimately, the relative amount of GFP mRNA per embryo was normalized by the copy number of GFP reporter construct per genome. The nucleotide sequences of primer sets used in QPCR experiment are described in Ochiai et al. (2008).
Data are presented as mean values ± SD. Differences were evaluated for significance by the Student's t-test, with significance defined as P < 0.01 using a two-tailed unpaired t-test.
The authors thank Dr. Kiyomoto (Tateyama Marine Laboratory, Ochanomizu University) for supplying live sea urchins. The authors also thank the Fisheries and Ocean Technology Center, Hiroshima Prefectural Technology Research Institute for supplying sea water. T.Y. was funded by a Grant-in-Aid for Scientific Research (C) from the Ministry of Education, Science, Sports and Culture, Japan.