Proteomics Recapitulates Ovarian Proteins Relevant to Puberty and Fertility in Brahman Heifers (Bos indicus L.)

High fertility and early puberty in Bos indicus heifers are desirable and genetically correlated traits in beef production. The hypothalamus–pituitary–ovarian (HPO) axis synthesizes steroid hormones, which contribute to the shift from the pre-pubertal state into the post-pubertal state and influence subsequent fertility. Understanding variations in abundance of proteins that govern steroid synthesis and ovarian signaling pathways remains crucial to understanding puberty and fertility. We used whole ovaries of six pre-pubertal and six post-pubertal Brahman heifers to conduct differential abundance analyses of protein profiles between the two physiological states. Extracted proteins were digested into peptides followed by identification and quantification with massspectrometry (MS) by sequential window acquisition of all instances of theoretical fragment ion mass spectrometry (SWATH-MS). MS and statistical analysis identified 566 significantly differentially abundant (DA) proteins (adjusted p < 0.05), which were then analyzed for gene ontology and pathway enrichment. Our data indicated an up-regulation of steroidogenic proteins contributing to progesterone synthesis at luteal phase post-puberty. Proteins related to progesterone signaling, TGF-β, retinoic acid, extracellular matrix, cytoskeleton, and pleiotrophin signaling were DA in this study. The DA proteins probably relate to the formation and function of the corpus luteum, which is only present after ovulation, post-puberty. Some DA proteins might also be related to granulosa cells signaling, which regulates oocyte maturation or arrest in ovaries prior to ovulation. Ten DA proteins were coded by genes previously associated with reproductive traits according to the animal quantitative trait loci (QTL) database. In conclusion, the DA proteins and their pathways were related to ovarian activity in Bos indicus cattle. The genes that code for these proteins may explain some known QTLs and could be targeted in future genetic studies.


Animal Selection and Sampling
All experimental procedures were approved by the Animal Ethics Committee of The University of Queensland, Production and Companion Animal Group (certificate number QAAFI/279/12). Twenty Brahman heifers were sourced as weanlings (<250 kg) from two commercial herds in Queensland, Australia, and managed together at the Gatton Campus facilities of the University of Queensland.
Heifers' ovarian activity was observed using ultrasonography every fortnight for eight consecutive months (HS-2000(VET), Honda Electronics Inc.). When a CL was identified the heifer was deemed post-pubertal and paired randomly with a pre-pubertal heifer (no CL observed) for same-day euthanasia [3]. Euthanasia was planned for approximately 21 days after detection of the first CL, as described before [23]. On euthanasia day, we also measured the size of the largest follicle in each ovary and found that there was no statistical difference between pre-and post-pubertal heifers according to this criterium. The entire ovaries were harvested and preserved by snap freezing in liquid nitrogen and kept at −80 • C until protein extraction. We recorded which ovary had the CL for post-pubertal heifers. When thawed, the entire ovaries were pulverized and homogenized using a mortar and pestle and liquid nitrogen to produce a uniform sample for each ovary. Subsequently, aliquots of pulverized tissue of the left and right ovaries of the same heifer were mixed to produce a sample that would represent all ovarian tissue for proteomics. Plasma levels of progesterone were measured to confirm the luteal phase status post-puberty (average 2.0 ± 0.7 ng/mL). The average concentration of progesterone in pre-pubertal heifers was significantly lower (0.4 ± 0.2 ng/mL), as has been previously described [29].

Proteomics, Mass Spectrometry, and Data Analysis
Tissue samples from both ovaries were ground in liquid nitrogen so that fragments would represent the entire tissue. Subsequently, the buffer having 6M Guanidine chloride, 50 mM Tris pH 8, and 10 Mm DTT was added to lobind tubes containing samples. Samples were then sonicated at 35% for 10 s, and vortexed at 30 • C for 1 h in order to denature the proteins and break disulfide bonds. Then, 25 mM acrylamide was added to the samples and incubated at 30 • C for 1 h followed by the addition of 5 mM DTT. A small aliquot of the samples was precipitated by incubation at −20 • C for 16 h after adding four volumes of 1:1 methanol:acetone to one volume of sample. Afterwards, the proteins were precipitated from aliquot samples by centrifugation at 18,000 rcf for 10 min and the supernatant was discarded. The dried protein pellet was resuspended in 0.1% SDS. Protein concentration was measured using Nanodrop. Aliquots were further processed by filter-assisted sample preparation, adding 100 µg total protein to 10 kDa Cut-off Amicon columns which were then centrifuged at 18,000 rcf for 30 min. Ammonium bicarbonate buffer (50 mM) was added to the Amicon columns and centrifuged again using the same conditions as above. Ammonium bicarbonate buffer (50 mM) and trypsin were then added to the Amicon columns and incubated in a chamber of saturated humidity to prevent evaporation at 37 • C overnight. The Amicon columns were transferred to new tubes and centrifuged at 18,000 rcf for 30 min, followed by another centrifugation (same conditions) after addition of 0.5 M NaCl. The Amicon columns were discarded and the soluble peptides in the tubes were desalted using C18 ZipTips (Millipore) and resuspended in 9.25% acetonitrile and 0.1% formic acid. Desalted peptides were analyzed by liquid chromatography electrospray ionization tandem Genes 2019, 10, 923 4 of 27 mass spectrometry (LC-ESI-MS/MS) with a TripleTof 5600 instrument as previously described [33], except in this case a 45 min LC gradient was used to separate the peptides. Proteins were identified using information-dependent acquisition analysis of one randomly chosen pre-puberty sample and one randomly chosen post-puberty sample with Protein Pilot v5.0.1 (SCIEX) blasted against a database of proteins downloaded from Uniprot on 28 March 2016, with a total of 43,813 entries assigned to Bos taurus, including 6870 reviewed entries from Swiss-prot and 36,948 unreviewed entries from TrEMBL. Sequential window acquisition of all theoretical mass spectra (SWATH)-MS relative quantitative proteomics [34] data was analyzed with PeakView v2.1 (SCIEX). Statistical analyses were performed using MSstats in R as previously described [35,36] to identify differentially abundant (DA) proteins with a p value lower than 0.05 adjusted for multiple testing [37].

Functional Enrichment and Protein Interaction Analysis
Enrichment analyses for gene ontology (GO) terms and biological pathways were performed using Uniprot accession identifiers of DA proteins as target lists and the identifiers of all detected proteins as our experimental background list using Database for Annotation, Visualization, and Integrated Discovery (DAVID) [38]. Protein-to-protein interaction analysis of specific proteins was done using Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) [39].

Matching DA Proteins to Female Reproduction Quantitative Trait Loci (QTL)
The genes of DA proteins were also searched for their association with reproductive phenotypes according to the cattle quantitative trait loci database (Cattle QTLdb) [40]. Ovarian function is relevant to female reproduction in general and DA proteins from this study could be linked to CL function as well as puberty. Therefore, we performed an inclusive QTL analysis: we looked for all the female reproduction traits in the QTL database.

Results
Seven hundred and sixty-nine proteins were identified by LC-ESI-MS/MS detection and ProteinPilot database searching in the ovarian samples of both pre-and post-pubertal animals. Among the identified proteins, 566 were significantly DA when post-and pre-pubertal heifers were compared (Table S1). Among the DA proteins, 321 were up-regulated while 245 were down-regulated in post-pubertal heifers, as compared to pre-pubertal heifers ( Figure 1). A high proportion (73%) of identified proteins were DA.
Genes 2019, 10, x FOR PEER REVIEW 4 of 25 acetonitrile and 0.1% formic acid. Desalted peptides were analyzed by liquid chromatography electrospray ionization tandem mass spectrometry (LC-ESI-MS/MS) with a TripleTof 5600 instrument as previously described [33], except in this case a 45 min LC gradient was used to separate the peptides. Proteins were identified using information-dependent acquisition analysis of one randomly chosen pre-puberty sample and one randomly chosen post-puberty sample with Protein Pilot v5.0.1 (SCIEX) blasted against a database of proteins downloaded from Uniprot on 28 th March 2016, with a total of 43,813 entries assigned to Bos taurus, including 6870 reviewed entries from Swiss-prot and 36,948 unreviewed entries from TrEMBL. Sequential window acquisition of all theoretical mass spectra (SWATH)-MS relative quantitative proteomics [34] data was analyzed with PeakView v2.1 (SCIEX). Statistical analyses were performed using MSstats in R as previously described [35,36] to identify differentially abundant (DA) proteins with a p value lower than 0.05 adjusted for multiple testing [37].

Functional Enrichment and Protein Interaction Analysis
Enrichment analyses for gene ontology (GO) terms and biological pathways were performed using Uniprot accession identifiers of DA proteins as target lists and the identifiers of all detected proteins as our experimental background list using Database for Annotation, Visualization, and Integrated Discovery (DAVID) [38]. Protein-to-protein interaction analysis of specific proteins was done using Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) [39].

Matching DA Proteins to Female Reproduction Quantitative Trait Loci (QTL)
The genes of DA proteins were also searched for their association with reproductive phenotypes according to the cattle quantitative trait loci database (Cattle QTLdb) [40]. Ovarian function is relevant to female reproduction in general and DA proteins from this study could be linked to CL function as well as puberty. Therefore, we performed an inclusive QTL analysis: we looked for all the female reproduction traits in the QTL database.

Results
Seven hundred and sixty-nine proteins were identified by LC-ESI-MS/MS detection and ProteinPilot database searching in the ovarian samples of both pre-and post-pubertal animals. Among the identified proteins, 566 were significantly DA when post-and pre-pubertal heifers were compared (Table S1). Among the DA proteins, 321 were up-regulated while 245 were downregulated in post-pubertal heifers, as compared to pre-pubertal heifers ( Figure 1). A high proportion (73%) of identified proteins were DA.   In our enrichment analyses using DAVID, no pathways or GO terms were significantly overrepresented in the DA gene list. However, when we analyzed the up-and down-regulated proteins as separated target lists, eight biological pathways terms were enriched: three in the up-regulated list of DA proteins, affecting different metabolic pathways, and five in the down-regulated list of DA proteins, affecting extracellular signaling through extracellular matrix and focal adhesion pathways ( Table 1). The genes in each pathway are reported in the Supplementary Material (Table S2). The 566 DA proteins were classified into 26 functional clusters using DAVID ( Table 2 and Table S2). Functional clustering of DA proteins allowed the observation that most of proteins relevant to glycolysis, Tricarboxilic acid (TCA) cycle, pentose phosphate pathway, oxidative phosphorylation, cell redox homeostasis, and glutathione metabolism were up-regulated in post-pubertal heifers compared to pre-pubertal heifers. On the other hand, proteins directly or indirectly related to the extracellular matrix and focal adhesions were down-regulated in post-pubertal heifers ( Table 2).  We observed the functional annotation of DA proteins, their associated GO terms, and pathways to manually curate protein interactions. Manual annotation of DA proteins allowed grouping into known interaction pathways that provide an overview of the ovarian proteome in the context of puberty. We identified a group of DA proteins associated with steroidogenesis (progesterone synthesis) at the luteal phase in post-pubertal heifers ( Figure 2). The DA proteins related to progesterone synthesis were up-regulated in post-pubertal heifers compared to pre-pubertal heifers, which was exactly as expected since samples were collected at the luteal phase after puberty and a functioning CL was present. In this sense, up-regulated proteins post-puberty would be mostly from increased expression by CL cells. Meanwhile, the down-regulated proteins would correspond to proteins that had their expression reduced when granulosa cells were luteinized.
Another group of DA proteins was related to regulation of oocyte maturation/arrest. Clustering by DAVID included nine proteins in progesterone-mediated oocyte maturation, estrogen, and cyclic adenine monophosphate (cAMP) signaling ( Figure 3). The DA proteins from these clusters, when manually curated with evidence from the previous literature, seem involved in the regulation of oocyte maturation/arrest upon puberty in post-pubertal heifers. The DA proteins linked to oocyte maturation would be present in granulosa cells, which regulate and signal to promote oocyte maturation or arrest depending on the phase of the estrus cycle. Granulosa cells of the ovulatory follicle are luteinized and become CL cells post-puberty and at the luteal phase oocyte arrest is observed in the non-ovulatory follicles.
As an overall trend, the up-regulated proteins were involved in metabolic pathways and steroidogenesis while down-regulated proteins seemed to be involved in the regulation of CL function and oocyte maturation/arrest through the extracellular matrix and focal adhesion proteins in post-pubertal heifers at the luteal phase compared to pre-pubertal heifers (Table 3). Also, twenty-two DA proteins were identified as "uncharacterized" according to Uniprot (Table S1). synthesis) at the luteal phase in post-pubertal heifers ( Figure 2). The DA proteins related to progesterone synthesis were up-regulated in post-pubertal heifers compared to pre-pubertal heifers, which was exactly as expected since samples were collected at the luteal phase after puberty and a functioning CL was present. In this sense, up-regulated proteins post-puberty would be mostly from increased expression by CL cells. Meanwhile, the down-regulated proteins would correspond to proteins that had their expression reduced when granulosa cells were luteinized.  adenine monophosphate (cAMP) signaling ( Figure 3). The DA proteins from these clusters, when manually curated with evidence from the previous literature, seem involved in the regulation of oocyte maturation/arrest upon puberty in post-pubertal heifers. The DA proteins linked to oocyte maturation would be present in granulosa cells, which regulate and signal to promote oocyte maturation or arrest depending on the phase of the estrus cycle. Granulosa cells of the ovulatory follicle are luteinized and become CL cells post-puberty and at the luteal phase oocyte arrest is observed in the non-ovulatory follicles. As an overall trend, the up-regulated proteins were involved in metabolic pathways and steroidogenesis while down-regulated proteins seemed to be involved in the regulation of CL function and oocyte maturation/arrest through the extracellular matrix and focal adhesion proteins in post-pubertal heifers at the luteal phase compared to pre-pubertal heifers (Table 3). Also, twentytwo DA proteins were identified as "uncharacterized" according to Uniprot (Table S1).   The processes involved in regulation of ribosome and translation were up-regulated, except for some specific down-regulated proteins like eukaryotic translation initiation factor-4B (EIF4B). The DA proteins involved in repression of transcription, such as various histones, LIM, and cysteine-rich domains protein-1 (LMCD1), hepatoma-derived growth factor (HDGF), and heterochromatin protein 1 binding protein-3 (HP1BP3), were down-regulated. Nine SERPINs, extracellular matrix proteins, were identified as DA, out of which eight were down-regulated in post-pubertal heifers compared to pre-pubertal heifers. Three of these proteins, SERPINE2, SERPINA3, and SERPINA3-3, were among the top twenty down-regulated proteins in our dataset (Table 3). This down-regulation of extracellular matrix (ECM) proteins may also be linked to the formation and function of the CL cells (see the discussion).
Analysis by STRING revealed interactions between proteins related to pleiotrophin signaling, cytoskeleton arrangement, focal adhesion, and spindle fiber arrangement (Figure 4). It is important to mention that pleiotrophin was a highly up-regulated protein in post-pubertal heifers at the luteal phase compared to pre-pubertal heifers. Pleiotrophin seems to play an important role while interacting with focal adhesion, cytoskeleton, and spindle assembly proteins during the onset of puberty and the establishment of progesterone signaling.  Nomenclature of proteins is given in Table 1.
Using Uniprot ID information we were able to annotate the coding genes for all the DA proteins. We investigated the location of these genes to discover if they could be linked to any reproductive QTL. Ten DA proteins were coded by genes associated with reproductive traits according to QTL position (Table 4).   Table 1.
Using Uniprot ID information we were able to annotate the coding genes for all the DA proteins. We investigated the location of these genes to discover if they could be linked to any reproductive QTL. Ten DA proteins were coded by genes associated with reproductive traits according to QTL position (Table 4). We also compared our dataset of DA proteins with the list of differentially expressed (DE) genes for the same ovarian samples from post-versus pre-pubertal heifers which was previously reported by our group [29]. Comparative analysis resulted in 126 features in common, i.e., DA proteins that were also DE genes in the previous RNA-sequencing analyses. We performed correlation and regression analysis for these 126 proteins and genes. The correlation coefficient (r) between expression of genes and abundance of respective proteins was positive: r = 0.54. In regression analysis, we regressed the fold change in differential expression of genes (independent variable) against the fold change in differential abundance of their respective proteins (dependent variable). Regression coefficient was a positive value of 0.61 while the coefficient of determination (R 2 ) was calculated as 0.28 with p value < 0.0001 ( Figure 5). In short, the abundance level of DA proteins was correlated to the expression of DE genes.
Genes 2019, 10, x FOR PEER REVIEW 11 of 25 by our group [29]. Comparative analysis resulted in 126 features in common, i.e., DA proteins that were also DE genes in the previous RNA-sequencing analyses. We performed correlation and regression analysis for these 126 proteins and genes. The correlation coefficient (r) between expression of genes and abundance of respective proteins was positive: r = 0.54. In regression analysis, we regressed the fold change in differential expression of genes (independent variable) against the fold change in differential abundance of their respective proteins (dependent variable). Regression coefficient was a positive value of 0.61 while the coefficient of determination (R 2 ) was calculated as 0.28 with p value < 0.0001 ( Figure 5). In short, the abundance level of DA proteins was correlated to the expression of DE genes.

Discussion
This study aimed to investigate the genes and pathways involved in ovarian progesterone production-a hallmark of CL function-at the onset of puberty. To this aim, homogenized ovarian samples that represent the entirety of both ovaries were used for proteomics. Ovarian granulosa and theca cells are crucial for steroid synthesis and oocyte maturation, and so their activity is relevant to

Discussion
This study aimed to investigate the genes and pathways involved in ovarian progesterone production-a hallmark of CL function-at the onset of puberty. To this aim, homogenized ovarian samples that represent the entirety of both ovaries were used for proteomics. Ovarian granulosa and theca cells are crucial for steroid synthesis and oocyte maturation, and so their activity is relevant to maintenance of subsequent pregnancy through progesterone signaling. After puberty (first ovulation), a CL is formed and it produces progesterone, which is necessary to maintain early pregnancy, and hence contributes to fertility [86,87]. The presence of the CL alters ovarian gene expression [88]. The identified DA proteins in post-pubertal heifers at the luteal phase compared to pre-pubertal heifers indicate their importance to CL function, as the presence of the CL cells represents the main difference between the two physiological stages. This DA proteins might have a role in puberty and subsequent fertility as they were linked in this work to essential ovarian activities. In the following sections, DA proteins are grouped and discussed according to their ovarian function.

Metabolic Pathways
Up-regulated DA proteins were found to be related to metabolic pathways such as glycolysis, oxidative phosphorylation, and beta-oxidation. Up-regulation of glycolysis and genes which involve glycolysis and steroidogenesis have been reported in the oocytes of post-pubertal cows when compared to pre-pubertal cows [29,89]. Oxidative phosphorylation was an enriched pathway in the list of up-regulated proteins in our results. Up-regulation of oxidative phosphorylation affects steroidogenesis through steroidogenesis acute regulatory protein (StAR) [90]. Stimulation of beta-oxidation of fatty acids for oocyte maturation through protein kinase-A signaling has also been reported in mice [91]. Up-regulation of proteins is linked to metabolic pathways in post-pubertal heifers at the luteal phase compared to pre-pubertal heifers and evidence from the cited literature suggests that these pathways play a role in progesterone synthesis. Progesterone signaling contributes to pubertal development according to the Gonodostattheory and the observation that a first short cycle is often necessary to the establishment of normal cycling in bovine species [14]. Therefore, and in the context of this study, DA proteins related to metabolic pathways might contribute to both CL function and pubertal development.

Transcription and Translation
We identified DA proteins that regulate transcription. Regulatory patterns of gene transcription and translation during oocyte growth, oocyte maturation, and luteinization of granulosa cells are timing specific [92,93]. Histones play a crucial role in transcriptional regulation [68]. Histone H3.3 was found to be one of the most down-regulated proteins post-puberty (fold change = −1.46). Specific methylation of H3 is associated with repression of α-inhibin, which is important for CL function and progesterone synthesis [58]. H3.3 and H4 interfere with meiosis and their depletion results in primary oocyte death and zygotic and early embryonic deaths [68]. Macro-H2A, HDGF, HP1BP3, and LMCD1 were down-regulated in post-pubertal heifers at the luteal phase in our data. These proteins are involved in the repression of transcription [94][95][96][97]. Down-regulation of specific histones and transcription repression proteins indicate enhanced transcription in post-pubertal heifers at the luteal stage compared to pre-pubertal heifers. These regulatory DA proteins probably have a role in progesterone synthesis by CL cells. These same proteins might be involved in meiotic oocyte maturation/arrest in pre-ovulatory follicles in coordination with progesterone signaling.
Increased translation and ribosomal activity, which is related to steroidogenic synthesis after LH stimulation, was observed in our data and has been seen in a previous work [98]. Some ribosomal proteins were DA in this study namely, RPL30, RPL27a, RPS8, RPS12, and RPLPO, and have also been found to be differentially expressed in mature buffalo oocytes compared to pre-pubertal oocytes [99]. Future work could use ovarian dissection to isolate granulosa cells and oocytes to investigate which cell type contributes to the significant difference observed for DA proteins. Overall, DA proteins associated with regulation of transcription and translation were found to be up-regulated, which could be explained by very active CL cells compared to the relatively less active cells in pre-pubertal ovaries.
All translation-and ribosome-related DA proteins were up-regulated in our data except for the eukaryotic initiation factor EIF4B. Translation of maternal mRNAs is regulated by secondary structures in UTRs [93,100]. EIF4B was down-regulated (FC = −1.24) in post-pubertal heifers. EIF4B stimulates EIF4A to unwind secondary structures from maternal mRNAs to enable attachment of ribosomes for translation initiation during oocyte maturation [60,101]. Time-specific translation of maternal mRNAs is necessary for oocyte maturation [60]. At luteal phase post-puberty, where no oocyte maturation is expected, the selective down-regulation of EIF4B indicates that this protein might regulates translation that is specific to oocyte maturation. Once more, isolation of oocyte gene expression would be important to confirm this idea.

Complement System
The list of down-regulated proteins post-puberty was enriched for the 'Complement and coagulation cascade pathway'. Maintenance of mature CL is dependent on lowering the immune system in CL cells [102]. Down-regulation of complement factors, which are part of the immune system, at the luteal phase in post-pubertal heifers is indicative of a mature CL being present and functional. Progesterone measurements also confirmed that post-pubertal heifers had functioning CLs in our dataset. Complement component C3 has also been identified as positively associated with oocyte maturation in pigs [103]. Another study has identified complement components C3, C4, C7, C8, C9, C-H, and C-I in human follicular fluid and associated these with oocyte maturation [104]. Differential abundance of complement components in post-pubertal heifers at the luteal phase compared to pre-pubertal heifers in our study supports the role of these proteins in CL maintenance and pubertal development. Both phenomena would have implications for subsequent fertility.

Extracellular Matrix Proteins
Extracellular matrix and focal adhesion proteins are important in mediating paracrine signaling in ovaries for follicular development and CL functions [105,106]. Fibronectin and asporin are examples of ECM proteins that were DA. Fibronectin was increased in abundance post-puberty, at the luteal phase. Increased fibronectin is indicative of a functioning CL [107]. Asporin (ASPN) was the most down-regulated protein in post-pubertal heifers at the luteal phase compared to pre-pubertal heifers. ASPN in theca cells is associated with growth of secondary follicles in the gonadotrophins independent stage [57]. The up-regulation of fibronectin and the down-regulation of asporin both tell a story of a functioning CL and reduced follicular growth, which could be expected by the presence of CL cells at the luteal phase which were not present before.
Some ECM proteins, such as SERPINs and Tenascin-X (TNXB), might be involved in oocyte maturation. Nine SERPIN proteins, including SERPINE2, were down-regulated in post-pubertal heifers compared to pre-pubertal heifers. ECM proteins, including SERPINs and ITIH1, are found to be specific to oocyte maturation and SERPINE2 has been proposed as a biomarker for oocyte maturation [108,109]. Tenascin-X (TNXB) was among the top down-regulated proteins (FC = −1.12) in post-pubertal heifers in our results. Transcription of the TNXB gene is increased in pre-ovulatory follicles and pregnancy competent oocytes [64,110]. As we did not separate specific cell types in our study, the role of these nine SERPINs and TNXB in oocyte maturation requires validation.
Agrin (AGRN), lamin-A (LMNA), lamin-B (LMNB), laminin, clusterin (CLU1) and vimentin (VIM) are also ECM proteins that were down-regulated in our data. Remodeling of extracellular matrix proteins and adhesion proteins has also been reported in the leutinization of cells that form the CL. Proteins such as integrin-vα and laminin are associated with leutinization [111]. Lamins and laminins are part of the basal lamina that separates granulosa cells from theca cells in pre-ovulatory follicles. They are important ECM proteins that mediate the interactions between granulosa cells, the thecal layer, and the oocyte. This basal lamina degrades during CL formation [107]. Decreased abundance of lamins in our study agrees with the presence of functioning CL cells post-puberty, a major contrast to the pre-pubertal samples. Lamins are increased in expression during meiotic oocyte maturation [112]. Laminin A and B have increased expression levels associated with oocyte maturation in mice [113]. The reverse, down-regulation, is expected in the luteal phase. Similarly, CLU1 was down-regulated in our data, which could be expected for the luteal phase since increased levels of CLU1 have been related to oocyte maturation in bovine previously [56]. Differential abundance of ECM proteins in post-pubertal as compared to pre-pubertal heifers probably relates to CL formation and function, with implications for pubertal development and subsequent fertility.

Steroidogenesis
The up-regulation of steroidogenic proteins in our study can probably be explained by the presence of functioning CL cells post-puberty which were absent pre-puberty. Steroidogenesis starts with the precursor acetyl-CoA, which leads to synthesis of cholesterol and finally steroid hormones. We observed 32 DA proteins between pre-and post-pubertal Brahman heifers that are relevant to steroidogenesis. Highlighting these DA proteins, we were able to annotate the steroidogenesis pathway in Bos indicus animals.
Acetyl-CoA results from different metabolic pathways, including glycolysis. Nine up-regulated proteins from our results play a role in glycolysis and result in pyruvate synthesis [114], which produces acetyl-CoA in a subsequent reaction [115]. Acetyl-CoA is converted to citrate by citrate synthase (CS). Citrate can enter the tricarboxylic acid cycle or be transported to cytosol by citrate transport protein (CTP). Citrate is cleaved and converted into acetyl-CoA and oxaloacetate by ATP citrate lyase (ACL) [49]. Acetyl-CoA can be used to synthesize fatty acids or cholesterol or to enter into thetricaboxylic acid cycle [115,116]. The above-mentioned proteins CS, CTP, and ACL were up-regulated in our data. Up-regulation of proteins in glycolysis and subsequent steps indicates their important role in contributing acetyl-CoA as a precursor for up-regulation of steroidogenesis at puberty.
The apolipoproteins A-I, A-II, and C-III were down-regulated post-puberty. Apolipoproteins sustain cholesterol homeostasis in cells, removing it when it becomes excessive to avoid cellular cholesterol toxicity [120][121][122]. Up-regulation of enzymes that contribute to cholesterol synthesis and down-regulation of cholesterol efflux proteins together indicates increased supply of cholesterol for steroidogenesis post-puberty.
Cholesterol side-chain cleavage enzyme (CYP11A1), ferredoxin (FDX), ferredoxin reductase (FDXR), and 3-beta-hydroxysteroid dehydrogenase (HSD3B1) were up-regulated while steroid 17-alpha-hydroxylase (CYP17A1) was down-regulated in our results. Conversion of cholesterol to pregnenolon is mediated by CYP11A1. FDXR and FDX help in electron transfer for this reaction. Pregnenolon is converted to progesterone by 3-beta-hydroxysteroid dehydrogenase (HSD3B1) [42,43]. CYP17A1 converts progesterone to androstenedione [123], which in further reactions is converted into estrogen [124]. Up-regulation of progesterone synthesis proteins and down-regulation of estrogen-related proteins indicate high levels of progesterone production, which is typical of the luteal phase.
The process of steroidogenesis is the same for both progesterone and estrogen up to the synthesis of cholesterol. The up-regulation of this process up to cholesterol synthesis may be important for both the estrogenic and the luteal phase post-puberty. To confirm the overall relevance of these DA proteins to the onset of puberty, future research could compare pre-pubertal heifers to post-pubertal heifers in all phases of the estrus cycle.

Progesterone Signaling Regulating Puberty and Fertility
Progesterone and its membrane receptors are involved in oocyte maturation in the pre-ovulatory follicle, in fertilization processes, and in the maintenance and development of the pregnancy [125][126][127]. These phenomena, which are affected by progesterone signaling, contribute to overall female fertility. Progesterone signaling is complex and some of the same pathways are relevant to mature CL cells and to oocyte maturation. The DA proteins G inhibitory protein GNAI2, adenylyl cyclase associated protein (CAP-1), adenylate kinase (ADK), and protein kinase A regulatory subunit II α and β (PRKAR2α and PRKAR2β) all contribute to activate the transcription factor cAMP responsive regulatory element binding protein (CREB). In turn, CREB regulates the transcription of proteins related to CL functioning [128,129]. The fact that these proteins are DA herein might be explained by the presence of the CL being unique to post-pubertal heifers.
Some of these DA proteins are also known for their roles in progesterone signaling related to oocyte maturation: GNAI2, CAP-1, and ADK are, for example. GNAI2 was down-regulated whileCAP-1 and ADK were up-regulated in post-pubertal heifers compared to pre-pubertal heifers. These proteins are involved in progesterone-mediated oocyte maturation/arrest. Progesterone through membrane progesterone receptors activates GNAI2, which inhibits adenylyl cyclase (AC) to suppress levels of cAMP in oocytes [24,[130][131][132][133][134]. Similarly, CAP-1 and ADK are also involved in the regulation of cAMP levels [135,136]. Suppression of cAMP relieves oocytes from meiotic arrest and hence causes oocyte maturation [130,132,133,137,138]. PRKAR2α and PRKAR2β were also down-regulated in post-pubertal heifers compared to pre-pubertal heifers. Interestingly, increased levels of PRKAR2α and PRKAR2β are associated with progesterone-, mediated oocyte maturation [24,131].
Granulosa cells contribute to oocyte maturation at the pre-ovulatory stage and then contribute to the maintenance of pregnancy after being transformed into luteal cells. Both of these phenomena are associated with progesterone signaling at two different stages of the estrous cycle. Low levels of progesterone (and other factors) produced by granulosa cells facilitate follicular growth and oocyte maturation. High levels of progesterone produced by the CL might contribute to oocyte arrest during the luteal phase. The differential abundance of these progesterone-signaling proteins between ovarian tissues of pre-pubertal heifers and post-pubertal heifers at the luteal phase indicates that these proteins are probably differentially regulated in luteinized granulosa cells in the CL compared to non-luteinized granulosa cells.

Retinoic Acid Signaling
Our results revealed increased abundance of retinal dehydrogenase (RALHD-1), protein phosphatase-2A (PPP2R1A), and cellular retinoic acid binding protein (CRABP1). These proteins are relevant to retinoic acid (RA) signaling. Retinoic acid is involved in stimulation of progesterone synthesis in the CL [142]. Also, the direct genomic action of RA signaling or its regulation of other effector pathways aids oocyte maturation [143,144]. CRABP1 protein is differentially expressed in different stages of bovine oocyte maturation [56]. It inhibits RA activity and also stimulates PPP2R1A via ERK1/2 kinases to delay the cell cycle [144,145]. Cell cycle regulation is probably also important to the formation of the CL via cellular differentiation and proliferation processes, as discussed above.

Pleiotrophin and Progesterone Sinaling on Focal Adhesion, Cytoskeleton, and Microtubules
We suggest a putative role of pleiotrophin in coordination with focal adhesion and cytoskeleton proteins at the luteal phase in post-pubertal heifers. Cytoskeleton and microtubules play a role in progesterone synthesis and oocyte maturation [155][156][157][158][159]. Pleiotrophin, ITGAv (an integrin), and ADD-1 were up-regulated, while filamin (FLNA) and actinin (ACTN1) were down-regulated in post-pubertal heifers at the luteal phase. Pleiotrophin signaling through ITGAvβ3 (heterodimer of ITGAv and ITGAβ) has been reported [160]. ITGAvβ3is involved in phosphorylation of cytoskeleton proteins and focal adhesion proteins, including FLNA and ACTN1 [161,162]. Signaling through integrins down-regulates cytoskeleton structures in cells [162]. Adducin heterodimers, composed of β-adducin (ADD-2) and α-adducin (ADD-1), regulate the actin-spectrin complex in the cytoskeleton [163]. Phosphorylation of ADD-2 by pleiotrophin signaling results in cytoskeletal dissociation [164]. Tubulin-α, tubulin-β subunits, microtubules-associated protein (MAP), dynein light and heavy chains (DYN), and dynactin (DCTN-2) were up-regulated in post-pubertal heifers compared to pre-pubertal heifers in our results. These proteins are involved in microtubule and spindle fiber assembly and function [165]. Dissociation of the cytoskeleton and microtubules is associated with more progesterone synthesis in granulosa cells [159]. Nuclear mitotic apparatus-1 (NuMA-1) was down-regulated in post-pubertal heifers. NuMA organizes microtubules at spindle poles in coordination with DYN and DCTN [166] and plays an important role in meiotic oocyte maturation by organizing tubulins [167]. Similarly, specific hyper-phosphorylation of ADD-1 causes abnormal assembly of spindle fibers in meiosis and mitosis [168,169]. Mitosis-associated proteins are very important in the context of forming the CL since the ceasing of cell division is required for cell differentiation and growth processes that transform granulosa cells into luteal cells [16]. Progesterone, through its membrane receptors, is involved in spindle fiber assembly arrangement at different stages of meiosis during oocyte maturation [126] and during mitosis of granulosa cells [170,171]. Differential abundance of proteins related to the cytoskeleton, spindle fibers, and cell division, corroborated by the cited literature, are evidence for the involvement of these proteins in the formation of the CL.

Association of DA Proteins with Reproductive Traits
Genes of 10 DA proteins from our study were associated with important reproductive traits according to the animal QTL database [40]. The DA protein ezrin (EZR) was associated with "length of reproductive life" and "calving ease" in cattle [79]. Functionally, ezrin is involved in cytoskeleton arrangement [172]. T-complex protein component-1 (TCP-1) and T-complex protein component-8 (CCT-8) were DA in our data and were associated with "conception rate" in cattle [82]. Both of these chaperone proteins help in folding actin and tubulin in cytoskeleton and spindle formation [173,174]. Calpain-1 (CAPN1) was another DA protein in our data that has been previously associated with two reproductive traits, namely "post-partum anestrous interval" and "IGF-1 levels" [83,175]. CAPN1 seems to play a role in cumulus cell-oocyte complex expansion and oocyte maturation [176] and from our data it seems it might also be involved in CL function. Leucine aminopeptidase (LAP3) is a DA protein associated with "calving ease" in cattle [84]. LAP3 expression increases in theca cells of mature pre-ovulatory oocytes in humans in the presence of high levels of progesterone. It has also been reported to play a role in oocyte maturation in Haemaphysalislongicornis tick [177,178]. FST, another DA protein, has been associated with "conception rate" [81]. Increased FST mRNA levels have been observed in mature oocytes, while decreased levels have been associated with oocyte arrest in cattle [179]. Serpine-2 is associated with fertilization rates in cattle [85]. It has been found to be differentially expressed in the granulosa cells of bovine ovaries [66]. Association of these DA proteins with reproductive trait QTLs seems to support their involvement with female fertility, which could be expected for proteins that are involved in the formation and function of the CL.

Conclusions
This study has provided a set of ovarian proteins that were found to have altered abundance levels as a result of CL activity, progesterone signaling, and puberty in Bos indicus heifers. Selective regulation of proteins involved in various metabolic pathways indicated their role in up-regulating overall steroidogenesis upon puberty. The reported proteins have enhanced our current understanding of pathways that are regulated in CL cells and therefore contribute to progesterone signaling, with implications for oocyte maturation, puberty, and subsequent fertility. Some DA proteins are coded by genes in QTLs for female reproductive traits and can be further scrutinized in the search for causative mutations.