Subcellular Metabolite and Lipid Analysis of Xenopus laevis Eggs by LAESI Mass Spectrometry

Xenopus laevis eggs are used as a biological model system for studying fertilization and early embryonic development in vertebrates. Most methods used for their molecular analysis require elaborate sample preparation including separate protocols for the water soluble and lipid components. In this study, laser ablation electrospray ionization (LAESI), an ambient ionization technique, was used for direct mass spectrometric analysis of X. laevis eggs and early stage embryos up to five cleavage cycles. Single unfertilized and fertilized eggs, their animal and vegetal poles, and embryos through the 32-cell stage were analyzed. Fifty two small metabolite ions, including glutathione, GABA and amino acids, as well as numerous lipids including 14 fatty acids, 13 lysophosphatidylcholines, 36 phosphatidylcholines and 29 triacylglycerols were putatively identified. Additionally, some proteins, for example thymosin β4 (Xen), were also detected. On the subcellular level, the lipid profiles were found to differ between the animal and vegetal poles of the eggs. Radial profiling revealed profound compositional differences between the jelly coat vitelline/plasma membrane and egg cytoplasm. Changes in the metabolic profile of the egg following fertilization, e.g., the decline of polyamine content with the development of the embryo were observed using LAESI-MS. This approach enables the exploration of metabolic and lipid changes during the early stages of embryogenesis.


Introduction
The early stages of embryonic development following fertilization in animals are characterized by synchronous cell divisions, the onset of transcription of genes that will pattern the embryo, and local signaling events that transition a ball of equipotent cells into regions that express different tissue fates [1,2]. These events are common across different species of animals, but can occur for different lengths of times and at different stages of morphogenesis.
The African clawed frog, Xenopus laevis, has been used extensively to study the cellular and molecular events of early embryogenesis. This species is used because they are easily bred in the laboratory, are disease resistant, have a reproductive response that is independent of season, and produce a large number of eggs per spawning. In particular, the large size of the oocytes, ova (unfertilized eggs) and embryos make them easy to manipulate and analyze by standard biochemical and molecular biology approaches [3,4,5,6,7].
X. laevis oocytes and eggs have been an important source of material for elucidating the molecular regulation of the cell cycle. For example, traditional protein chemistry approaches led to the identification of maturation promoting factor and the cyclin proteins [8,9]. More recently, the oocyte has been used to identify the proteins involved in nuclear transport and DNA replication [6]. To reveal the biochemical makeup, elucidate low abundance proteins, and study the kinetics of developmental and cellular events, X. laevis oocytes have been analyzed by high performance liquid chromatography (HPLC) [10,11] [12], time-of-flight (TOF) secondary ion mass spectrometry (SIMS) [13], gas chromatography (GC) mass spectrometry (MS) [14], 2D gel electrophoresis followed by electrospray ionization (ESI) MS [15], and liquid chromatography-tandem MS (LC-MS/MS) [2]. To study the types and amounts of lipids extracted from Xenopus eggs, HPLC methods were developed for the detection of multiple species [10,12]. Similarly, with an HPLC method for characterizing the carbohydrates in the jelly coat of Xenopus eggs, specific oligosaccharides were identified and assigned to the three specific layers of the coat [11,16]. In another study, LC-MS/MS was used to monitor the amounts of alanine in the early embryo, resulting in a new hypothesis for the increase in cell cycle time at the mid-blastula transition [2]. Using a GC-MS method for analyzing metabolites from volume-limited samples, e.g., X. laevis eggs, good reproducibility was shown for volumes as small as 10 nL. However, the method required silylation of some metabolites to increase their volatility for analysis using GC [14]. On a subcellular scale, 3D molecular images of freeze dried Xenopus oocytes were obtained under optimized sample preparation conditions using TOF-SIMS [13]. While all of these methods for X. laevis oocytes and embryo analysis have been shown to capture important aspects of its biochemistry, most of them require elaborate sample preparation prior to analysis that might alter subcellular localization of important developmental molecules.
An advantage of direct analysis methods is that they can report on the unperturbed composition of biological specimens, ideally in their living, natural state. Recently, the analysis of single oocytes and embryos of mice and bovine species by desorption electrospray ionization (DESI) mass spectrometry provided new insight into lipid metabolism changes during early embryonic development [17,18].
In another direct analysis technique, laser ablation electrospray ionization (LAESI) MS, the energy from a focused mid-infrared laser pulse is absorbed by the water molecules in a cell or tissue leading to the ablation sampling of adjacent cellular content [19]. Each consecutive laser pulse ablates a successive layer of the sample enabling analysis at a defined subcellular depth or of an entire cell [20,21]. Combination of LAESI with microdissection of plant cells reveals subcellular metabolite gradients between the cytoplasm and the nucleus [22]. Diverse applications of LAESI-MS include the analysis of virally infected immune cells as well as comparative analysis of plant cell phenotypes [23,24]. Because the developing egg is a spatially and temporally dynamic system that is sensitive to environmental and thus preparatory changes, we used the direct LAESI-MS analysis of individual X. laevis eggs and embryos to explore subcellular asymmetries and potential changes during significant developmental transitions. Using LAESI-MS to probe these specimens required minimal sample preparation, was performed at atmospheric pressure, and each ablation event occurred instantaneously (in less than a second) on the developmental time scale. The analyses were performed on living eggs and embryos that were minimally perturbed from their natural state.
Some of the metabolic profiles of the unfertilized egg obtained from the LAESI-MS analyses were compared to results reported previously in the literature and showed good agreement, thus validating the approach. In addition we identified subtle differences between the unfertilized egg, in which the first meiotic division is completed, and the fertilized egg and early cleavage stage embryos. Importantly, we found differences between the animal and vegetal poles of the egg, which are predisposed to different developmental fates. Since these early developmental events are shared across animals, a direct analysis of the cellular constituents should provide fundamental information about the common biochemical processes that regulate them. This simple approach using a single, live embryo in its natural state should be ideal for analyses of other embryos that are difficult to collect and are small in size.

Materials and Methods
Ova and Embryos X. laevis eggs and sperm were obtained from mature animals kept in 30 gallon aquaria at 21˚C in natural daylight and fed frog brittle 2 times per week. Oocyte maturation was induced by injecting frogs with 1000 U human chorionic gonadotropin (Sigma) 12 hours prior to removal of eggs as described elsewhere [25]. Mature eggs were gently squeezed from female frogs into Petri dishes; some were collected prior to fertilization and others were fertilized in vitro and collected at various stages after sperm entry. Sperm were obtained from minced testes of sacrificed males that were anesthetized by submersion in an ice bath with tricaine methanesulfonate. Eggs were fertilized in vitro by adding the released sperm to the eggs and adding 0.16 MBS (16588 mM NaCl, 1 mM KCl, 1 mM MgSO 4 , 0.7 mM CaCl 2 , 5 mM HEPES, pH 7.8 and 2.5 mM NaHCO 3 ) [26]. The jelly coat was removed from some samples by washing the eggs with 2% cysteine solution (pH 8.0, adjusted by 10 M NaOH) for 4 minutes followed by washing in 0.

Chemicals
All chemicals for egg preparation were reagent grade and obtained from Sigma. For the electrospray solution, highly purified spectral-grade methanol and water solvents were obtained from Acros Organics (Geel, Belgium), and glacial acetic acid was purchased from Fluka (Munich, Germany). All of the chemicals were used without further purification.

LAESI-MS
The LAESI ion source used was similar to those described earlier [19]. Briefly, mid-infrared laser pulses at 2940 nm wavelength and 20 Hz repetition rate were produced by a Nd:YAG laser driven optical parametric oscillator system (Opolette 100, Opotek, Carlsbad, CA). Laser pulses with an average 320 mJ/pulse energy were focused to about 200-300 mm diameter spot with either a 150 mm focal length CaF 2 lens or a 75 mm ZnSe lens. The eggs/embryos were placed on a precleaned microscope glass slide held 15 mm below the electrospray axis that was in line with the mass spectrometer orifice. In the home-built electrospray system, a syringe pump supplied 50% methanol solution containing 0.1% acetic acid through a tapered tip metal emitter (length 5 cm, tube OD 320 mm and tip ID 50 mm) at a flow rate of 300-400 nL/min. The electrospray emitter was held at 2800-3300 V. Positive ions were collected by either an orthogonal acceleration TOF mass spectrometer (Q-TOF Premier, Waters Co., Milford, MA) with a mass range of m/z 50-2000 and a typical resolution of 8,000 (FWHM) or with a high performance TOF mass spectrometer (Synapt G2S, Waters Co., Milford, MA) with a mass range of m/z 50-2000 and a typical resolution of 30,000 (FWHM) [27].

Data Analysis
The electrospray background ion signal was subtracted from the collected LAESI mass spectra in the MassLynx 4.1 software (Waters Co., Milford, MA). Putative peak assignments for metabolites and lipids were based on accurate mass measurements, isotope distribution patterns, database searches, data mining of the related literature, and in some cases tandem MS analysis. Multiple databases were used for the metabolite and lipid searches, including METLIN [28], MetaCyc [29], Lipid Maps [30], HMDB [31] and KEGG [32,33]. They were last accessed on August 12, 2014. The NIST Isotope Calculator program (ISOFORM, Version 1.02) was used to calculate monoisotopic masses.

Ovum Analysis
A single X. laevis egg is ,1.4 mm in diameter and has an approximate volume of 400 nL. A typical LAESI mass spectrum from a single unfertilized X. laevis egg without the removal of its jelly coat is shown in Fig. 1. Most of the ions below m/z 450 correspond to small metabolites, whereas many ions between m/z 450 and 900 are assigned to lipids. For a typical unfertilized X. laevis ovum, putative assignments of the ions generally found in the m/z ,550 region reveals 52 small metabolites (see Table 1). Many of these metabolites, e.g., amino acids, organic acids, and redox buffering agents, fulfill an essential role in cell development and serve as building blocks for cellular biosynthesis. Using liquid extraction and LC-MS/MS, a previous study followed 48 metabolites in X. laevis embryos [2]. In Table 1, sixteen metabolites found by both the LC-MS/MS and LAESI-MS methods are marked by asterisks. The advantage of LAESI-MS over conventional techniques is that it rapidly (within seconds) identifies metabolites in a single egg with minimum perturbation before analysis. It also provides complementary data for improved coverage of the small metabolites.
In the mass spectra we noticed the presence of multiply charged peaks at m/z 835, 1,002, and 1,251 with charge states of 6+, 5+ (see Fig. 1) and 4+, respectively. Deconvolution of these peaks revealed that they corresponded to thymosin b4 (Xen) (see the inset in Fig. 1). This 44-residue peptide in Xenopus has been highly conserved throughout evolution as it only differs from the human thymosin b4 in three amino acid residues [34]. A typical defolliculated oocyte contains 0.5 and 10 picomoles of thymosin b4 (Xen) in stage I and stage VI of the oogenesis, respectively [34]. Our results show that unfertilized eggs, the next step after stage VI oocytes, continue to contain thymosin b4 (Xen) in significant quantities.

Radial Profiling of Ovum
The X. laevis egg is surrounded by a jelly coat that must be penetrated by the sperm during fertilization. The LAESI mass spectra of the jelly coat were dominated by sodiated disaccharide, trisaccharide, and tetrasaccharide molecules at m/z 365.108, 527.192 and 689.286, respectively. The Xenopus egg jelly is known to contain a 21,073 Da protein, allurin, that serves as a chemoattractant for the sperm [35]. The LAESI mass spectra of the jelly coat did not show the ions for allurin, but exhibited strong signal for multiply charged ions corresponding to a protein with a molecular weight of 11,728¡1 Da (see Fig. 2) [36].
Radial profiling of an unfertilized X. laevis egg with a jelly coat was conducted by ablating the same area with multiple laser pulses, producing a mass spectrum for each shot (see Fig. 3). The first approximately six laser shots produced mass spectra similar to that of the jelly coat (see the top panel in Fig. 3). This observation was consistent with the ,30 mm ablation depth per laser shot and the ,200 mm thickness of typical jelly coats. Once the jelly coat-related signal diminished, multiply charged thymosin b4 (Xen) peptide ions were detected (see the middle panel in Fig. 3). From its radial localization, it seems that this peptide is associated with the vitelline and/or plasma membranes. After approximately two additional laser shots, metabolites and lipid species characteristic to the egg cytoplasm composition were detected in the mass spectra (see bottom panel in Fig. 3).

Lipid Composition of Unfertilized Ovum
The cytoplasm of the unfertilized egg contains a high percentage of yolk platelets composed of proteins and lipids for nourishment of the embryonic cells [37]. The major peaks observed between m/z 400 and 900 in the LAESI mass spectrum are attributed to lipids. Earlier studies on phospholipid extracts from X. laevis eggs showed that the major lipid components were phosphatidylcholines (PC) at ,20 nmol/egg, followed by phosphatidylethanolamines (PE) at ,10 nmol/egg, phosphatidylinositols (PI) at ,4 nmol/egg, lysophosphatidylcholines (LPC) at  ,2 nmol/egg, and phosphatidylserines (PS) at ,1 nmol/egg [12]. In addition to phospholipids, triacylglycerols (TAG) were the major neutral lipid components found in the Xenopus eggs along with fatty acids (FA), cholesterol and diacylglycerols (DAG) [38]. The LAESI mass spectra gave additional insight into the nature of the lipids present in the ovum. A wide variety of lipid classes were identified, including 14 FAs, 13 LPCs, 36 PCs and 29 TAGs. Most of the abundant lipid peaks in these spectra corresponded to PCs, where one of the acyl chains was palmitic acid (16:0), whereas the other acyl chain length and saturation varied between (14:0) and (22:6) (see Table 2). The acyl chain length assignments were based on the tandem mass spectra of the lithiated lipids, [M+Li] + , generated by reactive LAESI. In this technique Li + ions from the electrospray solution react with the lipid molecules from the ablated sample [39]. For example, tandem MS of a protonated lipid ion at m/z 760.582 produced a single fragment ion corresponding to its phosphocholine head group at m/z 184.073 resulting in the overall assignment of PC(34:1) showing only the combined length and number of double bonds for the acyl chains (see Fig. 4a). Further information on the individual acyl chains was obtained by producing and fragmenting the corresponding lithiated species, Fig. 3. Radial profiling of jellied unfertilized X. laevis egg by consecutive laser shots generated mass spectra at each shot corresponding to increasing depths. The first ,6 laser shots produced mass spectra similar to that of the jelly coat (top spectrum). This was followed by a mass spectrum that showed presence of multiply charged thymosin b4 (Xen) peptide (middle spectrum). After an additional ,2 laser pulses, other metabolites and lipids characteristic of the cytoplasm were detected (bottom spectrum).
doi:10.1371/journal.pone.0115173.g003 which showed fragments for the loss of oleic acid (18:1) and palmitic acid (16:0) at m/z 510.3631 and 484.3452, respectively (see Fig. 4b). Another major class of lipids in the spectra is LPC that is structurally similar to PC without the acyl group at sn-2. Phospholipase A2 is known to degrade PC into LPC and facilitate fertilization. Similar to PC, the major LPC is 16:0, and others include acyl chains between 14:0 and 22:6 (see Table 2). Because LPCs contain a free -OH group on the glycerol moiety, they undergo a loss of H 2 O during protonation and most ions are detected as [MH-H 2 O] + . These fragments are often more abundant than the protonated peaks. A few weak peaks corresponding to PE and PS can also be detected in the spectra. The major lipids found in the LAESI experiments and their fatty acid compositions are in good agreement with the literature reports based on extraction, HPLC separation and GC-based fatty acid analysis [12]. A putative list of the assigned lipids observed in the LAESI mass spectra of de-jellied unfertilized X. laevis egg is shown in Table 2. LAESI mass spectra with annotations for selected FA, LPC and PC species are shown in Fig. 5.

Animal Pole vs. Vegetal Pole
The unfertilized X. laevis egg is polarized, with a highly pigmented animal pole and a weakly pigmented vegetal pole separated by an unpigmented equatorial zone. During oogenesis, mRNAs and proteins are synthesized and differentially stored in the cytoplasm of these regions for use by the embryo after fertilization.
In particular, mRNAs that direct ectodermal development are enriched in the animal pole whereas mRNAs that direct endoderm and gamete development are enriched in the vegetal pole [40,41]. This molecular asymmetry is altered after fertilization by a cytoplasmic reorganization that occurs in response to sperm entry. This reorganization is critical for establishing the dorsal axis of the embryo. Because the mRNAs in the animal versus vegetal regions of the egg have very different developmental functions that are essential to establishing the vertebrate body plan, the corresponding metabolic and lipid profiles were investigated by LAESI-MS. As seen in Fig. 6, numerous lipids were found in higher abundance at the vegetal pole compared to the animal pole. This finding was consistent with magnetic resonance imaging studies of the animal and vegetal poles in X. laevis oocytes that had shown two to three times higher abundance for TAGs in the vegetal region [42]. Among the many lipids detected at the vegetal pole, we found an LPC at m/z 478, which was consistent with the presence of yolk platelets there. A previous report suggested that LPC lipids played a role in assisting fertilization [12]. In addition, there were many small metabolites found at both poles and their putative identifications can be found in Table 1. Although differences in the small metabolite abundances at the two poles were clearly discernable, they exhibited significant variance between eggs. As an example, the ion abundance ratios between vegetal and animal poles are shown in Table 3 for certain metabolites and lipids. To establish reliable abundance ratios, further analysis is needed.

Fertilized Ovum and Early Embryos
A number of critical developmental events occur after the X. laevis egg is fertilized, including: sperm nucleus entry into the cytoplasm and an influx of calcium, the completion of meiosis II, the initiation of synchronous cell cycles, and the initiation of low levels of gene transcription at the 32-cell stage [37]. In the unfertilized egg, the cell cycle is arrested in metaphase of the second meiotic division, but all of the components for cell division, cytokinesis, gene transcription and protein synthesis have previously been produced in the oocyte.
We directly analyzed unperturbed eggs by LAESI-MS at 30 minute intervals after fertilization, and cleavage stage embryos at 8-cells and 32-cells. Most of the major metabolites and lipids found during these early stages after fertilization were similar to those in unfertilized Xenopus eggs (see Fig. 7). These results are consistent with the prevailing concept that the synthesis of the major cellular components occurs during oogenesis. However, some subtle differences between the some key metabolites were observed in the spectra. Most notably, spermine was consistently detected in the unfertilized eggs, but was absent in the 32-cell stage embryos. It is well documented that the concentration of spermine decreases during oogenesis leading to the reduction of total polyamines required for maturation [43,44].
In both jellied and dejellied systems, the intensities for the following ions from stage 8 embryos were consistently lower than from unfertilized eggs: m/z 113. 884

Conclusions
There has been extensive of the transcriptome that regulates the major developmental transitions in X. laevis at the subcellular level. These investigations have been limited to the detection of mRNAs by in situ hybridization and RT-PCR approaches and to the detection of proteins by immunohistochemistry or Western blotting techniques. The latter have been hampered by the lack of antibodies against proteins of interest. In addition, these approaches reliably detect mRNAs and proteins that are in high abundance. Metabolites and lipids have not been studied in detail at in single eggs because of the lack of appropriately sensitive technologies that can assess these molecules in the living egg/embryo in its natural state. Therefore, evaluation of eggs and embryos with LAESI-MS was performed to determine whether subtle changes in metabolites and lipids and their subcellular profiles are detectable.
In this report the direct atmospheric pressure analysis of X. laevis eggs and early cleavage stages embryos is presented. The results of the LAESI-MS metabolic and lipidomic profiles of the Xenopus eggs were compared to previous literature results and showed a wide coverage for lipids and small metabolites. Reactive-LAESI MS/ MS, with in-plume cationization of PC lipids without sample preparation, was  utilized to identify the acyl chains. An investigation of the animal and vegetal poles of the X. laevis ovum showed increased abundance of lipids in the vegetal pole relative to the animal pole. Radial profiling of a jelly coated egg by LAESI-MS utilizing consecutive laser pulses revealed dramatic compositional changes between the jelly coat the vitelline/plasma membranes and the cytoplasm. Finally, LAESI-MS was used to demonstrate that subtle metabolic profile changes could be detected after the egg was fertilized and synchronous cell cycles were initiated. Of particular note, spermine was depleted after fertilization, which consistent with the reduction of total polyamines required for embryo maturation [43,44]. Major differences between the metabolites of unfertilized egg and the early cleavage stage embryos were not expected because all the needed organelles and biochemical building blocks needed to carry the embryo through the first 8 hours of development had previously been synthesized and stored in the oocyte. Nonetheless, the sensitivity of LAESI-MS enabled the detection of a few subtle differences, and uncovered novel molecules that are asymmetrically distributed between the animal, ectoderm-forming, and vegetal, endoderm-forming, regions of the egg. This simple approach, partially validated here for a vertebrate embryo  for which there is abundant traditional biochemical data, should be ideal for the direct analyses of other embryos that are rare and smaller in size.