Generation of NSE-MerCreMer Transgenic Mice with Tamoxifen Inducible Cre Activity in Neurons

To establish a genetic tool for conditional deletion or expression of gene in neurons in a temporally controlled manner, we generated a transgenic mouse (NSE-MerCreMer), which expressed a tamoxifen inducible type of Cre recombinase specifically in neurons. The tamoxifen inducible Cre recombinase (MerCreMer) is a fusion protein containing Cre recombinase with two modified estrogen receptor ligand binding domains at both ends, and is driven by the neural-specific rat neural specific enolase (NSE) promoter. A total of two transgenic lines were established, and expression of MerCreMer in neurons of the central and enteric nervous systems was confirmed. Transcript of MerCreMer was detected in several non-neural tissues such as heart, liver, and kidney in these lines. In the background of the Cre reporter mouse strain Rosa26R, Cre recombinase activity was inducible in neurons of adult NSE-MerCreMer mice treated with tamoxifen by intragastric gavage, but not in those fed with corn oil only. We conclude that NSE-MerCreMer lines will be useful for studying gene functions in neurons for the conditions that Cre-mediated recombination resulting in embryonic lethality, which precludes investigation of gene functions in neurons through later stages of development and in adult.


Introduction
Conventional gene knockout in mice can result in early lethality, which prevents study of gene functions through later stages of development and in adult. To circumvent early lethality, the use of Cre/loxP system to generate conditional gene deletion in a spatial and temporal manner has become the most popular method. Cre recombinase is a Type I topoisomerase from bacteriophage P1 that catalyzes the site-specific recombination of DNA between loxP elements in both bacteriophage and mammalian cells [1,2,3]. The loxP recognition element is a 34 base pair (bp) DNA sequence comprised of two 13 bp inverted repeats flanking an 8 bp spacer region which confers directionality [4]. Cre recombination could result in excision, inversion or translocation of the floxed (flanked with two loxP sites) genomic regions depending on the locations and orientations of the loxP elements [3,5]. By utilization of a specific promoter driving a spatially restricted Cre expression, conditional deletion of a gene or activation of transgene expression can be achieved in specific tissue in mice, thus offering the opportunity to study gene function with spatial control [3,6]. A further refinement of the Cre/loxP technology is the development of inducible Cre transgene which permits temporal control of gene recombination, allowing investigation of gene functions in a particular developmental stage of the entire life-span of mice.
The inducible Cre recombinase is consisted of mutated ligandbinding domain (LBD) of the mouse estrogen or progesterone receptor and Cre recombinase [3,7,8]. The mutated LBD fails to bind to estrogen or progesterone, but retains its ability in binding to synthetic ligands such as tamoxifen, 4-OHT (4-hydroxytamoxifen) and RU486 [3,7,9]. When ligand is absent, LBD-Cre is bound by HSP90 and retained in the cytoplasm. Upon ligand binding, LBD-Cre translocates into the nucleus and mediates genomic recombination [3,7,9,10]. Therefore, Cre-mediated recombination is induced by the administration of synthetic ligand, allowing a temporal control of the recombination event [3,11]. The tamoxifen inducible Cre recombinase protein (MerCreMer) is composed of (i) Cre recombinase (Cre) and (ii) two tamoxifen-binding domains (Mer) of mutated mouse estrogen receptor a (ERa), one at each end of the Cre recombinase, ensuring efficient binding of MerCreMer to tamoxifen and 4-OHT, but at the same time retaining maximal Cre activity [9].
Neural specific enolase (NSE) is a glycolytic enzyme enolase abundantly but specifically expressed in terminally differentiated neurons and neuroendocrine cells [12]. Transcript of mouse NSE was detectable from E12 onwards [13], and its expression was correlated with synaptogenesis [12,14]. The 1.8 kb rat NSE promoter DNA fragment has been shown to drive expression of target genes in brain neurons of the transgenic mice [13,15,16,17,18,19]. A NSE-Cre mouse line has been previously generated which exhibited spatially restricted Cre activity in neurons of the central nervous system [18]. However, such mouse line did not allow temporal control of Cre recombination.
In this study, we generated transgenic mouse line (NSE-MerCreMer) that expressed tamoxifen inducible Cre activity in neurons. We cloned the rat NSE promoter 59 to the cDNA encoding the MerCreMer and generated transgenic mouse lines that expressed tamoxifen inducible Cre activity in neurons. RT-PCR, Western blot analysis, immuno-histochemistry for MerCreMer were performed to investigate the temporal and spatial expression patterns of MerCreMer in different transgenic lines. Furthermore, we crossed our mouse lines to reporter mice Rosa26R (R26R) [20]. Immuno-staining for b-galactosidase in NSE-MerCreMer/R26R mice showed that Cre activity in neurons of the central and enteric nervous systems was induced by tamoxifen.

Establishment of NSE-MerCreMer Transgenic Mice
Transgenic construct NSE-MerCreMer was constructed by linking the rat NSE promoter 59 upstream of the cDNA encoding the tamoxifen inducible Cre recombinase (MerCreMer), which is composed of Cre recombinase and mutated ligand-binding domain (LBD) of the mouse estrogen receptor a (ERa) (Mer) on both ends (Figure 1). MerCreMer will only translocate into the nucleus and mediate genomic recombination in the presence of synthetic ligands such as tamoxifen and 4-OHT [3,7,9,10]. To ascertain the tamoxifen induction of Cre activity of NSE-MerCreMer, the transgenic construct was co-transfected together with the pCAG-CAT-LacZ plasmid into HeLa cells. The pCAG-CAT-LacZ plasmid carries a LacZ (b-galactosidase) gene downstream of a chicken beta-actin promoter (CAG) and a 'DNA stuffer' (CAT) flanked by two loxP sequences, so that LacZ is expressed only when the DNA stuffer is removed by the action of Cre recombinase. Addition of 4-OHT (0.8 mM; +OHT) induced a robust expression of b-galactosidase of the co-transfected cells. In contrast, no bgalactosidase could be detected in co-transfected cells without 4-OHT treatment (2OHT) (Figure 1). These data indicated that Cre activity was successfully induced in HeLa cells transfected with the NSE-MerCreMer and the pCAG-CAT-LacZ vectors upon treatment with 4-OHT.
To generate transgenic mice that express Cre recombinase in neurons, NSE-MerCreMer was micro-injected into fertilized eggs. A total of 54 mice were born from the micro-injection, and six transgenic founders (4 males and 2 females) were identified by PCR genotyping ( Figure 1B). All the six founders transmitted the transgene through germ-line, and 6 transgenic lines were established (Table 1). Brain and intestine contain neurons of the central nervous system (CNS) and the enteric nervous system (ENS), respectively. Therefore, brain and intestine were collected from the F3 transgenic mice (postnatal week-4) of each line, and expression of the transgene was analyzed by RT-PCR and Western blot. Transcript of MerCreMer was detected in four transgenic lines by RT-PCR ( Figure 1C). Expression of MerCre-Mer protein in transgenic mice was further confirmed by Western blot using the monoclonal anti-body (anti-ERa), which recognized the ligand binding domain of the estrogen receptor a. Therefore, anti-ERa could recognize both the endogenous ERa and the mutated ERa ligand binding domain in MerCreMer. As shown in Figure 1D, strong bands were detected in uterus and intestine, corresponding to endogenous ERa (67 kDa) and MerCreMer protein (113 kDa), respectively in two transgenic lines (#778, #805). A very weak band (corresponding to 113 kDa) was also observed in brain of mice in transgenic lines (#778, #805) upon prolonged exposure (data not shown). However, MerCreMer protein expression was not detected in the intestine and brain of the other 2 transgenic lines that were found expressing the MerCreMer transcript by RT-PCR (date not shown). Expression of MerCreMer protein in the neurons of the enteric nervous system (ENS) of the intestine in transgenic mice (from lines #778 and #805) was investigated by immuno-fluorescence using anti-bodies against estrogen receptor a (ERa) and Tuj1 (neuron marker). Immuno-reactivity of Tuj1 (green) was localized to the neurons and nerve fiber of the ENS of transgenic and nontransgenic intestines ( Figure 2C, D and data not shown). Immuno-reactivity of ERa (red) was detectable at the myenteric ganglion plexus between the circular and the longitudinal muscle layers of the intestine of transgenic mice ( Figure 2F). Furthermore, ERa immuno-reactivity was localized to the cytoplasm overlapping with that of neuron marker Tuj1 ( Figure 2H) in transgenic intestine, indicating the cytoplasmic expression of MerCreMer in neurons. The absence of ERa immuno-reactivity in non-transgenic control Figure 2. Localization of MerCreMer protein to the enteric neurons in the intestine of NSE-MerCreMer transgenic mice. Immunoreactivity for Tuj1 (green; neuronal marker) was localized to the enteric ganglion plexus at the myenteric region between the circular muscle and the longitudinal muscle of the intestine of non-transgenic mice (C), and transgenic mice (D). In contrast, immuno-reactivity for ERa (red) was only detected at the myenteric plexus in the intestine of transgenic mice (F), but not in non-transgenic mice (E). Superimposed photos of immunofluorescence for ERa (red) and Tuj1 (green) of transgenic intestine (H), and non-transgenic intestine (G) showed co-localization of ERa and Tuj1 immuno-reactivity in transgenic intestine. Dotted regions are magnified and showed as insets. Abbreviations: m, mucosa; cm, circular muscle; lm, longitudinal muscle. doi:10.1371/journal.pone.0035799.g002 ( Figure 2E), further confirmed the specificity of the ERa antibody for the detection of MerCreMer transgenic protein in the transgenic intestine.

Temporal and Spatial Expression of MerCreMer in NSE-MerCreMer Mice
RT-PCR analysis was performed on RNAs isolated from various tissues of postnatal week-4 transgenic mice to investigate if MerCreMer transcript was expressed in other tissues besides the central and the enteric nervous systems. High level expression of MerCreMer was observed in the brain and intestine of transgenic mice, nevertheless, weak to moderate expressions of MerCreMer transgene were detectable in the heart and liver of transgenic mice from line #778; and in the heart, liver and kidney of mice from line #805 ( Figure 3A). In line with low level expression of transgene in non-neural tissues as observed in our transgenic mice, the same NSE promoter has been previously shown to direct weak level of transgene expression in non-neural tissues [13,15,16,18].
In mouse brain, neurogenesis occurs from E12 to E17 [21,22]. In mouse intestine, ENS progenitor cells migrate along the gut, proliferate and differentiate into neurons and glia between E10.5 and E14.5 [23]. The expression of MerCreMer in brain and intestine of transgenic embryos between E12.5 and E16.5 was evaluated by RT-PCR to determine if the onset of the NSE-MerCreMer expression matched with the neurogenesis of the CNS and ENS in embryonic stages. The earliest detectable expression of MerCreMer transcript in brain and intestine was at E14.5 for transgenic line #778 ( Figure 3B), which correlated with the neurogenesis of CNS and ENS. However, for transgenic line #805, expression of MerCreMer in brain and intestine was not detectable till E16.5 ( Figure 3B). Transcript of mouse endogenous NSE was detectable from E12 onwards [13] and its expression was correlated with synaptogenesis [12,14]. However, the same NSE promoter has been shown to direct transgene expression in the embryonic brain as early as E9.5 [16]. The variations in the embryonic onset of transgene expression among different transgenic lines could be attributable to the (i) influences of the genetic loci at which the transgene integrated in different transgenic lines, and/or (ii) number of copies of the transgene integrated into the genome.

Tamoxifen Inducible Cre Recombinase Activity in Neurons
NSE-MerCreMer transgenic mice (from lines #778 and #805) were crossed with Rosa26R Cre-dependent lacZ reporter mice (R26R) (Soriano, 1999) to generate transgenic mice carrying both the NSE-MerCreMer and R26R transgenic loci (NSE-MerCreMer/ R26R). Double transgenic mice (postnatal week-4) were divided into two groups (experimental group and control group). For the experimental group, mice were given tamoxifen via intragastric gavage for 4 consecutive days. Corn oil was given to mice in the control group. Mice of both groups were sacrificed 3 days after the last administration, brain and intestine were processed to assay the tamoxifen inducible Cre activity in neurons of NSE-MerCreMer/ R26R transgenic adults by Western blot analysis and immunohistochemistry for b-galactosidase ( Figure 4).
As shown in Figure 5, Western blot analysis with anti-bgalactosidase serum demonstrated the presence of b-galactosidase in the brain, spinal cord and small intestine of NSE-MerCreMer/ R26R transgenic mice only after tamoxifen induction, but bgalactosidase was not detected in these tissues of the corn oil controls. A weak expression of b-galactosidase was detected in the brain of R26R mice, which indicated a low level of leaky expression of the b-galactosidase from the Rosa26R locus.
Spatial expression of b-galactosidase in the brain of tamoxifen treated NSE-MerCreMer/R26R double transgenic mice from both transgenic lines was investigated by co-immuno-fluorescence using anti-sera for b-galactosidase and NeuN (neuronal marker). Corn oil treated NSE-MerCreMer/R26R mice and R26R mice were included as controls to ascertain the specificity of the Cremediated induction of b-galactosidase upon tamoxifen administration in NSE-MerCreMer/R26R mice. Coronal sections from the forebrain, rostral and caudal midbrain, and cerebellum of NSE-  MerCreMer/R26R mice (both lines #778 and #805; tamoxifen and corn oil treated; at least two mice from each line were analyzed in different treatment groups) and R26R mice were co-immunostained for b-galactosidase (green) and NeuN (red). In tamoxifen treated NSE-MerCreMer/R26R mice from both lines, majority of the neurons (NeuN immuno-positive; red) at various regions of the cerebral cortex from the forebrain ( Figure 6A), rostral ( Figure 6B) and caudal midbrain ( Figure 6C) were also immuno-positive for bgalactosidase. In addition, neurons at the hippocampus, thalamus and hypothalamus were also immuno-positive for b-galactosidase in these tamoxifen treated NSE-MerCreMer/R26R mice. In corn oil treated NSE-MerCreMer/R26R mouse brains, with the exception that few neurons at the forebrain cerebral cortex of line #805 expressed very low level of b-galactosidase (arrows; Figure 6A), neurons at all the other regions of the brain were not immunoreactive for b-galactosidase. No b-galactosidase immuno-reactivity was detectable in all the brain sections of R26R mice ( Figure 6A-C). The weak expression of b-galactosidase in the forebrain cerebral cortex of corn oil treated NSE-MerCreMer/R26R mice from line #805 was probably attributed to low level of Cre activity in the forebrain region of the mouse line #805. Purkinje cells in the cortex of the cerebellum of the brain from the tamoxifen treated NSE-MerCreMer/R26R mice from both lines were also immuno-positive for b-galactosidase (arrowheads; Figure 6D). In contrast, Purkinje cells in the cerebellum of corn oil treated NSE-MerCreMer/R26R mice and R26R mice were not immuno-reactive for b-galactosidase.
In the intestine, expression of b-galactosidase (green) was specifically localized to the myenteric ganglion plexus of tamoxifen treated NSE-MerCreMer/R26R double transgenic mice ( Figure 7A, C). Very weak or a complete absence of b-galactosidase immunoreactivity was detected at the myenteric ganglion plexus in the corn oil treated double transgenic intestine from transgenic line #778 ( Figure 7B), and #805 ( Figure 7D).
RT-PCR revealed weak to moderate expression of MerCreMer transgene in some of the non-neural tissues from line #778 and line #805 (Figure 3A), immuno-fluorescence staining for bgalactosidase were performed on sections prepared from these non-neural tissues of NSE-MerCreMer/R26R mice (both lines #778 and #805; tamoxifen and corn oil treated). No specific staining could be detected in these non-neural tissues from either tamoxifen or corn oil treated NSE-MerCreMer/R26R mice (data not shown).
In conclusion, we established transgenic mouse lines NSE-MerCreMer that express tamoxifen inducible Cre activity, which allow a temporal control of the Cre/loxP recombination event in neurons. Our transgenic mice will be very useful in studying loss of gene function (Cre-mediated gene deletion) and/or gain of gene function (Cre-mediated gene expression) by Cre-mediated recombination in neurons, which results in embryonic lethality and precludes investigation of gene functions in neurons through later stages of development and in adult. Furthermore, as a gene expressed in neurons at different developmental stages may play distinct roles, NSE-MerCreMer mice allow investigation of different functions of a gene in neurons at different stages of nervous system development, providing critical information on the pathogenesis of neurodegeneration in normal aging and pathological conditions, which has been obsessing many individuals and levying a heavy burden to society.

Generation of Transgenic Construct NSE-MerCreMer
The 1.8 kb promoter of the rat neural specific enolase (NSE) gene was subcloned from the NSE-Cre transgenic construct [16] 59 upstream of the cDNA encoding the MerCreMer protein [9]. The insert of the transgenic construct NSE-MerCreMer was completely sequenced to ascertain that no mutation was introduced during the cloning steps.

Generation of Transgenic Line
The DNA insert of NSE-MerCreMer construct was released by NotI digestion, purified and micro-injected into fertilized oocytes collected from super-ovulated (CBA/129) F1 hybrid mouse. The micro-injected fertilized eggs were transferred to pseudo-pregnant wild type ICR foster mothers. NSE-MerCreMer transgenic founder mice were backcrossed to (CBA/129) F1 hybrid mice for germ-line transmission to establish transgenic lines. All experimental procedures were approved by the Committee on the Use of Live Animals at the University of Hong Kong (approval CULTRA 1044-05).

Genotyping of Transgenic Mice
Genomic DNA was extracted from 2 mm tail clip using PBND extraction method (http://www.jax.org/imr/tail_nonorg.html). In brief, mouse-tail was digested with 40 mg Proteinase K (Invitrogen TM ) in 200 ml PBND buffer at 55uC for 16 hours. After heating at 96uC for 10 minutes to inactivate the Proteinase K, the tail digest was used as template DNA for PCR analysis. For the detection of the MerCreMer transgene, the following primers were used: CreF1
To detect the expression of b-galactosidase in the brain, spinal cord and small intestine of NSE-MerCreMer mice upon tamoxifen induction, protein of each samples from week-4 NSE-MerCreMer mice (tamoxifen or corn oil treated; at least two mice were analyzed from each line in each treatment groups) and Rosa26R mice were prepared as described above. Samples from the week-4 Z/EG mice (Tg(ACTB-Bgeo/GFP)21Lbe/J; The Jackson Laboratory) that expressed b-galactosidase constitutively were included as a positive control. Protein (100 mg) of each samples were separated by electrophoresis in 6% (w/v) polyacrylamide gel, and were electro-transferred onto PVDF membrane. After blocking in TBS-T with 5% non-fat milk (w/v) for 2.5 hours at room temperature, the membrane was incubated with anti-b-galactosidase serum (1:5000 dilution; ab616; Abcam) for 16 hours at 4uC. After TBS-T rinse, and an incubation in HRP-conjugated secondary antibody (1:5000 dilution; P0399; Dako; 1 hour, room temperature), the membrane was washed in TBS-T. Signal was visualized by chemiluminescence (GE Health Amersham ECL Plus Western Blotting Detection System) and exposure to X-ray film (Kodak). Membrane was stripped and reprobed with anti-bactin serum as described above. Figure 6. Tamoxifen induction of b-galactosidase in neurons of the central nervous system. Coronal sections of the forebrain (A), rostral (B) and caudal (C) midbrain, and cerebellum (D) of NSE-MerCreMer/R26R double transgenic mice of lines #778 and #805 fed with tamoxifen (TM) or corn oil (CO) and Rosa26R (R26R) mice were analyzed by co-immunofluorescence staining for b-galactosidase (green) and NeuN (red). Superimposed photos of immuno-fluorescence for b-galactosidase (green) and NeuN (red) showed co-localization of b-galactosidase and NeuN immuno-reactivity in neurons in tamoxifen fed NSE-MerCreMer/R26R mice. Arrowheads indicated the Purkinje cells in the cortex of the cerebellum (D). Arrows indicated that few neurons in the corn oil treated forebrain cerebral cortex expressed weak b-galactosidase (A). Dotted square indicated the plans of the coronal sections being analyzed from different regions of the brain. The respective locations of the superimposed photos being taken from the sections were indicated by letters on the drawing of the sections. At least two NSE-MerCreMer/R26R mice from each line were analyzed for each treatment groups. doi:10.1371/journal.pone.0035799.g006

RT-PCR Analysis
Total RNA was isolated using TRIzol (Invitogen TM ) according to manufacturer's protocol. RNA (1 mg) was treated with RQ1 RNase-Free Dnase (Promega) to remove genomic DNA. Firststrand cDNA was synthesized according to the protocol provided by the Reverse Transcription System (Promega). Transcript of MerCreMer was detected by PCR using CreR1 and CreF1 primers pair. PCR amplification for mouse b-actin using b-actin specific primers: b-actinF1 (59-GAG AGG GAA ATC GTG CGT GAC-39) and b-actinR1 (59-AGC TCA GTA ACA GTC CGC CTA-39) was included as internal control to test the integrity of RNA. PCR amplification was performed as follow: 94 o C for 7 minutes, 40 cycles of (94uC for 45 seconds, 55uC for 45 seconds, 72uC for 45 seconds), 72 o C for 10 minutes. Amplification of b-actin generated a 534 bp fragment.

Immuno-fluorescence
Brain and intestine were fixed in 4% PFA in PBS at 4 o C for overnight. After PBS (Phosphate buffered saline pH 7.4) rinse, the specimens were incubated in 10% sucrose solution for 1 hour at 4 o C and then 30% sucrose solution at 4 o C for 24 hours. The specimens were embedded and frozen in O.C.T. (Tissue-Tek). Transverse sections (6 mm in thickness) were prepared and mounted onto microscopic glass slides coated with TESPA (3aminopropyl-triethoxysilane; Sigma). Antigen was retrieved by incubating in 10 mM sodium citrate buffer (pH 6.0) at 95uC for 10 minutes. After blocking in PBS-T (PBS with 0.1% Triton) supplemented with 10% normal goat serum (Dako) for 1 hour at room temperature, sections were incubated with either anti Invitrogen TM ) at room temperature for 2 hours. After PBS-T wash, sections were mounted in DAPI containing anti-fade solution (Vector Laboratories., Inc.). Images were taken with Nikon Eclipse 80i microscope mounted with SPOT RT3 microscope digital camera (DIAGNOSTIC instruments, Inc.), and photos were compiled using Adobe Photoshop 7.

Tamoxifen Induction
Tamoxifen stock (20 mg/ml of corn oil) was prepared by warming tamoxifen (Sigma) in corn oil at 65 o C (protected from light) until completely dissolved. Tamoxifen (0.15 mg per gram of body weight for four consecutive days) was administered by intragastric gavage to adult mice. For control, corn oil was given instead of tamoxifen.