The covalent attachment of ubiquitin (Ub) to target proteins recruits the cascade reactions of the E1 activating enzymes, E2 conjugation enzymes, and E3 ligases. Ub can also be post-translationally tagged by Ub itself or Ub-like proteins, phosphorylation, acetylation, ribosylation, and deamination (Mattiroli and Penengo, 2021; Swatek and Komander, 2016) (for details, see the references therein) (Fig. 1A). These post-translational modifications induce distinct topologies of the Ub molecule, leading to a vast range of cellular outcomes (cell division, death, development, etc.) via the creation of distinct words for the Ub code (Dittmar and Selbach, 2017; Komander and Rape, 2012; Kwon and Ciechanover, 2017). The identification of new types of Ub modification can reveal unrecognized Ub codes, which deepens our understanding of a vast range of Ub-associated biological processes. However, the post-translational modifications of Ub per se and their regulatory mechanisms remain incompletely understood.

Virtually all newly made polypeptides from ribosomes contain methionine (Met) due to the design of the AUG initiation codon. The initiator Met is frequently excised by Met-aminopeptidases (MetAPs), in cases where the residue at position 2 is Gly, Ala, Ser, Thr, Pro, or Val. Intriguingly, Nt-Met excision (NME) is an evolutionarily conserved and essential process that is applied to two-thirds of proteins (Giglione et al., 2004). NME also operates even in a set of proteins with a residue larger than Val at position 2, although the mechanism underlying this is still unclear (Chen and Kashina, 2021; Redman and Rubenstein, 1981; Sadis et al., 1995). Specifically in the yeast Saccharomyces cerevisiae, the Nt-Asn and Nt-Gln exposed after NME are deaminated to Nt-Asp and Nt-Glu, respectively, by Nta1 Nt-amidases and subsequently arginylated by Ate1 arginyltransferase (Nguyen et al., 2019). The resulting Nt-arginylated proteins are targeted for degradation by the Arg/N-degron pathway (previously known as the N-end rule pathway), the prototype of N-degron pathways (Fig. 1B) (Bachmair et al., 1986; Tasaki et al., 2012; Varshavsky, 2019). The hitherto established N-degron pathways of S. cerevisiae encompass the Arg/N-degron pathway (that targets unmodified Nt-Arg, His, Lys, Leu, Ile, Trp, Phe, Tyr, and Met-_Φ [Met followed by a hydrophobic residue]) (Kim and Hwang, 2014; Kim et al., 2014), the Ac/N-degron pathway (that targets acetylated Nt-residues) (Hwang et al., 2010; Lee et al., 2016; Nguyen et al., 2018; Shemorry et al., 2013), the Pro/N-degron pathway (that targets Nt-Pro) (Chen et al., 2017; 2021), and the fMet/N-degron pathway (that targets Nt-formylated Met) (Kim et al., 2018) (Figs. 1B-1E).

The 76-residue Ub starts with the Met-Gln (MQ) N-terminus. The MQ-starting Ub (MQ-Ub^3-76) is hereafter denoted simply as ^MQUb. Thus, we hypothesized that ^MQUb might be processed by these NME-provoked cascade reactions of Nt-deamination and Nt-arginylation to yield Nt-arginylated RE-Ub^3-76 (denoted here as ^REUb) in yeast and mammalian cells (Figs. 1A and 2A). Indeed, we here identify the previously unknown ^REUb in yeast and consider the ramifications of this discovery, which greatly expands the known complexity and biological significance of the Ub code.

Yeast strains, plasmids, and culture media

The yeast strains, plasmids, and oligomers used in this study are listed in Table 1. To construct pCH6913, pCH6914, pCH6915, or pCH6916 expressing ^MQUb_AA-GST, ^QUb_AA-GST, ^EUb_AA-GST, or ^REUb_AA-GST in the pHUE vector (Catanzariti et al., 2004), respectively, Ub variant-encoding open reading frames (ORFs) were polymerase chain reaction (PCR)-amplified from pCH5097 as a template using a set of primer pairs: OCH7003/OCH8110, OCH8089/OCH8110, OCH8088/OCH8110, or OCH6968/OCH8110. The resulting PCR products were digested with SacII/EcoRI and triply ligated into SacII/HindIII-cut pHUE with EcoRI/HindIII-cut GST-encoding DNA (that was PCR-amplified from pCH33 using OCH6991/OCH6512 primers). To construct pCH7047 expressing MQIFVKTLTGK-GST, the PCR-amplified DNA fragment using OCH8085/OCH6512 primers was digested with SacII/HindIII and ligated into SacII/HindIII-cut pHUE. To construct pCH7048, pCH7049, and pCH7050, the DNA fragments encoding QIFVKTLTGK-GST, EIFVKTLTGK-GST, or REIFVKTLTGK-GST were PCR-amplified from pCH7047 using the set of primers OCH8089/OCH6512, OCH8088/OCH6512 or OCH6968/OCH6512 and pCH7047, respectively. The resulting PCR products were digested with SacII/HindIII and ligated into SacII/HindIII-cut pHUE, yielding pCH7048, pCH7049, and pCH7050. Further details of the cloning procedures for the plasmids are available upon request. All final plasmids were verified by DNA sequencing.

Yeast cells were cultured in YPD (1% yeast extract, 2% peptone, 2% glucose) and synthetic complete (0.67% yeast nitrogen base with ammonium sulfate, 2% glucose, specific compounds essential for the growth of an auxotrophic strain) media. The construction of yeast strains and DNA transformation were performed following standard techniques (Gietz and Schiestl, 2007; Janke et al., 2004; Sherman, 2002).

Purification of Ub-GST fusions and their derivatives

pCH6913 (expressing ^6HisUb-^MQUb_AA-GST), pCH6914 (expressing ^6HisUb-^QUb_AA-GST), pCH6915 (expressing ^6HisUb-^EUb_AA-GST), pCH6916 (expressing ^6HisUb-^REUb_AA-GST), pCH7047 (expressing ^6HisUb-MQIFVKTLTGK-GST), pCH7048 (expressing ^6HisUb-QIFVKTLTGK-GST), pCH7049 (expressing ^6HisUb-EIFVKTLTGK-GST), and pCH7050 (expressing ^6HisUb-REIFVKTLTGK-GST) were transformed into BL21 (DE3) Escherichia coli. Five milliliters of overnight culture of these transformants was inoculated into 500 ml of Lysogeny Broth (LB) medium with ampicillin (final concentration of 100 _µg/ml), followed by growth at 37°C to A600 of ~0.6. Expression of the Ub fusions was induced by incubating E. coli cells with 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) (Cat. No. I2481C; GoldBio, USA) at 18°C overnight.

Cell pellets were thawed and resuspended in lysis buffer (50 mM HEPES, 150 mM NaCl, 10% glycerol, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride [PMSF, Cat. No. 10837091001; Sigma-Aldrich, USA], 1 mM DTT, 0.1% Triton X-100) with lysozyme (final concentration of 1 mg/ml). Resuspended cells were incubated on ice for 20 min and disrupted by sonication (VCX-130; SONICS, USA), five times for 30 s each, at 1 min intervals with 40% amplitude. Cell lysates were centrifuged at 25,000 × g for 20 min at 4°C and the supernatants were incubated with 0.5 ml of pre-washed Glutathione Sepharose 4B (Cat. No. GE17075605; GE Healthcare, USA) at 4°C for 4 h. After washing the beads three times in 10 ml of washing buffer (50 mM HEPES, 150 mM NaCl, 10% glycerol, 0.1% Triton X-100, 1 mM DTT), the bound proteins were eluted with 6 ml of elution buffer (50 mM HEPES, 150 mM NaCl, 10% glycerol, 0.1% Triton X-100, 1 mM DTT, 30 mM reduced glutathione, pH 8.0), followed by 4 h of dialysis against storage buffer (25 mM Tris-HCl pH 7.4, 150 mM NaCl, 10% glycerol, 1 mM DTT) and overnight incubation with the catalytic core Usp2 (Usp2-cc) at 4°C. The Usp2-cc treatment yielded free ^6HisUb and MQIFVKTLTGK-GST, QIFVKTLTGK-GST, EIFVKTLTGK-GST, REIFVKTLTGK-GST, ^MQUb_AA-GST, ^QUb_AA-GST, ^EUb_AA-GST, or ^REUb_AA-GST. To remove the free ^6HisUb molecules, the Usp2-cc-treated samples were incubated with Ni-NTA resins. Subsequently, the Ni-NTA bead-unbound proteins were re-dialyzed against the above storage buffer and further purified by gel filtration chromatography. The active fractions were collected, concentrated, and stored at −80°C.

Production of anti-^REUb

The synthetic peptide REIFVKTLTGKC was used for the production of anti-^REUb as the antigen. Rabbit polyclonal antisera to the peptide were raised by AbClon (Korea). Antibodies specific for REIFVKTLTGKC were “negatively” selected from antiserum by sequentially incubating 5 ml of clarified antiserum with 1 ml of Glutathione Sepharose beads that had been pre-incubated individually with 2 mg of the purified MQIFVKTLTGK-GST, QIFVKTLTGK-GST, and EIFVKTLTGK-GST. The flow-through of antisera was further “positively” selected by overnight incubation with REIFVKTLTGK-GST-conjugated Sulfo-Link Coupling Resin (Cat. No. 20401; Thermo Fisher Scientific, USA) at 4°C. Bound antibodies were eluted by adding 9 ml of 0.1 M glycine HCl (pH 3.5). The eluted antibodies were further incubated with Affi-Gel 10/15 (Cat. No. 1536098; Bio-Rad, USA) that were pre-conjugated with 30 mg of proteins in the extracts of CHY2014 (ate1∆) cells. The unbound anti-^REUb was concentrated by Protein A-Sepharose affinity chromatography (Amicogen, Korea).

Immunoprecipitation of ^REUb and its conjugated proteins

The indicated yeast cells were cultured in 100 ml of YPD medium to A₆₀₀ of ~1. Cells were harvested and washed twice with ice-cold phosphate-buffered saline (PBS). Cell pellets were resuspended in 2 ml of lysis buffer (25 mM Tris-HCl pH 7.4, 150 mM NaCl, 10 mM N-ethylmaleimide [NEM, Cat. No. E3876; Sigma-Aldrich], 10% glycerol, 1% nonidet P40 substitute [NP-40, Cat. No. 11754599001; Sigma-Aldrich], and 1× protease inhibitor cocktail [Cat. No. P8215, Sigma-Aldrich]) and disrupted by bead beating using Mini Beadbeater-24 (BioSpec Products, USA), 10 times for 10 s each at 1 min intervals on ice. Cell extracts were precipitated by centrifugation at 13,000 × g for 15 min at 4°C. The supernatants were incubated with 1 µg of anti-^REUb for 6 h at 4°C and then with 5 µl of Dynabeads Protein A (Cat. No. 10001D; Thermo Fisher Scientific) for 1 h. After washing the beads three times in the lysis buffer, the bound proteins were eluted by boiling in 1× SDS sample buffer and fractionated using 4%-20% Mini-PROTEAN TGX precast protein gels (Cat. No. 4561094; Bio-Rad). Fractionated proteins were transferred into PVDF membranes (polyvinylidene difluoride, Cat. No. 88518; Thermo Fisher Scientific) and ^REUb-conjugated proteins were identified by anti-Ub (P4D1) (Cat. No. Sc-8017; Santa Cruz Biotechnology, USA).

Preparation of SILAC (stable isotope labeling by amino acids in cell culture) samples

S. cerevisiae CHY3188 (arg4∆ lys2Δ) and CHY6056 (ate1Δ arg4Δ lys2Δ) were grown at 30°C to A₆₀₀ of ~1 in 300 ml of SC medium with 20 _µg/ml of either light isotope-labeled ¹²C₆¹⁴N₄-L-arginine (Cat. No. ULM-8347; Cambridge Isotope Laboratories, USA) and ¹²C₆¹⁴N₂-L-lysine (Cat. No. ULM-8766; Cambridge Isotope Laboratories) or 20 _µg/ml heavy isotope-labeled ¹³C₆¹⁵N₄-L-arginine (Cat. No. CNLM-539-H; Cambridge Isotope Laboratories) and ¹³C₆¹⁵N₂-L-lysine (Cat. No. CNLM-291-H; Cambridge Isotope Laboratories). The yeast strains were treated with 2 mM H₂O₂ for 1 h, followed by mixing an equal amount of each culture and subsequent centrifugation at 2,000 × g and 4°C for 10 min. Cell pellets were washed twice with ice-cold PBS, resuspended in 5 ml of lysis buffer (25 mM Tris-HCl pH 7.4, 150 mM NaCl, 10 mM NEM, 10% glycerol, 1% NP-40, 1× protease inhibitor cocktail), and disrupted by bead beating using Mini Beadbeadbeater-24, 10 times for 10 s each at 1 min intervals on ice. Cell lysates were centrifuged at 13,000 × g and 4°C for 15 min. The resulting supernatants were incubated with 2 _µg of anti-^REUb at 4°C for 6 h before adding 10 µl of Dynabeads-Protein A for an additional 1 h of incubation. After washing three times with 1 ml of the yeast lysis buffer, the binding proteins were eluted by boiling in 30 µl of 1× SDS-polyacrylamide gel electrophoresis (PAGE)-sample buffer. The eluted proteins were fractionated by 4%-20% Mini-PROTEAN TGX precast protein gels (Cat. No. 4561096; Bio-Rad) and stained with Coomassie Brilliant Blue (CBB).

Liquid chromatography and tandem mass spectrometry (LC-MS/MS)

The CBB-stained protein bands of interest were excised, destained in 50 mM ammonium bicarbonate with 50% acetonitrile (ACN), and dehydrated with 100% ACN, followed by drying in a vacuum evaporator. The destained proteins were proteolyzed with trypsin by the in-gel digestion method (Kim et al., 2013). The tryptic peptides were reconstituted in 7 µl of 0.1% formic acid, and a 5 µl aliquot was injected into a reverse-phase EASY-Spray PepMap RSLC C18 LC column (0.075 mm inner diameter × 500 mm length) on the Eksigent NanoLC Ultra system with an integrated column heater at 40°C; the column was pre-equilibrated with 96% buffer A (0.1% formic acid in water) and 4% buffer B (0.1% formic acid in ACN). The peptides were eluted with a 4%-35% gradient of buffer B over 150 min and a 32%-80% gradient of buffer B over 40 min. The total ion acquisition run time was set to 190 min at a flow rate of 250 nl/min. A Q-Exactive mass spectrometer (Thermo Fisher Scientific) was operated in data-dependent acquisition mode for the entire analysis. Full scans (m/z 300-1600) were acquired at a resolution of 70,000 using an automatic gain control (AGC) target value of 1e6 and a maximum ion injection time of 30 ms. Tandem mass spectra were generated for up to 12 precursors by high-energy collision dissociation using a normalized collision energy of 35%. The dynamic exclusion was set to 60 s. Fragment ions were detected in normal scan mode using an AGC target value of 5e4 and a maximum ion injection time of 120 ms. Source ionization parameters were as follows: spray voltage, 1.9 kV; capillary temperature, 275°C; and S-Lens RF Level, 50.

Proteomic analysis of SILAC-based LC-MS/MS data

The RAW files from Q-Exactive were directly conveyed to the proteomic data analysis program suite, Proteome Discoverer v2.2 (Thermo Fisher Scientific). We employed the adjusted exemplar workflow for the SILAC experiment and the basic Sequest HT search module. Parameters for the Sequest search module were as follows: enzyme, specifically trypsin; protein database, UniProt reference proteome database of yeast (UP000002311, released in 04/2019) plus modified cRAP contaminant database with accessions in the dust/contact category (https://www.thegpm.org/crap/); fragment mass tolerance, 0.05 Da; precursor mass tolerance, 20 ppm; dynamic modifications, acetylation of protein-N-term (+42.010565 Da), oxidation of methionine (+15.9949 Da), label: ¹³C₆¹⁵N₂ at lysine (+8.014199 Da), label: ¹³C₆¹⁵N₄ at arginine (+10.008269 Da), Gly–Gly at Lys (Kε-GG) or protein-N-term (+114.042927 Da); and static modification, carbamidomethylation of cysteine (+57.021464 Da). Next, the identified spectra were validated using Percolator module, which was set to use a concatenated target/decoy strategy with a target cut-off q-value of 0.01 at PSM (peptide spectrum match) level. To integrate chromatographic features of identified SILAC pairs, we used Minora Feature Detector module with maximal trace retention time window of 5 min. Peptide abundance from the feature detector module was normalized with total sum of the abundance values over all identified peptides. Each result of the replicates was treated separately.

Identification of mass spectrum matched to the Nt-peptide of ^REUb

The generated mass spectrum files were directly matched to the peptide sequence in the customized ^REUb protein fasta file using the MS-GF+ search algorithm (v20190418) (Kim and Pevzner, 2014). The ^REUb sequence is a virtual one that replaces the two N-terminal residues of Ub MQ with RE. The search engine settings were as follows: trypsin enzyme specificity with 1 as the number of tolerable termini; 20 ppm for MS1 tolerance; variable modifications: oxidation of methionine (+15.9949 Da), label: ¹³C₆¹⁵N₂ at lysine (+8.014199 Da), label: ¹³C₆¹⁵N₄ at arginine (+10.008269 Da); and fixed modification: carbamidomethylation of cysteine (+57.021464 Da). All spectra matched to the Nt-peptide of the ^REUb were manually inspected. For annotating peptide spectra, we used a freely accessible version of the Interactive Peptide Spectrum Annotator (IPSA) (Brademan et al., 2019).

In silico structure modeling of ^MQUb, ^QUb, ^EUb, and ^REUb

The amino acid sequence of S. cerevisiae Ub (UniProt accession: P0CG63) was seeded as a template for the structure modeling of ^MQUb, ^QUb, and ^EUb. In contrast, the ^REUb template was deduced from ^MQUb by modifying its Nt-MQ into Nt-RE. The three-dimensional structure models of these Nt-varied Ubs were predicted by ColabFold, which combines a protein homolog search MMseqs2 (Steinegger and Soding, 2017) with AlphaFold2 (Jumper et al., 2021). In this setting, the model of ^MQUb was used as a template for accurate comparison with those of ^REUb, ^QUb, and ^EUb. The predicted models with the highest pLDDT (predicted Local Distance Difference Test) value were selected for the structure comparison. In the case of ^REUb, the second-ranked model was chosen because the top-ranked one exhibited a covalent bond of Nt-Arg and Ser19, which is most unlikely to form in vivo. Surface structure modeling of the Ub variants was performed by Chimera (Pettersen et al., 2004) and the hydrogen bonds (H bonds) were displayed by PyMOL (http://www.pymol.org).

Antibody to recognize Nt-arginylated ^REUb

To detect the hypothetical ^REUb (Fig. 2A), we produced rabbit anti-^REUb polyclonal antibody using a synthetic peptide, REIFVKTLTGKC, the 11-residue Nt-sequences of ^REUb and Cys for the KLH (keyhole limpet hemocyanin) conjugation. Dot immunoblotting analysis with the affinity-purified anti-^REUb revealed the specific binding of the antibody to the REIFVKTLTGK-GST (glutathione S-transferase) fusion, but not to MQIFVKTLTGK-GST (with the 11-residue Nt-sequence of ^MQUb), QIFVKTLTGK-GST (bearing the 10-residue Nt-sequence of ^QUb), or EIFVKTLTGK-GST (bearing the 10-residue Nt-sequence of ^EUb) (Fig. 2B). The following single-letter abbreviations denote amino acid residues: M, Met; A, Ala; C, Cys; G, Gly; S, Ser; T, Thr; V, Val; N, Asn; Q, Gln; D, Asp; E, Glu; L, Leu; I, Ile; F, Phe; Y, Tyr; W, Trp; H, His; K, Lys; R, Arg; and P, Pro.

Moreover, the anti-^REUb antibody specifically recognized ^REUb_AA-GST (UbAA, a deubiquitylation-protective Ub variant with the mutation of C-terminal GG to AA), but not ^MQUb_AA-GST, ^QUb_AA-GST, or ^EUb_AA-GST (Fig. 2C). In agreement with these dot blotting results, immunoblotting with anti-^REUb further confirmed the specific recognition of purified ^REUb but not of ^MQUb; the ^REUb and ^MQUb were produced by the in vitro deubiquitylation of purified N-terminally 6 His-tagged ^His6Ub-^MQUb and ^His6Ub-^REUb, respectively, using the Ub fusion technique (Catanzariti et al., 2004) (Figs. 2D and 2E).

^MQUb undergoes NME, Nt-deamination, and Nt-arginylation to produce ^REUb

We next attempted to determine the levels of endogenous ^REUb and ^REUb-linked protein species by immunoblotting with the purified anti-^REUb. Notably, however, ^REUb and ^REUb–protein conjugates in wild-type S. cerevisiae were barely detectable under normal growth conditions (data not shown), suggesting that their endogenous levels are most likely near or below the detection limit of the antibody.

Since ^MQUb is a putative substrate of NatB Nt-acetylase that targets MQ, MN, ME, and MD at N-termini (Nguyen et al., 2018; Ree et al., 2018), we conjectured that Nt-acetylation might inhibit the NME-triggered Nt-modifications of ^MQUb. To rule out this possibility, we performed anti-^REUb-based immunoblotting in the extracts from naa20Δ cells that lacked a catalytic subunit of the NatB Nt-acetylase (Ree et al., 2018). Indeed, the high-molecular-weight ^REUb–protein adducts were detected in overnight-grown (stationary-phase) naa20Δ yeast cells (Fig. 2F, lane 4). As expected, the ^REUb and ^REUb–protein adducts were abolished in nta1Δ naa20Δ cells (lacking Nta1 Nt-amidase) and ate1Δ naa20Δ cells (lacking Ate1 arginyltransferase) (Nguyen et al., 2018), whereas total Ub and Ub-conjugated proteins were still detected (Fig. 2F). In sum, we conclude that, in vivo, ^MQUb can be converted into ^REUb through the consecutive reactions of NME, Nt-deamination, and subsequent Nt-arginylation (Fig. 2A).

Upregulation of ^REUb–protein adducts in the stationary-growth phase or by oxidative stress

To detect the ^REUb–protein adducts more effectively and sensitively, we performed anti-^REUb-based immunoprecipitation, followed by immunoblotting with a highly Ub-specific monoclonal anti-Ub (P4D1) antibody, in extracts from naa20Δ cells. The hypersensitive coimmunoprecipitation–immunoblotting assays revealed that those ^REUb–protein adducts were markedly induced in wild-type cells in the stationary-growth phase, but hardly at all in the exponential-growth phase (Fig. 2G; cf, lane 6 vs lane 5). In contrast, the ^REUb–protein adducts were not detected in ate1Δ cells regardless of their growth phase (Fig. 2G; lanes 7 and 8).

Ubr1 is as a key E3 Ub ligase of the Arg/N-degron pathway and directly detects the Nt-Arg of Nt-arginylated proteins via its UBR box (Choi et al., 2010). Interestingly, the levels of ^REUb and ^REUb-protein adducts were discernably decreased in ubr1Δ naa20Δ cells, compared with those in naa20Δ cells, indicating the involvement of Ubr1 in processing ^REUb (Fig. 2H; cf, lane 5 vs lane 6).

We also observed that ^REUb–protein adducts were strongly increased in wild-type cells in the presence of 2 mM H₂O₂, an oxidative stressor (Fig. 2I; lane 7). In agreement with results with the stationary-phase yeast, however, the ^REUb-conjugated proteins were hardly detected in ate1Δ cells, regardless of H₂O₂ treatment (Fig. 2I; lanes 6 and 8). Taken together, these results suggest that the ^REUb and ^REUb–protein adducts are upregulated in yeast cells in the stationary-growth phase or by oxidative stress in an Ate1-dependent manner.

Identification of ^REUb and ^REUb-interacting proteins

Given that oxidative stress increases the levels of ^REUb–protein conjugates, we next sought to identify the ^REUb-attached proteins by employing SILAC in the H₂O₂-treated wild-type cells and ate1Δ cells, followed by anti-^REUb-based coimmunoprecipitation and LC-MS/MS (Fig. 3A).

Out of 541 distinct proteins detected by LC-MS/MS, 83 proteins were identified as direct or indirect ^REUb-binding candidates in the H₂O₂-treated wild-type cells (Fig. 3B). STRING enrichment analysis (for protein–protein interaction networks) indicated the functional associations of these putative ^REUb-interacting proteins with a set of cellular processes including gene transcription, vesicle trafficking, DNA replication, RNA processing, and mitochondrial regulation (Fig. 3C).

The tryptic digestion of Ub-conjugated proteins produces a K-ε-GG (K_GG) Ub isopeptide, which allows mapping of the ubiquitylation sites of substrates by LC-MS/MS (Kirkpatrick et al., 2005). Using this Ub-remnant K_GG signature-based approach, in the H₂O₂-treated yeast cells we identified the (putative) ^REUb-conjugated proteins and determined their ubiquitylation sites: Ub (⁴³LIFAGK_GGQLEDGR⁵⁴); Fdc1 (a ferulic acid decarboxylase, ¹²⁶THILSEEK¹³³_GG); Ino80 (a chromatin remodeling ATPase, ⁹³⁹NVQSELGDK_GGIEIDVLCDLTQR⁹⁵⁹); Rpb3 (an RNA polymerase subunit, ¹⁵⁵LTCVAK_GGK¹⁶¹); Uso1 (a vesicle transport protein, ⁹¹²ITEIK_GGAINENLEEMK⁹²⁶); Sla1 (a cytoskeleton-binding protein, ³⁵⁸GIVQYDFMAESQDELTIK³⁷⁵_GG); Ufd4 (a Ub-fusion decay E3 ligase, ⁵⁴⁹AINDQLIK⁵⁵⁶_GG); Rad57 (a DNA repair protein, ⁴⁰⁵SFK_GGASTIIQR⁴¹⁴); Hda1 (a histone deacetylase, ⁵⁶¹SK_GGLNDELR⁵⁶⁸); and Rpl36a (a ribosome 60S subunit, ⁸⁵AK_GGVEEMNNIIAASR⁹⁸) (Fig. 3D). In this setting, however, it should be noted that ^REUb might be attached not only directly to the identified proteins but also indirectly to them by way of the Ub moiety of the already Ub-tagged proteins.

Strikingly, the exhaust manual survey of the tandem MS/MS database revealed the ^REUb-derived REIFVK Nt-peptide, verifying the actual existence of ^REUb in vivo. To the best of our knowledge, ^REUb and its conjugated proteins have not been documented until this study (Fig. 3E).

Nt-Arg of ^REUb is flexibly extended to the surface

To gain insight into the mechanistic role of ^REUb, we constructed its neural network-based 3D structure (for details, see Materials and Methods section). The overall surface structure of ^REUb was almost the same as that of ^MQUb, except its Nt-region (Figs. 4A and 4B). Specifically, the top and side views of the predicted models revealed that seven ubiquitylatable Lys residues and C-terminal Gly of ^REUb are located at almost the same surface positions as those of ^MQUb. Tellingly, however, Nt-Arg of ^REUb is stretched outward from the globular Ub protein, whereas ^MQUb retains its Nt-Met buried inward (Figs. 4A and 4B).

Given these findings, using the three-dimensional structure viewer programs we further estimated the H-bond acceptor–donor lengths of Nt-Met-Gln, Nt-Gln, Nt-Glu, or Nt-Arg-Glu of ^MQUb, ^QUb, ^EUb, or ^REUb with their surrounding residues, respectively (Figs. 4C-4F) (see Materials and Methods section for details). In ^MQUb, Nt-Met forms two H bonds with Val17 (each 2.69 Å in length), and Gln at position 2 forms two H bonds with Glu16 (2.54 Å) and with Glu64 (2.56 Å) (Fig. 4C). In contrast, both Nt-Gln of ^QUb and Nt-Glu of ^EUb form only one H bond with Glu64, resulting in H-bond acceptor–donor lengths of 2.77 Å and 2.67 Å, respectively (Figs. 4D and 4E). In the case of ^REUb, however, Nt-Arg forms two H bonds with Val17 (3.44 Å and 2.63 Å) and Glu at position 2 with Glu64 (2.66 Å). Notably, the H-bond acceptor–donor length between the amino group of Nt-Arg and the oxygen atom of Val17 in ^REUb was increased by 3.44 Å, relative to that in ^MQUb with 2.69 Å (Fig. 4F). These results suggest that Nt-Arg-Glu of ^REUb interacts more weakly with its surrounding residues Glu16, Val17, and Glu64 than Nt-Met-Gln of ^MQUb. Consequently, the Nt-Arg flexibly protruding from the Ub globule would increase the accessibility of ^REUb to the factors that recognize it.

In this study, we produced for the first time anti-^REUb-specific antibody and detected the previously unrecognized ^REUb or ^REUb-conjugated proteins in S. cerevisiae (Figs. 2B-2E). The scarce ^REUb and its conjugated proteins were strongly induced in the stationary-growth phase or by oxidative stress. Owing to the scarcity of ^REUb in vivo, its physiological roles and regulation can be confined to the localized pools of Ub that might suddenly arise in response to specific stresses, such as nutrient starvation or reactive oxygen species (Figs. 2G and 2I).

Using anti-^REUb-based immunoprecipitation–tandem mass analysis, this study provides direct evidence of the existence of ^REUb in vivo and identifies the ^REUb-conjugated proteins, suggesting the possible involvement of ^REUb in epigenetic regulation, DNA repair, and protein degradation (Fig. 3). However, further detailed analyses are needed to define the functions of ^REUb in these cellular processes.

We found that ^REUb can be produced in vivo via NME of ^MQUb, Nt-deamination of ^QUb, and subsequent Nt-arginylation of ^EUb (Figs. 2A and 2F). These Nt-modified Ub variants ^QUb, ^EUb, and ^REUb have different physicochemical properties (charge, hydrophobicity, topology, etc.) of their Nt-region (Fig. 4), possibly impeding the ubiquitylation processes of the E1/E2/E3 cascade reactions, the Ub-chain assembly and disassembly, and their interplay with other Ub modifications for versatile biological processes and outcomes. Specifically, Nt-modifications of Ub can directly block the Met1-mediated linear polyubiquitylation, which is important for immune response, cell death, and proteostasis (Iwai et al., 2014).

The actual X-ray crystal structure of ^MQUb (PDB:4XOL, resolution 2.91 Å) reveals that the B factors of ^MQUb, Met-1, and Gln-2 are ~88.1 Ų, ~78.1 Ų, and ~90.5 Ų, respectively, reflecting the rigid property of Nt-Met-Gln region in ^MQUb. Accordingly, the relatively flexible exposure of the Nt-Arg moiety of ^REUb to its surface in comparison to Nt-Met of ^MQUb (Figs. 4A-4C and 4F and 4F) may provide a binding site for the factors that recognize it, such as Ubr1 that targets basic destabilizing Nt-residues including Nt-Arg for the Arg/N-degron pathway (Nguyen et al., 2019; Varshavsky, 2019).

Ubr1 contains at least three substrate-binding sites for basic (type-1) Nt-residues, bulky hydrophobic (type-2) Nt-residues, and internal degrons (Nguyen et al., 2019; Varshavsky, 2019). ^REUb is highly upregulated by stationary-growth phase and oxidative stress that most likely induces Ubr1 activation for protein quality control (Szoradi et al., 2018). If so, is Ubr1 mechanistically or functionally linked to the metabolic regulation of ^REUb under stress conditions and vice versa? As one of parsimonious functions, Ubr1 might mediate the degradation of the apparently faulty ^REUb or its derivatives which can perturb the operation of normal Ub molecule. However, such possibility seems to be low, because ^REUb and ^REUb–protein adducts were rather downregulated in the absence of Ubr1 (Fig. 2H). Conversely, the Nt-Arg-bearing ^REUb might occupy the type-1 substrate-binding site of Ubr1, thereby modulating the degradation of a wide range of its substrates, similarly to the hydrophilin protein Roq1 (Szoradi et al., 2018). The precise roles of Ubr1 in ^REUb regulation and the mechanism behind them remain to be examined.

While no specific enzymes for ^MQUb NME have been identified thus far, Met-aminopeptidases (Map1 or Map2) in yeast most likely remove Nt-Met of ^MQUb, as in the case of other MQ- or MN-starting proteins (Nguyen et al., 2019). However, the possibility that other dedicated proteases might mediate the NME of ^MQUb under specific conditions including oxidative stress cannot be ruled out.

The identification of ^REUb in this study would enormously increase the repertoire and complexity of the Ub code. In addition, ^REUb would be present and actively operate in multicellular organisms as well because NME, Nt-deamination, and Nt-arginylation are conserved in nearly all eukaryotes (Sriram et al., 2011; Varshavsky, 2019).

We are grateful to the current and former members of the Hwang and Lee laboratories for their assistance and advice. We also thank Edanz (www.edanz.com/ac) for editing a draft of this manuscript. This work was supported by grants from the Korean Government (MSIP) NRF-2020R1A3B2078127 and NRF2017R1A5A1015366 (C.-S.H.) and NRF-2020 R1A2C2003685 (C.L.), and the BK21 Plus program (C.-S.H.).

C.-S.H., C.L., K.T.N., and S.J. designed the research. K.T.N., S.J., S.-Y.K., and C.-S.L. performed the research and all coauthors analyzed the data. C.-S.H., C.L., K.T.N., S.J., and S.-Y.K. wrote the paper.

The authors have no potential conflicts of interest to disclose.

Yeast strains, plasmids, and primers used in this study