Establishment of a conditional Nomo1 mouse model by CRISPR/Cas9 technology

The Nomo1 gene mediates a wide range of biological processes of importance in embryonic development. Accordingly, constitutive perturbation of Nomo1 function may result in myriad developmental defects that trigger embryonic lethality. To extend our understanding of Nomo1 function in postnatal stages and in a tissue-specific manner, we generated a conditional knockout mouse model of Nomo1. To achieve this, we used clustered regularly interspaced short palindromic repeats (CRISPR)/Cas9 technology in C57Bl/6J mouse zygotes to generate a new mouse model in which exon 3 of the Nomo1 gene is specifically flanked (or floxed) by LoxP sites (Nomo1f/f). Nomo1f/f mouse embryonic fibroblasts were transduced with a Cre adenovirus and efficiently recombined between LoxP sites. Genomic and expression studies in Nomo1-transduced MEFs demonstrated that the Nomo1 exon 3 is ablated. Western blot assay showed that no protein or early truncated protein is produced. In vivo assay crossing Nomo1f/f mouse with a Msi1-CRE transgenic mouse corroborated the previous findings and it showed Nomo1 exon 3 deletion at msi1+ cell compartment. This short technical report demonstrates that CRISPR/Cas9 technology is a simple and easy method for creating conditional mouse models. The Nomo1f/f mouse will be useful to researchers who wish to explore the role of Nomo1 in any developmental stage or in a tissue-specific manner.


Introduction
The Human Genome Project (HPG) was an international, collaborative research program whose goal was the complete mapping of all the genes of human beings. The HPG revealed that there are probably about 20,500 human genes [7] and subsequently much effort has been expended on understanding their functions and on studying how gene mutations affect the occurrence of human diseases. Since human and mouse genes are 95% homologous, it soon became apparent that mouse genes could serve as tools for understanding human gene function. In this context, researchers and various consortia initially used gene targeting or gene trapping in mouse stem cells to knock out the corresponding murine homologous loci [1,10]. Phenotypic characterization of these mutations is the most accurate and widely used method for elucidating in vivo gene functions and the genetics of diseases. Nevertheless, customary knockout methods can give rise to embryonic lethality from a gene mutation, which prevents researchers from studying the gene in postnatal stages [35]. Tissuespecific gene functions could remain unknown when the null mutation cannot be exclusively produced in a particular tissue. To overcome these obstacles, conditional gene knockout with the Cre-loxP recombination system [31] Electronic supplementary material The online version of this article (https ://doi.org/10.1007/s1103 3-019-05214 -7) contains supplementary material, which is available to authorized users. is widely used to eliminate a specific gene in a particular tissue at a certain development stage [15]. Many mouse stem cell clones have been developed by the International Knockout Mouse Consortium with putative conditional null mutations that are being used to generate conditional mice [1,8]. However, the process of generating the mouse model from the ES cell clone is not free of problems: it is laborious, time-consuming, and can only be carried out in certain mouse strains. Fortunately, the recently developed clustered regularly interspaced short palindromic repeats (CRISPR)/Cas9 system has changed the gene targeting landscape, making the use of transgenesis and mutagenesis technology more feasible, not only in zebrafish or mice as was first reported [28], but in all organisms [14,19,32]. This report describes in detail the generation by CRISPR/ Cas9 technology of a novel conditional Nomo1 mouse, which will allow Nomo1 gene function to be studied in a rigorous and specific way that avoids embryonic lethality. The Nodal subfamily belongs to the group of TGFβ proteins that play a key role in the embryonic development of vertebrates. Nodal signaling is essential for the patterning of the early embryo during mesoderm and endoderm induction as well as the specification of left-right asymmetry [26,27]. Gene expression studies showed that Nodal proteins act before gastrulation to specify the mesodermal and endodermal progenitors, and that the patterning depends on graded signaling with participation of Nodal inhibitors [5,9,29]. These inhibitors can directly bind and inhibit Nodal proteins or compete with them to bind to their coreceptors [3,6]. Nomo (Nodal modulator) and its binding partner, Nicalin, are components of a novel protein complex involved in Nodal signaling. This complex has been revealed as an antagonist of Nodal signaling during mesendodermal patterning in zebrafish. The evolutionary conservation of Nicalin and Nomo in metazoans and plants indicates that the Nomo/Nicalin complex could have a similarly important role. Nomo/Nicalin transcripts were detected in all postnatal human tissues, and an overlapping subcellular distribution was found in endoplasmic reticulum membranes. In zebrafish embryos, Nomo expression was detected in the endoderm and anterior mesendoderm, and its overexpression was found to impair forebrain patterning [11]. Other studies showed that the Nomo1 gene positively regulates the expression of the master regulator and other key development genes involved in bone formation and cartilage development, and that it is essential for cartilage development in zebrafish [2]. The NOMO1 gene is also known to be substantially downregulated in human ventricular septal defect myocardium and, thus, may be an important molecular pathway in human heart development [34]. In humans, tumors from the early-onset subgroup of colorectal cancers showed a homozygous deletion in NOMO1 and were associated with a microsatellite-stable feature and better prognosis, although the molecular mechanism was not identified [21].
Despite the study of Nomo in embryogenesis and pattern expression in embryo and adult tissues, its biological functions in mammalian postnatal tissues remain unknown. Since Nomo is required to produce proper axial mesoderm patterning, bone and heart formation, complete Nomo inhibition could impair embryo development. To study its functions during postnatal stages we decided to create a new mouse model to conditionally interfere with its functions.

Ethics statement
This study followed Spanish and European Union guidelines for animal experimentation (RD 1201/05, RD 53/2013 and 86/609/CEE). The study was approved by the Bioethics Committee of the University of Salamanca.

In silico analysis and CRISPR/Cas9 system design
The mouse Nomo1 gene contains 31 coding exons that produce 5 different transcripts (annotated as Nomo1-201 to Nomo1-205). Nomo1-201, with a length of 4257 bp, is the only one capable of translating into a protein, comprising 1214 residues. Nomo1 exon 3 is the first exon with a residue overlapping the splicing site. In silico analysis showed that an early truncated Nomo1 protein could be produced when exon 3 is removed. Two sgRNAs (Nomo1-sgRNA1 and Nomo1-sgRNA2), targeting the coding sequence of Nomo1 exon 3 and the adjacent intron sequence, were designed with the Spanish National Biotechnology Centre (CNB)-CSIC web tool (http://bioin fogp.cnb.csic.es/tools /break ingca s/) ( Fig. 1). Two complementary oligos corresponding to sgR-NAs were designed, including two 4-bp overhang sequences (Nomo1 UP and LO oligos, Table 1).
As earlier papers described, a long single-strand oligonucleotide of DNA (long ssDNA) (Integrated DNA Technologies, IDT) was considered as donor for homologous recombination to create the conditional KO allele [17,18,24]. A Nomo1 long ssDNA containing LoxP sequences in the same orientation flanking exon 3 was designed, adding silent mutations that destroy PAM sequences. This had an expanded total length of 789 bp ( Supplementary Fig. S1).

Generation of CRISPR/Cas9 reagents
Cas9 mRNA and sgRNA were generated by cloning into a plasmid vector and followed by in vitro transcription [12]. To clone the sgRNAs into the pX458 vector it was previously linearized with BpiI (NEB). The two complementary   [25]. Competent cells were transformed with 2 μl of the ligated plasmid, and single colonies were grown before plasmid extraction using a QIAprep Spin Maxiprep kit (Qiagen). The correct insertion of the sgRNA sequences was confirmed by Sanger sequencing. sgRNA sequences were PCR-amplified from px458based vector with primers carrying the T7 RNA polymerase promoter at the 5′ ends (Nomo1 T7 F and gRNA R, Table 1). PCR products were purified (NZYGelpure, NZYTech) and used as a template for in vitro T7 RNA polymerase transcription (MEGAshortscript T7 Transcription Kit, Thermo Fisher). The Cas9 nuclease ORF, including NLS, was also PCRamplified from pX458 vector [25] with primers carrying the T7 RNA polymerase promoter at the 5′ ends (T7 Cas9 F and R, Table 1). The Cas9 PCR product was purified and used as a template for in vitro transcription, 5′ capping (mMES-SAGE mMACHINE T7 Transcription Kit, Thermo Fisher), and 3′ poly(A) tailing [Poly(A) Tailing Kit, Thermo Fisher]. Transcription products were purified with RNeasy Mini Kit (Qiagen) and eluted in nuclease-free EmbryoMax microinjection buffer (Millipore).

In vitro test to analyze Nomo1-sgRNAs and Cas9 activity
Ten surplus C57BL/6J zygotes from routine processes carried out in the Transgenesis Facility of University of Salamanca were microinjected with 20 ng/µl of Nomo1 sgRNAs, 20 ng/µl of Cas9 mRNA, and 20 ng/µl of Cas9 protein into the cytoplasm and pronucleus. Microinjected embryos were grown on KSOM with non-essential amino acids (Millipore), 37 °C and 5% CO 2 atmosphere at the blastocyst stage. Genomic DNA from single blastocyst-staged embryos were extracted in 10 μl of lysis buffer (50 mM KCl, 10 mM Tris-HCl pH 8.5, 0.1% Triton X-100, and 4 mg/ml of proteinase K) at 55 °C overnight, then heated at 95 °C for 10 min. 2 μl of this DNA solution was used as a template for two rounds of PCR (30 cycles + 20 cycles) to amplify the target sequences using a specific primer for the Nomo1 target region (Nomo1 F and R oligos, Table 1). PCR products were purified using a High Pure PCR Product Purification Kit (Roche) and sequenced by the Sanger method using forward and reverse PCR primers.

Mice and embryo microinjection
Sixty-seven 1-cell-stage embryos from 6 superovulated C57BL/6J females were harvested and microinjected with 20 ng/µl of sgRNA, 20 ng/µl of Cas9 mRNA, 20 ng/µl of Cas9 protein and 25 ng/µl of long ssDNA template solution into the cytoplasm and pronucleus. Microinjected embryos were grown overnight on KSOM with non-essential amino acids (Millipore), 37 °C and 5% CO 2 atmosphere. Fifty-six two-cell-stage embryos were recovered and transferred to two pseudopregnant females. Embryo donor females were euthanized by cervical dislocation and given humanitarian care in accordance with the standards of the Bioethical Committee of University of Salamanca and with Spanish and European Union guidelines for animal experimentation.

Analysis of Cas9-system targets sites
Mouse genomic DNA was extracted from tail biopsy though proteinase K digestion, followed by phenol/chloroform purification. Genomic DNA from MEFs was extracted using the QIAamp DNA Micro Kit (Qiagen) following the manufacturer's protocol. To amplify target regions of mouse Nomo1 gene, PCR was performed with Nomo1 F and Nomo1 R oligos (Table 1). PCR products were purified using a High Pure PCR Product Purification Kit (Roche) and sequenced by the Sanger method using forward and reverse PCR primers.
To characterize the various alleles generated in mouse pups or in MEFs, PCR products were subcloned and transformed in bacteria. DNA from single clones was extracted with a QIAprep Spin Miniprep Kit (Qiagen) and Sanger-sequenced.

RT-PCR and quantitative PCR
To test Nomo1 mRNA expression, 150 ng of total mRNA from MEFs was in vitro-retrotranscribed using a SuperScript III First-Strand Synthesis Super Mix kit (Thermo Fisher). MEF cDNA was used as a template to PCR-amplify the specific sequence between exons 2 and 4 with oligos Nomo-1Ex2F and Nomo1Ex4R (Table 1). PCR product was purified and Sanger-sequenced.

Statistical analysis
Statistical analysis of annexin V expression was performed using GraphPad Prism version 6.00 for Mac OS X (Graph-Pad Software, La Jolla, California, USA, www.graph pad. com). Differences in Nomo1 expression were tested by two-way ANOVA, followed by Tukey's post hoc multiple comparison test. Experimental results were expressed as the mean ± standard deviation (SD). Statistical significance was concluded for values of p < 0.05.

CRISPR/Cas9 reagents efficiently induce mutations at the target site of the Nomo1 sgRNAs
Before injecting embryos with CRISPR/Cas9 reagents, we tested the activity of Nomo1 sgRNAs by injecting 10 C57BL/6J zygotes with a mixture of mRNA Cas9, Cas9 protein and Nomo1 sgRNAs. A nested PCR of material obtained from single blastocyst stage embryos was performed. Genomic DNA from all the embryos was analyzed by Sanger sequencing. A large number of mutations at the expected cut site for both Nomo1 guides was detected in most of the single embryos. A high proportion of mosaic embryos with different alleles was detected ( Fig. 2A); in most of them, the alleles exhibited a specific deletion corresponding to the target area (43 bp) flanked by the cut sites between Nomo1 guides (Fig. 2B) and other mutated (Fig. 2C) and wild-type (wt) alleles.

CRISPR/Cas9 technology produces Nomo1 mosaic mice with a wide variety of mutated alleles
Two pseudopregnant females were transferred, each with 28 two-cell embryos. An offspring of 5 live (#1 to #5) and 5 dead (#6 to #10) pups were born from them. The target region of Nomo1 locus was analyzed by PCR in all of them. All pups, except #3, showed the 786-bp band corresponding to the wt allele. Additionally, six of them displayed secondary bands with higher or lower molecular weights (Fig. 3A), indicating the presence of mosaic pups. Sanger sequencing of all pups confirmed the presence of mosaic animals with three or more edited alleles in most of them. Sanger sequencing revealed a high percentage of null alleles in dead pups due to frameshift mutations (Fig. 3B).

CRISPR/Cas9 technology efficiently produces a Nomo1 conditional allele
Three of the pups analyzed (#1, #9 and # 10) displayed 1 higher secondary band with the expected weight (854 bp), a result that was compatible with the conditional designed allele (Fig. 3A). Sanger sequencing revealed the presence of LoxP sites in two of the pups. In silico analysis showed that pups #9 and #10 contained 1 partially recombined allele with a single LoxP site and the mutated PAM sequence. However, pup #1 had four distinct alleles: one wt, two with insertions or deletions (indels), and one allele edited with the conditional design (Fig. 3C).

Mosaic Nomo1 conditional founder F0 establishes the colony and effectively segregates the conditional allele
Mosaic male #1 was viable and was selected as the founder (F0) of the colony. The F0 was mated with a C57Bl6/J female to segregate the conditional allele into an inbred strain. F1 litter sizes varied from 4 to 9 pups per delivery, and the conditional allele was segregated in a 1/3-1/4 ratio (Fig. 3D). Heterozygous (Nomo1 f/+ ) F1 were fertile and mated to generate homozygous mutants (Nomo1 f/f ) F2. The Nomo1 f/f F2 mice born had a pattern Mendelian inheritance; they were viable and fertile, and displayed no obvious alterations in size or behavior.

Cre recombinase efficiently removes Nomo1 exon 3 in vitro and in vivo
To test whether Cre activity effectively removes the LoxP exon 3 flanked region, we transduced E12.5-derived MEFs from Nomo1 +/+ , Nomo1 f/+ and Nomo1 f/f mice with GFP-Cre adenovirus. Nomo1 +/+ , Nomo1 f/+ and Nomo1 f/f MEFs expressed a high level of GFP after adenoviral infection, indicating an efficient expression of Cre recombinase (Fig. 4A). Genotyping of MEFs after adenoviral infection showed no differences in Nomo1 +/+ or non-transduced MEFs. However, all mutant MEFs (Nomo1 f/+ and Nomo1 f/f ) showed a lower band corresponding to the molecular weight expected after Cre activity (482 bp) and exon 3 removal (Fig. 4B). In order to corroborate these findings, we crossed a female Nomo1 f/f with a male Msi1-Cre transgenic mouse (Msi1-CRE tg/tg ) [22]. Intestinal and brain tissues from Nomo1 +/f ; Msi1-CRE +/tg double mutants were collected and genomic DNA was isolated. A specific lower band due to Cre activity at Msi1+ cells at Nomo1 floxed locus was obtained (Fig. 4C).

Cre recombinase activity in Nomo1 f/f cells generates a Nomo1 null allele
Expression of Nomo1 mRNA and protein was studied in adeno-CRE transduced and non-transduced MEFs from Nomo1 +/+ , Nomo1 f/+ and Nomo1 f/f mice. After adenoviral infection, Nomo1 mRNA expression was checked by RT-PCR in all types of MEFs. All MEFs showed an mRNA of 233 bp, corresponding to the specific amplified Nomo1 cDNA region containing exon 3. Nevertheless, Nomo1 f/+ and Nomo1 f/f MEFs showed an additional 187 bp lower band (Fig. 5A). PCR subcloning and Sanger sequencing of this lower band demonstrated the correct excision of exon 3. In silico analysis of the Nomo1 f/f allele after Cre activity revealed the presence of an early stop codon that Fig. 2 Analysis of SgRNAs and CRISPR/Cas9 activity in Nomo1 locus. A Sanger sequencing of Nomo1 exon 3 amplified from a single edited embryo with both sgRNAs. A mixed sequence is noted between the expected cleavage point (blue arrowhead) of the two sgRNAs (blue box). Subcloning and Sanger sequencing revealed B a 46-bp deletion in the sequence between cleavage points of the two sgRNAs, and C some indels due to the Cas9 action of just one cleavage point  triggered a truncated Nomo1 protein with a different translating frame due to absence of the exon 3 (Fig. 5A).
To study the effect of Nomo1 exon 3 excision by Cre recombinase, Nomo1 expression was measured by realtime PCR and western blot. Nomo1 mRNA expression was quantified in Nomo1 +/+ , Nomo1 f/+ and Nomo1 f/f MEFs before and after adeno-Cre transduction. We observed a substantial reduction (> 80%) of Nomo1 mRNA in Nomo1 f/f -transduced MEFs compared with Nomo1 +/+ (Fig. 5B). These results were corroborated at the protein level by western blot assay, which showed a lower level of Nomo1 expression in Nomo1 f/+ and minimal expression in Nomo1 f/f MEFs due to the presence of a minimal percentage of non-transduced Nomo1 f/f MEFs (Fig. 5C).

Discussion
Genetically engineered mouse models are the best tool for studying gene function and human diseases. In particular, the Cre-LoxP system is widely used to investigate gene functions in specific cellular compartments (spatial control) and/or at specific ages (temporal control). Hundreds of gene-floxed lines and constitutive or tissue-specific Credriver mouse lines have been generated to date and more continue to be developed. They are key to understanding the gene function in a specific cell context. In mouse models of human cancer, instead of using tissue-specific Cre mouse lines, a more real approximation might be to turn-off/turnon the tumor suppressor gene/oncogene in a small group of specific cells using Cre-viral vectors.
Until recently, the process of generating conditional alleles in mice involved manipulating embryonic stem cells, which is a laborious, time consuming and inefficient process. Nowadays, CRISPR/Cas9 technology has improved and simplified this process, allowing us to generate a conditional model that is quicker to use with a reduction in time and number of animals.
This technical report shows that is possible to create a conditional mouse model using a small group of females as embryo donors. However, a careful study of the gene sequence is always necessary. At this point, short exons in In order not to alter the branch site, the LoxP site should not be placed within the first 100 nucleotides upstream of the 3′ end of the intron to ensure that the splicing process is not affected. The first intron of a gene should be avoided because it could contain regulatory DNA sequences and disrupting a key transcription factor binding site could affect the expression of the gene. Regarding the size of the fragment floxed, although big deletions have been achieved with the Cre-LoxP-system flanking the full length of a gene, LoxP positions could certainly affect the efficiency of recombination if they are separated by a long distance.
Despite all these general instructions, an in vitro study to test sgRNAs and Cas9 activity is always recommended before use. In this paper we have used surplus embryos with an identical genetic background. However, mouse cell lines could be used for this purpose as long as they have the same sgRNA target sequence as the mouse embryo. Thus, the target areas would have to be sequenced in order to rule out polymorphisms and mutations that could interfere with the CRISPR/ Cas9 activity. Although the CRISPR/Cas9 system is a rapid, simple, and often extremely efficient gene-editing method, the main obstacle to producing gene-edited animals is mosaicism in the founders [16,23,33]. As our results show, most mice had three or more alleles and, fortunately, the conditional allele was segregated. However, in general, the major challenge of mosaicism is that the mutated allele will not be transmitted to the offspring. In any case, mosaicism also complicates genotype analysis and PCR cloning, and genome sequencing approaches are always required. Nevertheless, not all are disadvantages regarding mosaicism, and it can provide useful advantages, such as the rapid estimation of candidate gene functions in vivo, and the direct comparison of mutant and wt cells in the same organ of mosaic animals. Furthermore, the presence of null alleles in a small proportion of cells enables animals to survive beyond the lethal phase, thus enabling study of the null phenotype in specific groups of cells [36]. Additionally, a mosaic founder mouse can segregate different alleles, generating several mutants in one step.
Given all these challenges and strengths, CRISPR-Cas technology allows the genome mouse embryo to be modified in an easy and conditional manner. We have used this technology to generate a conditional mouse model to study Nomo1 gene function and have demonstrated that Cre activity removes exon 3 to produce a non-functional allele.