Comparative Phosphoproteomics of Zebrafish Fyn/Yes Morpholino Knockdown Embryos*S

The coordinated movement of cells is indispensable for normal vertebrate gastrulation. Several important players and signaling pathways have been identified in convergence and extension (CE) cell movements during gastrulation, including non-canonical Wnt signaling. Fyn and Yes, members of the Src family of kinases, are key regulators of CE movements as well. Here we investigated signaling pathways in early development by comparison of the phosphoproteome of wild type zebrafish embryos with Fyn/Yes knockdown embryos that display specific CE cell movement defects. For quantitation we used differential stable isotope labeling by reductive amination of peptides. Equal amounts of labeled peptides from wild type and Fyn/Yes knockdown embryos were mixed and analyzed by on-line reversed phase TiO2-reversed phase LC-MS/MS. Phosphorylated and non-phosphorylated peptides were quantified, and significant changes in protein expression and/or phosphorylation were detected. We identified 348 phosphoproteins of which 69 showed a decrease in phosphorylation in Fyn/Yes knockdown embryos and 72 showed an increase in phosphorylation. Among these phosphoproteins were known regulators of cell movements, including Adducin and PDLIM5. Our results indicate that quantitative phosphoproteomics combined with morpholino-mediated knockdowns can be used to identify novel signaling pathways that act in zebrafish development in vivo.

In vertebrate gastrulation morphogenetic cell rearrangements coordinate the movement of the three germ layers, ultimately creating the basic body plan of the developing embryo. One of the major cell movements during gastrulation is convergence and extension (CE). 1 During CE movements, cells converge toward the midline of the embryo and extend along the anterior-posterior axis, a process that requires the precise regulation of cell polarity, cell migration, and cell division (1,2). Forward and reverse genetics screens have provided insight into the many signaling pathways that drive CE cell movements during gastrulation (3). The non-canonical Wnt pathway has a central role in CE movements, and mutations in Wnt11 and Wnt5 result in CE cell movement defects (4,5). Moreover gain or loss of function of positive and negative regulators of CE signaling result in severe gastrulation defects, indicating that precise control of cell movements is required (6,7). Non-canonical Wnt signaling acts to activate several downstream effectors including the small GTPases, Rho, Rac, and Cdc42. In turn, Rho kinases mediate cytoskeletal dynamics by regulating actin-myosin contractility (8 -10). In addition, several other regulators of Rho and Rac that are required for proper CE movements such as guanine nucleotide exchange factors (11,12) and Has2 (13) have been identified. Morpholino-mediated double knockdown of Fyn and Yes also results in impaired CE cell movements. These knockdowns are specific as they are rescued by co-expression of Fyn and Yes. The Fyn/Yes knockdowns phenocopy mutations in Wnt11 and Wnt5 in zebrafish. These phenotypes are rescued by active RhoA, indicating that Wnt11 and Fyn/Yes converge on RhoA downstream in the signaling pathway (14). How all of these signaling pathways cooperate in CE cell movements remains to be determined.
Proteomics is a powerful method to study signaling pathways and associated post-translational modifications that propagate signals, such as protein phosphorylation. Recent advances in phosphoproteomics have made the global characterization of the phosphoproteome of a complete organism feasible (15)(16)(17)(18). These studies have revealed an enormous amount of new phosphoproteins and phosphorylation sites and the relationships between those phosphoproteins. In addition, the global analysis of the phosphoproteome is used to study specific signaling pathways (19). However, to detect dynamic changes in the phosphoproteome, quantitative analysis is indispensable. Quantitation methods vary and are based on two-dimensional gel electrophoresis (DIGE), chemical approaches (ICAT and isobaric tags for relative and absolute quantitation (iTRAQ)) or stable isotope labeling by amino acids in cell culture (SILAC) (20). Especially SILAC has proven to be successful to study changes in the phosphoproteome upon induction of signal transduction pathways. This labeling method has the advantage that after labeling cell states are mixed together and all other sample handlings are carried out simultaneously. However, the use of SILAC is limited to single cell cultures. Other methods use stable isotope metabolic labeling in which all amino acids and peptides are labeled by growing cells on media containing exclusively 14 N/ 15 N (21-23). This method is widely used for lower organisms such as yeast, but it was also shown to be effective in labeling higher organisms such as Caenorhabditis elegans and Drosophila by feeding them heavy and light labeled yeast (24). It is feasible to use metabolic labeling approaches for higher organisms such as rats (25). However, these approaches are not straightforward and are expensive. In contrast, relatively simple chemical labeling approaches have recently become available that, in principle, allow for the quantitation of any biological sample of interest.
Dimethyl labeling provides a relatively straightforward, cheap, and simple strategy to label samples. In stable isotope dimethyl labeling, primary amine groups (lysine and amino termini) of peptides are labeled with formaldehyde through reductive amination (26 -28). Primary amines are isotopically labeled using either light formaldehyde or heavy (deuteriumcontaining) formaldehyde. Because trypsin is usually used as the enzyme for digestion, most peptides will contain at least one label because each amino terminus and each lysine becomes labeled.
Here we investigated signaling pathways and phosphorylation events involved in CE cell movements. Fyn/Yes knockdown impairs CE cell movements during gastrulation in zebrafish embryos. We questioned whether there are changes in the phosphoproteome of Fyn/Yes knockdown embryos compared with wild type (WT) embryos. We used a quantitative phosphoproteomics approach using stable isotope labeling by amination of tryptic peptides to detect changes in protein expression and protein phosphorylation upon morpholinomediated knockdown of Fyn and Yes in zebrafish embryos. This approach led to the identification of phosphoproteins that were differentially regulated upon knockdown of Fyn and Yes, whereas the majority of phosphoproteins were not changed, indicating that the observed changes were specific. Interestingly among the differentially phosphorylated proteins were many proteins with known functions in gastrulation cell movements, such as Wnt signaling and Rho kinase signaling components.

MATERIALS AND METHODS
Zebrafish Embryos and Microinjections-Zebrafish were kept, and the embryos were staged as described before (29). Antisense mor-pholinos were ordered from GeneTools (Philomath, OR). The morpholinos were targeted close to the start ATG of the respective cDNAs, and their sequences have been described before (14). The Fyn and Yes morpholinos (4 ng each) were injected at the one-cell stage.
Sample Preparation-Embryos, 30 wild type (approximately 100 g) and 30 Fyn/Yes knockdown (approximately 100 g) (24 h postfertilization (hpf)), were manually dechorionated and deyolked with deyolking buffer ( 1 ⁄2 Ginzburg Fish Ringer) without calcium (30). Subsequently embryos were lysed in 8 M urea, 25 mM ammonium bicarbonate containing 5 mM sodium phosphate, 1 mM potassium fluoride, 1 mM sodium orthovanadate, pH 8.2, and EDTA-free protease inhibitor mixture (Sigma). Homogenized lysates were centrifuged at 14,000 ϫ g to pellet cellular debris. Lys-C (Roche Diagnostics) was added to the lysate, and digestion was performed for 4 h at 37°C. Samples were reduced with DTT at a final concentration of 2 mM at 56°C; subsequently samples were alkylated with iodoacetamide at a final concentration of 4 mM at 20°C. The eluate was diluted to 2 M urea, 50 mM ammonium bicarbonate, and trypsin (Roche Diagnostics) was added. Digestion was performed overnight at 37°C.
Stable Isotope Labeling by Amination of Tryptic Peptides-Tryptic peptides were desalted, dried in vacuo, and resuspended in 100 l of triethylammonium bicarbonate (100 M). Subsequently 4 l of formaldehyde-H 2 (4% (v/v) in water) was added and vortexed for 2 min followed by the addition of 4 l of freshly prepared sodium cyanoborohydride (600 mM). The resultant mixture was vortexed for 60 min at room temperature. 16 l of ammonium hydroxide (1% in water) was added to consume the excess formaldehyde, and 5% formic acid (in water) was added to acidify the solution. For deuteromethylation labeling, formaldehyde-D 2 (4% (v/v) in water) was used. The light 4Hand heavy 4D-dimethyl-labeled samples were mixed in a 1:1 ratio based on total peptide amount determined by running an aliquot of the labeled samples on a regular LC-MS run and comparing overall peptide signal intensities.
Strong Cation Exchange-Strong cation exchange was performed using Zorbax BioSCX-Series II columns (0.8-mm inner diameter ϫ 50-mm long; particle size, 3.5 m), a Famos autosampler (LC Packings, Amsterdam, the Netherlands), a Shimadzu LC-9A binary pump, and an SPD-6A UV detector (Shimadzu, Tokyo, Japan). Prior to strong cation exchange (SCX) chromatography, the mixed and labeled digest was desalted using a small plug of C 18 material (3M Empore C 18 extraction disk) packed into a GELoader Tip as described previously (31). The eluate was dried completely by vacuum centrifugation and subsequently reconstituted in 20% acetonitrile and 0.05% formic acid. After injection, the first 10 min were run isocratically at 100% solvent A (0.05% formic acid in 8:2 (v/v) water/acetonitrile, pH 3.0) followed by a linear gradient of 1.3% min Ϫ1 solvent B (500 mM NaCl in 0.05% formic acid in 8:2 (v/v) water/acetonitrile, pH 3.0). A total number of 25 SCX fractions (1 min each, i.e. 50-l elution volume) were manually collected and dried in a vacuum centrifuge.
Two-dimensional Nanoflow HPLC-Nanoflow LC-MS/MS was performed by coupling an Agilent 1100 HPLC system (Agilent Technologies, Waldbronn, Germany) to an Orbitrap mass spectrometer (Thermo Electron, Bremen, Germany). The trap column consists of three separate precolumns: a 30-mm long ϫ 100-m-inner diameter Aqua C 18 precolumn followed by a 5-mm long ϫ 100-m-inner diameter TiO 2 precolumn followed by a 30-mm long ϫ 100-m-inner diameter Aqua precolumn. The "sandwich" precolumn is then coupled to a 200-mm long ϫ 50-m-inner diameter ReproSil-Pur C 18 -AQ analytical column. Peptides are trapped at 5 l/min in 100% solvent A (0.1 M acetic acid ϩ 0.13 M formic acid in water) on the first 30-mm C 18 trap column. The subsequent H 2 O/ACN gradient elutes and separates bound peptides using the analytical column at a flow rate of ϳ100 nl/min. Phosphorylated peptides will pass through the TiO 2 precolumn at this flow rate and bind (32). All other peptides with no

Phosphoproteomics of Zebrafish Fyn/Yes Knockdown Embryos
TiO 2 affinity are chromatographically separated at ϳ100 nl/min in a 100-min gradient from 0 to 40% solvent B (80% acetonitrile, 0.1 M acetic acid, and 0.13 M formic acid). Elution of phosphorylated peptides is achieved by injection of 30 l of 100 mM ammonium hydrogen bicarbonate, pH 9.0 (adjusted with ammonia) containing 10 mM sodium phosphate, 5 mM sodium orthovanadate, and 1 mM potassium fluoride followed by an injection of 20 l of 5% formic acid. During a second H 2 O/ACN gradient phosphopeptides are separated using the 200-mm analytical column at ϳ100 nl/min in a 100-min gradient from 0 to 40% solvent B (80% acetonitrile, 0.1 M acetic acid, and 0.13 M formic acid). The eluent was sprayed via distal coated emitter tips (New Objective) butt-connected to the analytical column. In between the high voltage supply, the Orbitrap, and the electrospray needle an additional 33-megaohm resistor was placed to reduce ion current.
Mass Spectrometry-The mass spectrometer was operated in datadependent mode, automatically switching between MS, MS/MS, and neutral loss-driven MS 3 acquisition. Full-scan MS spectra (from m/z 300 to 1500) were acquired in the Orbitrap with a resolution of 60,000 at m/z 400 after accumulation to a target value of 500,000. The three most intense ions at a threshold above 5000 were selected for collision-induced fragmentation in the linear ion trap at a normalized collision energy of 35% after accumulation to a target value of 10,000. The data-dependent neutral loss settings were chosen to trigger an MS 3 event after a neutral loss of 49 Ϯ 0.5 m/z units was detected in the most intense fragment ion.
Data Analysis-All MS 2 and MS 3 spectra were converted to single .dta files using Bioworks 3.3. An in-house developed Perl script was used to assign the original and accurate parent mass to all MS 3 spectra, enabling an accurate mass database search. All first and second LC-MS runs of all SCX fractions were searched using an in-house licensed MASCOT search engine (MASCOT (version 2.1.0) software platform (Matrix Science, London, UK)) against the zebrafish International Protein Index (IPI) database version 3.27 (41,506) with carbamidomethylcysteine and "light" methylation of peptide amino termini and lysine residues set as fixed modifications and protein N-acetylation, oxidized methionines, phosphorylation of serine or threonine, and "heavy" methylation of peptide amino termini and lysine residues set as variable modifications. Trypsin was specified as the proteolytic enzyme, and up to two missed cleavages were allowed. The mass tolerance of the precursor ion was set to 5 ppm, and that of fragment ions was set to 0.9 Da.
Quantitation was performed using an in-house dimethyl-adapted version of MSQuant (SourceForge, Inc.). Briefly peptide ratios were obtained by calculating the extracted ion chromatograms (XICs) of the light and heavy forms of the peptide using the monoisotopic peaks only. The total XIC for each of the peptide forms was obtained by summing the XIC in consecutive MS cycles for the duration of their respective LC-MS peaks in the total ion chromatogram using FT-MS scans. This total XIC was then used to compute the peptide ratio. Heavy and light labeled peptides were found to largely co-elute. Quantified proteins were normalized against the log 2 of the median of all peptides quantified. An in-house developed program was used to determine standard deviation, and a Student's t test was performed to detect significant regulation (p Յ 0.05). Proteins for which the standard deviation and p value could be determined and with a MASCOT score Ն80 were accepted (listed in supplemental Table 1). At least one unique peptide per protein was identified, eliminating redundancy in the data set.
Individual MS/MS spectra from phosphopeptides were accepted for a MASCOT score Ն40 (p value Յ0.005). Justification of the threshold was as described before (18). All identified phosphopeptides that were differentially phosphorylated were manually validated.
To determine the false discovery rate in the data sets, we performed a MASCOT search against the decoy database. For a MASCOT peptide score of 40 (p Յ 0.005), the false discovery rate was estimated to be 0.41%. All raw data and annotated spectra were submitted to the PRIDE database.

RESULTS
Analysis of Differential Phosphorylation-To investigate signaling pathways that have a role in the regulation of CE cell movements, we set out to evaluate the changes in the phosphoproteome after morpholino-mediated knockdown of Fyn and Yes in 24-hpf zebrafish embryos. To quantitate the differences, we used stable isotope labeling by reductive amination of peptides derived from the two experimental conditions and on-line automated RP-TiO 2 -RP-LC-MS/MS. The experimental procedures are described in detail under "Materials and Methods," and a schematic overview of the complete experimental approach is depicted in Fig. 1.
Zebrafish embryos were injected with Fyn/Yes morpholinos at the one-cell stage, inducing specific knockdown of protein expression. As shown previously, Fyn/Yes knockdown induced severe gastrulation defects, characterized by reduced extension of the embryonic axis and improper extension of the forebrain anterior to the eyes. The Fyn/Yes knockdown phenotype is specific because it is rescued by co-injection of synthetic RNA encoding Fyn and Yes (14). Moreover Fyn/Yes knockdown resulted in specific changes in protein tyrosine kinase activity profiles that were also rescued by co-injection of Fyn and Yes RNA (33). A representative picture of a WT and a Fyn/Yes knockdown embryo at 24 hpf is shown in Fig. 1A. At 24 hpf, 30 WT and 30 Fyn/Yes knockdown embryos were manually dechorionated, deyolked, pooled, and lysed. Protein mixtures were digested with trypsin and subsequently labeled with normal formaldehyde (CH 2 O), light, for the WT embryos or deuteroformaldehyde (CD 2 O), heavy, for the knockdown embryos. Samples were mixed in a 1:1 ratio (total peptide content). The resulting peptide mixture was first separated and fractionated by SCX chromatography, which, under acidic conditions, causes phosphopeptides to elute earlier than regular tryptic peptides (34). Resulting SCX fractions were subsequently subjected to our on-line RP-TiO 2 -RP-LC-MS/MS enrichment setup (18,32). The resulting peptides were analyzed by mass spectrometric detection using a linear ion trap (LTQ) orbitrap, and the MS/MS spectra were searched against the IPI zebrafish database using MASCOT to identify peptide sequence and phosphorylation sites. Finally peptide/protein ratios were determined using the intensities of heavy and light peptide pairs.
More than 1400 proteins were identified, quantified, and manually checked (supplemental Table 1). Quantified proteins were normalized against the log 2 of the median of all peptides quantified. A Student's t test was performed to detect significant regulation. 555 proteins were shown to be significantly regulated (p Յ 0.05). Among the up-regulated proteins, Vitellogenins were found to be up-regulated in the Fyn/Yes knockdown. It is likely that this up-regulation of Vitellogenins is an artifact of the manual deyolking procedure, and therefore we omitted Vitellogenins from further analyses. Skeletal muscle myosins and muscle actin were shown to be significantly down-regulated in the Fyn/Yes knockdown. Actomyosin contractility is known to be involved in gastrulation cell movements, which are controlled by Wnt signaling (35,36). Although these results are potentially interesting, we focused on changes in protein phosphorylation upon knockdown of Fyn and Yes because these most likely reflect differences in signaling underlying CE cell movements.
All of the phosphopeptides that we detected were quantified by MSQuant and manually verified. Phosphopeptides and phosphoproteins were subsequently filtered based on their ratio and MASCOT score (Ն40 and p value Յ0.005) and checked manually. After manual verification of the MS spectra and normalization, 348 phosphoproteins were validated. Table I shows a selection of differentially phosphorylated phosphopeptides that were identified with high confidence. We also determined the ratio of control, non-phosphorylated peptides that were derived from the proteins corresponding to the phosphopeptides. This allowed us to determine the relative protein expression level and the phosphorylation level and thus to discriminate between changes in protein levels and in phosphorylation state. The ratio of phosphorylation was determined multiple times, and the average is given in Table I as well as the number of peptides that were used to determine the ratio, the standard deviation, and p value. Similarly the average ratio of protein levels, number of peptides, and respective standard deviation and p value are provided. Some phosphopeptides were only identified and quantified once, whereas the corresponding protein ratio was based on multiple non-phosphorylated peptides (Table II, upper panel). Some proteins were identified and quantified based on phosphopeptides without information on the relative protein levels (Table II). In the latter cases we cannot exclude protein up-or down-regulation, but the absolute amount of phosphopeptide did change. A list of all phosphoproteins and the ratio of phosphorylation in Fyn/Yes knockdown and WT is available as supplemental Table 2. A list of all identified phosphopeptides (MASCOT score Ն40 and p value Յ0.005) including their MASCOT score is shown in supplemental Table 3. All annotated spectra were uploaded to the PRIDE database. Spectra of all phosphopeptides in Tables I and II  The differences in phosphorylation ratio of the phosphoproteins we identified are not as high as reported in other recent studies (37)(38)(39); this can be explained by the fact that we analyzed total embryo lysates. The Fyn/Yes knockdown will only affect certain cell types, and therefore the detected effect is less pronounced than in homogenous tissue culture cells that are treated with large amounts of growth factor. Nevertheless we detected changes in phosphorylation of proteins.
Decrease in Phosphorylation-Among the peptides with a decrease in phosphorylation upon Fyn/Yes knockdown, we identified the phosphopeptide QMpSPGSGIPSK (where pS is phosphoserine) from insulin-like growth factor 2 mRNAbinding protein 1 (IGF2BP1) with high confidence (MS 2 spectrum in Fig. 2A). The phosphopeptides were less abundant in the Fyn/Yes knockdown (0.52 Ϯ 0.10). To monitor IGF2BP1 protein expression, we quantified the ratio of control nonphosphorylated peptides and found that the protein ratio was 0.93 Ϯ 0.19 (Table I). Indeed equal amounts of non-phosphorylated peptide MILEIMNQEAK were detected in Fyn/Yes knockdown and WT lysates (Fig. 2C). Fig. 2 shows both the MS spectrum and the XIC of the phosphorylated (Fig. 2B) and non-phosphorylated (Fig. 2C) peptide. It was evident that the MS spectrum of the phosphorylated peptide in the Fyn/Yes knockdown (gray) was lower in intensity than in the WT (black). This was reflected in the XIC (Fig. 2B). In the control, nonphosphorylated peptide, the intensities are similar; this was also reflected in the XIC (Fig. 2C), indicating that phosphorylation but not abundance of IGF2BP1 protein levels was affected by Fyn/Yes knockdown.
IGF2BP1 is a member of the VICKZ family of RNA-binding proteins, and it regulates mRNA nuclear export, localization, stability, and translation. IGF2BP1 binds to the 5Ј-untranslated region of the insulin-like growth factor 2 mRNA thereby regulating its subcellular localization and translation. VICKZ proteins are highly conserved among all vertebrates and are expressed in embryonic cells. It is suggested that VICKZ proteins play a role in cell migration via the ability to localize RNA. Many human tumors overexpress VICKZ proteins (40). Xenopus xVICKZ3 is required for migration of cells that form the roof plate of the neural tube and for migration of neural crest cells (41). Taken together, IGF2BP1 likely has a role in cell migration during zebrafish development.
The phosphopeptide TpSMGGTQQQFVEGVR, corresponding to ␤-catenin, was also identified in our screen. The protein ratio of ␤-catenin did not change significantly upon Fyn/Yes knockdown (0.96 Ϯ 0.03), whereas the identified phosphopeptides showed a decrease (0.73 Ϯ 0.13) ( Table I). The identified site is a reported Akt phosphorylation site (42). Akt is a direct downstream effector of phosphatidylinositol 3-kinase (PI3K). It is noteworthy that PI3K is involved in CE cell movements as treatment of zebrafish embryos with the PI3K inhibitor LY294002 induces defects in CE cell movements (43).
Another phosphopeptide that showed a significant decrease in phosphorylation was QEVVp(TT)PPPPSVGK (where pT is phosphothreonine) from Adducin 3 (0.76 Ϯ 0.09). The protein ratio remained unchanged upon Fyn/Yes knockdown (1.01 Ϯ 0.04) ( Table I). The identified phosphorylation site of Adducin has not been reported previously, and the upstream kinase is unknown. Adducin is a membrane skeletal protein that binds to Actin. Previously it has been shown that phos-

TABLE I Down-and up-regulated phosphopeptides after Fyn/Yes knockdown
Selected phosphopeptides and proteins showing differential phosphorylation are shown. Included in the table are phosphopeptide ratio, number of phosphopeptides used for quantitation, standard deviation (StDev), and p value. Also included are the protein ratio based on non-phosphorylated peptide, number of non-phosphorylated peptides used for quantitation, standard deviation, and p value. ‡ Calculated with only non-phosphorylated peptides.

Phosphoproteomics of Zebrafish Fyn/Yes Knockdown Embryos
phorylation of Adducin by protein kinase C and Rho kinase is implicated in cell motility, spreading, and migration (44). Therefore, it is likely that Adducin has a role in CE cell movements. Whether phosphorylation of the newly identified site also has a role in this process remains to be determined.
Vimentin protein expression and phosphorylation were both shown to be down-regulated upon Fyn/Yes knockdown (0.56 Ϯ 0.04 and 0.68 Ϯ 0.06, respectively) ( Table I). Vimentin shows a high degree of sequence homology between species and is highly phosphorylated. Phosphorylation of Vimentin has been related to its function in cell adhesion, migration, and signaling (45), and reduced Vimentin protein levels and phosphorylation may therefore have a role in CE cell movements.
Among the peptides that were less phosphorylated upon Fyn/Yes knockdown, we identified a single phosphopeptide of Sly1 (Suppressor of Ypt) (FY/WT ratio, 0.65; Table II) encompassing a phosphorylation site that had not been reported before. Sly1 protein levels were not affected by Fyn/ Yes knockdown (protein ratio, 1.13 Ϯ 0.20; Table II). Sly1 is the vertebrate homologue of Sly1p in yeast. It is known to bind and regulate syntaxin 5 in vesicle transport from the endoplasmic reticulum to the Golgi (46). It also interacts with the small GTPase Rab (Ypt1p in yeast), a key regulator of protein transport in eukaryotic cells. Sly1 is involved in vesicle motility and docking as well as in membrane remodeling and fusion. In zebrafish, Sly1 was shown to be necessary for cell proliferation in fin regeneration (47). It will be interesting to investigate the role of phosphorylation in Sly1 function.
Phosphorylation of the Kaiso phosphopeptide SQIQPD-SPNTASLpSPK was shown to decrease upon Fyn/Yes knockdown (FY/WT ratio, 0.76 Ϯ 0.08), but we did not detect non-phosphorylated control peptides for Kaiso (Table II). Kaiso is a dual specificity transcriptional repressor that was identified to associate with p120-catenin. Kaiso is required for Xenopus gastrulation movements, and depletion of Kaiso results in the increased expression of non-canonical Wnt11, a regulator of CE cell movements (48). The identified phos-

TABLE II Quantitation of single phosphopeptides and of phosphopeptides without protein ratios after Fyn/Yes knockdown
The upper panel shows selected phosphopeptides and proteins showing differential phosphorylation based on only one phosphopeptide. Included in the table are phosphopeptide ratio and total protein ratio based on non-phosphorylated peptides including number of non-phosphorylated peptides used for quantitation, standard deviation (StDev), and p value. The lower panel shows selected phosphopeptides showing differential phosphorylation, but non-phosphorylated peptides were not identified. Included in the table is phosphopeptide ratio including number of phosphorylated peptides used for quantitation, standard deviation, and p value.
* p value based on all phosphopeptides. X Standard deviation, p value, and protein ratio cannot be determined; ratio is based on one (phospho)peptide. ‡ Calculated with only non-phosphorylated peptides.
phopeptide is predicted to become phosphorylated by CDK5 (Scansite), and it will be interesting to see whether phosphorylation of this site results in functional changes in Kaisomediated signaling in gastrulation cell movements.
Increase in Phosphorylation-The phosphopeptide SQPp-SPNSISSPASNSAPKPSSGR from PDLIM5 was shown to be increased upon Fyn/Yes knockdown (1.53 Ϯ 0.01; Table I; see also MS 2 spectrum in Fig. 3A). The protein ratio, based on control non-phosphorylated peptides, also showed a slight up-regulation compared with WT although less strong than the phosphopeptides (1.22 Ϯ 0.17; Table I). The MS spectrum of the phosphorylated peptide in the Fyn/Yes knockdown (gray) was much higher in intensity than in the WT (black). This was also reflected in the XIC (Fig. 3B). In one of the non-phosphorylated peptides, LIEDTEDWHPR, the intensities were similar, and this was reflected in the XIC (Fig. 3C). PDZ/LIM proteins have important roles in various cellular processes. PDLIM5 belongs to the Enigma subfamily of PDZ/LIM proteins that contain one PDZ and three LIM domains (49). In zebrafish, Enigma knockdown results in embryonic lethality shortly after the end of gastrulation, and embryos show elongation defects and disorganized somites, suggesting a role for Enigma in cytoskeleton and somite organization during embryonic development (50).
Two phosphopeptides of Gravin/AKAP12 were identified. The ratio of one of the phosphopeptides increased (1.75), whereas the other phosphopeptide remained unchanged (1.05). The protein level increased somewhat (1.29 Ϯ 0.29) upon Fyn/Yes knockdown (Table II). Gravin was recently shown to be involved in axis elongation during zebrafish gastrulation. It is thought to function as a scaffolding protein that promotes cell changes. Upon morpholino-mediated knockdown of Gravin, cells migrate normally to the dorsal side of the embryo but fail to extend properly along the anterior-posterior axis (51). It is conceivable that Gravin mediates some of the effects of Fyn and Yes in gastrulation cell movements.
Two different phosphopeptides that we identified matched the Ras-GTPase-activating protein (GAP) SH3 domain-binding protein 2 (IPI00500409 and IPI004873772). Phosphorylation of one of these proteins was decreased (0.79 Ϯ 0.02), whereas phosphorylation of the other was increased (1.30 Ϯ 0.03) ( Table I). Control peptides of both of these proteins were slightly increased in response to Fyn/Yes knockdown (1.13 Ϯ 0.03 and 1.11 Ϯ 0.16, respectively; Table I). Alignment of the sequences of these two Ras-GAP SH3-binding protein 2 proteins indicated that they share 63% identity, and the phosphopeptides are clearly distinct. Ras-GAPs enhance the hydrolysis of bound GTP to GDP, leading to inactivation of Ras, and it is likely that these two Ras-GAP SH3 domain-binding proteins regulate Ras-GAP activity. Yet the role of (de)phosphorylation of these proteins in the process remains to be determined, and the finding that phosphorylation of one of these proteins is up-regulated and phosphorylation of the other is down-regulated in response to Fyn/Yes knockdown is intriguing.
Two phosphopeptides of two isoforms of brain-specific angiogenesis inhibitor 1-associated protein 2-like1 (BIAP2L1A/ BIAP2L1B) were identified, and phosphorylation of both peptides was enhanced upon Fyn/Yes knockdown (1.35 Ϯ 0.10 and 1.36 Ϯ 0.11, respectively) ( Table II). Only the phosphopeptides could be identified, so up-regulation of the protein itself cannot be excluded. The phosphosite in BIAP2L1A is a potential substrate of GSK-3 as predicted by Scansite. The BIAP2L1A/B protein contains an evolutionarily conserved insulin receptor tyrosine kinase substrate p53 homology domain (IRSp53/MIM) that is known to bundle actin filaments and interact with the small GTPase Rac (52).
Two unique phosphopeptides of DOCK7 were enhanced upon Fyn/Yes knockdown (1.34 and 1.34 Ϯ 0.18), but upregulation of the protein itself cannot be excluded because control peptides were not identified in the sample (Table II). DOCK7 is a Rac activator and is a key regulator of axon formation. DOCK7 and Rac activation lead to the phosphorylation and subsequent inactivation of stathmin/Op18, a microtubule-destabilizing protein (53,54). DOCK7 has been proposed to be a guanine nucleotide exchange factor not only for Rac but also for Rheb, transforming the GDP-bound Rheb into the GTP-bound state, although there is no direct evi-dence (55). From our data it is evident that phosphorylation of this regulator of small GTPases is affected by Fyn/Yes knockdown.
No Change in Phosphorylation upon Fyn/Yes Knockdown-It is noteworthy that the majority of phosphopeptides and protein levels did not change upon Fyn/Yes knockdown. The ratios of phosphorylation of 207 of the 348 phosphoproteins (62%) were shown to be distributed between 0.8 and 1.2 (Fig. 4). The phosphopeptide NLpSPGAVDTEVR (MS 2 spectrum shown in Fig. 5A) from the coatomer protein complex subunit ␣ (COPA) did not show significant changes upon Fyn/Yes knockdown (1.04 Ϯ 0.06) ( Table I and Fig. 5B). The peptide ratio of control peptides from COPA was also unchanged (1.21 Ϯ 0.05) ( Table I). For instance, the non-phosphorylated control peptide CPLSGACYCPK is shown in Fig.  5C. Both phosphorylated and non-phosphorylated peptides from COPA, which is involved in cell cycle control, were not affected upon knockdown of Fyn and Yes. DISCUSSION Here we probed in vivo changes in protein phosphorylation upon knockdown of Fyn and Yes using RP-TiO 2 -RP-LC-MS/ MS. For quantification, we used stable isotope labeling by reductive amination of peptides as an alternative to stable isotope labeling of an entire zebrafish (embryo). Our procedure was efficient and robust, and the ratio of most phos- phopeptides was between 0.8 and 1.2, i.e. phosphorylation of most proteins did not change significantly upon Fyn/Yes knockdown (Fig. 4). Yet a subset of phosphopeptides was differentially phosphorylated between wild type and Fyn/Yes knockdown embryos. For the identified phosphopeptides different, non-phosphorylated peptides of the same protein were quantified. This internal control for protein expression boosted confidence in the observed differences in phosphorylation. For some phosphopeptides, corresponding nonphosphorylated peptides were not detected (Table II). This is probably because of high enrichment of phosphopeptides by the TiO 2 column. Non-phosphorylated peptides of these proteins were not detected in the unbound fraction because of undersampling in the mass spectrometer. Low abundance phosphoproteins were therefore only quantified based on their phosphopeptides.
Taken together, stable isotope reductive amination is a fast and reliable alternative for metabolic stable isotope labeling of whole animals, and it can be used to quantify differences in protein phosphorylation between experimental samples. Moreover the reductive amination method is amenable to quantify and compare differences between any two sets of proteins or protein mixtures, and it can be used as an inexpensive alternative to SILAC and stable isotope labeling of whole animals (27,28).
Significant changes in both protein expression and protein phosphorylation were detected. We focused on the changes in phosphorylation upon Fyn/Yes knockdown because these are likely to represent the signaling pathways that have a role in CE cell movements. The changes in phosphorylation ratios were not as high as reported in previous studies (37)(38)(39); this presumably is the consequence of the mixed cell population that we used in this study as opposed to the homogenous cell cultures used in previous studies. The Fyn/Yes knockdown will not affect every cell type in the developing zebrafish embryo, and hence the differences in protein phosphorylation will be masked partially by the contribution of cells that are not affected by Fyn/Yes knockdown.
Nevertheless we detected significant changes in phosphopeptides. The phosphorylation of 141 of the 348 phosphoproteins was altered by Fyn/Yes knockdown (72 up and 69 down) (supplemental Table 2). Interestingly a subset of the phosphoproteins that were regulated by Fyn/Yes knockdown were known regulators of gastrulation cell movements (Kaiso, PDLIM5, and Gravin), actin cytoskeleton reorganization (Adducin and BIAP2L1A/B), and Wnt signaling pathways (␤-catenin, Sox-3, and Kaiso) or were implicated in small GTPase signaling (Sly1, Ras-GAP SH3BP2, and DOCK7) (Tables I and  II). The high amount of small GTPase regulators that showed a change in phosphorylation suggests a change in the balance of Rho and Rac activity upon Fyn/Yes knockdown. This is consistent with previous work demonstrating that CE cell movements are tightly regulated by small GTPases (3). CE cell movement defects in gastrulation upon Fyn/Yes knockdown are rescued by active RhoA (14), demonstrating that Rho signaling indeed is downstream of Fyn and Yes in gastrulation cell movements. Whether and, if so, how changes in the phosphorylation state of these proteins affects their activities remain to be determined. Nevertheless it is striking that differences in phosphorylation were detected in these known regulators of CE cell movements upon Fyn/Yes knockdown. The phosphorylation of many other proteins was altered upon Fyn/Yes knockdown as well. We analyzed our entire data set for predictions of involvement of kinases and kinase networks by bioinformatics approaches using the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) (European Molecular Biology Laboratory). These attempts did not result in a wealth of data most likely because of poor annotation of the zebrafish IPI database. This will certainly improve in the coming years, and with our data publicly deposited it will be interesting to see what the function is of the phosphoproteins that we identified in CE cell movements and/or Fyn/Yes signaling.
Fyn and Yes are members of the Src family of tyrosine kinases. However, we did not detect changes in tyrosine phosphorylation of downstream targets. This is primarily because of the low abundance of tyrosine phosphorylation in general and the mixed cell population used in this endogenous sample. We have shown previously using an immunoaffinity approach with anti-Tyr(P) antibodies that tyrosine phosphorylation is detectable in early zebrafish embryos (56). However, the amount of material that was required for those experiments (ϳ2000 embryos per sample) would either require a significant scale-up of the morpholino injections or further refinement of the technology to allow detection of phosphotyrosine-containing proteins in Fyn/Yes knockdown embryos. The aim of the work described here was not to detect direct substrates of Fyn and Yes but rather to identify components of signaling pathways that have a role in CE cell movements.
Future research will be aimed at elucidating the role of Fyn and Yes signaling in gastrulation cell movements. It will be interesting to compare signaling between the Fyn/Yes knockdown embryos and non-canonical Wnt knockdown embryos. We have found recently that there is a significant overlap in the kinase activity profiles of Fyn/Yes and Wnt11 knockdowns (33), suggesting that indeed Fyn/Yes and Wnt11 signaling overlap. The large number of signaling proteins and novel proteins with altered phosphorylation upon Fyn/ Yes knockdown warrants further investigation of the role of these proteins and of their phosphorylation in CE cell movements. Quantitative phosphoproteomics on zebrafish knockdown embryos as described here provides many new leads for the analysis of the role of these proteins in gastrulation cell movements.