Detection of nucleotide-specific CRISPR/Cas9 modified alleles using multiplex ligation detection

CRISPR/Cas9 genome-editing has emerged as a powerful tool to create mutant alleles in model organisms. However, the precision with which these mutations are created has introduced a new set of complications for genotyping and colony management. Traditional gene-targeting approaches in many experimental organisms incorporated exogenous DNA and/or allele specific sequence that allow for genotyping strategies based on binary readout of PCR product amplification and size selection. In contrast, alleles created by non-homologous end-joining (NHEJ) repair of double-stranded DNA breaks generated by Cas9 are much less amenable to such strategies. Here we describe a novel genotyping strategy that is cost effective, sequence specific and allows for accurate and efficient multiplexing of small insertion-deletions and single-nucleotide variants characteristic of CRISPR/Cas9 edited alleles. We show that ligation detection reaction (LDR) can be used to generate products that are sequence specific and uniquely detected by product size and/or fluorescent tags. The method works independently of the model organism and will be useful for colony management as mutant alleles differing by a few nucleotides become more prevalent in experimental animal colonies.

CRISPR/Cas9 genome-editing has emerged as a powerful and simple tool that is making many model organisms more genetically tractable 1,2 . The nucleotide-level precision of the technique is accompanied by complications in genotyping and colony management 3 . Genetically modified mice, for an example, represent an important model for understanding gene function in development and disease. Mutant mice conventionally generated by insertional mutagenesis or gene-targeting methods have relied on homologous recombination of exogenous DNA fragments 4 . The integrated exogenous DNA can be exploited to facilitate individual animal genotyping and colony management. More recently, CRISPR/Cas9 gene-targeting technology has been developed to efficiently disrupt genes in zygotes or ES cells for the purpose of creating animal models 5 , including organisms that have traditionally been challenging to manipulate genetically 2,4,5 . DNA double strand breaks created by Cas9 are repaired by non homologous end joining (NHEJ) when a homologous template is not provided 6 . However, in many cases, the small insertions or deletions (in-dels) created by NHEJ DNA repair do not generate a novel restriction site or change the size sufficiently enough to be easily resolved with traditional genotyping strategies 7 .
Based on the many benefits of this CRISPR/Cas9 gene targeting technology, the number of genome edited animal models is rapidly increasing, necessitating cost effective, rapid and sensitive genotyping strategies. Conventional polymerase chain reaction (PCR) followed by agarose gel electrophoresis lacks the sensitivity to detect products that differ by only a few nucleotides. Assays that have been used to detect the small genetic variants created at targeted loci include: Surveyor Assay (Cel1) [8][9][10] , T7 Endonuclease I 8,11,12 , high resolution melting curve analysis (HRMA) 13 , allele specific PCR (ASPCR) 14 , restriction fragment length polymorphism (RFLP) 15 and direct Sanger sequencing of PCR products. While Surveyor and T7 Endonuclease I specifically cleave heteroduplex DNA mismatch, HRMA utilizes the difference in melting curve of the heteroduplex, and RFLP is dependent upon creation or elimination of a defined restriction site within the targeted locus. These assays are expensive, time consuming, lack sequence specificity, and require costly capital equipment. Moreover, several of these techniques fail to distinguish biallelic combinations for all modes of inheritance; wild type, heterozygous, compound heterozygous and homozygous mutations 15,16 .
Here, we describe a cost-effective approach to accurately, reliably, and efficiently genotype CRISPR/Cas9 alleles generated in model organisms that only differ by small in-dels. We adapted a PCR ligation detection reaction (PCR-LDR) as a sequence-specific genotyping method that can be completed quickly with a small number of steps 17,18 . With this assay, sequence variance is detected when perfectly matched base-pair probes are ligated at the nick junction of two synthetic oligonucleotides using high-fidelity DNA ligase (Fig. 1a). Allele-specific ligation products can be uniquely distinguished by using primers of different lengths and/or fluorescent tag modifications. LDR is straightforward to implement in most molecular biology laboratories, and has the potential to replace extant protocols for routine laboratory genotyping of model organisms.

PCR-LDR for genotyping.
Mouse lines with frameshift truncating mutations in Asxl3 were generated using CRISPR/Cas9 genome editing technology. sgRNA targeted to exon 12 of Asxl3 directed double strand breaks that underwent NHEJ DNA repair creating a series of small in-dels (Fig. 1b). Four founder lines were established that include deletion 1 (del1; a two-nucleotide CA deletion with T > G transversion), deletion 2 (del2; a two-nucleotide CA deletion), insertion 1 (ins1; a single nucleotide C insertion) and insertion 2 (ins2; a two-nucleotide CA insertion) ( Supplementary Fig. S1). All four alleles are transmitted in independent mouse lines.
Locus specific PCR amplification generates an amplicon that incorporates the edited region of Asxl3 exon 12 (Fig. 1c). PCR primers are designed to co-amplify wild type and mutant products from genomic DNA, with the edited regions in the middle of the amplicon. Amplicons are prepared for LDR after protease treatment to inactivate PCR enzymes. LDR achieves linear target product amplification by ligating a pair of target-complementary probes (Fig. 1a). Each probe pair consists of one common universal and one allele-specific probe. The common probe sequence is shared by all Asxl3 alleles and designed with a 5′ phosphorylation to enable successful ligation. The allele-specific oligonucleotide probes hybridize distinct Asxl3 mutations and can be uniquely modified by sequence tags that vary in nucleotide length and fluorescent moieties. Our size selection tag sequence was chosen based on lack of complementarity to genomic sequence in the mouse genome. The modifications allow ligation Allele-specific primers are designed to complement a common primer that recognizes all alleles similarly. Ligation products are created only when the allele-specific and common primers completely anneal to PCR products amplified from organism to be genotyped. (b) Asxl3 mouse alleles created with CRISPR/Cas9 genome editing. Cas9 nuclease (light sky blue) was targeted to Asxl3 Exon 12 by a single guide RNA (sgRNA). Target specific double strand DNA cleavage was repaired by non-homologous end joining creating in-dels [CA deletion and T > G transversion (del1), C insertion (ins1), CA deletion (del2), and CA insertion (ins2)] defined by Sanger sequencing. (c) Allele-specific and a common primer pair are designed to distinguish specific alleles within each PCR amplicon. The allele-specific and common primers are ligated on complementary amplicons by thermostable ligase (skyblue). Each allele-specific primer is designed with a 5′ size-selection linker of varying nucleotides to modulate the lengths of ligation products and fluorescently labeled FAM-(yellow circle) or TET-tag (green circle). The common primer is phosphorylated at the 5′ end. (d) Products of LDR reaction can be separated on an acrylamide gel and uniquely detected based on size and fluorophore.
products corresponding to specific alleles to be distinguished by both size and/or fluorescence following acrylamide gel electrophoresis. Unique product sizes and fluorophores also permit multiplexing for simultaneous detection of all expected alleles in a single reaction (Fig. 1d). Size-selection sequence tags on the Asxl3 allele-specific oligonucleotide probes vary by ~10 nucleotides and fluorescent moieties on the 5′ end alternated between FAM or TET (Supplementary Table 1). If there is target-complementarity between the template PCR strand and probes, ligation of the common and allele-specific probes will be catalyzed by the thermo-stable DNA ligase after repeated cycles of denaturation and ligation (Fig. 1a).

LDR reaction for genotyping Asxl3 mutant alleles.
To genotype Asxl3 mutant mice, a 110 bp PCR product was amplified across the site of CRISPR/Cas9 genome editing and separated on a 1.5% agarose gel. Despite the small size of the PCR product, small in-dels < 5 nucleotides could not be distinguished from the wild type based on mobility (Fig. 2a). Analysis of the PCR products by polyacrylamide gel electrophoresis (PAGE) was also not sufficient to uniquely identify the different alleles for genotyping purposes (Supplementary Fig. S2). Protease treated PCR products were incubated with a ligase master mix that includes a common probe and allele-specific probes that recognize in-dels del1, del2, ins1 and ins2. Ligation products were analyzed by electrophoresis on a 10% polyacrylamide gel. For each Asxl3 mouse line, the wild type, heterozygous and homozygous mutant ligation products for each allele were easily distinguished by PAGE (Fig. 2b). Allele-specificity is demonstrated by the different nucleotide sizes of ligation products generated for each allele. The sequence specificity of LDR is indicated by the lack of off-target ligation products from the ligation master mix containing all of the allele-specific probes. Analysis of ligation products indicated that different alleles are easily distinguishable, demonstrating specificity with no evidence of crossover or false positive ligation between ligation primers for different in-del variants. The genotyping strategy clearly differentiates between alleles that differ by a single nucleotide substitution, as in the comparison of alleles del1 and del2 (Fig. 2b, lanes 1-4).
In contrast to the PCR-LDR method, T7 Endonuclease I cleavage techniques fail to resolve WT or mutant homozygous alleles. An ~850 base pair Asxl3 exon12 PCR product was amplified across the site of genome editing (Fig. 2c). Denatured amplicons were allowed to reanneal with heteroduplexes forming between products with mismatched nucleotides (Fig. 3b). T7 Endonuclease I cleaved PCR products at sites of heteroduplex mismatch, and the resulting smaller fragments were resolved on an agarose gel. T7 Endonuclease I digestion efficiently detects heterozygous animals in all four unique Asxl3 lines, with smaller cleaved DNA fragments observed in the corresponding lanes. T7 Endonuclease I does not cleave PCR products in homoduplexes and only full length PCR products were detected in homozygous animals. Consequently, the technique fails to distinguish WT or mutant homozygotes (Fig. 2c). The T7 Endonuclease I genotyping strategy is also unable to distinguish between the distinct Asxl3 alleles, even in heterozygous animals, since the size of the cleavage products are the same for the different Asxl3 alleles.
Zebrafish LDR-PCR genotying. To test the general application of our technique between species, this genotyping strategy was used for a zebrafish line that was CRISPR/Cas9 genome-edited at a unique sequence in the zebrafish genome. The PCR-LDR genotyping strategy was established to detect a four base pair deletion (Z-del) segregating in the colony (Supplementary Fig. S3a). For the LDR reaction, common primers were designed with the 6FAM florescent moiety and allele-specific primers for Z-wt and Z-del alleles were designed with size-selection linkers (Supplementary Table S1). PCR primers were designed to specifically amplify 104 bp fragments flanking the mutated region (Fig. 2d). PCR products resolved on an agarose gel did not distinguish the genotypes of the samples by size. PCR products of Z-het samples could be resolved by size when analyzed by PAGE, yet homozygous Z-wt and Z-del samples were difficult to differentiate by this method (Supplementary Fig. S3b). Zebrafish samples were sequenced and analyzed by PCR-LDR genotyping. Ligation products of LDR and sample genotypes were easily resolved by PAGE (Fig. 2d). Likewise, PCR-LDR and Sanger sequencing identified all genotypes with 100% consistency, but heterozygotes were more clearly genotyped by LDR ( Supplementary Fig. S3).
LDR distinguishes multiplexed alleles. The efficiency of CRISPR/Cas9 genome editing permits simultaneous targeting of both alleles of the same gene. We therefore investigated the potential of LDR genotyping to detect multiple alleles in a single reaction. DNA samples containing Asxl3 alleles del1, del2, ins1, ins2, WT were combined and the PCR-LDR assay was performed as described. Asxl3 ligation products were compared with DNA samples from animals homozygous for Asxl3 for the corresponding alleles. The multiplexed PCR-LDR reaction reliably resolved distinct alleles, as depicted by the ligation products generated after acrylamide gel electrophoresis (Fig. 3a). The signal intensity of each band was sufficient for detection and all of the bands were of similar signal intensity, suggesting that all ligation products are equivalently generated. Because each allele can be resolved by simple visual examination, this strategy can also be employed for genotyping compound heterozygotes when different Asxl3 lines are intercrossed.

Discussion
PCR-LDR is an inexpensive and timely genotyping strategy capable of detecting small genetic changes in multiple alleles (Table 1). PCR-LDR uniquely detects individual alleles so that compound heterozygotes can be genotyped in a single reaction, which is difficult with T7 Endonuclease I or Sanger sequencing (Fig. 3b). The PCR-LDR genotyping strategy overcomes the deficiency in distinguishing between different homozygous genotypes encountered with T7 Endonuclease I assays. Because T7 Endonuclease I indiscriminately cleaves heteroduplexed PCR products with nucleotide mismatches, genotyping based on T7 endonuclease I must be periodically confirmed by sequencing, adding cost and time to experiments and colony management. While sequencing is unparalleled in specificity to define sequence differences between alleles, it is the most expensive and time-intensive genotyping option (Table 1). In addition, sequencing of heterozygous alleles can be difficult to interpret due to overlapping Scientific RepoRts | 6:32048 | DOI: 10.1038/srep32048 Asxl3 allele specific LDR reaction products are uniquely assigned by size and fluorophore following polyacrylamide gel electrophoresis. Bands at the top of the gene represent allele specific ligation products (colored allele specific primer with black common primer) that resolve wild type (WT), heterozygous, and homozygous samples. Shared common bands at bottom of the gel represent unligated primers (colored primers) present in the LDR master mix. The relative fluorophore and size variation of ligated and unligated products are graphically depicted on the left side of the gel. (c) A 850 bp amplicon was PCR amplified from Asxl3 modified locus. Asxl3 alleles carrying small in-dels are not resolved relative to the WT allele in either heterozygous or homozygous samples following electrophoresis on a 1.5% agarose gel. Renatured PCR products were incubated with T7 Endonuclease I to detect heteroduplex formation. This assay does not differentiate between WT and homozygous Asxl3 samples as both form homoduplexes not sensitive to T7 Endonuclease I cleavage. Cleavage products (asterisks) are evident in heterozygous samples that form heteroduplexes between mismatched Asxl3 alleles. (d) A 104 bp amplicon was PCR amplified and products checked on 1.5% agarose gel. Ligation product for wt (Z-wt), heterozygous (Z-het) and homozygous 4 bp deletions (Z-del) in a unique sequence in the zebrafish genome were generated by PCR-LDR and uniquely identified by size following PAGE. Ligated LDR reaction products are uniquely identified by size at the top of the gel. The band at bottom of the gel is the unligated 6FAM labeled common primer in the LDR master mix.
Scientific RepoRts | 6:32048 | DOI: 10.1038/srep32048 fluorochrome peaks in chromatographs, a problem amplified for compound heterozygote genotypes. PCR-LDR is an inexpensive genotyping strategy that can differentiate between multiple alleles that differ only by small in-dels or single nucleotide polymorphisms. The LDR strategy is valuable for routine genotyping, however, it should be noted that it cannot be used for screening initial mutations as it requires prior sequence information. While the utility of the LDR genotyping strategy has been demonstrated for CRISPR/Cas9 mutated genomes, this strategy should be broadly applicable for other small mutations created by genome editing techniques such as TALENs and ZFNs. The LDR approach for genotyping small variations created by different nucleases has the potential to become an essential laboratory tool for regularly maintaining alleles and managing colonies. Genomic DNA extraction. Genomic DNA was isolated from mouse-tail biopsies using simple lysis-extraction procedures. Each tail clipping was lysed in 100 μ l of lysis buffer (NaOH, 25 mM, Na 2 EDTA.2H 2 O, 0.2 mM) at 95 °C for 30 mins-1 hr, and neutralized with an equal volume of Tris-Cl (40 mM), and briefly centrifuged. 1 μ l of final supernatant was used for each PCR reaction. Table S1) were designed to amplify a product size of ~850 bp flanking the targeted mutation site. T7 Endonuclease I assay was performed on the PCR products according to the manufacturer's instruction (Ipswich, MA, USA). The reaction products were resolved using 1.5% agarose gel electrophoresis.

T7E1 Assay. PCR primers (Supplementary
PCR and Probe Primer Design. PCR primers (Supp. Table1) were designed to amplify a region of Asxl3 exon 12 with the product size of ~110 bp. The genotyping specificity of the LDR assay is attained by ensuring the stringency of primer design and reaction conditions. LDR probes were designed to distinguish specific alleles within each PCR amplicon by the identity of the fine nucleotides in the probe. The sequence of each allele-specific primer differs for each allele and is labeled at the 5′ end with a 6FAM or TET fluorescent moiety. The inclusion of different fluorophore dyes on the specific 5′ primers enables simultaneous multicolor detection of the alleles. The locus-specific common LDR primers were phosphorylated at the 5′ end. Linker nucleotides were appended to the discriminating primers at their 5′ end to modulate the lengths of the expected ligation products. Both PCR and LDR primers were obtained from Integrated DNA Technologies (Coralville, IA).