Tau proteins undergo several post-translational modifications (PTMs) in physiological and disease conditions. In Alzheimer’s disease, O-linked β-d-N-acetylglucosamine (O-GlcNAcylation) modification of serine/threonine (S/T) residues in tau is reduced. In mouse models of tauopathy, O-GlcNAcase inhibitors lead to increased O-GlcNAcylation and decreased filamentous aggregates of tau. However, various non-filamentous tau conformations, linked to toxicity and neurodegeneration in tauopathies, involve processes like oligomerization, misfolding, and greater exposure of the phosphatase-activating domain in the amino-terminus of tau. Additionally, it is becoming clearer that PTMs may differently regulate tau pathobiology in an isoform-dependent manner. Therefore, it is crucial to investigate the effects of O-GlcNAcylation on non-filamentous conformations of both the 4-repeat (4R, e.g. hT40) and 3-repeat (3R, e.g. hT39) tau isoforms. In this study, we assessed how O-GlcNAcylation impacts pathological tau conformations of the longest 4R and 3R tau isoforms (hT40 and hT39, respectively) using recombinant proteins. Mass spectrometry showed that tau is modified with O-GlcNAc at multiple S/T residues, primarily in the proline-rich domain and the C-terminal region. O-GlcNAcylation of hT40 and hT39 does not affect microtubule polymerization but has opposite effects on hT40 (increases) and hT39 (decreases) binding to pre-formed microtubules. Although O-GlcNAcylation interferes with forming filamentous hT40 aggregates, it does not alter the formation of pathological non-filamentous tau conformations. On the other hand, O-GlcNAcylation increases the formation of pathological non-filamentous hT39 conformations. These findings suggest that O-GlcNAcylation differentially modulates microtubule binding and the adoption of pathological tau conformations in the longest 4R and 3R tau isoforms.

Preparation of recombinant unmodified and Glc tau proteins

Recombinant tau proteins were prepared by co-transforming BL21 bacteria (NEB, #C2527H) with two plasmids: a plasmid expressing tau under the T7 promoter as described previously (64) and the pHis-OGT plasmid. The purification procedure for Glc tau proteins was performed using 2 L terrific broth (TB) cultures grown in the presence of ampicillin (50 μg/ml) and kanamycin (25 μg/ml) as selection markers. Moreover, TB was supplemented with GlcNAc (2 mM; Sigma, #U4375) and PUGNAC (10 μM; Sigma, #A7229) to enrich GlcNAc and inhibit the activity of O-GlcNAcase enzyme, respectively. Unmodified tau proteins were grown in the same way with the exception that kanamycin, GlcNAc, and PUGNAC were excluded from TB. Bacterial pellets were lysed using 0.5 M NaCl, 10 mM Tris, and 5 mM Imidazole, pH 8 in the presence of protease inhibitors (as described previously in (64)) and PUGNAC (10 μM) at weight: volume ratio of 1:5. PUGNAC was not included in lysing the bacterial pellets for unmodified tau proteins. The bacterial lysate was subjected to centrifugation at 107,377 RCF for 45 min at 4 °C using a Type 70 Ti rotor (Beckman Coulter, #337922). Supernatant was collected, then the residual bacterial pellet was further lysed in RIPA buffer (10 ml; CST, #9806) supplemented with the same inhibitors by sonicating for 4 times, 30 seconds each. Another centrifugation step was performed to collect the supernatant extracted with the RIPA buffer, followed by pooling the two supernatants (lysis buffer and RIPA buffer) together for further purification. The rest of the purification procedure was performed as described previously (64). Briefly, three stages of fast protein liquid chromatography were performed:  heavy metal affinity chromatography using a 5 ml HiTrap Talon crude column (Cytiva, #28953767); size exclusion chromatography using HiPrep 16/60 Sephacryl S-500 HR (Cytiva, #28935606); anion exchange chromatography using 5 ml HiTrap Q HP (Cytiva, #17115401). The elution fractions containing the highly purified monomeric tau were concentrated to 2-4 mg/ml and supplemented with 1 mM DTT. The final unmodified and Glc tau proteins were aliquoted and frozen at -80 °C. The final concentration of recombinant tau proteins was determined using the SDS-Lowry method as described previously (64).

Recombinant tau protein preparation for tandem mass spectrometry (MS)

Unmodified and Glc tau proteins were digested using a combination of Asp-N (Promega, #V V1621) and rLysC (Promega, #V167A). First, each recombinant tau protein sample (10 µg, n = 1) was subjected to 5 rounds of buffer exchange with 25 mM ammonium bicarbonate (AmBic) pH 8 using 0.5 ml Amicon filter with 3K MWCO (15,000 RCF for 10 minutes; Millipore, #UFC500396). Then, recombinant tau proteins were recovered from the filter by centrifugation at 15,000 RCF for 2 minutes and vacuum dried using Vacufuge. The dried pellets of recombinant tau proteins were reconstituted in 50 ml of digestion buffer (12.5 mM AmBic pH 8 + acetonitrile (ACN) 50%) and incubated at 37 °C for 16-18 hours with Asp-N (150 ng of enzyme). The following day, digested protein samples were subjected to vacuum drying and stored at 4 °C until the second digestion was initiated. Lys-C (500 ng of enzyme) was added and incubated at 37 °C for 16-18 hours. The following day, digested protein samples were subjected to vacuum drying and stored at -20 °C until running on the MS.

Tandem MS of recombinant tau proteins

We utilized an approach like that described by Yang et al. (67). MS analysis was performed twice: initially for method development of recombinant Glc tau and subsequently to validate the final protein preparations used for experiments. The Vanquish Neo nanoHPLC system interfaced to a Thermo Scientific Orbitrap Eclipse MS (Thermo Fisher Scientific) was used for analysis. For each sample, 1 μg was injected and desalted with an Accalaim™ PepMap™ C18 Nano trap column (3 μm, 100 Å, 75 μm × 2 cm) in 100% Buffer A (0.1% formic acid in HPLC water) at 3 μl/min for 5 min. Samples were separated in a linear gradient of 5–35% Buffer B (80% ACN and 0.1% formic acid) over 105 min and washing at 90% Buffer B for 12 min using an Easy Spray PepMap™ RSLC C18 nano column (2 μm, 100 Å, 75 μm × 250 mm). Before each injection, the column was equilibrated at 1.0 % Buffer B for 5 min. Mass spectra were collected using data dependent MS analysis with a duty cycle of 2 sec. To collect precursor masses, orbitrap [resolution (R) of 120,000 at 200 m/z] with internal calibration was used. For precursors carrying charges between 2 and 8 and with intensities over 5 × 10⁴ at R = 30,000, stepped HCD spectra at HCD energies of 15, 25, and 35% were acquired with dynamic exclusion of 15 sec. The fragments are monitored for GlcNAc oxonium ions at m/z of 138.0545, 204.0867, 366.1396, 126.005, 144.0655, 168.0654, 186.076, 274.0921, and 292.1027 Da. If at least one GlcNAc oxonium ion was detected with 15 ppm mass accuracy, the corresponding precursor ion was used to collect an EThcD spectrum in the orbitrap at R of 30,000. For charges of 2 and 3, ETD target was 5.0 × 10⁵; for charges of 4 to 8, ETD target was 2.0 × 10⁵. Supplemental collision energy at 15% was also included. Reaction time of ETD was variable according to the precursor charge state. For a charge of 2, ETD reaction time was 125 msec; for a charge of 3, ETD reaction time was 100 msec; for a charge of 4, ETD reaction time was 75 msec; for charges ≥5, ETD reaction time was 50 msec.

MS data analysis to determine O-GlcNAc modification sites

RAW data files were analyzed with the MetaMorpheus software version 1.0.1 developed by the Smith laboratory (68). For hT40 proteins, the following FASTA files were downloaded from Uniprot (November 2021) and used for analysis: Escherichia coli (strain K12) (UP000000625), Asp-N (Q9R4J4), Lys-C (Q02SZ7), and full-length tau (2N4R isoform, P10636-8). The same FASTA files were used to analyze the hT39 proteins except full-length tau (2N4R isoform, P10636-8) was replaced with 2N3R tau isoform (P10636-5). A mass shift of +203.079 Da (C₈H₁₃NO₅) was used to search for O-GlcNAc modifications (69) on S and T. In addition, the following mass-to-charge-ratios (m/z) corresponding to diagnostic ions (DIs) were investigated: +126.055 Da (C₆H₇NO₂), +138.055 Da (C₇H₇NO₂), +144.066 Da (C₆H₉NO₃), +168.066 Da (C₈H₉NO₃), +186.076 Da (C₈H₁₁NO₄), and +204.087 Da (C₈H₁₃NO₅) (69).

The analysis sequence included mass calibration, global post-translational modification discovery (G-PTM-D) (70), and a classic search. Mass calibration was conducted using the following criteria: protease = Asp-N/Lys-C; maximum missed cleavages = 2; minimum peptide length = 7; maximum peptide length = unspecified; initiator methionine behavior = Variable; variable modifications = Oxidation on M; max mods per peptide = 2; max modification isoforms = 1024; precursor mass tolerance = ±15.0000 PPM; product mass tolerance = ±25.0000 PPM. The criteria utilized for G-PTM-D were protease = Asp-N/Lys-C; maximum missed cleavages = 2; minimum peptide length = 7; maximum peptide length = unspecified; initiator methionine behavior = Variable; max modification isoforms = 1024; variable modifications = Oxidation on M; G-PTM-D modifications count = 3; precursor mass tolerance(s) = ±5.0000 PPM around 0, 203.079372521 Da; product mass tolerance = ±20.0000 PPM. Finally, a classic search was conducted using the following criteria: protease = Asp-N/Lys-C; search for truncated proteins and proteolysis products = False; maximum missed cleavages = 2; minimum peptide length = 7; maximum peptide length = unspecified; initiator methionine behavior = Variable; variable modifications = Oxidation on M; precursor mass tolerance = ±5.0000 PPM; product mass tolerance = ±20.0000 PPM; report PSM ambiguity = True. Peptides were quantified through the FlashLFQ method for label-free quantification bundled into MetaMorpheus (71). At least two peptides were required to identify the protein. Sites of O-GlcNAc modification on tau detected at a false discovery rate (calculated using the target-decoy approach) of 1% are reported (Supplementary Table S1). Supplementary Table S2 demonstrates all quantified tau peptides in unmodified vs GlcNAc-modified tau samples. Supplementary Table S3 shows the quantified peaks of tau with their corresponding peptide masses, theoretical and observed m/z, retention time, and peptide spectral matches (PSMs). MetaDraw version 1.0.5 was utilized to review the PSMs of modified and unmodified tau peptides (samples of these peptides are included in Figures S1 and S2). Processed proteomics data on tau peptides are available in this manuscript (Supplementary Tables 1-3 and Supporting Information).