Deletion of conserved sequences in IG-DMR at Dlk1-Gtl2 locus suggests their involvement in expression of paternally expressed genes in mice.

Expression regulation of the Dlk1-Dio3 imprinted domain by the intergenic differentially methylated region (IG-DMR) is essential for normal embryonic development in mammals. In this study, we investigated conserved IG-DMR genomic sequences in eutherians to elucidate their role in genomic imprinting of the Dlk1-Dio3 domain. Using a comparative genomics approach, we identified three highly conserved sequences in IG-DMR. To elucidate the functions of these sequences in vivo, we generated mutant mice lacking each of the identified highly conserved sequences using the CRISPR/Cas9 system. Although mutant mice did not exhibit the gross phenotype, deletions of the conserved sequences altered the expression levels of paternally expressed imprinted genes in the mutant embryos without skewing imprinting status. These results suggest that the conserved sequences in IG-DMR are involved in the expression regulation of some of the imprinted genes in the Dlk1-Dio3 domain.

Genomic imprinting is an epigenetic mechanism that results in parental-specific expression of genes inherited from the father and mother. In vertebrates, genomic imprinting is found only in eutherian and marspials. Loss of genomic imprinting leads to embryonic lethality due to abnormal expression of imprinted genes, which are expressed parent-of-origin [1, 2]. Therefore, parent-of-origin-specific expression of imprinted genes is essential for mammalian development. Most imprinted genes form clusters and are regulated by cis-acting imprinting control regions (ICRs) [3]. ICRs are differentially methylated regions (DMRs) that exhibit allele-specific DNA methylation patterns [3] established during gametogenesis and maintained throughout embryonic development [4,5,6]. To date, three paternally methylated DMRs (inherited sperm-specific DNA methylation) and 19 maternally methylated DMRs (inherited oocyte-specific DNA methylation) have been identified in mice [7].

The Dlk1-Dio3 domain is an imprinted gene cluster located on mouse chromosome 12 (Chr12) and has been identified by producing mouse embryos with uniparental disomy of Chr12 [8]. The Dlk1-Dio3 domain contains paternally expressed protein-coding genes (Dlk1, Rtl1, and Dio3), maternally expressed non-coding RNAs, including microRNA clusters (Gtl2, Rtl1as, Rian, and Mirg), and the ICR, whose methylation pattern is derived from the sperm (Fig. 1A) [9,10,11]. It has been reported that the ~ 8-kb intergenic DMR (IG-DMR) located between Dlk1 and Gtl2 acts as the ICR of the Dlk1-Dio3 domain. Analysis of knockout (KO) mice lacking the 4.1-kb IG-DMR showed that this region regulates allele-specific expression of imprinted genes [12]. Maternal inheritance of the deleted allele resulted in embryonic lethality due to biallelic expression of paternally expressed genes and downregulation of maternally expressed genes. On the other hand, paternal inheritance of the deleted allele did not cause embryonic defects [12]. Therefore, hypomethylation of the maternal IG-DMR is functionally essential for embryonic development. However, the precise molecular mechanisms that determine how IG-DMR regulates allele-specific expression of imprinted genes remain largely unclear, given that the IG-DMR is too large for molecular analysis and affects expression of other genes that are located at up to a distance of ~ 1 Mb.

Fig. 1.

Identification of conserved sequences in IG-DMR. (A) Schematic representation of the mouse Dlk1-Dio3 domain. Genomic sequences are indicated with a horizontal black line. Imprinted genes and their expression alleles are indicated with gray boxes and gray arrows, respectively. Names of imprinted genes are shown with their expression alleles. Locations of IG-DMRs are indicated with vertical lines. Black and white circles indicate methylated and unmethylated statuses, respectively. (B) Analysis of sequence conservation of putative IG-DMRs in eutherians. (Top) Results of MultiPipMaker analysis. Alignment of putative IG-DMRs of 10 species is shown. Sequences exhibiting more than 50% and 80% conservation are colored light gray and dark gray, respectively. (Bottom) Identification of three highly conserved sequences in IG-DMR (CS1, CS2, and CS3). Mouse genomic sequence used as a reference is indicated by a black bar. Positions of CS1, CS2, and CS3 are indicated with white boxes. The 4.15-kb region deleted in KO mice [12] is shown by double headed arrows. (C, D) DNA methylation status in CS1, CS2, and CS3 in 14.5 dpc embryos (C) and sperms and MII oocytes (D) analyzed using bisulfite sequencing. The entire amplified sequences and positions of CpG sites are shown by horizontal and vertical lines at the top, respectively. SNP position is indicated with an arrowhead at the top. JF1 (paternal) and B6 (maternal) alleles aligned at the top and bottom, respectively. Each clone is indicated with a vertical line. Methylated and unmethylated CpGs are indicated with black and white circles, respectively. Percent methylation values of CS2 and CS3 are shown at the bottom. Positions of CS1, CS2, and CS3 in the amplified regions are indicated above the methylation status.

Considering that the Dlk1-Dio3 domain is imprinted in all eutherian species examined so far [13], the molecular mechanisms underlying IG-DMR function are likely to be conserved among eutherians. To gain insights into IG-DMR function, we searched for conserved sequences in the IG-DMR of eutherians based on the premise that genes crucial for normal development tend to be conserved. In this study, we identified three IG-DMR sequences that are highly conserved among eutherians. Next, to examine the functional roles of these conserved sequences in vivo, we generated KO mice lacking each of these regions using the clustered regularly interspaced short palindromic repeat (CRISPR)/CRISPR-associated protein 9 (Cas9) system [14, 15].

Materials and Methods

Comparative genomics analysis

Draft genome sequences were obtained from the University of California, Santa Cruz (UCSC) Genome Browser (http://genome.ucsc.edu/). Genome assemblies of each species are listed in Table 1. Repeat sequences in genomic sequence were identified using RepeatMasker (Smit A.F.A., Hubley R. & Green P., http://www.repeatmasker.org/cgi-bin/WEBRepeatMasker). Sequence homology search and alignment were carried out using MultiPipMaker (http://pipmaker.bx.psu.edu/pipmaker/) [16].

Table 1.

Genetic conservation of Dlk1-Gtl2 locus

Species	Location of putative IG-DMR	Length (bp)	Data set
Mouse (Reference)	chr12:110,761,962-110,770,898	8937	NCBI37/mm9
Human	chr14:101,271,066-101,282,625	11560	GRCh37/hg19
Chimp	chr14:100,497,937-100,509,499	11563	CGSC 2.1.3/panTro3
Orangutan	chr14:102,345,370-102,361,570	16201	WUGSC 2.0.2/ponAbe2
Gibbon	GL397390:288,960-299,501	10542	GGSC Nleu1.0/nomLeu1
Rhesus	chr7:164,700,679-164,712,080	11402	BGI CR_1.0/rheMac3
Marmoset	chr10:126,706,354-126,716,922	10569	WUGSC 3.2/calJac3
Rat	chr6:134,099,981-134,108,698	8718	Baylor 3.4/rn4
Cow	chr21:67,352,219-67,364,122	11904	Baylor 4.0/bosTau4
Panda	GL192537.1:392,649-402,737	10089	BGI-Shenzhen 1.0/ailMel1
Cat	chrB3:151,838,799-151,859,155	20357	NHGRI/GTB V17e/felCat4

Generation of mutant mice using the CRISPR/Cas9 system

Single-guide RNA (sgRNA) cloning vectors and human codon-optimized Cas9 (hCas9) plasmids were a generous gift from George Church (Addgene plasmid #41824 and #41815) [14]. sgRNAs were cloned into the sgRNA cloning vector as described previously [17] using the primers listed in Supplementary Table 1 (online only). sgRNA and hCas9 mRNA were prepared via in vitro transcription using mMESSAGE/mMACHINE T7 in vitro transcription kit (Ambion, Austin, TX, USA). PCR products were amplified using primers containing the T7 RNA promoter sequence and subsequently used as templates for in vitro transcription. For microinjection, fertilized eggs were collected from superovulated F1 hybrids of C57BL/6 × DBA/2 (BDF1) female mice crossed with BDF1 male mice (Sankyo Lab Service, Tokyo, Japan). sgRNAs and hCas9 mRNAs (167 ng/µl each) were combined and microinjected into zygotes. Embryos were cultured in KSOM medium for one day. Two-cell embryos were transferred to pseudopregnant ICR female mice. For genotyping, genomic DNA was extracted from the tail or fingertips of pups using DirectPCR Lysis Reagents (Viagen Biotech, Los Angeles, CA, USA). To determine individual mouse genotypes, genomic DNA was PCR-amplified using the primers listed in Supplementary Table 1. Sequences of the mutant alleles were determined by Sanger sequencing analysis using BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosciences, Foster City, CA, USA ). Founder mutant mice were crossed with C57BL/6 mice (Sankyo Lab Service) to obtain the F1 generation. To produce mice with the deleted allele inherited paternally and maternally, male and female F1 mutant mice were crossed with JF1 female and male mice, respectively. JF1 mice were obtained from the National Institute of Genetics (NIG, Mishima, Japan). All animal protocols were approved by the Animal Care and Use Committee of the National Research Institute for Child Health and Development, Tokyo, Japan. All experiments were conducted in accordance with the approved animal protocols.

Expression analysis of imprinted genes

Total RNA and genomic DNA were separately isolated from 14.5 days post coitum (dpc) embryos using AllPrep DNA/RNA/Protein Mini Kit (QIAGEN GmbH, Hilden, Germany). Total RNA (1 µg) was treated with TURBO DNase (Thermo Fisher Scientific, Waltham, MA, USA) and used for cDNA synthesis using SuperScript II (Invitrogen, Carlsbad, CA, USA) reverse transcriptase and oligo(dT) primers. Quantitative RT-PCR (qRT-PCR) analyses were carried out using Power SYBR Green PCR Master Mix (Applied Biosciences) using the primers listed in Supplementary Table 1 [18,19,20,21]. PCR was performed on a Thermal Cycler Dice Real Time System Single TP850 instrument (TaKaRa Bio, Shiga, Japan) using the following profile: 95°C for 10 min, followed by 40 cycles at 95°C for 15 sec, and 60°C for 1 min by. For expression analysis of the Rtl1 gene, cDNA was synthesized from 1 µg of DNase-treated total RNA using the 3′ RACE primer (5′-CTGATCTAGAGGTACCGGATCCGACTCGAGTCGACATCGTTTTTTTTTTTTTTTTT-3′) and SuperScript II reverse transcriptase. qRT-PCR analysis was performed using the Forward primer: 5′-ATACCCAGAGTACTGTGCCAAG-3′ and Reverse primer: 5′-AGAGGTACCGGATCCGACTCGAGTCGACATCG-3′.

Allelic expression analysis

To determine allelic expression of Dlk1, Gtl2, Rtl1, and Dio3, cDNAs were amplified with GoTaq DNA Polymerase (Promega, Madison, WI, USA) using the primers listed in Supplementary Table 1. Reaction conditions were as follows: 35 cycles of 98°C for 10 sec, 58°C for 30 sec, and 72°C for 1 min. PCR products were treated with ExoSAP-IT (USB; Affymetrix , Santa Clara, CA, USA) and sequenced using BigDye Terminator v3.1 Cycle Sequencing Kit.

DNA methylation analysis

Sperms were collected from adult C57BL/6N mice and lysed in extraction buffer (10 mM Tris HCl (pH 7.5), 1 mM EDTA, 0.1 M NaCl, 1% SDS, 0.1 mg/ml Proteinase K, and 10 mM DTT) at 50°C overnight. Sperm DNA was purified using phenol/chloroform extraction and EtOH precipitation. MII oocytes were collected from superovulated C57BL/6N adult female mice. Cumulus cells were completely removed by pipetting. DNA was extracted from 100–120 oocytes by incubation in DNA lysis buffer (1 mg/ml Proteinase K, 1% SDS, and 2 µg of yeast tRNA) at 37°C for 1 h. For bisulfite conversion, DNA was treated with EpiTect Bisulfite Kit (QIAGEN). PCR (nested PCR for oocyte DNA) was performed using EpiTaq HS (TaKaRa Bio) and the primers listed in Supplementary Table 1. PCR products were purified using QIAquick Gel Extraction Kit (QIAGEN) and subsequently cloned into the pGEM-T Easy vector (Promega). Colonies were selected and amplified using Illustra TempliPhi DNA amplification Kit (GE Healthcare, Piscataway, NJ, USA). Sequences were analyzed using M13 reverse primer. Obtained sequence data were aligned using the online web tool QUMA (http://quma.cdb.riken.jp/top/index.html) [22]. Statistical analysis of differences in methylation levels was performed using the Mann-Whitney U test.

Results

Identification of conserved sequences in IG-DMR among eutherians

To examine sequence conservation of IG-DMR in mammals, we first collected putative IG-DMR sequences from different mammalian species. The reference IG-DMR sequence used in this study was set between nearest outside HpaII or HhaI recognition sequences from mouse IG-DMR (8937 bp) as defined by Takada et al. (2002). Methylation levels of the HpaII and HhaI recognition sequences are similar between paternal and maternal alleles (Fig. 1A) [11]. Draft sequences of the entire Dlk1-Gtl2 region in 14 species (human, chimp, orangutan, rhesus, marmoset, rat, guinea pig, pig, cow, sheep, dog, cat, panda, and elephant) were obtained from the UCSC Genome Browser. We excluded wallaby (marsupial) and platypus (monotreme) from the analysis, since the gene corresponding to Gtl2 could not be identified in these two species via homology search using the obtained Gtl2 sequences of the 14 abovementioned species. Next, putative IG-DMRs were searched using BLAT using the following criteria: SCORE > 100, IDENTITY > 80%, and SPAN > 1000. Results showed that out of the 14 eutherians, ten species contained putative IG-DMRs (human, chimpanzee, orangutan, gibbon, rhesus, marmoset, rat, cow, panda, and cat) (Table 1). We aligned the obtained sequences using MultiPipMaker after masking the repetitive elements in mouse sequence using RepeatMasker. Results showed that the sequence corresponding to the conserved 4.1-kb region (KO region), contained conserved sequences among eutherians at more than 50% homology; however, only mouse and rat sequences showed more than 80% homology. On the other hand, we identified three sequences (referred to as CS1, CS2, and CS3) that exhibit more than 80% conservation in more than half of the 11 species, whereas and 50% in the rest at outside of 4.1 kb region deleted in previous report (Fig. 1B) [12].

To determine whether CS1, CS2, and CS3 are DMRs, we characterized their DNA methylation patterns. Since CS1, CS2, and CS3 have 0, 1, and 4 CpG sites, respectively, DNA methylation statuses of CS2 and CS3 were analyzed using bisulfite sequencing in wild-type C57BL/6 × JF1 hybrid F1 whole embryos at 14.5 dpc. In CS2, paternal and maternal alleles could be distinguished based on an intersubspecific polymorphism, as shown in Fig. 1C. Results showed that the CpG site at CS2 was biallelically hypermethylated (paternal: 80%; maternal: 82%), and parent-of-origin effect was not observed (Fig. 1C). Because no polymorphisms were identified in CS3, the analysis was performed without discriminating between paternal and maternal alleles. Results showed that CS3 has a mean methylation level of 72% in the CpG sites and did not possess DMR-like methylation patterns (Fig. 1C). In addition, to verify methylation status at the same loci in germ cells, we also determined methylation patterns in the genomes of sperms and MII oocytes. We observed that both CS2 and CS3 were hypermethylated in sperm but hypomethylated in oocytes (Fig. 1D). These results suggest that sperm-specific DNA methylation of CS1, CS2, and CS3 are established in germ cells, whereas the germline-derived DNA methylation patterns are lost during post-fertilization.

Generation of knockout mice lacking IG-DMR conserved sequences using CRISPR/Cas9

To elucidate the function of the conserved sequences of IG-DMR in vivo, we generated mice lacking the CS1, CS2, and CS3 sequences using the CRISPR/Cas9 system. A pair of single-guide RNAs (sgRNAs) was designed flanking each of CS1, CS2, and CS3 regions (Fig. 2A). sgRNAs for CS1, CS2, and CS3 were microinjected into zygotes with Cas9 mRNA to generate founder mice. We selected mouse lines with the following deletions: 438-bp deletion of CS1 (ΔCS1), 292-bp deletion of CS2 (ΔCS2), and 303 bp-deletion of CS3 (ΔCS3, Fig. 2B). The founder mice were fertile; pups in the F1 generation exhibited normal growth and were fertile. Next, we obtained heterozygous offspring with inherited deletion alleles from the father (+/ΔCS1-3) or mother (ΔCS1-3/+) in the next generation by crossing F1 mice with JF1 mice. Resulting mice also grew normal until adulthood, suggesting that deletion of conserved sequences of IG-DMR does not affect embryonic development and growth after birth (Fig. 3).

Fig. 2.

Generation of CS-deficient mice by the CRISPR/Cas9 system. (A) Design of sgRNAs for deletion of CS1, CS2, and CS3. Schematic representation of mouse IG-DMR. Genomic sequence of IG-DMR is shown as a black line. Positions of CS1, CS2, and CS3 are shown with white boxes. Enlarged region around CS1, CS2, and CS3 is also shown at bottom. Target sequences of sgRNA pairs for CS1 (sgRNA-L1/R1), CS2 (sgRNA-L2/R2), and CS3 (sgRNA-L3/R3) are indicated with gray lines. (B) Genotyping of mice carrying ΔCS1-3 alleles. Nucleotide sequences around sgRNA target sequences of CS1-3 and ΔCS1-3 are shown. Numbers in parentheses indicate counts of deleted and inserted nucleotides. Lowercase letters represent inserted sequences.

Fig. 3.

Growth of mutant mice carrying ΔCS1-3 alleles. Body weights of mice after birth are shown. Error bars indicate standard error.

Expression levels of paternally expressed genes are altered in CS-deficient mice

Next, to test whether the loss of CS regions affects the expression of imprinted genes in the Dlk1-Dio3 domain, we performed gene expression analysis of the paternally expressed genes Dlk1, Rtl1, and Dio3 and the maternally expressed gene Gtl2 in the +/ΔCS1-3 and ΔCS1-3/+ embryos. Quantitative RT-PCR analysis of +/ΔCS1-3, ΔCS1-3/+, and wild-type embryos at 14.5 dpc revealed no significant differences in Gtl2 expression levels in all genotypes. On the other hand, Dlk1 expression was 1.5-fold upregulated in ΔCS1/+ embryos compared to wild-type and the other genotypes. In addition, Dio3 expression levels were significantly lower in +/ΔCS1-3 but not in ΔCS1-3/+ embryos. Rtl1 expression was also downregulated in ΔCS2/+ embryos compared to the other genotypes (Fig. 4A). To elucidate whether the observed upregulation of Dlk1 and downregulation of Rtl1 and Dio3 were due to loss of imprinting, we examined allelic expression of imprinted genes in the +/ΔCS1-3 and ΔCS1-3/+ embryos. Results showed that all genes tested, including Dlk1, Rtl1, and Dio3, exhibit the same imprinting patterns as the wild-type, suggesting that the observed Dlk1 upregulation and Rtl1 and Dio3 downregulation were not caused by loss of imprinting (Fig. 4B). These results are consistent with that of allelic expression analysis, which showed that DNA methylation status of IG-DMR and promoter CpG island of Gtl2 gene (also known as Gtl2-DMR, Fig. 5A) in +/ΔCS1-3 and ΔCS1-3/+ embryos were similar to those of wild-type embryos (Fig. 5B).

Fig. 4.

Expression analysis of imprinted genes in CS-deficient embryos. (A) Quantitative RT-PCR analysis of imprinted genes in mutant embryos carrying ΔCS1, ΔCS2, and ΔCS3 alleles. Expression levels of imprinted genes in wild-type (+/+), maternally transmitted (ΔCS/+), and paternally transmitted (+/ΔCS) embryos are indicated with white, red, and blue bars, respectively. Expression levels are normalized by of the mean value of wild-type embryos. Error bars indicate standard error. * P < 0.01, t-test. Representative data are shown. Results were obtained from embryos in a single littermate. Similar results were obtained using embryos obtained from independent littermate. (B) Expression alleles of imprinted genes in mutant embryos carrying the ΔCS1, ΔCS2, and ΔCS3 alleles. Electropherograms obtained by direct sequencing of RT-PCR products of Dlk1, Gtl2, Rtl1, and Dio3 are shown. Genotypes are indicated at the left. Mating combination is indicated at the top of each electropherogram. Polymorphic nucleotides are shaded in blue. Representative data are shown. Similar results were obtained using embryos obtained from independent littermates.

Fig. 5.

DNA methylation analysis of IG-DMR and Gtl2-DMR in CS-deficient embryos. (A) Schematic representation of mouse IG-DMR and Gtl2-DMR. Genomic sequences are indicated with a black line. Positions of IG-DMR and Gtl2-DMR are shown in white boxes. CS1, CS2, and CS3 are indicated with black boxes. Transcription start site of Gtl2 is shown with an arrow. Regions amplified by bisulfite PCR are indicated with gray boxes. (B) Bisulfite sequencing analysis of IG-DMR and Gtl2-DMR in mutant embryos carrying ΔCS1, ΔCS2, and ΔCS3 alleles. White and black circles indicate unmethylated and methylated CpG sites, respectively. Asterisks show the positions of CpG sites harboring polymorphisms. Representative data are shown. Similar results were obtained using embryos obtained from independent littermates.

Taken together, these results suggest that conserved regions in IG-DMR enhance the expression of paternally expressed genes without skewing imprinting patterns.

Discussion

Deletion of the 4.1-kb sequence (KO region) in IG-DMR is known to result in similar phenotypes in humans and mice [12, 23]. It was previously assumed that the KO region constitutes the most functionally essential region(s) of IG-DMR and thus represents the most highly conserved sequences among eutherians. However, the three most highly conserved regions (CS1, CS2, and CS3) identified in this study, which exhibit more than 80% sequence conservation, were located outside the KO region. The KO region contains sequences with only 50-80% conservation, which is lower than when 8.9-kb region analyzed in this study is included in the alignment [11]. Thus, the functionally important sequences in IG-DMR that are located in the KO region may not be highly conserved among eutherians and act via different molecular mechanisms. To date, all studies show that the molecular functions of imprinted genes in the Dlk1-Dio3 domain are associated with the KO region. Kota et al. (2014) demonstrated that maternally expressed non-coding RNAs regulate the expression of imprinted genes in the Dlk1-Dio3 domain. In addition, Luo et al. (2016) showed that the AFF3 transcription factor is essential for maternal expression of Gtl2-Rian-Mirg genes in mice and acts by binding to allele-specific enhancers in IG-DMR [24, 25]. Some parts of the non-coding RNAs sequences and of the AFF3 binding sites are located more than 50% conserved region, but some are not. In addition, IPW, a non-coding RNA located in the Prader-Willi syndrome region on chromosome 15, was reported to regulate the epigenetic status of the DLK1-DIO3 domain in human induced pluripotent cells. However, the IPW gene sequence is not highly conserved between humans and mice [26]. Thus, the results point to species-specific variations in regulatory mechanisms in the DLK1-DIO3 domain.

In this study, we identified three highly conserved sequences (CS1, CS2, and CS3) in IG-DMR and generated knockout mice lacking each of these regions. Mutant mice lacking CS1, CS2, and CS3 did not exhibit gross phenotypes and uniparental expression in the Dlk1-Dio3 domain that are usually observed in wild-type mice, suggesting that these regions are not required for maintenance or regulation of imprinted expression in Dlk1-Dio3. However, Dlk1 was found to be upregulated in ΔCS1/+ embryos, whereas Rtl1 and Dio3 genes were downregulated in +/ΔCS2 and +/ΔCS1-3 embryos, respectively. These results suggest that the conserved sequences in IG-DMR can directly or indirectly modulate expression levels of paternally expressed imprinted genes in the Dlk1-Dio3 domain and act in a cis manner. Interestingly, maternal deletion of CS1 resulted in Dlk1 upregulation, suggesting trans modulation of paternal Dlk1 expression by the maternal CS1 in embryos. In addition, Dlk1 was 1.5-fold upregulated and Rtl1 and Dio3 were 0.5-fold downregulated in CS-deficient embryos, but no gross phenotype was observed at postnatal stages. These results suggest that the changes in expression levels occur either in the entire embryo or in a tissue-specific manner at levels that do not impair embryonic development. In either or mixed cases, results indicate that CSs are involved in expression modulation of the imprinted genes. Further studies are required to elucidate the detailed molecular mechanisms that determine how CSs modulate the expression of paternally expressed genes.

In summary, we identified three highly conserved sequences in the IG-DMR of eutherians. By generating knockout mice lacking these sequences using the CRISPR/Cas9 system, we showed that these conserved sequences in IG-DMR influence the expression of paternally expressed genes. These findings provide new insights into the molecular mechanisms underlying gene expression modulation of imprinted genes in the Dlk1-Dio3 domain.