A Photoprotein in Mouse Embryonic Stem Cells Measures Ca2+ Mobilization in Cells and in Animals

Exogenous expression of pharmacological targets in transformed cell lines has been the traditional platform for high throughput screening of small molecules. However, exogenous expression in these cells is limited by aberrant dosage, or its toxicity, the potential lack of interaction partners, and alterations to physiology due to transformation itself. Instead, primary cells or cells differentiated from precursors are more physiological, but less amenable to exogenous expression of reporter systems. To overcome this challenge, we stably expressed c-Photina, a Ca2+-sensitive photoprotein, driven by a ubiquitous promoter in a mouse embryonic stem (mES) cell line. The same embryonic stem cell line was also used to generate a transgenic mouse that expresses c-Photina in most tissues. We show here that these cells and mice provide an efficient source of primary cells, cells differentiated from mES cells, including cardiomyocytes, neurons, astrocytes, macrophages, endothelial cells, pancreatic islet cells, stably and robustly expressing c-Photina, and may be exploited for miniaturized high throughput screening. Moreover, we provide evidence that the transgenic mice may be suitable for ex-vivo bioimaging studies in both cells and tissues.


Introduction
Movements of Ca 2+ ions are fundamental for signal transduction in cells. Because the intracellular level of Ca 2+ is highly regulated and compartmentalized, transient alterations in Ca 2+ concentration are excellent signals, and are downstream targets of G-protein coupled receptors (GPCRs), ion channels and transporters, all important examples of therapeutic targets [1]. One effective tool to measure Ca 2+ mobilization sensitively, and noninvasively, are photoproteins, which release photons upon binding to Ca 2+ (in presence of a cofactor). These photoproteins are widely used in cell-based assays for high throughput screening (HTS) as reporter genes to monitor Ca 2+ movements associated with signals [2][3][4]. The sensitive detection, virtually undetectable background and high signal to noise ratio favour photoproteins over Ca 2+ -sensitive fluorescent dyes and permit small assayvolumes [5,6]. The cofactor, coelenterazine, is added to mammalian cells expressing the photoprotein and photon emission is detected as an indicator of intracellular Ca 2+ concentration [7].
Usually, cell-based assays exploit transformed cell lines, which express both a photoprotein and a target receptor. These cell lines have been selected for limited expression of other receptors, for their easiness to be cultured, and to be expanded. However, exogenous expression of targets and the transformed environment can create artefacts of gene dosage, toxicity, or stoichiometry of the receptor target itself, when it requires assembly of multiple subunits. An alternative to transformed cell lines are primary cells, isolated from mammals. They have a more physiological environment and may express targets endogenously, but they are frequently complicated to purify and culture in sufficient numbers, and furthermore, are sometimes very difficult to transfect with (reporter) genes.
The embryonic stem cells are a possible alternative. By maintaining the self-renewal property of the undifferentiated state, they can be cultured and expanded in vitro for long periods, and they are quite easily transfected [8]. Moreover, embryonic stem cells can differentiate into virtually any cell type, resembling primary cells [9,10]. Accordingly, they offer a natural environment for the receptor targets, and they can form stoichiometrically appropriate complex targets (like multi-subunit ion channels), that are regulated natively [11].
Therefore, we generated clones of mouse embryonic stem cells expressing a photoprotein as a Ca 2+ reporter system under the control of a ubiquitous promoter. We show that multiple types of cells differentiated from one of these clones report Ca 2+ signals in response to physiological stimuli. Furthermore, we exploited these undifferentiated photoprotein mES cells to produce a transgenic mouse, which may be useful for ex vivo imaging studies and as a source of differentiated primary cells expressing a Ca 2+ reporter gene.

c-Photina Photoprotein
To identify a photoprotein which combines sensitive detection of intracellular mobilized Ca 2+ and stable expression, we performed random mutagenesis of Clytin, a natural photoprotein isolated from Clytia gregaria jellyfish (syn. with Phialidin) [12,13]. One of those modified photoproteins displayed the desired characteristics and was called c-PhotinaH. To improve the transduction efficiency in mammalian cells, the c-Photina gene was optimized for mammalian codon usage (as described for PhotinaH in [6], and then fused to a mitochondrial tag (human Cytochrome C Oxidase, subunit VIII) [14], in an expression vector that contained no antibiotic resistance gene.
To investigate the function and stability of mito c-Photina expression, we transfected CHO-K1 cells, in order to create a stable clone [6]. The cells were kept in culture for more then 7 months and 58 passages, in absence of antibiotic selective pressure, and we confirmed function regularly by stimulating with agonists for endogenous GqPCRs (Gaq Protein Coupled Receptors), whose activation induce a Ca 2+ release from internal stores through the Gaq/phospholipase C pathway. The kinetics of the bioluminescent response indicated high affinity for Ca 2+ , and stable expression over time, as specified by the consistent EC 50 for ATP ( Figure 1). This was an important feature since several mammalian cells transfected with other natural or recombinant photoproteins, tended to loose function over time in the absence of selective pressure (unpublished). The stability of mito c-Photina function made this mutant the first choice for transfection into mES cells.

Generation of a mES Cell Line Expressing a Photoprotein
We electroporated mouse ES cells with the mito c-Photina gene. After neomycin selection, 114 drug-resistant colonies were picked, and expanded on mouse embryonic fibroblasts (MEFs), and screened for the ability to emit light after functional stimulation with histamine, which is known to activate the endogenously expressed GqPCR histamine-1 receptor in mES cells [15] and consequently to raise transiently the cytoplasmic Ca 2+ concentration. A typical GqPCR-mediated response curve [6] was obtained after injection of 100 mM histamine from almost all the clones (Figure 2A).
To verify that the response amplitude correlated with amount of photoprotein, cells, after primary measurement, were lysed to expose all coelenterazine-c-Photina complexes to Ca 2+ . This response was much higher than that of all clones (data not shown) and was not always correlated to the differences in amplitude observed after GqPCR stimulation, suggesting that total coelenterazine-c-Photina reacting complex was not limiting.
The final mES mito c-Photina clone was selected from 12 high responders on the basis of different parameters: primarily the ability to respond to histamine in a dose-responsive manner normalized for cell number, secondly the total photoprotein content after cell lysis, but also the number of copies of the transgene in the host genome, karyotype, cell morphology and growth rate. The number of copies in the genome was analyzed by Southern blot and quantitative PCR analysis. Clone 29 was selected on the basis of the parameters described before (see its histamine dose-responsiveness in Figure 2B) and because it has the transgene inserted into the genome in a single copy. Further confirmation of this was the impossibility to detect signals by FISH analysis (data not shown).

mES/mito c-Photina/29 Clone Is Pluripotent
Indirect immunofluorescence assays were performed on the mES mito c-Photina clone in order to evaluate the presence of specific markers of the undifferentiated pluripotent mouse embryonic stem cells, such as the stage specific embryonic antigen-1 (SSEA-1) [16] (Figure 3A-B) and the transcription factor oct 3/4 [17] (Figure 3C-D). As shown in Figure 3 both SSEA-1 and oct 3/4 are present selectively in stem cells and not in the surrounding feeder cells. Additionally, we could detect also alkaline phosphatase activity ( Figure 3E-F), another characteristic of undifferentiated stem cells [18].
Since the ''bona fide'' demonstration of stemness is the germline transmission test, clone 29 was injected into blastocysts of pregnant host female mice. Two chimeric mice, with a high degree of chimerism (almost 100%) and male phenotypes, were obtained. When these 2 mice reached sexual maturity, they were crossed with C57BL/6 female mice and gave rise to more than 95% agouti progeny (Table 1), indicating robust germline transmission.

In Vitro Differentiation Assays Performed with mES/ mito c-Photina/29 Clone
To show that the introduction of the transgene did not influence the ''in vitro'' differentiation capabilities of the mito c-Photina/29 clone, we cultured the cells under conditions to favour either cardiomyocyte or neuronal fates, employing well described protocols including suspension protocols for embryoid body (EB) formation, and adhesion protocols [19,20] all optimized for miniaturized formats.
4.1. Cardiomyocytes. Cardiomyocytes are one of the most important cell types for drug discovery projects and the hallmark of cardiomyocytes is the Ca2+ dependent contractility and its characteristic Ca2+ channel driven depolarisation curve. In addition, the heart rate is controlled by GqPCR-dependent Ca2+ release, for example mediated by the adrenergic receptors.
EBs were formed in hanging drops for two days and then in suspension for another three days. The fifth day, EBs were plated on gelatin-coated tissue culture dishes. Within one day, we observed spontaneously pulsating cardiomyocytes. The percentage of EBs containing pulsating areas was about 80% ( Figure 4A-B, and Video S1 and Video S2). The protocol was adapted to miniaturized formats, putting exactly a single EB per well of 96 or 384 micro titre plates (MTP). The cardiomyocyte development occurred directly in the micro titre plate format, maintaining the same proportion of pulsating areas.
To verify the presence of mature cardiomyocytes we stained for specific cardiomyocyte markers, such as the transcription factor, GATA-4 ( Figure 4C), and for the cytoskeleton protein, myosin heavy chain (MHC) ( Figure 4D-E). As seen in Figures 4C, D, and E all the markers were present and showed appropriate localization demonstrating proper cardiomyocyte development.
Next, functional tests were performed by disaggregating EBs and seeding 5,000 cells/well in a 384 MTP under the same differentiation conditions. 48 hours after seeding, the cells were stimulated with standard Tyrode's buffer as control, 50 nM endothelin-1, and 100 mM norephinephrine, which are agonists for the endogenously expressed GqPCR endothelin receptors and for the a1-adrenergic receptor, respectively. The cells were also stimulated with a depolarizing solution such as 60 mM KCl able to activate the voltage-gated channels, including the Ca 2+ ones. As shown in Figure 4F, all compounds led to characteristic kinetics indicating that these cells not only express markers of mature cardiomyocytes, but are also able to respond to stimuli for endogenous GPCRs and channels.
4.2. Neurons. Neurons are another cell type in which Ca2+ plays a fundamental role in signal transduction, particularly in the release of neurotransmitters.
For neuronal differentiation, the mES cells were differentiated in monolayer. The presence of cellular processes were visible 4-5 days after plating on gelatin-coated tissue culture dishes, and their length increased over time ( Figure 5A). The presence of specific markers in these cells was investigated by immunofluorescence. The nestin staining indicated the presence of neural precursors at day 13 of differentiation ( Figure 5B), together with betaIII tubulin and MAP-2 labelling which detected differentiated neurons ( Figure 5C-D), as well as staining for GFAP ( Figure 5E) which indicated the presence of astrocytes, among the population of differentiated cells. The functionality of these cells was explored with the Lumibox luminometer. At differentiation day 13, the cells were stimulated by injecting glutamate and NMDA plus glycine, in order to investigate the glutamate receptors, or with ATP and ab methylene ATP, for purinergic receptors, or depolarized with 60 mM KCl, for the activation of voltage-gated channels, including the Ca 2+ ones. All stimulations produced the expected response curves ( Figure 5F).
To validate the reporter characteristics of c-Photina in this context, we characterized activation of a GqPCR (group I metabotropic glutamate receptor), a Ca 2+ -permeable, ligand-gated ion TRP (Transient Receptor Potential) channel (vanilloid receptor-1), and voltage-gated Ca 2+ channels, in our neuronally differentiated cells (at differentiation day 13) by comparing the photoprotein-based luminescent read-out to an acetoxymethylester-coupled dye-based (Fluo4NW) fluorescent read-out on the FLIPR tetra instrument ( Figure S1). The similar results further confirm the ability of our system to detect both extracellular and intracellular Ca 2+ influxes.
With this differentiation protocol, we expected to have a high percentage of GABA-ergic and glutamatergic neurons [20]. Since GABA is a chloride channel and glutamate receptors permit passage of not only Ca 2+ but also Na + , we analyzed the cells with the FLIPR 384 using the voltage-sensitive dye (named Membrane Potential dye) able to detect non Ca 2+ ionic flux into the cells. At differentiation day 13, after stimulation with GABA and glutamate  compounds at various concentrations, we observed a dosedependent response with both agonists ( Figure 5G-H). By using picrotoxin, a specific GABA channel antagonist, we verified that this GABA-mediated signal was mainly due to activation of the GABA channels and not of the GABA transporters (data not shown). Note that the GABA response in Figure 5G is depolarizing. This is consistent with the previous observation that GABAergic reversal potential (E GABA ) is more liable to depolarize the membrane potential in immature as compared to adult neural cells. This difference is due to a variation of the intracellular Cl 2 concentration which changes during development, due to the presence of the K + -Cl 2 cotransporter KCC2 [21].

Systematic Characterization of Ca 2+ Signals in mES As Compared to Neural Cells
Extrinsic signals from 'niche' play an important role in maintenance of multipotency in stem cells. mES cells are pluripotent and model the cell intrinsic response to signals that maintain multipotency. But the cellular pathways known to transduce these signals are few [22], and Ca 2+ signalling has been only partially characterized in embryonic stem cells [23]. Therefore, we screened undifferentiated and neuronal differentiated cells with an unbiased library of pharmacologically active compounds (LOPAC 1280TM ). This library contains all the major pharmacological target classes, including GPCRs, and ion channels active compounds. We asked which receptors and ions were functional in undifferentiated mES-c-Photina cells, as compared to neural differentiated c-Photina cells. The results were expressed as ''percent activity'' with respect to ATP (''max'' signal for the undifferentiated cells) and glutamate (''max'' signal for the differentiated cells). ATP and glutamate were selected as reference compounds, since these agonists show the highest response in the cell populations tested. The data were then analyzed with the two-sample unequal variance, one-tailed t-Student test. All the compounds showing a t-test value less than 0.05 and a percent activity mean value higher than the one showed by the mean of the min signals were selected as positives and included in table 2 (see also Table S1 and Figure S2). The data revealed very few extrinsically activated Ca 2+ signalling pathways in the undifferentiated mES cells. The Ca 2+ related signals that were present are mainly associated with histamine and purinergic compounds, indicating that histamine and purinergic receptors are present. Interestingly, one of these receptors, the adenosine 1, in the P1 purinergic class, has not been previously noted in undifferentiated mES cells. There were a number of additional active compounds able to induce intracellular Ca 2+ elevations in the neural cells. Most of these substances suggested receptors compatible with the expected prevalence of receptors in neurons. These data offer additional evidence for the robust and appropriate differentiation of the c-Photina neurons.

Photoprotein Transgenic Mouse
In addition to providing cells differentiated in culture, the mES cells can be exploited to generate a transgenic mouse, which might serve as a direct source of primary cells that could express both the photoprotein transgene, and an endogenous pharmacological target in the native physiological context. The 2 chimeric mice obtained by germline transmission (as described above) were crossed with C57BL/6 female mice and gave rise to agouti progeny. All the litters were genotyped in order to check for the presence of the transgene. As expected, half of the mice born from these crosses were heterozygous for the c-Photina photoprotein gene (32/64) ( Table 1). We named these animals PhotoTopoH mice. The heterozygous mice were crossed, in order to obtain a homozygous population. One fourth of the offspring were homozygous and phenotypically normal, demonstrating that the transgene did not disrupt any gene crucial for survival.

c-Photina Expression and Activity in PhotoTopo
To determine the c-Photina m-RNA expression profile, we performed a TaqManH qPCR analysis on transgenic and control samples. The results were normalized to the amount of 18S rRNA level ( Figure 6A). Expression was detected in most tissues and at high levels. In order to check for functional c-Photina, 8 transgenic mice containing the c-Photina gene and 6 negative mice from the same litter were sacrificed in 3 different experiments. Several tissues were removed from mice, and all samples were incubated in an isotonic solution containing coelenterazine to form the active photoprotein complex. All the isolated tissues were tested in triplicates, injecting a 1% TritonH X-100 plus 250 mM CaCl 2 solution, in order to discharged all the photoprotein-coelenterazine active complexes into an enriched Ca 2+ environment ( Figure 6B). We found a good correlation between c-Photina mRNA expression and light emission across tissues. In addition, light emission remained approximately stable in all expressing tissues of animals of 3, 6, or 10 months old (data not shown).
Furthermore, we investigated the bioavailability of coelenterazine after intravenous systemic injection via the tail vein [24]. After 3 hours, one transgenic and one non-transgenic animal were sacrificed and several tissues/organs were removed. Half of the material was tested immediately with the Lumibox luminometer after cell lysis and injection of a Ca 2+ solution ( Figure 6C). The other half of the material was incubated for another 3 hours with a solution containing coelenterazine, and tested in the same way ( Figure 6D). We found that the profile of luminescence across tissues was comparable between intravenous coelenterazine injection-in vivo formation of the photoprotein-coelenterazine complex, or after incubation ex vivo. The tissue samples with highest luminescent signals were: spleen, heart, lung, testis, kidney, and skeletal muscle ( Figure 6).

Primary Cells Cultured from PhotoTopo
8.1. Aortic endothelial cells. Aortic endothelial cells are very important cells for cardiovascular diseases, and are difficult to obtain in primary culture. Since Ca2+ has a fundamental role also in these endodermally-derived cells, we decided to isolate these cells. Seven animals (4 positives and 3 negatives) were sacrificed and aortas explanted. After plating on MatrigelTM and culture for 11 days in presence of endothelial cell growth supplement and heparin, we obtained endothelial cells [25]. The presence of these cells was demonstrated by flow cytometry and immunofluorescence analysis using von Willebrand Factor (vWF) and CD31/PECAM-1 markers. The cells were also tested functionally by seeding 50,000 cells/well in a 96 MTP and stimulated with endothelin-1 and TRAP 10 and 6 peptides, agonists for the endothelin receptor and proteinase-activated receptors, respectively, both of which are highly expressed in endothelial cells. As shown in Figure 7A-D activation of both receptors induced a Ca2+ mobilization from internal stores giving rise to typical kinetics of light emission.

Bone marrow-derived monocytes/macrophages.
The PhotoTopo mice are also a source of stem cells and precursors. To verify that the c-Photina was active in hematopoietic monocyte lineage, bone marrow-derived monocyte/macrophage precursors were isolated from the femurs of 4 positive and 3 negative transgenic mice and cultured for 10 days in the presence of M-CSF (Macrophage -Colony Stimulating Factor) [26]. We confirmed the presence of mature macrophages in cell culture, by staining for specific markers F4/80 and scavenger receptor type III (CD204) in flow cytometry and immunofluorescence analysis ( Figure 7E-G). Functional studies were performed in these cells by injecting a solution of 100 mM UTP, UDP and ATP in order to stimulate the purinergic receptors. As reported in Figure 7H, all agonists were shown to induce appropriate Ca2+-mediated light emission. 8.3. Micro-organs: beta islets. Next we analysed if c-Photina would trace the glucose-triggered and Ca2+-mediated secretion of insulin in islet cells, representing micro-organs. A pancreatic islet isolation and purification was performed from PhotoTopo animals and from negative controls. Islets were cultured overnight at 37uC, and, the following day transferred to 96 MTP (10 islets/well). After incubation with Krebs-Ringer's buffer in presence of coelenterazine, they were stimulated with 11 mM glucose in order to activate the Ca2+-mediated insulin pathway. As control, mannitol which does not induce the Ca2+mediated insulin response was injected at the same final concentration in order to maintain the same osmotic concentration ( Figure 7I). We observed waves of Ca2+-mediated luminescence, induced only after stimulation with glucose and not with mannitol. The islets were then stimulated with a depolarizing agent (60 mM KCl) which induces a massive Ca2+ influx through the voltage-gated Ca2+ channels and produced robust light emission ( Figure 7J).

Ex Vivo Bioimaging from the PhotoTopo
To further explore the potentiality of c-Photina, we performed luminescence-based bioimaging studies on pancreatic islets. In order to detect topographical light emission in islets, we exploited a microscope-based device, equipped with an intensified CMOS camera (Photron Fastcam). Subsequent to a polarized injection of 60 mM KCl solution, we observed light emission representing Ca 2+ moving across the entire beta islet ( Figure 8A-B-E and Video S3). After acquisition, it was possible to retrieve either the response kinetics recorded from the whole islet ( Figure 8C), or define the response kinetics from single areas ( Figure 8D).

Discussion
The generation of a pluripotent embryonic stem cell line containing a Ca 2+ -activated photoprotein offers many opportunities to study Ca 2+ -based signals. In fact, we demonstrated that c-Photina stem cells can be differentiated into two specific cell types and can be used as source of multiple primary-like cells for Ca 2+ functional studies. Furthermore their pluripotency allowed the generation of transgenic mice, which can be an interesting reserve of cells, such as the adult stem cells (for example haematopoietic stem cells), committed progenitors, and also primary cells containing the photoprotein, for pharmacological or bioimaging studies. Interestingly, the animals can additionally be crossed with other animal models, in order to exploit these possibilities in the context of disease.
Monitoring Ca 2+ signalling with a Ca 2+ -sensitive reporter gene in mES cells, primary cells, and in a whole animal model, opens many opportunities to understand the development, function, and plasticity of many crucial Ca 2+ -mediated pathways. Several techniques have been described for measuring intracellular Ca 2+ . Patch-clamp and Ca 2+ selective microelectrodes allow quantitative measurements of Ca 2+ fluxes in single-cell analysis. These ionselective microelectrodes (ISMs) are highly sensitive and selective, but suffer from a slow response time, and high levels of noise [27]. Furthermore this technology can be applied only to a restricted number of cells. On the other hand, large populations of cells can be investigated for intracellular Ca 2+ dynamics with fluorescent probes [28]. In addition to fluorescent dyes there are also genetic tools, which provide fluorescent-based methods for Ca 2+ monitoring, and are basically divided in two groups. The first category uses the principle of fluorescence resonance energy transfer (FRET) between two variants of the green fluorescent protein (GFP), covalently linked with Ca 2+ binding proteins like calmodulin [29,30]. The second category is composed of bioluminescent proteins such as aequorin [31] fused with a GFP molecule. This latter approach is used both in single cells and in transgenic animals [32,33]. The configuration of the construct with the fusion of the two proteins allows intramolecular chemiluminescence resonance energy transfer (CRET). In fact, after Ca 2+ binding, aequorin, in presence of coelenterazine, emits a quantum of light that is transferred to GFP, which works as an acceptor and emits green light [32]. The advantage of using CRET approach, instead of using only the aequorin bioluminescent signal, is to overcome the low light quantum yield of the photoprotein. The combination of the two proteins in CRET, in fact, permits detection of the Ca 2+ -mediated signal, even with unavoidable loss of energy during the transfer.
Here we reported the development of a more direct and simpler system to measure efficiently Ca 2+ movements without any loss of energy during the transfer from the photoprotein to the GFP proteins. To do this we transfected directly only the photoprotein gene in mES cells and then we used these cells to develop a transgenic animal. The choice to use only the bioluminescent reporter gene avoids fusion proteins that could induce, even in presence of tethers, problems of folding and translation. This might be of particular relevance if we consider the larger dimension of the fusion construct compared with the single photoprotein gene. We strategically used the new photoprotein c-Photina and we demonstrated its successful application to detection of Ca 2+ -induced quantum of light using optical systems both in stem cells and in organs from transgenic animals.
An important additional feature of the mitochondrial tagged c-Photina photoprotein is its cellular stability, both in mammalian cells even in absence of a selective pressure, and in an embryonic stem cell line, and in a transgenic mouse model. This is not trivial since from our unpublished observation the expression of many natural and recombinant photoproteins often decreases during time, especially in the absence of selective pressure. The photoprotein was targeted to mitochondria, given the crucial role of Ca 2+ homeostasis in them. In fact, besides a central function in cell energy metabolism, mitochondria are able to modulate cytosolic Ca 2+ concentration and participate in Ca 2+ signalling. Moreover, mitochondrial Ca 2+ uptake is a phenomenon involved not only after Ca 2+ release from intracellular stores, as happens after stimulation of GqPCRs, but also after Ca 2+ influx through specific Ca 2+ channels from extracellular space [34,35],(for review see [36,37]). Accordingly, our mitochondrial tagged photoprotein actually efficiently recorded cytoplasmic Ca 2+ variations, after activation of both Ca 2+ channels, such as the voltage-gated Ca 2+ channels, P2X purinergic channels, and TRP channels such as the vanilloid receptors.
We demonstrated that the presence of the photoprotein does not interfere with the pluripotency of mES cells. In fact, these cells still express stemness markers like oct 3/4 transcription factor and the SSEA-1 surface antigen and possess alkaline phosphatase activities [16][17][18]. Furthermore, the c-Photina mES cells, when injected into the blastocyst of a recipient surrogate mother, gave rise to germline transmission with a high efficiency ( Table 1). The pluripotency of the c-Photina stem cells was also confirmed by the ability of the c-Photina mES cells to differentiate in vitro into cell types derived from different germ layers, such as cardiomyocytes and neurons.
The possibility to generate primary cells such as cardiomyocytes and neurons containing a bioluminescent system for monitoring Ca 2+ movements finds many applications. Since, Ca 2+ has a crucial role, for example, in controlling cardiac rhythm, one could investigate Ca 2+ behaviour in many models of different cardiac diseases. Also in neurons it could be interesting to visualize Ca 2+ movements, in normal development or during neurodegeneration, or address synaptic function. We optimized the differentiation protocols also in miniaturized formats, like the 96 and the 384 MTP, sine qua non condition for the feasibility of their usage in high throughput screening. Cells differentiated in these formats showed all the characteristics of primary-like cells, such as the expression of cell-specific targets (demonstrated by immunofluorescence analysis) or the ability to spontaneously pulse or to respond to agonist for receptors highly expressed in these cells. Moreover, the differentiation processes were shown not to interfere with the expression of the reporter gene. The functional presence of the photoprotein in the differentiated cells was demonstrated by the ability of the cells to respond to different stimuli, inducing Ca 2+ movements, confirming their utilization suitable for the development of cell-based assays.
This approach was exploited to address which Ca 2+ receptors or channels might have a functional role in mouse embryonic stem cells. We screened an unbiased library of 1280 pharmacological active compound (LOPAC 1280 TM ) modulating all of the major classes of important receptors and channels. This signalomic approach allowed high throughput identification of all cell surface receptors and channels whose activation induce a variation of intracellular Ca 2+ in undifferentiated mES cells. As a positive control for known Ca 2+ signalling a parallel approach was performed on neuronally-differentiated cells. The results indicated sparse activation of Ca 2+ response in the undifferentiated mES cells. The only classes of receptors that were shown to be activated were the one of histamine-1 receptor (H 1 ) and those of purinergic receptors, both already described in literature [15,23,38]. The presence of the H 1 receptor in these cells suggests a role of histamine in early mammalian development. Furthermore, histamine was shown to have a role in regulating neural stem cells proliferation and the expression of the H 1 receptor was shown to favour neuronal fate [39]. Also the presence of P2 purinergic receptors was already described in mouse embryonic stem cells. In particular it was proposed a role of extracellular ATP in stimulating mouse embryonic stem cell proliferation [38]. Here the authors revealed the presence of P1 purinergic receptors too by the functional response to 2-chloroadenosine only in undifferentiated and not in neuronally-differentiated cells. Its activation suggests that the role of the P1 receptor in pluripotent mouse embryonic stem cells should be further investigated.
Thanks to germline transmission of c-Photina mES cells, a transgenic animal containing the c-Photina photoprotein (Photo-Topo) was derived. The cells cultured from photoprotein transgenic animals can be used as positive controls for the ''primary-like'' cells obtained after differentiation of mES cells, and directly as primary cells, for a pharmacological screening process per se. As proof of principle, we isolated from PhotoTopo animals primary endothelial cells and hematopoietic precursors, which we differentiated into bone marrow-derived monocytes/ macrophages. We demonstrated that these cells contain the photoprotein and that they can be used in miniaturized format for functional assays. Moreover, the organism-wide expression of the transgene in PhotoTopo mice ( Figure 6A-B) suggests the availability of a larger spectrum of useful primary cells. Interestingly, the expression of the reporter gene was shown not to decrease with animal age, further confirming the stability of this photoprotein and indicating that the transgene is not inserted in a position subject to chromatin inactivation over the course of time.
The c-Photina transgenic mouse, can be used in combination with optical microscope systems, like CCD cameras, which are able to detect in real time light emission of bioluminescent reporter within the animal's cells. This application opens the possibility to charge the photoprotein by systemic injection of coelenterazine, representing a suitable model for monitoring modulation of intracellular Ca 2+ levels, and for the generation of in vivo bioluminescence imaging (BLI) based studies. As proof of principle we isolated PhotoTopo pancreatic islets since they are a perfect source of material for studying complex Ca 2+ exchanges, occurring between different cell types. In fact they are multicellular structures, in which Ca 2+ plays a fundamental role in insulin secretion. Actually, the entrance of glucose through the type 2 glucose transporters induces the activation of voltage-gated Ca 2+ channels. The consequential entry of Ca 2+ ions from the extracellular space induces insulin release from insulin-storing granules exocytosis. We demonstrated the feasibility of Ca 2+ movement observation in PhotoTopo islets, after a glucose and a depolarizing stimulus, through the photoprotein activation, not only by a CCD camerabased luminometer but also by using an intensified CMOS-based camera with 5126512 pixel resolution. The potentiality to study Ca 2+ movements in whole islets and within single cells opens very interesting potential applications for diabetes research.
mES Cell Culture TBV2 (129S2/SvPas) mouse embryonic stem cells [40] were cultured in the undifferentiated state on a monolayer of Mitomycin C treated mouse embryonic fibroblasts in the presence of leukemia-inhibiting factor (LIF) (Chemicon) [41].

Photoprotein mES Cell Clone Generation and Selection
The c-Photina gene was cloned into the pcDNA3.1+ vector (Invitrogen) downstream the mitochondrial tag (mito) of the human Cytochrome C Oxydase, subunit VIII [14]. 7610 6 mES cells were electroporated using 30 mg of the mito c-Photina DNA, linearized with BglII (New England Biolabs). Positive clones were selected with 200 mg/mL G418 (geneticin, SIGMA) [40]. Four hours before the test the medium was replaced with 50 ml/well of Tyrode's buffer (130 mM NaCl, 5 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , 5 mM NaHC0 3 and 20 mM HEPES, pH 7.4, 2 mM Ca 2+ ) and 10 mM coelenterazine, in the dark, and incubated at 37uC in a humidified atmosphere with 5% CO 2 in order to reconstitute the active photoprotein. The number of photons emitted after injection of the different ligands for 60 seconds was measured on the Lumibox CCD camera-based luminescence detector designed and built by Bayer Technologies GmbH (Wuppertal, Germany), and expressed as RLU (Relative Luminescence Units).
DNA from mES cells plated on gelatin-coated dishes was extracted with standard methods [42]. 10 mg of ES genomic DNA of ES/mito c-Photina cells was digested with the restriction enzymes, HindIII, XbaI, BamHI, HindIII/XbaI (New England Biolabs), transferred to a positively-charged nylon membrane (Roche) for Southern blot analysis. The [ 32 P]dCTP-labelled c-Photina coding sequence [42] was used as the probe. The Southern blot analysis was performed by digesting the genomic DNA with restriction enzymes which cut only once in the transfected vector (to discriminate concatamers).
All QPCR experiments were run on an ABI Prism 7700 Sequence Detector (Applied Biosystems). To calculate the number of copies we used the following formula:

Production of Transgenic Mice and Germline Transmission Test
All experiments involving animals were performed in strict accord with experimental protocols approved by the San Raffaele Institutional Animal Care and Use Committee and with Italian National Health Ministry regulations. ES 29 clone was injected into C57BL/6 blastocysts to generate chimeric mice (estimated by hair color and tail genomic DNA analysis) [40]. Two chimerae were mated with C57BL/6 mice to obtain germline transmission of the mutation (see Table 1). Genotype analysis was performed on genomic DNA prepared from tail snips using the primer pair: Primer for: AACTTCGACAACCCCAAGTG, Primer rev: TGTCGAAGATGTCGAACACG, that recognize the c-Photina coding sequence.

Cardiomyocyte Differentiation Protocol with the Embryoid Bodies (EBs) Formation Step
The cardiomyocyte differentiation protocol was obtained using hanging drops procedure, using 300 cells/drop for 2 days then putting the cells in suspension into bacterial grade dishes for other 3 days in DMEM medium plus 15% fetal calf serum (FCS) (Invitrogen), as described [19]. EBs were dissociated at differentiation day 12 with Accutase TM (Chemicon), plated in 384 MTP plates and tested 48 h after seeding at the Lumibox luminometer. Four hours before the test the medium was replaced with 25 mL/ well of Tyrode's buffer and 10 mM coelenterazine, in the dark, and incubated at 37uC in a humidified atmosphere with 5% CO 2 .

Neuronal Differentiation Protocol
The mES cells are seeded at 1,500 cells/cm 2 on gelatin-coated dishes in Knock Out-DMEM at 15% KSR (Knock-out Serum Replacement) (Invitrogen), ), at day 7 they were collected with 0.05% trypsin/EDTA solution and replated at 5,000 c/w in gelatin-coated 384 MTP dishes for in total 13-14 days with medium changing every two-three days [20]. The luminescent test was performed using the Lumibox luminometer, and incubating four hours before the differentiated cells with 25 mL/well of Tyrode's buffer and 10 mM coelenterazine, in the dark at 37uC in a humidified atmosphere with 5% CO 2 .
For the fluorescent test with the Membrane Potential dye (Molecular Devices) the differentiated cells were incubated in 25 ml/well of the dye solubilised in Tyrode's buffer for 30 min at 37uC plus 30 min at room temperature. The fluorescence signals were recorded for 250 sec and expressed as RFU (Relative Fluorescence Units) at FLIPR 384 H fluorescent reader. GABA and glutamate solution ligands were injected 3X concentrated (12.5 mL/well) at different concentration.
For the comparison between luminescent and fluorescent readouts the cells were or incubated with with 40 mL/well of Tyrode's buffer and 10 mM coelenterazine four hours before the test or with 40 ml/well of the Fluo4NW dye (Molecular Devices) solubilised in Tyrode's buffer for 60 min at 37uC. The signals were recorded at FLIPR tetra H reader, for 220 sec for fluorescence and 60 sec for luminescence read-out.
The different antibodies were incubated in 10% normal goat serum 0.1% Triton X-100 in 1X PBS.
The Hoechst 33342 dye (Invitrogen) (2 mg/mL final concentration) was incubated for 5 min at room temperature.
Images were acquired by using either an Olympus IX51 microscope equipped with a F-View II camera and the dedicated software cell-F (Olympus) or an Olympus IX70 microscope coupled to a Leica digital camera with a customized acquisition system. In some cases the brightness and/or the contrast were modified in order to reduce the background signal deriving from the white wall 384 plates used for the experiments (Matrix-Thermo Scientific 384 well plates-polystyrene white/clear).

Alkaline Phosphatase Staining
The alkaline Phosphatase activity was measured with the ELFH Phosphatase staining kit (ATCC) following manufacturer's instructions.

TaqmanH PCR
Total RNA was isolated by TRIzolH (Gibco/BRL, Gaithersburg, MD). Reverse transcription-PCR (RT-PCR) was performed with the Invitrogen Superscript II RT-PCR kit (Invitrogen), as recommended by the manufacturer. Total cDNA content normalized by simultaneous quantification (by multiplex PCR) of the 18S ribosomal RNA. All experiments were performed on an ABI Prism 7900HT Sequence Detection System (Applied Biosystems), using the comparative C T method [43]. The relative expression units (REU) were calculated as: 1 REU = 2ˆ-(average Ct Target -average Ct 18S) * 10ˆ7. The range of variation was determined by evaluating the expression: {2ˆ-[(Ct cPh -Ct 18S 6 standard deviation]}10ˆ7, where the standard deviation is calculated as !(s cPh +s 18S ).

LOPAC 1280 TM Screening
The 1280 compounds of LOPAC 1280 TM library were reformatted in 384 MTP and diluted at 50 mM in tyrode plus 2.5% DMSO (5X concentrated), the undifferentiated mES cells were seeded in triplicate on 384 gelatin-coated plates 24 h before the tests at a concentration of 20,000 cells/well.
Neural differentiated mES cells were seeded in quadruplicate at a concentration of 5,000 cells/well 6 days before the test (done at day 13 of differentiation).
Prior to the test, all the samples were incubated in Tyrode's buffer containing 10 mM coelenterazine for 3 h.
The LOPAC library was used at the final concentration of 10 mM. Negative control wells contained Tyrode's buffer plus 2.5% DMSO (that is the same concentration as test wells) (defined as ''min'' signal). The test was performed on the FLIPR tetra H instrument.
Spotfire Decision SiteH version 9.0 was used for analysis and curve-fitting of the results obtained from the activity determination experiments. The results were expressed as ''percent activity'' with respect to ATP (''max'' signal for the undifferentiated cells; final test concentration: 100 mM) and glutamate (''max'' signal for the differentiated cells; final test concentration: 100 mM). ATP and glutamate were selected as reference compounds, since these agonists show the highest response in the cell populations tested.
For differentiated cells, percent activity was computed based upon the median response value of min signal wells and glutamate wells on each plate. Then the percent activity mean and standard deviation for the quadruplicate wells were computed for each compound. For undifferentiated mES cells, percent activity was computed based upon the median response value of the test wells and ATP wells on each plate. Then the percent activity mean and standard deviation for the triplicate wells were computed for each compound.
The large symbols in Figure S2 indicate that the %Activity was statistically significant based upon a t-test, in comparison to the percent activity of min signal wells (taking all of the min signal wells as a large group). The t-test was performed using Excel, twosample unequal variance (heteroscedastic), two-tailed. The logic used to define a ''significant'' activity was:

PhotoTopo CCD Camera Functional Tests
Mice were perfused with a physiological solution and tissues were harvested and incubated in a reaction solution containing 20 mM Tris-HCl pH 7.5, 150 mM NaCl, 5 mM DTT, 1 mM EDTA, 0.1% BSA, 20 mM coelenterazine plus protease inhibitor cocktails (Roche), for 3 h at room temperature. Luminescence was determined on the Lumibox, after injection of a solution of Triton X-100 and 250 mM CaCl 2 . 300 ml of coelenterazine (373 mM coelenterazine, 3.3% DMSO, 990 nM glutathione in physiological solution or 2.8 mg of coelenterazine/kg), was injected via the tail vein. Tissues were explanted, triturated with a scissors, and divided into two batches. One part was put in the reaction solution (20 mM Tris-HCl pH 7.5, 150 mM NaCl, 5 mM DTT, 1 mM EDTA, 0.1% BSA, plus protease inhibitor cocktails) without coelenterazine and tested immediately on the Lumibox; the other part was incubated with the same reaction solution in presence of 20 mM coelenterazine for 3 h at room temperature before the Lumibox luminometer test, after injection of a solution of Triton X-100 and 250 mM CaCl 2 .

Flow Cytometry
Cells were resuspended in 100 mL of a blocking solution containing 4% FCS serum, 1 mM EDTA in PBS at a cell concentration of 1610 6 cells/mL for 10 min, then incubated with 5 mg/10 6 cells of the primary antibodies for other 30 minutes.
Cells were washed twice with blocking buffer. After 30 min of incubation with the secondary antibody and another 2 washes, cells were resuspended at 10 6 /mL in blocking buffer.
Cell acquisition was performed with FACSort Becton Dickinson in a region (R1) defined gating out only debris by forward and side scatter characteristics. 40,000 gated events were analysed with CellQuest software (BD).
The list of the antibodies used is:

PhotoTopo Endothelial Cell Preparation
Endothelial cells were isolated from mouse aorta. The aorta was removed from anesthetized and heparinised mice, and after cleaning was placed (with the intima side down) on Matrigel TM (BD)-coated plates in DMEM medium plus 10% FCS, glutamine, not essential amino acids, and 75 mg/mL endothelial cell growth supplement (ECGS, Sigma). See Suh et al., 1999 [25] for extensive details.

PhotoTopo Bone Marrow Derived Monocytes/ Macrophages Preparation
Mice were sacrificed in order to isolate the bone marrow. The haematopoietic precursors were isolated from bone marrow flushed from femurs and differentiated in vitro into macrophages as described [26].
Ten mito c-Photina transgenic mice islets/well were put in a white 96 MTP and incubated in Krebs-Ringer's solution (125 mM NaCl, 5 mM KCl, 1.2 mM MgSO 4 , 1.2 mM KH 2 PO 4 , 2 mM CaCl 2 , 25 mM HEPES pH 7.4, 0.1% BSA and 3 mM glucose) with 10 mM coelenterazine for 4 h at 37uC. The islets Ca 2+ kinetic responses were measured at the Luminoskan Ascent (Labsystems) luminometer after stimulation with a glucose stimulus (11 mM), or with mannitol (11 mM), as the negative control. The glucose concentration was then normalized to 3 mM and the islets were then stimulated with a depolarizing stimulus (40 mM KCl) on the Lumibox. The total photoprotein content in the islets was measured after cell were lysed with a Triton X-100-based buffer.

PhotoTopo Bioimaging
Islets were isolated as described above and seeded 30 islets/ 35 mm Matrigel TM (BD)-coated glass for 4 h at 37uC in Krebs-Ringer's solution containing 10 mM coelenterazine. Islets were then stimulated with a depolarizing stimulus (60 mM KCl) and light recorded with a set-up based on an Axiovert 200 inverted epifluorescence microscope (Zeiss, Oberkochen, Germany) positioned over an anti-vibration table and equipped with a 406/ 1.3NA EC Plan-Neofluar oil immersion objective lens. The light detector was the CMOS technology-based camera Fastcam (Photron, Tokyo Japan).

Supporting Information
Video S1 Pulsating cardiomyocytes at differentiation day 15. Embryoid body containing a spontaneous beating area of cardiomyocytes obtained 15 days after differentiation of mES c-Photina cells.  ). For all the compounds is indicated their complete name, the class, the action, the selectivity and the description (information provided directly by SIGMA LOPAC). Found at: doi:10.1371/journal.pone.0008882.s006 (0.11 MB DOC)