Dual-FRET imaging of IP3 and Ca2+ revealed Ca2+-induced IP3 production maintains long lasting Ca2+ oscillations in fertilized mouse eggs

In most species, fertilization induces Ca2+ transients in the egg. In mammals, the Ca2+ rises are triggered by phospholipase Cζ (PLCζ) released from the sperm; IP3 generated by PLCζ induces Ca2+ release from the intracellular Ca2+ store through IP3 receptor, termed IP3-induced Ca2+ release. Here, we developed new fluorescent IP3 sensors (IRIS-2s) with the wider dynamic range and higher sensitivity (Kd = 0.047–1.7 μM) than that we developed previously. IRIS-2s employed green fluorescent protein and Halo-protein conjugated with the tetramethylrhodamine ligand as fluorescence resonance energy transfer (FRET) donor and acceptor, respectively. For simultaneous imaging of Ca2+ and IP3, using IRIS-2s as the IP3 sensor, we developed a new single fluorophore Ca2+ sensor protein, DYC3.60. With IRIS-2s and DYC3.60, we found that, right after fertilization, IP3 concentration ([IP3]) starts to increase before the onset of the first Ca2+ wave. [IP3] stayed at the elevated level with small peaks followed after Ca2+ spikes through Ca2+ oscillations. We detected delays in the peak of [IP3] compared to the peak of each Ca2+ spike, suggesting that Ca2+-induced regenerative IP3 production through PLC produces small [IP3] rises to maintain [IP3] over the basal level, which results in long lasting Ca2+ oscillations in fertilized eggs.

As a partner of IRIS-2s, we developed a FRET based Ca 2+ indicator with single fluorophore to avoid fluorescent overlapping with IRIS-2s. We introduced a non-fluorescent mutation (Y145W 26 ) into a yellow fluorescent protein, cp173Venus, of YC3.60 27 (lower panels in Fig. 1a,b). The resultant protein have a fluorescent spectrum as same as ECFP, and addition of 100 μM Ca 2+ decreased its emission by FRET quenching. The peak fluorescent amplitude was 71 ± 3% (n = 3) reduced after addition of Ca 2+ in DYC3.60 (lower panel in Fig. 1c). Fluorescence from the three fluorophores used in IRIS-2s and DYC3.60 can be easily separated (Fig. 1e). Figure 1f shows time course changes of fluorescence from DYC3.60 and IRIS-2 in glutamate stimulated mGluR5-expressing HeLa cells. Less overlaps of excitation and emission spectra of IRIS-2 and DYC3.60 allowed dual-FRET imaging of Ca 2+ and IP 3 even without spectral unmixing 28 (Fig. 1f).
Characterization of IRIS-2s and DYC3.60 expressed in cultured mammalian cells.  and DYC3.60 were uniformly distributed within the cytosol when expressed in HeLa cells (Fig. 2a-e). Halo-TMR staining increased fluorescent signal detected by a 573-613-nm emission filter (Fig. 2b,d). The frequency of Ca 2+ oscillations monitored with Indo-5F in mGluR5-expressing HeLa cells stimulated with 10 μM glutamate were not significantly different among IRIS-2-, IRIS-2-Dmut-, and DYC3.60-expressing cells (50 ± 16 mHz for IRIS-2, n = 9; 55 ± 6 mHz for IRIS-2-Dmut, n = 4; 47 ± 10 mHz, n = 6 for DYC3.60) (Fig. 2f-h). IRIS-2 TMR signals did not return to its basal level during the intervals between Ca 2+ transients, and its fluctuation was synchronous with Ca 2+ oscillations (Fig. 2f). These characteristic IP 3 dynamics monitored with IRIS-2 TMR in HeLa cells are almost same as those recorded with other FRET-based IP 3 (Fig. 2i,j). Figure 2k shows a phase plane trajectory of [IP 3 ] and [Ca 2+ ] imaging data. [IP 3 ] gradually increased from 1st to 4th Ca 2+ spikes, and then,   (Fig. 2k). In the range of [IP 3 ], the trajectory cycled at almost the same orbit, suggesting that the trajectory is in a limit cycle (Fig. 2k). After termination of agonist stimulation, [IP 3 ] decreased below the range of limit cycle maintenance, which resulted in the termination of Ca 2+ oscillations (Fig. 2i,k). In the initial phase of Ca 2+ oscillations, [IP 3 ] increase precedes Ca 2+ spikes (Fig. 2k), suggesting that [IP 3 ] increases induce Ca 2+ spikes. In the limit cycle phase, Ca 2+ spikes occur without marked [IP 3 ] increases ( Fig. 2k), suggesting that Ca 2+ induced positive and negative feedbacks to IP 3 R autonomously induce Ca 2+ spikes 24 . Ca 2+ oscillations last as long as [IP 3 ] maintained in the range of limit cycle. Termination of agonist stimulation induces [IP 3 ] decrease below to the range maintaining the limit cycle. [IP 3 ] necessary to induce Ca 2+ spike should be different at the initial state and later state of Ca 2+ oscillations because Ca 2+ directly or indirectly inactivates IP 3 R 28,31 . Thus, Ca 2+ disappears even [IP 3 ] above the basal level at the termination of Ca 2+ oscillations (Fig. 2i).
Characterization of IRIS-2 in UV-uncaging experiments. Next, we checked the compatibility of IP 3 sensors with UV-uncaging. Caged-compounds are light-sensitive probes that functionally encapsulated biomolecules in an inactive form. The active compounds can be released from caged-compounds with UV light in most of caged-compounds. IRIS-1 or IRIS-2 were expressed in HeLa cells and irradiated by UV pulses ( Supplementary  Fig. 2a,b). We found UV irradiation caused temporal reduction of fluorescence of both ECFP and Venus in IRIS-1 expressing cells (Supplementary Fig. 2a). Because of the difference of the signal reduction between those fluorescent proteins, the fluorescent ratio of IRIS-1 was significantly reduced (−1.9 ± 0.7%, n = 22). In contrast, the fluorescent signals from EGFP and HaloTag-TMR were stable after the UV irradiation ( Supplementary Fig. 2b), which resulted in successful detection of [IP 3 ] changes after UV-uncaging of caged-IP 3 ( Supplementary Fig. 2c).

Detection of Ip 3 concentration changes in fertilized mouse eggs.
To detect IP 3 dynamics in fertilized mouse eggs, IRIS-1, IRIS-2, or IRIS-2.3 was expressed in eggs by cRNA injection. For the simultaneous monitoring of [Ca 2+ ] changes, we first used Indo-5F as a Ca 2+ indicator according to the method described previously 25 . As shown in Supplementary Figure 3, we did not detect any changes of IRIS-1 signals in fertilized eggs. Not only the fails of the detection of IP 3 changes, it was difficult to detect [Ca 2+ ] changes after addition of www.nature.com/scientificreports www.nature.com/scientificreports/ sperm into the culturing media. Even in the experiments with successful detection of fertilization-induced [Ca 2+ ] changes, the number of Ca 2+ transients was less compared to IRIS-2-Dmut (number of Ca 2+ spikes during 30 min after the first Ca 2+ spikes: IRIS-1: 1.91 ± 0.13 (n = 3); IRIS-1-Dmut: 3.75 ± 0.5 (n = 4); p = 0.008, Student's t-test), suggesting that IRIS-1 works as a significant IP 3 buffer. We also tested IRIS-2 expressing eggs for in-vitro fertilization assay and found IRIS-2 expressing eggs had normal Ca 2+ spikes after fertilization (Fig. 3a). However, it was also hard to detect clear increases in FRET signals in IRIS-2-expressing eggs during the first Ca 2+ transient evoked after fertilization, while small repetitive transients of IRIS-2 TMR signals synchronous with Ca 2+ oscillations were observed approximately 30 min after the onset of the first Ca 2+ transient (Fig. 3a). On the other hand, we clearly detected IP 3 increases during the all Ca 2+ transients, including the first Ca 2+ transient, in IRIS-2.3-expressing eggs (Fig. 3b). During the first large Ca 2+ transient, [IP 3 ] continues to increase, and all the following Ca 2+ transients accompanied with a rapid increase and a following slow decline on the elevated level of [IP 3 ] (Fig. 3b). Three independent experimental results of [IP 3 ] and [Ca 2+ ] imaging with IRIS-2.3 and Indo-5F at the onset of first Ca 2+ spike were shown in Supplementary Figure 4. We did not find significant difference of numbers of Ca 2+ spikes during 30 min after 1st Ca 2+ spike between IRIS-2 and IRIS-2.3 expressing eggs (IRIS-2: 5.17 ± 1.72, n = 6; IRIS-2.3: 6.33 ± 5.72, n = 9; p = 0.58, student's t-test). ] changes experimentally, we used Fura-2, whose affinity is higher than that of Indo-5F (Fura-2: Kd = 135 nM; Indo-5F: Kd = 470 nM), as a Ca 2+ indicator to detect the timing of the onset of the first Ca 2+ transient as precise as possible. As shown in Fig. 3c Fig. 3c) as reported previously 33 . The peak amplitude and the rising speed of [IP 3 ] increased after the shoulder point of the first Ca 2+ transient (Fig. 3c,d), suggesting acceleration of IP 3 production via Ca 2+ -induced activation of PLC isozymes. positive feedback loop to produce rising phase of Ca 2+ spikes. Each Ca 2+ spike of Ca 2+ oscillations usually form as a result of an initial slow pacemaker rise in [Ca 2+ ] followed by a rapid rise in [Ca 2+ ] [34][35][36] . The accelerated rise of [Ca 2+ ] is suggested that a regenerative process is involved in the generation of the abrupt upstroke 35 . Such regenerative processes require a positive-feedback element 22 , and CICR from IP 3 R and Ca 2+ -induced IP 3 production through PLC have been proposed as candidates of the positive feedback element. In the previous   (Fig. 4). In the early phase (from first to 5th transients) of fertilization-induced Ca 2+ oscillations, the amplitudes of IP 3 fluctuations were relatively small (Fig. 3b), and the rate of [IP 3 ] rise did not increase during the rising phase of the Ca 2+ transients, as found in cultured HeLa cells 25 (Fig. 4a,b). The amplitudes of IP 3 fluctuations were gradually increased during the later phase of Ca 2+ oscillations (Fig. 3a,b), and contrary to the early phase, the onset of the rate of [IP 3 ] rise precedes that of [Ca 2+ ] (Fig. 4c). The result suggests that Ca 2+ -induced IP 3 production through PLC may work as a part of the positive feedback loop to produce abrupt [Ca 2+ ] rise at Ca 2+ spikes in later phase of Ca 2+ oscillations. However, the peak of the rate of [IP 3 ] rise always delayed from that of [Ca 2+ ] (Fig. 4c), suggesting that CICR from IP 3 R has major role to produce the rising phase of Ca 2+ spikes and [IP 3 ] rises.

Initiation of [Ip 3 ] and [Ca
[IP 3 ] stayed at the elevated level and did not return to the basal level through Ca 2+ oscillations (Fig. 3b).  Fig. 5a, these fluorescent probes were distributed evenly in the egg. Well separation of excitation and emission spectra of these fluorophores enabled simultaneous detection of these fluorescence (Figs 1e, 5a,b). As same as the results we obtained with the pair of Indo-5F and IRIS-2.3 TMR , we successfully detected fertilization-induced [Ca 2+ ] and [IP 3 ] changes with DYC3.60 and IRIS-2.3 TMR (Fig. 5b and Supplementary video 1). As same as HeLa cells, [IP 3 ] at the termination was higher than that at the initiation of Ca 2+ oscillations in fertilized mouse eggs ( Fig. 2i and Supplementary Fig. 5).

Dual-FRet imaging of [Ip 3 ] and [Ca
Ca 2+ -induced regenerative Ip 3 production. We also detected delays in the peak of [IP 3 ] compared to the peak of each Ca 2+ spike (17 ± 11 sec, n = 63, Fig. 5c,d) www.nature.com/scientificreports www.nature.com/scientificreports/ the mouse egg, we stimulate unfertilized mouse eggs with 100 μM carbachole (Fig. 5f,g). The stimulation caused Ca 2+ spikes and a monotonic [IP 3 ] rise ( Fig. 5f). At fertilization, [IP 3 ] changes always follow after Ca 2+ spikes. On the other hand, Ca 2+ spikes did not accompany with delayed [IP 3 ] rises in carbachole stimulated unfertilized eggs (Fig. 5f). Particularly, IP 3 peak at the first Ca 2+ spike preceded Ca 2+ peak (Fig. 5g). These data showed that Ca 2+ -induced IP 3 producing activity is not strong in unfertilized eggs, suggesting that sperm derived PLCζ should participate Ca 2+ -induced IP 3 production. As we showed in Figs 3a,b and 4, Ca 2+ -induced [IP 3 ] rises increased later phase of fertilization-induced Ca 2+ oscillations, suggesting that fertilization induces quantitative or qualitative changes of PLC in later phase of Ca 2+ oscillations.

Discussion
In this study, we developed a dual-FRET pair of biosensors for the detection of [IP 3 ] and [Ca 2+ ] in mammalian cells. The uniqueness of our dual-FRET pair is using single fluorophore for one of the pairs. Replacement of Y145 to W in EYFP is known to produce a non-fluorescent chromoprotein that retains its absorption of emission light 26 . Introduction of the Y145W mutation into cp173Venus of YC3.60 27 resulted to produce single fluorophore with fluorescent quencher in the Ca 2+ FRET sensor, DYC3.60. Usually, four fluorophores are necessary for dual-FRET imaging. Most of FRET sensors have cyan and yellow fluorescent proteins 45 , and these fluorescent proteins cover a broad spectral profile. Thus, using FRET sensor with cyan and yellow proteins, it is difficult to find a partner FRET sensor for dual-FRET imaging without using spectral unmixing to distinguish each fluorescent signal mathematically from significantly overlapped fluorescent signals 46,47 . We offer a dual-FRET imaging with three fluorophores, which gives easier detection and separation of fluorescent signals.
The new FRET sensors enabled imaging of [IP 3 ] and [Ca 2+ ] at fertilization of mouse eggs. We have succeeded to detect [IP 3 ] changes in fertilized mouse eggs using a second-generation fluorescent IP 3 sensor, IRIS-2.3, which has an improved dynamic range and a high IP 3 sensitivity. Simultaneous monitoring of both Ca 2+ and IP 3 in fertilized mouse eggs showed that the [IP 3 ] increase was detected approximately 3 min before the onset of the first www.nature.com/scientificreports www.nature.com/scientificreports/ Ca 2+ transient. The result is consistent with the expectation that highly Ca 2+ sensitive PLCζ produces IP 3 at the basal level of [Ca 2+ ] in the egg cytosol after sperm-egg fusion 12 . Mehlmann and Kline reported microinjection of small amount of IP 3 (8.6 nM) is able to induce Ca 2+ spike in unfertilized mouse eggs 48 . Our measurements showed the same results that the amount of IP 3 produced in mouse eggs is small even after the fertilization because only IRIS-2.3, which shows the highest IP 3 sensitivity (Kd = 47 nM) among the IP 3 24 . In this and previous reports, we showed sustained [IP 3 ] increase during Ca 2+ oscillations in HeLa cells and fertilized mouse eggs 25 (Figs 3 and 4), and the same results were obtained with other IP 3 sensor proteins 29,30 . Consistently with our results, Mehlmann and Kline reported single microinjection of IP 3 induces Ca 2+ oscillations in unfertilized mouse eggs 48 . Jones et al. also reported Ca 2+ oscillations with continuous low level caged-IP 3 photolysis in unfertilized mouse eggs 49 . PLCζ is a smallest and simplest PLC isoform 9 . The activity of PLCζ is regulated by Ca 2+ and localization into nucleus after pronuclear formation, and other regulations are not known 50 . PLCζ has highest Ca 2+ sensitivity compared to the other PLC isoforms and is 70% active at the basal level [Ca 2+ ] in cells 12 . Thus, PLCζ should be continuously active after fertilization until pronuclear formation 51 , which should sustain continuous [IP 3 ] increase during fertilization-induced Ca 2+ oscillations (Figs 3 and 4). We previously found that CICR dominantly work as a positive feedback loop to produce the rising phase of Ca 2+ spikes in HeLa cells 25 . Our data suggest that the mechanism elicits the rising phase of Ca 2+ spikes in fertilized mouse eggs is more complex. Initially, CICR dominantly works as the positive feedback loop, and Ca 2+ -induced IP 3 production gradually participates to produce Ca 2+ spikes cooperatively with CICR in the later phase of Ca 2+ oscillations. Ca 2+ -induced IP 3 production through PLC produces [IP 3 ] rises at each Ca 2+ spike to help keeping [IP 3 ] over the basal level, which results in long lasting Ca 2+ oscillations in fertilized eggs. In

Materials and Methods
Animals. Experiments used ddY mice for preparation of oocytes and sperm. All animal experiments were performed in accordance with the guidelines approved by the Animal Experiments Committee of RIKEN Brain Science Institute. All experiments were carried out in accordance with the approved ethical guidelines and regulations.
Gene construction. The FRET donor and acceptor of IRIS-1 were replaced with mEGFP and Halo-protein (Promega), respectively, to produce IRIS-2. Amino acid residues 224-575 of mouse IP3R1 in IRIS-2 were replaced with amino acid residues 224-579 of mouse IP3R1 to produce IRIS-2.3. The Y145W mutant 26 of circular permutated Venus (cp173V-Y145W) 27 was generated using the site-directed mutagenesis. The FRET acceptor of YC3.60 27 was replaced with cp173Venus-Y145W to produce DYC3.60. IRIS-2, IRIS-2.3 and DYC3.60 cDNAs were cloned into the NheI and XbaI sites of pcDNA3.1 zeo(+) (Invitrogen) for the expression in HeLa cells. The cDNAs were cloned into the XbaI site of pTNTTM (Promega) with extended poly(A) tail (57 residues) and synthesized cRNAs were injected into mouse oocytes.
Protein expression and purification. The full-length cDNA of IRIS-2 was isolated from pcDNA3.1 zeo-(+)-IRIS-2 by using NheI and XbaI sites and was cloned into the XbaI site of baculovirus transfer vector pFast-Bac1 (Invitrogen). The recombinant baculovirus was used for the large-scale expression of IRIS-2 in Sf9 cells as described previously 52 . The expressed proteins were purified on a HiTrap heparin HP column (GE Healthcare Life Sciences) as described previously 53 .
Cell culture and transfection. HeLa cells were cultured in Dulbecco's modified Eagle medium supplemented with 10% heat-inactivated fetal bovine serum. HeLa cells were transfected with expression vectors by transfection reagent (Mirus TransIT). One day after the transfection, cells were used for imaging experiments. preparation of RNA. Plasmids carrying IRISs or DYC3.60 were digested by NdeI, and linearized DNA fragments were purified with Wizard SV Gel and PCR clean-up Kit (Promega). They were used as the templates for RNA transcription by T7 polymerase using T7 mMESSAGEmMACHINE Kit (Ambion). RNA was purified using RNeasy MinElute Cleanup Kit (Qiagen) and stored at −80 °C until use. preparation of gametes. Full grown immature oocytes were collected from the follicles in the ovaries of female mice 47-49 h after the injection of pregnant mare serum gonadotropin. Isolated oocytes were freed from cumulus cells mechanically by pipetting in M2 medium, and then cRNAs of IRIS-1, IRIS-2, IRIS-2.3, DYC3.60 or dKeima570 were injected as described below. Sperm was collected from the caudal epididymides and were incubated in M16 medium 54 supplemented with 4 mg/ml BSA (Sigma) at 37 °C (5% CO2) for >5 h for capacitation and acrosome reaction 55 . Microinjection and insemination. RNA solutions were diluted to 130 ng/μl with the intracellular medium (150 mM KCl, 5 mM Tris-KOH, pH 7.0). Immature oocytes were injected with 20 pl of RNA solutions and incubated in the M16 medium for 16 h at 37 °C with 5% CO 2 . Only eggs maturated normally to metaphase II with the www.nature.com/scientificreports www.nature.com/scientificreports/ first polar body were used in the following experiments. After loaded with 2 μM of Indo-5F or Fura-2 for 30 min in the M2 medium, eggs were freed from the zona pellucida by brief treatment with acidic Tyrode's solution (pH 2.5) 56 for insemination. Sperm was added during imaging experiments.
Imaging. After loading HeLa cells with 10 μM Indo-5F-AM (AnaSpec), imaging was performed under the constant flow (2 ml/min) of the balanced salt solution containing 20 mM Hepes, pH 7.4, 115 mM NaCl, 5.4 mM KCl, 1 mM MgCl2, 1.3 mM CaCl2, and 10 mM glucose as an imaging media at 37 °C through an inverted microscope (IX71 or IX81; Olympus) with a cooled charge-coupled device (CCD) camera (ORCA-ER; Hamamatsu Photonics) and a 40x, 1.35 NA, oil-immersion objective (Olympus). For the fluorescent images of IRIS-1 and Indo-5F, an emission splitter (W-view; Hamamatsu Photonics) was used with a light source exchanger (DG-4; Sutter Instrument Co.) on the IX71 inverted microscope. Sequential excitation of IRIS-1 and Indo-5F was performed by using a 450-nm dichroic mirror and two excitation filters (a 425-445 nm filter for IRIS-1 and a 360-nm filter for Indo-5F). Emissions from IRIS-1 and Indo-5F were split with a 460-510-nm filter (for IRIS-1 and Indo-5F), a long-path 520-nm (for IRIS-1) barrier filter, and two 505-nm dichroic mirrors equipped in W-view.
Image acquisition was performed with MetaFluor (Molecular Devices). Data analysis was performed with MetaFluor and Igor Pro (WaveMetrics) softwares. The EGFP/TMR emission ratio (IRIS-2s), the ECFP/Venus emission ratio (IRIS-1s), the dKeima570/ECFP emission ratio (DYC3.60), the 420-440 nm/460-510 nm emission ratio (Indo-1) and the ratio of 510-550 nm emission excited at 340 nm and 510-550 nm emission excited at 380 nm (Fura-2) were defined as R. ∆R was defined as R -Rbase, where Rbase is the basal level of R. Baseline drift in each experiment was corrected with subtracting the trend line which is calculated with the line around the beginning of each experiment.