Ubiquitin Ser65 phosphorylation affects ubiquitin structure, chain assembly and hydrolysis

The protein kinase PINK1 was recently shown to phosphorylate ubiquitin (Ub) on Ser65, and phosphoUb activates the E3 ligase Parkin allosterically. Here, we show that PINK1 can phosphorylate every Ub in Ub chains. Moreover, Ser65 phosphorylation alters Ub structure, generating two conformations in solution. A crystal structure of the major conformation resembles Ub but has altered surface properties. NMR reveals a second phosphoUb conformation in which β5-strand slippage retracts the C-terminal tail by two residues into the Ub core. We further show that phosphoUb has no effect on E1-mediated E2 charging but can affect discharging of E2 enzymes to form polyUb chains. Notably, UBE2R1- (CDC34), UBE2N/UBE2V1- (UBC13/UEV1A), TRAF6- and HOIP-mediated chain assembly is inhibited by phosphoUb. While Lys63-linked poly-phosphoUb is recognized by the TAB2 NZF Ub binding domain (UBD), 10 out of 12 deubiquitinases (DUBs), including USP8, USP15 and USP30, are impaired in hydrolyzing phosphoUb chains. Hence, Ub phosphorylation has repercussions for ubiquitination and deubiquitination cascades beyond Parkin activation and may provide an independent layer of regulation in the Ub system.

Peak intensities of (auto) signals of the major form (blue) and minor form (lightblue) (left column) are fitted simultaneously to cross peaks as result of major to minor exchange (blue) and minor to major exchange (lightblue) (right column) for residues Ile23, Phe45, Ser57 and Leu73 using the methods described in (Latham et al, 2009). This compensates for the loss of signal intensity due to longitudinal T 1 relaxation (apparent from the decaying auto peaks). The exchange rate was calculated to be 1.76 ± 0.09 s -1 . Backbone chemical shifts (HN, N, CA, CB and HA) were submitted to TALOS+ (Shen et al, 2009) for secondary structure prediction. Confidence in the prediction of α-helix (blue) or β-sheet (red) is given for the major (A) and minor (B) phosphoUb species. See Figure 2E for annotation.          (Dou et al, 2012)). Ser65 is not contacted by the E2 or E3 enzyme. B) Composite model that combines the crystal structure of NEDD4 bound to Ub~UBE2D (pdbid 3jw0, (Kamadurai et al, 2009), UBE2D is omitted for clarity) and the crystal structure of NEDD4L with Ub bound to HECT N-lobe (2xbb, (Maspero et al, 2011)). In either position of Ub, Ser65 is not contacting the E3 ligase. C) Structure of the catalytic core of HOIP, comprising the catalytic IBR (CBR) domain (also known as RING2) and the C-terminal LDD domain (4ljo, (Stieglitz et al, 2013)), shown with acceptor and donor Ub (cyan) bound. Ser65 does not contact the CBR-LDD in either Ub. D) Ubiquitination reaction as in Figure 6A with GST-cIAP1 and UBE2D3. E) Ubiquitination reaction with UBE2D3 and GST-TRAF6 50-285. F) Ubiquitination reaction as in E with or without TRAF6, demonstrating that chain formation is TRAF6-dependent also with phosphoUb.

G) Ubiquitination reaction with NleL and UBE2L3. This combination makes free
Lys6/Lys48-linked polyUb . A) The BEST-TROSY spectrum of wild-type Ub (red) overlaid with the BEST-TROSY spectrum of Ub S65E (blue) and Ub S65D (green). The phosphomimetic Ub mutants shows no signs of the Ub retraCT conformation as observed for phosphoUb. B) A minimal map of the weighted chemical shift perturbations (CSPs) between the peaks of S65E Ub and wild-type Ub. Resonances for Thr9, Glu24 and Ala46 are presumed to be missing due to line broadening.    Reaction progress was monitored by ESI-MS. Reactions were stopped with apyrase, PINK1 was removed using glutathione sepharose 4B resin (GE Healthcare), and the reaction mixture was buffer exchanged to 20 mM Tris, pH 8.7. PhosphoUb was purified by anion exchange using a pH gradient from 20 mM Tris pH 8.7 to 50 mM Tris pH 7.4. Phosphorylated but not unphosphorylated Ub binds to MonoQ anion exchange resin at pH 8.7 and elutes at lower pH. Fractions containing phosphoUb were concentrated using VivaSpin 3.5K concentrators and frozen. For NMR analysis, samples were generated as described below.

Phosphorylation analysis of Ub
Phosphorylation assays were performed by mixing 5 µM GST-PINK1 (species as indicated), 0.2 mg/ml Ub or Ub chains as indicated, 10 mM ATP in reaction buffer (40 mM Tris pH 7.5, 10 mM MgCl 2 , 0.6 mM DTT) and incubated at room temperature for the specified time. The reaction was quenched in 4x LDS sample buffer and samples were applied on a 12% polyacrylamide gel containing 50 µM Phos-tag acrylamide (Wako Chemicals) and 0.77 mM ZnCl 2 . MS analysis was performed as described below.

Intact protein MS analysis
Ub and phosphoUb protein stocks were diluted to 1 µM (50% ACN, 0.1% FA) prior to MS analysis. Samples were directly injected into the Q-Exactive (Thermo Fisher Scientific) mass spectrometer at a flow rate of 10 µl min -1 .

Parkin activity assays
For Parkin activation assays a mixture of 0.1 µM E1, 9 µM Parkin, 10 mM Proteins were separated on NuPAGE 4-12% gradient Bis-Tris gels and Western blotting was performed by transfer on a nitrocellulose membrane and detection using a monoclonal anti-Ub FK2 antibody (Millipore).

PhosphoUb preparation for NMR analysis
Isotope labeled Ub was expressed and purified as described previously .
NMR acquisition was carried out at 298 K on Bruker Avance III 600 MHz and Avance2+ 700 MHz spectrometers equipped with cryogenic triple resonance TCI probes. Topspin (Bruker) and Sparky (Goddard & Kneller, UCSF; http://www.cgl.ucsf.edu/ home/sparky/) software packages were used for data processing and analysis, respectively. 1 H, 15 N 2D BEST-TROSY experiments (Favier & Brutscher, 2011) were acquired with in-house optimized Bruker pulse sequences incorporating a recycling delay of 400ms and 512*64 complex points in the 1 H, 15 N dimension, respectively. High quality 2D data sets were acquired in ~8 min.
Backbone chemical shift assignments were completed using Bruker triple resonance pulse sequences. CBCACONH and HNCACB spectra were collected with 1024*32*55 complex points in the 1 H, 15 N and 13 C dimensions.
HNCO and HNCACO experiments were collected with Non Uniform Sampling (NUS) at a rate of 25% of 1024*50*47 complex points in the 1 H, 15 N and 13 C dimensions, respectively. HA and HB proton shifts were obtained from an HBHACONH spectrum collected with 50% NUS and 512*40*80 in the 1 H, 15 N and indirect 1 H dimensions, respectively. These data sets were processed with Multi-Dimensional Decomposition or Compressed Sensing using the MddNMR software package (Orekhov & Jaravine, 2011;Kazimierczuk & Orekhov, 2011).
Secondary structure calculations were completed using TALOS+ (Shen et al, 2009) incorporating HN, N, CA, CB and HA shifts.
15 N{ 1 H}-heteronuclear NOE (hetNOE) measurements were carried out using a Bruker pseudo 3D pulse program, applying a 120º 1 H pulse train with a 5 ms interpulse delay for a total of 5 s interleaved on-or off-resonance saturation.
The hetNOE values were calculated from peak intensities according the equation I on /I off .
The rate of exchange between the major and minor forms of phosphoUb was established using ZZ exchange spectroscopy. Mixing times of 6,18,30,48,66,96,132,192,258,324,372,426 and 492 ms were used in the pseudo 3D data set. Peak intensities of the major and minor forms (auto) and their exchange peaks (cross) of Ile23, Phe45, Ser57 and Leu73 were fitted in Mathematica 9 (Wolfram) using the methods described in (Latham et al, 2009).
Differences in the hydrogen bonding network were established using the long range TROSY-based HNCO (trHNCO) experiment described by (Cordier et al, 2008

Ubiquitin chain composition mass spectrometry analysis
Chain assembly reactions were resolved on NuPAGE 4-12% gradient Bis-Tris gel prior to in-gel digestion and the addition of 400 fmoles AQUA peptide standards according to (Kirkpatrick et al, 2006) and (Ordureau et al, 2014).  (Kirkpatrick et al, 2006).
For ligase reactions, 5-10 µM of respective E3 ligases were added to the E2 mixture. Western blotting was performed using rabbit polyclonal anti-Ub antibody (Millipore).

UBD pull-down assay
Pull-down assays were essentially performed as previously described (Kulathu et al, 2009). 30 µg of GST-tagged TAB2 NZF was immobilized on 25 µl of Glutathione Sepharose 4B (GE Life Sciences) and washed three times with pull-down assay buffer (PDAB; 50 mM Tris pH 7.4, 150 mM NaCl, 2 mM βmercaptoethanol, 0.1 % NP-40). Then, 1.5 µg of the indicated tetraUb species (see section above; Generation of phosphoUb) was incubated with the immobilized TAB2 NZF overnight at 4 ºC in a total volume of 450 µl pull-down assay buffer containing 0.2 mg/ml BSA. The beads were then washed five times with PDAB prior to separation by SDS-PAGE. Proteins were transferred to PVDF and blotted using a polyclonal rabbit anti-Ub antibody (Millipore).

Disassembly of phosphorylated polyubiquitin
DUBs were either kind gifts from Marc Pittmann, purchased, or purified according to published procedures (Mevissen et al, 2013). Polyubiquitinated cIAP substrate was generated from a ligase reaction with GST-tagged cIAP1 and UBE2D1, which was stopped with 0.1 U apyrase. 10 µl of this reaction were used in a 30 µl DUB reaction, that contained 3 µl 10 x DUB buffer (500 mM sodium chloride, 500 mM Tris pH 7.5, 50 mM dithiothreitol) and DUBs at indicated concentrations. During incubation at 37 ºC, aliquots of 6 µl of the reaction were taken at the time points indicated and mixed with 6 µl 4 x LDS loading buffer (Invitrogen) to stop the reaction. Samples (10 µl) were resolved by SDS-PAGE as above and silver stained using the Bio-Rad Silver Stain Plus kit according to the manufacturer's protocol.