Activity‐Based Probes for HECT E3 Ubiquitin Ligases

Abstract Activity‐based probes (ABPs) have been used to dissect the biochemical/structural properties and cellular functions of deubiquitinases. However, their utility in studying cysteine‐based E3 ubiquitin ligases has been limited. In this study, we evaluate the use of ubiquitin‐ABPs (Ub‐VME and Ub‐PA) and a novel set of E2–Ub‐ABPs on a panel of HECT E3 ubiquitin ligases. Our in vitro data show that ubiquitin‐ABPs can label HECT domains. We also provide the first evidence that, in addition to the RBR E3 ubiquitin ligase Parkin, E2–Ub‐ABPs can also label the catalytic HECT domains of NEDD4, UBE3C, and HECTD1. Importantly, the endogenous proteasomal E3 ligase UBE3C was also successfully labelled by Ub‐PA and His‐UBE2D2–Ub‐ABP in lysate of cells grown under basal conditions. Our findings provide novel insights into the use of ABPs for the study of HECT E3 ubiquitin ligases.


Introduction
Ubiquitin (Ub) is as mall (76 aminoa cids) and highly conserved protein, best knownf or its role in mediating protein degradation as part of the ubiquitin-proteasome system. [1,2] It is conjugated through an enzymatic cascadei nvolving an E1-activating enzyme( E1), E2-conjugating enzymes (E2s), and E3 ubiquitin ligases (E3 ligases) onto lysine(s)o faprotein target. [3,4] During the ubiquitin cascade,t he C-terminal Gly76 residue of Ub is adenylated by an E1 in an ATP-dependentr eaction prior to the molecule being transferred onto the catalytic cysteine (Cys) residue of the E1.U bi st hen transferred by a trans-thioesterification reactiono nto the catalytic Cys residueo fa nE 2. E3 then facilitates the transfer of ubiquitin from E2 onto the substrate. E3 ligases are categorised into different groupsb ased on the mechanism by whicht hey catalyset his transfer: RING, [5] Ubox, [6] ring-between-ring (RBR), [7] and HECT. [8] HECT and RBR accept Ub on ac atalytic Cys residue, whereas RING and U-box ubiquitin ligases act as scaffolds to provide optimal orientation for the transfer of ubiquitin from E2 to the substrate. The RBR E3s represents an unusual hybrid family,a nd shares features with RING and HECT ligases. In RBR ligases, the enzyme activity is contributed to by an intrinsic C-terminald omain,b ut is also able to recruit thioesther-bound E2 enzymes at aR ING domain. [7] The HECT family was first discovered when investigating the functiono fE 6AP,aprotein that formsacomplex with human papillomavirus E6 oncoprotein types 16 and 18. [9] HECT E3 ligases feature ah ighly conserved C-terminal HECT domain of around3 50 amino acids;t his consists of the E2-binding Nlobe separated from aClobe (containing the active-site Cys) by as hort hinge loop that provides flexibility ( Figure 1). [10] The conserved catalytic Cys receives Ub from the E2 prior to establishing an isopeptide bond between the C-terminal Gly76 of Ub and aL ys e-amino group on the target protein. [11] The HECT family contains 28 members in humans and, based on the N-terminal domain architecture, can be subdivided into three groups: [12] NEDD4 (neural precursor cell-expressed developmentally downregulated 4) has nine members;t he HERCs ubfamily has six members;t he other subfamily is composed of 13 members including the candidate ubiquitin-fusion degradation E3 ubiquitinl igases TRIP12, [13] HECTD1, [14] and HUWE1 (Mule), [15] the proteasomal E3 ligase UBE3C (KIAA10), [16] and the N-end-rule pathway E3 ligases, UBR1-7. [17] Recently,p hylogenetica nalysisl ed to the classification of HECT E3 ubiquitin ligases into class It oc lass VI. [18,19] Some HECT ligasesa re involved in the DNA damage response, apoptosis, and cell proliferation;t his is interesting given that HUWE1,U BR5, and NEDD4 are overexpressed in certain cancers. [20] Therefore, targeting the Cys-based catalytic activity of enzymesi nt he ubiquitin system representsan ew and exciting avenue for anticancer therapy.
To this end, the expansion of the Ub toolbox with activitybased probes (ABPs)h as contributed to ar apid increasei no ur understanding of deubiquitinases (DUBs). [21,22] First-generation ABPs (Ub-ABPs) comprise aU bf or selective recognition by DUBs, with the C-terminal G76c hemically modified with an electrophilic warhead to covalently label the catalytic Cys. Ub-ABPs have been designed with different C-terminal warheads (aldehyde, [23] vinyl sulfone, [24] vinyl methylester (Ub-VME), [25] and propargylamide (Ub-Prg or Ub-PA)). [26] These have been used to identify novel DUB families [25] and to monitorD UB ac-Activity-based probes (ABPs) have been used to dissect the biochemical/structuralp roperties and cellular functions of deubiquitinases.H owever,t heir utility in studying cysteine-based E3 ubiquitin ligasesh as been limited. In this study,w ee valuate the use of ubiquitin-ABPs (Ub-VME and Ub-PA) and an ovel set of E2-Ub-ABPs on ap anel of HECT E3 ubiquitin ligases. Our in vitro data show that ubiquitin-ABPs can label HECT domains. We also providet he first evidence that, in addition to the RBR E3 ubiquitin ligase Parkin, E2-Ub-ABPs can also label the catalytic HECT domains of NEDD4,U BE3C,a nd HECTD1. Importantly,t he endogenous proteasomal E3 ligase UBE3Cw as also successfully labelled by Ub-PAa nd His-UBE2D2-Ub-ABP in lysate of cells grown under basal conditions. Our findings provide novel insights into the use of ABPs for the study of HECT E3 ubiquitin ligases.
tivity. [27][28][29] In addition, they can be used as powerful structural tools for profiling DUB inhibitors and to dissect the mechanisms of action of DUBs. [30,31] For example, the identification of S1 and secondary S2 ubiquitin-binding sites on the catalytic domain of USP21 [32] and OTU (ovarian tumour)f amily DUBs [33,34] has led to an ew generation of Ub-ABPs based on diubiquitin. [35] In contrast to DUBs, the Ub toolboxr emains largelyu nderexplored for Cys-basedE3ubiquitin ligases. E3 ubiquitin ligases have been proposedt oc ontain less reactive Cys nucleophiles, thus suggesting that these enzymes might be less reactive to existing Ub-ABPs. [26] However,r ecent evidence suggests that Cterminal electrophilicp robesc an indeed label Cys residuesi n HECT ligase. Structurals tudies have shown that the HECT domain can be boundb yU bo ni ts Clobe. Furthermore, the flexibility between the Cand Nlobes, important for the catalytic function of HECT domains, can orientatet he C-terminal elec-trophilicw arheado fU b-ABPs and lead to labelling of the catalytic Cys residue in HECTs, in addition to DUBs. [36,37,38] For example, the recombinant HECT domain of HUWE1 was found to react with the Ub-VME-ABP,a nd interestingly multiple Cys residues on the HECT domain were found to be labelledb yt he probe. [39] Importantly,m utation of probe-labelled Cys residues resulted in reduced, but not total loss of, ubiquitin ligase activity,t hus suggesting that Cys residues othert han the catalytic Cys might contribute to enzymatic activity.F urthermore, HUWE1,E 6AP,a nd TRIP12 were labelled by Ub-VME by using mousea nd human cell-line lysates. [39] More recently,r ecombinant HUWE1 was labelledb yU b-PA. [26] However,a sE 2-Ub is the native substrate of E3s, Ub-ABPs do not necessarily provide am echanistically relevant measure of E3 activity;w eq uestion the biological validity of using these probesf or studyingE 3 ligases. The recent development of E2-based ABPs (referred to here as E2-Ub-ABPs or E2-ABPs) that mimic au biquitin- Figure 1. Sequence alignment of HECT domains of human E6AP,HUWE1, NEDD4, UBE3C, HECTD1, and UBR5. Residuess trictly conserved (red boxes) and partially conserved (yellow boxes) are shown. Blue circles indicateresidues mediating E2 binding in the structure of E6AP:UBCH7; [10] the conserved catalytic cysteine is shownasag reen circle, and the flexible hinge loop connecting the Na nd Clobes is shown as an orange box. chargedE 2w ith aCterminal electrophilicw arhead has provided further selectivity for the labelling of catalytic Cys, in particular for RBR and HECT E3 ubiquitin ligases. [40] E2-Ub-ABPs can be customised with defined E2 components. [41] The activated vinylsulfide warhead employed is also less electrophilic. These properties enabled E2-Ub-ABPs to be successfully used to profile the transthiolation of the RBR E3 ligase Parkin, NEDD4L, and ab acterial HECT-like E3 ligase. [40] The crystal structure of the HECT domain of NEDD4L in complex with E2-ubiquitin shows how the Cterminus of Ub is positioned between UBE2D2( bound to the Nlobe) and the active-site Cys in the Clobe. [37,38] This therefore, makes E2-Ub-ABPs more specific in targetingt he catalytic Cys residues of HECT ligasesc ompared to Ub-ABPs.
In this study,w ee valuated the use of Ub-ABPs (Ub-VME and Ub-PA) and E2-Ub-ABPs on ap anel of HECT ligases (NEDD4, UBE3C,H UWE1, HECTD1,a nd UBR5). Our work demonstrates that both types of ABPs can be used effectively to study HECT E3 ligases, either when expressed recombinantly in vitro or at endogenous levels in cell lysates in the case of UBE3C.

Results
We aimed to determinew hich ABPs could be used for HECT E3 ubiquitinl igases:H ECTD1( class II), UBR5 (class IV), UBE3C (class V), NEDD4 (class VI), and HUWE1 (class VI). [19] These HECT ligases show high conservation ( Figure 1) of the N-and Clobes, the hinge loop connecting the N-and Clobes, as well as the catalytic Cys. We first used in vitro auto-ubiquitylation to assess whether GST-taggedH ECT ligases domains (GST-HECT) expressed in Escherichiac oli were active. In the absence of substrates, HECT ligases can auto-ubiquitylate with Ub. We therefore used this to determine whether the recombinant GST-HECTs were functional and active. E2-binding elements originally mapped on E6AP differ between HECT ligases, but they retain the hydrophobic nature,w hich is key for E2 binding. [10] We included HUWE1 because it is the only HECT E3 ubiquitinl igase shown to be labelled by ubiquitin-VME [39] and ubiquitin-PA. [26] First, we explored whether UBE2L3 and UBE2D2 showed any preference for supporting the ligase activity of specific HECTs. UBE2L3 (UBCH7)a nd the UBE2D family E2s show conservation of the catalytic Cys (C86 in UBE2L3, C85 in UBE2Ds), as well as conservation of the criticalr esidue for binding to the Nlobe of HECT ligases( F63 in UBE2L3 and F62 in UBE2Ds;F igure2A). [43] UBE2L3h as been shown to support the activity of HECT ligases including NEDD4,E 6AP,a nd UBE3C. [44][45][46] In agreement with published work, UBE2L3 supported the E3 ubiquitin ligase activity of GST-NEDD4, GST-UBE3C, and GST-HUWE1( Figure 2B-D). Furthermore, UBE2L3 supported the activity of GST-UBR5 ( Figure 2F). Although UBE2L3 led to auto-ubiquitylation of GST-HECTD1 ( Figure 2E,l ane 2, upper panel), it showedw eak functional interaction, as shown by the absence of an autoubiquitylated smear ( Figure 2E,l ane 2, lower panel). This interesting observation emphasises the need to understand better E2:HECTp airs. [47] In contrast to UBE2L3, all UBE2D family members (UBE2D1-3) werea ble to support the activity for all GST-HECTsi no ur panel. Although GST-HUWE1 activity was evident as as trongu biquitylated smear,t here was only am ain higher molecular weight band and aw eak smear by detection with GST.T his could be due to the fact that the GST tag might be modified with ubiquitin therebya ffecting recognition of the epitope by the anti-GST antibody.
Having confirmed that recombinantly expressed GST-HECTs were active, we screened ad ifferent set of ABPs for their potentialu se in labelling the catalytic domain of five HECT ligases. We first tested E2-Ub-ABPs( Figure 3). [40] These novel probesh ave been successfully used to labelt he catalytic Cys of the RBR E3 ubiquitin ligase Parkin. [40] We tested the ability of two E2-Ub-ABPs probes( 7 and 8)f or labelling recombinant GST-HECTs .P robes 7 ( Figure 3A)a nd 8 ( Figure 3B)a re two variants of ap robe design. They differ in the electrophilic warhead, with probe 8 being more reactive. This might be explained by increased electrophilicity,o rt hat its Cterminus more closely mimics the native Ub Cterminus, or both. As expected, incubation of GST-HECTs with these probesr evealed that GST-NEDD4 and GST-UBE3Cs howedastrong labelling with 8 andw eaker labelling with 7.I mportantly,t he labelling was markedly decreased with the corresponding F63A mutant probes( 7F and 8F)f or both GST-NEDD4 and GST-UBE3C, thus validatingt he specificityo fe ach His-UBE2DL3-Ub-ABP. In contrast, no labelling was observed for either GST-HUWE1 or GST-UBR5, whereas aw eak labellings ignal was obtained with GST-HECTD1( Figure 3D). Data obtained with Coomassies taining was confirmed by immunoblotting with an HRP-conjugated anti-His antibody that specifically detected free ABPs as well as ABP-labelled GST-HECT domains of NEDD4 and UBE3C, and to al esser extent GST-HECTD1 ( Figure 3E).
In order to address the lack of labellingo fG ST-HUWE1 and GST-UBR5w ith E2-Ub-ABPs, we next tested whether probe labelling is more efficient under conditions of active ligase activity.I no rder to ensuret hat the modified E2 used for E2-Ub-ABPs still supports HECT ligase activity,w ef irst carried out in vitro auto-ubiquitylationw ith UBE2D2,o rm odified versions of this E2 in which the three noncatalytic Cys residues are mutat-ed (C21S, C107S, and C111S): His-UBE2D2_3xCys_WT and His-UBE2D2_3xCys_F62A ( Figure 4A). These assays revealed that His-UBE2D2_3xCys_WT supports HECTD1a nd HUWE1 but not UBR5 ligase activity ( Figure 4A,l ane 11). In agreement with our previousd ata, His-UBE2D2_3xCys_F62A did not support HECT ligase activity ( Figure 4A,l anes 4, 8, and 12). We next tested whether addition of the probe at the start of the auto-ubiquitylation assay or 30 min later affected the labellingo fH ECTD1.
In order to determine the specificity of ABPs, we analysed the labelling of DUBs, includingt he catalytic domain (CD) of TRABID (AnkOTU) [48] and of OTUD3. We observed strong labelling of TRABID_CD and OTUD3_CD with Ub-VME and Ub-PAa s expected, and virtually no labelling with His-UBE2L3-or His-UBE2D2-ABPs ( Figure 6A). In order to provide as pecific context where the labelling of HECTso verD UBs might be important, we next determined the specificity of ABPs on three proteasomal DUBs:y Ubp6,U CH37, andy Rpn8-11. [49] We found that yUbp6 and UCH37 could be labelled by Ub-ABPs but not by E2-Ub-ABPs. Specifically,y Ubp6 was efficientlyl abelledw ith Ub-VME and Ub-PA, whereas UCH37w as preferentially modified by Ub-VME. The heterodimer yRpn8-11, which contains the Zn 2 + -dependentD UB Rpn11, wasu sed as an egative control as its DUB activity is not Cys-based. [50] Importantly,n either His-UBE2D2-ABP nor His-UBE3L3-ABP convincingly labelled proteasomal DUBs ( Figure 6B). This, together with the fact that E2-Ub-ABPs efficiently modifiedt he HECT domain of the proteasomal E3 ubiquitin ligase UBE3C (Figure 3), suggestst hat the specificity of E2-Ub-ABPs could be particularly useful for dissecting the activitya nd function of Cys-based E3 ubiquitin ligases, for example, in the contexto ft he proteasome, where both E3 ligasesa nd DUBs are present. [51]  Having established that E2-Ub-ABPs and Ub-ABPs can label the HECT domains of ap anel of E3 ubiquitin ligases, we next set out to determine the use of these probes in ac ellular context. Previous studies have reported that E6AP and TRIP12 can be labelled by Ub-VME, [39] whereas HUWE1 was shown to be labelledb yb oth Ub-VME [39] and Ub-PA. [26] Although the recombinant HECT domain of NEDD4L, the RBR domain of HOIP, and the bacterial HECT-like E3 ligase NleL were shown to be la-belled by E2-Ub-ABPs,t he labellingo ft hesee nzymes has not been shown in cell lysate. [40] We tested two methodsf or lysing cells (Figure 7). The first method used 0.5 %T riton X-100 (detergent), and with this lysis mode we werea ble to detect a weak yet convincing level of labellingo fe ndogenous UBE3C in HEK293T( Figure 7A)a nd HeLa cells ( Figure 7B). In contrast, we could not detect any labellingf or NEDD4,H ECTD1, or HUWE1. Importantly,E 2-Ub-ABPs retaineds pecificity,a ss hown by the Figure 7. ABP labelling of endogenous UBE3C in cell lysate. Lysates from HEK293T and HeLa cells were incubated for 3hwith 5 mm probe and analysed by western blotting( WB specifies antibody). Ashift in the molecularweight of the ligase indicates labelling (arrows). An asterisk indicates ac ontaminating band detected by the antibody.Weak yet specific labelling of endogenous UBE3C with His-UBE2D2_WT-ABP (but not the F62A mutant) was detected underb asal/ unstimulated conditionsinA)HEK293T and B) HeLa cells lysed in buffer with a0.5 %T riton X-100. An HRP-conjugated His antibody was used as aloading control for His-UBE2D2-ABPs. In contrasttoU BE3C, labelling of otherH ECT ligases was not detected. C) Western blot showsimproved labelling of endogenous UBE3Cunder basal/unstimulated condition after sonication (no detergent). Additionof2m m DTT during labelling resultedinaslight increase in labelling of endogenous UBE3C. Similarly to A) and B), no labelling is detected for other HECT ligases. D) Western blot analysisofH eLacelllysates obtainedb ys onication and incubation with E2-Ub-ABPs (His-UBE2D2_WT-ABPand the F62A mutant) or Ub-ABPs (Ub-VMEa nd Ub-PA). As in C), UBE3C was efficiently labelled by His-UBE2D2_WT-ABP. We also observeds ome weak labelling with biotin-Ub-PA but not with Ub-VME. HRP-conjugated His and streptavidin-HRP antibodies were used as loading control for the probes. Note that Ub-VMEd id not have at ag. E) Cell lysate from HEK293T overexpressingH A-taggedfull-length-mouse HECTD1w as incubated with ABPs. UBE3C was used as ap ositive control for probe functionality.However,labelling of HA-FL-mHECTD1was not detected in this assay. absence of labelling with the corresponding F62A mutant probe,w hich failed to produce the higherm olecular-weight band for UBE3C (indicative of successful labelling). In the second method, we used as onication protocol without detergent, ando btainedm ore-robust labelling fore ndogenous UBE3C.T his is in line with other studies that used sonication for cell lysis. [40,42] Although our assay is non-quantitative, addition of 2mm DTT during labelling seemedt os lightly increase UBE3C labelling in both HEK293T and HeLa cells ( Figure 7C). As in Figure 7A,H is-UBE2D2_F62A-ABP did not result in any labelling, thus emphasising the specificity of the probe. In contrast to UBE3C,w hich could be readily labelled withouta ny cell stimulation, labellingc ould not be detected for NEDD4, HECTD1, HUWE1, or UBR5.
Havingv alidatedaprotocol for labelling endogenous UBE3C in cell lysate with His-UBE2D2-ABP,w et hen evaluated the labelling of Ub-ABPs with HeLa cell lysate ( Figure 7D). In addition to His-UBE2D2-ABP,w hich showed strong labelling, we also observedaband that (given its size) would correspond to endogenous UBE3C modified with Ub-PA, but not with Ub-VME. This is in line with the in vitro assays,w hich also showed that Ub-PAm ight be better at labellingG ST-UBE3C ( Figure 5C). Although labelling of endogenousH ECTD1 could not be detected,w ew ondered whether overexpression of the fulllength enzyme might improve the detection of any labelled pool of the enzyme. HA-FL-mouse-HECTD1 was transiently overexpressed in HEK293T cells. Followings onication, the cell lysate was incubated with each ABP (5 mm, Figure 7E). UBE3C was used as positive control. However,w ec ould not detect any labelling of HA-FL-mHECTD1,s ot his suggestst hat (under basal conditions at least) the active-site Cys is not accessible for efficient labelling. This might imply that HECTD1 and the other HECT ligasesh ave low intrinsic activity in cells under basal/unstimulatedc onditions.

Discussion and Conclusion
The development of novel chemical biology tools over the last two decades has contributed to our understanding of the cellular functions and the biochemical/structuralp roperties of DUBs. Ubiquitin-aldehyde, the first Ub-based ABP,w as initially used as ag eneral inhibitor of Cys-based DUBs. [23] More recently,n ew-generation Ub-ABPs have been developed with different C-terminal electrophilic warheads. In addition, N-terminal modifications with biochemical (e.g.,H A, His) or cellular (e.g., Cy5 fluorophore) tags have enabled these chemical tools to be used in av ariety of assays aimed at deciphering the function of DUBs. [27,41,52] Ub-VME has been used for DUB profiling in order to identify novel putative deubiquitylating enzymes in cells. The development of novel warheads, in particularU b-PA, has improved the specificityo fA BPs towards DUBs. [26] Structural studies of DUBs have revealed multiple Ub-binding interfaces on the catalytic domain to account for DUB-linkage specificity,a nd these efforts have led to the generation of new diubiquitin-based ABPs, whicha lso offer increased specificity for the targeted DUB. [53] The latest Ub-based ABP,U b-Dha, undergoes trans-thioesterification, and cascades from E1 to E2 and fi-nally E3 without being transferred to the substrate. Therefore, this mechanism-based ligase probe offers au nique opportunity to monitort he activities of componentso ft he ubiquitylation cascade upon drug treatment and for certain pathological cues and stresses. [42] In order to assessw ith more specificity the activity of HECT and RBR ubiquitin ligases, an ew type of ABP based on E2 rathert han ubiquitin wasr ecently developed. [40] These E2-Ub-ABPs are based on an activated vinylsulfide electrophile( AVS), which is included between E2 and Ub and forms the basis for the activity-basedl abellingo fc atalytic Cys nucleophiles in E3 ligases. Although these probes have great potential to provide novel insights on RBR E3 ubiquitin ligasest hrough functional cellular assays and structural studies, thesep robesh ave not yet been extensively explored for their use with HECT E3 ubiquitin ligases. Our data show that the HECT domains of NEDD4, UBE3C,a nd HECTD1c an be labelledi nv itro by E2-Ub-ABPs. Given that specific E2 ligasesp referentially support the activity of particularH ECT ligases, these E2-based ABPs could be used to target HECT ligases with more specificity. [40,41] For example, our data show that UBE2D2 is more efficient at supporting HECTD1l igase activity than UBE2L3. In vitro assays confirmed previousf indings that the HECT domain of HUWE1 can be labelledb yU b-VME [39] and Ub-PA. [26] We now extendt hese observations to show that Ub-VME efficiently labels the HECT domainso fN EDD4,U BE3C, HECTD1, and UBR5;U b-PA labels NEDD4,U BE3C, and HUWE1.T he use of ABPs is summarised in Ta ble 1. The time-course labelling experiment shows that E2-Ub-ABPs quicklya nd efficiently label the HECT domain of NEDD4,t hus suggestingt hat these reagents are particularly powerful to capturet ransient E3 ligase activity under specific cellular conditions. UBE3C is found at the proteasome where it functions as an E4 to extend short Ub chains on difficult-to-degrade sub- Table 1. Summary of the compatibility between ABPs (including recently developed Ub-Dha) [42] and HECT ubiquitinl igases.

NEDD4
His-UBE2L3-Ub( 7)t his study His-UBE2L3-Ub( 8)t his study His-UBE2D2-Ubt his study Ub-VME this study biotin-Ub-PAt his study Ub-Dha His-UBE2L3-Ub( 7)t his study His-UBE2L3-Ub( 8)t his study His-UBE2D2-Ubt his study (also with cell lysate) Ub-VME this study biotin-Ub-PAt his study (also with cell lysate) Ub-Dha [42] HUWE1Ub-VME [39] (also with cell lysate);t his study biotin-Ub-PA [ 26];this study HECTD1His-UBE2D2-Ubt his study Ub-VME this study Ub-Dha [42] (also with cell lysate) UBR5 Ub-VME this study strates. [51,54,55] The fact that His-UBE2D2-ABP can efficiently label UBE3C,b ut not proteasomal DUBs yUbp6,U CH37o r yRpn8-11, suggests that these E2-Ub-ABPs might act as specific tools to inhibita nd/ord etect the activity of this proteasome-resident E3 ubiquitin ligase. Furthermore, we have validated the use of ABPs in cell lysate. Specifically,w es howed that endogenous UBE3C can be labelled with E2-Ub-ABPs (and to al esser extentU b-PA) under basal/unstimulated culture conditions in HEK293T andH eLa cells. Therefore, both ABPs could potentially be used as tools to dissect furthert he fate and functiono fU bc hains extendedb yt his E4 at the proteasome. [51] In contrastt oU BE3C,w edidn ot detect labellingo f endogenousN EDD4, HUWE1,H ECTD1, or UBR5 in HEK293T or HeLa cells under basal/unstimulated condition. This could be attributablet he limit of sensitivity of western blotting, especially if only as mall pool of enzyme has been modified. However, the absence of labelling more likely reflects the low activity of someH ECT ligasesu nder basal/unstimulated conditions. In support of this, for the first time, E2-Ub-ABPs were successfully usedt od emonstrate activation of the en-dogenousR BR E3 ubiquitin ligase Parkin in response to mitochondrial depolarization. [40] Furthermore, NEDD4 ligase activation wass hown to require release of the C2 domain, which can be achieved through NEDD4-interacting proteins NDFIP1 and NDFIP2 or by calcium. [56,57] More recently,H UWE1 ligase activity was shown to be regulated by conformational changes and that p14ARF could regulate HUWE1 ligase activity by maintaining it in an autoinhibited state. [58] The activity of Itch was shown to be activated through ap hosphorylation-induced conformational change, [59] and the activity of WWP2 was regulated by the polymerizings tate of Dishevelled, ak ey component of Wnt signalling. [60] Therefore, ab etter understanding of how the activity of individual HECT ligases is regulated at the molecular level, together with the development of novel ABPs, will be instrumental in dissecting the cellular roles and functions of these enzymes in different cellular contexts. Furthermore, althought he crystal structures of some members of the HECT family have been solved (E6AP, [10] WWP1, [61] SMURF2, [43] NEDD4-L, [38] HUWE1, [62] yeast Rsp5, [36] NEDD4, [37] and WWP2), [63] those of N-end-rule pathway (UBR5)a nd the ubiquitin fusion degradation (UFD) pathway( TRIP12, HECTD1) lack structural knowledge.A BPs could acceleratet his, and in doing so they could increase our understanding of how specific Ub chains are assembled. The recent emergenceo fE 2-Ub-ABPs andt he novel Ub-ABP Ub-Dhag reatly expand the Ub toolboxa nd provide new ways to decipher the cellular functions and structural/biochemical properties of HECT ligases.
Our work provides important new insights form onitoring HECT ligase activity in specific cellular contexts as well as potentially in normal and disease states. [40] For example, UBE3C appearst ob eo verexpressed in glioma tissue, and this is thought to promoteg lioma progression through inhibition of the tumour suppressor gene ANXA7. [64] ABPs might be particularly useful here to monitor UBE3C ligase activity in the search for better therapeutics and to furtherd issect its mechanism of action during cancerprogression.
Immunoblotting: Samples were resolved on 4-12 %S DS-PAGE under reducing conditions and transferred to ap olyvinylidene difluoride membrane (PVDF,0 .45 mm, Thermo Fisher Scientific) or Im- mobilon FL (Merck Millipore) for near-infrared fluorescence detection by an Odyssey Clx (LI-COR, Bad Homburg, Germany). Membranes were blocked in non-fat dried skimmed milk powder (5 %, w/v)i nP BST with Tween-20 (0.1 %) for 1h at RT.M embranes were then probed with the appropriate primary antibodies in blocking buffer overnight at 4 8C. Detection was performed by incubating membranes with HRP-conjugated or IRDye secondary antibodies in blocking buffer at RT for 1h.E nhanced chemiluminescence (ECL; Thermo Fisher Scientific) was used for anti-ubiquitin western blots, and images were acquired on aF USION-SL imager (Vilber Lourmat, Marne-la-VallØe, France). Anti-GST blots were visualized on aLI-COR Clx.
Sonication: Cells were grown and harvested as described above. Cell pellets were resuspended in ice-cold labelling buffer and subjected to sonication in aB ranson Sonifier (15 cycles of 0.3 so n1s off, amplitude 55 %). Lysates were then clarified by centrifugation as described above (13 000 rpm, 15 min, at 4 8C). Clarified lysate (25 mL) was used in ar eaction mixture (30 mL) containing activitybased probe (5 mm) (or no ABP), in the presence or absence of DTT (2 mm).F ollowing incubation for 3h at 30 8C, labelling reactions were stopped by addition of 4 LDS/DTT(30 mL). Samples were then run on 4-12 %S DS-PAGE and detected by western blotting.
Labelling of overexpressed HECTD1 in cell lysate: HEK293T cells were grown in 10 cm dishes and transfected by using Lipofectamine 2000, at 80 %c onfluency,w ith HA-Full-Length-mouse-HECTD1 (5 mg). After 48 h, cells were pelleted, washed in PBS, and lysed by sonication. Labelling and analysis of samples were as above.