Single base editing of rs12603332 on Chromosome 17q21 with a Cytosine Base Editor regulates ORMDL3 and ATF6α expression

Abstract

Background:

Genetic association studies have demonstrated that the SNP rs12603332 located on chromosome 17q21 is highly associated with the risk of the development of asthma.

Methods:

To determine whether SNP rs1260332 is functional in regulating levels of ORMDL3 expression we used a Cytosine Base Editor (CBE) plasmid DNA or a CBE mRNA to edit the rs12603332 C risk allele to the T non-risk allele in a human lymphocyte cell line (i.e. Jurkat cells) as well as in primary human CD4 T cells that carry the C risk alleles.

Results:

Jurkat cells with the rs12603332 C risk allele expressed significantly higher levels of ORMDL3 mRNA, as well as the ORMDL3 regulated gene ATF6α as assessed by qPCR compared to Jurkat clones with the T non-risk allele. In primary human CD4 T cells, we edited 90 ± 3 % of the rs12603332-C risk allele to the T non-risk allele and observed a reduction in ORMDL3 and ATF6α expression. Bioinformatic analysis predicted that the non-risk allele rs12603332-T could be the central element of the E-box binding motif (CANNTG) recognized by the E47 transcription factor. An EMSA assay confirmed the bioinformatics prediction demonstrating that a rs12603332-T containing probe bound to the transcription factor E47 in vitro.

Conclusions:

SNP rs12603332 is functional in regulating the expression of ORMDL3 as well as ORMDL3 regulated gene ATF6α expression. In addition, we demonstrate the use of CBE technology in functionally interrogating asthma-associated SNPs using studies of primary human CD4 cells.

Graphical Abstract

• SNP rs12603332 on chromosome 17q21 is highly associated with the development of asthma.

• Using a Cytosine Base Editor, the rs12603332 C risk allele was edited to the T non-risk allele in primary human CD4 T cells.

• This change regulated the expression of ORMDL3 and ORMDL3 regulated gene ATF6α expression.

INTRODUCTION

ORM1 like 3 (ORMDL3) located on chromosome 17q21 is highly linked to asthma in genetic association studies^1,2. SNPs in the ORMDL3 gene have been associated with expression levels of ORMDL3 in Epstein–Barr virus-transformed lymphoblastoid cell lines from asthmatic children¹. ORMDL3 has 4 exons and 3 introns and codes for a 153 amino acid protein expressed in the endoplasmic reticulum of several cell types important to the pathogenesis of asthma including CD4 cells¹, epithelial cells³, airway smooth muscle cells⁴, and eosinophils⁵. ORMDL3 is expressed at increased levels in the lungs and lung cells of individuals with asthma^1,6,7, while transgenic mice which express increased levels of human ORMDL3 have an asthma phenotype characterized by increased airway responsiveness⁸ and increased airway remodeling (increased airway smooth muscle, increased peribronchial fibrosis, increased mucus)⁸.

This study of ORMDL3 and chromosome 17q21 focuses on SNP rs12603332 (located in intron 1 of ORMDL3) as several studies have reported that the C risk allele for rs12603332 was significantly associated with asthma in different ethnic groups including Caucasians^1,9, Hispanics¹⁰, African Americans¹⁰, and Han Chinese¹¹, but not Iranian Northwestern Azeris¹². Although these genetic association studies established that the SNP rs12603332 is linked to asthma, as yet no studies have directly demonstrated that the SNP rs12603332 is functional and regulates levels of ORMDL3. For example, in genetic association studies, linkage disequilibrium patterns can result in multiple SNPs at a locus being associated with an asthma phenotype, even if only one of the SNPs is causal. The 17q21 locus harbors a dense haploblock of 136 SNPs in tight linkage disequilibrium that overlap six gene loci including IKZF3, ZPBP2, GSDMB, ORMDL3, LRRC3C and GSDMA¹³, and thus not all the SNPs in this region may be functional in regulating levels of ORMDL3. As SNP rs12603332 is located in a non-coding region in intron 1 of ORMDL3, we hypothesized that it is likely to be regulatory if functional.

In this study we used single base editing to test the hypothesis that the C risk allele of SNP rs12603332 on chromosome 17q21 regulates levels of ORMDL3. Two methodologic approaches (CRISPR-Cas9, or a Cytosine Base Editor) have been utilized to carry out single base editing of genomic DNA from C to T¹⁴. In this study we used a Cytosine Base Editor (CBE) approach¹⁵ to change the C risk allele for rs12603332 to the T non-risk allele in human T cells (Jurkat and primary human CD4 cells) as CD4+ T cells are important to the pathogenesis of asthma and CD4 cells express ORMDL3¹. To precisely edit the C risk allele for rs12603332 to the T non-risk allele in the genome of live cells, we used a CBE, which uses a catalytically impaired CRISPR-associated nuclease [nCas9(D10A)] complexed with a guide RNA (gRNA) for sequence-specific genomic DNA targeting. This ribonucleoprotein complex is additionally fused to the cytosine deaminase enzyme rAPOBEC1 for targeted C-U conversion, and two tandem Uracil glycosylase inhibitor (UGI) units to inhibit Uracil-DNA glycosylase (UNG)¹⁵ (Figure 1). The rAPOBEC1 component of CBE changes the rs12603332 C allele into a U, and the nCas9 nicks the opposite strand to bias DNA repair factors to use the edited U as a template during downstream DNA repair. The UGI domain “protects” the U by inhibiting the enzyme UNG, further biasing DNA repair mechanisms to repair the nicked strand (containing the G) and keep the U-containing strand (originally a C), ultimately changing a C: G basepair to T: A following DNA replication (Figure 1). In this study we have therefore used a CBE to perform single base editing of C to T to test the hypothesis that the C risk allele of SNP rs12603332 on chromosome 17q21 regulates levels of ORMDL3, as well as ORMDL3 downstream gene Activating Transcription Factor α (ATF6α)³.

Figure 1: Genomic single base editing of chromosome 17q21 SNP rs12603332 with Cytosine Base Editor.

A. The BE4 Cytosine Base Editor (CBE) has a rAPOBEC1 deaminase (labeled rAPOBEC1), Cas9 D10A nickase (labeled Cas9) and two Uracil Glycosylase Inhibitor (UGI) in tandem. When coupled with a gRNA, the Cas9 opens up the DNA double helix and nicks the strand base paired with the gRNA, while rAPOBEC11 edits Cytosines (C) on the unpaired strand to Uracils by deamination. B. The ORMDL3 gene is located in a cluster of 9 genes localized to the chromosome 17q12–21 region. C. The human ORMDL3 gene has 4 exons and 3 introns with SNP rs12603332 in intron 1. D. SNP rs12603332 has a risk allele C and a non-risk allele T. E. The SNP rs12603332 risk allele C is highlighted in red. The designed gRNA recognizes a 20bp genomic DNA (underlined, called the protospacer) with a GGG PAM (marked in grey). F. CBE edited rs12603332C to U, and with its Cas9 D10A nickase, CBE created a nick on the complementary strand (the strand that is not edited). G. With UGI (Uracil Glycosylase Inhibitor), DNA mismatch repair mechanism preferably replaces the nicked strand, and changes the U:G mismatch to a U:A pair. H. During DNA replication, the U:A pair is permanently altered to a T:A pair.

METHODS

Jurkat T Cells

The Jurkat T cell clone E6–1 cell line was purchased from ATCC (Catalog # TIB-152) and maintained in RPMI 1640 media supplemented with 1% penicillin/streptomycin, 10% FBS and L-glutamine.

Human Peripheral Blood CD4 cells

Human peripheral blood leukopaks (total n=46 different donors) were purchased from either HemaCare (Northridge, CA) (n=15 different donors) or iXSells Biotechnologies (San Diego CA) (n=31 different donors) who collected the leukopaks under either a Hemacare or iXSells Biotechnologies IRB-approved donor consent. Aliquots of the leukopaks were frozen in liquid nitrogen and thawed upon use for CD4 cell isolation. Human primary CD4+ cells were isolated and purified from the leukopak aliquots using a human CD4 T cell negative selection kit (StemCell Technologies). CD4+ cells were maintained in ImmunoCult human T cell expansion medium (StemCell Technologies) with 100 unit/ml human recombinant IL-2 (R&D Systems). To activate CD4+ T cells, 1:40 ImmunoCult human CD3/CD28 T cell activator (StemCell Technologies) was added to the medium, and expanded cells were diluted to 1 million cells per ml at day 3 and day 5 after activation by adding more ImmunoCult human T cell expansion medium with IL-2 (complete medium). CD4+ T cells were cultured for 3–5 days before transfecting with the CBE single base editor mRNA and gRNA.

CBE plasmid DNA preparation for Single Base Editing of Chromosome 17q21 SNP rs12603332

We used the cytidine base editor (i.e. BE4) to convert a target C:G base pair to T:A. The rAPOBEC1-nCas9(D10A)-UGIx2-P2A-GFP (or BE4-P2A-GFP) was obtained from Addgene (Watertown, MA). To minimize unwanted random RNA editing and unguided DNA editing, we swapped the rAPOBEC1 with the Selective Curbing of Unwanted RNA Editing (SECURE) version of the deaminase¹⁴. P2A is a self-cleaving peptide that allows simultaneous expression of the CBE (i.e. BE4) and GFP under one promotor. GFP thus becomes a selection marker to monitor CBE transfection. The single guide RNA (sgRNA) expressing plasmid EMX1 was derived from Addgene plasmid #47511 and uses a U6 promoter to drive high level of sgRNA expression. The sgRNAs were cloned into the EMX1 backbone using a PCR method as described¹⁵. The BE4-P2A-GFP and the sgRNA plasmids were purified with maxiprep kits (ZYMO Research) and filtered by Endozero columns to remove residual endotoxins that could affect transfection. The plasmids were stored at −20°C until use.

BE4-P2A-GFP in vitro transcription mRNA and sgRNA synthesis.

BE4-P2A-GFP mRNA was synthesized by HiScribe T7 ARCA mRNA kit (New England Biolab) following the manufacturer’s instructions. The quality of the synthetic mRNA was checked by Tapestation. The modified sgRNA that targets SNP rs12603332 was synthesized by Synthego (Menlo Park, California). The in vitro transcription (IVT) mRNA and sgRNA were reconstituted in RNase-free water and aliquoted for long-term storage at −80°C.

Genomic single base editing of chromosome 17q21 SNP rs12603332 with BE4-P2A-GFP in Jurkat cells

We cloned the 20nt guide RNA (gRNA) sequence to the sgRNA expressing plasmid EMX1. Jurkat T cells were resuspended to 0.5 ×10⁶ cells/mL one day before electroporation to maximize editing efficiency. On the day of the experiment, 2.2 × 10⁵ Jurkat cells were washed in PBS, diluted in 10 μL Resuspension Buffer R (Neon Transfection System, Thermo Fisher Scientific), and added to 0.8μg BE4-P2A-GFP and 0.3 μg sgRNA plasmid in 1.2μL TE buffer. We used program #16 (1350V, 10ms pulse width, and 3 pulses) on the Neon Transfection System (Thermofisher) in a 10μL tip to transfect Jurkat cells. Electroporated cells were immediately mixed with 1 mL pre-warmed media in a 24-well plate. Two days after transfection, live cells (staining 7-AAD negative) that were positive for GFP were FACS sorted by BD AriaII into a U-bottom 96-well plate, with 200 uL complete medium (10% FBS, 1% Penn/Strep, 10mM HEPS in RPMI-1640) in each well. Cell colonies were visible in about 10% of the wells after 2 weeks. These clones were sequenced and the clones that were positive for rs12603332 C-T editing were expanded.

Genomic single base editing of chromosome 17q21 SNP rs12603332 in primary human CD4 cells with CBE

Primary human CD4 cells were washed twice by PBS, and resuspended in Buffer R (Thermo Fisher). 2×10⁶ CD4 cells were mixed with 5μg BE4-P2A-GFP mRNA and 1.5μg one-piece sgRNA in a 100μL tip, and electroporated by Neon Transfection system with program #20 (1600V, 10ms and 3 pulses). CD4 cells were then suspended in 2ml pre-warmed complete medium (ImmunoCult human T cell expansion medium with 100 unit/ml human recombinant IL-2), and cultured in 37°C incubator for a day. GFP positive CD4+ live cells were sorted, washed and cultured in complete medium in the presence of ImmunoCult CD3/CD28 T cell activator. CD4 cells were cultured for another 7 days before harvesting for PCR and qPCR experiments detailed below. As negative controls, primary CD4 cells were also transfected with either BE4-P2A-GFP mRNA alone or just the transfection reagent (Buffer R).

PCR and Sanger sequencing for chromosome 17q21 SNP rs12603332 detection

Genomic DNA samples were extracted by Quick-DNA Microprep Plus kit (Zymoreserch), according to the manufacturer’s protocol. TaqMan SNP genotyping assay was performed using predesigned qPCR primers (Thermofisher) on a QuantStudio^™ 3 Real-Time PCR System (Thermofisher). The rs12603332 genotype (could be either CC, TT or TC) of each sample was determined by the QuantStudio Design and Analysis Software available with the qPCR machine (Thermofisher). One μL of purified DNA samples were amplified by a PCR Master Mix (Thermofisher) with forward primer GCC CCA GCC AAT TGG GAA AT and reverse primer CCT GCC TCC AAA ACC TAG CA. The PCR products were then sequenced at GeneWiz (La Jolla, CA) using a Sanger sequencing primer TAC CAA ATTA GTC GGG GGT G. Trace files from Sanger sequencing were used to determine genomic editing and their frequencies with ICE CRISPR analysis online tool (Synthego).

qRT-PCR: ORMDL3, and ORMDL3 regulated gene ATF6α mRNA

We assessed the mRNA expression of chromosome 17q21 gene ORMDL3 as well as mRNA expression of ORMDL3 downstream gene ATF6α³ in the Jurkat human T cell line and in primary human CD4 T cells by qRT-PCR. Cells were harvested and washed with PBS, then stored in RNA-STAT-60 (Tel-Test) at −80°C. Total RNA was extracted by Quick-RNA Miniprep kit (Zymoreserch) and reverse transcribed with Maxima reverse transcription kit (Thermofisher). qRT-PCR was performed with Gene Expression Master Mix and either ORMDL3 or ATF6α primers (synthesized by Thermofisher). The relative amounts of mRNA transcripts were normalized to those of housekeeping gene HPRT mRNA and compared between the different genes by the ΔΔ cycle threshold method as previously described in the laboratory⁸.

qRT-PCR: TGFβ1 and Th2 cytokine mRNA in primary human CD4 cells

As we have previously demonstrated that ORMDL3 transgenic mice expressing increased levels of human ORMDL3 express increased levels of TGFβ1⁸, we have used qRT-PCR as described above to quantitate levels of TGFβ1 in primary human CD4 T cells. In addition, as another SNP rs8076131 on chromosome 17q21 has been linked to expression of Th2 cytokines¹⁶ we also used qRT-PCR to quantitate levels of Th2 cytokines including IL4, IL5, and IL13 in primary human CD4 T cells. Catalog numbers of all the primers used are described in the Supporting Information.

Bioinformatic Analysis of potential Transcription Factor binding to Chromosome 17q21 rs12603332 sequence.

The online transcription factor analysis tool TRANSFAC professional version (http://genexplain.com/transfac) was used to predict potential transcription factor binding sites in the chromosome 17q21 rs12603332 region. A 100bp long DNA sequence surrounding the rs12603332 SNP was surveyed and TRANSFAC matrix table version 2019.32019.3 was used to match transcription factor binding sites.

Electrophoretic mobility shift assay (EMSA)

The infrared dye IR700 labeled oligos and the unlabeled competitor DNA oligos corresponding with both the chromosome 17q21 rs12603332-T and rs12603332-C were synthesized by IDT DNA (San Diego, CA). The oligo pairs were annealed to create duplexes and kept at −20°C until use. Human recombinant E47 protein (a member of the helix-loop-helix family of transcription factors)^17,18 and E47 antibody (catalog number TA306926) were purchased from Origene. 50 ng E47 protein was incubated with either 50 fmol of the labeled rs12603332-T probe, or 50 fmol labeled rs12603332-T probe combined with 2.5 pmol (50x) competing unlabeled rs12603332-T probe, for 20 minutes at room temperature in a 1x binding buffer (10mM Tris, 50mM KCl, 2.5mM DTT, 0.25% Tween-20, 2.5% glycerol, 5 mM MgCl₂, 50 ng/μL Poly(dI*dC), pH 7.5). Experiments were also performed in which the E47 protein was incubated with either 50 fmol of the labeled rs12603332-C probe, or 50 fmol labeled rs12603332-C probe combined with 2.5 pmol (50x) competing unlabeled rs12603332-C probe, as described above for the rs12603332-T probe.

In “supershift” experiments 400ng E47 antibody was added to the mixture and incubated for another 20 minutes. Samples were then loaded to a 6% TBE polyacrylamide gel (Novex, Thermofisher) in native 0.5x TBE buffer, electrophoresed at 4°C until the orange loading dye (Licor, Lincoln, Nebraska) migrated out of the gel. The gel cassette was then imaged on a Licor Odyssey Imaging System. Exposure time was set to 1 minute.

Statistical analysis

All comparisons were analyzed using Student’s t-test, while ANOVA was used for multiple comparisons. Results are reported as the mean and standard error of the mean (SEM). Statistical analysis was performed using Prism software (Graphpad, La Jolla, CA) and a two-sided P value of 0.05 was considered statistically significant.

RESULTS

Single Base Editing of chromosome 17q21 SNP rs12603332 in Jurkat human T cell line

We sequenced the genomic DNA from Jurkat T cells and demonstrated that the original Jurkat cells carry the chromosome 17q21 SNP rs12603332 risk allele C/C located in ORMDL3 intron 1 (Figure 1 B–D). The Jurkat rs12603332-C has a downstream GGG protospacer adjacent motif (PAM) site (Figure 1E) which is required for Cas9:gRNA binding to a locus of interest. The cytosine base editor BE4 has a narrow editing window and facilitates maximal editing efficiency when the target C is located in positions 5–8 within the protospacer (counting the PAM-distal end as position 1, and the PAM site as positions 21–23). We therefore chose the protospacer/PAM sequence in Figure 1E which places the rs12603332-C at C6 position, as this positioned all nearby Cs (potential “bystander” Cs that could get edited along with the target C) outside of the CBE editing window as compared to other potential protospacers. Finally, we also used the recently engineered Selective Curbing of Unwanted RNA Editing (SECURE) variant of BE4 to minimize unwanted RNA editing to background levels¹⁴.

After co-transfection of Jurkat cells with the BE4-P2A-GFP and sgRNA plasmids, the GFP fluorescence signal was detected in Jurkat cells after 6 hours, with the GFP fluorescence intensity peaking at day 2 in approximately 40% of the live Jurkat T cells, and was still detectable for over a week.

Sanger sequencing of chromosome 17q21 SNP rs12603332 in FACS sorted GFP⁺ Jurkat T cells

We FACS sorted GFP positive live Jurkat cells (7-AAD negative) into a 96-well U bottom plate to form single-cell colonies. In addition, we sorted the rest of the GFP positive live Jurkat cells into a collection tube to evaluate the base-editing efficiency of the batch. We extracted DNA from the sorted Jurkat cells, amplified by PCR and sequenced the chromosome 17q21 SNP rs12603332 region. On the Sanger sequencing trace file, we observed a mixture of T and C at rs12603332 position, with no editing of nearby cytosines, suggesting the editing was precise and effective. Since we edited rs12603332 C to T, the ratio of T vs T + C would be the editing efficiency. We compared the sequence trace files before and after editing with the ICE CRISPR HDR analysis online tool and demonstrated the editing efficiency for the batch at rs12603332 was 91%, with off-target “bystander” editing of Cs adjacent to rs12603332 below the detection limit (Figure 2 A).

Figure 2. CBE edited chromosome 17q21 SNP rs12603332 in Jurkat cells.

A. ICE analysis of the Sanger sequencing chromatogram shows that after electroporation with CBE and gRNA plasmids, 91% of the DNA from live, GFP-positive Jurkat cells has rs12603332-T, with only 9% of DNA containing the rs12603332-C. B. Sanger sequencing chromatograph demonstrating that the original Jurkat cells carry the SNP 12603332-C risk allele (left panel). In 8 out of 11 clones, CBE treatment of Jurkat cells permanently changed the rs12603332-C to T on both chromosomes (right panel). Note that the two “bystander” Cs remain unedited. C. Levels of mRNA (ORMDL3, left panel. ATF6α, right panel) were quantified by qPCR in Jurkat clones with the non-risk allele rs12603332-T and the risk allele rs12603332-C. (n=4/group) *p<0.05. Statistical analysis was performed with a two tailed T test.

After incubating the FACS-sorted GFP-positive live Jurkat cells (7-AAD negative) for 2–3 weeks in the 96-well expansion plate, we counted 11 clones growing from single cells. After sequencing, we demonstrated that 8 of the 11 Jurkat cell colonies were positive for rs12603332 C/C to T/T editing (Figure 2 B). The remaining 3 clones were rs12603332 C/C (no heterozygous T/C clones were detected). As we only had 4 C/C clones (3 not edited and 1 original), to match the numbers of T/T clones with the number of C/C clones (n=4) we randomly selected 4 T/T clones from the 8 edited T/T clones to form the T/T group. As can be seen in Figure 2C, the original one Jurkat C/C line functioned similarly to the three non-edited Jurkat C/C clones.

Expression of chromosome 17q21 gene ORMDL3 in Jurkat T cells with base edited SNP rs12603332

We performed qRT-PCR to determine whether single base editing of the chromosome 17q21 SNP rs12603332 from the risk allele (C) to the non-risk allele (T) influenced mRNA expression levels of ORMDL3 in Jurkat T cells. Single base editing of the chromosome 17q21 SNP rs12603332 from the risk allele (C) to the non-risk allele (T) regulated ORMDL3 mRNA expression, as levels of ORMDL3 mRNA was significantly increased in Jurkat cells with the rs12603332-C risk allele compared to the non-risk allele (p= 0.03) (Figure 2C). The expression of a known ORMDL3 regulated downstream gene (i.e. ATF6α)³ was also modulated, as ATF6α mRNA expression was also significantly lower in Jurkat cells with the rs12603332-T non-risk allele compared to the C risk allele (p= 0.02)(Figure 2C).

Single Base Editing of chromosome 17q21 SNP rs12603332 in primary human CD4 cells

To edit chromosome 17q21 SNP rs12603332 in primary human CD4 cells from the C/C (risk allele) to the T/T (non-risk allele) we initially needed to identify homozygous C/C donors. Our screening, in which we sequenced SNP rs12603332 in 46 different healthy normal human subject donors, identified five of these donors (Supporting Information, Table 1) as homozygous for C/C at rs12603332. These 5 donors were thus used in our CD4 cell editing experiments.

We used CBE mRNA rather than a CBE plasmid DNA to edit the rs12603332 C risk allele to the T non-risk allele in primary human CD4 T cells as pilot studies demonstrated that the CBE plasmid mRNA editing efficiency (approximately 90%) was greater rather than the CBE plasmid DNA editing efficiency (approximately 40%). We initially electroporated in vitro transcribed BE4-P2A-GFP mRNA and synthetic sgRNA into the primary human CD4 cells of donor 1. One day after electroporation, we observed a GFP signal in 70% of the live human CD4 cells (Figure 3A). Those cells were sorted and cultured in complete medium with Immunocult CD3/28 T cell expander. Seven days after sorting, we sequenced the edited cells and determined that 90 ± 3 % (n=3) of the rs12603332-C had been edited to T, while bystander editing near rs12603332 was not detectable (Figure 3B and 3C), suggesting that the CBE complex worked very efficiently in primary human CD4 cells. Similar results were obtained with the other 4 donors. Seven days after editing primary human CD4 cells with CBE plasmid mRNA to the non-risk T allele (n=5 donors), we detected significantly reduced levels of ORMDL3 mRNA expression (Figure 3D), as well as significantly reduced levels of the ORMDL3 regulated gene ATF6α mRNA expression (Figure 3D). In contrast, control primary CD4 cells (n=5 donors) transfected with either BE4-P2A-GFP mRNA alone (with no sgRNA), or just the transfection reagent (Buffer R) had no reductions in ORMDL3 or ATF6α mRNA expression (Figure 3D).

Figure 3. Cytosine Base Editor editing of chromosome 17q21 SNP rs12603332 in human primary CD4 T lymphocytes.

A. Flow cytometric analysis of human primary CD4 T cells (donor 1) not transfected with CBE mRNA containing EGFP (left panel), demonstrating that all cells are EGFP negative. After electroporation with CBE mRNA containing EGFP (right panel), approximately 70% of the 7-AAD live cells were EGFP positive, representing successful CBE delivery. B. Sanger sequencing demonstrates editing of rs12603332 (red arrow) in human primary CD4 cells (donor 1) from C (upper panel, original sample) to mostly T (lower panel, base edited sample) by CBE. Note that bystander Cs within the protospacer near rs12603332 are not edited. Neighboring Cs outside of the protospacer also remain unedited. C. ICE analysis of the Sanger sequencing chromatogram shows that 93% of the DNA in these human primary CD4 T cells (donor 1) had the non-risk allele rs12603332-T (red arrow), whereas the risk allele rs12603332-C only consisted of approximately 7% of the total amplified DNA. D. Levels of ORMDL3 and ATF6α mRNA were quantitated by qPCR in human primary CD4 T cells (n= 5 blood donors) that were either untreated, mock treated (electroporation buffer only), control treated (CBE mRNA only; no gRNA), or CBE edited (CBE mRNA and gRNA). Results of each of the individual 5 donors are represented in the figure as the mean of a triplicate experiment. CD4 cells with the non-risk allele rs12603332-T (labeled edited) had significantly reduced levels of ORMDL3 and ATF6α mRNA expression compared with control samples (untreated, mock treated and control treated; n=5/group) *p<0.05, **p<0.01. Statistical analysis was performed with a two way ANOVA.

Editing of the chromosome 17q21 SNP rs12603332 in primary human CD4 cells (n=5 donors) showed a trend to regulate levels of TGFβ1(p=0.09)(Supporting Information, Figure 1A), but did not regulate levels of Th2 cytokines (IL4, IL5, IL3)(Supporting Information, Figure 1B–D).

Chromosome 17q21 SNP rs12603332 is a binding site for the Transcription Factor E47

We investigated potential molecular mechanisms to determine how the chromosome 17q21 SNP rs12603332 polymorphism alters the expression of ORMDL3 mRNA. We hypothesized that the palindromic sequence CANNTG (N can be any base pair) where rs12603332-T is located in chromosome 17q21 could be a possible transcription factor binding site. We used the TRANSFAC online tool¹⁹ to analyze both sequences containing rs12603332-T and rs12603332-C. The results demonstrated that transcription factor E47 could bind to the CANNTG sequence within rs12603332-T (Figure 4A), but not to the sequence within rs12603332-C. There is a 100% match in the 6 base pair CANNTG sequence within rs12603332-T with the Ebox CANNTG motif that binds E47²⁰.

Figure 4. Chromosome 17q21 SNP rs12603332 is a binding site for the Transcription Factor E47.

A. The TRANSFEC bioinformatic tool demonstrated that rs12603332-T could be contained within a transcription factor E47 binding site with the T located in the core binding site of the CANNTG motif (N can be any base pair). B. EMSA: Lane 1–3 examines E47 binding to rs12603332-C DNA. Lane 1 contains the labeled rs12603332-C DNA probe alone, Lane 2 contains the labeled rs12603332-C DNA probe and E47, and Lane 3 contains the labeled rs12603332-C DNA probe, E47 and the non-labelled competitor DNA probe. None of Lanes 1–3 exhibited any E47 binding to rs12603332-C DNA. Lane 4–7 examines E47 binding to rs12603332-T DNA. Lane 4 contains the rs12603332-T DNA probe alone (No E47; No non-labelled competitor DNA probe), Lane 5 contains the rs12603332-T DNA probe and E47 (No non-labelled competitor DNA probe), and Lane 6 contains the rs12603332-T DNA probe, E47 and E47 antibody, while Lane 7 contains the rs12603332-T DNA probe, E47 and the non-labelled competitor DNA probe. Lane 5 demonstrated that recombinant human E47 slowed down the migration of the rs12603332-T DNA probe, while in Lane 6 with the E47 antibody the migration was delayed even further. In Lane 7, with the non-labelled competitor DNA probe, the binding of labeled probe and E47 was abolished.

To confirm the bioinformatic findings, we performed Electrophoretic Mobility Shift Assay (EMSA) to demonstrate the binding of rs12603332-T to E47 (Figure 4B). The results demonstrated that in vitro, recombinant E47 delayed the migration of the infrared dye IR-700 DNA probe with rs12603332-T suggesting that E47 can interact with the rs12603332-T non-risk allele.

The addition of a competing unlabeled rs12603332-T probe prevented binding of the labeled rs12603332-T probe to E47 (Figure 4B), while the addition of an anti-E47 antibody induced a supershift (Figure 4B). Moreover, E47 only bound to the rs12603332-T non-risk allele and not to the rs12603332-C risk allele (Figure 4B) which only differ from each other by one base pair.

DISCUSSION

The 17q21 locus harbors a dense haploblock of 136 SNPs in tight linkage disequilibrium¹³ some of which SNPs may or may not be functional in regulating ORMDL3 expression. As the functional status of SNP rs12603332 linked to asthma is at present unknown, in this study we used a novel approach to edit the chromosome 17q21 SNP rs12603332 from the risk allele C to the non-risk allele T using CBE mRNA, a new genome editing methodology which has not been reported in studies of SNPs in any primary human cell type including CD4 cells in asthma. Using CBE to precisely edit a single base on chromosome 17q21 in the SNP rs12603332 from the risk allele C to the non-risk allele T significantly reduced levels of ORMDL3 mRNA expression, indicating that SNP rs12603332 (located in intron 1 of ORMDL3) is functional in terms of regulating levels of ORMDL3. ORMDL3 SNP rs12603332 not only regulated levels of ORMDL3 but also regulated levels of the ORMDL3 downstream gene ATF6α³. As asthma is associated with increased levels of ORMDL3 mRNA expression in the lung⁷, ORMDL3 SNPs such as rs12603332 may explain why the C allele of this SNP is the risk allele as it increases levels of ORMDL3 and its downstream gene ATF6α.

In addition, using bioinformatic tools and EMSA we demonstrated that the 6 base pair sequence (CANNTG) surrounding rs12603332-T in chromosome 17q21 was a binding site for the transcription factor E47. ORMDL3 gene transcription has previously been noted to be cooperatively regulated by multiple transcription factors in vitro, including CTCF¹³, USF¹⁶, Ets-1²¹, p300²¹, and CREB²¹, but ORMDL3 has not previously been linked to E47 a transcription factor encoded by the E2A gene^17,18. E47 recognizes a DNA consensus sequence known as the E-box (CANNTG) and plays a crucial role in T cell growth, survival and development²². In thymocytes, E47 decreased the expression of CD25 (the α chain of the IL2 receptor), a process that is critical for T lineage maturation²³. In addition, E47 can repress the expression of cell cycle regulators like Cdk6 and Rb²³. As E47 is known to inhibit the expression of several genes (i.e. CD25, Cdk6, Rb) our studies suggest that ORMDL3 may be another gene target for inhibition by E47. For example, we demonstrated that E47 binds to the E-box (CANNTG) motif in vitro present in the chromosome 17q21 non-risk allele SNP rs12603332-T (associated with reduced ORMDL3 expression), but does not bind to the risk allele SNP rs12603332-C (associated with increased ORMDL3 expression). Further study is needed to directly demonstrate that binding of E47 to the non-risk allele SNP rs12603332-T is present in vivo in asthmatics with this SNP, and how this E47 binding inhibits ORMDL3 expression.

In silico analysis of transcription factor binding to rs12603332 has also been performed by other groups²⁴ using the on-line Promo tool²⁴ (as opposed to the TRANSFAC tool used in this study)¹⁹. These prior studies of rs12603332 demonstrated differential transcription factor binding to either the predicted major or minor allele of rs12603332 for AP-2αA and E2F-1 to the C allele, and for USF2 to the T allele²⁴. In addition, this prior study²⁴ analyzed allele-specific effects for four ORMDL3 linked SNPs i.e. rs12603332 (ORMDL3, bin 1), as well as rs8079416 (ORMDL3, bin 15), rs4795405 (ORMDL3, bin 5), and rs3902920 (ORMDL3 region, bin 1) on levels of ORMDL expression in PBMCs of the non-asthmatic subjects studied. These results demonstrated that in unstimulated PBMCs only one of these four SNPs (rs8079416) had a statistically significant association with levels of ORMDL3 expression at the p<0.05 threshold, whereas the p value for rs12603332 trended (p=0.11) but did not reach significance²⁴. In contrast, in our study using single base pair editing of rs12603332 in unstimulated CD4 cells we were able to demonstrate that rs12603332 regulates levels of ORMDL3 expression. It is likely that differences in methods used in the two studies (SNP association with levels of ORMDL3 in prior study²⁴; single base pair SNP editing in this study to prove SNP causation) contributed to the p value for rs12603332 trending (p=0.11) but not reaching significance in the prior study. The prior study²⁴ also differs from our study in investigating ORMDL3 expression in PBMCs (mixture of different subsets of T cells, mononuclear cells) which differs from this study of purified CD4 cells which have undergone single base pair editing.

Previous studies of other SNPs in chromosome 17q21 have investigated SNP rs4065275 and rs8076131 located in the intron 1 region of ORMDL3¹⁶. Although the SNP rs12603332 we examined is also located in intron 1 of ORMDL3, the SNP rs12603332 investigated in this study is approximately 1.9kb away from rs4065275 and rs8076131 which are located 47bp apart¹⁶. Studies of SNP rs4065275 and rs8076131 have demonstrated that their risk alleles regulate downstream luciferase activity in studies utilizing transfection of >1,000 bp of the ORMDL3 promoter region containing either the risk or non-risk allele of either SNP rs4065275 or rs8076131 into a human embryonic kidney cell line (HEK293 cells)¹⁶. In contrast, in this study we used a novel alternative single base editing approach with CBE to edit the SNP of another chromosome 17q21 SNP (i.e. rs12603332) from a risk allele to a non-risk allele in both a human T cell line (Jurkat) as well as primary human CD4 cells. Thus, this CBE approach allowed us to demonstrate that the chromosome 17q21 SNP rs12603332 was functional in regulating levels of ORMDL3, and an ORMDL3 regulated downstream gene ATF6α, in primary human CD4 cells which has not been demonstrated using CBE for any chromosome 17q21 SNP in prior studies. In addition, the CBE approach has potential advantages over using a SNP promotor region transfection approach¹⁶, as experimental approaches using SNP promotor region transfection still allows continued endogenous cell SNP promotor region expression and regulation of ORMDL3. Moreover, levels of the transfected promotor region may not accurately reflect endogenous cellular levels of the promotor region.

Although we were able to demonstrate that single base editing of SNP rs12603332 from the risk allele C to the non-risk allele T regulated levels of ORMDL3 and ATF6α mRNA, a limitation of this study is that we did not demonstrate that this editing regulated levels of ORMDL3 and ATF6α protein. The ability to measure ORMDL3 protein is constrained by the very high level of protein sequence positional identity of ORMDL3 with ORMDL1 (84% identity)²⁶ and of ORMDL3 with ORMDL2 (80% identity)²⁶. We have previously reported that antibodies to ORMDL3 cross react with ORMDL1 and ORMDL2 in western blots even when using commercial antibodies reportedly specific for ORMDL3³. As CD4 T cells express not only ORMDL3, but also ORMDL1 and ORMDL2, the cross reactivity of ORMDL3 antibodies would make it difficult to detect differences in ORMDL3 protein levels induced by SNP editing in CD4 cells also expressing ORMDL1 and ORMDL2. In addition, although our strategy to use EMSA to demonstrate binding of a transcription factor to a DNA sequence is a well recognized method²⁵, an alternative method such as a Chromatin immunoprecipitation (ChIP) assay could have been used. Alternative approaches to strengthen the conclusions of the EMSA could have included use of an expression vector to express E47 (rather than our use of recombinant E47)

In summary, we successfully used CBE plasmid DNA and a novel approach with CBE mRNA to edit a single base on chromosome 17q21 in the SNP rs12603332 from the risk allele C to the non-risk allele T in T cells (human Jurkat cell line and primary human CD4 cells). This allowed us to demonstrate that SNP rs12603332 (located in intron 1 of ORMDL3) is functional in terms of regulating levels of ORMDL3 with significantly higher levels of ORMDL3 mRNA being present with the risk allele rs12603332-C SNP of ORMDL3 compared to the non-risk allele rs12603332-T SNP. The risk allele rs12603332-C SNP of ORMDL3 not only regulated levels of ORMDL3 but also regulated levels of ORMDL3 downstream gene ATF6α to a greater degree compared to the non-risk allele rs12603332-T SNP. As asthma is associated with increased levels of ORMDL3, ORMDL3 risk allele SNPs such as rs12603332-C which increases levels of ORMDL3 and its downstream genes including ATF6α may help to explain this increased risk. In addition, we demonstrated using bioinformatic tools and EMSA that a 6 base pair sequence (CANNTG) where rs12603332-T is located in chromosome 17q21 was a binding site for the transcription factor E47. As E47 is known to inhibit the expression of several genes (i.e. CD25, Cdk6, Rb) our studies suggest that ORMDL3 may be another gene target for inhibition by E47. Overall, this study demonstrates that CBE (in particular CBE mRNA) can be used to characterize which SNPs are functional in regulating genes and their downstream pathways important to the pathogenesis of asthma not only in cell lines but also in primary human CD4 cells.

Abbreviations:

ATF6α

activating transcription factor α

C

cytosine

Cas9

CRISP-associated protein 9

Chr 17q12–21

chromosome 17q12–21 region

ERBB2

erb-b2 receptor tyrosine kinase 2

GRB7

growth factor receptor bound protein 7

gRNA

guide RNA

GSDMA

gasdermin A

GSDMB

gasdermin B

IKZF3

IKAROS family zinc finger 3

LRRC3C

leucine rich repeat containing 3C

ORMDL3

ORM1 like

PGAP3

post-GPI attachment to proteins phospholipase 3

rAPOBEC1

rat apolipoprotein-B-editing enzyme, catalytic polypeptide-1

SNP

single nucleotide polymorphisms

T

thymine

UGI

uracil glycosylase inhibitor

ZPBP2

zona pellucida binding protein 2

Abbreviations:

ATF6α

Activating Transcription Factor alpha

CRISPR

Clustered regularly interspaced short palindromic repeats

Cas9

CRISPR associated protein 9

CBE

Cytosine Base Editor

EMSA

Electrophoretic mobility shift assay

HDR

Homology Directed Repair

IVT

in vitro transcription

GWAS

Genome-wide association study

ORMDL3

ORM1 like 3

SECURE

Selective Curbing of Unwanted RNA Editing

sgRNA

single guide RNA

SNP

Single nucleotide polymorphisms

UGI

Uracil glycosylase inhibitor

UNG

Uracil-DNA glycosylase

USF

Upstream stimulatory factor

EBox

Enhancer box