A versatile new ubiquitin detection and purification tool derived from a bacterial deubiquitylase

Protein ubiquitylation is an important post-translational modification affecting an wide range of cellular processes. Due to the low abundance of ubiquitylated species in biological samples, considerable effort has been spent on developing methods to purify and detect ubiquitylated proteins. We have developed and characterized a novel tool for ubiquitin detection and purification based on OtUBD, a high-affinity ubiquitin-binding domain derived from an Orientia tsutsugamushi deubiquitylase. We demonstrate that OtUBD can be used to purify both monoubiquitylated and polyubiquitylated substrates from yeast and human tissue culture samples and compare their performance with existing methods. Importantly, we found conditions for either selective purification of covalently ubiquitylated proteins or co-isolation of both ubiquitylated proteins and their interacting proteins. As a proof-of-principle for these newly developed methods, we profiled the ubiquitylome and ubiquitin-associated proteome of the yeast Saccharomyces cerevisiae. Combining OtUBD affinity purification with quantitative proteomics, we identified potential substrates for E3 ligases Bre1 and Pib1. OtUBD provides a versatile, efficient, and economical tool for ubiquitin researchers with specific advantages over other methods, such as in detecting monoubiquitylation or ubiquitin linkages to noncanonical sites.


183
A valine-to-aspartate mutation in OtUBD (V203D) severely impairs its binding to 184 ubiquitin (41). To further validate the specificity of the OtUBD resin towards ubiquitylated 185 proteins, we made an OtUBD(V203D) affinity resin and tested its ability to purify ubiquitin and 186 ubiquitylated proteins. This single mutation greatly diminished the resin's ability to enrich 187 ubiquitylated species (Fig. 2E) and also strongly reduced the total bound protein eluate from the 188 resin (Fig. 2F). This indicates that the ability of OtUBD resin to enrich for ubiquitylated species 189 is based on its binding affinity towards ubiquitin.

190
Taken together, these results indicate the OtUBD resin specifically enriches ubiquitin and 191 ubiquitylated polypeptides as well as proteins that interact with ubiquitin-containing proteins. 201 Under such conditions, the OtUBD resin concentrated ubiquitylated proteins with efficiencies 202 similar to those seen under native conditions (Fig. 3C). At the same time, the denaturing 203 treatment greatly reduced the total amount of proteins eluted compared to native conditions, and 204 the spectrum of purified protein species also changed (Fig. 3D). This suggests that ubiquitylated 205 proteins were specifically enriched by the urea treatment. To verify that OtUBD pulldown following a denaturation step is specific for proteins 207 covalently modified with ubiquitin, we utilized a yeast strain whose endogenous ubiquitin-208 coding sequences were all deleted and replaced with a single plasmid-borne His 6 -tagged 209 ubiquitin sequence (27). The eluted fractions from OtUBD resin pulldowns done after either 210 denaturing or nondenaturing treatments of lysates (Fig. 3A) were then denatured again by 211 incubation with urea or guanidine-HCl (Fig. 3B). The denatured proteins were then applied to a 212 Co 2+ (Talon) resin for immobilized metal affinity chromatography (IMAC) via the His 6 -tagged 213 ubiquitin. If the eluate from the OtUBD resin had contained only (His 6 -)ubiquitylated proteins, 214 most or all of the total proteins should bind to the resin. We observed that when OtUBD 215 pulldowns were done following a denaturing lysate treatment, most of the eluted proteins were 216 indeed bound to the Co 2+ resin (Fig. 3E). By contrast, a large portion of proteins from a "native" 217 OtUBD pulldown remained in the flow-through of the Co 2+ resin (Fig. 3E). The overall levels of 218 ubiquitylated species recovered, however, were comparable between the two treatments ( Fig. 219 3F). Consistent with these findings with bulk ubiquitin conjugates, when we tested whether the 220 proteasome, which binds noncovalently to many polyubiquitylated substrates (49), was in the 221 OtUBD eluates, we readily detected proteasome subunits in the native pulldowns but not 222 pulldowns from denatured lysates (Fig. S2A).

223
OtUBD-based affinity purifications, under either native or denaturing conditions, were 224 also effective with human cell lysates. Both conditions led to similarly enrichment of ubiquitin 225 conjugates (Fig. 3G), but the denaturing pretreatment greatly reduced the amounts of co-226 purifying nonubiquitylated proteins (Fig. 3H). Congruent with this, nonubiquitylated human 227 proteasomal subunits were only present at substantial levels in eluates from native lysates ( Fig.   228 3I, Fig. S2B). Interestingly, low amounts of presumptive ubiquitylated proteasome subunits were 11 229 discovered in OtUBD pulldowns from both native and denatured lysates, and these species were 230 strongly enriched over the unmodified subunits under the latter condition (Fig. S2B).

231
Overall, these results indicate that OtUBD-based protein purification under denaturing 232 conditions can specifically enrich proteins that are covalently modified by ubiquitin.

263
As a final example of single protein analysis, we used OtUBD to detect ubiquitylated 286 Notably, the latter definition will exclude cases where a subpopulation of a protein is 287 ubiquitylated while the non-ubiquitylated population of the same protein interacts noncovalently 288 with ubiquitin or other ubiquitylated proteins. For example, some proteasomal subunits are 289 known to be ubiquitylated (56), but proteasome particles where these subunits are unmodified 290 still interact noncovalently with ubiquitylated proteins. Proteins such as these proteasomal 291 subunits will be excluded from the ubiquitin interactome as defined here. Nevertheless, these 292 definitions provide a general picture of the ubiquitylome and ubiquitin interactome.

293
We performed OtUBD pulldowns of whole yeast lysates with and without prior

321
GO analysis indicated that the yeast ubiquitylome defined by OtDUB binding spans 322 proteins from a wide variety of cellular processes, including multiple biosynthesis pathways, 323 protein localization, vesicle-mediated transport, and protein quality control pathways (Fig. 5C).
324 By contrast, the ubiquitin interactome, as defined above, appeared to yield greater representation 325 in nucleic acid-related processes such DNA replication, RNA transcription, ribosome biogenesis 326 and noncoding RNA processing (Fig. 5D).

327
We also performed OtUBD pulldowns with denatured HeLa cell lysates and compare the 328 data side by side with results obtained from immunoprecipitation using the FK2 antibody, a 329 monoclonal antibody raised against ubiquitin (20). Both OtUBD and FK2 antibody resins 330 efficiently enriched ubiquitylated proteins from HeLa cell lysates (Fig. S4D). The majority of 331 ubiquitylated proteins identified by the FK2 antibody resin were also found in the ubiquitylated 332 proteome identified by OtUBD resin (Fig. 5E). Compared to the FK2 immunoprecipitation, 333 OtUBD pulldowns identified 700 additional ubiquitylated proteins, the majority of which have at 334 least one reported ubiquitylation site in a previous study using diGly antibodies (60).

335
These proteomics experiments demonstrated that the OtUBD affinity resin can be used to

358
We harvested WT, bre1 and pib1 yeast cells and performed OtUBD pulldowns 359 following lysate denaturation. Proteins eluted from the OtUBD resin were subject to label-free 360 quantitative proteomics (Fig. 6A, Fig. S5A, B). Three biological replicates were examined for 361 each group, and each replicate was analyzed by two separate LC-MS/MS runs. Quantitation was 362 performed using total TIC (total ion current) after normalization among the analyzed samples. As 363 expected, histone H2B (identified as Htb2) presented at a much higher level in the ubiquitylome 364 of WT cells compared to that of bre1 cells (Fig. 6B). Interestingly, we identified two different 365 ubiquitylation sites on histone H2B (Htb2) in different samples (Fig. 6C, Fig. S6A (58), the function of this ubiquitylation remained to be 371 studied. Besides histone H2B, we also identified 16 other proteins present in significantly higher 372 levels in the WT cell ubiquitylome compared to bre1 cells (Fig. 6D). In addition, 35 proteins 373 were only detected in the ubiquitylome of WT cells but not bre1 cells (Fig. 6E). Taken 374 together, these proteins are considered potential Bre1 substrates. Interestingly, some of these 375 proteins (Fig. 6D, E, green) have been shown to be metabolically stabilized in bre1 cells in an 376 earlier study (68), which indicates that they could be direct or indirect proteolytic ubiquitylation 377 substrates of Bre1.

378
Analogous to the Bre1 data, we identified three proteins whose ubiquitylated forms were 379 found at significantly higher levels in WT cells versus pib1 cells (Fig. 6F, G) and 38 proteins 380 that were detectably ubiquitylated only in WT cells but not pib1 cells (Fig. 6H). Of these 381 proteins, six have been shown previously to be stabilized in pib1 cells (Fig. 6G, H) Data 2), substantiating the idea that OtUBD can enrich proteins with non-lysine 390 ubiquitylation sites. We confirmed one of these sites by manual validation of the spectrum 391 assignment (Fig. S6C).

394
Protein ubiquitylation continues to be of great interest due its vital contributions to many 514 supplemented with lysozyme and DNaseI, incubated on ice for 30 minutes and lysed using a 515 French press. Lysates were clarified by centrifugation for 1 h at 4C at 10,000 rcf before being 516 subjected to Ni-NTA (Qiagen) affinity purification following the manufacturer's protocol.

517
For purification of OtUBD variants using pRT498-based plasmids, the proteins eluted from 518 the Ni-NTA resin were subject to buffer exchange in a 50 mM Tris•HCl, pH 7.5, 150 mM NaCl 519 buffer supplemented with 10 mM tris(2-carboxyethyl)phosphine) (TCEP) (from a 1M TCEP stock 520 neutralized with NaOH to pH 7) using a centrifugal filter device (Amicon, 3000 MWCO) 521 following manufacturer's protocol. His-tagged TEV protease was added to remove the His 6 -MBP 522 tag, and the mixture was incubated on ice overnight. The cleavage mixture was then allowed to 523 flow through a column of clean Ni-NTA resin to capture the cleaved His 6 -MBP tag. The flow- 550 For rabbit primary antibodies, the HRP-linked anti-rabbit IgG secondary antibody (GE Healthcare, 551 NA934) was used at a dilution of 1:5000 or 1:10000. For mouse primary antibodies, the HRP-linked 552 anti-mouse secondary antibody (GE Healthcare, NXA931V) was used at a dilution of 1:10000.
574 The resin was pelleted by centrifugation at 5,000 x g for 30 seconds, and the supernatant was 575 removed. One mL of yeast lysate (1-2 mg/mL) prepared in column buffer freshly supplemented 576 with protease and DUB inhibitors (cOmplete mini EDTA-free (Roche), 10 mM NEM, 2 mM 577 PSMF) was added to the beads. (For detailed methods of lysate preparation, see section "Ubiquitin 578 pulldown with protein-linked resins" below.) The mixture was incubated with rotation at 4°C for 579 2 hours. The resin was washed 5 times with 1 mL column buffer and then eluted by incubating 580 with column buffer containing 50 mg/mL maltose for 2 hours at 4°C with rotation. Alternatively, 581 bound proteins could be eluted by incubating with SDS sample buffer for 15 minutes at room 582 temperature.

583
In the alternative incubation method described in Fig. S1B, MBP-OtUBD was first 584 incubated with yeast lysate for 4 hours at 4°C with rotation. The mixture was then added to the 585 amylose resin and incubated with rotation at 4°C for another 2 hours, followed by the same 586 washing and elution steps described above.