Expression of PEG11 and PEG11AS transcripts in normal and callipyge sheep
© Bidwell et al. 2004
Received: 17 March 2004
Accepted: 06 August 2004
Published: 06 August 2004
Skip to main content
© Bidwell et al. 2004
Received: 17 March 2004
Accepted: 06 August 2004
Published: 06 August 2004
The callipyge mutation is located within an imprinted gene cluster on ovine chromosome 18. The callipyge trait exhibits polar overdominant inheritance due to the fact that only heterozygotes inheriting a mutant paternal allele (paternal heterozygotes) have a phenotype of muscle hypertrophy, reduced fat and a more compact skeleton. The mutation is a single A to G transition in an intergenic region that results in the increased expression of several genes within the imprinted cluster without changing their parent-of-origin allele-specific expression.
There was a significant effect of genotype (p < 0.0001) on the transcript abundance of DLK1, PEG11, and MEG8 in the muscles of lambs with the callipyge allele. DLK1 and PEG11 transcript levels were elevated in the hypertrophied muscles of paternal heterozygous animals relative to animals of the other three genotypes. The PEG11 locus produces a single 6.5 kb transcript and two smaller antisense strand transcripts, referred to as PEG11AS, in skeletal muscle. PEG11AS transcripts were detectable over a 5.5 kb region beginning 1.2 kb upstream of the PEG11 start codon and spanning the entire open reading frame. Analysis of PEG11 expression by quantitative PCR shows a 200-fold induction in the hypertrophied muscles of paternal heterozygous animals and a 13-fold induction in homozygous callipyge animals. PEG11 transcripts were 14-fold more abundant than PEG11AS transcripts in the gluteus medius of paternal heterozygous animals. PEG11AS transcripts were expressed at higher levels than PEG11 transcripts in the gluteus medius of animals of the other three genotypes.
The effect of the callipyge mutation has been to alter the expression of DLK1, GTL2, PEG11 and MEG8 in the hypertrophied skeletal muscles. Transcript abundance of DLK1 and PEG11 was highest in paternal heterozygous animals and exhibited polar overdominant gene expression patterns; therefore, both genes are candidates for causing skeletal muscle hypertrophy. There was unique relationship of PEG11 and PEG11AS transcript abundance in the paternal heterozygous animals that suggests a RNA interference mechanism may have a role in PEG11 gene regulation and polar overdominance in callipyge sheep.
The mutation responsible for the callipyge trait is located within an imprinted gene cluster on the distal end of ovine chromosome 18 [1–3]. The callipyge phenotype is associated with an altered carcass composition including a 30–40% increase in muscle mass, a 6–7% decrease in carcass fat, decreased organ weights, and a more compact skeleton, all without a net affect on animal growth [4–7]. There is a pronounced hypertrophy of muscles in the loin and pelvic limbs and a lesser degree of hypertrophy in muscles of the thoracic limbs [6–8]. The callipyge phenotype is inherited in a non-Mendelian mode termed polar overdominance [9, 10] in which only animals that inherit a normal allele (wild type; +) from the dam and the mutant callipyge allele from the sire (CLPG Pat ) exhibit the callipyge phenotype. Maternal heterozygotes (CLPG Mat /+ Pat ) and callipyge allele homozygotes (CLPG Mat /CLPG Pat ) have muscling and carcass compositions that are similar to normal sheep (wild type homozygotes; +/+).
The causative mutation for callipyge is a single base transition of A (wild type; +) to G (CLPG) in the intergenic region located between DLK1 and GTL2  (Figure 1). This mutation has been shown to be 100% concordant with all animals of the + Mat /CLPG Pat genotype based on haplotype analysis . Analysis of sheep from 19 different breeds as well as 13 mammalian species revealed a highly conserved 12 base sequence that includes the single nucleotide polymorphism . The G polymorphism is unique to direct descendents of the first known callipyge animal, a ram named "Solid Gold". This animal was mosaic for the mutation , providing strong evidence that this single nucleotide polymorphism is the causative mutation. Initial results indicate that the mutation alters the expression of several of the genes within the imprinted cluster [25, 26] when they are inherited in cis without altering their parent-of-origin-specific expression .
In this study, we analyzed the expression of five genes within the callipyge cluster in the muscles of lambs of all four genotypes. Quantitative analysis of gene expression using a series of orthogonal contrasts showed that DLK1, PEG11 and MEG8 exhibited a polar overdominant pattern of gene expression. The expression of PEG11 and PEG11AS transcripts in the muscles of paternal heterozygous callipyge lambs (+ Mat /CLPG Pat ) was different from the other three genotypes. A sense/antisense interaction of PEG11 and PEG11AS, such as an RNA interference mechanism, would be consistent with a trans interaction between reciprocally imprinted genes that has been previously proposed as a mechanism for polar overdominance [27, 28].
Two probes from upstream of the PEG11 coding sequence (riboprobes F and G) did not detect any PEG11 transcripts, indicating that transcription of the 6.5 kb PEG11 was initiated within 350 bp of the start codon of the open reading frame. Riboprobe F detected PEG11AS transcripts in the gluteus medius and supraspinatus of all four genotypes, with the lowest expression in the paternal heterozygous (+ Mat /CLPG Pat ) animals. The PEG11AS riboprobe G shows three protected fragments, suggesting variable splice junctions or transcription termination sites for PEG11AS transcripts.
Statistical analysis of gene expression by quantitative PCR
Orthogonal Contrast P-Values
Effect of Genotype
The effect of the callipyge mutation on DLK1 and PEG11 expression in the supraspinatus was different (Table 1). No differences in DLK1 transcript abundance in the supraspinatus were found among the four genotypes (Figure 4C). Although not statistically analyzed, the level of DLK1 expression was similar between the gluteus medius and supraspinatus (Figure 4A and 4C). PEG11 expression in the supraspinatus had a different genotype-specific pattern than in the gluteus medius. PEG11 expression was elevated 150-fold in + Mat /CLPG Pat animals (p < 0.05) but was not significantly changed in the other three genotypes (Figure 4D). The transcript abundance of PEG11 in supraspinatus was generally much lower than in gluteus medius of the same genotype. PEG11 expression in the supraspinatus was 12-fold lower than in the gluteus medius of paternal heterozygous animals (Figure 4B and 4D).
Maternal inheritance of the callipyge allele significantly altered the expression of MEG8 but not PEG11AS in the gluteus medius (Table 1). Transcript abundance of MEG8 in gluteus medius was 6-fold greater (p < 0.05) in the CLPG Mat /+ Pat and CLPG Mat /CLPG Pat animals relative to the + Mat /CLPG Pat and +/+ animals (Figure 4A). Expression of MEG8 in the supraspinatus was also affected by the callipyge mutation but to a lesser degree (Figure 4C). The homozygous callipyge animals had significantly higher MEG8 transcript abundance than paternal heterozygotes, but those two genotypes were not significantly different from maternal heterozygotes and homozygous wild type lambs.
Orthogonal contrasts were used to analyze different models of gene action for the genes that had a significant effect for genotype (Table 1). The polar overdominance contrast was significant for transcript abundances of DLK1, PEG11 and MEG8 in gluteus medius and for PEG11 and MEG8 in the supraspinatus. The additive contrasts were also significant for DLK1, PEG11 and MEG8 in gluteus medius but were only significant for PEG11 in supraspinatus. The maternal dominance contrast was significant for DLK1 and PEG11 gluteus medius.
Our results show a clear pattern of increased expression for genes within the imprinted callipyge cluster in muscles that become hypertrophied, whereas expression of these genes was either reduced or absent in a muscle that does not become hypertrophied. The increased gene expression occurred when the mutation was inherited in cis and was dependent on each gene's imprinting status, consistent with previous reports [25, 26]. These results support the hypothesis that the mutation has disrupted a long range control element . Two paternally expressed genes, DLK1 and PEG11, had significantly increased transcript abundance when a callipyge allele was inherited from the sire. One maternally expressed gene, MEG8, showed significantly increased transcript abundance when the callipyge allele was inherited from the dam. The changes in gene expression were sustained until 12 weeks of age, when the differences in growth and body composition between callipyge and normal lamb are established and are subsequently maintained .
Northern blot analysis suggests that expression of GTL2 and PEG11AS was increased by maternal inheritance of the callipyge allele. Quantitative analysis was not done for GTL2 in this study due to the expression of numerous alternatively spliced transcripts that have been reported for mice and sheep [26, 29]. The 2.4 kb GTL2 mRNA seen in this study consisted of a heterogeneous population of alternatively spliced mRNAs. Similarly, two PEG11AS transcripts were detected over a 5.6 kb area extending from beyond the 5' end of the PEG11 transcript to the 3'UTR. The 1.7 and 0.8 kb PEG11AS transcripts detected by northern blot analysis using probe P would not be protected by probes C or G unless the PEG11AS transcripts undergo intron splicing or there are other transcripts that were not detected in the northern blots. Therefore, the effect of the callipyge mutation on GTL2 and PEG11AS will require a more extensive analysis to fully elucidate their expression patterns in the four genotypes and determine their role in the callipyge model.
Due to their paternal allele-specific expression, the DLK1 and PEG11 genes are both candidates for an effector gene that is responsible for the skeletal muscle hypertrophy exhibited by paternal heterozygous (+ Mat /CLPG Pat ) animals. In this study, both genes showed a polar overdominant expression pattern in that paternal heterozygotes had significantly higher levels of gene expression than the other three genotypes. The major differences between DLK1 and PEG11 were the magnitude and the muscle specificity of the up-regulation. DLK1 was readily detectable in all muscles and genotypes by northern blot, but the up-regulation in paternal heterozygous animals was restricted to muscles of the loin and pelvic limb. The quantitative results showed a 6-fold increase in DLK1 transcript abundance in the gluteus medius and no change of DLK1 abundance in the supraspinatus. This pattern of gene expression is consistent with studies on individual muscle growth that show significant muscle hypertrophy in the gluteus medius but not in the supraspinatus [6, 8]. DLK1 transcripts were significantly increased (2.5-fold) in the gluteus medius of CLPG Mat /CLPG Pat animals, which do not exhibit muscle hypertrophy. The lack of a phenotype in CLPG Mat /CLPG Pat animals could be due to a threshold effect that requires more than a 2.5-fold increase in DLK1 transcript abundance to change muscle growth.
The expression of PEG11 was very low in the gluteus medius and supraspinatus of normal sheep and was induced in both muscles of callipyge lambs. Northern blot analysis, ribonuclease protection assay and quantitative PCR results all show that the expression of PEG11 and PEG11AS transcripts was much lower in the supraspinatus than the other muscles. The high level of expression of PEG11 in the gluteus medius relative to the supraspinatus of callipyge lambs indicates that PEG11 could also be the gene responsible for muscle hypertrophy if there is a threshold level required to change muscle growth. The PEG11 gene has a long intronless open reading frame but it is not known if a protein is produced or what function it may have.
The overdominant nature of the callipyge phenotype through the lack of muscle hypertrophy in animals with the CLPG Mat /CLPG Pat genotype is one of the more intriguing aspects of the trait and has led to a hypothesis of trans effects by other reciprocally imprinted genes in the callipyge region [25, 27]. If the PEG11 gene has a direct role in muscle hypertrophy, either alone or in concert with DLK1, then PEG11AS may have a role in a trans effect on PEG11 expression. In + Mat /CLPG Pat animals, PEG11 transcripts were 14-fold and 4-fold more abundant than PEG11AS transcripts in the gluteus medius and supraspinatus respectively. PEG11AS transcripts were more abundant than PEG11 transcripts in the other three genotypes for both muscles. Therefore, the relative abundance of PEG11 transcripts to PEG11AS transcripts was unique in paternal heterozygous animals. MicroRNAs are a central component of RNA interference mechanisms. In the mouse, two antisense microRNA have been identified for the orthologous rtl1 locus . RNA interference mechanisms have been shown both to repress transcription by inducing heterochromatin formation [30–32] and to cause post-transcriptional silencing through nuclear retention or targeted degradation [33–35]. MicroRNAs may be produced from post-transcriptional processing of PEG11AS RNA and be involved in normal regulation of the locus and in generating overdominance . Expression of PEG11AS may normally cause repression of the paternal PEG11 locus since very little PEG11 mRNA was detectable in the muscles of normal animals. In the animals with a paternally inherited callipyge allele, (+ Mat /CLPG Pat and CLPG Mat /CLPG Pat ), the normal repression of the PEG11 locus has been disrupted by the mutation in a putative long range control element . PEG11 transcripts only accumulate in significant excess of PEG11AS in the paternal heterozygous animals, whereas in the CLPG Mat /CLPG Pat animals the PEG11AS transcripts remain in excess and may prevent the accumulation of the threshold level of PEG11 mRNA required to produce a muscle hypertrophy phenotype. Although the expression of PEG11AS was not affected by genotype in quantitative PCR, the northern blots and transcript mapping indicate there are multiple transcripts that are likely to undergo different intron splicing. Further analysis of PEG11AS expression will be necessary to determine its role in the PEG11 locus regulation.
The effect of the callipyge mutation has been to increase transcript abundance of four genes, DLK1, GTL2, PEG11 and MEG8, within the imprinted cluster in skeletal muscles that become hypertrophied. The increase in transcript abundance was consistent with each gene's parental allele-specific expression. The DLK1 and PEG11 genes were both expressed at their highest levels in paternal heterozygous animals and exhibited polar overdominant gene expression patterns. Therefore, both genes are candidates for causing muscle hypertrophy. DLK1 expression was only elevated in muscles that undergo hypertrophy, so its muscle-specific increase was consistent with the callipyge phenotype. PEG11 was 12-fold more abundant in hypertrophied muscle than non-hypertrophied muscle and only paternal heterozygous animals had PEG11 transcript levels in excess of PEG11AS transcript levels. The unique relationship of PEG11 and PEG11AS in paternal heterozygous animals suggests that an RNA interference mechanism may have a role in regulating the PEG11 locus and polar overdominance in callipyge sheep.
A series of planned matings were conducted to produce the four possible callipyge genotypes. The genotypes of all lambs were verified using the single nucleotide polymorphism [23, 24] and several markers that flank the callipyge region on chromosome 18. Lambs were slaughtered in accordance with humane practices approved by the Utah State University Institutional Animal Care and Use Committee. Samples were collected from the longissimus dorsi, semimembranosus, gluteus medius and supraspinatus, and preserved in RNAlater (Ambion Inc., Woodlands, TX USA). The tissue samples were homogenized in 4 M guanidinium thiocyanate, 25 mM sodium citrate, 50 mM EDTA, 1% sodium-N-lauroyl-sarcosine and total RNA was sedimented by ultracentrifugation of the homogenate on a cushion of 5.7 M CsCl, 50 mM EDTA . Purified RNA was quantified by spectrophotometry and the use of a constant mass of RNA for each quantitative assay was based on absorbance at 260 nm. Total RNA was treated with DNase I using DNA free™ reagents (Ambion Inc.) to remove trace genomic DNA prior to use in the ribonuclease protection assay and quantitative PCR.
Northern blots were prepared using denaturing formaldehyde gel electrophoresis (NorthernMax™; Ambion Inc) of 10 μg of total RNA and transferred to positively charged nylon membranes using standard methods . Primer sequences used to amplify probes from the callipyge region are given in Charlier et al. , and the PCR products were verified by DNA sequencing. Strand specific DNA probes were synthesized by 40 cycles of asymmetric PCR with Strip-EZ™ nucleotides (Ambion Inc.) and 50 μCi of α-[32P]-dATP (Amersham-Pharmacia, Piscataway, NJ USA). Unincorporated nucleotides were removed by spin column chromatography (BioSpin P30; Bio-Rad Inc., Hercules, CA USA). The probes were hybridized to the membranes without denaturation using Ultrahyb™ (Ambion Inc.) at 42°C overnight. After hybridization, the membranes were washed in 2X SSC (0.3 M sodium chloride, 0.03 M sodium citrate)/0.5% SDS followed by 3 washes in 1X SSC/ 0.1% SDS at 65°C for 30 min and a final high stringency wash in 0.1X SSC/ 0.1%SDS at 65°C for 30 min. The northern blots were exposed to Kodak XAR autoradiography film for 18 to 72 h at -80°C. After autoradiography, the probe was degraded and removed from the membranes using Strip-EZ™ reagents (Ambion Inc.).
Primer sequences for ribonuclease protection assay probes and quantitative PCR
Statistical methods for genetic models
Orthogonal Contrast Values by Genotype
CLPG Mat /+ Pat
+ Mat /CLPG Pat
CLPG Mat /CLPG Pat
Paternal Polar Overdominance
Number of Animals
This project was supported by funds from the USDA/NRI (2002-35205-11602) and the Purdue University Agricultural Research Program (paper No. 17430). The authors acknowledge Michel Georges and Carole Charlier, University of Liège, Belgium, for providing DNA sequence and PCR primer data and thank Christine Bynum, Anne Danielsson and David Forrester for their assistance with this experiment.
This article is published under license to BioMed Central Ltd. This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.