Composition and Localization of Lipids in Penaeus merguiensis Ovaries during the Ovarian Maturation Cycle as Revealed by Imaging Mass Spectrometry

Ovary maturation, oocyte differentiation, and embryonic development in shrimp are highly dependent on nutritional lipids taken up by female broodstocks. These lipids are important as energy sources as well as for cell signaling. In this study, we report on the compositions of major lipids, i.e. phosphatidylcholines (PCs), triacylglycerols (TAGs), and fatty acids (FAs), in the ovaries of the banana shrimp, Penaeus merguiensis, during ovarian maturation. Thin-layer chromatography analysis showed that the total PC and TAG signal intensities increased during ovarian maturation. Further, by using gas chromatography, we found that (1) FAs 14∶0, 16∶1, 18∶1, 18∶2, 20∶1, and 22∶6 proportionally increased as ovarian development progressed to more mature stages; (2) FAs 16∶0, 18∶0, 20∶4, and 20∶5 proportionally decreased; and (3) FAs 15∶0, 17∶0, and 20∶2 remained unchanged. By using imaging mass spectrometry, we found that PC 16∶0/16∶1 and TAG 18∶1/18∶2/22∶6 were detected in oocytes stages 1 and 2. PCs 16∶1/20∶4, 16∶0/22∶6, 18∶3/22∶6, 18∶1/22∶6, 20∶5/22∶6, and 22∶6/22∶6 and TAGs 16∶0/16∶1/18∶3, 16∶0/18∶1/18∶3, 16∶0/18∶1/18∶1, and 16∶0/18∶2/22∶6 were present in all stages of oocytes. In contrast, the PC- and TAG-associated FAs 20∶4, 20∶5, and 22∶6 showed high signal intensities in stage 3 and 4 oocytes. These FAs may act as nutrition sources as well as signaling molecules for developing embryos and the hatching process. Knowledge of lipid compositions and localization could be helpful for formulating the diet for female broodstocks to promote fecundity and larval production.


Introduction
The banana shrimp, Penaeus merguiensis, is one of the most important commercial shrimps in the Indo-Pacific region [1]. In Thailand, more than 90% of the shrimp production is cultured, and the most popular shrimp species which have been selected for cultured are Penaeus monodon and Litopenaeus vannamei. The former is considered to have superior meat quality and commands higher price, while the latter is an imported species having lower quality meat. However, the large scale and high density of P. monodon farming caused increases of disease outbreaks. An alternative indigenous species of equal quality is the banana shrimp, P. merguiensis, which is also highly favored by consumers [1,2]. This species has a large natural population that can provide sufficient broodstocks; however, to have any commercially viable aquaculture the broodstocks must be domesticated and provided with suitable feeds which must be formulated based on the lipid compositions of the developing and ripe gonads. Recent attempts to culture this species employed practical procedures based entirely on the knowledge of P. monodon. Since the details of gonadal lipids profiles of this species could be different from other marine species, to improve the production of P. merguiensis, especially female broodstocks, information on ovarian lipid composition is important as it could be used for formulating a suitable feed to be used as a convenient diet for the broodstock. During ovarian maturation, lipids are one of the essential molecules taken up by female broodstocks and stored in the oocytes to nourish subsequent embryonic development. The composition of lipids in the ovaries of several shrimp species have been studied, including Penaeus semisulcatus [3], Chorismus antarcticus [4], L. vannamei [5], Pandalus montagui [6], Penaeus esculentus [7], and Macrobrachium rosenbergii [8]. In general, dietary lipids are the essential sources for fatty acid, phospholipid, sterol, and carotenoid syntheses [9]. Most of the lipids are synthesized and stored in the hepatopancreas before being transported to the ovary to be incorporated into oocytes at various stages during the ovarian cycle [3]. However, the ability to synthesize certain essential lipids, including polyunsaturated fatty acids (PUFAs), is limited in shrimps, so that these lipids must be acquired from the diet [10]. These lipids have effects on growth and reproductive success in female shrimps through stimulating ovarian maturation and oocyte differentiation [10,11]. In addition, phosphatidylcholine (PC) and triacylglycerol (TAG) are reported to be predominant lipids in shrimp ovaries and are required for embryogenesis, improvement in hatching, and nauplii production [12]. A number of studies on fatty acid (FA), PC, and TAG compositions in the ovaries of several shrimp species has been reported [5,7,13,14]; however, no similar studies have been conducted in P. merguiensis. In this study, we focused on the compositions of 3 kinds of lipids,i.e., PCs, TAGs, and total FAs including PUFAs, particularly linoleic acid (18:2), linolenic acid (18:3), arachidonic acid (ARA, 20:4), eicosapentaenoic acid (EPA, 20:5), and docosahexaenoic acid (DHA, 22:6). Moreover, we examined the localization of PCs, TAGs, and associated FAs in the P. merguiensis ovaries using imaging mass spectrometry (IMS). IMS is a technique that is useful for detecting the distribution of analytes directly on the tissue surface. The analytes that can be detected from this technique include pharmaceuticals, lipids, peptides, proteins, metabolites, and polymers [15,16,17,18,19,20]. Since small metabolites are highly abundant in living organisms and easily detected by mass spectrometry, especially phospholipids and glycerolipids [21,22], the IMS technique is a powerful tool for detecting these types of molecules and can performed without the requirement for labels [23,24]. According to this technique, after the laser irradiation of a tissue surface, analyte maps were constructed from the several hundred types of molecules that were detected [25]. Lipid accumulation is an important process for promoting successive hatching and embryo development; however, this technique has not been used to examine the distribution of individual lipids in a penaeid shrimp. The distribution of lipids, as determined in this study using the IMS technique, provided information on the changes in lipid accumulation during ovarian maturation. This information could be used to formulate the diet of the female broodstock of this shrimp, thereby help to promote optimal ovarian development and oocyte differentiation, and to improve embryo development and egg production and quality.

The histology of 4 ovarian stages during the ovarian cycle
To obtain the ovarian histology, hematoxylin-and eosin (H&E)stained ovarian sections from 4 ovarian stages were examined and photographed by a Nikon light microscope equipped with a digital camera E600. Each ovary was observed to be separated into lobules by connective tissue trabeculae (Tr) (Fig. 1A, C, E, G). In addition, each lobule was in contact with an oogenetic zone (OZ), which is an epithelial component lining the central ovarian cavity and containing the oogonia (Og) (Fig. 1A, C). Depending on the stage, each lobule might contain variable areas of a previtellogenic zone (PZ) (Fig. 1C, D), a vitellogenic zone (VZ) (Fig. 1E, F), and a mature zone (MZ) (Fig. 1G, H). In the stage I (the spent stage) (Fig. 1A, B), the OZ was quite large, and each lobule contained a small PZ area with stage 1 oocytes (Oc1) as the predominant germ cells and some stage 2 oocytes (Oc2). The Og and Oc1 diameters were approximately 10-14 and14-21 mm, respectively. Hematox-ylin stained the Og cytoplasm blue and the Oc1 a deeper blue. In the stage II ovary (the proliferative stage) (Fig. 1C, D), each lobule contained mostly the PZ, and the predominant germ cells were Oc2. These oocytes increased in size (diameter ranging between 25-40 mm) over the span of the PZ; the cytoplasm stained light blue, and the chromatin was less densely packed in the nucleus when compared to that of Og and Oc1. In the stage III ovary (the premature stage) (Fig. 1E, F), each lobule primarily contained the VZ. The majority of cells were stage 3 oocytes (Oc3) that increased in diameter, ranging between 45-77 mm. The cytoplasm stained purple pink because of increased eosinophilia; an increased accumulation of lipid droplets was also evident. The nuclei could not be clearly observed at this stage, owing to the beginning of germinal vesicle break down. In the final stage or stage IV ovary (Fig. 1G, H), also known as the mature stage, each lobule contained mostly the MZ. The most abundant germ cells were stage 4 (Oc4) or mature oocytes. The cytoplasm stained a deep pink because of very high eosinophilia; the nuclei could not be observed in these oocytes due to complete breakdown of the germinal vesicle. The most prominent feature unique to Oc4 was the presence of cortical rods (Cr) that were arranged radially throughout the cytoplasm. The Oc4 were the largest oocytes in the ovaries, with diameters ranging between 115-197 mm (Fig. 1G, H), and were surrounded by follicular cells, which are the supporting cells in the ovary.

Identification of PCs and TGs by thin-layer chromatography
Extracted lipids were separated by thin layer chromatography (TLC) to estimate the changing amounts of PCs and TAGs during the ovarian cycle. Figure 2 shows that the intensity of the PC band ( Fig. 2A) dramatically increased from stage I to stage IV, whereas the TAG band ( Fig. 2B) was absent in the stage I ovary, became increasingly prominent in stages II and III, and reached a maximum in stage IV. We validated that these bands were real PCs and TAGs by using TLC-Blot-MALDI MS (data not shown). Figure 2C and 2D show histograms of the calculated signal intensities of the PCs and TAGs. The results indicated that PC levels increased from stage I to stage III ovaries and remained constant from stage III to IV. On the other hand, during stage I, TAGs were barely detectable, but dramatically increased from stage II to IV.

Identification of fatty acid (FA) compositions by GC-MS
We found that FAs in the P. merguiensis ovary were composed of both saturated and unsaturated FAs. The saturated  Table 1. To estimate the FA accumulation rates, the ratio of each FA to the total FAs was calculated (Fig. 3B). From these data, the

Localization of lipids by imaging mass spectrometry
In this study, we observed specific PC and TAG signals that indicated their distributions in 3 different patterns, i.e. lipids localized mainly in Oc1,2 or Oc3,4 or present in all oocyte stages (Tables 2, 3). The first group of lipids detected in Oc1,2 comprised a PC at m/z 756.5 and a TAG at m/z 951.7. The second group detected with high signal intensities in Oc3,4 comprised PCs at m/ z 780.5, 782.5, 804.5, 806.5, 808.5, 820.5, 826.5, 830.5, 832.5, 844.5, 852.5, and 870.5 (Table 2) and TAGs at m/z 823.6, 825.7, 851.7, 853.7, 877.7, and 879.7 ( Table 3). The last group were lipids present in all oocyte stages and comprised PCs at m/z 802.5 828.5, 850.5, 854.5, 874.5, and 900.5 (Table 2) and TAGs at m/z 827.6, 855.6, 881.6, and 925.7 (Table 3). In this group, the signal intensities were found to be higher in Oc3 and Oc4 than in Oc1 and Oc2. Figure 4 shows the merged H&E-stained ( were selected to illustrate the distribution patterns. In Figure 4B, the signal of the molecule at m/z 780.5 was detected in Oc3,4 and dramatically increased in these cells during ovarian maturation. In contrast, the molecule at m/z 828.5 was distributed in all oocyte stages, with an especially high signal intensity detected in Oc3,4. For TAG, the molecules at m/z 881.6 and 925.7 were distributed in all oocyte stages, and the signal intensity slightly increased from ovarian stages I to IV. We have also provided the IMS images of other lipid molecules in Figures S1 and S2.  [26,27], the molecule at m/z 780.5 showed a fragmented major peak at m/z 146.9, which represented a neutral loss of the PC head group adducted with Na + and corresponded to [(CH 2 ) 2 PO 4 H+Na] + . Moreover, we were able to detect two minor peaks at m/z 441.2 and 465.2, which corresponded to neutral losses of trimethylamine ((CH 3 ) 3 N) from the PC head group plus FAs (59+280 and 59+256). The molecular weights of 256 and 280 Da could be assigned as FAs 16:0/18:2. This identification was also matched with an online database search (http://www.hmdb.ca/labm/jsp/mlims/ MSDbParent.jsp) and Simlipid software version 1.0, which helped to increase the confidence of the lipid identification. Similarly, the molecule at m/z 828.5 showed neutral losses of FAs by the presence of 2 peaks at m/z 441.2 and m/z 513.2, corresponding to trimethylamine ((CH 3 ) 3 N) from the head group plus FAs (59+328 and 59+256). The identification of the other lipid molecules was performed in a similar manner; and the identified lipid species are shown in Table 2 for the PCs. In case of the TAGs, the molecules at m/z 881.6 ( Fig. 5C) and 925.7 (Fig. 5D) are shown as fragmented ions after MS/MS analysis. The identification of TAGs is different from that of PCs. According to a previous study [27], the peak corresponding to the neutral loss of a FA can be observed directly in the mass spectrum. For the molecule at m/z  Table 3 for TAG identification. The underlined molecules represent those that were observed as fragmented ions from MS/MS, whereas the nonunderlined molecules could be obtained as partially fragmented ions. These partially fragmented ions were further verified to validate their identities by an online database search and SimLipid software.

Discussion
As shown in our TLC analysis (Fig. 2), the increasing PC and TAG intensities during ovarian maturation in P. merguiensis were similar to those of another penaeid shrimp, P. semisulcatus.  (Fig. 3). Thus, the composition of the major FAs in P. merguiensis ovary detected in this study are quite similar to those reported in L. vannamei [5] and P. monodon [28]. However some FAs were found to be different to those reported in M. rosenbergii [8] and L. Vannamei. In P. Merguiensis and L. Vannamei the proportion of FAs 16:0 and 18:0 found to be decreased whereas in M. rosenbergii the first FA found to be increased and the later FA remained unchanged. In addition proportion of FAs 22:6 found to be decreased in M. rosenbergii and L. Vannamei but not in P. Merguiensis. The different proportion of some FAs among shrimp species indicating that the requirement to synthesis and store each FA to utilize during ovarian maturation in each shrimp species is different. Therefore to promote the successive ovarian maturation to the specific shrimp species the knowledge of FA requirement is important as to be used to formulate the suitable feeding diet. Moreover, because of the sensitivity of the technique used in our study, we were also able to detect the minor FAs comprising 14:0, 15:0, 17:0, and 20:1.
Based on our findings, the changing FA levels in relationship to the total lipids during the ovarian maturation cycle exhibited 3 patterns: decreasing, increasing, and unchanging (Fig. 3B). In the first group FAs comprised 16:0, 18:0, 20:4, and 20:5. We observed high levels of the saturated FAs 16:0 and 18:0 in the stage I ovary as well as in early stage oocytes (Oc 1, 2). In a previous study, FAs 16:0 and 18:0 were found to be the major saturated FAs in mammals and fishes [29]. Further, these FAs were suggested to be nutritional sources and to serve as precursors for the syntheses of other FAs. In M. rosenbergii, relatively high levels of these FAs are present in the early stages of ovarian development, indicating they may be required for the development of the earlier oocyte stages, as they could be effectively digested in this prawn [30]. Furthermore, the prawns were also able to convert these saturated FAs to monounsaturated FAs to be stored in the ovary and subsequently utilized by the embryo [30]. Therefore, the decreased proportional levels of these FAs as the ovary becomes more mature may be due to their increased utilization to satisfy the energy requirement of the developing oocytes, and also due to their conversion to unsaturated lipids. Moreover, we found that the relative amounts of FAs 20:4 and 20:5 also decreased (Fig. 3B) as in P. monodon [11]. During the progression of the ovarian cycle, the reproductive hormone prostaglandin is required to enhance ovarian maturation and ovulation [11]. Therefore, the reduced levels of 20:4 and 20:5 in P. merguiensis may be due to the high requirement of prostaglandin biosynthesis, which uses these FAs as precursors. This is supported by the same study showing that high prostaglandin in the feed shortened the period of gonadal maturation in this shrimp [11].
In the last group, the FAs displaying unchanged levels comprised 15:0, 17:0, and 20:2. We classified the level of FA 17:0 as an unchanged level, since first we observed slightly decreased proportion of FA 17:0 from ovaries stage I to III and finally the level of this FA at stage IV remained stable. This finding agree with the previous study in L. vannamei which found the level of FA 17:0 slightly fluctuated [5]. All FAs in this group are minor components, indicating that they might not be directly involved in reproductive performance.
In addition to examining the fatty acid compositions and their changing levels in all 4 ovarian stages, the present study is the first one to report on the spatial distribution of lipids in various oocyte stages during ovarian maturation in a penaeid shrimp (Fig. 4, S1, S2). Normally, the lipid composition is reported as total FAs extracted from a tissue that usually includes FAs from PCs and TAGs [3,5,6,7,13,14,31]; however, the specific localization of each lipid has never been investigated. In this study, we were able to detect the distribution of PCs, TAGs, and associated FAs directly in the ovarian tissue sections by IMS and thereby provide information on the accumulation of these lipids during oocyte maturation, and enabling us to identify the FAs that might be required for each stage of oocyte development. Previous study in penaeid shrimp has shown that during sexual maturation, the remarkable increases of PCs and TAGs were observed [12]. A study in Penaeus japonicus [32] showed that the ovarian maturation period was lengthened when the diet did not contain phospho- lipids. Hence PC may be essential for oocyte maturation. As for TAG, this molecule was incorporated into the eggs and act as energy source to supply embryogenesis, hatching, egg quality and nauplii development [12]. Furthermore in our present study, by using IMS technique, we could localize and classify PC-and TAGassociated FAs that were selectively stored in each oocyte stages during ovarian maturation ( We suggest that these molecules might act as the essential nutritional and energy source for developing embryo as they were accumulated in the late and mature stages of oocytes. The reproductive capacity of fish and shrimp depends on many factors, such as hormones [33], dietary intake [34], and the amounts and types of lipids in the diet [35]. Amongst lipids it has been reported that polyunsaturated (PUFA) and highly unsaturated fatty acids (HUFA) play the key roles in stimulating female fecundity. In many species of fish it has been clearly demonstrated that HUFA (DHA, 22:6; EPA, 20:5 and ARA, 20:4) are essential for female reproduction and fecundity, as well as the quantity and quality of the offspring, even-though the exact mechanism and targets of each fatty acid are not known [9,28,34,35]. Further-  Table 2 (the PCs) and 3 (the TAGs). The black scale bars represent 400 mm. doi:10.1371/journal.pone.0033154.g004 more, in M. rosenbergii feeding female broodstocks with high level of 18:2n-6 together with n-3 HUFA increased fecundity, egg hatchability and larval quality [30]. It is known that some PUFA, for examples, linolenic acid (LNA), 18:3n-3, and linoleic acid LA, 18:2n-6, can serve as precursors for EPA/DHA and ARA, respectively [30]. Taken together the high level of PUFA in P. merguiensis ovary indicates that this is an essential group of fatty acids that must be included in the broodstock feed. It is notable that LNA is lacking in this prawn, hence inclusion of EPA and DHA in the feed is also essential. Whereas the detection of high level of saturated FAs 16:0 and 18:0 suggested that these molecules could serve as major energy source for the maturing oocytes. Thus, information on lipid compositions and localization, as performed in this study, could provide basic knowledge on FA requirements for oocyte differentiation during ovarian maturation in this shrimp species, and could be used in formulating feed for female broodstocks to enhance their fecundity and larval production.

Histology of the developing ovary
Thirteen female P. merguiensis shrimp (weighing between 18-22 g, with length between 20-25 cm.) were obtained for each of the 4 stages of the ovarian cycle from the Coastal Aquatic Feed Research Institute, Petchburi Province, Thailand. Shrimp ovaries were quickly dissected and immediately fixed in Bouin's fixative overnight. The fixed samples were processed for paraffin embedding by using a LEICA TP 120 tissue processor (Leica, Germany). The samples were finally embedded in paraffin blocks and cut by a rotary microtome (Leica RM2235, Germany) into5mm thick sections, which were mounted on glass slides and stained with H&E. The H&E images of the ovaries were captured by a Nikon ECLIPSE E 600 light microscope equipped with an E66 digital camera (Nikon, Tokyo, Japan).

Identification of lipids by thin-layer chromatography
Extraction of lipids. The extraction procedure was performed as follows: ovarian samples were weighed, homogenized, and extracted with 0.1 g/ml of extraction solution (chloroform:methanol 2:1, v/v). The homogenates were placed in a glass tubes (pre-treated with chloroform:methanol 1:1, v/v), and sonicated more than 10 times by a Microson Ultrasonic Cell Disruptor XL-2000 (Wakenyaku Co. Ltd., Kyoto, Japan) for a duration of 10 sec per sonication, with 5-sec stops. The sonicated samples were incubated overnight at room temperature. The samples were centrifuged at 3,0006g, and the supernatants were collected into new glass tubes. Afterwards, the glass tubes were tightly wrapped with parafilm and stored at 280uC until sample analysis.

Thin layer chromatography
The separation procedure followed a previously described method [36,37,38]. In brief, 3 ml of the extracted lipids were applied to a 10-610-cm high performance thin layer chromatography (HPTLC) plate (Merck, Darmstadt, Germany) and then completely dried before being placed into the TLC chamber. To enable better separation and prevent lipid diffusion, the applied areas were prevented from being immersed in the separation buffer. For PC separation, the HPTLC plate was incubated in a separation buffer composed of methylacetate:n-propanol:chloroform:0.25% KCl at the ratio 25:25:10:9 (v/v/v/v), whereas for TAG separation the plate was incubated in a buffer composed of N-hexane:diethylether:acetic acid at the ratio 80:30:1 (v/v/v), for 30 min. After separation, the PC and TG plates were completely dried. To observe the separated lipids, primuline solution (1 mg of primuline in 100 ml of 80% acetone in water) was sprayed onto the HPTLC plates, which were then completely dried, and the

Analyses of fatty acid composition by gas chromatography-mass spectrometry
Methylation and purification of FAs. The extracted FAs from each ovarian stage were subjected to methylation using an acid methylation kit (NacalaiTesque, Inc., Kyoto, Japan) before being analyzed by gas chromatography-mass spectrometry (GC-MS). The internal standard arachidicacid (20:0) was added to the extracts at concentration 0.4 mg/ml. The extracts were then completely dried by a TurboVap LV Evaporation System (Caliper Life Sciences, Hopkinton, MA, USA) and continuously aerated with nitrogen gas to prevent the oxidative reaction. After the methylation procedure, the samples were purified by using a fatty acid methyl ester purification kit (Nacalai Tesque, Inc.), according to the manufacture's instruction. The purified FAs were stored at 280uC until analyzed. D.F., Agilent technologies). The column temperature was held at 210uC. The column pressure was increased from 110 kPa to 380 kPa at a rate of 7 kPa/min (from 110 kPa to 180 kPa) and 10 kPa/min (from 180 kPa to 380 kPa). Significant differences (p,0.05) among the lipids of the ovarian stages were tested using Scheffe's method.

Localization of lipids by imaging mass spectrometry
The fresh frozen ovaries from each stage were serially cut at 220uC by a CM 1950 cryostat (Leica Microsystems) at a thickness of 10 mm, and the sections mounted onto indium tin oxide (ITO)coated microscope glass slides (BrukerDaltonics, Bremen, Germany). The ovarian tissues were not embedded in any embedding medium lest any residual polymers on the tissue sections might interfere and degrade the mass spectra [39]. The tissues sections were stored at 280uC until being analysed. Before applied matrix, the tissue sections were completely dried by a hair dryer. The matrix, 2,5-dihydroxybenzoic acid (DHB) (Bruker Daltonics), was dissolved at 50 mg/ml in 70% methanol and 0.1% TFA, and then sprayed onto the tissue sections by using a 0.2-mm nozzle caliber airbrush (Procon Boy FWA Platinum, Tokyo, Japan). The sprayed samples were analysed using an Ultraflex II MALDI TOF/TOF mass spectrometer (Bruker Daltonics) in positive ion mode. The mass spectra were acquired in the mass ranges between m/z 500-1,500, and the laser frequency was optimized to 200 Hz. The number of laser irradiations was 200 shots in each spot. For obtaining a clear signal distribution on the tissue sections, the resolution of the images was optimized by setting the laser raster width at 35 mm per laser spot. After analysis, the images were constructed by using the imaging software tool, Flex imaging 2.1 (Bruker Daltonics). The H&E stained was performed by using the same tissue sections after IMS analysed (Fig. 4A).

Identification of lipid species by MS/MS analysis
The extracted lipids were subjected to tandem mass spectrometry (MS/MS) analysis. To validate the identity of each lipid species, the QSTAR Elite high-performance, hybrid quadrupole Time-of-Flight mass spectrometer (Applied Biosystems, Foster city, CA) was used to acquire fragment ion spectra. The collision energy was optimized and adjusted between 40 to 50 V during fragmentation of the lipid molecules. An online lipid database search, using Metabolite MS Search (http://www.hmdb.ca/ labm/jsp/mlims/MSDbParent.jsp) and the software SimLipid Version 1.0 (PREMIER Biosoft International, Palo Alto, CA, USA), was performed to identify molecules from their partial fragment spectra. Figure S1 Imaging mass spectrometry showing the distribution of each PC molecule in various ovarian stages. The distributions of these molecules exhibit 3 patterns: the first group is mainly distributed in Oc 1,2; the second group, Oc 3,4; the last group, all oocytes. The molecular species and their distributions are shown in Table 2. (TIF) Figure S2 Imaging mass spectrometry showing the distribution of TAG molecules in various ovarian stages. Similar to the PCs, the TAGs also show 3 patterns of distribution, i.e., mainly in Oc1,2, in oocyte 3,4, or in all stages. The molecular species and their distributions are shown in Table 2.