Catalytic deficiency of O-GlcNAc transferase leads to X-linked intellectual disability

Significance Protein O-GlcNAcylation is a posttranslational modification essential for development. Recently, mutations in the O-GlcNAc transferase (OGT) substrate binding domain have been described that lead to intellectual disability, but the mechanisms underpinning pathogenesis remain to be explored. This work describes the first point mutation in the OGT catalytic domain leading to effects on O-GlcNAcylation and Host cell factor 1 processing in vitro and in a stem cell/Drosophila model, resulting in delayed neuronal differentiation. This establishes a potential link between OGT activity and intellectual disability.


5'
3' CAG repeat (a) Schematic representation of the methodology. Each sample is digested separately with two enzymes: HhaI (3 cut sites) and HpaII (2 cut sites). These sites are expected to be methylated on the inactive X chromosome and unmethylated on the active X chromosome.
Both enzymes are methylation sensitive and will not cut at methylated sites.   Modification highlighted in orange is responsible for the N595K mutation, nucleotides highlighted in green are silent wobble sites. HinfI restriction site is lost in sxc N595K flies.

K P D H M I K P V E V T E S A
Genomic DNA sequence of mouse wild type and N567K OGT gene and translated protein product is shown. Modification highlighted in orange is responsible for the N567K mutation, nucleotides highlighted in green are silent wobble sites. HinfI restriction site is lost in N567K cells.  * -upper case letters denote introduced changes in gene-blocks (GB).

X-inactivation analysis
DNA (200 ng) was digested with methylation-sensitive restriction enzymes followed by PCR for the CAG repeat in exon 1 of the AR gene. Each sample was digested separately with two methylation sensitive restriction enzymes: HhaI (3 cut sites) and HpaII (2 cut sites), and an undigested sample was used as control. PCR primers flanking the entire CAG repeat region including these restriction sites were used to amplify the CAG repeat sequence. PCR amplification occurred only on the inactive X chromosome where the methylation sensitive restriction enzymes were unable to cut the DNA. Intensities of digested and undigested DNA were compared in order to assess the Xinactivation pattern. Ratios and results are reported in Supplementary Fig. S1. equilibrated with base buffer. The peak fractions were concentrated to 10 mg/ml. For crystallisation, truncated OGT was used fresh at 7 mg/ml concentration. For all other purposes, proteins were concentrated to 10 mg/ml, mixed 1:1 with 50% glycerol, snap-frozen and stored at -80˚C until use.

Protein crystallisation
Briefly, experiments were performed at 22 ºC using 24-well hanging drop crystallization plates, by Proteolytic assays was performed using HCF1-rep1 fragment (residues 867-1071) with GST and His tags at the N-terminus and C-terminus, respectively. HCF1-rep1 (2.5 µM) was combined with full length OGTWT or OGTN567K (1 µM) in the presence of 1 mM UDP-GlcNAc. Reaction mixtures were incubated at 37 ˚C for 2-8 h with gentle agitation. Reactions were stopped by addition of LDS loading buffer (4X) (Thermo Fisher Scientific). Proteins were resolved by SDS-PAGE (4-12% acrylamide, Life Technologies) and transferred onto nitrocellulose membranes (GE Healthcare).
After antibody treatment, progress of the reaction was visualised using LI-COR Odyssey scanner and associated quantification software.  Table S3). The size of the PCR product (546 bp vs 450 bp for wild type) was used to screen for successful integration. To further confirm, the PCR product was then digested using PstI or BfmI. Clones negative for the restriction enzyme assay were then sequenced to confirm the presence of the modification. For generation of the 3HA-OGT N567K mESCs, we transfected the previously generated 3HA mESCs with pBABED puro U6 and pX335 (Cas9 D10A) vectors containing the gRNA sequences using the same procedure we previously used for the insertion of the 3HA-tag. Restriction enzymes and genomic DNA sequencing analysis were used for mutation identification. For the restriction fragment length polymorphism assay MouseGlasDiag_F and MouseGlasDiag_R primers were used to amplify by PCR the mutated site and the silent mutation which eliminates a HinfI restriction site (Supplementary Table S4). The size of the bands following digestion (167 bp and 433 bp for wild type versus 600 bp for mutant) was used to screen for successful integration. Clones negative for cutting in the restriction enzyme assay were then sequenced to confirm the presence of the modification. Zeiss 710 confocal microscope and were processed using Image J (NIH) and OMERO.figure. Total neurite length per image was measured using the semi-automatic neuron tracing ImageJ plugin NeuriteTracer (6) and DAPI stained nuclei were counted in ImageJ using automated method.

Immunocytochemistry
Average neurite length was calculated as the ratio between total neurite length and total nuclei number. Three biological replicates were performed and quantified. Student t-test were used to determine the significance of the difference.

Western Blot from mESCs
Cells were washed twice in PBS and harvested in RIPA buffer and the amount of protein was quantified using Pierce™ 660 nm Protein Assay Reagent (Thermo). 15-20 µg of cell lysate was loaded in NuPAGE 3-8% Tris-Acetate gels (Invitrogen) and transferred to Nitrocellulose membranes using wet transfer system. For neural differentiation, 15-20 µg of cell lysate was loaded on NuPAGE 4-12% Bis-Tris gels (Invitrogen) and transferred to PVDF-P membranes using wet transfer system.  Supplementary Table S1.

Cloning of vectors coding for the guide RNA and repair template DNA for Drosophila CRISPR
The sxc N595K Drosophila melanogaster line was generated using CRISPR as previously described (7) with some modifications. Briefly, a guide RNA site was selected with the help of the crispr.mit.edu online tool search and the annealing primer pair (gRNA_sxcT595_fwd and gRNA_sxcT595_rev) with appropriate overhangs for BpiI restriction digestion were cloned into pCFD3-dU63gRNA plasmid (8). A vector coding for repair template DNA of 2160 bp was generated from Drosophila Schneider 2 cell genomic DNA by PCR using GoTaq G2 Polymerase (Promega), T59fix_BAM_fwd and T59fix_NOT_rev primers (Supplementary Table S2). The PCR product was inserted into pGEX6P1 plasmid. The desired mutation, in addition to four silent mutations (Supplementary Figure S3) was introduced by the BA_N595K_GLAS_F primer in a PCR reaction with BA_N595K_GLAShelp_R primer generating a 384 bp product. The silent mutations removed a HinfI restriction site, thus enabling genotyping based on restriction digestion. The PCR product was subsequently utilized for restriction-free cloning (5) using KOD Hot start polymerase (Novagen).
DNA products of cloning and mutagenesis were confirmed by sequencing.

Restriction fragment length polymorphism assay for genotyping sxc N595K Drosophila line
To  Table S2). 5 μl of PCR product was used for restriction fragment length polymorphism assay with HinfI followed by agarose gel electrophoresis of the digested products. Samples that were resistant to HinfI indicated CRISPR/Cas9 gene editing event and were sequenced using T595_DIG_SEQ primer.
Precise incorporation of the repair template into the right position of the genome was confirmed by sequencing a second round of PCR products obtained from potential homozygous CRISPR mutants with mixed T595_DIG and T595_oob primer pairs (Supplementary Table S2).

Generation of the sxc N595K lines, micro-injection and genetics
The sxc N595K Drosophila lines were generated by CRISPR/Cas9 mutagenesis; a mixture of 100 ng/μl guideRNA plasmid with 300 ng/μl repair construct were injected into Vasa::Cas9 embryos in-house  Table S2).

Western blotting from Drosophila samples
To prepare total lysates for Western blotting, embryos were collected on apple juice agar plates at