Single embryo and oocyte lipid ﬁ ngerprinting by mass spectrometry

Methods used for lipid analysis in embryos and oocytes usually involve selective lipid extraction from a pool of many samples followed by chemical manipulation, separation and characterization of individual components by chromatographic techniques. Herein we report direct analysis by matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) of single and intact embryos or oocytes from various

CAMP), which follows the Ethical Principles of Animal Research established by the Brazilian College for Animal Experimentation (COBEA) under protocol number 1752-1. The use of human unfertilized oocytes was approved by the São Paulo Federal University (UNIFESP) Institutional Committee for Ethics under the protocol number 0411/07. Immature bovine ( Bos taurus ) and sheep ( Ovis Aries ) oocytes were obtained by post mortem follicular aspiration of ovaries from cows and ewes slaughtered at commercial slaughterhouses. Ovaries were transported in 0.9% (w/v) saline solution at 25-30°C to the laboratory and follicles were aspirated using an 18-gauge needle attached to a 20-ml syringe. Cumulus oocyte complexes with at least three layers of cumulus cells and homogeneous cytoplasm were denuded of cumulus cells by gentle pipetting in 0.5% hyaluronidase.
Bovine in vivo-derived embryos were obtained from superovulated heifers kept in pasture ( Brachiaria decumbens ). The emergence of the animal's follicular wave was synchronized by one intramuscular (i.m.) injection of 2.0 mg estradiol benzoate (Estrogin, Farmavet, São Paulo, Brazil) and insertion of an intravaginal controlled internal drug release device (CIDR), containing 1.9 g progesterone (Pfi zer, Hamilton, New Zealand) on day 0. On day 4.5, the superstimulatory treatments were initiated (follicle stimulating hormone [FSH]; Folltropin-V, Bioniche Animal Health, Belleville, Canada) and given in decreasing doses of 28, 21, 14 and 7 mg FSH twice daily, over a 4-day period, for a total dose of 70 mg. At the time of the fi fth and sixth injections of FSH, 25 mg of dinoprost tromethamine (Lutalyse, Pfi zer, Paulinia, Brazil) was injected intramuscularly. The CIDR was removed at the time of the seventh superstimulatory injection. Ovulation was induced with an i.m. injection of 0.05 mg gonadotrophin releasing hormone (GnRH; Gestran Plus; ARSA S.R.L., Buenos Aires, Argentina) 12 h after the last superstimulatory injection. All heifers were artifi cially inseminated (AI) with frozen/thawed semen from the same bull 12 and 24 h after GnRH injection. Seven days after the fi rst AI, embryos/ova were recovered using a nonsurgical uterine fl ushing technique ( 20 ).
For bovine in vitro embryo production, oocytes obtained by post mortem follicular aspiration were in vitro fertilized and cultured as described previously ( 21 ), but with changes in culture medium supplementation and incubator atmosphere. After IVF, embryos were cultured in four different in vitro conditions: 5% O 2 and BSA (Group 1); 20% O 2 and BSA (Group 2); 20% O 2 and FCS (Group 3); and 5% O 2 and FCS (Group 4).
Human unfertilized oocytes from women submitted to transvaginal oocyte retrieval were provided by the Human Reproduction Service of São Paulo Federal University (UNIFESP) in Brazil. These samples would normally be discarded.
Mullet ova ( Mugil spp.) were collected from fresh fi sh in the City Fish Market of Santos-SP, Brazil, and transported at 4°C to the laboratory.
Fire ant ( Solenopsis spp.) eggs were collected in the fi eld after species identifi cation and were immediately transported to the laboratory.
Oocyte and embryo samples were stored in microtubes containing 100 l of a 50% aqueous methanol solution at -80°C until analysis. Sample preparation involved placing each oocyte, egg or embryo in a given spot of the target plate under the stereomicroscope. Samples were allowed to dry at room temperature, and their location was recorded in order to place the laser at the correct location during analysis. Just before analysis, 1 l of 1.0 mol/l 2,5-dihydroxybenzoic acid (DHB) in methanol was placed in each target spot and allowed to dry at room temperature. from patients undergoing cancer treatment ( 8 ); gametes and embryos for conservation of animal genetic resources; and to facilitate the international transit of genetic material ( 9 ). Cytoplasm lipid accumulation has been shown to affect the rate of post-thaw development of bovine embryos (10)(11)(12), which suggests that the embryo culture process causes lipid metabolism changes that affect the properties and stability of cell membranes ( 13 ). In laboratories for assisted reproduction for humans, embryos and oocytes are routinely cryopreserved, but there is still considerable need of research aiming to improve knowledge of the role of lipids and the effi ciency of protocols used (14)(15)(16).
To date, changes in PL composition profi les of mammalian embryos and oocytes are poorly understood. Gas chromatography (GC) has been applied to study triacylglycerol (TAG) composition in human and bovine oocytes and embryos (17)(18)(19), but this approach requires chemical transformation via saponifi cation and derivatization, providing free fatty acyl residue profi les but not TAG or PL identifi cation. A main drawback for GC analyses is the need of a large pool of samples (more than 10 oocytes or embryos for each lipid analysis), especially when human embryos or oocytes are investigated ( 19 ).
MALDI-MS offers, however, the advantage of detecting intact lipids, and due to its unmatched high spatial resolution, allows targeting single species of microscopic dimensions (100-200 m) such as mammalian oocytes and embryos. Herein we show that direct analysis by MALDI-MS can contribute signifi cantly to embryo and oocyte lipidomic studies. With no solvent extraction and no chemical manipulation, direct MALDI-MS of single and intact embryos or oocytes of different species (from invertebrates to mammalian) is shown to provide reproducible lipid profi les. Besides species recognition, the method described herein is also shown to be sensitive and, hence, to reveal important changes in lipid profi les during bovine preimplantation development and in bovine embryo in vitro culture conditions.

Reagents
Unless mentioned, chemicals and culture media for handling bovine and ovine oocytes were purchased from Sigma (St. Louis, USA). Media used for human oocytes handling (modifi ed Human Tubal Fluid [HTF]) was purchased from Irvine Scientifi c (Santa Ana, CA, USA). Phosphate buffer saline (PBS) solution was supplied by Nutricell (Campinas, SP, Brazil).
Methanol (ACS/HPLC) grade was purchased from Burdick and Jackson (Muskegon, MI, USA) and 2,5-dihydroxybenzoic acid (DHB) was purchased from ICN Biomedicals (Aurora, OH, USA). Ultrapure water, purifi ed by a Direct-Q water system (Millipore, Bedford, MA, USA) was used for the preparation of solvents.

Sample collection and storage for analysis
The care and use of animal samples (bovine, sheep, fi sh and fi re ant) was approved by the Institutional Committee for Ethics in Animal

Lipid fi ngerprinting in different species
MALDI-MS analysis and sample preparation (described in detail in the "Material and Methods" section) involved no lipid extraction and no chemical manipulation which, as already mentioned, would require substantial sample pooling of these highly valuable and sometimes unavailable (particularly in large numbers) microscopic structures. An intact single embryo or oocyte was collected and then placed in one of the many spots of the MALDI target plate placed under a stereomicroscope ( Fig. 1A ). Then, the target plate spot was moistened with the MALDI matrix before spectra acquisition ( Fig. 1B ). Fig. 2 displays representative MALDI-MS spectra from single embryos or oocytes from the fi ve species investigated in this work. The spectra were acquired in the m/z 700-950 range, where most PL and TAG should be detected by MALDI(+)-MS mainly as either their protonated [M + H] + or sodiated molecules [M + Na] + , or both ( 22 ). Note that the MALDI-MS profi les are rather characteristic, and in some cases, very distinctive with unique PC, SM and TAG ions (see below) with contrasting abundances or ratios. Group comparison by principal component analysis (PCA; see below) was also found to allow their prompt characterization.
The PL species were fi rst attributed ( Table 1 ) based on previous lipid profi le studies by MALDI-MS and are described by the class abbreviation followed by the total number of carbons and double bounds in the acyl residues attached to the glycerol backbone (in parenthesis). The MALDI-MS of the human oocyte is characterized

Lipid analysis by MALDI-MS
MALDI-MS and MALDI-MS/MS spectra were acquired in the positive ion and refl ectron modes using a Q-ToF Premier mass spectrometer (Waters, Manchester, UK) equipped with a 200-Hz solid-state laser in the m/z range of 700-950. The principal operating condition used was 10 V (sample plate), and laser irradiation consisted of diverse shots during the time of 60-90 s in the region were the sample had been placed on the target plate, until signals in the region of interest were observed and disappeared due to the consumption of the microscopic sample ( Fig. 1 ). MALDI-MS/MS were manually acquired by increasing the collision energy until extensive dissociation of the precursor ion was observed. Argon was used as the collision gas. Spectra were centered and aligned using the MassLynx 4.0 software (Waters, Manchester, UK). From each spectrum, after the exclusion of isotopic peaks, the fi fty most intense ions were considered as the starting point for searching m/z values corresponding to lipids. After attribution, only the m/z values which were clearly distinct from noise level in the spectra were included in the principal component analysis (PCA), which was performed using Pirouette v.3.11 (Infometrix Inc., Woodinville, WA, USA).

RESULTS
As proof-of-principle cases, we investigated the potential of MALDI-MS fi ngerprinting to provide characteristic lipid profi les from a representative set of single embryos and oocytes from mammals and from two distantly related species: insect and fi sh. We then evaluated the ability of the technique to reveal changes in the lipid profi les during embryo development and under different in vitro culture conditions for bovines as a representative biological model. The MALDI-MS of the insect embryo ( Fig. 2F ) is even more distinct and characteristic, displaying a much diverse and broader (in m/z range) set of abundant PC, SM and TAG ions. Indeed, about 30-40% of the dry weight of insect eggs are known to correspond to lipids, mainly of TAG (24)(25)(26).
PCA analysis ( Fig. 3A ) shows that all classes of samples can be resolved via their MALDI-MS lipid profi les. As could be expected, the mammalian samples [human (a1), bovine (a2) and sheep (a3)] are placed far away from the fi sh (a4) and ant (a5) samples, but they form a group located close together in the 3D PCA domain. They are, however, clearly resolved and grouped properly by species with little overlap.
When the MALDI-MS data for the mammalian samples were analyzed separately ( Fig. 3B ), group individualization in the 3D PCA plot was even more pronounced, with no overlaps. The fi sh and insect (ant) samples, as expected from their more distinct lipid distributions, form the bestresolved groups with a greater dispersion for the insect  Comparisons were performed considering the culture medium supplementation (Group 1 versus Group 2 and Group 3 versus Group 4) and atmosphere infl uence (Group 1 versus Group 4 and Group 2 versus Group 3). Regarding culture medium supplementation ( Fig. 4A , B  and Fig. 5A ), most embryos from BSA at low oxygen concentration presented less intense ions of m/z 760.6 and 782.6 [protonated and sodiated forms of PC 34:1, respectively] and of m/z 725.5 (sodiated SM 16:0) compared with higher oxygen concentration. The FCS culture medium supplementation effect is clearly noted at different oxygen concentrations ( Fig. 4C-D and Fig. 5B ), where the embryos cultured in high oxygen concentration present less intense PC 36:1 ions and higher PC 34:1 and SM 16:0 ions compared with lower oxygen concentration. Regarding the atmosphere infl uence, at low oxygen concentration, no infl uence of BSA or FCS was observed ( Fig. 4A, D and  Fig. 5C ). At higher oxygen concentration, PCA clearly separates BSA from FCS-supplemented embryos ( Fig. 4C, D and Fig. 5D ). The principal difference is that embryos from Group 1 presented more intense PC 36:1 ions than embryos from Group 4, which presented more intense PC 34:1 and the ion of m/z 725.6. Together with GC data on the fatty acyl residue composition of bovine oocytes ( 18,27 ), these results indicate that low oxygen concentration in the incubator and BSA as a medium supplement favor accumulation of PC containing 18:1 (PC 36:1) and not 16:0 (PC 34:1 and SM 16:0). Table 1 Table II lists m/z values for lipid ions followed by minimal and maximal relative ion abundances, as well as the percentage of samples containing the specifi c lipid in the fi ve different species, as individual variation is expected to occur in biological samples.

DISCUSSION
This article has described a practical and novel approach for embryo and oocyte lipid studies. Reproducible and characteristic MALDI-MS fi ngerprints can be obtained from a single oocyte or preimplantation intact embryo samples, probably due to the much greater diversity of the lipid ions detected in their MALDI-MS.
A 2D PCA ( Fig. 3C ) compared separately lipid profi les from bovine embryos and oocytes. Both types were very well resolved by proper grouping, which suggests an interesting application of single embryo and oocyte MALDI-MS fi ngerprinting, that is, monitoring changes in PL and TAG profi les during different developmental stages.
To investigate whether MALDI-MS of single embryos could monitor lipid changes owing to in vitro culture conditions, bovine embryos were cultured using four different in vitro conditions regarding incubator atmosphere and culture medium supplementation, respectively: 5% O 2 Reports of fatty acyl residue composition in oocytes describe the palmitic (16:0), stearic (18:0) and oleic (18:1) groups as the predominant fatty acyl residues in immature bovine oocytes, in unfertilized human oocytes ( 17,18 ) and in sheep oocytes ( 27 ). Herein, we have reported the structural composition of some PL species containing these fatty acyl residues, and we show that embryo and oocyte lipid fi ngerprinting by MALDI-MS allows differentia-sample in a procedure involving no extraction, no chemical manipulation and no preseparation. MALDI-MS is a powerful tool in lipidomics ( 34 ), generating PL and TAG profi les due to detection of singly charged molecular ions, which greatly facilitates data interpretation. We estimate about 10-15 pmol as the sensitivity for lipids from single embryo or oocyte MALDI-MS analyses as each bovine oocyte is known to contain this amount of PL ( 18 ). Oocytes from other mammals have similar sizes, and embryos have greater amounts of PL due to their higher amount of cells.
Research on the biological role of lipids in cells requires the determination, usually from numerous samples, of detailed lipid composition and changes during organism development. Methods traditionally used in oocyte and embryo lipid analysis, such as GC (17)(18)(19)27 ) and radioactive labeling ( 35 ), always demand the pooling of 10-400 ( 17,18,27 ) of these microscopic structures for lipid extraction with organic solvents, followed by chemical ma- tent of oocytes or embryos infl uence their cryosensitivity ( 43,44 ). Oocytes of all species are especially sensitive to low temperatures ( 45 ). Adding linoleic acid to bovine embryo culture medium was shown to enhance cryosurvival of embryos ( 42 ), increasing interest in studies involving the role of lipid composition in embryos with different cryosurvival rates.
In vitro culture conditions of bovine embryos are recognized to infl uence embryo developmental success, as well as embryo cryosensitivity (46)(47)(48). We evaluated the infl uence of the culture medium supplement (BSA or FCS) and incubator atmosphere (high oxygen concentration: 20% O 2 , and low oxygen concentration: 5% O 2 ) on the MALDI-MS lipid profi le (represented by SM, PC and TAG) of single bovine embryos. FCS is widely used as a supplement for embryo in vitro culture, but it alters embryo metabolism, stimulating lipid accumulation, as observed by staining methods and electron microscopy ( 47 ). Regarding incubator oxygen concentration, it has been suggested that lower oxygen concentrations (5% O 2 ) mimic physiological conditions and signifi cantly decrease oxidative stress in embryos ( 48 ). tion among species. Also, developmental stage differences in PC profi les between bovine oocytes and embryos (at the developmental stage of blastocyst) were indicated to occur. It is known that blastocyst formation involves cell differentiation into trophoectoderm and the inner cell mass. Metabolomic, gene expression and proteomic studies in bovine blastocysts have been performed (36)(37)(38). Nonetheless, membrane PC and SM changes and their roles during embryo development have not been reported. Changes in lipid patterns during oocyte development have been observed in Bufo arenarum oogenesis and early embryo development ( 39,40 ). Although used as a model for understanding oogenesis and embryogenesis, due to their high numbers of gametes and external fecundation, frog reproduction differs signifi cantly from that of mammalians species. MALDI-MS monitoring of single samples seems to be able to provide an important tool for studies of mammalian samples.
Fatty acyl residues have previously been shown to be associated with oocyte maturation and fertility ( 18 ) as well as with embryo development ( 19,35,41,42 ) in mammalian species. As an example, storage and membrane lipid con- known to present the high sensitivity needed to detect the few picomols of lipids present in these samples.
We believe that the fi gures of merit presented for MALDI-MS of single embryo analysis can still be improved. Altering the laser source (nitrogen or a Nd:YAG laser source) and particularly the MALDI matrix may allow more selective or broader detection of different lipids according to the specifi c biological question investigated. The use of other matrices, such as paranitroaniline ( 51 ), 2-(2-aminoethylamino)-5-nitropyridine ( 52 ), and 9-aminoacridine ( 53 ), has been reported for PL analysis. DHB (2,5-dihydroxybenzoic acid), although widely used as a matrix for lipid analysis, is not optimal for PL detection in the negative mode due to its acidic character and more effectively detects PC and SM in the positive ion mode ( 29 ). Recently, the use of 9-aminoacridine with improved preparation protocols for Our results indicate that low incubator oxygen concentration and BSA as a culture medium supplement favor PC containing 18:1 (PC 36:1) and not 16:0 (PC 34:1 and SM 16:0). These results correlate well with studies showing the benefi cial effects of low oxygen concentrations and serum-free medium ( 46,48 ) and the negative effect of palmitic acid (16:0) supplementation on embryo cryosurvival ( 49 ).
The high quality and selectivity of the fi rst single embryo MALDI-MS data presented herein point already to the suitability of the technique. Note that other direct desorption/ionization ambient techniques could be envisaged for single embryo MS analysis such as desorption electrospray ionization (DESI) and direct analysis in real time [DART ( 50 )], but they lack the desirable high spatial resolution characteristic of MALDI-MS, which is a vital feature for microscopic samples. Besides, MALDI-MS is The proof-of-principle data presented herein with characteristic spectra for fi ve species at different developmental stages confi rms the broad and successful application of direct single embryo lipid fi ngerprinting by MALDI-MS.

CONCLUSIONS
Single embryo and oocyte MALDI-MS of intact samples allows nearly direct lipid profi ling (represented herein by SM, PC and TAG species) with no extraction, chemical manipulation or pre-separation steps. Mass measurements are made with accuracy, high reproducibility and high sample throughput. Typifi cation and monitoring of some lipid changes during the developmental stages were shown to be dependent on species specifi city and were related to the environment. MAL-DI-MS of intact embryos and oocytes is also open to further improvements as other lasers and matrices may boost lipid detection. MALDI-MS, by providing lipid profi les for single embryos and oocytes from many different species, should contribute to improve cryopreservation and in vitro culture conditions of embryos and oocytes in general.
The authors are grateful to Prof. Carol Hollingworth Collins for English review, and to Amadeu Hoshi Iglesias, Juliana Hayashi Tannura, Dr. Aguinaldo Pereira Cedenho, Amanda Begatti Victorino and Juliana Stevanato for technical assistance during sample collection and spectra acquisition. Supplemental Material can be found at: