Characterization of single nucleotide polymorphisms for a forward genetics approach using genetic crosses in C57BL/6 and BALB/c substrains of mice.

Forward genetics is a powerful approach based on chromosomal mapping of phenotypes and has successfully led to the discovery of many mouse mutations in genes responsible for various phenotypes. Although crossing between genetically remote strains can produce F2 and backcross mice for chromosomal mapping, the phenotypes are often affected by background effects from the partner strains in genetic crosses. Genetic crosses between substrains might be useful in genetic mapping to avoid genetic background effects. In this study, we investigated single nucleotide polymorphisms (SNPs) available for genetic mapping using substrains of C57BL/6 and BALB/c mice. In C57BL/6 mice, 114 SNP markers were developed and assigned to locations on all chromosomes for full utilization for genetic mapping using genetic crosses between the C57BL/6J and C57BL/6N substrains. Moreover, genetic differences were identified in the 114 SNP markers among the seven C57BL/6 substrains from five production breeders. In addition, 106 SNPs were detected on all chromosomes of BALB/cAJcl and BALB/cByJJcl substrains. These SNPs could be used for genotyping in BALB/cJ, BALB/cAJcl, BALB/cAnNCrlCrlj, and BALB/cCrSlc mice, and they are particularly useful for genetic mapping using crosses between BALB/cByJJcl and other BALB/c substrains. The SNPs characterized in this study can be utilized for genetic mapping to identify the causative mutations of the phenotypes induced by N-ethyl-N-nitrosourea mutagenesis and the SNPs responsible for phenotypic differences between the substrains of C57BL/6 and BALB/c mice.

Introduction

Mouse models are important tools for understanding the phenotypes of human diseases, genes responsible for diseases, and molecular mechanisms for the development of disease phenotypes. The approaches to developing mouse models include reverse genetics and forward genetics [1,2,3,4]. Reverse genetics is a powerful approach based on gene targeting and genome editing. This approach has successfully produced null mutations in almost all genes and mimics mutations associated with several human diseases. Although reverse genetics is a major approach in mouse genetics, forward genetics is a powerful approach based on chromosomal mapping of phenotypes and has led to the discovery of numerous mutations in genes responsible for various spontaneous phenotypes [4]. Moreover, the combination of forward genetics and large-scale N-ethyl-N-nitrosourea (ENU) mutagenesis studies has uncovered the functions of unknown genes and mutations [1, 5,6,7].

To identify genes associated with spontaneous and ENU-induced phenotypes using the classical forward genetic approach, researchers first produced experimental populations of chromosomal recombinants—F2 and backcross (N2) mice—by crossing between the mutant and wild-type mice. Wild-type mice must be chosen from other strains with different genetic backgrounds and mutants, and they are preferred over genetically diverse strains with mutants. This is because the mutations map to a chromosomal region via genotyping of the polymorphic markers, such as microsatellites and single nucleotide polymorphisms (SNPs), between both parental strains. Therefore, interspecific, intersubspecific, and interstrain genetic crosses have been used for classical chromosomal mapping of mutants of interest because of the availability of many polymorphic markers for genotyping in F2 and N2 mice from these genetic crosses [8].

However, partner strains that show highly similar phenotypes should be chosen to create a genetic cross in cases involving the mapping of causative mutations for mild phenotypes, which are easily affected by the genetic background of the partner strain. Crosses with similar phenotypic strains are effective for the mapping of behavioral and sensorial traits. Since there are individual differences in these phenotypic traits, phenotypic noise caused by modifiers in the genetic background of the partner strain should be reduced. Substrains may be candidate partners for genetic crosses. Genetically different substrains have been established in most classical inbred strains [9,10,11,12]. Although there are far fewer genomic polymorphisms between substrains than between strains, previous studies have successfully identified the mutations associated with phenotypes via linkage and quantitative trait loci (QTL) analysis using intersubstrain genetic crosses [13,14,15,16,17].

To perform linkage and QTL analyses using intersubstrain genetic crosses, it is important to set up polymorphic markers between substrains of mice without large gaps on the chromosomes. Therefore, we investigated the SNPs that are available for immediate use in linkage and QTL analyses using substrains. In this study, we chose substrains of C57BL/6 and BALB/c mice as the research targets. These are very common inbred mice, and several substrains have been established. Moreover, there are potential mutants available as human disease models because several large-scale ENU mutagenesis studies have been carried out based on ENU-treated male mice from both strains [18,19,20,21]. Therefore, marker sets based on SNPs between substrains can be a useful tool for identifying new phenotypes associated with human diseases.

Materials and Methods

Mice and ethics statement

The C57BL/6 and BALB/c substrains used in this study were purchased from each their respective production breeders (Table 1). This study was reviewed and approved by the Institutional Animal Care and Use Committee of the RIKEN Tsukuba Branch and the Animal Care and Use Ethics Committee of the Tokyo Metropolitan Institute of Medical Science.

Table 1.

C57BL/6 and BALB/c substrains used in this study

Strain	Substrain	Breeder
C57BL/6	C57BL/6J	The Jackson Laboratory via Charles River Laboratories Japan
	C57BL/6JJcl	CLEA Japan
	C57BL/6JJmsSlc	Japan SLC
	C57BL/6NTac	Taconic Biosciences via CLEA Japan
	C57BL/6NJcl	CLEA Japan
	C57BL/6NCrl	Charles River Laboratories Japan
	C57BL/6NCrSlc	Japan SLC
BALB/c	BALB/cAJcl	CLEA Japan
	BALB/cByJJcl	CLEA Japan
	BALB/cAnNCrlCrlj	Charles River Laboratories Japan
	BALB/cCrSIc	Japan SLC

Preparation of PCR templates

All PCR templates were obtained from the pinna and/or tail samples of adult mice euthanized by carbon dioxide inhalation. The pinna (diameter: 2 mm) and tail (length: less than 2 mm) samples were incubated in 50 µl of lysis buffer solution (50 mM Tris-HCl, pH 8.0; 20 mM EDTA, pH 8.0; 100 mM NaCl; and 1% sodium dodecyl sulfate) with 0.5 µl of proteinase K solution (20 mg/ml) at 37°C overnight and then incubated at 95°C for 10 min. The solution samples were diluted with water (1:200) and used as PCR templates.

Validation of SNPs between C57BL/6J and C57BL/6N substrains

To prepare novel SNP markers, we selected SNPs between the C57BL/6J and C57BL/6N substrains, based on published data [16, 22,23,24]. For several SNPs, the flanking sequences including the SNPs were amplified by PCR using AmpliTaq Gold DNA Polymerase (Thermo Fisher Scientific, Waltham, MA, USA) or TaKaRa Ex Taq (Takara Bio, Inc., Kusatsu, Japan) and the primers listed in Supplementary Table 1 according to the manufacturer’s protocol. The SNPs were validated via Sanger sequencing of the PCR products. The PCR products were purified using the ExoSAP-IT (Thermo Fisher Scientific) and sequenced using an Applied Biosystems 3130xl Genetic Analyzer and 3730xl DNA Analyzer (Thermo Fisher Scientific) by the Genetic Analysis Division of Fasmac Co., Ltd. (Atsugi, Japan).

SNP genotyping

A set of validated SNPs was used to design TaqMan assays using a custom TaqMan SNP Genotyping Assay (Thermo Fisher Scientific; Supplementary Table 2). PCR template solutions (1 µl) were dispensed into each well of 384-well plates and dried using a speed-vacuum drying system. The reaction mixtures for genotyping using TaqMan Genotyping Master Mix (Thermo Fisher Scientific) and the custom TaqMan assays were dispensed into 384-well plates and amplified via PCR following the manufacturer’s instructions. The PCR products were analyzed using the TaqMan Genotyper software with an ABI PRISM 7900HT Sequence Detection System (Thermo Fisher Scientific) or QuantStudio 12K Flex Real-Time PCR System (Thermo Fisher Scientific) according to the manufacturer’s instructions.

Detection of SNPs between BALB/cA and BALB/cByJ substrains

To obtain information on SNPs between BALB/cAJcl (BALB/cA) and BALB/cByJJcl (BALB/cByJ) mice, next-generation sequencing (NGS) technology was used. For sequencing, genomic DNA was extracted from the spleens of female BALB/cA and BALB/cByJ mice using a QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany). Whole genome sequencing and listing of mutations were performed by Hokkaido System Science Co., Ltd. (Hokkaido, Japan). Paired-end reads of 150 bp were generated with HiSeq X with HiSeq X Ten Reagent Kit v2.5 (Illumina, San Diego, CA, USA) to achieve more than 30× genome coverage. The sequence reads were mapped on the reference genome sequence (mouse genome reference assembly, GRCm38.p6) and analyzed using Samtools/BCFtools version 1.2 [25] and SnpEff version 4.1g [26]. The quality of a given base call is measured as a Phred quality score, which indicates the probability that the base is correctly called. The scores generally range from 2 to 40, with higher scores indicating greater confidence in the call. The candidate SNPs between BALB/cA and BALB/cByJ were confirmed via Sanger sequencing.

Results

Construction of marker sets for genetic mapping of genetic crosses between C57BL/6J and C57BL/6N substrains

We previously developed 101 primer and probe sets (Supplementary Table 2) for TaqMan assays and succeeded in linkage mapping of several phenotypes using these sets [14,15,16,17]. To place markers on chromosomal regions for linkage mapping at equal intervals without the absence of markers in the long-range chromosomal region, SNPs were screened from databases. Thirteen SNPs were identified as available for linkage mapping in genetic crosses between the C57BL/6J and C57BL/6N substrains via Sanger sequencing (Supplementary Tables 1 and 2). The locations of all markers for TaqMan assays are shown in Fig. 1. The markers were assigned at 20.58 Mb intervals on the chromosomes on average in the case of choosing either one from the neighboring markers (D2SNP38/D2SNP39, D3SNP39/D3SNP48, D4SNP38/D4SNP42, D12SNP38/D12SNP41, D17SNP34/D17SNP32, and D19SNP23/D19SNP26) located within 10 Mb (Fig. 1 and Supplementary Table 2). Although the SNP markers were absent in several long-range genomic regions (e.g., D6SNP32–D6SNP33, 50.18 Mb, and DXSNP58–DXSNP52, 48.43 Mb), we predicted that the marker sets for identifying the loci linked with phenotypes could be assigned to all chromosomes.

Fig. 1.

Chromosomal map of the 114 SNP markers for genotyping of genetic crosses between the C57BL/6J (C57BL/6J, C57BL/6JJcl, and C57BL/6JJmsSlc) and C57BL/6N (C57BL/6NTac, C57BL/6NJcl, C57BL/6NCrl, and C57BL/6NCrSlc) substrains. The vertical gray lines and black diamonds indicate chromosomes and locations of SNP markers, respectively. The SNP markers common between C57BL/6J and C57BL/6N substrains are labeled in black. The intrasubstrain polymorphisms were detected in 25 markers, which are labeled in red, blue, green, and magenta.

The marker sets were designed based on the SNPs between C57BL/6J mice from The Jackson Laboratory and C57BL/6NTac mice from Taconic Biosciences. C57BL/6J is the most commonly used substrain in biomedical research, and its genomic sequence was first published as a reference genome sequence by the Mouse Genome Sequencing Consortium (MGSC) [27, 28]. C57BL/6NTac was selected by the Knockout Mouse Project and the International Mouse Phenotyping Consortium as a background strain to produce genetically engineered mutants [29,30,31]. However, previous studies have reported that there are several SNPs at the intrasubstrain level in both C57BL/6J and C57BL/6N mice from each breeder [22, 32]. Therefore, we investigated SNPs in C57BL/6J and C57BL/6N mice purchased from three major breeders (Table 1) in Japan using TaqMan assays. SNPs were detected in 25 markers, and SNPs in 13 markers were detected within the C57BL/6J substrains. The genotypes of 13 markers in C57BL/6JJmsSlc mice and 5 markers in C57BL/6JJcl mice were identical to those of all C57BL/6N mice (Fig. 1 and Table 2). In C57BL/6N mice, SNPs were detected in 12 markers and were identical to those of all C57BL/6J mice (Fig. 1 and Table 2). Twelve SNPs were detected in C57BL/6NCrSlc mice, and only one SNP was detected in C57BL/6NCrl mice. There were no SNPs in C57BL/6NTac and C57BL/6NJcl mice.

Table 2.

SNPs among the C57BL/6 substrains within the markers developed in this study

Identification of SNPs for genetic mapping of the genetic crosses between BALB/cA and BALB/cByJ substrains

Next, we performed NGS whole-genome sequencing to screen SNPs used for genetic mapping of the genetic crosses between the BALB/cA and BALB/cByJ substrains. In total, 912,105,404 and 636,697,536 reads of BALB/cA and BALB/cByJ mice, respectively, were mapped to the reference genome sequence. A large number of SNPs were detected in BALB/cA and BALB/cByJ mice via this analysis. However, the quality of the sequence data was low because the read coverage was poor in many genomic regions, including these SNPs. Therefore, candidate SNPs for genotyping between BALB/cA and BALB/cByJ mice were filtered according to several criteria that included read coverage, allele frequency, and genotype information (Supplementary Table 3). In addition, we selected SNPs that were assigned dbSNP IDs from the filtered SNPs to increase the reliability of SNPs because a large amount of SNP data generated using oligonucleotide arrays in BALB/cByJ mice have previously been reported [33]. Finally, 632 candidate SNPs were screened between BALB/cA and BALB/cByJ mice (Supplementary Table 3). To construct the marker sets for genetic mapping using F2 and N2 progeny between the BALB/c substrains, we used Sanger sequencing to validate SNPs selected based on the genomic locations. One hundred six SNPs were detected on all chromosomes in BALB/cA and BALB/cByJ mice (Supplementary Table 4). Figure 2 shows the locations of the 106 SNPs. Most of the SNPs were assigned locations on all chromosomes, as planned. The average interval between the SNPs was 21.18 Mb when either SNP was chosen from the neighboring markers (D10SNP302/D10SNP303, D16SNP304/D16SNP305, D18SNP301/D18SNP302, D19SNP301/D19SNP302, D19SNP302/D19SNP303, and DXSNP303/DXSNP304) located within 10 Mb. However, we could not detect the SNPs in the long-range genomic region (~39.12 Mb) of the centromere on chromosome 1 because of difficulty of designing primer sets for Sanger sequencing to validate them. This was attributed to the presence of repetitive sequences and long stretches of a single nucleotide in the flanking sequence of all candidate SNPs. We also could not detect the SNPs on the proximal region (1–28.12 Mb) of chromosome 18 because there were no candidate SNPs in this region (Supplementary Table 3).

Fig. 2.

Chromosomal map of the 106 SNP markers for genotyping of genetic crosses between the BALB/c (BALB/cJ, BALB/cA, BALB/cByJ, BALB/cAn and BALB/cCr) substrains. The vertical gray lines and white diamonds indicate chromosomes and locations of SNP markers, respectively. The genotypes of the SNP markers are shown in different colors (black, red, and blue) for each marker name.

Regarding BALB/c substrains, the whole-genome sequence of BALB/cJ mice has been published by the Wellcome Trust Sanger Institute [23]. In addition, 106 SNPs were genotyped in BALB/cAnNCrlCrlj (BALB/cAn) and BALB/cCrSIc (BALB/cCr) mice. Table 3 shows comparisons of the genotypes of the 106 SNPs among the five substrains. The genotypes of BALB/cJ and BALB/cA mice were identical at 90 SNPs. In contrast, the same genotypes were detected in only 16 positions in BALB/cJ and BALB/cByJ mice. The genotypes of the 106 SNPs matched exactly in BALB/cAn and BALB/cCr mice. The genotypes at 73, 57, and 49 SNPs of BALB/cAn and BALB/cCr were identical to those of BALB/cJ, BALB/cA, and BALB/cByJ, respectively.

Table 3.

SNPs among the BALB/c substrains within the markers developed in this study

Locations of SNPs are according to the mouse genome reference assembly (GRCm38.p6). SNPs with the genomic sequence (Mouse Genome Project: https://www.sanger.ac.uk/sanger/Mouse_SnpViewer/rel-1505) of BALB/cJ (cJ) are highlighted in gray. cA, BALB/cAJcl; cByJ, BALB/cByJJcl; cAn, BALB/cAnNCrlCrlj; and cCr, BALB/cCrSlc.

Discussion

This study provides information on available SNP marker sets for genotyping of C57BL/6 and BALB/c substrains. As mentioned in the Introduction, a useful application of these marker sets is linkage and QTL analyses of the phenotypes produced by ENU mutagenesis. Large-scale ENU mutagenesis studies in mice have been powerful discovery platforms for identifying novel genes or gene functions associated with disease phenotypes [1, 5]. We also identified several genes and gene functions associated with several disease phenotypes, such as hair formation [34, 35], amelogenesis [36], development of single positive thymocytes [37], and abnormal behavior due to neurological disorders [38,39,40]. However, we could not identify the loci associated with several phenotypes by linkage and QTL analyses. In particular, the genetic mapping of several mild morphological and behavioral phenotypes, which were detected using a modified SHIRPA (SmithKline, Harwell, Imperial College, Royal Hospital, Phenotype Assessment) protocol, failed [20]. We have previously used interstrain genetic crosses between ENU-treated C57BL/6J and DBA/2J (or C3H/HeJ) mice for linkage and QTL analyses. We predicted that these mild phenotypes might be overlooked by the genetic background effects of DBA/2J and C3H/HeJ mice. In contrast, several studies have identified causative mutations associated with disease phenotypes using intersubstrain genetic crosses. Hossain et al. (2016) identified obesity-related genes based on genetic mapping of the N2 progeny between ENU-treated C57BL/6J and C57BL/6N mice [14]. Funato et al. (2016) and Miyoshi et al. (2019) identified sleep-related genes by a forward genetics approach using the same intersubstrain cross [13, 15]. As it was assumed that target phenotypes in these studies are susceptible to genetic background effects, we predict that substrain crosses will become a key strategy for the successful identification of these causative genes.

In addition, we expect that the marker sets developed in the present study will be useful tools for identifying loci and mutated genes associated with phenotypic differences between substrains of C57BL/6 and BALB/c mice. Kirkpatrick et al. (2017) identified a robust difference in binge eating between C57BL/6 substrains and a major genetic factor associated with the phenotype by QTL mapping using C57BL/6J × C57BL/6N-F2 mice [41]. Kumar et al. (2013) confirmed that substrain differences in the locomotor response to cocaine in C57BL/6J and C57BL/6N mice were caused by an SNP [16]. Previous studies have reported several phenotypic differences among substrains of C57BL/6 and BALB/c mice [9, 11, 42, 43]. Although almost all causative genes and mutations are still unknown, they may be identified by linkage and QTL analyses using intersubstrain crosses and SNP markers in both substrains.

The phenotypic differences are estimated to be associated with approximately 10,000 SNPs between the C57BL/6J and C57BL/6N substrains [11, 23, 42]. In addition, Mekada et al. (2015) reported genetic differences among substrains of C57BL/6J and C57BL/6N mice produced by breeders [44]. We also detected several SNPs in both the C57BL/6J and C57BL/6N substrains from different breeders. Although we investigated only Japanese breeders, the genetic differences in mice from different breeders would probably be detected in mice from different breeders in other countries. Therefore, it is essential to pay attention to the genetic background of the mutant and control mice for the production of genetic crosses to perform linkage and QTL analyses using SNP markers between the C57BL/6J and C57BL/6N substrains.

In this study, we investigated and confirmed the SNPs between the BALB/cByJ and BALB/cA substrains purchased from CLEA Japan for linkage and QTL analyses, since this approach has been successful in discovering several phenotypic differences between these substrains. However, most reports have described phenotypic differences between BALB/cByJ and BALB/cJ substrains [45,46,47,48]. Although 90 markers within our marker sets are available for the genotyping of these substrains, SNP markers were absent in several long-range genomic regions (e.g., D6SNP302–305, D9SNP302–304, D11SNP302–305, and proximal and distal regions on chromosome 16; Fig. 2 and Table 3). Additional markers will need to be developed at several positions to perform detailed linkage and QTL analyses using genetic crosses between BALB/cByJ and BALB/cJ substrains. In addition, we screened the genotypes of SNPs in BALB/cAn and BALB/cCr mice purchased from the Charles River Laboratories Japan and Japan SLC, respectively; however, the genotypes of these strains were the same for all SNPs. Our SNP markers might be used for linkage and QTL analyses to cross these mice (BALB/cAn and BALB/cCr) and BALB/cByJ mice. However, whole-genome genotyping is difficult for crosses between BALB/cA, BALB/cJ, BALB/cAn, and BALB/cCr mice using only our SNP markers because of the lack of polymorphisms in many chromosomal regions. Whole-genome sequencing of BALB/cAn and BALB/cCr mice can help screen the SNPs and compare them with those in other BALB/c substrains.

In the present study, we developed a genotyping system on 384-well plates using TaqMan assays. The approach can promptly distinguish the homozygous and heterozygous genotypes in progeny from the genetic crosses between C57BL/6 substrains, as shown in Supplementary Fig. 1. This genotyping system can provide researchers with the ability to identify the responsible and modifier genes via linkage and QTL analyses using genetic crosses between the C57BL/6 substrains. We are currently developing a similar high-throughput genotyping system for BALB/c substrains using the SNPs identified in the present study.