| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ANIMAL GENETICS |
Department of Animal Science, University of Nebraska, Lincoln 68583-0908
Abstract
Differential display PCR (ddPCR) and complementary DNA microarray analyses were used to evaluate gene expression differences in porcine ovarian follicles between a line of pigs selected for an index of ovulation rate and embryo survival (Line I) and its randomly selected control line (Line C). Follicles (4.0 to 7.0 mm) were dissected from ovaries of multiparous sows (n = 27) at either 2 or 4 d following PGF2
analog injection on d 12 to 14 of the estrous cycle. Using ddPCR, differentially expressed bands (n = 282) were excised from gels and 107 were sequenced, yielding 84 unique porcine follicle expressed sequence tags. Northern hybridization confirmed differential expression (between lines, days, or follicle sizes) for messenger RNA representing the calpain I light subunit, cytochrome C oxidase subunit III, cytochrome P450 aromatase, and cytochrome P450 side chain cleavage genes. For microarray analysis, two mRNA pools representing follicles (d 2; 4.50 to 4.75 mm) from Line I and Line C sows were hybridized to the Incyte UniGEM V1.0 human chip (approximately 7,000 gene probes). A second analysis was performed using mRNA from follicles (d 2; 4.50 to 5.00 mm) hybridized to the Incyte UniGEM V2.0 human chip (approximately 9,100 gene probes). A total of 33 and 21 genes were identified with significant expression differences using UniGEM V1.0 and V2.0, respectively (twofold or greater relative expression following adjustment for expression of control probes). However, there was little overlap between results of the two hybridizations. Expression differences between lines for two genes, follistatin and nuclear receptor subfamily 4, group A, member 1, were confirmed using Northern hybridization. These results demonstrate changes in follicular gene expression as the result of long-term selection for enhanced reproduction. These correlated responses may directly represent allelic variation utilized by selection (e.g., quantitative trait loci), or more likely, transcriptional changes in other genes that interact with reproductive QTL. This work represents one of the first applications of gene expression analysis to evaluate long-term selection response in livestock populations.
Key Words: Differential Display Polymerase Chain Reaction • Microarray • Ovulation Rate • Pig • Selection
Introduction
Female reproductive traits in the pig are generally polygenic in nature and have relatively low heritabilities. Genetic characterization of litter size and its components (e.g., ovulation rate and embryo survival) will increase our understanding of the underlying physiology and could enhance genetic improvement through use of marker-assisted selection or transgenic modification.
Long-term selection for an index of ovulation rate and embryo survival has successfully increased litter size (Johnson et al., 1999
). After 11 generations of selection for an index of ovulation rate and embryonic survival, Index line (Line I) females ovulated 7.4 more ova than randomly selected Control line females (Line C). Yen (1999) concluded that Line I gilts maintain a larger pool of healthy medium-sized follicles to d 4 of the follicular phase, which can be selected and matured into ovulatory follicles.
Previous quantitative trait locus and candidate gene analyses (Cassady et al., 2001; Linville et al., 2001) in these lines were relatively unsuccessful in elucidating the underlying genes involved in ovulation rate differences between the selection lines. To better understand the genetic and physiological consequences of long-term selection for enhanced female reproduction, we employed a gene expression approach to identify differences between Lines I and C in the ovarian transcriptome. Initial evaluation utilized differential display PCR, and subsequently human cDNA microarrays were used to evaluate differences in thousands of known genes in parallel. Based on previous information in these selection lines (Yen, 1999), ovarian follicles at 2 and 4 d following synchronization of follicular development were selected for gene expression analysis. Our primary hypothesis was that selection for components of litter size has led to correlated responses in altered mRNA levels for genes expressed in the ovarian follicle.
Materials and Methods
Population and Experimental Design
A sample of 13 Line I and 14 Line C sows from the 16th generation of selection were used after they had farrowed their first litter. Line I was selected for increased ovulation rate, increased embryonic survival, and increased litter size, and Line C was selected randomly. Selection procedures and responses to selection are in Neal et al. (1989) and Johnson et al. (1999).
Sows were checked twice daily for expression of estrus and were injected with the prostaglandin F2 analog Lutalyse (Upjohn, Kalamazoo, MI) 12 to 14 d later. Sows were killed either 2 or 4 d (d 2 or d 4, respectively) after Lutalyse injection at the UNL Meat Laboratory under USDA supervision. The experimental design included six Line I and seven Line C sows at d 2, and seven each of Lines I and C sows at d 4. Days 2 and 4 of the follicular phase were chosen because Yen (1999) previously found differences between Line I and C sows at d 4 in dynamics of follicular growth. Number of live piglets and number of fully formed piglets at birth had been recorded on a previous litter for each sow. Live weight was measured for each pig.
Tissue Collection
After being killed, a slit was made in the abdominal region of each sow, and the ovaries were removed for dissection. Ovaries were kept in ice-cold physiological saline (0.9% wt/vol NaCl) during dissection of follicles. Caliper measurements were taken from three different directions across the follicle and averaged to calculate follicle size in increments of 0.25 mm. Follicles were snap-frozen in liquid nitrogen within a 20- to 35-min period after the killing of the sows, and were stored at -80°C.
Statistical Analyses
The phenotypic traits number of live piglets, number of fully formed (FF) piglets at birth, and sow weight were analyzed using PROC GLM of SAS (SAS Inst. Inc., Cary, NC) with the model Y = µ + L + e, where Y is the response variable, µ is the mean, L is the fixed effect of line, and e is the residual. Follicle size was analyzed using PROC MIXED of SAS with the model Y = µ + L + D + (LD) + e, where Y is the response variable, µ is the mean, L is the fixed effect of line, D is the random effect of day of prostaglandin analog injection, LD is the interaction between line and day, and e is the residual effect.
RNA Extraction
Total RNA was extracted from ovarian follicles using TRIZOL LS reagent (Gibco, Grand Island, NY) and treated with DNase I (Gibco) in the presence of RNasin (Promega, Madison, WI). The concentration of RNA was determined with a TD-700 Laboratory Fluorometer (Turner Designs, Sunnyvale, CA) using the average readings from three separate dilutions.
Differential Display PCR
RNA Pooling for Differential Display PCR.
Follicles from Line I and Line C sows were each grouped into six pools representing three sows each from d 2 and d 4 (Table 1). The pools were further divided into size classifications, with d-2 pools having follicle diameters of 4.0 to 4.25 mm, 4.5 to 4.75 mm, and 5.0 to 5.25 mm, and d-4 pools having follicle diameters of 4.0 to 4.5 mm, 5.0 to 5.5 mm, and 6.0 to 7.0 mm. Each pool consisted of a balanced contribution of RNA from three sows. Sows in each pool were selected based on their litter size. Line I females utilized had 
12 FF piglets, whereas Line C females used had 
11 FF piglets in their previous litter. All d-2 pools were comprised of samples from the same three sows (within line), as the size range of follicles was not very large. By d 4, follicle size range had increased, so samples from five sows were required to populate the within-line pools. The pooling strategy enabled comparison of six pools per line for addressing the objective of identifying differences between lines. There was also replication of six pools per day for the objective of identifying differences between days. For examining the effect of follicle size, there was replication of two pools per class. The interaction of line x treatment day x follicle size was not examined.
|
Differential Display PCR.
Fluorescent differential display PCR (ddPCR) was utilized to evaluate gene expression differences. Each sample representing a pool was amplified with 200 different primer combinations, consisting of each of the 10 3' fluorescent anchor primers in combination with each of 20 5' arbitrary primers (5'–3' sequences: M13 reverse–48 + a random 10-base pair sequence). Ten-microliter PCR were run with 1 µL of cDNA, 3.75 mM MgCl2, 50 µM each dNTP, 0.35 µM arbitrary primer, 0.35 µM TMR-labeled fluorescent anchor primer (Genset Oligos, La Jolla, CA), and 0.5 U of Taq polymerase with 1x accompanying buffer (Promega). Thermal cycling conditions were initial denaturation at 95°C for 2 min, followed by three cycles of denaturation at 92°C for 15 s, annealing at 50°C for 30 s, and extension at 72°C for 2 min. This was followed by 29 cycles of denaturation at 92°C for 15 s, annealing at 60°C for 30 s, and extension at 72°C for 2 min with a final extension period at 72°C for 7 min.
After fluorescent ddPCR, samples were mixed with 1.5 µL of loading buffer (formamide and dextran blue), denatured, and electrophoresed on a 5.8% polyacrylamide denaturing gel containing urea for 5 h at 3,000 V and 100 W in a GenomyxLR DNA sequencer (Beckman, Fullerton, CA). Gel images were analyzed using a GenomyxSC Fluorescent Imaging Scanner (Beckman) and Adobe Photoshop 4 (Adobe Systems Incorporated, San Jose, CA). Differentially expressed bands were selected for excision based on visual appraisal, suspended in 50 µL of TE (10 mM Tris•Cl, pH 7.4), and incubated at 37°C for 30 min.
Reamplification of Differential Display PCR Products.
Excised band lysate was used as template for PCR using M13rev–48 and T7 primers. Reaction conditions included annealing at 50°C for two initial cycles followed by annealing at 60°C for 24 cycles.
Sequence Characterization of Differential Display PCR Products.
Cloning of ddPCR products for sequencing was performed using the TOPO TA Vector Cloning Kit (Invitrogen, Carlsbad, CA). Colony PCR using product-specific anchor and arbitrary primers was utilized to analyze and select clones for sequencing. Products of PCR were evaluated in 4% metaphor gels stained with ethidium bromide. Product sizes were compared with expected values based on the ddPCR gel and the reamplification products.
Glycerol stocks were prepared from cultures inoculated with the selected clone. Plasmid was extracted from the bacteria using the Concert High Purity Plasmid Miniprep System (Gibco) using the remainder of the inoculated medium. Five microliters of the extracted plasmid was subjected to EcoRI (Promega) digestion to test the length of the insert. After 6 h of digestion at 37°C, electrophoresis in a 2% agarose gel was used to visualize products. Product sizes were compared with expected values based on the ddPCR gel.
Automated dideoxy terminator cycle sequencing was performed at the University of Nebraska DNA Sequencing Core Facility. Expressed sequence tags (EST) were evaluated using the BLAST program (www.ncbi.nlm.nih.gov/BLAST/) and were classified as matching known genes, matching unknown EST, or as being a novel gene with no significant hits to other sequences. GenBank accession numbers were obtained for each submission. Sequences with ambiguous BLAST results due to matches with repetitive elements were further analyzed using RepeatMasker (www.genome.washington.edu/UWGC/analysistools/repeatmask.htm). Sequences were clustered using the TIGR porcine gene index (www.tigr.org/tdb/tgi/ssgi/).
Confirmation of Differential Expression Using Northern Hybridization.
Northern hybridization was used to confirm differences found by ddPCR. Because the use of individual samples was limited, due to a low volume of RNA, pools of total RNA were created using the same sows that contributed RNA for the ddPCR study. For Northern hybridizations between different-sized follicles, Line I and Line C samples were merged to create pools of follicles of sizes 4.0 to 4.75 mm, 5.0 to 5.75 mm, 6.0 to 6.75 mm, and 7.0 mm.
To create membranes for hybridization, 15 µg total RNA was utilized. Membranes were prepared as described previously (Allan et al., 2000). Probes for hybridization were produced by random priming using the MegaPrime DNA Labeling Systems kit (Amersham Life Sciences, Buckinghamshire, England). Twenty five nanograms of reamplified cloned ddPCR product was labeled with 8 µCi [-32P]dCTP using random nonamer primers. The reaction was incubated for 10 min and then quenched using 0.2 M EDTA, pH 8.0.
Hybridization of labeled probes to the membranes was as described previously (Allan et al., 2000). Following hybridization, membranes were exposed to Kodak BioMax MS autoradiography film (Eastman-Kodak Company, Rochester, NY) to test for probe binding and intensity. Next, membranes were exposed to a PhosphoImager cassette for quantification using the PhosphoImager:SF (Molecular Dynamics, Sunnyvale, CA).
Membranes were stripped using a solution of 0.1% SDS and 0.1x SSC with incubation at 95°C for 20 min (twice). Membranes were then probed with cytochrome C oxidase III (COIII) to standardize loading. This gene was shown to be equally expressed by hybridizing it and ß-actin on the same membrane. After quantification of the gene probe and COIII, the standardized differential expression was determined using the following equation, where subscript prb is the intensity of the probe and std is the intensity of the COIII standard:
 |
Microarray Analysis
Two separate hybridizations to human microarrays were performed by Incyte Genomics (St. Louis, MO). For the first analysis, 1,800 ng of mRNA from each line was used and cDNA from Lines I and C were labeled with Cy5 and Cy3 fluorescence, respectively. Samples were hybridized to the human UniGEM V1.0 microarray, consisting of 7,075 probes, including 192 control probes (48 control sequences from the yeast genome, each arrayed four times).
A second analysis was performed using the human UniGEM V2.0 microarray, using different mRNA samples (1,800 ng) from each line. In this case, cDNA from Line I was labeled with Cy3 fluorescence, and cDNA from Line C was labeled with Cy5 fluorescence. This second-generation microarray contained 9,182 probes, of which 2,748 were novel relative to the probes on UniGEM V1.0, whereas 641 probes had been removed.
For analysis with human UniGEM V1.0, Lines I and C were compared using d-2 follicles in size classifications of 4.5 to 4.75 mm. Two mRNA pools were created, and each comprised two sows within line. Sows used for this initial array analysis were not used in the ddPCR evaluation. For the subsequent analysis on the UniGEMV2.0 human microarray, lines were compared using d-2 follicles ranging in size from 4.5 to 5.0 mm. Two mRNA pools were created, and each comprised two sows within line. The Line I pool comprised four sows, including the two sows used in the previous array study and two sows used in the ddPCR study. The Line C pool comprised five sows, including the two sows used in the previous array study, one sow from the ddPCR study, and two additional sows.
Microarray Hybridization.
Samples were labeled with Cy3 and Cy5 and the reaction pair combined. This solution was hybridized to the GEM microarray followed by scanning. Background was subtracted based on fluorescence from the control probes, and the ratio of Cy3:Cy5 was determined. Results were analyzed using the Incyte GEMtools software (Incyte Pharmaceuticals, Palo Alto, CA). The statistical significance threshold was set as a twofold or greater difference in expression.
Confirmation of Microarray Analysis Using Northern Hybridization.
Selected significant results from the microarray analyses were tested for confirmation using Northern hybridization as described above. These were selected based on the biological function of the gene product. Thus, results were tested as lines across all follicle sizes and the two treatment days, instead of testing only the experimental samples used in the microarray analyses. Cloning and sequence verification of the UniGEM clones were performed as described above using Incyte universal primers.
Results
The 27 pigs used in this experiment were selected from a population of 90 sows, 45 per line. Litter data were recorded on all 90 sows. Analysis of the 90 sows indicated 2.73 more fully formed and 2.22 more live pigs in Line I (P < 0.0001); the differences were 2.39 fully formed and 2.35 live pigs in the sample of 27 sows selected for this study (P < 0.001). Weight of the 27 sows before tissue collection did not differ between lines (P > 0.1). Follicle size tended to be larger in the Index line (P = 0.08), and d-4 follicles were 0.56 mm larger than d-2 follicles; line x day interaction was not significant. Least squares means are presented in Table 2.
|
. Eighty-four of the 107 EST (78.5%) represent unique sequences. Forty-five of the matches were to known genes (five of these were EEG), 32 have partial matches to sequenced clones or EST (two were EEG), and 7 had no match in the database representing novel EST. Twenty-three (21.5% of total and 27.3% of unique) EST contained repetitive elements.
|
shows examples of each type of expression difference, and Table 3
summarizes potential expression differences for each EST.
|
). Of the Northern hybridizations performed on DEG, three were confirmed to be differentially expressed. The expression of calpain I light subunit (CAPN4) was greater in Line C (Figure 2). Expression of cytochrome P450 aromatase (CYP19; Figure 2) and cytochrome P450 side chain cleavage (CYP11A; Figure 2) increased with increasing follicle size. Expression of cytochrome P450 side chain cleavage was more variable across follicle sizes than that of CYP19.
|
|
). Four of these probes had greater expression in the Index line (regulator of G-protein signaling 12, an EST weakly similar to CGI-69 protein, an EST weakly similar to HKL, and an EST with no identification). The other 29 probes showed lower expression in Line I. Hybridization of samples to UniGEM V2.0 resulted in 21 probes showing differential expression (Table 5). Four of these probes had higher expression in Line I (cytochrome P450, subfamily XIX, Nr4a1, cathepsin L2, and retinoic acid receptor responder (tazarotene induced) 1), and the remaining 17 probes exhibited lower expression as a result of Index selection. For UniGEM V1.0, 97% of the probes had signal bound to them. For UniGEM V2.0, hybridization signal was detected for 84.5% of the probes. The vast majority of the microarray results clustered around a ratio of 1.0 (Figure 3).
|
|
Only one probe had significantly different expression in both of the hybridizations, an EST with weak similarities to ORFYNR007c (Saccharomyces cerevisiae) with reduced mRNA levels in Line I. Seven probes significant in UniGEM V1.0 were not present on UniGEM V2.0, and six probes significant in UniGEM V2.0 were new to the microarray (Table 5). Twenty expression changes from one array were verified, albeit at a lower magnitude, by the other array, whereas fifteen probes only demonstrated significant ratios on one array or the other (Table 5).
Northern hybridization was performed for follistatin (FST), early growth response 1 (EGR1), and nuclear receptor subfamily 4, group A, member 1 (NR4A1; Figure 2). Reduced expression of FST in Line I was confirmed with an average differential expression of 3.5-fold on two different membranes, whereas greater expression of NR4A1 in Line I was confirmed with an 8.4-fold difference. However, the NR4A1 results should be interpreted with caution due to high background in the hybridization. Early growth response 1 was not confirmed to have differential expression between lines.
Discussion
Cassady et al. (2001) performed a genomewide scan to search for QTL influencing various reproductive traits in F2 females of a Line I and Line C cross. Although several putative QTL were identified for litter size, ovulation rate, and number of stillborn piglets, the mapping resolution of this study and similar studies (e.g., Rohrer et al., 1999, and Wilkie et al., 1999) was insufficient to enable identification of the underlying genes. Linville et al. (2001) evaluated six candidate genes for reproduction in advanced selection lines derived from Line I (Ruíz-Flores and Johnson, 2001) and Line C. Although allele frequency differences suggested a role for three of the loci in selection response, estimates of additive and dominance effects at these loci did not differ from zero. Thus, to better elucidate the genetic and physiological consequences of long-term selection for enhanced female reproduction, in this study we employed a gene expression approach to identify differences between Lines I and C in the ovarian transcriptome, using differential display PCR and human cDNA microarrays.
Phenotypic results for litter size traits in the I and C selection lines were in general agreement with Johnson et al. (1999). As expected, there was a significant difference between treatment days for follicular size, with d-4 follicles being larger than those harvested on d 2, and follicle size tended to be slightly larger in the Index Line. These results indicate that a representative sampling of pigs was obtained from the two selection lines and that follicular development following treatment with PGF2 was taking place following an expected pattern.
This ddPCR study identified 108 genes expressed in the porcine ovarian follicle. A total of 45 of the potentially differentially expressed EST found by ddPCR matched to known genes. The 3' UTR is highly variable between species, so comparative information may not always be helpful in sequence identification. Currently there are 
120,000 pig sequences deposited in GenBank, clustering into 17,000 expressed genes represented by two or more EST. A significant number of these originated from a large ovarian-follicle cDNA library (Caetano et al., 2003). Based on the current database status, this present study has contributed seven previously unknown EST to the databank. The value of the EST submissions resulting from this study is enhanced given that each EST potentially represents a gene with differential expression during an important period of follicular development, with relevance to litter size.
Two EST were confirmed to have greater expression as follicle size increases, representing genes for CYP19 and CYP11A. These results validate the use of ddPCR to identify gene expression changes, given that they confirm previous findings (Garrett and Guthrie, 1996; Yuan et al., 1996).
In regard to correlated responses to selection in the ovarian transcriptome, the primary finding of this ddPCR study was that levels of mRNA for the gene coding for calpain I light subunit decreased as a result of Index selection. Another member of the calpain small subunit subfamily is apoptosis-linked gene product (ALG-2; Lo et al., 1999). The cellular role of ALG-2 has been demonstrated to be involved in apoptosis. Calpains are most widely noted in livestock as being proteinases involved in postmortem muscle digestion and thus are associated with meat tenderness (Koohmaraie et al., 1992; Parr et al., 1999). Speculatively, Index selection may have reduced expression of CAPN4 allowing for decreased follicular degradation and apoptosis, which would lead to a larger pool of follicles available for recruitment for ovulation. However, Mellgren et al. (1996) found that decreased calpain activity in Chinese hamster ovary cells resulted in reduced proliferation and growth. Alternatively, changes in expression of CAPN4 in the Index line may be a result of genetic drift.
The high percentage of probes showing hybridization signals for both UniGEM V1.0 and V2.0 indicate successful utilization of cross-species hybridization between pig and human, confirming results presented by Moody et al. (2002) using nylon membranes. The second-generation microarray (UniGEM V2.0) was used as a confirmation of the results from the initial human UniGEM V1.0 microarray. However, results between the two arrays were highly variable, indicating that the arrays and/or hybridizations were not of high quality, or, more likely, that much greater replication was required. This relative lack of replication also limited the use of statistical analyses for evaluation of expression changes to rankings based on magnitude of ratios. Furthermore, samples used across the two arrays were not identical. Thus, results from this study should be interpreted with great caution, and future confirmation of individual expression differences is required to differentiate between results caused by genetic differences between lines and those obtained by chance.
The gene encoding follistatin was confirmed to have lower expression as a correlated response to Index selection. One function of FST is to bind to activin, thereby inactivating its function of stimulating granulosa cell proliferation and differentiation (Hasegawa et al., 1994). A second function of FST is to stimulate FSH receptor formation in granulosa cells (Findlay, 1993; Hasegawa et al., 1994). If selection has decreased the amount of follicular follistatin produced by Line I sows, there would be less follistatin to bind to activin in the follicle. With an increase in bioavailable activin in the ovarian follicle, follicular growth and differentiation may be accelerated in Line I. However, there was no other indication of Line I follicles being further differentiated than Line C follicles because we saw no differences in aromatase, a predictor of follicle differentiation, between lines. Additionally, an increase in FSH receptors in Line I granulosa cells may enhance the ability of Line I follicles to respond to FSH, also potentially increasing growth of the follicle.
Alternatively, Li et al. (1997) found that a decrease in follistatin normally occurred in follicles during advanced stages of the estrous cycle. This should also indicate that follicles from Line I are further advanced in their development. This does not support the results by Yen (1999), which indicated that greater development of follicles was seen in Line C until d 4. Only after d 4 did the Line I animals have a greater number of large follicles. At this point, the role of follistatin in determining differences in ovulation rate between Lines I and C remains unclear.
Ignoring the possibility of genetic drift, changes in expression between Lines I and C are correlated responses to selection. In this context, lower expression in Line I indicates that selection acted to decrease expression of a gene, whereas higher mRNA levels in Line I indicate that selection acted to increase expression of a gene. Changes in expression also could be the direct result of allelic variation in the base population (e.g., the differentially expressed gene is the direct transcriptional consequence of a QTL), or, alternatively, changes in expression can be caused indirectly as the result of interaction between a QTL and the differentially expressed locus.
The gene FST maps to Sus scrofa chromosome (SSC) 16 (Ellegren, 1993). In this population, there are no significant QTL for ovulation rate, or for any of the other reproductively related traits studied by Cassady et al. (2001), on SSC16. None of the other studies searching for QTL affecting ovulation rate or litter size show any QTL on SSC16 (e.g., Milan et al., 1998; Rohrer et al., 1999), suggesting that the line difference in expression of FST is a response to alterations in other genes. Identification of the trans-acting QTL responsible for expression changes in FST would greatly enhance our understanding of the underlying genetic control of reproduction.
Higher expression of the NR4A1 gene was also confirmed in Line I. Nuclear receptor subfamily 4, group A, member 1 is a member of the steroid receptor superfamily with 20% amino acid sequence similarity to estrogen receptor (Chang et al., 1989). Steroid hormones, especially estrogen, are locally produced and critical for follicular development. Increased gene expression of a nuclear steroid hormone receptor in Line I may indicate that enhanced ovulation rate is in response to increased sensitivity to steroid hormones.
At the time this study was conducted, ddPCR was considered the most powerful tool available for a genomewide investigation of differential gene expression. This is because it does not rely on any previous knowledge of expressed genes in the tissue of interest (Wang et al., 1999). However, the small number of genes identified and confirmed by ddPCR as having differential gene expression indicates that this method may not be sensitive enough to detect the slight changes in gene expression that may be occurring in selection lines. In addition, experimentation by Ledakis et al. (1998) showed that ddPCR might be covering only a small proportion of the total mRNA population. Furthermore, ddPCR is prone to a high incidence of false-positive results (Ledakis et al., 1998; Matz and Lukyanov, 1998). We utilized multiple lanes of the same treatment run together in order to try to identify false-positives; yet, over 50% of the Northern hybridizations did not confirm the ddPCR result in this experiment. Given the status of EST identification in the pig and the ability to generate pig-specific microarrays, the methods used in the current study (ddPCR and cross-species hybridization experiments) are no longer attractive options for analysis of the pig transcriptome. For these reasons, further confirmation of the many additional putatively differentially expressed genes that were identified in both the ddPCR and the cross-species microarray analyses, as well as an enhanced large-scale expression analysis, are currently being conducted using microarrays based on the pig ovarian-follicle cDNA normalized library that we have developed (Caetano et al., 2003). This array-based approach will also rectify the primary shortcoming of the present study by simultaneously enabling significant biological and experimental replication.
Implications
Differential display PCR and cross-species microarray hybridization were used to analyze gene expression differences between pigs selected for increased ovulation rate and embryo survival and pigs of a randomly selected control line. Our results indicate that a potentially large number of gene expression changes may be involved in long-term selection response for a complex trait, such as reproduction, but that the magnitude of these changes may be relatively small. Evaluation of gene expression changes within the context of selection line analysis may reveal direct results of selection in the transcriptome due to allelic variation in the base population, but more likely represent trans-acting correlated responses to selection within a relevant pathway. Highly sensitive and thorough transcriptional analysis will be required for examining selection response and helping to understand the genetic architecture of complex traits with low heritability.
Footnotes
1 Published as paper No. 14021, Journal Series, Nebraska Agric. Research Division, Univ. of Nebraska, Lincoln 68583-0908. 
2 The authors thank M. Allan, L. Messer, and S. Hassler for their help in data collection; D. Cheleen and D. Aherin for experimental animal care assistance; C. Schrock for assistance in processing pigs; and J. Ford and G. Rohrer for providing peer review on an earlier version of this manuscript. Support for initial work with the UniGEM V1.0 cross-species hybridization was provided through a grant to C. Tuggle from the Office of the Provost, Iowa State University.
3 Scholarship funded by CAPES, Brazil. Researcher of Embrapa Swine and Poultry, Brazil.
4 Correspondence: A218 Animal Sciences (phone: 402-472-6416; fax: 402-472-6362; e-mail: dpomp{at}unl.edu).
Received for publication March 25, 2003. Accepted for publication August 8, 2003.
Literature Cited
Allan, M. F., M. K. Nielsen, and D. Pomp. 2000. Gene expression in hypothalamus and brown adipose tissue of mice divergently selected for heat loss. Physiol. Genomics 3:149–156.
Caetano, A. R., R. K. Johnson, and D. Pomp. 2003. Generation and sequence characterization of a normalized cDNA library from swine ovarian follicles. Mamm. Genome 14:65–70.[Medline]
Cassady, J. P., R. K. Johnson, D. Pomp, G. A. Rohrer, L. D. Van Vleck, E. K. Spiegel, and K. M. Gilson. 2001. Identification of quantitative trait loci affecting reproduction in pigs. J. Anim. Sci. 79:623–633.
Chang, C., J. Kokontis, S. S. Liao, and Y. Chang. 1989. Isolation and characterization of human TR3 receptor: a member of steroid receptor superfamily. J. Steroid Biochem. 34:391–395.[Medline]
Ellegren, H. 1993. Abundant (A)n. (T)n mononucleotide repeats in the pig genome: linkage mapping of the porcine APOB, FSA, ALOX12, PEPN and RLN loci. Anim. Genet. 24:367–372.[Medline]
Findlay, J. K. 1993. An update on the roles of inhibin, activin, and follistatin as local regulators of folliculogenesis. Biol. Reprod. 48:15–23.[Abstract]
Garrett, W. M., and H. D. Guthrie. 1996. Expression of androgen receptors and steroidogenic enzymes in relation to follicular growth and atresia following ovulation in pigs. Biol. Reprod. 55:949–955.[Abstract]
Hasegawa, Y., Y. Eto, Y. Ibuki, and H. Sugino. 1994. Activin as autocrine and paracrine factor in the ovary. Horm. Res. 41(Suppl. 1):55–62.
Johnson, R. K., M. K. Nielsen, and D. S. Casey. 1999. Responses in ovulation rate, embryonal survival, and litter traits in swine to 14 generations of selection to increase litter size. J. Anim. Sci. 77:541–557.
Koohmaraie, M. 1992. The role of Ca(2+)-dependent proteases (calpains) in post mortem proteolysis and meat tenderness. Biochimie 74:239–245.[Medline]
Ledakis, P., H. Tanimura, and T. Fojo. 1998. Limitations of differential display. Biochem. Biophys. Res. Commun. 251:653–656.[Medline]
Li, M. D., L. V. DePaolo, and J. J. Ford. 1997. Expression of follistatin and inhibin/activin subunit genes in porcine follicles. Biol. Reprod. 57:112–118.[Abstract]
Linville, R. C., D. Pomp, R. K. Johnson, and M. F. Rothschild. 2001. Candidate gene analysis for loci affecting litter size and ovulation rate in swine. J. Anim. Sci. 79:60–67.
Lo, K. W., Q. Zhang, M. Li, and M. Zhang. 1999. Apoptosis-linked gene product ALG-2 is a new member of the calpain small subunit subfamily of Ca2+-binding proteins. Biochemistry 38:7498–7508.[Medline]
Matz, M. V., and S. A. Lukyanov. 1998. Different strategies of differential display: areas of application. Nucleic Acids Res. 26:5537–5543.
Mellgren, R. L., Q. Lu, W. Zhang, M. Lakkis, E. Shaw, and M. T. Mericle. 1996. Isolation of a Chinese hamster ovary cell clone possessing decreased mu-calpain content and a reduced proliferative growth rate. J. Biol. Chem. 271:15568–15574.
Milan, D., J. P. Bidanel, P. Le Roy, C. Chevalet, N. Woloszyn, J. C. Caritez, J. Gruand, H. Lagant, M. Bonneau, L. Lefaucheur, C. Renard, M. Vaiman, P. Mormède, C. Désautés, Y. Amigues, F. Bourgeois, J. Gellin, and L. Ollivier. 1998. Current status of QTL detection in Large White X Meishan crosses in France. Proc. 6th World Cong. Genet. Appl. Livest. Prod., Armidale, Australia 26:414–417.
Moody, D. E., Z. Zou, and L. McIntyre. 2002. Cross-species hybridisation of pig RNA to human nylon microarrays. BMC Genomics 3:27–37.[Medline]
Neal, S. M., R. K. Johnson, and R. J. Kittok. 1989. Index selection for components of litter size in swine: response to five generations of selection. J. Anim. Sci. 67:1933–1945.
Parr, T., P. L. Sensky, G. P. Scothern, R. G. Bardsley, P. J. Buttery, J. D. Wood, and C. Warkup. 1999. Relationship between skeletal muscle-specific calpain and tenderness of conditioned porcine longissimus muscle. J. Anim. Sci. 77:661–668.
Rohrer, G. A., J. J. Ford, T. H. Wise, J. L. Vallet, and R. K. Christenson. 1999. Identification of quantitative trait loci affecting female reproductive traits in a multigeneration Meishan-White composite swine population. J. Anim. Sci. 77:1385–1391.
Ruíz-Flores, A., and R. K. Johnson. 2001. Direct and correlated responses to two-stage selection for ovulation rate and number of fully formed pigs at birth in swine. J. Anim. Sci. 79:2286–2297.
Wang, K., L. Gan, E. Jeffery, M. Gayle, A. M. Gown, M. Skelly, P. S. Nelson, W. V. Ng, M. Schummer, L. Hood, and J. Mulligan. 1999. Monitoring gene expression profile changes in ovarian carcinomas using cDNA microarray. Gene 229:101–108.[Medline]
Wilkie, P. J., A. A. Paszek, C. W. Beattie, L. J. Alexander, M. B. Wheeler, and L. B. Schook. 1999. A genomic scan of porcine reproductive traits reveals possible quantitative trait loci (QTLs) for number of corpora lutea. Mamm. Genome 10:573–578.[Medline]
Yen, H.-W. 1999. Follicular development, maturation and atresia during the estrous cycle in gilts expressing high and low ovulation rates. Ph.D. Diss., Univ. Nebraska, Lincoln.
Yuan, W., M. C. Lucy, and M. F. Smith. 1996. Messenger ribonucleic acid for insulin-like growth factors-I and -II, insulin-like growth factor-binding protein-2, gonadotropin receptors, and steroidogenic enzymes in porcine follicles. Biol. Reprod. 55:1045–1054.[Abstract]
This article has been cited by other articles:
|
|
 |
|
  C. S. Lin and C. W. Hsu Differentially transcribed genes in skeletal muscle of Duroc and Taoyuan pigs J Anim Sci, September 1, 2005; 83(9): 2075 - 2086. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. R. Caetano, R. K. Johnson, J. J. Ford, and D. Pomp Microarray Profiling for Differential Gene Expression in Ovaries and Ovarian Follicles of Pigs Selected for Increased Ovulation Rate Genetics, November 1, 2004; 168(3): 1529 - 1537. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. R. Bertani, C. D. Gladney, R. K. Johnson, and D. Pomp Evaluation of gene expression in pigs selected for enhanced reproduction using differential display PCR: II. Anterior pituitary J Anim Sci, January 1, 2004; 82(1): 32 - 40. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
