Differential regulation of translational stress responses by herpesvirus ubiquitin deconjugases

Abstract

The strategies adopted by viruses to counteract the potential antiviral effects of ribosomal quality control (RQC) that regulates the fidelity of protein translation, ribosome recycling, and the activation of ribosomal and integrated stress responses are poorly understood. Here, we investigated the capacity of the viral ubiquitin deconjugase (vDUB) encoded in the large tegument protein of human pathogenic herpesviruses to interfere with the triggering of RQC upon the induction of translational stress in cytosolic and endoplasmic reticulum (ER)‐associated ribosomes. We found that the vDUBs encoded by Epstein–Barr virus (EBV), human cytomegalovirus (HCMV), and Kaposi sarcoma virus (KSHV) share the capacity to counteract the ubiquitination of RPS10, RPS20, and RPS3, and the UFMylation of RPL26 in cells treated with the translation elongation inhibitor anisomycin (ANS), which resulted in the rescue of model RQC and ER‐RQC substrates from proteasome‐ and lysosome‐dependent degradation, readthrough of stall‐inducing mRNAs, and inhibition of ER‐phagy. In contrast, while inhibiting the ubiquitination of RPS10, RPS20, and RPS3, and rescuing RQC substrates almost as efficiently as the homologs, the herpes simplex virus‐1 (HSV1) encoded vDUB failed to counteract RPL26 UFMylation. Furthermore, it was unable to rescue the ER‐RQC substrate or inhibit ER‐phagy, nor did it promote ZAKα phosphorylation or activate the ISR. Our findings pinpoint important differences in the strategies adopted by these human viruses for regulating translational stress responses.

Translating viral mRNAs is challenging due to structural features that may slow translation or induce ribosome stalling. The viral ubiquitin deconjugases encoded by human pathogenic herpesviruses regulate the cellular response to ribosomal stress by inhibiting various branches of the Ribosomal Quality Control (RQC) and activating Ribosomal Stress Response (RSR) and Integrated Stress Response (ISR), which contribute to remodeling the mRNA translation machinery and promote virus production.

Abbreviations

DUB

deubiquitinating enzyme

EBV

Epstein–Barr Virus

ER

endoplasmic reticulum

ER‐phagy

reticulophagy

HCMV

human cytomegalovirus

HSV1

herpes simplex virus‐1

ISR

integrated stress response

KSHV

Kaposi Coma Virus

RQC

ribosome quality control

RSR

ribosomal stress response

Introduction

The cells have developed sophisticated mechanisms to monitor and maintain the proteome in physiological and stress conditions via the coordinated regulation of protein synthesis and degradation [1]. During viral infection, proteostasis is severely challenged by the antagonistic need of viruses to reprogram the translation machinery toward the rapid production of large amounts of viral proteins and the cellular attempt to curb infection through the activation of the protein kinase RNA‐activated (PKR) that, upon detection of viral nucleic acids, phosphorylates the translation initiation factor eIF2α and represses translation [2]. The capacity of viruses to induce a profound decline of host protein synthesis via selective host shutoff [3] and simultaneously inhibit PKR underscores the need to maintain high levels of active ribosomes capable of sustaining the translation of viral mRNAs. Although a remarkable array of viral tactics for rapidly subverting translation initiation, elongation, and termination has been described [4], very little is known about the molecular events and viral effectors that may counteract the potential antiviral effect of translational stress responses.

The translation of viral mRNAs is likely to be challenging due to the presence of long repetitive sequences, complex secondary structures, suboptimal codon usage, and the frequent occurrence of frameshifting and nucleotide misincorporations that can slow down translation or induce ribosome stalling [5]. While pausing the elongation cycle may resolve minor issues, persistent stalling and ribosome collision trigger the ribosome stress response (RSR) that, via ZAKα‐dependent activation of the JNK and p38 kinases, elicits inflammation and may ultimately cause cell death [6]. The cells have developed rescue strategies that, via activation of the ribosome‐associated quality control (RQC), split the stalled ribosomes, promote the disposal of both damaged mRNA and aberrant translation products, and allow the recycling of ribosome subunits [7, 8]. The RQC is initiated by the recognition of collided di‐ and trisomes [9] or arrested ribosomes [10] by ligases that perform site‐specific mono‐ubiquitination of the 40S particle [11, 12, 13, 14, 15]. In the elongation block‐induced RQC (eRQC), the RING domain ligase ZNF598 ubiquitinates RPS10 (eS10) and RPS20 (uS10) [11, 12, 16, 17, 18], which elicits splitting of the ribosome subunits by the ASC‐1 complex [19, 20], recognition of the 60S‐peptidyl‐tRNA conjugate by NEMF for CAT tailing [21], ubiquitination of the nascent peptide by Listerin [22], its extraction from the conjugate by the VCP/p97 chaperone [23], and subsequent degradation by the proteasome [7, 24]. When ribosomes stall at the initiation codon, RPS2 (uS5) and RPS3 (uS3) are ubiquitinated by the RNF10 ligase, which targets the 40S particle for proteasomal degradation in a process known as initiation RQC (iRQC) that is inhibited by the USP10 ubiquitin deconjugase [15, 25]. RPS2 and RPS3 are also ubiquitinated upon triggering of the integrated stress response (ISR) via phosphorylation of eIF2α by the GCN2 kinase [26]. GCN2 is activated by multiple signals involving stalled translation, including direct binding to uncharged tRNA during amino acid starvation [27] and the detection of collided ribosomes by the ZAKα kinase [28] or the GCN1‐GCN20 regulatory complex [29].

The stalling of ribosomes that synthesize proteins cotranslationally inserted into the endoplasmic reticulum (ER) activates a branch of the eRQC, named ER‐RQC, that is dependent on post‐translational modification of the 60S subunits by the ubiquitin‐like modifier‐1 (UFM1) [30, 31]. Although the components, recognition signals, and detailed organization of the ER‐RQC are still poorly understood, the UFMylation of RPL26 (uL24) by the UFL1/DDRGK1/CDK5RAP3 ligase complex [32, 33, 34, 35] was shown to be essential to initiate a signaling cascade leading to either extraction of the nascent polypeptide to the cytosol for proteasomal degradation [36] or to lysosomal degradation of the peptide alone or together with a small portion of the ER in a process known as ER‐phagy [37, 38, 39]. Failure of this RQC pathway following the ablation of UFL1 [40, 41], the ER‐bound ligase subunit DDRGK1 [32, 37, 42], or the ER‐phagy receptor CDK5RAP3 (also known as C53/LZAP) [34, 43], caused the accumulation of immature polypeptides in the ER, which initiated the unfolded protein response (UPR) and triggered the ISR. It has been proposed that activation of the ISR in response to a failure of the RQC may counteract the pro‐apoptotic effects of RSR by inhibiting global translation while reprogramming the translation machinery in favor of mRNAs containing internal ribosomal entry site (IRES) or upstream open reading frame (uORF) that encode products with pro‐survival and anti‐apoptotic activity [43, 44, 45]. It is noteworthy that, although potentially deadly, delays in the activation of RQC may benefit the cell because the restart of translation after prolonged pausing facilitates the readthrough of stall‐inducing mRNAs [18].

The challenging structure of viral mRNA and the depletion of tRNAs induced by the mass production of comparatively small proteomes [46] suggest that virus‐infected cells may be exquisitely dependent on fine‐tuning of the RQC to secure levels of functional ribosomes sufficient to support virus production. We have previously shown that the viral deubiquitinase (vDUB) encoded in the N‐terminal domain of Epstein–Barr virus (EBV) large tegument protein BPLF1 inhibits both RQC [47] and the ER‐RQC [48] by regulating the ubiquitination and UFMylation of stalled ribosomes, which promoted the translation of viral proteins and enhanced the release of infectious virus particles. EBV, also known as human herpesvirus‐4 (HHV‐4), is a lymphotropic virus that establishes latent infections in most adults worldwide and participates in the pathogenesis of a broad spectrum of lymphoid and epithelial cell malignancies [49] and autoimmune diseases [50]. Despite poor genome sequence conservation and significant differences in host ranges and virus life cycles, the large tegument protein of all animal and human herpesviruses investigated encodes a functional vDUB [51]. Here, we set out to examine whether the vDUBs encoded by EBV, the human cytomegalovirus (HCMV, HHV5), Kaposi sarcoma virus (KSHV, HHV8), and herpes simplex virus (HSV, HHV1) share the capacity to regulate the RQC by interfering with the ubiquitination and UFMylation of ribosomal proteins. We found that the catalytic domain of EBV‐BPLF1, HCMV‐UL48, and KSHV‐Orf64 inhibit the ubiquitination of RPS10, RPS20, RPS3, and UFMylation of RPL26, rescue RQC and ER‐RQC substrates, and impair ER‐phagy. In contrast, while exhibiting comparable inhibition of ribosome ubiquitination and RQC substrate rescue, HSV1‐UL36 failed to induce ZAKα phosphorylation and activate the ISR, did not prevent RPL26 UFMylation, and did not rescue ER‐RQC substrates or inhibit ER‐phagy.

Results

The herpesvirus vDUBs inhibit 40S ribosome ubiquitination, rescue a model eRQC substrate, and promote the readthrough of stall‐inducing mRNAs

Different types of translational stress trigger RQC responses characterized by the site‐specific ubiquitination of 40S ribosome proteins. Upon hindrance of translation elongation, the eRQC is initiated by the recognition of collided ribosomes by the ZNF598 ubiquitin ligase that ubiquitinates RPS10 and RPS20 [18]. To assess whether the vDUBs may counteract these ubiquitination events, the expression of ZNF598 was reconstituted in a previously described ZNF598‐KO cell line [47] by transfection of a ZNF598‐expressing plasmid together with previously described FLAG‐tagged versions of the catalytic domains of EBV‐BPLF1 (aa 1–235), HCMV‐UL48 (aa 1–264), KSHV‐Orf64 (aa 1–205), and HSV1‐UL36 (aa 1–287) [47] or, as a control, the empty FLAG vector (FLAG‐ev). As expected, the overexpression of ZNF598 resulted in significant ubiquitination of RPS10 and polyubiquitination of RPS20 in cells cotransfected with FLAG‐ev. Ubiquitination was strongly inhibited in cells co‐expressing each of the vDUBs (Fig. 1A,B), supporting the conclusion that the viral enzymes share the capacity to counteract the activity of the ZNF598 ligase. In line with the effect observed in reconstituted ZNF598‐KO cells, the vDUBs inhibited the ubiquitination of endogenous RPS10 in HEK293T cells treated for 30 min with 50 ng·mL⁻¹ of anisomycin (ANS) under conditions previously shown to promote random ribosome collision and activation of the RQC [17] (Fig. 1C,D). The presence of a substantial proportion of nontransfected cells likely explains the weaker effect of the vDUBs in this experimental setup.

Fig. 1.

Herpesvirus deconjugases inhibit 40S ribosome ubiquitination. (A) The vDUBs inhibit the ubiquitination of 40S ribosome proteins by ZNF598. ZNF598 knockout HEK293T cells were cotransfected with a ZNF598‐expressing plasmid and FLAG‐ev/vDUBs. After 24 h, the cells were lysed in a buffer containing NEM and iodoacetamide to inhibit DUB activity, and western blots were probed with the indicated antibodies. Blots from one representative experiment out of three are shown in the figure. Unmodified and ubiquitinated species are indicated by arrows. The areas of the blots included in the quantification are indicated by red dotted boxes. (B) The intensities of the Ub‐RPS10 and Ub‐RPS20 bands identified by the red box areas were quantified by densitometry in three independent experiments. Dots indicate the values recorded in individual experiments (n = 3). The mean ± SE% inhibition in vDUB‐transfected versus FLAG‐ev‐transfected cells is shown. (C) The vDUBs inhibit the ubiquitination of endogenous RPS10 in Anisomycin (ANS)‐treated cells. FLAG‐ev/vDUBs transfected HEK293T cells were cultured for 24 h and then treated with 50 ng·mL⁻¹ ANS for 30 min. Endogenous unmodified and ubiquitinated RPS10 are indicated by arrows. The areas of the blots included in the quantification are indicated by red dotted boxes. Representative blots from one out of four independent experiments are shown in the figure (n = 4). (D) Densitometry quantification of the intensity of the ubiquitinated species identified by the red box areas. The mean ± SE percentage inhibition in four independent experiments is shown relative to ANS‐treated FLAG‐ev transfected cells. (E) The vDUBs inhibit the ubiquitination of RPS3. HEK293T cells transfected with FLAG‐ev/vDUBs were treated with 50 ng·mL⁻¹ ANS for 30 min, followed by lysis in a buffer containing NEM to inhibit DUB activity. Western blots were probed with the indicated antibodies. Endogenous unmodified and ubiquitinated RPS10 are indicated by arrows. The area of the scan included in the quantification is indicated by a red dotted box. Blots from one representative experiment out of four are shown in the Fig. (F). The intensities of the Ub‐RPS3 bands identified by the red dotted box were quantified by densitometry in four independent experiments (n = 4). Dots indicate the values recorded in individual experiments. The mean ± SE% inhibition in vDUBs transfected versus FLAG‐ev transfected cells is shown.

The stalling of ribosomes at the initiation codon triggers the iRQC, where the RNF10 ligase ubiquitinates the 40S proteins RPS3 and RPS2 [52]. To assess whether the vDUBs may interfere with this branch of the RQC, the ubiquitination of RPS3 was investigated in HEK293T cells transfected with FLAG‐ev/BPLF1/UL36/UL48/Orf64 following induction of translation arrest by ANS treatment. A significant inhibition of RPS3 ubiquitination was again observed in cells expressing each of the vDUBs (Fig. 1E,F). It is noteworthy that, although HSV1‐UL36 inhibited ribosome ubiquitination somewhat less efficiently compared to the homologs, the differences did not correlate with the strength of the enzymatic activity as assessed by labeling with the Ub‐VS functional probe (Fig. 2A,B).

Fig. 2.

Assessment of enzymatic activity by Ub‐VS labelling. (A) HEK293T cells were transfected with the indicated FLAG‐vDUB vectors. After 24 h, the cells were lysed, and equal amounts of cell lysates were incubated with 1 μm of the HA‐Ub‐VS functional probe at 37 °C. The reactions were stopped at the indicated times by adding NuPAGE loading buffer, and western blots were probed with the FLAG antibody. Enzymatic activity is indicated by a 10 kDa migration shift corresponding to the covalent attachment of the probe to the catalytic Cys residue. One representative experiment out of two is shown. (B) Densitometry quantification of the percentage of the Ub‐VS conjugated vDUB relative to the total (labeled plus unlabeled) in the experiments shown in (A).

Inhibition of the RQC prevents the degradation of aberrant translation products by the proteasome and, upon prolonged ribosome stalling, may allow the restart of translation, leading to the readthrough of the stall‐inducing region [53]. To test the capacity of the vDUBs to functionally interfere with these RQC outcomes, we used a set of reporter plasmids encoding in frame the green fluorescent protein (GFP) and cherry fluorescent protein (ChFP) separated by linkers containing 20 AAA encoded Lys residues (K20). In the K20 reporter (Fig. 3A), the progression of translating ribosomes through the poly(A) stretch causes ribosome stalling and collision, which activates the eRQC and promotes the degradation of the arrested polypeptide by the proteasome [11]. In agreement with our previous finding that BPLF1 and the homologs inhibit the proteasomal degradation of a reporter expressing a GFP construct lacking the stop codon [47], cotransfection of the K20 reporter with FLAG‐tagged BPLF1, UL36, UL48, or Orf64 resulted in significant stabilization of a polypeptide of size slightly larger than full‐length GFP (indicated by a red arrow in Fig. 3B) that is derived from translational arrest at the poly(A) sequence (Fig. 3B,C). Thus, confirming the shared capacity of the vDUBs to rescue a bona fide RQC substrate.

Fig. 3.

The vDUBs rescue a model eRQC substrate and promote the readthrough of stall‐inducing mRNAs. (A) Schematic illustration of the RQC reporter (K20). The reporter expresses in‐frame an N‐terminal GFP following a stretch of Lys residues encoded by AAA codons (K20) and ChFP. Stalling of the ribosome at the poly(A) sequence traps the nascent polypeptide during translation elongation. (B) HEK293T cells were cotransfected with the K20 and FLAG‐ev/vDUBs plasmids for 24 h, and equal amounts of lysates were analyzed in western blots and probed with the indicated antibodies. Blots from one representative experiment out of three are shown in the figure. The sizes of the arrested GFP product and expected size of regular GFP are indicated by black and red arrows, respectively. (C) The Mean ± SE relative intensity of the GFP bands in FLAG‐vDUBs transfected versus in FLAG‐ev transfected cells after normalization to loading control in three independent experiments is shown (n = 3). Dots indicate the values recorded in individual experiments. Significance was calculated by unpaired, two‐tailed Student's t‐test. **P < 0.01, ***P < 0.001. (D) Schematic illustration of the translation‐readthrough dual fluorescence reporter. The reporter expresses, from a single mRNA, the GFP and RFP separated by a linker encoding the FLAG‐tagged villin headpiece (VHP) alone (GFP‐VK0‐RFP) or fused to a stall‐inducing stretch of 20 consecutive lysine residues (GFP‐VK20‐RFP). The linker regions are flanked by Ribosomal 2A skipping sequences (P2A), which allow independent assessment of translation before and after the linker. (E) The vDUBs promote the readthrough of stall‐inducing mRNAs. FLAG‐ev/vDUBs transfected HEK293T cells were cotransfected with either the GFP‐VK0‐RFP (upper panels) or the GFP‐VK20‐RFP reporters (lower panels) before GFP and RFP fluorescence analysis by FACS. After excluding nonfluorescent and dead cells, a readthrough region (RT) was defined in the plots of GFP‐VK0‐RFP transfected cells by gating cells exhibiting a linear correlation between GFP and RFP fluorescence. Cells falling in the stalling region (ST) exhibited decreased RFP:GFP fluorescence ratios. FACS plots from one representative experiment out of two are shown in the figure (n = 2). (F) Mean ± SD of the percentage of transfected cells falling in the readthrough and stalling quadrants in two independent experiments.

To further investigate whether the halted RQC response may also lead to enhanced mRNA readthrough, we utilized a dual fluorescence reporter that expresses GFP and RFP separated by a linker encoding the villin headpiece (VHP) alone (GFP‐VK0‐RFP) or fused to a K20 stall‐inducing domain (GFP‐VK20‐RFP) (Fig. 3D). The presence of ribosomal 2A skipping sequences (P2A) on each side of the VHP region allows the independent assessment of translation before and after the linker. The GFP‐VK0‐RFP and GFP‐VK20‐RFP reporters were cotransfected in HEK293T cells with FLAG‐ev/BPLF1/UL36/UL48/Orf64, and the intensity of GFP and RFP fluorescence was measured by FACS. A readthrough region was gated in cells expressing the GFP‐VK0‐RFP reporter based on linearly correlated GFP and RFP fluorescence, corresponding to an average RFP/GFP ratio close to 1 (Fig. 3E, upper panels). Upon cotransfection of the GFP‐VK20‐RFP reporter with control FLAG‐ev, robust repression of translation downstream of K20 caused the accumulation of approximately 80% of the cells into a stalling region where the RFP/GFP fluorescence ratio fell well below 1 (Fig. 3E, lower left panel). Expression of the vDUBs rescued significant levels of RFP fluorescence (Fig. 2E, lower panels). The results were highly reproducible, with 50–80% of the vDUB‐transfected cells found in the readthrough quadrant (Fig. 3F). Notably, despite the high frequency of RFP rescue, the intensity of RFP fluorescence remained lower in GFP‐VK20‐RFP/vDUB‐transfected cells compared to GFP‐VK0‐RFP/vDUB‐transfected cells (compare Fig. 3D, upper and lower panels), which is likely explained by the restart of translation in different frames [19], resulting in the decreased expression of fluorescent RFP species.

HSV1‐UL36 does not inhibit the ER‐RQC

Having established that the vDUBs share the capacity to inhibit the eRQC and iRQC branches of the RQC, we then ask whether the ER‐RQC is similarly affected. To distinguish between cytosolic and ER‐associated events, the K20 reporter was modified by N‐terminal insertion of an ER‐localization signal and glycosylation site (ER‐K20, Fig. 4A), which selectively promotes the stalling of ribosomes that translate ER‐targeted proteins. HEK293T cells were cotransfected with the FLAG‐ev/BPLF1/UL36/UL48/Orf64 plasmids and the ER‐K20 reporter, and the abundance of the translated products was assessed in immunoblots probed with the GFP antibody. As controls for lysosome and proteasome‐dependent degradation, FLAG‐ev‐cotransfected cells were treated overnight with Bafilomycin A1 (BafA1) or Carfilzomib, respectively. Two faint GFP bands of approximately 50 and 48 kDa were detected in FLAG‐ev‐transfected HEK293T cells (Fig. 4B, black arrows). The lower band corresponds in size to a signal peptide‐containing precursor peptide arrested at the poly(A) sequence. This, together with the resistance to Endo‐H treatment, indicates that the peptide is either released in the cytosol or accumulates at the cytosolic face of the ER. Based on the migration shift induced by Endo‐H treatment (Fig. 4B, red arrow), the upper band corresponds to an ER‐localized glycosylated peptide that lacks the signal peptide. In line with the notion that luminal ER‐RQC substrates are targeted for lysosome‐dependent degradation, the 50 kDa species accumulated in cells treated with BafA1, whereas both Carfilzomib and BafA1 promoted the accumulation of the nonglycosylated precursor peptide. The stabilization of the nonglycosylated precursor peptide by BafA1 suggests that the fraction of the peptide that remains associated with the ER membrane, possibly trapped in the translocon, is subjected to lysosome‐dependent degradation. Expression of FLAG‐BPLF1, FLAG‐UL48, and FLAG‐Orf64 rescued the precursor and glycosylated peptides, confirming that the vDUBs inhibit both RQC and ER‐RQC. In contrast, FLAG‐UL36 has only marginal effects on the ER‐associated and cytosolic versions of the reporter (Fig. 4B,C), suggesting that the HSV1‐encoded enzyme does not regulate ribosomal stress responses at the ER.

Fig. 4.

The HSV1 encoded vDUB does not inhibit the ER‐RQC. (A) Schematic illustration of the ER‐RQC reporter (ER‐K20). The reporter expresses in‐frame an N‐terminal ER‐targeting signal sequence followed by an N‐glycosylation site, GFP, and a stretch of Lys residues encoded by AAA codons (K20) and RFP. Stalling of the ribosome at the poly(A) sequence traps the nascent ER‐inserted polypeptide in the translocon. (B) HEK293T cotransfected with the ER‐K20 reporter and FLAG‐ev/BPLF1/UL36/UL48/Orf64 were cultured for 24 h. For control of lysosome‐ or proteasome‐dependent degradation, aliquots of FLAG‐ev/ER‐K20 transfected cells were treated with 100 nM Bafilomycin A1 (Baf A1) or 100 nm Carfilzomib overnight before harvesting. Bands stabilized by lysosome or proteasome inhibition were indicated as ER‐GFP or Cyto‐GFP, respectively (black arrows). Endo H treatment was used to verify the glycosylation of ER‐GFP (red arrow). Immunoblots from one representative experiment out of five are shown in the Fig. (C) Densitometry quantification of the ER‐GFP or Cyto‐GFP bands. Fold increase was calculated as the ratio between the band's intensity in FLAG‐vDUBs transfected versus FLAG‐ev transfected cells after normalization to loading control. Dots indicate the values recorded in individual experiments (n = 4). The mean ± SE fold change in four independent experiments is shown. Statistical analysis was performed using an unpaired two‐tailed Student's t‐test. *P < 0.05, **P < 0.01, ***P < 0.001, ns, nonsignificant.

HSV1‐UL36 fails to inhibit ribosome UFMylation and ER‐phagy

The UFMylation of 60S protein RPL26 is a distinctive feature of ribosomes that stall at the ER and was shown to be required to activate the ER‐RQC [36]. To investigate whether the vDUB regulates RPL26 UFMylation, we first cotransfected HEK293T cells with S‐tagged RPL26 and FLAG‐ev/BPLF1/UL36/UL48/Orf64, followed by RQC activation by treatment with 50 ng·mL⁻¹ ANS for 1 h. The cells were then lysed in a denaturing buffer containing 20 mM NEM and 10 mM iodoacetamide to inhibit the activity of Ub/UbL deconjugases, and S‐tag immunoprecipitates were analyzed by immunoblot. Three bands of sizes corresponding to RPL26 conjugated to one, two, and three UFM1 moieties were readily detected by the UFM1‐specific antibody in the S‐tag immunoprecipitates of ANS‐treated FLAG‐ev transfected cells (Fig. 5A). The UFMylation bands were virtually absent in the lysates of cells expressing FLAG‐BPLF1, FLAG‐UL48, or FLAG‐Orf64, with a corresponding increase in the intensity of the unmodified RPL26‐S band (Fig. 5A,B). In contrast, FLAG‐UL36 had only a marginal effect, suggesting that this vDUB may lack the capacity to interfere with ribosome UFMylation. Similar results were obtained when the UFMylation of endogenous RPL26 was assayed in ANS‐treated vDUB‐transfected (Fig. 5C,D). Also in this case, the presence of a significant proportion of nontransfected cells likely explains the less prominent effect of the vDUBs compared to the MYC‐RPL26 cotransfection and immunoprecipitation setup.

Fig. 5.

HSV1‐UL36 does not prevent the UFMylation of ribosomes stalled at the ER. (A) HeLa cells transfected with RPL26‐S Tag alone or together with FLAG‐ev/BPLF1/UL36/UL48/Orf64 were treated with 50 ng·mL⁻¹ Anisomycin (ANS) for 30 min, followed by lysis under denaturing conditions to destroy noncovalent interactions and immunoprecipitation with anti‐S‐coupled beads. Arrows indicate mono‐, di‐, tri‐, and tetra‐UFMylated RPL26 detected by UFM1 antibody. Immunoblots from one representative experiment out of three are shown in the figure. The area of the scan included in the quantification is indicated by a red dotted box. (B) Densitometry quantification of the UFMylated bands identified by the red dotted box. The mean ± SE of percentage inhibition was calculated relative to FLAG‐ev transfected cells. Dots indicate the values recorded in individual experiments (n = 3). The significance of the differences between the effect of HSV1‐UL36 versus the homologs was calculated by unpaired two‐tailed Student's t‐test. **P < 0.01. (C) HEK293T cells transfected for 24 h with FLAG‐ev/BPLF1/UL36/UL48/Orf64 were treated with 50 ng·mL⁻¹ ANS for 30 min, followed by lysis in a buffer containing NEM to inhibit DUB activity. High molecular weight species corresponding to the size of mono‐, di‐, and tri‐UFMylated RPL26 (indicated by arrows) were detected in immunoblots probed with antibodies specific for UFM1 or RPL26. Blots from one representative experiment out of four are shown in the Fig. (D) Densitometry quantification of the UFMylated species. The area of the scan included in the quantification is indicated by a red dotted box in (C). The intensity was normalized to the intensity of the RPL26 band in a short exposure of the same blots. The mean ± SE percentage inhibition relative to FLAG‐ev transfected cells was calculated, and dots indicate values recorded in individual experiments (n = 4).

The UFMylation of ribosomes and ER‐membrane proteins was shown to regulate ER‐phagy, which participates in the clearance of ER‐RQC substrates [54]. To test whether the effect of the vDUBs on ribosome UFMylation impacts ER‐phagy, we used a previously described subline of HCT116 that constitutively expresses a Dox‐regulated ER‐phagy dual fluorescence reporter constructed by in‐frame fusion of the coding sequences of the RAMP4 subunit of the ER translocon, enhanced GFP (eGFP), and mCherry Fluorescent Protein (ChFP) [48] (Fig. 6A). Upon Dox treatment, the RAMP4 domain is inserted in the ER membrane with GFP and ChFP facing the cytosol and emitting equal fluorescence, whereas, due to the selective loss of GFP fluorescence at low pH, ER‐loaded autophagolysosomes appear as distinct red fluorescent dots that become yellow upon inhibition of autophagosome acidification by treatment with Baf A1 (Fig. 6B). To assess the effect of the vDUB on this process, the reporter cell line was transfected with plasmids expressing FLAG‐BPLF1/UL36/UL48/Orf64 and cultured for 24 h in the presence of Dox before starvation overnight in EBSS medium, followed by visualization of autophagosomes by confocal microscopy. Analysis of red cargo‐loaded phagolysosomes in transfected and nontransfected cells from the same slide revealed the presence of red dots in the nontransfected cells and cells transfected with FLAG‐UL36, whereas only a diffuse yellow fluorescence was observed in cells expressing FLAG‐BPLF1, FLAG‐UL48, or FLAG‐Orf64 (Fig. 7A). Quantification of the number of red dots in positive and negative cells from the same transfection revealed highly significant suppression of ER‐phagy in cells expressing FLAG‐BPLF1/UL48/Orf64 (Fig. 7B). In contrast, red dots were still detected in the majority of FLAG‐UL36‐expressing cells.

Fig. 6.

Characterization of the ER‐Autophagy tandem reporter. (A) Schematic illustration of the ER‐Autophagy Tandem Reporter (EATR). The reporter expresses in‐frame the coding sequence of the RAMP4 subunit of the ER translocon complex, followed by the coding sequences of eGFP and ChFP. Upon ER insertion of the RAMP4 domain, eGFP and ChFP face the cytosol and emit equal fluorescence, whereas, due to the selective loss of eGFP fluorescence at low pH, ER‐loaded autophagosomes appear as distinct red fluorescent dots. (B) Characterization of the ER‐phagy reporter cell line. HCT116‐EATR were grown overnight in complete medium, or starvation‐inducing EBSS medium in the presence or absence of 100 nm Bafilomycin A1 (Baf A1). When the cells are grown in complete medium, they exhibit a diffuse yellow signal, indicating equal expression of GFP and mCherry. Upon starvation‐induced ER‐Phagy, red dots corresponding to ER‐loaded autophagosomes became apparent due to the quenching of GFP fluorescence caused by low pH conditions. When acidification was hindered by the presence of BafA1, yellow dots were observed (n = 2). Scale bar: 10 μm.

Fig. 7.

HSV1‐UL36 does not inhibit ER‐phagy. (A) Representative confocal images illustrating the failure to accumulate red fluorescent dots in cells expressing vDUBs starved overnight, respectively. Stable HCT116‐EATR cells were transfected with plasmids expressing FLAG‐vDUBs and then starved overnight in EBSS medium before visualizing the formation of ER‐loaded autophagosomes by confocal microscopy. Scale bar: 10 μm. (B) Quantification of the number of red fluorescent dots in vDUBs positive and negative cells from the same transfection experiments. The cumulative data from two independent experiments where approximately 50 vDUB‐positive and 50 negative cells were scored from the same slide are shown (n = 2). Significance was calculated using an unpaired two‐tailed Student's t‐test. ***P < 0.001.

HSV1‐UL36 does not promote the activation of ZAKα

Inhibition of the RQC leads to activation of the MAPKKK ZAKα that is autophosphorylated upon interaction with collided ribosomes and orchestrates the triggering of RSR via phosphorylation of the stress kinases JNK and p38 and ISR via phosphorylation of GCN2 [55]. We have previously shown that ZAKα is activated in cells expressing the EBV‐BPLF1, and this correlates with the phosphorylation of JNK and p38 and activation of a GCN2‐mediated ISR that can be conveniently monitored by the upregulation of ATF4 [48]. To test whether the homologs share the capacity to activate this pathway, the expression of phosphorylated ZAKα, JNK, and p38 and the upregulation of ATF4 were monitored in HEK293T cells transfected with FLAG‐ev/BPLF1/UL36/UL48/Orf64. Like EBV‐BPLF1, the expression of HCMV‐UL48 and KSHV‐Orf64 induced a phosphatase‐reversible shift in the migration of ZAKα of magnitude comparable to that induced by ANS treatment (Fig. 8A). In contrast, transfection of the HSV1‐UL36 had no appreciable effect. The failure to promote the activation of ZAKα correlated with weaker phosphorylation of JNK and p38 and much‐reduced upregulation of ATF4 in cells expressing HSV1‐UL36 compared to the homologs (Fig. 8B,C). Interestingly, the same pattern of vDUB‐induced JNK and p38 phosphorylation and ATF4 activation was observed upon Crispr/Cas9 knockout of ZAKα, suggesting that additional mechanisms are involved in the detection of collided ribosomes and activation of the ribosomal and integrated stress response in cells expressing the vDUBs.

Fig. 8.

HSV1‐UL36 does not promote the phosphorylation of ZAKα and does not trigger the ISR. (A) Lysates of HEK293T transfected overnight with FLAG‐ev/BPLF1/UL36/UL48/Orf64 were left untreated or treated for 45 min at 37 °C with 1 U·mg⁻¹ protein of alkaline phosphatase (P‐ase). Lysates of cells treated with 50 ng·mL⁻¹ Anisomycin (ANS) for 30 min were included as references for ZAKα phosphorylation. Western blots were probed with the ZAKα antibodies. Western blots from one representative experiment out of three are shown (n = 3). (B) Western blots of cell lysates from FLAG‐ev/BPLF1/UL36/UL48/Orf64 transfected control or ZAKα‐knockout HEK293T cells were probed with the indicated antibodies. Images from one representative experiment out of four are shown (n = 4). (C) The intensity of the p‐JNK, p‐p38, and ATF4 specific bands HEK293T transfected cells was quantified by densitometry in four independent experiments. The mean ± SE fold change relative to FLAG‐ev transfected cells is shown. Dots indicate the values recorded in individual experiments (n = 4).

Discussion

The site‐specific post‐translational modification of ribosomal proteins by the covalent attachment of ubiquitin and ubiquitin‐like polypeptides plays a pivotal role in the regulation of the RQC, RSR, and ISR that are activated upon hindrance of translation initiation, elongation, or termination [56]. The conservation of stress‐induced modifications across species points to distinct roles in the regulation of various steps of the translation cycle. Indeed, while the coordinated ubiquitination of RPS10 and RPS20, and RPL26 UFMylation was shown to be critical for activation of the eRQC and ER‐RQC that are triggered by hindrances in translation elongation, RPS3 and RPS2 are ubiquitinated upon ribosome stalling at translation initiation, leading to activation of the iRQC [52], or by events that trigger the ISR [14, 15]. The relationship between these stress responses, their triggering signals, and downstream effects is poorly understood. However, it is generally acknowledged that while activation of the eRQC or ER‐RQC promotes the degradation of aberrant translation products and the recycling of ribosome subunits, the coordinated ubiquitination of RPS3 and RPS2 is accompanied by disposal of the 40S subunit via proteasome‐dependent degradation [15].

Here, we have found that the deconjugases encoded in the N‐terminal domain of the large tegument protein of the human pathogenic herpesviruses EBV, HCMV, and KSHV share the capacity to hinder various branches of the RQC by impairing the ubiquitination of RPS10, RPS20, and RPS3 and UFMylation of RPL26, while the HSV1 encoded vDUB selectively failed to prevent RPL26 UFMylation and the activation of ER‐phagy. The coordinated inhibition of different types of RQC responses may offer several advantages to the viruses. One likely outcome is boosting the proficient translation of viral mRNA containing a variety of translational hindrances, as illustrated by the capacity of the vDUBs to promote the readthrough of stall‐inducing mRNAs. Conceivably, the stabilization of paused/stalled ribosomes by the combined inhibition of disassembly via the recruitment of ASC‐1 to ubiquitinated RPS10/RPS20 and inhibition of 40S degradation mediated by RPS2/RPS3 ubiquitination may be critical for enabling the restart of translation after a prolonged pause. In line with this possibility, we have previously shown that endogenous expression of catalytically active EBV‐BPLF1 is required for efficient translation of the viral genome maintenance protein EBNA1 [48] whose mRNA contains a long G‐C rich domain forming G‐quadruplexes that inhibit translation elongation [57]. Notably, a similar long repeat is found in the KSHV‐encoded genome maintenance protein LANA1 [58], and shorter G‐quadruplex‐forming repeats are present in the protein‐encoding mRNAs of several herpesviruses, including HCMV and HSV1 [59], pointing to a conserved function of the vDUBs in promoting the translation of viral proteins.

We have found that in addition to inhibiting eRQC and iRQC, the vDUBs encoded by EBV, HCMV, and KSHV share the capacity to hamper the ER‐RQC, as illustrated by the inhibition of RPL26 UFMylation, rescue of a model ER‐RQC substrate from both proteasome and lysosome‐dependent degradation, and inhibition of ER‐phagy. The inhibition of this branch of the RQC may provide an additional advantage to the viruses by enhancing the translation of viral glycoproteins essential for the assembly of infectious virus particles. Although its involvement in ER‐RQC and ER‐phagy is firmly established [36, 54, 60], how RPL26 UFMylation regulates these events is not fully understood. It has been proposed that UFMylation may be required to relax the ribosome‐translocon junction, exposing the trapped polypeptide to degradation by the cytosolic RQC [36]. This scenario aligns with the finding that RPL26 UFMylation plays a physiological role in promoting the release of ribosomes from the ER upon translation termination [61, 62]. However, the event that triggers RPL26 UFMylation remains unclear since we have previously shown that it is not affected by the ablation of ZNF598, which serves as the prime reader of ribosome stalling and collision [48]. Based on the observations that RPL26 UFMylation is abolished by treatment with the ubiquitin‐activation enzyme inhibitor TAK243 and that the ubiquitination of ribosomes and associated proteins is severely impaired in EBV‐BPLF1 expressing cells, we have previously proposed that EBV‐BPLF1 may act by reversing constitutive or ribosome‐stall‐induced ubiquitination events that enable UFMylation [48]. In this context, the failure of HSV1 UL36 to regulate the UFMylation of RPL26, stabilize ER‐RQC substrates, and inhibit the activation of ER‐phagy was surprising, given the comparable deubiquitinase activity of viral enzymes. Conceivably, the discrepancy may be explained by differences in the binding repertoire of the HSV1 encoded enzyme, as illustrated by the finding that, unlike the homologs, UL36 fails to regulate IFN responses via interaction with 14‐3‐3 proteins and TRIM25 [63] and does not inhibit macroautophagy via deubiquitination of SQSTM/p62 [64]. The possibility is substantiated by striking differences in the length and charge distribution of helix‐2 within the conserved catalytic domain of the vDUBs [63]. Indeed, helix‐2 of EBV‐BPLF1 was shown to mediate the binding to cullins [65], 14‐3‐3 [63], and presumably other interacting partners.

We have previously reported that vDUBs encoded by EBV, HCMV, and KSHV activate the ISR, while the vDUB encoded by HSV1 failed to do so [47]. Together with the stabilization of ribosomal subunits suggested by the inhibition of RPS3 ubiquitination, the activation of the ISR is likely to play an important role in the reprogramming of translation toward mRNA with alternative start sites such as 5′ untranslated open reading frames (uORFs) and internal ribosomal entry sites (IRES) that are frequently found in viral mRNAs [44, 66] and in the mRNAs in cellular genes with growth‐promoting and anti‐apoptotic functions [44, 45, 67]. In support of this scenario, the effect of BPLF1 on the translation of the EBNA1 mRNA expressed from a lytic promoter that contains both uORFs and IRES signatures [66, 68] was selectively blocked by the inhibition of GCN2 [47]. Interestingly, our findings suggest that, although HSV1‐UL36 inhibits the ubiquitination of RPS3, mimicking thereby the 40S subunit stabilizing activity of the cellular DUB USP10 [52], its failure to activate the ISR correlates with failure to promote the activation of ZAKα. This has several interesting implications. First, it suggests that RPS3 ubiquitination is triggered independently of the ISR, possibly as a consequence of stalling of the 43S preinitiation complex during mRNA scanning or collision of the scanning complex with a stalled 80S ribosome [52]. In addition, the uncoupling of RQC inhibition from ZAKα phosphorylation observed in cells expressing UL36, and the vDUB‐mediated triggering of the ISR in ZAKα‐KO cells, suggest that additional regulatory steps link the detection of collided ribosomes by ZAKα or other readers to the activation of GCN2‐dependent ISR.

During productive infections, viruses are strictly dependent on the cellular mRNA translation machinery for the synthesis of proteins required for virus replication and the assembly of new virus particles. Viral strategies for diverting the cellular translation apparatus toward the mass production of comparatively small viral proteomes have been intensely studied. However, we know very little about how viruses deal with the potentially antiviral effects of quality control systems that ensure the fidelity of ribosome translation and the disposal of aberrant translation products. Our findings provide a first insight into this issue in the context of infection by human pathogenic herpesviruses, pointing to an important contribution of the virus‐encoded vDUBs in regulating ubiquitination events that control different types of RQC responses. The failure of the HSV1‐vDUB to counteract the ER‐RQC and activate the ISR is in line with the finding that this virus has developed unique strategies to counteract the ISR including inhibition of PKR mediated by the Unique short region protein 11 (Us11), inhibition of PERK by physical association with glycoprotein B (gB), inhibition of the GCN1‐mediated activation of GCN2 activation by interaction with gH, and inhibition of all eIF2a‐kinases by the association of ICP34.5 with the PP1c phosphatase [69]. It is tempting to speculate that the distinct behavior of this member of the herpesvirus family may reflect the adaptation to a different range of host cells and the different requirement for activation of cell survival pathways imposed by the substantially shorter replication cycle.

Materials and methods

Reagents

For a complete list of commercial cell lines, reagents, kits, primers, and commercially available or donated plasmids with source identifiers, see Table 1. The identity of the cell lines obtained from commercial sources was certified by the providers.

Table 1.

Reagents used in this paper.

Reagent		Source	Identifier
Antibodies	Working dilutions
Mouse monoclonal anti‐β‐Actin clone AC‐15	1 : 5000	Sigma‐Aldrich (Saint Louis, MO, USA)	Cat# A5441 RRID:AB_476744
Mouse monoclonal anti‐GAPDH	1 : 10 000	Millipore (Burlington, MA, USA)	Cat#CB1001
Mouse monoconal anti‐α‐Tubulin	1 : 1000	Sigma‐Aldrich	Cat# CP06 RRID:AB_2617116
Mouse monoclonal anti‐FLAG	1 : 10 000	Sigma‐Aldrich	Cat# F3165 RRID:AB_259529
Rabbit polyclonal anti‐FLAG	1 : 5000	Sigma‐Aldrich	Cat#F7425 RRID:AB_439687
Mouse monoclonal anti‐S Tag	1 : 1000	Millipore	Cat#71549‐3 RRID:AB_11210600
Mouse monoclonal anti‐HA Tag	1 : 1000	Sigma‐Aldrich	Cat# H9658 RRID:AB_260092
S‐Protein Agarose		Novagen (Madison, WI, USA)	Cat#69704
Goat Anti‐Rabbit IgG (H + L) Antibody, Alexa Fluor 647 Conjugated	1 : 1000	Thermo Fisher Scientific (Walthanm, MA, USA)	Cat#A21245 RRID:AB_141775
Rabbit Monoclonal anti‐UFM1	1 : 2000	Abcam (Cambridge, UK)	Cat# ab109305 RRID:AB_10864675
Rabbit Polyclonal anti‐RPL26	1 : 3000	Abcam	Cat# ab59567 RRID:AB_945306
Rabbit polyclonal anti‐ZNF598	1 : 5000	Abcam	AB241092
Rabbit monoclonal anti‐RPS10	1 : 2000	Abcam	Cat# ab151550 RRID:AB_2714147
Rabbit monoclonal anti‐RPS20	1 : 3000	Abcam	Cat# ab133776 RRID:AB_2714148
Rabbit monoclonal anti‐RPS3	1 : 2000	Abcam	Cat# ab128995 RRID:AB_11145466
Mouse monoclonal anti‐GFP	1 : 2000	Santa Cruz (Dallas, TX, USA)	Cat# sc‐9996 RRID:AB_627695
Rabbit polyclonal anti‐ZAK	1 : 3000	Proteintech (Rosemont, IL, USA)	Cat# 28761‐1‐AP RRID:AB_2918199
Rabbit polyclonal anti‐Phos‐p38 (Thr180/Tyr182)	1 : 1000	Cell Signaling Technology (Danvers, MA, USA)	Cat# 9211 RRID:AB_331641
Rabbit polyclonal anti‐p38	1 : 1000	Cell Signaling Technology	Cat# 9212 RRID:AB_330713
Rabbit polyclonal anti‐ATF4	1 : 1000	Cell Signaling Technology	Cat# 11815 RRID:AB_2616025
Rabbit recombinant anti‐Phospho‐JNK (Tyr185)	1 : 2000	Proteintech	Cat# 80024‐1‐RR RRID:AB_2882943

Chemicals, peptides, and recombinant proteins
IGEPAL CA‐630	Sigma‐Aldrich	I3021; CAS: 9002‐93‐1
Ciprofloxacin	Sigma‐Aldrich	Cat#17850; CAS: 85721‐33‐1
MgCl2	Merck (Uden, Netherlands)	Cat# M1028
Sodium dodecyl sulfate	Sigma‐Aldrich	L3771; CAS:151‐21‐3
Sodium deoxycholate monohydrate	Sigma‐Aldrich	D5670; CAS:145224‐92‐6
Triton X‐100	Sigma‐Aldrich	T9284; CAS:9002‐93‐1
Bovine serum albumin	Sigma‐Aldrich	A7906; CAS:9048‐46‐8
Tween‐20	Sigma‐Aldrich	P9416; CAS: 9005‐64‐5
Trizma base	Sigma‐Aldrich	93 349; CAS:77‐86‐1
Doxycycline cyclate	Sigma‐Aldrich	D9891; CAS: 24390‐14‐5
Anisomycin	Sigma‐Aldrich	A5862; CAS:22862‐76‐6
N‐Ethylmaleimide	Sigma‐Aldrich	E3876; CAS: 128‐53‐0
Iodoacetamide	Sigma‐Aldrich	I1149; CAS: 144‐48‐9
Carfilzomib	Medchem Express (Monmouth Junction, NJ, USA)	Cat# 253339
Bafilomycin A1	Sigma‐Aldrich	B1793; CAS:88899‐55‐2
Complete protease inhibitor cocktail	Roche Diagnostic (Indianapolis, IN, USA)	Cat#04693116001
DAPI	Sigma‐Aldrich	Cat#9542 CAS:28718‐90‐3
Mowiol	Calbiochem (Uden, Netherlands)	Cat# 475904 CAS:9002‐89‐5
Dabco(1,4‐Diazabicyclo[2.2.2]octane)	Sigma‐Aldrich	Cat#D2522 CAS:280‐57‐9
HA‐Ubiquitin‐VS	R&D Systems (Minneapolis, MN, USA)	Cat# U‐212‐025
Kits
jetOPTIMUS DNA transfection reagent	Polyplus (Illkirch‐Graffenstaden, France)	Cat#101000006
Q5® Site‐Directed Mutagenesis Kit	New England Biolabs (Ipswich, MA, USA)	E0554S
DC Protein Assay quantification kit	Bio‐Rad (Hercules, California, USA)	Cat#A500‐0116
SuperSignal™ WestPico PLUS Chemiluminescent Substrate	Thermo Scientific (Waltham, MA, USA)	Cat#XE356732
Endo H	New England Biolabs (Ipswich, MA, USA)	Cat# P0702S
Medium and buffer
DMEM	Sigma‐Aldrich	Cat# D6429
EBSS	Gibco (Waltham, MA, USA)	Cat# 1854705
Experimental models: cell lines
HEK293T	ATCC (Manassas, VI, USA)	RRID:CVCL_0063
HeLa	ATCC	RRID:CVCL_0030
HCT116‐EATR	Liu et al., Autophagy, 2025	N/A
HEK293T ZNF598 knockout	Liu et al., Nat Comm, 2023	N/A
HEK293T ZAK knockout	This paper	N/A
Plasmids
TetOn‐mCherry‐eGFP‐RAMP4	Gift from Jacob Corn	Addgene #109014
ER‐K20	Gift from Yihong Ye	Addgene # 133861
K20	This paper	N/A
pmGFP‐P2A‐VK0‐P2A‐RFP	gift from Ramanujan Hegde	Addgene # 105686
pmGFP‐P2A‐VK(AAA)20‐P2A‐RFP	gift from Ramanujan Hegde	Addgene # 105688
pCMV10‐3xFLAG‐BPLF1	Li et al. PloS Path, 2021	N/A
pCMV10‐3xFLAG‐UL36	Gupta et al. Front Immunol, 2021	N/A
pCMV10‐3xFLAG‐UL48	Gupta et al. Front Immunol, 2021	N/A
pCMV10‐3xFLAG‐Orf64	Gupta et al. Front Immunol, 2021	N/A
pCDNA3.1‐RPL26‐S	Gift from Ron R. Kopito	N/A

Primers used for cloning K20
Forward: GTGAGCAAGGGCGAG	Reverse: CATGGTGGCGACCGG

Cell culture and transfection

The HEK293T, HeLa, and HCT116 cell lines were cultured in Dulbecco's minimal essential medium (DMEM) supplemented with 10% FBS and 10 μg·mL⁻¹ ciprofloxacin (complete medium) and grown in a 37 °C incubator with 5% CO₂. The cell lines were regularly tested for mycoplasma contamination, and only mycoplasma‐free cells were used in the experiments. The cells were transfected using the jetOPTIMUS® DNA transfection reagents according to the protocols recommended by the manufacturers. HEK293T‐ZNF598 knockout cell line was described previously [63]. HEK293T‐ZAKα knockout cell line was generated by transfecting early passages of HEK293T cells with the plasmids ZAK(MLTK) CRISPR/Cas9 KO (Santa Cruz Biotechnology, sc‐405472, Dallas, TX, USA), followed by the selection of GFP‐positive cells by fluorescence‐activated cell sorter (FACS). The knockout efficiency was validated by probing western blots with a ZAKα‐specific antibody.

Immunoblotting and coimmunoprecipitation

The cells were lysed for 30 min on ice in NP‐40 buffer (50 mM Tris, 150 mM NaCl, 5 mM MgCl₂, 1 mM EDTA, 1% Igepal CA‐630, 5% glycerol) supplemented with a protease inhibitor cocktail. To detect UFMylation, the lysis buffer was supplemented with 10 mM N‐Ethylmaleimide (NEM). After centrifugation at 20 000  g for 15 min at 4 °C, the protein concentration of the supernatants was measured with a protein assay kit. Equal amounts of lysates were fractionated in acrylamide Bis‐Tris 4–12% gradient gels (Life Technologies Corporation, Carlsbad, CA, USA). After transfer to PVDF membranes (Millipore Corporation, Billerica, MA, USA), the blots were blocked in Tris‐buffered saline (TBS) containing 0.1% Tween‐20 and 5% nonfat milk. The membranes were then incubated with the primary antibodies diluted in blocking buffer for 1 h at room temperature or overnight at 4 °C, followed by washing and incubation for 1 h with the appropriate horseradish peroxidase‐conjugated secondary antibodies. The immunocomplexes were visualized by enhanced chemiluminescence. For immunoprecipitation, the cells were harvested 24 h after transfection and lysed in IP lysis buffer (20 mM Tris–HCl, pH 7.6, 150 mM NaCl, 1 mM EDTA, 1% Igepal CA‐630, 1% Triton X‐100) supplemented with a protease inhibitor cocktail for 30 min on ice. For immunoprecipitations under denaturing conditions, the cells were lysed for 10 min on ice in a small volume of lysis buffer supplemented with 1% SDS, followed by dilution to 0.1% SDS. The lysates were incubated with 50 μL anti‐FLAG or anti‐S Tag conjugates agarose affinity gel for 3 h at 4 °C with rotation. The beads were then washed with lysis buffer, and the immunocomplexes were eluted by boiling in 2 × NuPAGE Loading buffer supplemented with a sample‐reducing agent. All images were acquired using a ChemiDoc Imaging system (Bio‐Rad), and the intensity of target bands was quantified using the imagelab IM software.

Ub‐VS labeling assay

HEK293T cells were transfected with the vDUBs expressing plasmids in six‐well plates and harvested after 24 h by lysis in 100 μL NP‐40 lysis buffer supplemented with a complete protease inhibitor cocktail. Ten μg of cell lysates was incubated with or without 1 μm HA‐Ub‐VS probe in reaction buffer (50 mM Tris–HCl, pH 7.4, 5 mM MgCl₂, 1 mM DTT, 0.5% sucrose) for the indicated time at 37 °C. The reaction was stopped by adding 4× NuPAGE sample buffer followed by fractionation in NuPAGE 4–12% gradient gel, as described above.

Translation‐readthrough assay

HEK293T cells were plated into 6‐well plates the day before transfection. Semiconfluent monolayers were cotransfected with the FLAG‐ev/BPLF1/UL36/UL48/Orf64 and pmGFP‐P2A‐K0‐P2A‐RFP or pmGFP‐P2A‐K20‐P2A‐RFP and cultured 24 h before detection of GFP and RFP fluorescence by flow cytometry using a BD LSR II SORP apparatus. The data were analyzed using the flowjo software.

ER and ER‐RQC reporter assay and Endo‐H treatment

The RQC K20 reporter was adapted from the ER‐K20 reporter by removing the signal peptide and glycosylation site sequences using the Q5® Site‐Directed Mutagenesis Kit according to the recommended protocol. The K20 or ER‐K20 reporters were cotransfected in HEK293T cells with either FLAG‐Tagged vDUBs or the empty FLAG‐vector as a control. Where indicated, 100 nm Bafilomycin‐A1 (BafA1) or 100 nm Carfilzomib were added to the cultures overnight to inhibit lysosome‐ or proteasome‐dependent degradation, respectively. To assess peptide glycosylation, aliquots of the cell lysates were treated with Endoglycosidase H (Endo H) (New Englands BioLabs, Cat#P0702S, Ipswich, MA, USA) according to the manufacturer's recommendation. Briefly, 10 μg of total cell lysates was denatured at 100 °C for 10 min in the kit's glycoprotein denaturing buffer. The denatured lysates were treated with 500 units of Endo H at 37 °C for 1 h. The reactions were stopped by adding NuPAGE loading buffer supplemented with a reducing agent and boiling. Samples were analyzed using NuPAGE 4–12% gradient gel electrophoresis.

ER autophagy tandem reporter (EATR) assay

The expression of the EATR reporter was induced in HCT116‐EATR cells [48] grown on cover slides by treatment for 24 h with 2 μg·mL⁻¹ Doxycycline before FLAG‐Tagged vDUBs transfection. After 8 h of transfection, the cell culture medium was removed by repeated PBS washing, and the cells were starved by culture for 16 h in Earl's balanced salt solution (EBSS) medium with or without Bafilomycin A1. The cells were then fixed and stained with the anti‐FLAG antibody as described. Images were acquired using a Zeiss LSM900 confocal microscope, and the number of red dots in 45 vDUBs positive or negative cells was counted manually.

Phosphatase treatment

Cells were lysed by incubation for 30 min on ice in NP‐40 lysis buffer (50 mM Tris–HCl, pH 7.6, 100 mM NaCl, 5 mM MgCl₂, 1 mM DTT, 1% Igepal/CA‐630, 5% glycerol) supplemented with EDTA‐free protease inhibitor cocktail. After centrifugation at 20 000  g for 15 min at 4 °C, the protein concentration of the supernatants was measured with a protein assay kit. The lysates were treated with 1 U·mg⁻¹ protein of bovine intestine alkaline phosphatase for 45 min at 37 °C, and the reaction was stopped by adding loading buffer.

Statistical analysis

Plotting and statistical tests were performed with data obtained in two or more independent experiments using the microsoft excel software or graphpad prism 10. No assumptions about data normality were made, and an unpaired two‐tailed Student's t‐test was used to determine statistical significance. Statistical significance is indicated in figures as ns P > 0.05, *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001.

Conflict of interest

The authors declare no conflict of interest.

Author contributions

JL: conceptualization, formal analysis, investigation, visualization, and original draft writing. NN and CMA‐T: investigation, data analysis. MGM: administration, data analysis, writing, review, and editing.

Data availability statement

All the data that support the findings of this study are available in this published article.