Toward defining the phosphoproteome of Xenopus laevis embryos

Abstract

Phosphorylation is universally used for controlling protein function, but knowledge of the phosphoproteome in vertebrate embryos has been limited. However, recent technical advances make it possible to define an organism's phosphoproteome at a more comprehensive level. Xenopus laevis offers established advantages for analyzing the regulation of protein function by phosphorylation. Functionally unbiased, comprehensive information about the Xenopus phosphoproteome would provide a powerful guide for future studies of phosphorylation in a developmental context. To this end, we performed a phosphoproteomic analysis of Xenopus oocytes, eggs, and embryos using recently developed mass spectrometry methods. We identified 1,441 phosphorylation sites present on 654 different Xenopus proteins, including hundreds of previously unknown phosphorylation sites. This approach identified several phosphorylation sites described in the literature and/or evolutionarily conserved in other organisms, validating the data's quality. These data will serve as a powerful resource for the exploration of phosphorylation and protein function within a developmental context.

INTRODUCTION

Many proteins are regulated by the reversible phosphorylation of specific serine, threonine, or tyrosine residues, including proteins that control the diverse and complex cellular processes that govern vertebrate development. For example, during Xenopus embryogenesis, the β-catenin protein preferentially accumulates in the nuclei of cells destined to form the dorsal and anterior structures of the embryo. Regulated phosphorylation of specific serine residues at the N-terminus of the β-catenin protein in response to Wnt signaling are required for this spatially controlled accumulation of β-catenin (Amit et al., 2002; Liu et al., 2002). Phosphorylation of CPEB (Cytoplasmic Polyadenylation Element Binding), a protein that binds to the 3’ Untranslated Regions (UTRs) of maternal mRNAs, leads to the translational activation of these stored mRNAs in response to the appropriate developmental cues (Hake and Richter, 1994; Mendez et al., 2000; Sarkissian et al., 2004). A myriad of other examples exist that establish phosphorylation as a universal mechanism for controlling the functions or levels of proteins and the activity of key cellular pathways.

Xenopus embryos have served as a powerful model for examining the role of protein phosphorylation in the pathways and processes governing early vertebrate development. However, the studies of many important phosphorylation events described in the literature were guided initially by discoveries made in other systems. That is, although phosphorylation plays clear and important roles during the development of Xenopus embryos, only a small number of the phosphorylation events examined in this model organism have been identified directly from studies of this organism. Given the probability that many more phosphorylation events critical to vertebrate development remain undiscovered, a large and functionally unbiased data set identifying both phosphoproteins and the precise sites of their phosphorylation would provide a powerful foundation for future functional studies. Here we begin to define the breadth and potential temporal and spatial dynamics of protein phosphorylation within the context of developing Xenopus embryos using recent advances in mass spectrometry and protein preparation that enable the accurate identification of phosphopeptides.

The methodological advances that facilitate the identification of specific phosphopeptides from complex mixtures include the use of metal-ion chromatography to enrich phosphorylated peptides and improvements in peptide fragmentation. In particular, the development of an electron-based dissociation method, electron capture dissociation (ECD) (Zubarev et al., 1998), and more recently the related electron transfer dissociation (ETD) (Coon et al., 2004; Syka et al., 2004; Coon et al., 2005a; Coon et al., 2005b) has allowed efficient cleavage of peptide backbone bonds without regard to phosphorylation and that leave phosphorylated residues intact (Good and Coon, 2006; Chi et al., 2007; Good et al., 2007; Khidekel et al., 2007; Lecchi et al., 2007; Molina et al., 2007; Swaney et al., 2007; Phanstiel et al., 2008). Thus, the new methods allow both the phosphopeptide and the precise site(s) of phosphorylation to be identified.

We used ETD to identify 1,441 different phosphorylation sites present on 654 different proteins from Xenopus oocytes, eggs, and embryos. Additional analysis of gastrula stage embryos identified phosphorylation sites on an additional 541 proteins. The Xenopus phosphoproteins were involved in a diverse array of cellular functions. In particular, many of the identified phosphoproteins are known regulators of embryonic development including translational regulatory proteins, signaling proteins and transcription factors. Our results provide a data-rich foundation for exploiting the experimental advantages of Xenopus to examine regulatory phosphorylation events that direct vertebrate embryonic development and cellular function(s).

RESULTS AND DISCUSSION

Xenopus phosphopeptide analysis with tandem mass spectrometry

To identify Xenopus phosphoproteins and phosphorylation sites, proteins from oocytes, eggs, and embryos were analyzed (Fig. 1A). Proteins were extracted from oocytes, eggs, stage 7 and stage 10.5 embryos (Fig. 1B) and treated with trypsin to generate peptides. The peptides were modified to create their corresponding methyl ester derivatives (Ficarro et al., 2002). This peptide modification serves to eliminate non-specific binding during subsequent steps. Phosphopeptides were isolated by immobilized metal affinity chromatography (IMAC) (Ndassa et al., 2006; Ficarro et al., 2002) and separated by reversed-phase chromatography before elution into the mass spectrometer by electrospray ionization (Martin et al., 2000). The mass spectrometer was set to continuously cycle through the acquisition of a full MS spectrum followed by subsequent ETD-MS/MS scans of the ten most abundant phosphopeptide cations observed. The resulting spectra were analyzed using the Open Mass Spectrometry Search Algorithm (OMSSA) (Geer et al., 2004) to identify the relevant phosphopeptides. Identified peptides with an e-value greater than zero were validated with a target-decoy searching method (Käll et al., 2008) and identified phosphopeptides with less than 5% false discovery rate were included in this current analysis. We report all hits at this false discovery rate to include potentially relevant regulatory proteins, and we feel it is warranted given the potential for functional analysis and validation in Xenopus. In addition, we have partitioned the data such that phosphopeptides with less than 0.5% false discovery rate have been indicated (SI Table 1).

Figure 1. Phosphoproteomic analysis of Xenopus oocytes, eggs and embryos.

(A) Diagram depicting the stages of Xenopus laevis development examined. (B) Diagram of the phosphoproteomic strategy used to analyze proteins from Xenopus laevis oocytes, eggs and embryos. (C) Freon extraction of protein samples preferentially removes vitellogenin proteins. Increasing amounts of protein samples from Xenopus eggs were analyzed by SDS-PAGE before (lanes 1 to 3) and after Freon extraction (lanes 4 to 6). Two of the major vitellogenin proteins lipovitellin 2B (31 kDa) and lipovitellin I (120 kDa) were removed upon treatment with Freon.

In our initial analyses the predominant phosphoproteins identified were vitellogenins; the abundant phosphoproteins stored in the cytoplasm of oocytes, eggs, and embryos for nutritive purposes during development. The presence of the vitellogenins prevented the efficient identification of other phosphoproteins. Therefore, we added another step to our protocol. Specifically, prior to trypsin treatment, the vitellogenins were removed using Freon (Fig. 1C) (Evans and Kay, 1991).

Identification of phosphoproteins from Xenopus oocytes, eggs and embryos

Protein samples from Xenopus oocytes, eggs, stage 7 and stage 10.5 embryos were analyzed to identify 1,441 sites of phosphorylation from 1,133 phosphopeptides derived from 654 proteins (Fig. 1B, Table 1, SI Table 1). As predicted from proteomic studies in other systems (Olsen et al., 2006) the most prominent phosphorylation events occurred on serine (76% frequency) followed by threonine (21% frequency). Tyrosine phosphorylation events were rare, constituting only 2% of the sites identified (Table 1). These percentages did not change significantly over the course of development. Most of the phosphopeptides possessed a single phosphorylated residue (Table 1). However, 22% of the peptides contained at least two and as many as six phosphorylated residues (Tables 1 and 3). Such multi-site phosphorylations often occur in regulatory domains of proteins that must integrate multiple signaling inputs (Gingras et al., 1999; Liu et al., 2002; Zeng et al., 2005). For example, we observed that a peptide derived from the APC protein was simultaneously phosphorylated on six closely spaced residues (Table 3). APC is a component of the Wnt signaling pathway and activation of Wnt signaling leads to the phosphorylation of APC on multiple residues (Brocardo and Henderson, 2008).

Table 1. Identification of Xenopus phosphoproteins from oocytes, eggs and embryos Phosphorylation events sorted by stage.

Summary of data from the phosphoproteomic analysis of Xenopus oocytes, eggs, stage 7 embryos and stage 10.5 embryos. All the results for individual proteins are found in SI Table 1.

		Oocyte	Egg	Blastula	Gastrula	All Stages
STY		298	350	462	331	1441
S		80%	73%	77%	75%	76%
T		18%	24%	21%	22%	21%
Y		2%	2%	2%	3%	2%
Unique peptides		241	272	360	260	1133
phosphorylation events on each peptide	1	193	208	280	204	885
	2	46	56	64	42	208
	3	3	10	18	13	44
	4	1	0	0	1	2
Unique proteins		195	224	281	199	654

Table 3.

Multiphosphorylated peptides.

Xenopus protein	Accession	Phosphosites:	Phosphopeptide:
adenomatous polyposis coli (APC)	AAB41671	S2523, S2525, S2527	HDIsRsHsESPSR
adenomatous polyposis coli (APC)	AAB41671	S2804, T2806, T2808, T2809, S2811, S2814	KRDsKtEttDsSGsQSPK
B-Raf (MGC86346)	AAZ06667	S355, S357	sSsAPNVHINTIEPVNIDDLIQDP
Casein kinase epsilon	NP001084228	S405, S408	ISAsQAsVPFDHLGK
CDK7	AAI30135	S170, T176	SFGsPNRIYtHQVVTR
NDRG2	Q7ZY73	S324, T326, S328	LSRsRtAsLSSEGNR
NDRG3	Q6GQL1	T329, S331	SRtHsASSSGSMEIPR
PDCD4	AAH56125	S42, S43	RTssRDSAR
RSK2	AAH80017	S366, S372	DsPGIPAsANAHQLFR
4E-BP2	AAH68624	T33, T37	TIPISDDDQLPHDYCTtPGGtLFS
4E-BP2	AAH68624	S61, T66, S79	RTsPLAQtPPRRLPDIPGVTsPNTVVE
4E-BP2	AAH68624	T66, T78, S79	RTSPLAQtPPRRLPDIPGVtsPNTVVE
Translation elongation factor 2 (eef2)	AAH60025	T57, T59	AGETRFtDtRKDEQER

Some of the earliest mechanisms that regulate embryonic patterning (Weaver and Kimelman, 2004; Thisse and Thisse, 2005) occur during the gastrula stage of Xenopus development. Our interest in the regulation of these events (Lane and Sheets, 2000; Mitchell and Sheets, 2001; Lane et al., 2004; Mitchell et al., 2007) motivated us to perform additional analysis of proteins from gastrula stage embryos manipulated to perturb embryonic patterning. In particular, we analyzed phosphoproteins present in organizer and non-organizer cells manually dissected from gastrula embryos (Table 2). Also, we analyzed the phosphoproteins from embryos treated with LiCl that enhances Wnt signaling and anterior development, and proteins from embryos injected with a β-catenin morpholino that disrupts Wnt signaling (Table 2). From these analyses we identified an additional 541 Xenopus phophoproteins (Table 2, SI Table 1).

Table 2. Identification of Xenopus phosphoproteins from oocytes, eggs and embryos.

Summary of data from the phosphoproteomic analysis of gastrula stage embryo fragments (organizer and non-organizer) and manipulated embryos (LiCl treated and β-catenin morpholino injected). All the results for individual proteins are found in SI Table 1.

	Manipulated gastrula stage embryos
	Organizer	Non-organizer	LiCl treatment	β-catenin morpholino	totals
STY	147	361	884	609	2001
S	126	281	646	470	1523
T	18	75	186	121	400
Y	3	5	52	18	78
Unique peptides	99	333	588	497	1517
Unique proteins	85	272	439	357	1153

Notably, many of the Xenopus phosphorylation events we identified corresponded to those observed in protein orthologs from mice and humans (Fig. 2, Tables 4 and 5). For many of these proteins multiple lines of analysis, including cell-labeling experiments, the use of phospho-specific antibodies and mutation of the specific amino acids were used to demonstrate that these specific residues were phosphorylated primarily in the context of cultured cells (see Table 4, column 4). For example, serine 369 (S-369) of the Xenopus RSK-2 kinase was identified as a site of phosphorylation in our analysis (Fig. 2). The same residue in the mouse RSK-2 protein is phosphorylated and this modification triggers RSK-2 activation in response to specific signaling events (Hauge et al., 2006; Hauge et al., 2007).

Figure 2. Evolutionary conservation of Xenopus phosphorylation sites.

A subset of the Xenopus phosphoproteins identified in this study (see Tables 4 and 5) were compared to their mouse and human orthologs. Here, six modified peptides corresponding to the designated full-length proteins are shown. Identified sites of phosphorylation are lowercase and bold. The phosphorylation sites in the Xenopus proteins identified in this study were also phosphorylation sites in the mouse and human proteins (Rsk-2 (Han et al., 2006), Fox-K2 (Dephoure et al., 2008), Casein Kinase II (β) (Lowery et al., 2007), HnRNP-K (Villen et al., 2007), Pumilio (Villen et al., 2007), Mek2 (Beausoleil et al., 2006).

Table 4.

Evolutionary conserved Xenopus phosphorylation sites and events^*

Xenopus Protein	Conserved phosphorylation site(s)*	Accession
RPS6	S-235, S-236, S-240, T-241, S-242, S244	P39017
Api5	S-465	AAH77529
MCM4	S-31	P30664
c-RAF	S-621	AAH72748
MCM2	S-139	P55861
DHA1	S-298, S-300, S-305	AAI06671
stmn1	S-37	AAH54159
CaRHSP1	S-36 S-38	AAH44047
aldolase	S-35	AAB31152
RSK-2	S-369	AAF15553
CIRP2	Y-160	AAH54250
casein kinase II (β)	S-209	NP_001084126
hnRNP-K	S-257	AAH44711.1
FOX-K2	S-359	Q7ZX03
pumilio	S-706	BAC57980
MEK2	S-388	NP_001080299.1

^*A partial list of phosphorylation sites conserved in mouse and human. Phosphorylation site data was obtained from PhosphoSitePlus (www.phosphosite.org).

Table 5.

Validation of Xenopus phosphorylation sites from the analysis of human and mouse proteins

Xenopus protein	Accession	Phosphosites:	Analogous human phosphosites		Analogous mouse phosphosites
			site	methods	site	methods
A-Raf	BAD04840	S573	S582	ms, pa, mt	S580	ms
B-Raf (MGC86346)	AAZ06667	S355	S363	ms	S346
B-Raf (MGC86346)	AAZ06667	S357	S365	ms, pa, mt	S348
C-Raf	AAH72748	S612	S621	ms, pa, cl, mt	S621	ms, pa, cl, mt
CAF-1	ABB04523	S417	S410	ms	S417
CAM kinase delta	NP_001083858	T286	T287	ms, pa	T287	ms
Casein kinase epsilon	NP_001084228	S405	S405	ms, pa, cl, mt	S405
Casein kinase epsilon	NP_001084228	S408	S408	ms, pa, cl, mt	S408
b-catenin	AAI08765	T551	T551	ms	T551	ms
CDC21/MCM4	AAH72870	S3	S3	ms, pa	S3
CDC21/MCM4	AAH72870	S83	S88	Ms, pa	S87	pa
CDC21/MCM4	AAH72870	T102	T102	ms	T101
CDC21/MCM4	AAH72870	T110	T110	ms, pa, mt	T109	pa
CDC21/MCM4	AAH72870	S120	S120	ms	T119	pa
CDK7	AAI30135	S170	S164	ms	S164	ms, mt
CDK7	AAI30135	T176	T170	ms	T170	ms, mt
CNOT2	AAH73075	S166	S165	ms	S165
DDI2 (LOC379186)	NP_001079499	S188	S194	ms
FoxK2	Q7ZX03	S359	S398	ms	S389	ms
Giant lethal larvae 1	NP_001084898	S656	S664		S662	ms, cl, mt
HnRNP-K	AAH44711	S257	S284	ms, pa, cl, mt	S284	ms
HP1 gamma/CBX3	AAO39117	S101	S93	ms, pa, mt	S93
Interleukin binding factor 3	NP_001083930	S385	S384	ms	S384
NDRG2	Q7ZY73	S324	S328	ms	S328	ms
NDRG2	Q7ZY73	T326	T330	ms	T330	ms
NDRG2	Q7ZY73	S328	S332	ms	S332	ms, pa, cl, mt
NDRG3	Q6GQL1	T329	T329	ms	T329	ms
NDRG3	Q6GQL1	S331	S331	ms	S331	ms
NDRG3	Q6GQL1	S334	S334	ms	S334	ms
Nsfli	AAH41297	S140	S140	ms, pa, cl, mt	S140	ms
Nsfli	AAH41297	S176	S176	ms	S176	ms
Programmed cell death 4	AAH56125	S42	S67	ms, pa, mt	S67	ms, pa, mt
Programmed cell death 4	AAH56125	S457	S457	ms, pa, mt	S457	ms, pa, mt
Pumilio	BAC57980	S704	S709	ms	S710	ms
RSK2	AAH80017	S369	S369	ms, pa, mt	S369	ms, pa, mt
RSK2	AAH80017	S375	S375	ms	S375
4E-BP2	AAH68624	T33	T36	ms, pa, cl, mt	T35	ms, pa
4E-BP2	AAH68624	T37	T40	ms
4E-BP2	AAH68624	T42	T45	ms, pa, cl, mt	T44	ms, pa
4E-BP2	AAH68624	S61	S64	ms, pa, cl, mt	S63	ms, pa
4E-BP2	AAH68624	T66	T69	ms, pa, cl, mt	T68	ms, pa
4E-BP2	AAH68624	T78	T81	ms
4E-BP2	AAH68624	S79	S82	ms
4E nuclear import protein (4E-T)	AAH77338	S372	S564	ms	S563	ms
4E nuclear import protein (4E-T)	AAH77338	S395			S586	ms
4E nuclear import protein (4E-T)	AAH77338	T575	T769	ms
Translation elongation factor 2 (eef2)	AAH60025	T57	T56	ms, pa, cl	T56	ms, pa
Translation elongation factor 2 (eef2)	AAH60025	T59	T58	ms	T58
TRA-2	AAH44990	T197	T204	ms	T202

Abbreviations: MS mass spectrometry; PA phosphospecific antibody; CL cell labeling; MT mutagenesis and functional analysis. Phosphorylation site data was obtained from PhosphoSitePlus (www.phosphosite.org).

In addition several residues of the Xenopus 4E-BP2 protein were identified as sites of phosphorylation in our analysis (Tables 3, 4 and SI Table 1). 4E-BPs inhibit translation by binding eIF-4E translational initiation factors (Gingras et al., 1999). Studies in cultured cells indicate that the binding of 4E-BPs to initiation factors is controlled by phosphorylation. Our analysis demonstrated that the Xenopus 4E-BP2 was phosphorylated at several different sites, analogous to known phosphorylation sites in human and mouse 4E-BP2 proteins (Table 4). As an independent method to detect 4E-BP2 phosphorylation events, we analyzed 4E-BP2 from oocytes, eggs, stage 7 and stage 10.5 embryos by protein immunoblotting. Analysis with an antibody that recognizes 4E-BP2 when it is phosphorylated at either Thr33 or Thr42 indicated that 4E-BP2 is phosphorylated to varying extents at all stages examined (Fig. 3, lanes 1-4). Analyzing the same filter with an antibody that detects 4E-BP2 regardless of its modification state revealed the presence of non-phosphorylated protein (Fig. 3, lanes 5-8). These results together with the data summarized in Table 4 increase confidence in the data obtained with mass spectroscopy.

Figure 3. Phosphorylation of Xenopus 4E-BP2.

Proteins from equivalent numbers of Xenopus oocytes eggs, stage 7 embryos and stage 10+ embryos were separated by SDS-PAGE and transferred to PVDF membrane. The membrane was probed with an anti-4E-BP2 antibody that detects 4E-BP2 when it is phosphorylated at Thr-33 or Thr-42. Antibodies were stripped from the filter and the same filter was probed with a second anti-4E-BP2 antibody that detects total 4E-BP2 regardless of its phosphorylation state.

Identified phosphoproteins represented a wide range of expression levels

A major challenge of proteomic analysis using mass spectrometry arises because of the extremely large dynamic range of protein levels expressed in specific cells. Many regulatory proteins such as signaling proteins and transcription factors are expressed at low levels. The detection of such important regulatory molecules can be occluded by the presence of much more abundant proteins in a sample. To determine whether our results were heavily biased towards the detection of abundant proteins we used the Xenopus tropicalis EST (expressed sequence tags) resources (http://www.sanger.ac.uk/Projects/X_tropicalis/) (Gilchrist et al., 2004). Xenopus tropicalis ESTs from different stages of development have been organized into clusters in which each cluster represents the ESTs from a single mRNA. The number of ESTs in a cluster provides an estimate of relative mRNA abundance and therefore, at present and by inference, the best approximate estimate of protein levels. For example, the CIRP2 RNA binding protein is encoded by an abundant mRNA that is highly expressed and the CIRP2 EST cluster contains hundreds of ESTs. In contrast, the EST clusters representing mRNAs expressed at low levels, such as the denticleless mRNA, contain fewer than ten ESTs. Xenopus tropicalis mRNAs and EST clusters were identified for 73 of the phosphoproteins we identified from gastrula embryos (SI Tables 2 and 3.). Notably, the mRNAs encoding Xenopus phosphoproteins were predominantly from EST clusters representing the non-abundant and moderately abundant mRNAs (Fig. 4). These data provided evidence that the Xenopus phosphoproteins we identified were encoded by mRNAs with varying levels of expression and were not enriched for the most abundant mRNA class.

Figure 4. Identified phosphoproteins represented a wide range of expression levels.

For each of seventy Xenopus laevis gastrula phosphoproteins (SI Tables 1, 2 and 3) the corresponding Xenopus tropicalis mRNA was identified along with the associated ESTs. Xenopus tropicalis ESTs for specific mRNAs have been sorted into clusters where the number of ESTs per cluster provides an estimate of mRNA expression levels (http://www.sanger.ac.uk/Projects/X_tropicalis/) (Gilchrist et al., 2004). EST clusters were identified for each of the mRNAs encoding a Xenopus laevis phosphoprotein. The number of ESTs per cluster was compared to the number of phosphoprotein encoded mRNAs that map to each cluster and plotted. The Xenopus phosphoproteins identified in this study corresponded to both the abundantly expressed mRNAs (clusters with large numbers of ESTs) and minimally expressed mRNAs (clusters with few ESTs).

The Xenopus phosphoproteins represented a diverse array of functional categories

Functional annotation revealed that the identified Xenopus phosphoproteins were involved in a diverse array of cellular processes (Fig. 5). In addition, many of the identified phosphoproteins encoded proteins of unknown function (42%). Of note, was the substantial number of proteins that were categorized as “hypothetical.” These data make it clear that many of these proteins are indeed bona fide proteins expressed in living embryos, underscoring a value in this analysis beyond presenting the functionally unbiased information about the Xenopus phosphoproteome.

Figure 5. Functional annotation of Xenopus phosphoproteins.

The function of each phosphoprotein was assigned and grouped using accession numbers. Each fragment on the pie chart represents a functional category. The percentage of phosphoproteins identified in this study that fall into each functional category is indicated. “Unknown proteins” have no assigned function, whereas “Other” proteins have an assigned function that lies outside of the functional categories shown.

Many of the Xenopus phosphoproteins function to regulate processes that play a central role in vertebrate embryogenesis (Table 6, SI Table 3). For example, in oocytes, eggs and early stage embryos normal development depends heavily on post-transcriptional processes such as mRNA translation and mRNA localization. Consistent with the notion that such processes must be robustly regulated in developing embryos, phosphorylation sites were identified on several different Xenopus proteins involved in translation and RNA metabolism (Table 6, SI Table 3). For example, the translational regulatory protein embryonic poly (A) binding protein (ePABP) was phosphorylated at multiple sites (Table 6, SI Table 3). ePABP binds to mRNA poly (A) tails to stimulate translation and protect mRNAs from degradation (Wilkie et al., 2005). Interestingly, serine 255 of ePAB was phosphorylated in eggs; this serine residue resides within one of the ePABP RRM domains important for poly (A) binding, raising the possibility that phosphorylation may regulate ePABP's ability to bind poly (A). The identification of this and many other previously unexamined phosphorylation sites sets the stage for future functional studies in a wide variety of research areas relevant to vertebrate embryogenesis and cell function.

Table 6. Developmentally relevant Xenopus phosphoproteins Phosphorylation of developmentally relevant proteins^*.

This table shows a partial set of phosphoproteins that are known to function as developmental regulatory proteins in translation, RNA metabolism, signal transduction and transcription. The complete list of this type of proteins is found in SI. Table 3.

Protein	Accession	Sample	Phos.sites
Proteins involved in translation or RNA processing:
4E-BP2	AAH68624	oocyte	T:34 T38
4E-T initiation factor 4E import factor 1	AAH77338	egg	S:395
Cap methyltransferase (xCMT)	Q9I8S2	blastula	S:50
EiF4G1	AAH89197	blastula	S:45
Embryonic poly A binding protein (ePAB)	AAK29408	blastula	S:461
MASKIN	AAF19726	gastrula	S:638
ePABP2	AAO33927	gastrula	T:71
Pumilio	AAL14121	gastrula	S:213
Vg1 RNA binding protein	AAC18597	gastrula	S:192, S:193

Proteins involved in signaling or transcription:
Casein Kinase II (Beta)	P28021	egg	S:109
Erb-b2	NP_001089062	egg	Y:440
Eyes absent-3	NP_001089994	gastrula bcat	S:286
FOXK2	Q7ZX03	oocyte	S:359
ISWI (imitation switch)	AAG01537	blastula	S:1024
Mi2 deacetylase	AAD55392	egg	S:79
Sox3	AAH72222	gastrula nonorganizer	S:244
Suppressor of fused	AAH73701	gastrula	S:291
Treacle	AAW56574	gastrula	S:700
V-ets	AAI55947	egg	S:239
β-TRCP	AAA02810	egg	S:19

^*A partial list. For a full list see SI Table 3

Conclusion

The data presented here provides an information-rich foundation for investigating the role of specific phosphorylation events by exploiting the many established experimental advantages of Xenopus oocytes, eggs, embryos and in vitro extracts. Despite the large number of undocumented phosphorylation sites uncovered in this study, the data presented here remains some distance from saturation. Further developments in mass spectrometry technology that couple high mass measurement accuracies with ETD promise an even deeper exploration of the Xenopus phosphoproteome in the future (Hogan et al., 2005; McAlister et al., 2007; Hubler et al., 2008; McAlister et al., 2008). Our data clearly illustrate that the study of regulatory phosphorylation in controlling vertebrate development is rich with unexamined questions.

METHODS

Xenopus laevis eggs, oocytes, embryos and gastrula stage modified tissues

Xenopus laevis oocytes, eggs and embryos were generated using standard methods (Lane and Sheets 2000; Mitchell and Sheets 2001; Lane et al., 2004). Two hundred defolliculated oocytes (st. VI), unfertilized eggs, embryos (st. 7 and st. 10+) were flash frozen in liquid nitrogen.

In addition samples of gastrula stage embryos (st. 10+) subjected to several different manipulations were also collected for analysis. These included LiCl treatment to activate the β-catenin pathway and injection of a morpholino to β-catenin to block β-catenin expression and signaling. For LiCl treatment, 32-cell embryos were exposed to 0.3 M LiCl in 0.1x MMR for 12 minutes. After LiCl treatment, embryos were washed four times every 15 minutes with R/10 buffer at pH 7.4. Embryos were allowed to develop to stage 10+ and flash froze in liquid nitrogen. β-catenin injected embryos were prepared by injecting 4-8 cell embryos with 10 ng of β-catenin morpholino (Heasman et al., 2000). These embryos were then allowed to develop to stage 10+, where they were flash frozen in liquid nitrogen. Organizer and non-organizer tissues were prepared by selective tissue enrichment of stage 10+ embryos using a Gastromaster microsurgery instrument. An equivalent of two hundred embryos for each tissue sample were flash frozen for protein extraction.

Protein extraction

Tissue was homogenized in 15 mM Tris pH 8, 75 ug PMSF, 50 mM NaF, 1 mM NaOV, and 25 mM β-glycophosphate. An equal volume of 1,1,2-trichlorotrifluoroethane (Freon) was mixed with homogenate, centrifuged and the lower phase containing yolk proteins removed. The sample was mixed with Trizol (400 μl), incubated for 3 minutes at room temperature and separated by centrifugation. Isopropanol was added to the bottom phase and incubated for 10 minutes to precipitate the proteins, and centrifuged to pellet the proteins. The pellet was washed three times with 0.3 M guanidine-HCl (in 95% ethanol), washed once with 100% ethanol and dried to remove residual ethanol. The pellet was resuspended in 6 M guanidine-HCl, 25 mM NH₄HCO₃, 50 mM NaF, 1 mM NaOV, and 25 mM β-glycophosphate. A buffer exchange using 6 M urea followed by 25 mM ammonium bicarbonate (pH 8) was done to lower guanidine hydrochloride concentration to 0.1 M, and the urea concentration to 2 M.

Phosphopeptide enrichment

Extracted proteins were prepared for proteolysis by reducing with 500 uM DTT for one hour, 1.2 mM iodo-acetate for 45 minutes in the dark followed with 1.25 mM DTT. The proteins were then treated with sequencing grade trypsin for one hour at 37 °C and quenched with glacial acetic acid (pH 3). The sample was desalted using a C₁₈ column, and the peptide containing eluent brought to dryness using a vacuum centrifuge. Esterification of peptides was completed by adding 40μl of thionyl chloride to 1 ml of anhydrous methanol and adding this solution to the dried peptides for 1 hr with mixing at room temperature. The modified peptides were brought to dryness again, and resuspended in a reconstitution buffer of equal parts methanol, acetonitrile, and water, with a 0.01% acetic acid concentration. Modified peptides were loaded onto a 560 μm I.D. × 8 cm fused-silica column packed with Poros MC particles (Applied Biosystems) activated with 100 mM FeCl₃. Non-phosphorylated peptides were removed by rinsing the column with the reconstitution buffer. Captured phosphopeptides were then eluted onto a nHPLC pre-column (360 μm × 75 μm, packed with 10-30 μm C₈ particles) using 15 μL of 50 mM phosphate buffer, pH 8.

LC-MS/MS

The pre-column was connected to an nHPLC analytical column (360 μm o.d. × 50 μm i.d.) packed with 5 cm of 5μm C8 reversed-phase packing material and equipped with an integrated, electrospray emitter tip (Martin et al. 2000). Peptides were eluted into a Thermo LTQ mass spectrometer modified to perform electron transfer dissociation (ETD) (Coon et al., 2004; Syka et al., 2004). Peptides were eluted at a flow rate of 60 nL/min with the following gradient: 0-7% B in 5 min, 7% - 45% B in 70 min, 45-100% in 15 min, 100-0% B in 5 min, (A = 0.1 M aqueous acetic acid, B = 70% acetonitrile in 0.1 M aqueous acetic acid). The mass spectrometer was set to continuously cycle through the acquisition of a MS1 spectra from which the 10 most abundant precursors were automatically selected for tandem MS interrogation via ETD (Haas et al., 2006).

LiCl treated and β-catenin morpholino injected embryos were prepared and analyzed in a consistent manner with protocol above, with the exception that the LiCl treated samples were analyzed by ETD only, and the β-catenin morpholino injected samples were analyzed by ETD and collision activated dissociation (CAD) on a hybrid linear ion trap-orbitrap mass spectrometer (McAlister et al., 2007).

Data analysis

All tandem mass spectra were searched against the Xenopus laevis subsection of the NCBI database using the Open Mass Spectrometry Search Algorithm (OMSSA). Static modifications of 58.01 Da on cysteine (carbamidomethylation), and 14.02 Da on glutamic acid, aspartic acid, and the peptide c-terminus to account for the o-methyl ester modification. Dynamic modifications of +79.9663 Da at serine, threonine, and tyrosine residues for phosphorylation, and +16 Da at methionine for oxidation were also employed. Confidence values were assessed by a target-decoy search strategy (Käll et al., 2008) to identify phosphopeptides in two groups (<0.5% false positive rate (FPR), and <5% FPR). Identified peptides were grouped into 12 functional categories according to their predicted parent protein according to accession number using a functional annotation database clustering program called Database for Annotation Visualization and Integrated Discovery (DAVID) available at (http://david.abcc.ncifcrf.gov/home.jsp).

Immunoblotting

Total protein from equivalent numbers of Xenopus oocytes eggs, stage 7 embryos and stage 10+ embryos were loaded onto each lane and electrophoresed on 12.5% SDS-polyacrylamide gels. Proteins were transferred to PVDF membrane, the membrane was blocked and probed with an anti-4E-BP2 antibody that detects 4E-BP2 when it is phosphorylated at Thr-33 or Thr-44 (Cell Signaling, phospho-4E-BP1 (Thr37/46) #2855). The membrane was incubated with goat anti-rabbit antibodies coupled to horseradish peroxidase. Antibody binding was detected using a chemiluminescent substrate and exposure to x-ray film. Antibodies were stripped from the filter by incubation in a solution of 0.2 M Glycine-HCl pH2.5, 0.05% Tween-20, 100 mM 2-mercaptoethanol for 60 min. The same filter was probed with a second anti-4E-BP2 antibody that detects total 4E-BP2 regardless of its phosphorylation state (Santa Cruz). Antibody binding was detected as described above.

EST cluster comparison

Seventy Xenopus laevis gastrula phosphoproteins were randomly chosen for analysis from SI. Tables 2, 3 and 4; APC-binding protein (EB2) (NP_001083449), Ataxin2 (AAH97692), β-catenin (AAI08765), Brg1 (AAX08100), BRM (AAQ02780), Interleukin Enhancer Binding Factor 3 (NP_001083930), Casein Kinase II (Beta) (P28021), Cnot2 (AAH73075), Denticleless (AAH73015), ELAV (NP_001081035), enhancer of zeste-2 (EZH2) (AAK30208), Eyes absent-3 (NP_001089994), FOXD3a (Q9DEN4), FoxN3a (NP_001090178), Fzd7 (AAH42228), ID2 (BAA76634), ISWI (imatation switch) (AAG01537), Jumonji (AAH88951), Lethal giant larvae 2 (AAH72924), Lin-28 (AAH42225), Lunitic fringe (AAI29635), MASKIN (AAF19726.1), Mastermind1 (NP_001080927), Mi2 deacetylase (AAD55392.1), mRNA-4 p56 (FRGY2) (P21574), Poly (A) binding protein-1 (AAO33927), Programmed cell death 4 (AAH56125), Pumilio (AAL14121), RAP55 (AAH42251), Ski2 (NP_001084112), Sox3 (AAH72222), Tra-2 (AAH44990), β-TRCP (AAA02810), Treacle (AAW56574), V-ets (AAI55947), Vg1 RNA binding protein (AAC18597), Xeek1 (AAC59904), zinc finger protein 36 (AAH84221), zinc finger transcription factor SALL4 (AAH95923), Api5 (AAH77529), death associated protein (AAI33257), deoxyuridine triphosphatase 3 (NP_0001091360), echinoderm microtubule associated protein 4 (AAI10955), eukaryotic translation elongation factor 2 (EF-2) (AAH44327), eukaryotic translation initiation factor 4 gamma (EIF4G) (AAH89197), eukaryotic initiation factor 4E binding protein 2 (4E-BP2) AAI33809, heat shock protein (90kDa) NP_001086624, high mobility group AT-hook 2 (AAI24963), histone binding protein N1/N2 (CAA28419), insulin-like growth factor 2 mRNA binding protein 1 (AAH57700), NDRG (NP_001080389), NOL1/NOP2/Sun domain 2 (Nsun2) (AAH68818), Nsfl1 (p47) (AAH41297), Nucleoplasmin (CAA28460), PAICS (AAH41276), pyruvate dehydrogenase alpha (PDHA) (AAI06671), programmed cell death 4 (AAH56125), RAD23 (AAH44115), Ribosomal protein L15 (AAH46569), Ribosomal protein L34 (AAH99259), Ribosomal protein S6 (AAH54151), serpin (AAA49703), Ribosomal protein L28 (AAH53798), stathmin 1/oncoprotein 18 (NP_001080672), eukaryotic translation initiation factor 4H (EiF4H) (NP_001083502), MCM4 (AAH83031), nuclear autoantigenic sperm protein (NASP1) (NP_001089907), nucleolin (AAI57416), nucleoporin (CAB53357), Ribosomal protein L24 (AAH78474), Ribosomal protein S1a (CAA40592), Ribosomal protein L13 (AAH75140), Upstream of NRAS (NP_001090026), CIRP2 (AAH54250), eukaryotic translation initiation factor 4E nuclear import factor 1 (4E-T) (NP_001086710). For each Xenopus laevis phosphoprotein, the analogous Xenopus tropicalis mRNA was identified along with the associated ESTs. The Xenopus tropicalis ESTs for specific mRNAs have been sorted into clusters where the number of ESTs per cluster provides an estimate of mRNA expression levels (http://www.sanger.ac.uk/Projects/X_tropicalis/) (Gilchrist et al., 2004) EST clusters were identified for each mRNA corresponding to a Xenopus laevis phosphoprotein. The number of ESTs per cluster was compared to the number of phosphoprotein encoded mRNAs that map to each cluster and plotted.