Dual effect of a single nucleotide polymorphism in the first intron of the porcine Secreted phosphoprotein 1 gene: allele-specific binding of C/EBP beta and activation of aberrant splicing
- 5.9k Downloads
Secreted phosphoprotein 1 (SPP1 or Osteopontin, OPN) is a multifunctional matricellular glycoprotein involved in development and regeneration of skeletal muscle. Previously, we have demonstrated that porcine SPP1 shows breed-related differential mRNA expression during myogenesis. With the aim to identify putative contributing cis-regulatory DNA variation we resequenced the 5' upstream region of the gene in the respective breeds Pietrain and Duroc. We found two single nucleotide polymorphisms (SNP; [GenBank:M84121]: g.1804C>T and g.3836A>G). We focused our investigation on the SNP g.3836A>G, because in silico analysis and knowledge about the regulation of SPP1 suggested an effect of this SNP on a CCAAT/enhancer binding protein beta (C/EBPβ) responsive transcriptional enhancer.
Using electrophoretic mobility shift assay we demonstrated that, similar to human SPP1, the 3' terminal end of the first intron of porcine SPP1 harbors a C/EBPβ binding site and showed that this binding site is negatively affected by the mutant G allele. Genotyping of 48 fetuses per breed revealed that the G allele segregated exclusively in Duroc fetuses with a frequency of 57 percent. Using real-time quantitative PCR we showed that, consistent with its negative effect on a transcriptional enhancer element, the G allele tends to decrease mRNA abundance of SPP1 in the fetal musculus longissimus dorsi (~1.3 fold; P ≥ 0.1).
Moreover, we showed that the SNP g.3836A>G leads to ubiquitous aberrant splicing of the first intron by generating a de novo and activating a cryptic splice acceptor site. Aberrantly spliced transcripts comprise about half of the SPP1 messages expressed by the G allele. Both aberrant splice variants differ from the native transcript by insertions in the leader sequences which do not change the reading frame of SPP1.
At the 3' terminal end of the first intron of the porcine SPP1 we identified a unique, dually functional SNP g.3836A>G. This SNP affects the function of the SPP1 gene at the DNA level by affecting a C/EBPβ binding site and at the RNA level by activating aberrant splicing of the first intron, and thus represents an interesting DNA-marker to study phenotypic effects of SPP1 DNA-variation.
KeywordsSplice Variant Aberrant Splice Transcriptional Enhancer Unlabelled Probe Musculus Longissimus Dorsi
Secreted phosphoprotein 1 (SPP1 or Osteopontin, OPN) is a matricellular glycoprotein mediating cell-adhesion, -migration, -survival and -signalling via integrin and CD44 receptors . In line with its versatility and widespread expression SPP1 has been linked with various physiological and pathological events, amongst others development and regeneration of skeletal muscle [2, 3]. SPP1 is regulated by the muscle regulatory factors MYOD and MYF5  and has been shown to be expressed in vitro in myoblasts and myotubes [3, 5]. Uaesoontrachoon et al.  demonstrated that soluble SPP1 stimulates proliferation of myoblasts whereas SPP1 deposited in extracellular matrix promotes their differentiation. Recently we showed upregulation of SPP1 mRNA expression in prenatal musculus longissimus dorsi (M.l.d.) at 35 and 63-77 days post conception (dpc) in pigs, i.e. at the time points of the first and second myogenic wave respectively, thus providing additional, in vivo, evidence supporting involvement of SPP1 in myogenesis. In addition we showed consistent elevation of SPP1 mRNA level during myogenesis in the pig breed Duroc compared to breed Pietrain, the latter showing higher muscularity and higher proportion of fast twitch glycolytic fibers postnatally . Moreover we mapped quantitative trait loci for the proportion of fast twitch glycolytic fibers and pH of M.l.d. 45 minutes post mortem on porcine chromosome 8 close to SPP1 position . To detect polymorphisms that contribute to the breed-related differential mRNA expression of SPP1 and which are associated with differential microstructural and biophysical muscle properties we performed SNP screening of the upstream regulatory region. We identified and functionally characterized a SNP located in the 3' terminal end of the first intron in an evolutionarily conserved transcriptional enhancer [8, 9].
Identification and in silico characterization of SNP g.3836A>G
The C/EBPβ binding site at the 3' terminal end of the first intron of SPP1 is evolutionarily conserved in the pig and is negatively affected by the SNP g.3836A>G
SNP g.3836A>G decreases mRNA abundance of SPP1 in fetal M.l.d
SNP g.3836A>G leads to ubiquitous aberrant splicing of porcine SPP1 in the 5' untranslated region
The total number of myofibers and to some extent their metabolic and contractile properties are determined during the process of myogenesis. Insights into the regulation of this process will consequently bring about better knowledge of factors affecting postnatal growth and function of skeletal muscle. We previously discovered breed-related differences in mRNA expression of porcine SPP1 during myogenesis  and hypothesized that this variation might be caused by cis-regulatory DNA-variation. Cis-regulatory DNA variation has been shown to affect a large proportion (~20%) of genes  and in fact has already been described in human SPP1 [13, 14]. We identified an A>G SNP at position g.3836 at the 3' terminal end of the first intron of porcine SPP1. In silico analysis of the polymorphic sequence and knowledge about the regulation of SPP1 in pigs and human suggested an effect of the SNP on a C/EBPβ responsive transcriptional enhancer. Using competitive EMSA and supershift assay we showed that the C/EBPβ binding site found in human is present also in the pig and is negatively affected by the mutation. As a consequence of the reduced affinity of the C/EBPβ binding site the activity of the transcriptional enhancer is reduced, at least in fetal M.l.d., as indicated by the downregulation of SPP1 mRNA expression by fetuses carrying the G allele. Conservation of the C/EBPβ binding site at the 3' terminal end of the first intron of human and porcine SPP1 points to an important function of this element in the regulation of SPP1. The relevance of C/EBPβ in positive regulation of mRNA expression of SPP1 is further emphasized by the identification of additional functional C/EBP transcriptional enhancer elements in the promoter of murine and human SPP1 [15, 16]. C/EBPs are a family of transcription factors involved in the regulation of proliferation and differentiation of diverse cell types, and play pivotal roles in a number of processes including adipogenesis . However, an involvement or a precise function of C/EBPs in myogenesis has not been described so far.
We demonstrate that the SNP g.3836A>G is functional also on the RNA level and causes aberrant splicing of the first intron. We found that this effect is ubiquitous with regard to temporo-spatial distribution which is in accord with the fact that it creates a general splicing signal. Naturally occurring DNA-variation affecting splicing represents a valuable resource to identify sequence signals involved in the regulation of this process. The g.3836A>G SNP exhibits several features of de novo 3'ss activating aberrant splicing identified by Kralovicova et al.  and Vorechovsky . It creates an AG dinucleotide in intron in polypyrimidine tract of the authentic 3'ss by introducing a guanine residue, has an uridine in position +1 relative to the de novo 3'ss and activates a cryptic 3'ss. Vorechovsky  has shown that authentic counterparts of de novo 3'ss are intrinsically weak. In the case of the affected 3'ss of SPP1 this is likely a consequence of the two concurrent functional constraints on the sequence of the 3' terminal end of the first intron; on one hand to provide a C/EBPβ responsive transcriptional enhancer and on the other to provide the canonical splicing acceptor signals. The in silico predicted strength of the mutated authentic 3'ss compared to the cryptic 3'ss and their utilization in vivo are reversed. An explanation for the lower utilization of the cryptic 3'ss compared to the mutated authentic 3'ss in vivo may be a suboptimal sequence of the putative cryptic branch site, in contrast to the predicted authentic branch site which perfectly matches with the consensus sequence YNYURAY (Figure 5B). Another explanation might be that the cryptic 3'ss is silenced by a splicing silencer or vice versa that the authentic 3'ss might be augmented by a splicing enhancer. Kralovicova and Vorechovsky  showed that in vivo selection of aberrant splice sites is extensively controlled by auxiliary splicing signals.
Mutations that affect hnRNA splicing account for up to 50% of disease-causing gene alterations in human and potentially represent the most frequent cause of hereditary disorders. Some 218 unique aberrant 3'ss, activated by disease-causing mutations in 131 genes, are presently known in human . Our results show that about half of the SPP1 messages expressed by the G allele comprise aberrantly spliced transcripts. Considering the versatile function of SPP1 and the ubiquitous expression of the aberrant splice variants it could be speculated that the SNP g.3836A>G might have pleiotropic effects on various traits including growth, reproduction and immune defence. However the relatively high frequency of the G allele argues against a major negative phenotypic effect, because in that case the G allele would be quickly eliminated given the high selection pressure on commercial pigs. Furthermore, the aberrant splicing induced by the SNP g.3836A>G does not change primary structure of the SPP1 protein. In fact, the aberrant splicing may counteract the negative effect of the G allele on mRNA expression of SPP1 by enhancing translational efficiency or RNA stability. Aberrantly spliced mRNA isoforms of the human insulin gene with longer 5'-leader sequence, induced by a SNP in its first intron, were reported to generate more proinsulin in vitro compared to the native transcripts .
Functional characterization of the SNP g.3836A>G revealed that it has two effects; it negatively affects a C/EBPβ binding site and activates aberrant splicing of the first intron. However, although the SNP g.3836A>G is functional it most likely does not represent the causative mutation responsible for the previously observed breed related differences in SPP1 mRNA expression. Nevertheless, the dual effect of the SNP on SPP1 function renders it as an interesting DNA-marker for association studies concerning muscle-related, growth, reproduction, and immune defence traits.
Identification of SNP and splice variants
The target sequence of the porcine SPP1 gene [GenBank:M84121] was amplified in six overlapping PCR fragments using standard PCR conditions and each six Pietrain and Duroc DNA samples respectively. A standard PCR reaction mixture contained 100 ng genomic DNA, 0.2 μM of each primer, 50 μM of each dNTP and 0.5 U SupraTherm Taq Polymerase in 1× supplied PCR-buffer containing 1.5 mM MgCl2 (Genecraft, Lüdinghausen, Germany). The temperature profile consisted of 40 cycles of denaturation at 95°C for 15 s, annealing at Ta for 30 s and extension at 72°C for 30 s for each <0.5 kb. Amplification products were subsequently pooled within breed and purified using the NucleoSpin Extract II kit (Macherey-Nagel, Düren, Germany).
To detect alternative splice variants a cDNA fragment spanning exons 1-6 was amplified in a standard PCR reaction and cloned using the pGEM-T vector (Promega, Mannheim, Germany). The PCR products and plasmids were sequenced using Big Dye Terminator Cycle sequencing kit V1.1 (Applied Biosystems, Darmstadt, Germany) and analyzed on ABI 310 or MegaBACE 750 automated sequencer.
The g.3836A>G SNP was amplified in a standard PCR reaction and genotyped using single strand conformation polymorphism (SSCP) visualized by silver staining after electrophoresis performed for 5 hours at 5°C on a 12% native polyacrylamide (49:1 AA: Bis) gel in 0.5 × TBE buffer. Sequences and annealing temperature (Ta) of primers used for comparative sequencing, genotyping and RT-PCR are given in Additional File 1: Table S1.
Tissue collection, RNA isolation and cDNA synthesis
Sampling of fetal M.l.d. was described in detail previously . Briefly, immediately after exsanguination of the sows the uteri were recovered and the embryos/fetuses were quickly removed, weighed and the M.l.d. dissected. Adult tissue samples from performance tested animals were collected in our experimental abattoir. After dissection samples were quickly frozen in liquid nitrogen and stored at -80°C for later analysis. Total RNA was isolated using TRI Reagent (Sigma, Taufkirchen, Germany). After DNaseI treatment (Roche, Mannheim, Germany) the RNA was cleaned up using the NucleoSpin RNA II Kit (Macherey-Nagel). First strand cDNA was synthesized using SuperScriptIII MMLV reverse transcriptase (Invitrogen, Karlsruhe, Germany) in a reaction containing 1 μg RNA and 500 ng of oligo (dT)11VN primer, according to the manufacturer's protocol.
Quantification of total transcript level and relative amounts of splice variants
Total transcript level of SPP1 and of the reference gene RPL32 were quantified by real-time quantitative PCR (qPCR) performed on a LightCycler 1.0 System using the LightCycler FastStart DNA Master SYBRplus Green I (Roche). The amplification was conducted in duplicate according to manufacturer's instructions using 200 μM of each primer. The temperature profiles consisted of an initial denaturation step at 95°C for 10 min and 45 cycles consisting of denaturation at 95°C for 10 s, annealing at 60/55°C for SPP1/RPL32 and extension/fluorescence acquisition at 72°C for 15 s. For both assays threshold cycles were converted to copy numbers using a standard curve generated by amplifying serial dilutions of an external PCR standard (107 - 101 copies). To account for variation in RNA input and efficiency of reverse transcription the calculated SPP1 mRNA copy numbers were normalized by dividing with a normalization factor derived from the expression of the reference gene.
To determine relative amount of splice variants the cDNA fragment described above was amplified using a FAM labelled primer R1F (Additional file 1: Table S1), separated on a MegaBACE 750 capillary sequencer and peak heights were measured using the MegaBACE Fragment Profiler v1.2 software (GE Healthcare, Munich, Germany). The relative quantity of a splice variant was calculated by dividing the corresponding peak height by the sum of peak heights corresponding to all three splice variants.
Electrophoretic mobility shift assay
Nuclear proteins from M.l.d. of 91 old fetuses were prepared essentially as described by Fürbass et al. . Cells (murine mammary epithelial cell line HC11) overexpressing N-terminally truncated bovine C/EBPβ (ΔN-C/EBPβ) and nuclear extracts from these were prepared as described by Yang et al. . Double-stranded probes were prepared by annealing a sense oligo with a shorter antisense oligo (Additional file 1: Table S2) serving as a primer of a Klenow fill-in reaction containing 5U enzyme (Fermentas, St. Leon-Rot, Germany), 1× buffer supplied by the manufacturer and 100 μM each dNTP or 20 μCi [α-32P]dATP for labelling. Nuclear extracts (~3 μg) were incubated with 80 fmol labelled probes (40 fmol labelled probes for EMSA using ΔN-C/EBPβ) in a binding mixture containing 10 mM HEPES-KOH pH 7.9, 50 mM KCl, 0.1 mM EDTA, 0.5 mM DTT, 1 μ poly(dI-dC), 10% glycerol and 1× protease inhibitor cocktail (Roche) at 20°C for 20 min. For competition experiments unlabelled probes were added 10 min prior to addition of labelled probes. For supershift assay 4 μg of an antibody against C/EBPβ (sc-150X, Santa Cruz) was included into the binding reaction. Samples were subsequently subjected to electrophoresis using native 6% polyacrylamide (30:1 AA: Bis) gels in 0.5 × TBE buffer at 20°C. After electrophoresis gels were dried on Whatmann paper, exposed overnight to phosphor storage screens and analysed on a STORM 840 PhosphorImager (Molecular Dynamics, Krefeld, Germany). Band intensities (peak heights) were measured using the ImageQuant TL v2005 software (GE Healthcare). The difference in the binding affinity between the allelic probes was estimated as the average of the ratio of the intensity drop caused by addition of wild type A competitor divided by the intensity drop caused by mutant competitor G for each competitor concentration.
Phylogenetic footprinting was performed using the Mulan online tool accessible at the NCBI DCODE.org Comparative Genomics Developments website http://mulan.dcode.org/. The corresponding bovine, canine, human and murine sequences were retrieved from USCS genome browser http://genome.ucsc.edu. Simple and interspersed repeats were identified using Repeatmasker http://www.repeatmasker.org. Transcription factor binding sites and the effect of the g.3836A>G SNP on these were predicted by the SNPInspector software http://www.genomatix.de.
The strength of the potential splice donor sites in silico was examined using MaxEntScan http://genes.mit.edu/burgelab/maxent/Xmaxentscan_scoreseq_acc.html. To detect open reading frames of the sequences of the different splice variants ORF finder was employed http://www.ncbi.nlm.nih.gov/projects/gorf/.
The effect of g.3836A>G genotype on mRNA expression of SPP1 was analyzed using general linear model (PROC GLM; SAS V9.1, SAS Inst. Inc., Cary, NC) including fixed effects of genotype, stage and their interaction for across stage analysis and fixed effect of genotype for the separate analysis within stage 91 dpc. Least square mean values of the genotypes were compared by a t-test, and the P-values were adjusted by a Tukey-Kramer correction.
The authors thank Angela Garve and Marlies Fuchs for excellent technical help and Dr. Reiner Fürbass for his valuable advice on EMSA technique. This research was supported by German Research Foundation (Deutsche Forschungsgemeinschaft, DFG; Forschergruppe 'DRIP', FOR 753).
- 2.Hirata A, Masuda S, Tamura T, Kai K, Ojima K, Fukase A, Motoyoshi K, Kamakura K, Miyagoe-Suzuki Y, Takeda S: Expression profiling of cytokines and related genes in regenerating skeletal muscle after cardiotoxin injection: a role for osteopontin. Am J Pathol. 2003, 163: 203-215.PubMedCentralCrossRefPubMedGoogle Scholar
- 9.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. 10.1111/j.1432-1033.1992.tb17092.xCrossRefPubMedGoogle Scholar
- 10.Shi X, Liu S, Metges CC, Seyfert HM: C/EBP-beta drives expression of the nutritionally regulated promoter IA of the acetyl-CoA carboxylase-alpha gene in cattle. 2009Google Scholar
- 12.Serre D, Gurd S, Ge B, Sladek R, Sinnett D, Harmsen E, Bibikova M, Chudin E, Barker DL, Dickinson T, et al: Differential allelic expression in the human genome: a robust approach to identify genetic and epigenetic cis-acting mechanisms regulating gene expression. PLoS Genet. 2008, 4: e1000006- 10.1371/journal.pgen.1000006PubMedCentralCrossRefPubMedGoogle Scholar
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.