DNA methylation in Friedreich ataxia silences expression of frataxin isoform E

Epigenetic silencing in Friedreich ataxia (FRDA), induced by an expanded GAA triplet-repeat in intron 1 of the FXN gene, results in deficiency of the mitochondrial protein, frataxin. A lesser known extramitochondrial isoform of frataxin detected in erythrocytes, frataxin-E, is encoded via an alternate transcript (FXN-E) originating in intron 1 that lacks a mitochondrial targeting sequence. We show that FXN-E is deficient in FRDA, including in patient-derived cell lines, iPS-derived proprioceptive neurons, and tissues from a humanized mouse model. In a series of FRDA patients, deficiency of frataxin-E protein correlated with the length of the expanded GAA triplet-repeat, and with repeat-induced DNA hypermethylation that occurs in close proximity to the intronic origin of FXN-E. CRISPR-induced epimodification to mimic DNA hypermethylation seen in FRDA reproduced FXN-E transcriptional deficiency. Deficiency of frataxin E is a consequence of FRDA-specific epigenetic silencing, and therapeutic strategies may need to address this deficiency.

www.nature.com/scientificreports/ CpG island that also encompasses exon 1 and exon 1b (Fig. 1A). Analysis of a 10 × Genomics single-cell ATACseq dataset from a healthy donor showed that the Ensembl-defined "FXN promoter" contains two discrete sites of open chromatin, in cis (Fig. S1), thus further supporting the existence of an independent promoter in intron 1. RNAseq data per EPDnew indicate that FXN_2 is expressed in the central nervous system and a few other cell types, albeit at lower levels and with a less-constrained transcription start site compared with FXN_1 (Fig. S2). RT-PCR analysis using HEK293T cells and a non-FRDA lymphoblastoid cell line showed that while exon 1 splices to exon 2 to form the FXN-M transcript (isoform I; which does not contain any exon 1b sequence), exon 1b independently splices to exon 2 (isoform II) via three alternate splice donor sites, denoted as IIa, IIb and IIc . Whereas isoform IIa uses the canonical "GT" splice donor sequence, isoforms IIb and IIc use the non-canonical "GC" splice donor sequence (Fig. 1C). The predicted translational initiation codon for all variants of isoform II (including IId; Fig. S3) is located within exon 2, and corresponds to the methionine at position 76 in frataxin-M (Fig. 1D). Thus, isoform II is predicted to encode frataxin-E (amino acids 76 to 210), which does not contain most of the mitochondrial targeting sequence found in frataxin-M ( Fig. 1D; henceforth, isoforms IIa, IIb, and IIc are collectively referred to in the singular, as the "FXN-E" transcript). Altogether, these data indicate that frataxin-E is encoded by the FXN-E transcript, which originates in intron 1 of the FXN gene, and is distinct from the FXN-M transcript that codes for frataxin-M.
Frataxin-E protein is deficient in FRDA. Blood samples (n = 5 non-FRDA controls, and n = 5 FRDA patients) were purified by immunoprecipitation (IP) using a commercial anti-human frataxin mAb and further analyzed by western blot. The presence of frataxin-M in the non-FRDA and FRDA blood samples was confirmed using the commercial anti-human frataxin mAb which detected both a His-frataxin-E standard and a His-frataxin-M standard ( Fig. 2A). Endogenous frataxin-E (MW = 14,953 Da) and endogenous frataxin-M (MW = 14,268 Da) isoforms were separated by PAGE using MES with SDS as the running buffer, although they only differ in mass by 685 Da. There was clearly a deficiency of both isoforms in blood samples from FRDA patients compared to non-FRDA controls ( Fig. 2A).
To determine if frataxin-E protein was deficient in FRDA, the IP purified blood samples were analyzed by western blot using a specific anti-human frataxin-E mAb generated in house. The anti-frataxin-E mAb detected a His-frataxin-E standard with < 5% cross-reaction to a His-frataxin-M standard (Fig. 2B) demonstrating specificity. Frataxin-E was hardly detectable in blood samples from FRDA patients and severely deficient compared to blood samples from non-FRDA controls (Fig. 2B).
Frataxin-E deficiency correlates with the length of the expanded GAA triplet-repeat. A clear relationship exists between the canonical frataxin-M protein and the length of the shorter of the two GAA repeat expansions (GAA1) in FRDA. Determining if a similar relationship exists with frataxin-E and GAA1 required a highly quantitative and accurate method for quantifying frataxin-E. Blood samples from a series of 32 FRDA patients (with a representative distribution of GAA repeat lengths 4,33 ; Table S1) were examined for correlation Frataxin-E is deficient in FRDA. Western blot analysis utilizing (A) a commercial anti-frataxin monoclonal antibody and (B) a specific anti-frataxin-E monoclonal antibody. Blood samples from n = 5 non-FRDA controls (C1 to C5; 0.3 mL) and n = 5 FRDA patients (F1 to F5; 0.5 mL) were immunopurified with anti-frataxin antibody and analyzed via western blot. Lanes "E" and "M" show results for histidine-tagged/ purified frataxin-E and frataxin-M protein, respectively. Arrows and labels (right-hand side) show location of proteins of interest. Although there is a small difference in size (Δ 685 Da) between endogenous frataxin-E (MW = 14,953 Da) and frataxin-M (14,268 Da), these proteins were resolved on PAGE using MES and SDS as the running buffer. Frataxin-E and frataxin M together with their His-tagged standards were detected by the anti-frataxin antibody; whereas, only frataxin-E and His-frataxin E were detected by the anti-frataxin-E mAb. Protein plus protein dual color standards were run on each gel and visualized in black and white by the ImageQuant LAS 4000 camera. Western blots of the frataxin proteins were visualized separately by the camera using the SuperSignal West femto luminol enhancement reagent. The two separate images were then combined. www.nature.com/scientificreports/ between GAA repeat length and frataxin-E deficiency using a highly quantitative and previously validated LC-MS assay 30 . Frataxin-E was measured in erythrocytes (where it is known to be expressed, and because they lack mitochondria) and frataxin-M and GAA repeat lengths were measured in PBMCs from the same venous blood sample. Similar to the western blot results, both frataxin-E and frataxin-M were found to be deficient in FRDA patients with the LC-MS assay (n = 32 vs. n = 11 non-FRDA controls; Fig. 3A and Fig. S4A), and the deficiency was significantly correlated with the length of GAA1, which accounted for approximately half the variability ( Fig. 3B and Fig. S4B). Whereas the relationship of frataxin-M deficiency with the expanded GAA repeat is well established, these data also suggest a direct relationship between the expanded GAA repeat and frataxin-E deficiency in FRDA. Consistent with previous observations 20,26,27,34 , the GAA2 allele did not correlate with either frataxin-E or frataxin-M levels (Fig. S5A,B). FXN-E transcript was highly deficient in patient-derived lymphoblastoid cell lines compared with non-FRDA controls (< 5% of non-FRDA; t test p = 0.0001; Fig. 3C), but the levels were unfortunately too low in lymphoblastoid cell lines (and essentially undetectable in FRDA patient PBMCs and whole blood) to permit a similar quantitative analysis for correlation with the expanded GAA repeat. As expected, the FXN-M transcript was also deficient in the same cell lines, but its deficiency was less severe compared with that of FXN-E (Fig. S4C).

Deficiency of frataxin-E correlates with FXN DNA hypermethylation in FRDA. The expanded
GAA triplet-repeat in FRDA induces DNA hypermethylation in intron 1, in a region with 11 contiguous CpG sites, which forms a defined FRDA-specific differentially methylated region (FRDA-DMR) located immediately downstream from the FXN CpG island (Fig. 1A) 22 . Bisulfite deep sequencing (n = 1000 sequence reads per CpG) in PBMCs from our cohort of 32 FRDA patients was used to determine the methylation status at all 39 CpG sites (numbered 57 to 95 per Rodden et al. 22 along the X-axis in Fig. 4A) that span the region of intron 1 from the 3' end of the CpG island (and exon 1b; Fig. 4A) to the expanded GAA triplet-repeat. This showed the typical hypermethylation of the FRDA-DMR, which ranged from 69 to 95% in FRDA patients (n = 32; Fig. 4A,B), compared with 4% in non-FRDA controls (n = 14; p < 0.0001; Fig. 4A,B; Fig. S6). Frataxin-E protein level in FRDA was inversely correlated with methylation in the FRDA-DMR ( Fig. 4C; interestingly, this correlation was stronger than for frataxin-M, Fig. S7). Almost all patients with frataxin-E levels of < 5 ng/ml had > 85% methylation, and the few patients with > 10 ng/ml had < 85% methylation (Fig. 4C). Methylation levels at each of the 39 CpG sites were individually assessed for correlation with frataxin-E levels in FRDA. This unbiased approach revealed that frataxin-E levels correlated specifically with all of the CpG sites within the FRDA-DMR (numbered 72-82; Fig. 4D), and the correlation weakened and eventually disappeared at CpGs away from the FRDA-DMR, thus spatially localizing the correlation within the FRDA-DMR. These data tie frataxin-E deficiency to the downstream epigenetic consequences of the expanded GAA triplet-repeat, and suggest a causal relationship with FRDA-specific DNA hypermethylation.   Table S2). We next analyzed cerebellum, heart and skeletal muscle from a humanized mouse model, wherein the human FXN transgene, containing either 9 (Y47R) or 480 (YG8sR) GAA triplets, rescues the lethality of the FXN-null Gray bars indicate correlations that are not significant (n.s.). The relative locations of the CpG island, CpG island shore, Alu/GAA, and Ex1b are displayed below the graph.  6A). In the Y47R mouse, FXN-E was found to be expressed in the cerebellum and heart but was seen at very low levels in skeletal muscle (Fig. 6B), which contrasted with FXN-M, which was detectable in all three tissues, and expressed at a relatively high level in the heart (Fig. 6C). Both FXN-E and FXN-M transcripts were deficient in all tissues in the YG8sR mouse (GAA-480), although the deficiency of FXN-M transcript was less pronounced in the heart (Fig. 6B,C). The YG8sR mouse clearly showed increased DNA methylation in the FRDA-DMR in all tissues, but it was notable for being rather low (20-30% methylation, with somewhat lower levels in the heart) compared with what is seen in human cell types with expanded GAA repeats (Fig. 6D). There were no obvious differences in the level of FRDA-DMR methylation in male versus female mice, and in young (1 month) versus older (12 month) mice ( Fig. S11A-C). Somatic instability of the expanded GAA repeat was ruled out as an explanation for the variability in levels of transcript or DNA methylation, as the expanded repeat remained mostly unchanged in all three tissues, across different mice ( Fig. S12; note: slight instability was noted in the cerebellum, consistent with previous studies [36][37][38] ).   Fig. 7A,B; the location of the targeting gRNA is depicted by an arrow alongside the X-axis in Fig. 7C). Methylation analysis of all 39 CpG sites from the 3' end of the CpG island to the GAA repeat showed that the normal (non-FRDA) pattern of methylation, i.e., very low in the FRDA-DMR and rising slightly towards the GAA repeat, was seen in untreated and scramble-treated HEK293T cells (Fig. 7C, which was comparable to non-FRDA PBMCs). The FXN targeting gRNA produced a level of methylation in the FRDA-DMR that was approximately two-thirds of the level seen in PBMCs from a heterogeneous group of FRDA patients ( Fig. 7C; FRDA PBMC data are depicted as a composite of the 32 patients in Fig. 3A), and it was mostly located within the FRDA-DMR. Despite this modest level of CRISPR-mediated DNA methylation in the FRDA-DMR, FXN-E transcript was significantly suppressed (Fig. 7D). Given that HEK293T cells do not have the expanded GAA triplet-repeat 22 , this indicates that DNA hypermethylation of the FRDA-DMR is sufficient to induce FXN-E transcriptional deficiency. In contrast, the FXN-M transcript showed a very slight trend towards deficiency (Fig. 7E). These results, i.e., convincing suppression of FXN-E but only a slight trend for FXN-M transcript, were confirmed in another complete experiment, also performed in triplicate (Fig. S13A,B). www.nature.com/scientificreports/

Discussion
The major product of the human FXN gene is the FXN-M transcript (which codes for 210 amino acids) and frataxin-M protein, which consists of residues 81-210 and is localized in the mitochondria 40,41 . The terms extramitochondrial and cytosolic frataxin have been used over the years [42][43][44] to suggest a possible cytoplasmic localization and functional role for the form of frataxin containing residues 81-210. Frataxin-E, on the other hand, is a distinct isoform consisting of residues 76-210 (with an acetylated N-terminus) and is found at relatively high levels in erythrocytes 30 , and is extramitochondrial because it does not possess a mitochondrial targeting sequence. In addition, erythrocytes lose their mitochondria during maturation. In our original study, we showed the presence www.nature.com/scientificreports/ of isoform-E protein in erythrocytes and whole blood samples from healthy subjects using PAGE and western blot analysis with a mouse anti-human frataxin mAb 30 . We have now used a specific mouse anti-frataxin-E mAb to show that isoform E levels are lower in whole blood samples from FRDA patients compared to non-FRDA controls (Fig. 2B). Our specific quantitative stable isotope dilution IP-LC-MS assay showed that frataxin-E levels were threefold higher in erythrocytes compared with frataxin-M in PBMCs/platelets 30 . In erythrocyte samples, only 1.3-2.0% of total frataxin is frataxin-M, with the likely sources being reticulocytes and/or low levels of nonerythrocytic cells. While the function of frataxin-E is not yet well defined, it is expressed at levels that suggest a physiological function in the erythrocytic lineage. Moreover, the transcript that codes for frataxin-E (FXN-E), is detected at relatively high levels in the cerebellum and heart, and early studies suggest that the cytosolic product expressed by this transcript plays a role in modulating mitochondrial function 32,45 . We focused our study on frataxin-E and FXN-E precisely because of its origin in intron 1, i.e., in close proximity to the site of maximal DNA methylation in FRDA. The FRDA-DMR, which shows GAA repeat-dependent hypermethylation, is predictive of age of onset in FRDA 22 , and FXN gene reactivation 46 , and here we show that it is also involved in regulating expression of frataxin-E. A caveat of our study is that in patient-derived blood samples we measured frataxin-E protein in erythrocytes, but DNA methylation (and GAA repeat length) was necessarily assayed in nucleated blood cells (PBMCs). How erythrocytes in FRDA patients end up with deficiency of frataxin-E is unclear. However, a reasonable explanation is that it originated in a nucleated erythrocytic precursor(s) that was susceptible to the epigenetic silencing signals seen in FRDA, thus ultimately resulting in partitioning of deficient quantities of frataxin-E in mature erythrocytes.
While DNA methylation in the FRDA-DMR clearly suppressed the FXN-E transcript, the suppression of FXN-M transcript was not very convincing. The reason for this is unclear, but it may be due to one or more of the following reasons: the relatively modest level of methylation induced by the CRISPR strategy; the relative proximity of the FRDA-DMR to the FXN_2 promoter; and that DNA methylation per se may be insufficient to silence the FXN-M transcript. FXN_2 is a distinct promoter from FXN_1, and the transcripts they produce show tissue-specific differences in expression, so while there may be some co-regulation (for instance, on account of sharing the same CpG island), it would not be particularly surprising that they are differentially regulated by DNA methylation in the FRDA-DMR.
As not much is known about the function of frataxin-E, it is difficult to attribute specific phenotypic aspects of the disease to its deficiency. Extra-mitochondrial frataxin seems to play a role in mitochondrial function 32,[43][44][45] , and in vitro studies of a cytosolic version of frataxin showed that it was capable of interacting with the Fe-S cluster assembly machinery and was protective against oxidative damage of cytosolic aconitase 32 . Carefully designed studies will be needed to determine the function of the endogenously expressed version of this isoform separately from mitochondrial frataxin. The vast majority of patients (~ 95%) are homozygous for the expanded GAA repeat 6 , and the remaining patients have one expanded allele and another pathogenic variant in the other FXN allele [47][48][49][50][51] . Among these compound heterozygous individuals, those with some missense variants and those with truncating variants located upstream of exon 1b (e.g. changes affecting the translation initiation codon of FXN-M) are likely to have at least half the normal level of frataxin-E. Genotype-phenotype correlations in FRDA have so far not considered such differential effects on frataxin-E expression, and we suggest that doing so could be a potentially useful way to delineate the phenotypic contribution of this lesser known product of the FXN gene.
Given that expression is detectable in cell types and tissues that are relevant to FRDA pathophysiology, it is imperative to uncover any potential deleterious effects of frataxin-E deficiency in FRDA. This is critical because gene and protein replacement therapies are actively being developed with the specific goal of replenishing frataxin-M 52-54 , i.e., inadvertently not addressing the deficiency of frataxin-E. It is noteworthy that gene therapy constructs containing varying lengths of intron 1 sequence have been designed 55 , and these could potentially deliver both isoforms of frataxin. Indeed, FXN gene reactivation strategies, also being developed 10,28,56,57 , may have an added advantage of utilizing the endogenous gene regulatory elements that could permit reactivation of both isoforms.

Materials and methods
Materials. Protein G Dynabeads (catalog 10009D), NuPAGE 12% bis-Tris protein gels (catalog NP0341PK2), www.nature.com/scientificreports/ Study participants and samples. Lymphoblastoid cell lines were purchased from Coriell Institute for Medical Research, including those from FRDA individuals (GM16209, GM14518, GM16197, GM16204, GM16207) and non-FRDA controls (GM22647, GM22671). Blood samples were obtained from Friedreich ataxia patients with a confirmed DNA diagnosis (homozygous for the expanded GAA repeat) and from healthy donors enrolled in an ongoing natural history study of FA in accordance with the Declaration of Helsinki and IRB approval from the Children's Hospital of Philadelphia (CHOP IRB 01-002609) and the University of Oklahoma Health Sciences Center (OUHSC IRB 8071). Informed consent was obtained from all subjects and/or legal guardians. Creation of iPS-derived proprioceptive neurons has been previously described 58 . HEK293T cells were purchased from ATCC® (CRL-11268 tm ).
Mouse tissues. YG8sR (480 GAA repeats) and Y47R mice (9 GAA repeats) were bred, euthanized and autopsied at 1 and 12 months of age at Brunel University London (U.K.) under humane conditions in accordance with the U.K. Home Office "Animals (Scientific Procedures) Act 1986" and with approval from the Brunel University London Animals Welfare and Ethical Review Board (Project license number PPL303031). Mice were euthanized for dissection and tissue collection by a Schedule 1 method (cervical dislocation) in accordance with the UK Animals (Scientific Procedures) Act 1986. Frozen tissue samples were transported on dry ice via overnight courier service to the University of Oklahoma Health Sciences Center in Oklahoma City for analysis. All relevant information pertaining to a controlled, non-interventional study is provided in the relevant sections of Results and Methods in accordance with ARRIVE guidelines.
Estimation of GAA triplet-repeat by long-range PCR analysis. GAA repeat lengths were measured using a long-range PCR assay (AccuStart Long Range SuperMix kit, Quantabio) with primers 104F and 629R flanking the GAA repeat in intron 1 7 . Preparation of mouse anti-frataxin-E mAb. The mAb was generated in collaboration with GenScript from a frataxin-E acetylated N-terminal peptide antigen, which contained amino acids 76-85 (acetyl-MNL-RKSGTLGC) linked through a C-terminal cysteine residue to keyhole limpet hemocyanin. Briefly, BALB/c mice were immunized with the frataxin-E peptide antigen, spleens from immunized animals were isolated, placed in a sterile disposable petri dish, and ground with a syringe, DMEM media was added, the B-lymphocytes washed with DMEM, and isolated by centrifugation. SP2/0 myeloma cells were then combined with the B-lymphocytes. The tubes were centrifuged, the cell pellets collected, and suspended in complete fusion media containing hypoxanthine-aminopterin-thymidine (HAT) medium to a density of 2 × 10 5 cells/mL. The cell suspension (100 μL) was added to each well of the 96-well fusion plate coated with feeder cells and the plates incubated at 37 °C with 6% CO 2 in a humidified incubator for 7-days. Hybridoma cell supernatants were screened after 7-days and the plates maintained after the screening. Analyses using His-frataxin-E as the antigen were conducted to identify positive clones. Two of the positive clones were sub-cloned with HT media lacking aminopterin. Plates were incubated at 37 °C with 6% CO 2 in a humidified incubator for 7-days. The supernatants were transferred from the wells (with the cell colony) for screening after 7 days cell growth. The most specific sub-clone that was obtained (19F1-1) was expanded first in a 25 cm 2 flask and then in 75 cm 2  www.nature.com/scientificreports/ ficity of the mAb obtained from expansion of the 19F1-1 clone was confirmed by western blot analysis using His-frataxin-E as the antigen.

Characterization of the FXN-E transcript. RT-PCR products generated with primers spanning FXN
Western blot analysis. The His-frataxin-E standard (2 ng), His-frataxin-M standard (2 ng) and a portion from each whole blood eluate (20 μL) were mixed separately with 5 μL of NuPAGE LDS sample buffer (4X) containing 8% BME. The samples were then heated to 95 °C for 10 min before loading on a 12% NuPAGE Bis-Tris protein gel. NuPAGE MES SDS buffer was used for optimal separation of proteins in the 10-30 kDa range. The gel was run under 150 V for 1.5 h until the blue dye ran to the bottom of the gel. The proteins were transferred to a nitrocellulose membrane using the iBlot 2 gel transfer device and an iBlot 2 transfer stack. The membrane was probed with either an Abcam (Ab113691) anti-human frataxin mouse mAb (diluted 1:1000 with 5% milk in PBS containing 0.1% Tween-20) or an anti-human frataxin isoform E mAb generated by GenScript (diluted 1:1000 with 5% milk in PBS containing 0.1% Tween-20). A goat anti-rabbit HRP IgG (diluted 1:5000) was used as the secondary antibody for chemiluminescence detection. Chemiluminescence was generated using a 1:1 mixture of SuperSignal West femto stable peroxide buffer and luminol enhancer solution. Western blot images were captured on an ImageQuant LAS 4000 (GE Healthcare, Piscataway, NJ).

Quantification of frataxin-M and frataxin-E by LC-MS analysis of frataxin-derived peptides.
The expression of unlabeled and stable isotope labeling by amino acids in cell culture (SILAC)-labeled mature frataxin was performed in Escherichia coli BL21 DE3 as described previously 30,61 . They each had GSG-SLEHHHHHH carboxy-terminal His-tags. All blood samples were thawed at room temperature, and 500 µL of each sample was mixed with 750 µL NP-40 lysis buffer (150 mM NaCl, 50 mM Tris/HCl pH 7.5, 0.5% Triton X-100, 0.5% NP-40, 1 mM DTT, 1 mM EDTA) containing protease inhibitor cocktail. The same amount of SILAC-labeled mature frataxin (20 ng) was spiked in each sample as an internal standard. Samples were lysed and incubated with pre-made DMP-crosslinked anti-frataxin protein G beads for immunopurification as described previously 61 . Samples were analyzed by liquid chromatography-mass spectrometry (LC-MS) analysis as described previously 30  DNA methylation analysis. The DNA methylation assay, analysis, and validation were previously described in detail 22 . Briefly, genomic DNA (0.5 µg) was bisulfite converted and prepared for targeted deep sequencing. Four amplicons were designed to cover all CpG dinucleotides (numbered 57 to 95) between the 3′ end of the CpG island and the Alu element containing the GAA triplet-repeat in intron 1. The amplicons were dual-indexed and pooled to create a library which was sequenced using the Illumina MiniSeq platform. n = 1000 sequence reads were used to calculate the percentage of methylated cytosines at individual CpG dinucleotides (CpGs 57 to 95) and plotted with LOWESS smoothing to generate trendlines. FRDA-DMR methylation values were calculated using n = 1000 sequencing reads of a single amplicon containing CpGs 72 to 82. The methylation panels depicting FRDA-DMR methylation were generated by stacking the sequence reads (n = 300 rows), with columns representing the n = 11 CpG dinucleotides in the FRDA-DMR, and marking each coordinate black if methylated and white if unmethylated (the individual reads [rows] were sorted for high methylation at the bottom). Note that for ease of visualization FRDA-DMR methylation panels were generated with n = 300 reads, however, all data were analyzed at n = 1000 sequence read depth.
CRISPR-mediated DNA methylation of the FRDA-DMR. An all-in-one plasmid system, pdCas9-DNMT3A-PuroR_v2, a gift from Vlatka Zoldoš (Addgene plasmid #74407) expressing a gRNA separately from a dCas9-DNMT3A fusion protein was used to epimodify the FRDA-DMR in HEK293T cells. gRNA target sequences were designed to target FXN intron 1 in the vicinity of FRDA-DMR using the web tool CHOPCHOP 62 . All Multiple gRNAs were synthesized by oligo annealing 63 , and cloned into pdCas9-DNMT3A-PuroR_v2 using Bbs I and tested for their ability to guide dCas9-DNMT3A to the FRDA-DMR. The gRNA overlapping CpGs 66 and 67 (depicted by an arrow alongside the X-axis in Fig. 5C; gRNAF: 5′-CAC CGG GCA CGG GCG AAG GCA GGG C-3′ and gRNAR: 5′-AAA CGC CCT GCC TTC GCC CGT GCC C-3′) resulted in DNA methylation of the FRDA-DMR that was comparable to what is seen in FRDA and was therefore chosen. A "scramble" gRNA, with