Differential Gene Expression in Uterine Endometrium During

0 downloads 0 Views 1MB Size Report
Dec 17, 2014 - endometrium, gene expression, litter size, microarray, QTL. INTRODUCTION ... GSE51970. 2Correspondence: E-mail: agwang@cau.edu.cn .... functional annotation chart and functional annotation clustering tools of .... Hardy-Weinberg equilibrium (HWE) in the studied herds was tested by comparing ...
BIOLOGY OF REPRODUCTION (2015) 92(2):52, 1–14 Published online before print 17 December 2014. DOI 10.1095/biolreprod.114.123075

Differential Gene Expression in Uterine Endometrium During Implantation in Pigs1 Xiao Chen,4 Aiyun Li,5 Wencheng Chen,4 Julong Wei,4 Jinluan Fu,3,4 and Aiguo Wang2,4 4

College of Animal Sciences and Technology, National Engineering Laboratory for Animal Breeding & Key Laboratory of Animal Genetics, Breeding and Reproduction, Ministry of Agriculture, China Agricultural University, Beijing, China 5 Ankang Yangchen Animal Husbandry Technology Co., Ltd., Ankang, China

Embryonic mortality during the implantation period strongly affects litter size in pigs. To analyze the differentially expressed genes (DEGs) in the endometrium during implantation and further to identify candidate genes for litter size, tissues of endometrial attachment sites and intersites were collected from nine pregnant sows on Days 13, 18, and 24 of pregnancy. Endometrium tissue was also collected from another three nonpregnant sows. Samples were hybridized to the porcine Agilent GeneChip microarray. The analysis of gene expression patterns over the implantation period revealed 858 DEGs at endometrial attachment sites. Comparisons of the gene files of attachment sites and intersites revealed 12, 51, and 89 DEGs on Days 13, 18, and 24 of pregnancy, respectively. Annotated function was used to identify overrepresented genetic processes, and several biological processes were considered as the most enriched. Genes related to vascular development, proteolysis, RNA metabolism and translation, protein modification, immune response, and hormone-related are discussed in detail. Then we combined microarray technology and linkage analysis to select powerful candidate genes for quantitative trait loci affecting pig litter size. Eighty-seven DEGs were located in quantitative trait loci related to litter size, that is, total number born and number born alive. Those candidate genes were thought to affect litter size by influencing embryonic implantation. Furthermore, single nucleotide polymorphism of VEGFA was shown to be associated with litter size in pigs. This study identified candidate genes for litter size that were related to embryonic implantation and could be a resource for target studies of genetic markers for litter size in pigs. endometrium, gene expression, litter size, microarray, QTL

INTRODUCTION Most reproductive traits are complex in terms of their genetic architecture [1]. As one of the most important economic traits in pig production, litter size is a quantitative trait with low heritability (h2, 0.1–0.15) [2]. In addition, the expression of litter size cannot be measured until the age of 1

Supported by National Natural Science Foundation of China (No.31172176), China Agriculture Research System (No. CARS-36), Program for Changjiang Scholar and Innovation Research Team in University (IRT1191) and National Scientific and Technical Supporting Programs of China (No. 2011BAD28B01). All the microarray data are available through the GEO database with accession number GSE51970. 2 Correspondence: E-mail: [email protected] 3 Correspondence: E-mail: [email protected] Received: 9 July 2014. First decision: 2 August 2014. Accepted: 5 December 2014. Ó 2015 by the Society for the Study of Reproduction, Inc. eISSN: 1529-7268 http://www.biolreprod.org ISSN: 0006-3363

1

Article 52

Downloaded from www.biolreprod.org.

sexual maturity. However, these biological constraints can be potentially ameliorated by a better knowledge of the genetic regulation of litter size [3]. With the development of high-throughput technologies for large-scale expression studies, the analysis of gene expression on a global scale has become feasible. Microarray technology allows us to identify genes involved in a particular biological process and cluster them based on closely related expression profiles [4]. Litter size is affected by the ovulation rate, fertilization rate, embryonic mortality, uterus volume, and placental efficiency. Increasing the number of ovulated follicles is the basis of improved litter size. The ovulation rate responds to direct selection in pigs, but the returns in terms of improved litter size have been minimal [5, 6]. Moreover, the ovulation rate is negatively correlated with embryonic mortality (0.56) [7]. The rate of fertilization is generally greater than 95%. Therefore, the loss of potential piglets is predominately the result of early embryonic (Days 10–30) death and fetal (Days 31–70) death [8]. Porcine embryos begin to attach to the endometrium on Days 13 and 14 of gestation, and attachment is completed on Days 18 to 24 of gestation [9]. Implantation is a sensitive period for embryonic mortality. During the firstphase of implantation, embryonic mortality can reach 22%, representing the major period of embryonic mortality [10]. Embryonic mortality is 17% on Day 18 of gestation, and up to 30% on Day 25 of gestation [11, 12]. Meishan pigs demonstrate high prolificacy. On Day 30 of gestation, embryonic mortality of Meishan pigs was lower (21%) than that of Large Whites (45%) (P , 0.01) [13]. For the above reasons, we have chosen the 13th, 18th, and 24th day of pregnancy to analyze gene expression patterns at different stages of implantation. We expect that these differences in gene expression patterns to help identify candidate genes that contribute to differences in litter size. There have been some expression studies carried out in pigs looking at changes in endometria [14–18], ovaries [3], and placentas [19], and genes differently expressed at one stage of pregnancy have been identified. Furthermore, important biological processes and pathways were identified. However, successive gene expression patterns over the implantation period are poorly understood. In this study, we examine the gene expression profiles in porcine endometria on Days 13, 18, and 24 of gestation using cDNA arrays and identify genes that are differentially expressed on these three days. Furthermore, we combine microarray technology and linkage analysis to select powerful candidate genes for quantitative trait loci (QTLs) that could affect pig litter size. Those candidate genes are thought to affect litter size by influencing embryonic implantation. Next, association analysis between genotypes and phenotypes (litter size) in large populations can reveal how those genes would be used as genetic markers to improve litter size in pigs.

ABSTRACT

CHEN ET AL.

MATERIALS AND METHODS

used to measure the degree of similarity between the expression profiles of the samples [21]. All the microarray data are available through the GEO database with accession number GSE51970. Functional annotation of the obtained DEGs. Enrichment analysis was carried out using the database for annotation, visualization, and integrated discovery (DAVID) [22, 23]. However, not all DEGs were used in the functional annotation because some of them lacked sufficient biological annotation. The pig functional annotation database is not as complete as the human functional annotation database; generally, only about 33% of pig gene identifications (IDs) could be annotated. Comparably, human functional annotation database allows more than 80% of gene IDs to be functionally annotated. Therefore, we converted the pig gene IDs (National Center for Biotechnology Information [NCBI] database Sus scrofa 10.2) into human gene IDs using the Ensembl BioMart program (http://asia.ensembl.org/index.html) by homology screening. Then the homologous human Ensembl IDs were submitted to DAVID database for functional annotation. Integrated analysis of the different functional database was done using the functional annotation chart and functional annotation clustering tools of DAVID database. In DAVID analysis, GO_BP_FAT (gene ontology_biological processes_functional annotation terms) terms and KEGG (Kyoto Encyclopedia of Genes and Genomes) pathways were selected [24]. Functional annotation was also used to illustrate the representation of functional categories in individual SOM expression clusters.

Animal Materials The Animal Care and Use Committee of China Agricultural University reviewed and approved the experimental protocol used in this study (Code: SYXK [Jing] 2009-0030). Multiparous Large White sows (fifth parity) were observed daily for standing heat in the presence of a boar. The sows of the pregnant group were inseminated twice, 12 and 24 h after heat detection [14]. The sows of the nonpregnant group were treated with inactivated sperm from the same boar [14]. Nine pregnant sows were slaughtered by electrocution on Days 13, 18, and 24 after insemination (the pregnant group, three sows every period). The nonpregnant sows were slaughtered on Day 13 after insemination. For the pregnant sows, samples of the endometrium attachment sites and intersites were taken. Samples were taken from three locations of each uterine horn: proximal (the end, close to the ovaries), medial, and distal (next to the corpus uteri) [14]. Samples from the endometrium of the nonpregnant sows were taken from comparable locations. Endometrial tissue sampling was carried out according to the procedure of Lord et al. [20] with minor modifications. All the samples were collected immediately, snap frozen in liquid nitrogen, and stored at 808C. Animals used to identify candidate genes for litter size were from Beijing Huadu Swine Breeding Company LTD. Rearing and feeding conditions were the same for all the sows. Ear tissue samples of 574 Large White, Landrace, and Duroc sows were collected in centrifuge tubes (1.5 ml) with 70% ethanol and stored at 48C until DNA extraction. DNA was extracted by phenol:chloroform (1:1) extraction. There were eight sire families in Large White, eight sire families in Landrace, and seven sire families in Duroc. A total of 1211 litter records were used for statistical analysis. Litter size records such as the total number born (TNB) and number born alive (NBA) were recorded by parity.

Real-Time Quantitative PCR

RNA Isolation Trizol reagent (Invitrogen) was used to extract total RNA, according to the manufacturer’s instructions. The quality of the samples was assessed using an Agilent Bioanalyzer 2100 (Agilent Technologies). The RNA integrity number values obtained were in the range of 7.0–9.0, which assured the homogeneity and high quality of the samples. For each animal, total RNA consisted of a mix of an equal quantity of total RNA from three locations of each uterine horn: proximal (the end, close to the ovaries), medial, and distal (next to the corpus uteri).

Agilent Porcine Microarray Study The RNA was further purified using RNeasy mini kit (74106; Qiagen), RNeasy micro kit (74004; Qiagen), and RNase-Free DNase Set (79254; Qiagen) according to the manufacturer’s instructions. The RNA was amplified and labeled using Low Input Quik Amp Labeling Kit, One-Color (5190-2305, Agilent technologies). Labeled cRNA were purified by RNeasy mini kit (74106; Qiagen). The RNA was then hybridized to the Agilent Whole Porcine Genome Microarray (4344K, 026440; Agilent) using an Agilent Gene Expression Hybridization Kit (5188–5242; Agilent) as recommended by the manufacturer. Following microarray hybridization and washing, the processed slides were scanned with an Agilent DNA microarray scanner (G2565CA; Agilent).

Chromosomal Localization of DEGs in QTLs for Litter Size The localization of the DEGs (FC . 2, P , 0.05) on porcine chromosomes was accessed from the NCBI database. Each gene location was estimated in centimorgans and was compared with the location of significant QTLs for TNB and NBA (http://www.animalgenome.org/cgi-bin/QTLdb/SS/ index) [25–28]. The QTL was extended by 5 Mb on either side. Genes located within QTL confidence intervals have been accepted as candidate genes for litter size.

Microarray Data Analysis and Statistical Analysis

Detection of Single Nucleotide Polymorphisms of the DEG VEGFA and Association Analysis

The raw gene expression data were extracted from Agilent Feature Extraction Software (version 10.7) and imported into Agilent GeneSpring GX software (version 11.0) for further analysis. The microarray data from the 21 samples were normalized in GeneSpring GX using the Agilent Feature Extraction one-color scenario. All of the data were interpreted using the logratio setting. Two independent analyses were applied to identify genes that were significantly different. First, the GeneSpring pairwise comparison (Welch t-test with ANOVA, P , 0.05) was conducted between all the possible stages. The criteria for significance of the differentially expressed genes (DEGs) was established as a fold change (FC) . 2 with an adjusted P , 0.05. Second, a one-way ANOVA for microarrays (default parameters: false discovery rate [FDR] , 0.1, P , 0.05) was also conducted to analyze gene expression patterns at attachment sites or intersites over implantation time. Clustering of gene expression profiles over implantation time was also performed with the self-organizing maps (SOMs) based on the result of one-way ANOVA (FDR , 0.1, P , 0.05). To compare gene expression profiles between different implantation times and between attachment sites and intersites, principalcomponent analysis (PCA) and unsupervised hierarchical clustering was performed. In unsupervised hierarchical clustering, the Euclidean distance was

VEGFA was selected to be the candidate gene for litter size based on its biological functions on embryonic implantation and its positions within the intervals of QTLs. Polymerase chain reaction single-strand conformation polymorphism (PCR-SSCP) was used to identify its mutations. In PCR-SSCP analysis, the target sequence in genomic DNA or cDNA is simultaneously amplified and labeled by using radioactive-labeled primers or nucleotides. The amplified product is then denatured to a single-strand form and subjected to nondenaturing polyacrylamide gel electrophoresis. Bands of the single-stranded DNA at different positions in the autoradiogram indicate the presence of mutations [29]. Genomic DNA was extracted by the standard phenol:chloroform (1:1) extraction method. The primer sequences used for PCR were shown in Supplemental Table S2. The genomic DNA samples were used as templates to individually and specifically amplify these regions. Amplifications were carried out on an Eppendorf Mastercycler gradient 5331 PCR System (Eppendorf). The PCR amplification was performed using 50–100 ng of genomic DNA and 12.5 ll of 23Taq PCR MasterMix (0.05 units/ll Taq DNA polymerase; 4 mM/ll MgCl2; and 0.4 mM/ll dNTPs) for 5 pM of each primer

2

Article 52

Downloaded from www.biolreprod.org.

To confirm the microarray results, the expression of eight genes was validated by RT-PCR using the same 21 porcine endometrium samples used for the microarrays. For each sample, first-strand cDNA was synthesized using 1 lg of total RNA. An M-MLV FIRST STRAND KIT (Invitrogen) and an oligo (dT)18 primer were used in a reverse transcription reaction of 20 ll, following the supplier’s instructions (see Supplemental Table S1; Supplemental Data are available online at www.biolreprod.org). All the PCR fragments were analyzed on an AlphaImager system to check amplification of a single product. The amplifications were sequenced to verify the obtained PCR product. Transcript quantification was performed using SYBR Green mix (Roche Diagnostics GmbH, Roche Applied Science) in a Roche LightCylcer 480. PCR efficiency of each gene was estimated by standard curve calculation using four points of cDNA serial dilutions. Ct values were quantified using the comparative Ct method, setting the relative quantities of nonpregnant group for each gene to 1 (quantity ¼ 10DCt/slope). Data normalization was carried out using GAPDH as the reference gene. Data were analyzed with SAS 8.02 software. ANOVA was used to compare the time sequence events in endometrium attachment sites and intersites. Then the Welch t-test was used to compare differences between attachment sites and intersites. The correlation between the result of microarray and RT-PCR was calculated using a correlation test [3].

GENE EXPRESSION IN UTERINE TISSUE OF PREGNANT SOWS

in a 25 ll final volume (the reagents all came from the National Laboratories for Agrobiotechnology, China Agricultural University). The PCR amplification was performed under the following conditions: a denaturation step at 958C for 5 min, 30 cycles at 958C for 30 sec, 628C at P1 (see below in Results) for 30 sec, 648C at P2 (see below in Results) for 30 sec, and 728C for 15 sec, and a final extension step of 728C for 7 min. The mixture, which included 5 ll of the PCR product and 10 ll of the loading buffer, was denatured for 10 min at 988C and then cooled on ice for 10 min. Polymorphisms were detected by PCR-SSCP in 10% polyacrylamide gel electrophoresis (110 V, 25 mA, 15 h). After migration, the gels were stained with 0.1% silver nitrate and visualized with 2% NaOH solution (supplied with 0.1% formaldehyde) according to Zhang et al. [30].

Fragments displaying different PCR-SSCP patterns were purified with Qiaquick spin columns (Qiagen) and sequenced on both strands with the primers used for the PCR-SSCP analysis, the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems) according to the manufacturer’s instructions. Electrophoresis was carried out with an automated DNA sequencer (ABI Prism3730 Genetic Analyser; Applied Biosystems). Followup sequencing of all the different genotypes confirmed that they indeed differed by one or more mutations; while on the other hand, sequencing of up to three samples containing identical genotypes confirmed that they displayed strictly identical sequences.

FIG. 2. Number of genes differentially expressed between any two groups during implantation. The up arrow indicates up-regulated, and the down arrow indicates down-regulated.

3

Article 52

Downloaded from www.biolreprod.org.

FIG. 1. Principal-component analysis (PCA) of the endometrial samples on Days 13, 18, and 24 of pregnancy. Samples from the same period and same sites are shown as symbols with the same color. NP, endometrium of nonpregnant sows; D13a, endometrial attachment sites on Day 13 of gestation; D13b, the endometrial intersites on Day 13 of gestation; D18a, endometrial attachment sites on Day 18 of gestation; D13b, the endometrial intersites on Day 18 of gestation; D24a, endometrial attachment sites on Day 24 of gestation; D24b, the endometrial intersites on Day 24 of gestation.

CHEN ET AL. TABLE 1. Selected results of DAVID functional annotation chart for up- and down-regulated genes in porcine endometrium on Days 13, 18, and 24 of Pregnancy. Tab Day 13 Attachment sites

Intersites Day 18 Attachment sites

Day 24 Attachment sites

Intersites a b

Up-regulated Steroid metabolic process Steroid biosynthetic process Oxidation reduction Sterol biosynthetic process Reproductive process in a multicellular organism Down-regulated Regulation of cell proliferation Positive regulation of cell proliferation Cell adhesion Biological adhesion Positive regulation of macromolecule metabolic process No significantly enriched GO terms Up-regulated Immune response Antigen processing and presentation Antigen processing and presentation of peptide antigen Antigen processing and presentation of peptide antigen via MHC class I Response to wounding Down-regulated Sex differentiation Fertilization Upregulated Steroid metabolic process Response to peptide hormone stimulus Response to endogenous stimulus Response to hormone stimulus Regulation of cell proliferation Down-regulated Steroid hormone receptor signaling pathway Blood vessel morphogenesis Response to drug Cell-cell signaling Intracellular receptor-mediated signaling pathway Up-regulated Homeostatic process Response to wounding Chemical homeostasis Inflammatory response Response to hormone stimulus Down-regulated Regulation of transcription, DNA-dependent Regulation of RNA metabolic process Proximal/distal pattern formation Response to wounding Wnt receptor signaling pathway No significantly enriched GO terms

Gene

P value

FEa

FDRb

16 10 24 7 19

2.30E08 8.50E07 2.70E06 3.50E06 2.60E05

6.5 9.6 3.1 16.4 3.2

3.80E05 1.40E03 4.60E03 5.90E03 4.30E02

27 14 18 18 20

2.40E08 1.70E04 3.80E04 3.90E04 5.30E04

3.5 3.5 2.7 2.7 2.4

4.00E05 2.80E01 6.20E01 6.30E01 8.60E01

27 11 8 6 15

4.60E14 9.30E11 3.20E10 5.10E08 5.00E06

6.2 20.8 44.9 55.5 4.5

7.40E11 1.50E07 5.20E07 8.20E05 8.10E03

5 4

7.00E03 7.30E03

6.5 9.9

1.10Eþ01 1.10Eþ01

14 11 16 15 19

1.00E08 5.30E07 1.10E06 1.80E06 6.80E05

8.4 8.6 4.8 4.9 2.9

1.70E05 8.70E04 1.70E03 2.90E03 1.10E01

5 8 8 14 5

0.0022 0.0041 0.0046 0.0048 0.0057

9 3.9 3.9 2.4 6.9

3.6 6.4 7.2 7.4 8.8

45 36 34 26 27

3.20E10 1.20E09 6.60E09 1.70E08 4.60E08

2.9 3.3 3.2 3.9 3.5

5.60E07 2.00E06 1.20E05 2.90E05 8.00E05

31 31 4 14 7

6.30E04 9.10E04 1.20E03 1.20E03 1.50E03

1.9 1.8 18.7 2.8 5.7

1.00Eþ00 1.50Eþ00 1.90Eþ00 2.00Eþ00 2.30Eþ00

FE, Fold enrichment. FDR, false discovery rate.

RESULTS

The genotype frequencies for each herd were calculated directly after SSCP. Hardy-Weinberg equilibrium (HWE) in the studied herds was tested by comparing expected and observed genotype frequencies using a chi-square test. The GLM procedure of SAS 8.02 software was used to compute the leastsquare means of TNB and NBA. After analyzing, the effect of sire and dam on litter size was not significant, so the linear model was established to analyze the genotype effects of VEGFA.

Unsupervised Hierarchical Clustering of All Samples Similar within-stage and different between-stage expression patterns were supported by PCA (Fig. 1). Interestingly, samples of attachment sites were clustered together. But samples of intersites were not clustered together. Unsupervised hierarchical clustering was performed based on the target values of the DEGs from the pregnant versus nonpregnant sows at each time point (see Supplemental Fig. S1). The dendrogram showed the relationships between the expression levels of the samples. The difference between the attachment sites and intersites was significant. Both the attachment sites and the intersites displayed variation at different implantation periods. These data further demonstrated that the changes in the

Yijkl ¼ l þ HYSi þ Pj þ Gk þ eijkl where Yijkl is the traits of TNB and NBA, l is the overall mean, HYSi is the effect of herd-year-season (i ¼ 1 to 52), Pj is the effect of parity (j ¼ 1, 2, and 3), Gk is the effect of genotype (k ¼ 1 to 3), and eijkl is the random residual. The data were analyzed separately for the first parity, the second parity, the third and following parities, and all the parities. The additive effect and the dominant effect were calculated according to the methods of Rothschild et al. [31].

4

Article 52

Downloaded from www.biolreprod.org.

Intersites

Functional term

GENE EXPRESSION IN UTERINE TISSUE OF PREGNANT SOWS

Downloaded from www.biolreprod.org. FIG. 3. Clustering of expression profiles over implantation time at endometrial attachments sites. SOMs were used to obtain groups of genes with similar expression profiles. Vertical axis is the expression value. Gene expression profiles corresponding to Figure 3 are shown in Supplemental Table S9.

endometrium at the gene expression level occurred in a timedependent manner during gestation.

Integrated Analysis of DEGs in the Endometrium During Embryonic Implantation Statistically significant expression differences between any two groups were compared. The numbers of DEGs (FC . 2, P , 0.05) are shown in Figure 2. In the endometrial attachment sites and intersites, the differences between different implan5

Article 52

CHEN ET AL. TABLE 2. Overrepresented DAVID annotation clusters for the obtained SOM expression clusters.a Expression cluster Cluster 0

Representative enriched functional termc

9.2 2.1 1.6 1.6

Translational elongation (12, 28), translation (16, 11), ribosome (12, 18) Generation of precursor metabolites and energy (8, 6) Protein folding (5, 6.6) n.r. of cellular protein metabolic process (5, 6.5), n.r. of protein metabolic process (5, 6.2), regulation of cellular protein metabolic process (7, 3.4) Generation of precursor metabolites and energy (8, 6), cellular respiration (3, 7.2), electron transport chain (3, 6.1) mRNA metabolic process (6, 3.8), RNA splicing, via transesterification reactions (4, 6.1) p.r. of apoptosis (7, 3.8), p.r. of programmed cell death (7, 3.8), p.r. of cell death (7, 3.8) Protein targeting (3, 3.3), intracellular transport (5, 1.8), macromolecular complex assembly (5, 1.8) Transcription (20, 1.4), regulation of transcription (22, 1.2), regulation of transcription, DNA-dependent (14, 1.2) Chromosome organization (6, 1.8), chromatin organization (5, 1.9), chromatin modification (3, 1.6) Pattern specification process (5, 2.8), regionalization (4, 3), anterior/posterior pattern formation (3, 3.2) Vasculature development (5, 2.9), blood vessel morphogenesis (4, 2.8), blood vessel development (4, 2.4) Regulation of autophagy (9, 28), autoimmune thyroid disease (9, 20), cytosolic DNA-sensing pathway (9, 18) Chemokine signaling pathway (5, 3), taxis (4, 3.7), chemotaxis (4, 3.7) Covalent chromatin modification (8, 9.3), histone modification (7, 8.4), histone acetylation (3, 9.2) Transcription, DNA-dependent (4, 4.2), RNA biosynthetic process (4, 4.2), transcription from RNA polymerase II promoter (3, 3.9) Glucose metabolic process (3, 6), hexose metabolic process (3, 4.8), monosaccharide metabolic process (3, 4.2) Transmembrane receptor protein tyrosine kinase signaling pathway (4, 5.5), enzyme linked receptor protein signaling pathway (4, 3.6), cell surface receptor linked signal transduction (7, 1.2) Regulation of cell growth (6, 3.8), regulation of cell development (6, 3.6), regulation of growth (6, 2.2) Golgi vesicle transport (6, 5.6) Steroid hormone receptor signaling pathway (4, 8.5), intracellular receptor-mediated signaling pathway (4, 6.6), p.r. of gene expression (6, 1.3) Nuclear-transcribed mRNA catabolic process, nonsense-mediated decay (3, 12), nuclear-transcribed mRNA catabolic process (3, 11), mRNA catabolic process (3, 8.6) n.r. of cellular protein metabolic process (6, 4.1), n.r. of protein metabolic process (6, 3.9), n.r. of macromolecule metabolic process (9, 1.5) Organic acid biosynthetic process (6, 4.8), carboxylic acid biosynthetic process (6, 4.8), cellular amino acid biosynthetic process (3, 7.2) Response to hormone stimulus (8, 2.7), response to endogenous stimulus (8, 2.4), response to organic substance (11, 1.9) Protein modification by small protein conjugation or removal (6, 4.8), protein deubiquitination (3, 15), protein modification by small protein removal (3, 13) Endocytosis (6, 3.5), membrane invagination (6, 3.5), membrane organization (8, 2.7) Protein modification by small protein conjugation or removal (6, 4.8), ubiquitin-dependent protein catabolic process (7, 3.7), modification-dependent macromolecule catabolic process (11, 2.4) Regulation of cell growth (6, 3.9), regulation of growth (7, 2.6), n.r. of cell growth (3, 4.2) Lipid biosynthetic process (8, 3.2), steroid biosynthetic process (3, 4.5), steroid metabolic process (3, 1.9) RNA processing (11, 2.6), RNA splicing (6, 2.7), mRNA processing (6, 2.4) Protein oligomerization (5, 6.3), protein homooligomerization (4, 9.2), protein complex biogenesis (7, 3) Glycoprotein metabolic process (4, 4.3), glycoprotein biosynthetic process (3, 4.1), Cellular response to stress (6, 2.3), DNA repair (3, 2.3), DNA metabolic process (4, 1.7) Behavior (6, 3), regulation of synaptic transmission (3, 5.1), regulation of transmission of nerve impulse (3, 4.8) Behavior (6, 3), regulation of secretion (3, 3.5), regulation of cellular localization (3, 2.8) Cell projection organization (5, 3.2), neuron differentiation (5, 2.7), neuron development (4, 2.8) Regulation of proteolysis (5, 34), n.r. of proteolysis (4, 65), n.r. of cellular protein metabolic process (4, 7.9) p.r. of response to stimulus (5, 7.5), regulation of response to external stimulus (4, 9), response to steroid hormone stimulus (4, 7.4) p.r. of response to stimulus (5, 7.5), regulation of response to external stimulus (4, 9), regulation of cell migration (4, 8.4) Phagocytosis (3, 22), immune response (6, 3.1), membrane invagination (3, 4.9) Steroid hormone biosynthesis (4, 26), metabolism of xenobiotics by cytochrome P450 (4, 20), oxidation reduction (5, 2.8) Bile acid and bile salt transport (3, 110), carboxylic acid transport (4, 9.7), organic acid transport (4, 9.6) Transmembrane transport (6, 3.8), cation transport (4, 2.6), ion transport (4, 1.9) Protein homooligomerization (3, 11), protein oligomerization (3, 6.1), macromolecular complex assembly (4, 2.1) Response to wounding (5, 3.4), cell adhesion (3, 1.5), biological adhesion (3, 1.5)

1.4

Cluster 1

1.2 1.2 1 1.04 0.91 0.81 0.76

Cluster 2

6.86

Cluster 3

2.22 2.11 1.04 0.93 0.92

Cluster 4

1.36 1.3 1.3 1.18 1.17 1.14 1.06

Cluster 5

1.84 1.52 1.29 1.28 1.22

Cluster 6

1.07 1.57

Cluster 7

1.32 0.89 1.05

Cluster 8

0.85 0.85 1.69 1.53 1.16

Cluster 9

1.1 3.15 2.46 0.94 0.75 0.58

6

Article 52

Downloaded from www.biolreprod.org.

Enrichment scoreb

GENE EXPRESSION IN UTERINE TISSUE OF PREGNANT SOWS TABLE 2. Continued. Expression cluster Cluster 10

Enrichment scoreb 3.17 2.01 1.57 1.48 1.37 1.14 1.03 1.02 0.8 0.76

Cluster 11

1.51 0.82

a c

Steroid hormone biosynthesis (5, 39), metabolism of xenobiotics by cytochrome P450 (4, 24), steroid metabolic process (5, 8.6) Steroid metabolic process (5, 8.6), regulation of hormone levels (4, 9.2), cellular hormone metabolic process (3, 18) Response to bacterium (4, 7.2), response to molecule of bacterial origin (3, 12), response to organic substance (6, 2.9) Monocarboxylic acid transport (3, 21), carboxylic acid transport (3, 7.1), organic acid transport (3, 7) Coenzyme biosynthetic process (3, 15), cofactor biosynthetic process (3, 11), coenzyme metabolic process (3, 6.8) Response to hormone stimulus (5, 4.7), response to endogenous stimulus (5, 4.3), response to organic substance (6, 2.9) Regulation of cell migration (3, 6.2), regulation of locomotion (3, 5.4), regulation of cell motion (3, 5.4) Regulation of neurogenesis (3, 6.3), regulation of nervous system development (3, 5.4), regulation of cell development (3, 5.1) Acute inflammatory response (3, 11), defense response (5, 2.8), inflammatory response (3, 3.2) Regulation of leukocyte activation (3, 6.3), regulation of cell activation (3, 5.9), p.r. of response to stimulus (3, 4.4) Energy reserve metabolic process (3, 29), energy derivation by oxidation of organic compounds (3, 8.8), generation of precursor metabolites and energy (3, 4.1) Response to peptide hormone stimulus (3, 8.2), response to hormone stimulus (3, 3.5), response to organic substance (4, 2.3)

p.r., positive regulation; n.r., negative regulation. Geometric mean of member’s P values of the corresponding annotation cluster (in log 10 scale). The number of genes and fold enrichment of the function term.

tation times were compared. In the endometrial attachment sites, 232, 121, and 412 up-regulated DEGs and 198, 108, and 214 down-regulated DEGs were identified on Days 13, 18, and 24 of pregnancy, respectively. In the endometrial intersites, 37, 187, and 34 up-regulated DEGs and 102, 163, and 18 downregulated DEGs were identified on Days 13, 18, and 24 of pregnancy, respectively. Interestingly, there were more upregulated DEGs than down-regulated DEGs except in endometrial intersites on Day 13. The gene expression files of attachment sites and intersites were also compared to identify the DEGs. A total of 12, 37, and 72 up-regulated DEGs were identified on Days 13, 18, and 24 of pregnancy, respectively, while 14 and 17 down-regulated DEGs were identified on Days 18 and 24 of pregnancy, respectively. No down-regulated genes was identified on Day 13 of pregnancy.

to steroid hormone receptor-signaling pathway was found as specific for down-regulated genes. The predominant group of functional terms specific for genes with higher expression in Day 24 endometrial attachment sites were related to homeostatic process, response to wounding, and response to hormone stimulus (Table 1 and Supplemental Table S5). Categories related to transcription and RNA metabolic process were found as specific for downregulated genes in Day 24 endometrial intersites (Table 1 and Supplemental Table S5). Analysis of Gene Expression Patterns over Implantation Time A dynamic analysis of gene expression for Days 13, 18, and 24 of implantation and the nonpregnant stage was performed for those genes with significant differences between different implantation times. The results of ANOVA analysis (FDR , 0.1, P , 0.05) (see Supplemental Table S8) revealed 858 DEGs at endometrial attachment sites. The result was used for SOM clustering. This analysis revealed 12 clusters of similar expression profiles over the attachment site samples from nonpregnant stage and Days 13, 18, and 24 of pregnancy (Fig. 3 and Supplemental Table S9). Particularly, clusters 8–10 contained genes obviously up-regulated over implantation time. The ANOVA analysis of intersites over implantation time revealed few genes differently expressed (FDR , 0.1, P , 0.05) (less than 5), so the result of this part was removed from our analysis. Next, a DAVID functional annotation clustering analysis was performed for the individual SOM clusters (Table 2). Strong enrichment of functions related to response to stimulus and immune response was found for genes contained in cluster 8. Overrepresented functional terms of cluster 9 were steroid hormone biosynthesis and transmembrane transport. Highly overrepresented terms for cluster 10 were related to steroid hormone biosynthesis and regulation of hormone levels. Overrepresented functional terms of clusters 0–7 and cluster

Gene Ontology Analysis of Up- and Down-Regulated Genes at Days 13, 18, and 24 of Pregnancy To find specific overrepresented functional terms for upand down-regulated genes, a comparative GO analysis was performed (Table 1 and Supplemental Table S3-S7). The predominant group of functional terms specific for genes with higher expression in Day 13 endometrial attachment sites was related to steroid metabolic and biosynthetic process as well as oxidation reduction (Table 1 and Supplemental Table S3). Categories related to cell proliferation and cell adhesion were found as specific for down-regulated genes in Day 13 endometrial attachment sites (Table 1 and Supplemental Table S3). In Day 18 endometrial attachment sites, some immune function categories and categories related to antigen processing and presentation were found as specific for up-regulated genes (Table 1 and Supplemental Table S4). In addition, categories related to sex differentiation and fertilization were found as specific for down-regulated genes. In Day 18 endometrial intersites, categories related to response to peptide hormone stimulus was found as specific for up-regulated genes (Table 1 and Supplemental Table S4). In addition, the category related 7

Article 52

Downloaded from www.biolreprod.org.

b

Representative enriched functional termc

NP, endometrium of nonpregnant sows; D13a, endometrial attachment sites on Day 13 of gestation; D13b, the endometrial intersites on Day 13 of gestation; D18a, endometrial attachment sites on Day 18 of gestation; D13b, the endometrial intersites on Day 18 of gestation; D24a, endometrial attachment sites on Day 24 of gestation; D24b, the endometrial intersites on Day 24 of gestation.

Functional Analysis of DEGs Between Attachment Sites and Intersites As the number of DEGs was small, there were no significant GO BP terms enriched in the DEGs list of attachment sites versus intersites on Day 13 of pregnancy. The enriched GO terms on Day 18 could be roughly grouped into one cluster: cell proliferation (see Supplemental Table S10). Overrepresented GO terms at Day 24 could be mainly grouped into two clusters: steroid biosynthetic and metabolic process and immune response (see Supplemental Table S10). Quantitative RT-PCR Validation of Microarray Results To validate the microarray results, quantitative RT-PCR was used to test the expression of eight genes. The results are shown in Table 3. RT-PCR showed that all eight genes were differentially expressed at the three time points of implantation. The results obtained from the microarray were statistically confirmed at .85% correlation for the genes tested by RTPCR. In every case, the relative gene expression FC was in the same direction as the microarray data and always had a higher magnitude by RT-PCR than by microarray technology. Chromosomal Localization of DEGs in QTL for Litter Size The other approach of our study was to combine microarray technology and linkage analyses to select powerful candidate genes for QTLs affecting pig litter size, as DEGs located in a QTL are obvious cis-acting positional candidate genes for the QTL [3, 32]. The most recent draft of the porcine genome (S. scrofa 10.2) and the QTL database (http://www.animalgenome. org/cgi-bin/QTLdb/SS/index) were used [25–28]. The temporal gene expression profiles of implantation (FC . 2, P , 0.05) and the different sites gene expression profiles (FC . 2, P , 0.05) were studied. Genes that were located within a QTL (a region extended by 5 Mb either side of the QTL) were accepted as candidate genes. We identified 87 candidate genes for TNB and NBA (Table 4 and Supplemental Table S11) located in nine out of the 18 porcine autosomes. Two of them were among the ones tested by RT-PCR (LIF and VEGFA). VEGFA as the Candidate Gene for Litter Size VEGFA was selected to be the candidate gene for litter size based on its biological functions during embryonic implantation and its positions within the intervals of the QTLs. VEGFA A/G at 191 bp (P1 locus) and C/T at 103 bp (P2 locus) mutations in the 5 0 -regulation domain were genotyped by PCRSSCP method. The representative SNP-sequencing output for genotypes is shown in Figure 4. The genotype frequencies and allele frequency at each polymorphic locus in Large White, Landrace, and Duroc are shown in Table 5. At the P1 locus, the genotype frequencies of AA and AG in Large White were 0.91 and 0.09, 0.91 and 0.09 in Landrace, and 0.85 and 0.15 in Duroc. All three breeds were found to be in HWE. At the P2 locus, the genotype frequencies of TT, TC, and CC in Large White were 0.79, 0.19 and 0.02, respectively,0.81, 0.16, and 0.03, respectively, in Landrace, and 0.84, 0.16, and 0, respectively, in Duroc. All three breeds were found to be in HWE. The data for TNB and NBA were observed for first, second, third and the following parities and all parities. The leastsquare means in Large White, Landrace, and Duroc are shown

a

1.43 24.16 6 0.06 55.65 6 0.39 16.40 6 1.98 1.22 4.76 6 0.37 7.30 6 0.89 2.81 6 1.28 0.06 3.41 6 0.35 2.58 6 0.07 1.77 6 0.62 0.32 1.51 6 0.44 2.68 6 0.35 1.51 6 0.40 0.48 2.21 6 0.26 3.74 6 0.90 1.37 6 0.20 0.05 5.16 6 0.42 0.99 6 0.90 1.30 6 1.01 0.07 5.60 6 0.01 2.74 6 0.14 2.18 6 0.95 0.24 1.05 6 0.17 1.19 6 0.33 0.89 6 0.48 6 6 6 6 6 6 6 6 69.47 10.79 3.60 3.04 6.10 1.20 8.27 1.14 0.33 50.79 6 1.76 6.87 6 0.09 0.48 5.82 6 0.09 1.69 6 0.05 0.65 1.98 6 0.08 1.37 6 0.78 0.13 2.43 6 0.32 1.17 6 þ0.19 0.30 7.11 6 0.24 2.29 6 0.53 0.69 4.19 6 0.03 10.26 6 0.57 0.26 2.23 6 0.09 3.68 6 0.04 0.63 1.96 6 0.17 0.97 6 0.48 6 6 6 6 6 6 6 6 12.73 1.47 1.55 1.37 1.71 7.87 4.24 5.22 1.59 0.45 0.06 0.50 0.23 0.10 0.06 0.11 6 6 6 6 6 6 6 6 56.17 4.25 2.09 2.53 5.92 2.09 3.67 3.76 RBP4 49.52 6 1.54 7.37 6 0.13 51.68 6 1.77 5.03 6 3.00 LIF 3.48 6 0.62 1.83 6 0.35 3.01 6 0.39 1.44 6 0.98 VEGFA 2.02 6 0.08 1.06 6 0.50 2.15 6 0.05 1.08 6 0.48 ACSL1 1.86 6 0.9 1.76 6 0.39 1.98 6 0.28 1.02 6 0.35 ANGPT2 5.62 6 0.26 1.63 6 0.27 5.62 6 0.24 1.65 6 0.02 IL8 3.21 6 0.15 14.32 6 1.19 2.27 6 0.08 10.68 6 1.40 PGR 2.10 6 0.03 4.29 6 0.07 1.73 6 0.01 2.95 6 0.49 CD163 1.65 6 0.16 1.34 6 0.64 3.68 6 0.32 2.00 6 0.92

Microarray RT-PCR Microarray RT-PCR Microarray RT-PCR Microarray RT-PCR Microarray RT-PCR Microarray RT-PCR

D18a vs. NP D13b vs. NP D13a vs. NP Gene

TABLE 3. Validation of eight DEGs by quantitative PCR.a

11 were mainly about translation, transcription, vasculature development, and protein modification.

8

Article 52

Downloaded from www.biolreprod.org.

D18b vs. NP

D24a vs. NP

D24b vs. NP

CHEN ET AL.

GENE EXPRESSION IN UTERINE TISSUE OF PREGNANT SOWS TABLE 4. Genes located in litter size traits (TNB and NBA). Trait TNB

NBA

TNB and NBA

a

Gene FAM213B(100134955a), TAS1R3(100127440), TNNI3(100049696), TNNT1(396579), CCND3(780426), CPNE5(100152591), FKBP5(100155423), GNMT(397444), HSPCB(396742), KLHDC3(100125975), LOC100154712(100154712), MOCS1(100154835), MTCH1(100152609), PI16(100155010), SRPK1(100155422), VEGFA(397157) DPPA5(100301558), SYNCRIP(100154370), ZUFSP(100156556), AWN(396783), CEBPA(397307), CYP2A19(403149), FXYD3(397413), HAMP(397207), LGALS4(397041), LOC100526171(100526171), SLC7A9(100037973), ABHD4(100154266), ADCY4(100152067), IL25(100154041), IPO4(100157703), LOC100152841(100152841), LOC100156489(100156489), LOC396635(396635), LOC780415(780415), TINF2(100155296), EDNRB(414911), LOC100152028(100152028), GAL3ST1(100155265), GGT1(397095), LIF(399503), LIMK2(100156359), LOC100152325(100152325), LOC100152687(100152687), LOC100157687(100157687), MMP11(100153503), PATZ1(100153931), PIK3IP1(100155147), PRR14L(100156356), RANBP1(100151896), SDF2L1(100156780), SEC14L2(100152451), SELM(100302025), SLC5A1(397113), SLC5A4(397376), TUBA3D(100155138), HAND1(541594), SPARC(595124) LOC100154785(100154785), LOC100156022(100156022), LOC100157673(100157673), POPDC3(100151910), REPS1(100154803), ADM(397195), DKK3(664653), SLC27A1(100037269), SLC5A5(399542), ANPEP(397520), BCL2A1(100156860), CIB2(100156043), CRISP2(397369), CRISP3(100158244), DNAJA4(397613), INSIG2(100170127), LOC100156419(100156419), LOC100156882(100156882), LOC100157078(100157078), LOC100169745(100169745), MFGE8(397545), MUT(399535), SH3GL3(100157660), TMED3(100155219), ZNF710(100152649), EIF4A1(100101929), ENO3(692156), SLC2A4(396754),ANGPT2(396730),

S. Scrofa Entrez Gene ID.

comparison with the TT genotype, respectively. The sows with the TT, TC, and CC genotypes had the following trend for TNB and NBA: TC . TT . CC (P , 0.05). DISCUSSION In the present study, we successfully identified temporal gene expression profiles over implantation time and gene expression profiles of different sites of implantation. However, successful embryonic implantation is dependent on complex molecular mechanisms. Genes Associated with Vascular Development and Proteolysis The porcine embryos appose and subsequently attach to the uterine luminal epithelium on Days 13 and 14 of gestation [33]. A pronounced vascularization in the endometrium is evident already from Day 13 of gestation [34, 35]. In the functional analysis, the SOM result showed that cluster 1 contained many

FIG. 4. PCR-SSCP results of swine VEGFA gene and sequence image of the different genotypes. A) Genotypes of the SSCP marker of P1 PCR products. B) Sequence image of AA and AG genotypes at P1 locus. C) Genotypes of the SSCP marker of P2 PCR products. D) Sequence image of CC and TT genotypes at P2 locus.

9

Article 52

Downloaded from www.biolreprod.org.

in Table 6 (P1 locus) and Table 7 (P2 locus). At the P1 locus, in Landrace in third and the following parities, the TNB and NBA of AG were 0.57 and 0.56 piglets higher than those of genotype AA, respectively, but not statistically significant. In Large White in the first parity, TNB and NBA were significantly increased for AG genotype with 1.34 (P , 0.05) and 1.44 (P , 0.05) piglets in comparison with the AA genotype, respectively. In Duroc in the second parity and all parities, NBA of AA were 2.59 (P , 0.05) and 0.79 (P , 0.01) piglets higher than those of genotype AG, respectively. The sows with the AG and AA genotypes had the following trend for TNB and NBA: AG . AA (P ,0.05 or P ,0.01). At the P2 locus, in Landrace in the first parity, sows with the genotype of TT had an advantage of 2.24 (P , 0.05) for TNB per litter over the CC genotype sows. In all parities, TNB of TC was 1.5 (P , 0.05) piglets higher than that of genotype CC. In Duroc in the third and the following parities, NBA of TC was 1.84 (P , 0.05) higher than that of genotype TT. In all parities, TNB and NBA were significantly increased for TC genotype with 0.84 (P , 0.05) and 0.84 (P , 0.05) piglets in

CHEN ET AL. TABLE 5. Genotype frequency and allele frequency at each polymorphic locus of VEGFA. P1* Allele frequencies Breed Landrace Large White Duroc

P2* Allele frequencies

Genotype frequencies 2

Genotype frequencies

No. of sows

A

G

AA

AG

v

T

C

TT

TC

CC

v2

240 229 105

0.95 0.95 0.92

0.05 0.05 0.08

0.91 0.91 0.85

0.09 0.09 0.15

0.51ns 0.54ns 0.71ns

0.91 0.88 0.92

0.09 0.12 0.08

0.81 0.79 0.84

0.16 0.19 0.16

0.03 0.02 0

2.42ns 2.36ns 0.72ns

* Means with ns in column are not significantly different (P . 0.05); df ¼ 2, chi-squared0.05 ¼ 5.99, chi-squared0.01 ¼ 9.21.

Hormone-Related Genes The implantation process is regulated by many kinds of hormones that affect fetal development and alter maternal physiology to support gestation [37]. From the uterus receiving the fertilized embryo and providing the right environment for implantation of the embryo to subsequent placentation, each step requires hormones. In the present study, genes of SOM clusters 8–10 and up-regulated files at Days 13, 18, and 24 contained many hormone related genes. These genes show an upward expression tendency during pregnancy, which agrees with the result of Fernandez-Rodriguez et al. [3]. Examples include ADAM metallopeptidase domain 9 (ADAM9), Indian hedgehog (IHH), bone morphogenetic protein 2 (BMP2), transforming growth factor, beta 3 (TGFB3). Many genes of the cytochrome P450 subfamily were detected differently expressed in clusters 9 and 10, such as CYP17A1, CYP19A1, CYP19A3, CYP11A1, and CYP26A1. An overrepresented pathway, metabolism of xenobiotics by cytochrome P450 (7.8 3 104), was enriched in clusters 9 and 10. Genes of the cytochrome P450 subfamily encode enzymes catalyzing many reactions involved in the synthesis of cholesterol, steroids, and other lipids, such as progesterone. The level of progesterone is statistically significantly higher during pregnancy. So metabolism by cytochrome P450 plays an important role for progesterone synthesis during pregnancy. This indicates that the uterine environment undergoes a subtle change in hormone regulation during implantation. Other biological processes were also enriched, including cell proliferation, oxidation reduction, homeostasis regulation, and sexual reproduction. These processes interact with one another and are coordinated and indispensable for successful implantation.

Immune Response-Related Genes The functional category immune response was highly overrepresented for genes specifically up-regulated in endometrial attachment sites on Days 18 and 24 of pregnancy (Table 1). Furthermore, the SOM result showed genes of cluster 8–10 contained many immune response-related genes. This agrees with many studies of pig implantation and pig reproductive performance [3, 14, 16]. A study showed that sows with lower immune system activation are less prone to maternal-fetal rejection, resulting in a more successful implantation rate [3]. In the present study, genes of SOM clusters 8–10 and up-regulated files at Day 18 contained many genes participating in the immune process with increased expression during pregnancy, for example, CD180 molecule (CD180), CD274, major histocompatibility complex, class II, DO alpha (HLA-DOA), HLA-DRA, and HLA-DQA. Many genes of the interleukin family, integrin family and the chemokine family were detected, including interleukin 4 (IL4), IL7, IL8, IL13, and IL21, chemokine (C-C motif) ligand 27 (CCL27), chemokine (C-X-C motif) ligand (CXCL2), and CXCL12. Two genes, leukemia inhibitory factor (LIF) and secreted phosphoprotein 1 (SPP1), were detected in the immune response process, and both LIF and SPP1 are known as candidate genes of litter size in pigs. In addition, the enrichment of functional categories positive regulation of

Analysis of DEGs with Known or Suspected Roles in Embryonic Implantation Gene function analysis revealed a number of DEGs that have known or suspected roles in embryonic implantation. Genes with known roles in embryonic implantation or litter size included SPP1, MUC1, VEGFA, LIF, ESR, progesterone 10

Article 52

Downloaded from www.biolreprod.org.

response to stimulus, response to wounding, and antigen processing and presentation can also be included within immune processes in the endometrial attachment sites on Day 18 of pregnancy. The study of Samborski et al. showed that pregnancy-induced immune cell infiltration begins as early as Day 12 of pregnancy [16]. The present study indicates that the immune system activation was strongest at Day 18 of pregnancy. Furthermore, the immune-related processes were only enriched with genes differently expressed in endometrial attachment sites. This indicates that the immune system selfregulates to prevent fetal rejection, resulting in a more successful implantation.

genes involved in vascular development. For example, vascular endothelial growth factor (VEGFA), angiopoietin 2 (ANGPT2), heart and neural crest derivatives expressed 1 (HAND1) and HAND2. HAND2 was considered as an aid to support the implantation window because HAND2 transmits the progesterone signals to stromal cells [36]. This triggers a signaling cascade that blocks estrogen receptor (ESR) signaling in uterine epithelial cells, thereby creating an epithelium that is receptive for embryo implantation [36]. The results also indicated that genes involved in blood vessel development were highly expressed in the attachment sites, an area where blood vessels are more abundant. This agrees with previous studies in pig that vascular development is essential for successful embryonic implantation, the formation of the placenta, and fetal development [3, 37]. A recent study by Samborski et al. revealed that there was a complex process network between vascular remodeling and proteolysis during the preimplantation phase [14, 16]. In the present study, the functional category regulation of proteolysis was significantly overrepresented for the genes with increased expression level on Day 24 of pregnancy (SOM cluster 8). Most prominent DEGs were members of the matrix metallopeptidase family (MMP1, MMP7, MMP11, MMP15, and MMP25) and TIMP metallopeptidase inhibitor family (TIMP1, TIMP2). These genes showed an upward expression tendency during pregnancy, which was accordance with the vascular development-related genes.

208 21 218 22 89 16

Litters

11.06 11.38 10.85 12.19 10.42 10.65 6 6 6 6 6 6

0.27 0.60 0.18A 0.48B 0.28 0.47

TNB 10.79 10.48 10.19 11.63 9.65 10.45 6 6 6 6 6 6

0.85 1.87 0.18A 0.48B 0.30 0.49

NBA 102 14 105 19 32 7

Litters

Values with different superscripts show significant levels within columns:

AA AG AA AG AA AG

Genotype 6 6 6 6 6 6 A, B

0.50 0.89 0.44 0.79 0.40 0.74

P  0.05;

a, b

11.20 11.70 10.85 11.41 10.85 11.97

TNB 10.22 10.21 9.98 10.50 9.96 12.55 6 6 6 6 6 6

0.49 0.85 0.43 0.76 0.49a 0.91b

NBA

P  0.01.

Second parity

127 26 124 23 42 16

Litters 11.60 12.17 12.43 12.06 9.82 11.09 6 6 6 6 6 6

TNB 0.58 0.79 0.47 0.78 0.54 084

TT TC CC TT TC CC TT TC CC

Genotype

180 43 6 200 36 4 88 17 0

Litters

11.09 11.48 8.85 11.11 10.63 9.77 10.44 10.71 6 6 6 6 6 6 6 6

b

0.29 0.42ab 1.10a 0.19 0.38 1.12 0.27 0.50

TNB 10.99 10.49 8.72 10.41 10.16 9.97 9.77 10.16 6 6 6 6 6 6 6 6

NBA 0.92 1.33 3.47 1.13 0.38 0.19 0.29 0.53

Values with different superscripts show significant levels within columns:

Duroc

Large White

Landrace

Breed 6 6 6 6 6 6 6 6

TNB 11.25 11.46 10.74 10.97 10.61 9.67 11.01 11.40

P  0.05.

a, b

86 26 4 95 26 3 27 12 0

Litters 0.52 0.71 1.54 0.45 0.72 1.73 0.39 0.60

Second parity

10.21 10.60 8.46 10.08 9.81 9.51 10.72 10.18 6 6 6 6 6 6 6 6

NBA 0.49 0.67 1.46 0.44 0.70 1.69 0.51 0.77

109 36 8 124 22 1 33 25 0

Litters

12.04 11.64 10.40 12.43 12.07 15.50 9.49 10.98

6 6 6 6 6 6 6 6

TNB 0.62 0.75 1.14 0.46 0.80 3.00 0.59 0.64

6 6 6 6 6 6 6 6

0.56 0.76 0.44 0.75 0.53 0.82

0.59 0.72 1.09 0.44 0.76 2.8 0.55a 0.61b

NBA 11.29 10.72 9.66 11.52 10.74 14.57 8.73 10.57

Third to sixth parity

Downloaded from www.biolreprod.org.

11 First parity

6 6 6 6 6 6

NBA 10.82 11.38 11.49 11.24 9.33 10.16

Third to sixth parity

TABLE 7. Effects of the VEGFA polymorphism at P2 locus on TNB and NBA in Landrace, Large White, and Duroc (least-square means 6 SEM).

a,b,A,B

Duroc

Large White

Landrace

Breed

First parity

TABLE 6. Effects of the VEGFA polymorphism at P1 locus on TNB and NBA in Landrace, Large White, and Duroc (least-square means 6 SEM).

375 105 18 419 84 48 149 53 0

Litters

437 61 447 64 163 39

Litters 0.36 0.47 0.27 0.42 0.26a 0.37b

All parities

6 6 6 6 6 6

11.69 11.79 10.34 11.37 10.89 10.33 9.74 10.58

6 6 6 6 6 6 6 6

0.37 0.42a 0.70b 0.27 0.39 0.96 0.28a 0.32b

TNB

11.50 11.91 11.30 11.60 9.87 10.66

TNB

All parities

6 6 6 6 6 6

0.75 0.99 0.27 0.41 0.27A 0.38B

10.72 10.42 8.87 10.28 9.83 9.96 9.17 10.01

6 6 6 6 6 6 6 6

0.78 0.88 1.47 0.27 0.38 0.93 0.29a 0.33b

NBA

10.43 10.47 10.20 10.69 9.22 10.27

NBA

GENE EXPRESSION IN UTERINE TISSUE OF PREGNANT SOWS

Article 52

CHEN ET AL. TABLE 8. Representative DEGs with known or suspected roles in embryonic implantation or litter size. Fold changea Day 13 vs. NP

Day 18 vs. NP

Day 24 vs. NP

GenBank

Symbol

Description

At-sites

Intersites

At-sites

Intersites

At-sites

Intersites

Reference

NM_214023 NM_214084 XM_001926883 NM_214402 NM_001166488 NM_214220 NM_001001533 NM_213875 NM_001001868 NM_214057 XM_001924386 NM_214273 NM_214020 NM_214311

SPP1 VEGFA MUC1 LIF PGR ESRA ESRB FSHB PRLR RBP4 RARG GNRHR EGF FOLR1

Secreted phosphoprotein 1 Vascular endothelial growth factor Mucin 1 Leukemia inhibitory factor Progesterone receptor Estrogen receptor alpha Estrogen receptor beta Follicle-stimulating Prolactin receptor Retinol-binding protein 4 Retinoic acid receptor-gamma Gonadotropin-releasing hormone receptor Epidermal growth factor Folate receptor 1 (adult)

2.28 1.06 1.49 1.83 4.23** 2.56* 1.12 1.05 4.35 7.37* 1.01 14.37 8.06327* 1.08

3.09 1.08 1.17 1.44 2.95* 3.33* 1.12 1.11 1.05 5.03 1.15 1.02 3.37 0.85

12.69* 1.55 2.04 1.47 4.24** 10.00** ND ND ND 12.74* ND 0.99 3.6 1.09

1.87 1.37 1.05 1.69 3.68 7.14** 2.77 1.12** 4.76 6.87* 1.45 0.90 2.54 1.01

47.05** 6.32** 1.41 4.76** 5.60** 9.09** 1.11 46.67 4.54 24.16* 2.34 0.99 1.63 7.85*

13.06* 1.91 1.39 2.81 2.18 3.03 1.02 1.85 3.13 16.4 1.78 1.14 1.35 1.36

[45–47] [48, 49] [50] [1, 38, 39] [51] [52] [53] [54–56] [42, 43] [1, 57] [58, 59] [60] [61] [62]

a At-sites, endometrial attachment sites; Intersites, endometrial intersites of the conceptus; negative values, down-regulated; ND, not detected; NP, endometrium of nonpregnant sows. *P , 0.05; **P , 0.01.

important role in maintaining a successful pregnancy, especially in the early stages of implantation. Genes Colocalized in QTL for Litter Size The DEGs were located on the pig genome available at the NCBI database (S. scrofa 10.2) and the results were linked to significant published QTLs (http://www.animalgenome.org/ cgi-bin/QTLdb/SS/index). However, linking QTLs is quite challenging because the genome of the target species has not been fully sequenced. The porcine genome (S. scrofa 10.2) is still partial and contains numerous gaps. We extended the QTL, for TNB and NBA, to 5 Mb on either side of the QTL. Thus, 87 genes were identified that colocalized in QTLs for litter size, including 41 genes for NBA, 17 genes for TNB, and 29 genes for TNB and NBA. These genes were mostly involved in blood vessel development, nutrient transportation, and tissue or organ development processes. One gene has been previously associated with prolificacy traits (LIF) [38, 39]. Some genes, such as ANGPT2 and VEGFA, have been enriched in important processes to implantation. Numerous genes in the solute carrier family, such as SLC27A1, SLC2A4, SLC5A1, SLC5A4, SLC5A5, and SLC7A9, are all located in the TNB or NBA regions. These genes were detected as differentially expressed and involved in nutrient transportation process. Moreover, SLC16A3 was previously identified as being associated with litter size traits (TNB) [40]. However, other genes identified in the present study have not been previously related with prolificacy (e.g., EIF4A1) or have unknown function (e.g., LOC100134955). A complex network of interacting genes regulates litter size. The information obtained from the expression analysis together with the QTL analyses will be useful for identifying candidate genes for litter size. The DEGs elsewhere in the genome that share pathways with genes in the QTL may have downstream effects on the QTL [3]. Thus, studying the genes that interacted with the QTL genes was also important to identify candidate genes affecting litter size. Further studies will involve fine mapping of these candidate genes and identifying polymorphisms that may be related to the litter size of pigs.

Comparison with Other Implantation Results The porcine endometrial transcriptome during early pregnancy has been extensively studied. Samborski et al. used RNA-Seq to identify gene expression in the porcine endometrium on Days 12 and 14 of pregnancy [14, 16]. We compared the gene file on Day 13 in the present study with the gene files on Days 12 and 14 in Samborski et al.’s result [14, 16]. There were 375 Ssc Entrez Gene IDs both detected on Days 12 and 13 (see Supplemental Table S12). Because the gene list was small, there were no GO terms enriched. There were 232 Ssc Entrez Gene IDs both detected on Days 13 and 14. The main biological processes enriched were immune response (see Supplemental Table S13). Erhualian pigs, a Chinese Taihu pig breed, are known to have the largest recorded litter size in the world [15]. Studies have demonstrated that Taihu sows exhibit a strong maternal effect and that their large litter sizes are mainly caused by maternal genes [15]. These observations confirmed that immune suppression regulation plays an 12

Article 52

Downloaded from www.biolreprod.org.

receptor (PGR), follicle-stimulating hormone, beta polypeptide (FSHB), prolactin receptor (PRLR), retinol-binding protein 4, plasma (RBP4), retinoic acid receptor, gamma (RARG), gonadotropin-releasing hormone receptor (GNRHR), epidermal growth factor (EGF), and folate receptor 1 (adult) (FOLR1) (Table 8). Genes such as VEGFA, LIF, RBP4, and RARG, which promoted implantation, were all up-regulated during the implantation period and reached their highest levels in the endometrial attachment sites at Day 24 of pregnancy. The expression of VEGFA and RBP4 were always higher in the endometrial attachment sites than in the intersites. The expression levels of GNRHR and EGF were highest at Day 13 of implantation and then decreased gradually. The expression level of MUC1 was down-regulated at Day 18 of pregnancy, which may be related with its inhibitory action on implantation. However, other genes in the mucin family did not show the same expression pattern, such as MUC13. The MUC13 expression level was very high (P , 0.01) and increased over time in the endometrial attachment sites and intersites. However, studies on the effect of MUC13 on implantation are not available. Further studies are required for the gene families with known roles in embryonic implantation.

GENE EXPRESSION IN UTERINE TISSUE OF PREGNANT SOWS

The Variations of VEGFA and Its Association with Litter Size 9. 10.

11. 12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

ACKNOWLEDGMENT The authors greatly appreciate Shijiazhuang Qingliangshan Swine Breeding Farm for providing samples.

23.

REFERENCES

24.

1. Spotter A, Muller S, Hamann H, Distl O. Effect of polymorphisms in the genes for LIF and RBP4 on litter size in two German pig lines. Reprod Domest Anim 2009; 44:100–105. 2. Johnson RK, Nielsen M, Casey DS. Responses in ovulation rate, embryonal survival, and litter traits in swine to 14 generations of selection to increase litter size. J Anim Sci 1999; 77:541–557. 3. Fernandez-Rodriguez A, Munoz M, Fernandez A, Pena RN, Tomas A, Noguera JL, Ovilo C, Fernandez AI. Differential gene expression in ovaries of pregnant pigs with high and low prolificacy levels and identification of candidate genes for litter size. Biol Reprod 2010; 84: 299–307. 4. Green JA, Kim JG, Whitworth KM, Agca C, Prather RS. The use of microarrays to define functionally-related genes that are differentially expressed in the cycling pig uterus. Soc Reprod Fertil Suppl 2006; 62: 163–176. 5. Cunningham PJ, England ME, Young LD, Zimmerman DR. Selection for ovulation rate in swine: correlated response in litter size and weight. J Anim Sci 1979; 48:509–516. 6. Lamberson WR, Johnson RK, Zimmerman DR, Long TE. Direct responses to selection for increased litter size, decreased age at puberty, or random selection following selection for ovulation rate in swine. J Anim Sci 1991; 69:3129–3143. 7. Neal SM, Johnson RK, Kittok RJ. Index selection for components of litter size in swine: response to five generations of selection. J Anim Sci 1989; 1933–1945. 8. Gentry BE Jr, Anderson LL, Melampy RM. Exogenous progesterone and

25.

26.

27.

28.

29. 30.

31.

32.

13

Article 52

Downloaded from www.biolreprod.org.

Marker-assisted selection in conjunction with tradition selection methods is most effective for traits such as litter size, which are either expressed later in life, are sex-dependent, or are of low heritability [41]. The candidate gene approach has led to notable success in demonstrating reproduction-related genetic markers or major genes, such as ESR, PRLR, the erythropoietin receptor (EPOR), and so on [31, 42–44]. Differently expressed genes or transcripts that fulfill one of several criteria—having a chromosomal location containing a QTL, being found in a tissue of interest involved in regulating the expression of a specific quantitative trait, and being important in a critical development stage responsible for differences in a specific quantitative trait—can be evaluated as potential candidate genes for litter size in pigs [24]. In the present study, we selected the DEG VEGFA as the candidate gene for litter size in pigs due to its chromosomal location and biological function. Two mutations were genotyped in Landrace, Large White, and Duroc populations. The results showed that at the P1 locus, AG genotype was the favorable genotype. At the P2 locus, the T allele was a favorable allele to increase litter size and TC was a favorable genotype. Because the mutations were located in the 5 0 regulation domain, the mutations may regulate the expression of VEGFA in the target tissues. The effect of VEGFA on litter size is possibly associated with its expression in endometrium during embryonic implantation. In the present study, gene expression profiles of the time course of implantation and of the different sites of implantation were evaluated and compared. Candidate genes for litter size were obtained. Polymorphism of VEGFA, one of the candidate genes, was significantly associated with litter size in pigs. Further studies should be investigated in association analysis of other candidate genes to find their effect on litter size as genetic markers.

estradiol benzoate on early embryonic survival in the pig. J Anim Sci 1973; 37:722–727. Kyriazakis I, Whittemore C. Whittemore’s Science and Practice of Pig Production, 3rd ed. Oxford: Blackwell; 2006:105–147. Tayade C, Black GP, Fang Y, Croy BA. Differential gene expression in endometrium, endometrial lymphocytes, and trophoblasts during successful and abortive embryo implantation. J Immunol 2006; 176:148–156. Anderson LL. Growth, protein content and distribution of early pig embryos. Anat Rec 1978; 190:143–153. Day BN, Anderson LL, Emmerson MA, Hazel LN, Melampy. Effect of estrogen and progesterone on early embryonic mortality in ovariectomized gilts. J Anim Sci 1959; 18:607–613. Bazer FW, Thatcher WW, Martinat-Botte F, Terqui M. Conceptus development in large white and prolific Chinese Meishan pigs. J Reprod Fertil 1988; 84:37–42. Samborski A, Graf A, Krebs S, Kessler B, Bauersachs S. Deep sequencing of the porcine endometrial transcriptome on day 14 of pregnancy. Biol Reprod 2013; 88:84. Zhang H, Wang S, Liu M, Zhang A, Wu Z, Zhang Z, Li J. Differential gene expression in the endometrium on gestation day 12 provides insight into sow prolificacy. BMC Genomics 2013; 14:45. Samborski A, Graf A, Krebs S, Kessler B, Reichenbach M, Reichenbach HD, Ulbrich SE, Bauersachs S. Transcriptome changes in the porcine endometrium during the preattachment phase. Biol Reprod 2013; 89:134. Franczak A, Wojciechowicz B, Kotwica G. Transcriptomic analysis of the porcine endometrium during early pregnancy and the estrous cycle. Reprod Biol 2013; 13:229–237. Ross JW, Ashworth MD, White FJ, Johnson GA, Ayoubi PJ, DeSilva U, Whitworth KM, Prather RS, Geisert RD. Premature estrogen exposure alters endometrial gene expression to disrupt pregnancy in the pig. Endocrinology 2007; 148:4761–4773. Su L, Zhao S, Zhu M, Yu M. Differential expression of microRNAs in porcine placentas on days 30 and 90 of gestation. Reprod Fertil Dev 2010; 22:1175–1182. Lord E, Murphy BD, Desmarais JA, Ledoux S, Beaudry D, Palin MF. Modulation of peroxisome proliferator-activated receptor delta and gamma transcripts in swine endometrial tissue during early gestation. Reproduction 2006; 131:929–942. Zhang R, Fang H, Chen Y, Shen J, Lu H, Zeng C, Ren J, Zeng H, Li Z, Chen S, Cai D, Zhao Q. Gene expression analyses of subchondral bone in early experimental osteoarthritis by microarray. PLoS One 2012; 7: e32356. Dennis G Jr, Sherman BT, Hosack DA, Yang J, Gao W, Lane HC, Lempicki RA. DAVID: database for annotation, visualization, and integrated discovery. Genome Biol 2003; 4:P3. Huang da W, Sherman BT, Lempicki RA. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 2009; 4:44–57. Sun X, Mei S, Tao H, Wang G, Su L, Jiang S, Deng C, Xiong Y, Li F. Microarray profiling for differential gene expression in PMSG-hCG stimulated preovulatory ovarian follicles of Chinese Taihu and Large White sows. BMC Genomics 2011; 12:111. Hu ZL, Reecy JM. Animal QTLdb: beyond a repository. A public platform for QTL comparisons and integration with diverse types of structural genomic information. Mamm Genome 2007; 18:1–4. Hu ZL, Park CA, Wu XL, Reecy JM. Animal QTLdb: an improved database tool for livestock animal QTL/association data dissemination in the post-genome era. Nucleic Acids Res 2013; 41:D871–D879. Hu ZL, Dracheva S, Jang W, Maglott D, Bastiaansen J, Rothschild MF, Reecy JMA. QTL resource and comparison tool for pigs: PigQTLDB. Mamm Genome 2005; 16:792–800. Animal QTLdb [Internet], PigQTLdb. Ames, IA: Iowa State University, the NAGRP Bioinformatics Coordination Program. http://www. animalgenome.org/cgi-bin/QTLdb/SS/index. Accessed 20 August 2012. Hayashi K. PCR-SSCP: a method for detection of mutations. Genet Anal Tech Appl 1992; 9:73–79. Zhang C, Wang Y, Chen H, Lan X, Lei C. Enhance the efficiency of single-strand conformation polymorphism analysis by short polyacrylamide gel and modified silver staining. Anal Biochem 2007; 365:286–287. Rothschild M, Jacobson C, Vaske D, Tuggle C, Wang L, Short T, Eckardt G, Sasaki S, Vincent A, McLaren D, Southwood O, van der Steen H, et al. The estrogen receptor locus is associated with a major gene influencing litter size in pigs. Proc Natl Acad Sci U S A 1996; 93:201–205. Jouffe V, Rowe S, Liaubet L, Buitenhuis B, Hornshoj H, Sancristobal M, Mormede P, de Koning DJ. Using microarrays to identify positional

CHEN ET AL.

33. 34.

35.

36. 37.

38.

39.

40.

42.

43.

44.

45.

46.

47.

14

Article 52

Downloaded from www.biolreprod.org.

41.

48. Kaczmarek MM, Blitek A, Kaminska K, Bodek G, Zygmunt M, Schams D, Ziecik AJ. Assessment of VEGF-receptor system expression in the porcine endometrial stromal cells in response to insulin-like growth factorI, relaxin, oxytocin and prostaglandin E2. Mol Cell Endocrinol 2008; 291: 33–41. 49. Kaczmarek MM, Kiewisz J, Schams D, Ziecik AJ. Expression of VEGFreceptor system in conceptus during peri-implantation period and endometrial and luteal expression of soluble VEGFR-1 in the pig. Theriogenology 2009; 71:1298–1306. 50. Xiao C, Jinluan F, Aiguo W. Effect of VNTR polymorphism of the Muc1 gene on litter size of pigs. Mol Biol Rep 2012; 39:6251–6258. 51. Mathew DJ, Sellner EM, Green JC, Okamura CS, Anderson LL, Lucy MC, Geisert RD. Uterine progesterone receptor expression, conceptus development, and ovarian function in pigs treated with RU 486 during early pregnancy. Biol Reprod 2011; 84:130–139. 52. Ellegren H, Chowdhary BP, Fredholm M, Hoyheim B, Johansson M, Brauner Nielsen PB, Thomsen PD, Andersson L. A physically anchored linkage map of pig chromosome 1 uncovers sex- and position-specific recombination rates. Genomics 1994; 24:342–350. 53. Munoz G, Ovilo C, Amills M, Rodriguez C. Mapping of the porcine oestrogen receptor 2 gene and association study with litter size in Iberian pigs. Anim Genet 2004; 35:242–244. 54. Mellink C, Lahbib-Mansais Y, Yerle M, Gellin J. PCR amplification and physical localization of the genes for pig FSHB and LHB. Cytogenet Cell Genet 1995; 70:224–227. 55. Zhao Y, Li N, Xiao L, Cao G, Chen Y, Zhang S, Wu C, Zhang J, Sun S, Xu X. FSHB subunit gene is associated with major gene controlling litter size in commercial pig breeds. Sci China C Life Sci 1998; 41:664–668. 56. Du LX, Liu SF, Yan YC, Jiang YL. Research on Alu element inserted mutation in porcine FSH beta subunit gene. Yi Chuan Xue Bao 2002; 29: 977–982 [in Chinese]. 57. Rothschild MF, Messer L, Day A, Wales R, Short T, Southwood O, Plastow G. Investigation of the retinol-binding protein 4 (RBP4) gene as a candidate gene for increased litter size in pigs. Mamm Genome 2000; 11: 75–77. 58. Messer L, Wang L, Yelich J, Pomp D, Geisert R, Rothschild MF. Linkage mapping of the retinoic acid receptor-gamma gene to porcine chromosome 5. Anim Genet 1996; 27:175–177. 59. Messer LA, Wang L, Yelich J, Pomp D, Geisert RD, Rothschild MF. Linkage mapping of the retinol-binding protein 4 (RBP4) gene to porcine chromosome 14. Mamm Genome 1996; 7:396. 60. Jiang Z, Gibson JP, Archibald AL, Haley CS. The porcine gonadotropinreleasing hormone receptor gene (GNRHR): genomic organization, polymorphisms, and association with the number of corpora lutea. Genome 2001; 44:7–12. 61. Mendez EA, Messer LA, Larsen NJ, Robic A, Rothschild MF. Epidermal growth factor maps to pig chromosome 8. J Anim Sci 1999; 77:494–495. 62. Vallet JL, Freking BA, Leymaster KA, Christenson RK. Allelic variation in the secreted folate binding protein gene is associated with uterine capacity in swine. J Anim Sci 2005; 83:1860–1867.

candidate genes for QTL: the case study of ACTH response in pigs. BMC Proc 2009; 3(Suppl 4):S14. Dantzer V. Electron microscopy of the initial stages of placentation in the pig. Anat Embryol (Berl) 1985; 172:281–293. Keys JL, King GJ. Morphological evidence for increased uterine vascular permeability at the time of embryonic attachment in the pig. Biol Reprod 1988; 39:473–487. Keys JL, King GJ, Kennedy TG. Increased uterine vascular permeability at the time of embryonic attachment in the pig. Biol Reprod 1986; 34: 405–411. Hewitt SC, Korach KS. Cell biology. A hand to support the implantation window. Science 2011; 331:863–864. Zhou QY, Fang MD, Huang TH, Li CC, Yu M, Zhao SH. Detection of differentially expressed genes between Erhualian and Large White placentas on day 75 and 90 of gestation. BMC Genomics 2009; 10:337. Spotter A, Drogemuller C, Hamann H, Distl O. Evidence of a new leukemia inhibitory factor-associated genetic marker for litter size in a synthetic pig line. J Anim Sci 2005; 83:2264–2270. Spotter A, Drogemuller C, Kuiper H, Brenig B, Leeb T, Distl O. Molecular characterization and chromosome assignment of the porcine gene for leukemia inhibitory factor LIF. Cytogenet Cell Genet 2001; 93: 87–90. Noguera JL, Rodriguez C, Varona L, Tomas A, Munoz G, Ramirez O, Barragan C, Arque M, Bidanel JP, Amills M, Ovilo C, Sanchez A. A bidimensional genome scan for prolificacy traits in pigs shows the existence of multiple epistatic QTL. BMC Genomics 2009; 10:636. Soller M. Marker assisted selection, overview. Anim Biotech 1994; 5: 193–207. van Rens BT, Evans GJ, van der Lende T. Components of litter size in gilts with different prolactin receptor genotypes. Theriogenology 2003; 59: 915–926. Vincent V, Goffin V, Rozakis-Adcock M, Mornon JP, Kelly PA. Identification of cytoplasmic motifs required for short prolactin receptor internalization. J Biol Chem 1997; 272:7062–7068. Li N, Zhao YF, Zeng FJ, Chen YZ, Dai RJ, Zang JS, Shen SQ, Chen YF, Wu, CX. Candidate gene analysis for identification of genetic loci controlling litter size in swine. In: Proceedings of the 6th World Congress of Genetics Applied to Livestock Production 1998; 183–186. Zhang Q, Wrana JL, Sodek J. Characterization of the promoter region of the porcine opn (osteopontin, secreted phosphoprotein 1) gene. Identification of positive and negative regulatory elements and a ‘silent’ second promoter. Eur J Biochem 1992; 207:649–659. Short TH, Rothschild MF, Southwood OI, McLaren DG, de Vries A, van der Steen H, Eckardt GR, Tuggle CK, Helm J, Vaske DA, Mileham AJ, Plastow GS. Effect of the estrogen receptor locus on reproduction and production traits in four commercial pig lines. J Anim Sci 1997; 75: 3138–3142. Korwin-Kossakowska A, Reed KM, Pelak C, Krause E, Morrison L, Alexander LJ. Radiation hybrid mapping of 118 new porcine microsatellites. Anim Genet 2002; 33:224–227.