Mouse preimplantation embryo responses to culture medium osmolarity include increased expression of CCM2 and p38 MAPK activation

Background Mechanisms that confer an ability to respond positively to environmental osmolarity are fundamental to ensuring embryo survival during the preimplantation period. Activation of p38 mitogen-activated protein kinase (MAPK) occurs following exposure to hyperosmotic treatment. Recently, a novel scaffolding protein called Osmosensing Scaffold for MEKK3 (OSM) was linked to p38 MAPK activation in response to sorbitol-induced hypertonicity. The human ortholog of OSM is cerebral cavernous malformation 2 (CCM2). The present study was conducted to investigate whether CCM2 is expressed during mouse preimplantation development and to determine whether this scaffolding protein is associated with p38 MAPK activation following exposure of preimplantation embryos to hyperosmotic environments. Results Our results indicate that Ccm2 along with upstream p38 MAPK pathway constituents (Map3k3, Map2k3, Map2k6, and Map2k4) are expressed throughout mouse preimplantation development. CCM2, MAP3K3 and the phosphorylated forms of MAP2K3/MAP2K6 and MAP2K4 were also detected throughout preimplantation development. Embryo culture in hyperosmotic media increased p38 MAPK activity in conjunction with elevated CCM2 levels. Conclusion These results define the expression of upstream activators of p38 MAPK during preimplantation development and indicate that embryo responses to hyperosmotic environments include elevation of CCM2 and activation of p38 MAPK.


Background
Culture medium osmolarity is one of the primary parameters that must be considered when formulating an optimized medium for the production of preimplantation embryos. Even brief exposure of preimplantation embryos to 300 mOsm/kg culture media (in the absence of osmolytes) results in impaired development [1][2][3].
Most mammalian embryo culture media formulations have employed osmolarities around 250 mOsm/kg. Greater culture medium osmolarities may be employed but only in the presence of osmolytes such as glycine, betaine, proline and glutamine [2,4]. Preimplantation embryos express a number of transporters that serve to regulate and maintain embryonic cell volume [5][6][7]. In somatic cell systems, activation of p38 MAPK is a common response of these osmoregulatory pathways [8,9].
The discovery of a class of compounds called cytokinesuppressive anti-inflammatory drugs (CSAIDs) has allowed for the specific pharmacological inhibition of MAPK14/p38α and MAPK11/p38β isoforms [20]. The most extensively characterized CSAIDs are the pyridinyl imidazoles SB203580 [21] and the more potent SB220025 [22]. We have reported that all four p38 MAPK isoforms are expressed throughout mouse preimplantation development [23]. In addition, embryos treated with CSAIDs experience a reversible blockade of development at the 8-16 cell stage which is accompanied by a reversible loss of filamentous actin (F-actin) [23,24]. These results point towards an essential role for MAPK14/11 in directing development of the mouse embryo past the 8-16 cell stage [23][24][25].
Among its possible roles in the early embryo, p38 MAPK signaling is likely to mediate embryonic responses to hyperosmotic stimuli. Recently, a novel scaffolding protein called Osmosensing Scaffold for MEKK3 (OSM) was characterized [26]. OSM binds to F-actin, the GTPase, RAC, and the upstream kinases MAP3K3/MEKK3 and MAP2K3 in the p38 MAPK phospho-relay module, recruiting these proteins to sites of active membrane ruffling and newly polymerized actin ( Figure 1) [26]. Downregulation of OSM by RNA interference demonstrated that MAP3K3 and OSM were required for p38 MAPK activation in response to sorbitol-induced hypertonicity [26]. The current mouse gene name for OSM is cerebral cavernous malformation 2 homolog (human) (CCM2).
The present study was conducted to investigate whether CCM2 and upstream p38 MAPK pathway constituents are expressed during preimplantation development and to determine whether changes in CCM2 expression are associated with p38 MAPK activation following exposure of preimplantation embryos to hyperosmotic stimuli. Our results indicate that CCM2 is expressed throughout mouse preimplantation development and its protein levels are elevated in response to hypertonic culture conditions and correlate with increased p38 MAPK activity in the early embryo.

Results
Detection of mRNA transcripts encoding Ccm2, Map3k3, Map2k3, Map2k6, and Map2k4 during mouse preimplantation development Qualitative RT-PCR methods resulted in the detection of mRNA transcripts encoding Ccm2, Map3k3, Map2k3, Map2k6 and Map2k4 producing the expected size amplicons of 371 bp, 374 bp, 387 bp, 395 bp and 363 bp respectively throughout mouse preimplantation development in all 3 experimental replicates ( Figure 2A). The identity of each RT-PCR product was confirmed through direct sequencing and BLAST ® analysis. RT-PCR products amplified using Ccm2, Map3k3, and Map2k3 primers possessed 100% sequence identity with their respective Gen-Bank ® mouse nucleotide sequences. RT-PCR products amplified using Map2k6 and Map2k4 primers possessed 98% and 99% sequence identity with their corresponding GenBank ® mouse nucleotide sequences, respectively.
Relative abundance of mRNAs encoding Ccm2, Map3k3, Map2k3, Map2k6, Map2k4 and Mapk14 during mouse preimplantation development Ccm2 ( Figure 2B) and Map2k6 ( Figure 2E) mRNA steady state levels were significantly highest at the blastocyst stage (P ≤ 0.05) and did not vary significantly between the morula and 1-cell stages (P > 0.05). However, both Ccm2 and Map2k6 mRNA levels decreased significantly from 1cell levels to those observed at the 2-cell, 4-cell, and 8-cell stage (P ≤ 0.05). Map3k3 ( Figure 2C) and Map2k3 ( Figure  2D) steady state mRNA levels were significantly highest at the morula and blastocyst stages (P ≤ 0.05). In contrast, Map2k4 ( Figure 2F) steady state mRNA levels were highest at the 1-cell stage and declined significantly at the 2-cell stage (P ≤ 0.05). Map2k4 mRNA levels then remained low throughout preimplantation development, but did increase significantly (P ≤ 0.05) at the morula and blastocyst stages ( Figure 2F). Finally, Mapk14 ( Figure 2G) steady state mRNA levels were significantly highest (P ≤ 0.05) at the blastocyst stage. Mapk14 mRNA levels gradually declined from the 1-cell stage to reach their lowest level at the 8-cell stage before increasing significantly (P ≤ 0.05) again at the morula and blastocyst stages.   CCM2 as a Scaffold for p38 MAPK Activation. As described in [26] the Rho-GTPase, RAC, is recruited to actin membrane ruffles by CCM2 following hyperosmotic stress, facilitating the activation of p38 MAPK. CCM2 acts as a scaffold binding to RAC, MAP3K3, and MAP2K3 to organize these components into a functional signaling module. We have included in this model the two additional upstream MAP2Ks to p38 MAPK: MAP2K6, a specific activator of p38 MAPK related to MAP2K3; and MAP2K4, which primarily activates the JNK/SAPK pathway, but can also phosphorylate p38 MAPK in vitro. These two kinases may contribute to activating p38 MAPK in response to hyperosmotic stress but unlikely via interactions with CCM2  Data are normalized to Luciferase control (0.025 pg/embryo) and relative to 1-cell target gene mRNA levels. Relative mRNA levels are presented as the mean ± s.e.m. representative of three independent replicates. Bars with different letters represent significant differences in relative mRNA levels between embryo stages (P ≤ 0.05). 1-cell to blastocyst) were also detected throughout preimplantation development. Phosphorylated MAP2K3/MAP2K6 was detected in the cytoplasm and the nucleus at all preimplantation stages of mouse development although the nuclear fluorescence was more intense than the cytoplasmic fluorescence. Phosphorylated MAP2K4 was primarily confined to the nucleus of each blastomere throughout preimplantation development ( Figure 3 p-MAP2K4 1-cell to blastocyst). In some cases faint phosphorylated-MAP2K4 immunofluorescence was detected in the cytoplasm. This was not consistent and was not restricted to a particular embryo stage ( Figure 3 p-MAP2K4 1-cell to blastocyst). For all proteins these distribution patterns were consistently observed in both trophectoderm and inner cell mass cell types of the blastocyst.

Effect of culture in 1800 mOsm hyperosmotic medium on blastocyst morphology and p38 MAPK activation
We initially employed treatment with high culture osmolarity (i.e. 1800 mOsm) to explore the effects of an extreme treatment paradigm on phospho-MAPKAP2 fluorescence. Blastocyst culture in KSOMaa media containing 10% glycerol or 1.4 M sucrose resulted in an instantaneous decrease in blastocyst volume. Those placed in medium + 10% glycerol displayed a rapid recovery (5 minutes) back to normal volume, however, this recovery was not observed in embryos placed in 1.4 M sucrose medium (data not shown). Phospho-MAPKAPK2 immunofluorescence increased in blastocysts cultured in KSOMaa + 1.4 M sucrose for 10 minutes ( Figure 4C) or for 30 minutes ( Figure 4G) when compared to blastocysts cultured in KSOMaa + 10% glycerol for 10 minutes ( Figure  4I). The RSS of blastocysts cultured in KSOMaa + 10% glycerol displayed was not significantly different from normal KSOMaa cultured controls ( Figure 4I).  Figure  4J).

Effect of 460 mOsm hyperosmotic treatment on blastocyst morphology and p38 MAPK activation
To investigate the effects of reduced levels of hyperosmolarity on phospho-MAPKAPK2 fluorescence we next conducted the following experiment using culture medium adjusted to 460 mOsm. Exposure of cultured blastocysts to a 1.4% glycerol or 0.2 M sorbitol medium did not visibly affect blastocyst volume (data not shown). While no perceptible differences in blastocyst volume were observed between embryos following 15 minutes and 30 minutes of 460 mOsm hyperosmotic treatment, variations in the levels of phospho-MAPKAPK2 fluorescence were detected ( Figure 4K-R). Following 15 minutes of exposure to 460 mOsm hyperosmotic medium treatments, blastocysts cultured in KSOMaa + 0.2 M sorbitol displayed a significant increase (P ≤ 0.05) in RSS of phosphorylated MAPKAPK2 immunofluorescence (1.75 ± 0.18, n = 33) when compared to blastocysts cultured in normal KSOMaa culture medium (1.00 ± 0.11, n = 33) or KSOMaa + 1.4% glycerol (1.16 ± 0.13, n = 34) ( Figure 4S). The RSS of phosphorylated MAPKAPK2 immunofluorescence in blastocysts cultured in KSOMaa + 1.4% glycerol did not differ from KSOMaa only controls ( Figure 4S). At the 30 minute time-point, the RSS of blastocysts cultured in KSOMaa + 0.2 M sorbitol was significantly higher (P ≤ 0.05) than that observed for blastocysts cultured in KSO-Maa + 1.4% glycerol and KSOMaa only controls [1.46 ± 0.12 (n = 35), 0.911 ± 0.082 (n = 35), and 1.00 ± 0.08 (n = 33), respectively] ( Figure 4T). Once again there was no significant difference in the RSS displayed by blastocysts cultured in the glycerol medium and normal KSOMaa medium controls ( Figure 4T).

Effect of 460 mOsm hyperosmotic treatment on Ccm2 transcripts and CCM2 protein
To explore a possible mechanism for the induction of p38 MAPK following incubation in hyperosmotic medium we next examined the influence of exposure to hyperosmotic medium on Ccm2 expression. We employed the 460 mOsm culture medium treatment for these experiments since we had now established that this treatment was sufficient to significantly increase phospho-MAPKAPK2 immunofluorescence levels. Cultured blastocyst stage mouse embryos were exposed to hyperosmotic treatment for 3, 6, 9, 12, or 24 hours in KSOMaa + 0.2 M sorbitol (460 mOsm) and compared with cultured blastocysts in normal KSOM medium (control group, 260 mOsm). No significant differences (P > 0.05) in relative Ccm2 mRNA transcript levels were observed between blastocysts cultured in the KSOMaa + 0.2 M sorbitol and blastocysts cultured in the normal KSOMaa medium control group at any of the investigated time-points ( Figure 5A). Despite Relative signal strengths are presented as the mean ± s.e.m. representative of three independent replicates. Bars with different letters represent significant differences in relative signal strength between treatment groups (P ≤ 0.05).  Figure 5F).

Discussion
Our overall objective was to increase our understanding of the intracellular signaling pathways that respond to the external environment and regulate preimplantation development. Our recent characterization of the expression of all four isoforms of p38 MAPK (MAPK14/p38α, MAPK11/p38β, MAPK12/p38γ, and MAPK13/p38δ) and a number of its downstream substrates provided the foundation for our current studies [23]. We have recently demonstrated a requirement for p38 MAPK signaling during preimplantation development by characterizing the reversible developmental blockade that occurs at the 8-to 16-cell stage following inhibition of MAPK14/p38α and MAPK11/p38β isoforms [23]. Furthermore, p38 MAPK regulates F-actin polymerization during preimplantation development via phosphorylation of downstream substrates MAPKAPK2 and PRAK/MAPKAPK5, and subsequently through the small heat shock proteins, HSPB1/2 [23,24]. Our present study has extended these initial investigations into p38 MAPK function during preimplantation development by addressing the upstream events in the three-tiered MAPK pathway module. Our first objective was to characterize the expression of regulatory kinase gene products upstream of p38 MAPK during preimplantation development. In this regard we have demonstrated that transcripts and proteins encoding all known MAPKKs upstream of p38 MAPK (i.e. MAP2K3, MAP2K6, and MAP2K4) are present throughout mouse preimplantation development. Moreover, we have also identified transcripts and proteins encoding the MAP3K, MAP3K3, as well as the novel scaffolding protein, CCM2.
Each gene of interest displayed a consistent increase in transcript abundance as embryo development advanced towards the blastocyst stage. There was a significant increase in relative mRNA levels from the 8-cell stage to the morula and blastocyst stages for all transcripts investigated. These results suggest that mRNAs for p38 MAPK pathway constituents are rapidly accumulating in the post 8-cell stage embryo and this observation correlates well with outcomes from p38 MAPK inhibition studies which suggest that p38 MAPK signaling is required to maintain development beyond the 8-16 cell stage [23][24][25]. Of particular note is the nearly 40-fold increase over 1-cell levels in Map3k3 expression at the morula stage, and the 12-fold increase over the same period in Map2k3 expression which indicates that MAP3K3 and MAP2K3 may be the primary upstream activators of p38 MAPK during preimplantation development. Conversely, the relatively low expression of Map2k4 mRNA transcripts past the 1-cell stage suggests a reduced role for this kinase during early development.
Our results indicate that all of these gene products are derived from both oogenetic (maternal) and embryonic origins since they are found in both pre-and post-maternal zygotic transition (MZT) stage embryos. The predominant pattern observed for transcripts encoding p38 MAPK upstream regulators was one where mRNA levels clearly increased with advancing cleavage stage. This accumulation of mRNAs with advancing embryo stage could only occur via contributions from embryonic transcriptional activity.
At the protein level, the CCM2 and MAP3K3 cytoplasmic distribution pattern observed following application of whole-mount indirect immunofluorescence was as expected considering the role of CCM2 as an actin binding scaffold protein and MAP3K3 as an upstream MAP3K that is responsive to extracellular stimuli at the cell surface. The diffuse cytoplasmic distribution mirrors the localization pattern displayed by these proteins in HEK293 and COS7 cell lines [26,27]. We also determined the distribution of phosphorylated MAP2K3/MAP2K6 proteins and MAP2K4 proteins throughout mouse preimplantation development. The use of antisera recognizing the phosphorylated forms of these proteins allowed us to not only characterize their distribution throughout preimplantation development but also report their activation during preimplantation development. Activated MAP2Ks were detected in the cytoplasm and nucleus of embryos collected from the 1cell to the blastocyst stage, indicating that the p38 MAPK pathway is activated from the earliest stages of development onward. While phospho-MAP2K3/MAP2K6 proteins were detected in the cytoplasm and the nucleus, phospho-MAP2K4 proteins were confined to the nuclei, suggesting that p38 MAPK phosphorylation by MAP2K4 may occur predominantly in the nucleus. This nuclear localization is supported by studies that investigated the activation of JNK/SAPK in P19 embryonic carcinoma cells [28]. MAP2K4 was detected in both the nucleus and cyto-Effect of 460 mOsm Hyperosmotic Treatment on Ccm2 mRNA and CCM2 Immunofluorescence   Figure 1) [31].
Since it is well known that the roles played by p38 MAPK extend well beyond the regulation of the actin cytoskeleton our present study investigated whether embryonic p38 MAPK signaling is affected by exposure to hyperosmotic stimuli as is reported for somatic cells. Our results clearly demonstrate that hyperosmotic treatment during culture increased p38 MAPK activity as assessed by increased phosphorylation of its specific downstream substrate, MAPKAPK2. These findings are supported by numerous studies that have described the responsiveness of this pathway to hyperosmotic stress, and have indicated that increases in p38 MAPK activity are a common cellular response to hyperosmotic stimuli (reviewed by [9]. While the increased p38 MAPK activity and phosphorylation of MAPKAPK2 in itself is not a novel discovery, the responsiveness of p38 MAPK to hyperosmotic stress has not until now been demonstrated to occur during preimplantation development. One of the most exciting outcomes from our study is the differential activation of p38 MAPK that occurred in response to the solute compound used to increase the osmolarity of embryo culture media. At both 1800 mOsm and 460 mOsm hyperosmotic treatment conditions, the addition of glycerol to KSOMaa culture medium did not result in a significant increase in phospho-MAPKAPK2 immunofluorescence in blastocysts. This is in contrast to the 1.4 M sucrose (1800 mOsm) and 0.2 M sorbitol (460 mOsm) treatments in which a significant increase in phospho-MAPKAPK2 immunofluorescence was detected at two treatment timepoints. We propose that this difference in p38 MAPK activation is a result of the alleviation of osmotic stress and restoration of osmotic equilibrium across the cell membrane due to aquaporin (AQP; water channel) mediated glycerol permeability. This likelihood is supported by our previous studies which demonstrated the selective permeability of water and glycerol, but not sucrose, through apical and basolateral AQPs located in the blastocyst trophectoderm [32,33]. Since glycerol can permeate the cell membrane relatively freely via AQPs, the initial osmotic gradient (from 260 mOsm to 1800 mOsm or 460 mOsm) produced by the addition of glycerol to culture media is rapidly reversed in these treatment groups, alleviating "osmotic stress" that might induce p38 MAPK activity. This also permits the observed re-expansion of blastocyst volume that does not occur with 1.4 M sucrose treatment [32]. The osmotic gradient produced by 1.4 M sucrose (1800 mOsm) or 0.2 M sorbitol (460 mOsm) persists for the duration of treatment, enough to induce p38 MAPK activity. Our results may therefore have an impact on studies directed at optimizing embryo cryopreservation protocols as the cyroprotectant used (ie glycerol or other) would be expected to have differential effects on p38 MAPK activation which may influence embryo survival and recovery post-thaw. Investigation of this possibility could result in improved outcomes following embryo cryopreservation.
Interestingly, hyperosmotic treatment of cultured rat astrocytes by mannitol or sorbitol increases the expression of AQP4 and AQP9 mRNAs and proteins, which is suppressed by treatment with the p38 MAPK inhibitor SB203580 [34]. Glycerol treatment, however, had no effect on the expression of these two AQPs [34]. Similar to our study, this was attributed to the absence of an osmotic gradient as a result of glycerol movement across cell membranes. Other AQP family members were also differentially expressed in sucrose, sorbitol, or mannitol treatment (in which an osmotic gradient is established) [35,36]. Mouse preimplantation embryos are capable of regulating AQP mRNA abundance in response to environmental changes in osmolarity and changes to blastocoel volume following puncture [37], however, the potential role of p38 MAPK activity in mediating these events was not investigated. Based upon our results, we would hypothesize that a prolonged shift in osmotic gradient induces p38 MAPK activity, which in turn regulates AQP expression as a compensatory mechanism in preimplantation embryos. This would also represent an important direction for future studies to pursue.

Conclusion
In conclusion, we have demonstrated for the first time that transcripts and polypeptides encoding MAP3K3, MAP2K3, MAP2K6, MAP2K4 and CCM2 are expressed and localized throughout mouse preimplantation development. We have discovered that p38 MAPK activity is regulated by exposure to hyperosmotic stimuli, and that the response to hyperosmotic stress in the early embryo includes increased CCM2 levels. These outcomes provide a basis for understanding the mechanisms controlling osmotic induction of p38 MAPK activity during preimplantation development. The outcomes therefore further our knowledge of the regulation of intracellular signaling pathways within the mammalian embryo and understanding of how culture environments can affect embryo development. The preimplantation embryo is highly sensitive to the environment in which it develops. With the extension of embryo culture to the blastocyst stage emerging as a routine treatment paradigm in human infertility clinics, it becomes even more important that we study the short-term and long-term effects of culture on embryo development. Our results may therefore contribute to advancements in the development of improved embryo culture systems with the ultimate goal of increasing our ability to produce healthy embryos for embryo transfer.

Reverse Transcription and Polymerase Chain Reaction (RT-PCR)
Embryo total RNA was reverse transcribed (RT) using Random Primers (Invitrogen Life Technologies, Burlington, ON, Canada) and RNaseOUT™ Ribonuclease Inhibitor (Invitrogen Life Technologies, Burlington, ON, Canada) along with Sensiscript RT (Qiagen Inc., Mississauga, ON, Canada) according to the manufacturer's suggested protocol. Following one hour incubation, the sample was diluted to a concentration of 1 embryo equivalent per μL (embryo/μL) and subjected to PCR amplification of H2A histone family, member Z (H2afz), Luciferase, or both to determine the efficiency of the RNA extraction and reverse transcription prior to investigation of expression of the target genes through standard and Real-Time PCR amplification. Polymerase chain reaction (PCR) was performed using one embryo equivalent of cDNA from each stage per reaction in a 50 μL volume consisting of 1. were repeated a minimum of three times using cDNA prepared from embryos at each of the indicated stages and isolated from a minimum of three separate developmental series. Positive (tissue cDNA: heart, liver, and kidney) and negative control (no cDNA template) samples were included for each primer set in every experiment.

Custom TaqMan ® gene expression assay design for realtime PCR
The Custom TaqMan ® primer and probe sets for Ccm2 and Luciferase were designed using the Assays-by-Design File Builder program (Applied Biosystems, Foster City, CA, USA). The probe sequence for CCM2 was designed against a target site 847 bp into the full-length sequence obtained from the Ensembl mouse transcript database (Transcript ID: ENSMUST00000000388). The probe sequence for luciferase was directed against a target site 550 bp into the full-length sequence used to generate the Luciferase Con-trol RNA (Promega Corporation, Madison, WI, USA). The target site specifies an approximate location for generation of a TaqMan ® probe, and each target site was verified to be unique by performing BLAST ® analysis. Dual-labeled probes were synthesized (Applied Biosystems, Foster City, CA, USA) to contain the reporter dye 6-carboxyfluorescein (6-FAM) at the 5' end and a non-fluorescent quencher dye at the 3' end.

Quantitative real-time PCR analysis
Real-Time PCR reactions were performed using the ABI PRISM ® 7900HT sequence detection system (Applied Biosystems, Foster City, CA, USA) and TaqMan  Relative quantification of target gene expression levels was performed using the comparative C T (threshold cycle) method (ABI PRISM ® Sequence Detection System, version 2.1, Applied Biosystems, Foster City, CA, USA). The quantification was normalized to the control luciferase RNA levels [44]. Within the log linear phase region of the amplification curve, the difference between each cycle was equivalent to a doubling of the amplified product of the PCR. The ΔC T value was determined by subtracting the control C T value for each sample from the target gene C T value of the sample. Calculation of ΔΔC T used either the 1-cell or control sample as a standard. Fold-changes in the relative mRNA expression of the target gene were determined using the formula 2 -ΔΔCT .

Image analysis and quantification of immunofluorescence intensity
To quantify immunofluorescence results we employed the method developed and reported by [45]. All microscope and image capture settings remained constant during the digital capture of confocal micrographs between embryos and between treatment groups. Acquired micrographs were saved in TIFF image format and processed using Adobe Photoshop CS2 (Adobe Systems Inc., San Jose, CA, USA), for recognition, selection and separation of the desired chromogen signal. Quantitative analysis began by separation of the green image channel representing FITC fluorescence of phospho-MAPKAPK2 from the red (representing rhodamine-phalloidin-labelled F-actin) and blue (representing DAPI-stained nuclei) channels. The green channel image was converted to "Grayscale", discarding the other two color channels to ensure the signal remaining represented only FITC fluorescence. The overall lumi-nance of each pixel in the converted image corresponded to the intensity of FITC fluorescence. Each converted micrograph was then inverted so that gray and black pixels represented areas of FITC immunofluorescence on a white background and saved as a new TIFF image file for Scion Image analysis.
The Scion Image program, Version 4.0.3.2 (Scion Corporation, Frederick, MD, USA), is a freely distributed commercial software program that mimics the performance of NIH Image program for quantification of the chromogen signal strength [45]. The grayscale inverted micrographs were opened in Scion Image and following the methods of [45], the mean density of the chromogen signal strength (SS) of each image was measured and recorded. The average SS from at least three micrographs representing no primary antibody controls was subtracted from the SS values of the measured images from each treatment group to produce an adjusted relative SS value. Relative signal strength (RSS) values were obtained by a ratio comparing the adjusted SS values of each treatment to the average adjusted SS of the control group. . For Real-Time PCR analysis, embryos were treated for 3, 6, 9, 12, or 24 hours prior to determination of transcript levels by real-time RT-PCR as described above. For whole-mount indirect immunofluorescence assays, embryos were treated for times ranging from 10 to 30 minutes and then were processed for whole-immunofluorescence methods as described above.

Statistical analysis
Statistical analysis was performed using SPSS ® , Version 14.0 (SPSS Inc., Chicago, IL, USA) or SigmaStat ® 2.0 (Jandel Scientific Software, San Rafael, CA, USA) software packages. Real-Time PCR results are presented as the mean ± s.e.m. for relative mRNA transcript levels from three independent replicates. Real-time RT-PCR data were square-root transformed, subjected to one-way analysis of variance (ANOVA), and followed by Tukey's Multiple Comparison Test or non-parametric Mann-Whitney Rank Sum Test. Results from Scion Image analysis are presented as the mean ± s.e.m. for relative signal strength (RSS) from three independent replicates with the number of embryos in each treatment group indicated. All data was tested for homogeneity of variances by Levene's Test for Equality of Variances. In instances of equal variance, data was subjected to one-way ANOVA followed by Fisher's Least Significant Different (LSD) test for comparing three means of unequal group size. When equal variances were not observed, data was subjected to Welch's varianceweighted ANOVA followed by the Games-Howell (GH) Post-hoc Test for Multiple Comparisons, appropriate for situations of unequal (or equal) sample sizes and unequal or unknown variances. For all data analysis, P ≤ 0.05 was considered statistically significant.