Identification of RING ligases in Simkania negevensis

To identify ubiquitin ligase candidates in the sequenced genome of Sne [41], the Sne subset of UniProt [42] was searched with generalized profiles [43] representing classical RING finger domains and several RING finger variants. When using a profile derived from a multiple alignment of established RING domains, only a single Sne ORF, SNE_A12920 (SneRING), reached a significant score of p < 0.01. To validate this finding, BLAST and profile searches were used to find other bacterial proteins related to SneRING. A small family was identified, of which representative members include Sim-KFB93 and Sim-QNJ27 from bacteria belonging to the Simkaniaceae family, and Chl-HKST.2, belonging to an unclassified Chlamydiia (Fig 1A). An Alphafold [44] model of SneRING shows a two-domain architecture with a RING-like fold forming the first domain (residues 14–72), followed by a linker region and a C-terminal α/β-fold domain without informative sequence or structural similarities (Fig 1B). Like canonical RING fingers, the RING-like domain of SneRING is predicted to coordinate two Zn²⁺ ions, albeit using somewhat atypical ligand residues. In the model, the first Zn²⁺ ion is coordinated by Cys-17, Cys-20, His-40, and Cys43, while the second Zn²⁺ ion is contacted by Asn-32, Cys-35, Cys-54, and Cys57.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. SneRING is a predicted RING domain ligase.

(A) Multiple alignment of the RING domains of SneRING (SNE_A12920), some bacterial relatives (Sim-KFB93: Simkaniaceae bacterium RefSeq:QVL56959; Sim-QNJ27: Simkaniaceae bacterium RefSeq:MDJ0651792; Chl-HKST: Chlamydiia bacterium Uniprot: A0A960X5Q0), and sequences of the two best DALI hits (pdb:3LRQ (TRIM37 from Homo sapiens) and pdb:2CKL (RING2 from Mus musculus)). Residues invariant or conserved in at least 50% of the sequences are shown on black and grey background, respectively. Residues involved in the coordination of Zn₁ and Zn₂ are highlighted in red and magenta, respectively. (B) Alphafold model of SneRING shows the N-terminal RING-like domain (blue) and a C-terminal domain of unknown function (DUF, green). The Zn₁ and Zn₂ ions coordinated by the RING-like domain are shown in red and magenta, respectively. (C) Alphafold model of the RING domain of SneRING (blue) in contact with UbcH5b (grey). The Zn₁ and Zn₂ ions coordinated by the RING-like domain are shown in red and magenta, respectively. The predicted contact residues Glu-22 (E22) and Ser-19 (S19) and their interaction partners are labeled.

https://doi.org/10.1371/journal.ppat.1013626.g001

A DALI [45] search using the RING-like region of SneRING identified as significant top-hits the established RING fingers of TRIM37 (pdb:3LRQ, Z-score 9.1) and RING2 (pdb:6WI7, Z-score 8.9). RING2 possesses a typical RING finger, where Zn₁ is coordinated by two CxxC dyads and Zn₂ is bound by a CxH and a CxxC dyad. In SneRING, both coordination spheres are slightly altered: His-40 replaces Cys as the 3^rd ligand of Zn₁, while the first two ligands of Zn₂ are NxxC rather than the canonical CxH (Fig 1A). Functionally, these changes might be inconsequential, as the Alphafold model of the SNE_A12920 RING domain in contact with UbcH5 predicts SneRING to form a complex with E2 enzymes resembling that of other RING ligases (Fig 1C). As with other RING fingers, the most important contacts are made by the region surrounding the first Zn-binding dyad. While most other RING domains contact E2s by a conserved hydrophobic residue before the second Zn-binding cysteine, SneRING carries the atypical Ser-19 at this position, which does not form a specific E2 interaction in the model. Instead, Glu-22 forms short-distance salt bridges to Lys-4 and Lys-8 of UbcH5 (Fig 1C). We reasoned that a mutation of Glu-22 combined with a Ser-19 mutation might abrogate E2-binding. Indeed, when subjecting an S19R_E22R variant of SneRING to Alphafold modeling with UbcH5, no productive complex was formed.

Besides the significantly-scoring SneRING, three further RING-like domains could be identified bioinformatically. A closer inspection of the borderline-significant matches identified the uncharacterized ORF SNE_A08700 and its close relative (and genomic neighbor) SNE_A08690 to have a conspicuous arrangement of cysteine and histidine residues. The assumed RING-relationship of these two proteins is supported by the conservation pattern of their sequence family and by Alphafold modeling: SNE_A08700 is predicted to use eight canonical ligands for coordinating two Zn²⁺ ions (S1A Fig); the best DALI match to an established RING-finger was RNF125 (pdb:5DKA, Z-score 5.6). Sequence conservation between RNF125 and the SNE_A08700 family shows a big insertion after the 3^rd Zn ligand (S1B Fig), which explains the poor score in the initial profile search. The closely related protein SNE_A08690 has lost three ligands for the first Zn²⁺ ion and probably coordinates only a single ion. A 4^th RING candidate (SNE_A19470) was picked up in profile searches using the SP-RING family, a subgroup of RING-like domains usually involved in SUMO conjugation [46]. However, a closer inspection suggested that SNE_A19470 does not contain an SP-RING domain but rather a (degenerate) conventional RING that has lost its ligands for Zn₁ coordination (S1C Fig). The best DALI matches were TRIM69 (pdb:6YXE, Z-score 8.5) and TRIM2 (pdb:8A38, Z-score 8.3), which both contain a structurally similar RING domain, but with a complete set of eight ligand residues (S1D Fig).

Considering that SneRING was the only protein with a RING-like domain that has been identified in a previously published proteomics analysis of the ER/SnCV membrane of infected HeLa229 cells on day 3 p.i. [24], we focused on further characterization of this enzyme.

SneRING is an active E3 RING ligase with an E2 enzyme-interacting region within its predicted RING domain

We next tested the activity of the SneRING using an in vitro autoubiquitination assay. For this, we purified SneRING and incubated it with a ubiquitin-activating (E1) and a ubiquitin-conjugating enzyme (E2), as well as free ubiquitin and ATP to allow the enzyme cascade to covalently attach ubiquitin to SneRING. Already after 30 min of incubation, we could detect a strong ubiquitin signal that increased over time, while in the negative control without SneRING, only monoubiquitin could be detected (Fig 2A). The resulting signal represented an autoubiquitinated SneRING because a similar pattern could be observed when the reaction was analyzed using SneRING antibodies (S2A Fig). This shows that the SneRING is an active ubiquitin ligase.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. SneRING shows ubiquitination activity in an in vitro assay.

(A) Purified recombinant SneRING was mixed with recombinant E1 ubiquitin-activating enzyme, E2 ubiquitin-conjugating enzyme (UbcH5b), ubiquitin, and ATP for the indicated periods. A reaction without the ligase served as a control. Samples were analyzed by SDS-PAGE and western blot, using an antibody against ubiquitin. (B) In vitro autoubiquitination reaction of wildtype SneRING and SneRING-ligase with mutated E2 enzyme interacting region (SneRING S19R_E22R) was performed for 5 h as in A and analyzed by SDS-PAGE and western blot, using anti-ubiquitin antibodies. (C) In vitro autoubiquitination reaction of wildtype SneRING was performed as in A in combination with indicated E2 enzymes for 5 h, followed by SDS-PAGE and western blot analysis using ubiquitin antibody. (D) SneRING was incubated with E2 enzyme UbcH5b or UBE2T in an in vitro autoubiquitination assay as described in A. Samples were taken at indicated time points and tested by immunoblot using the primary antibody against ubiquitin. Poly-Ub, polyubiquitin; Di-Ub, diubiquitin; Mono-Ub, monoubiquitin.

https://doi.org/10.1371/journal.ppat.1013626.g002

Since Alphafold modeling indicated the importance of Ser-19 and Glu-22 for the interaction with an E2 enzyme, we replaced these two amino acids with arginine to generate a catalytically inactive form and performed an in vitro autoubiquitination assay with the purified mutated enzyme for 5 h. While a strong ubiquitin signal could be detected in the presence of the wild-type SneRING, no ubiquitin chains were formed in the reaction containing SneRING S19R_E22R (Fig 2B). This confirms that the N-terminal domain of the SneRING interacts with the E2 ubiquitin-conjugating enzyme and is necessary for its function.

To test whether SneRING can generate ubiquitin chains when combined with other E2 enzymes besides UbcH5b, we performed an in vitro autoubiquitination assay for 5 h using selected ubiquitin-conjugating enzymes. Only UbcH5b- and, to a lesser extent, UBE2T-containing reactions showed the presence of the polyubiquitin signals, while no ubiquitin chain assembly occurred with E2 enzymes UbcH7, UBE2A, UBE2B, and UBE2W (Fig 2C). The time course analysis of the autoubiquitination reaction in combination with UbcH5b showed strong ubiquitination after 30 min of incubation, increasing over time (Fig 2A and 2D). In contrast, in the presence of UBE2T, nearly no reaction was observed after 30 min, and some polyubiquitin signals were detectable only 2 h and 5 h later (Fig 2D). In conclusion, among the tested E2 enzymes, SneRING appears to be most active in combination with UbcH5b, while a moderate reaction also occurs in the presence of the ubiquitin-conjugating enzyme UBE2T.

Ubiquitin chains generated by SneRING are cleaved by deubiquitinating enzymes specific for the K11/K63 linkage type

To characterize the type of ubiquitin chains formed by SneRING, we performed linkage-specific chain-cleaving assays (UbiCRest assays) [47]. Ubiquitin chains generated by SneRING after in vitro autoubiquitination for 5 h were incubated with a panel of different DUBs that either process any chain type, such as ubiquitin-specific peptidase 21 (USP21) [48], or have specificity against certain ubiquitin chain linkages. When incubating the reaction with the USP21, no remaining polyubiquitin signal could be detected. Treatment with the K11-specific DUB Cezanne [49] or with the K63-specific DUB AMSH (associated molecule with an Src homology 3 domain of signal transducing adaptor molecule, STAM) [50] reduced the polyubiquitin signal. In contrast, no cleavage was observed upon incubation with the DUB SneOTU that specifically cleaves linear M1 ubiquitin chains or SneVTD that shows cleavage specificity against K6-linked ubiquitin chains [28]. Furthermore, also the K48-specific DUB OTUB1 [51] did not visibly reduce the ubiquitin signal (Fig 3A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. SneRING generates K6-, K11-, K48-, and K63-linked ubiquitin chains that can be digested by specific DUBs.

(A) An in vitro autoubiquitination reaction, as in Fig 2A, was incubated for 5 h, after which a UbiCRest assay was performed using the indicated chain type-specific DUBs for 1 h at 37 °C. All DUBs were present at a final concentration of 1.5 µM, except OTUB1, which was present at 3 µM. Ubiquitin chain cleavage was analyzed by SDS-PAGE and western blot using a primary antibody against ubiquitin and SneRING as a control. (B) An in vitro autoubiquitination reaction, as in A, was incubated with Cezanne (K11-linked ubiquitin chain-specific DUB) and AMSH (K63-linked ubiquitin chain-specific DUB). Degradation of ubiquitin chains was detected by SDS-PAGE and western blot using anti-ubiquitin antibody. Poly-Ub, polyubiquitin; Di-Ub, diubiquitin; Mono-Ub, monoubiquitin. (C, D) A 5 h in vitro autoubiquitination reaction of SneRING with UbcH5b (C) or UBE2T (D) was analyzed by mass spectrometry. The graphs show log₁₀ intensities of GlyGly-modified residues for the different linkage types, as determined by MaxQuant.

https://doi.org/10.1371/journal.ppat.1013626.g003

To test whether the ligase only generates these two linkage types or if other ubiquitin chain types are produced at a lower level, a di-UbiCRest assay was performed. For this, an in vitro SneRING autoubiquitination reaction was incubated with the K11-specific DUB Cezanne, with the K63-specific DUB AMSH, or with both enzymes simultaneously. As already observed, Cezanne and AMSH led to a reduction of the ubiquitin signal, demonstrating the presence of K11- and K63-linked ubiquitin chains. However, the combined activity of both DUBs did not completely degrade ubiquitin chains generated by the SneRING (Fig 3B), indicating that other linkage types are formed by the SneRING in addition to K11- and K63-linked ubiquitin chains. Mass spectrometry of the in vitro SneRING autoubiquitination reaction in the presence of UbcH5b (UBE2D2) mainly detected K11- and K63-linked GlyGly-modified peptides, with about 5-fold lower signal detected for K6- and K48-modified peptides, along with small amounts of K33-modified peptides (Fig 3C). Similar results with overall weaker signal strength were obtained for the reaction containing the E2 enzyme UBE2T (Fig 3D). In conclusion, SneRING forms in vitro mostly K11- and K63-, in addition to K6- and K48-linked ubiquitin chains.

To further validate this finding, in vitro autoubiquitination assays performed with UbcH5b were tested by immunoblot analysis for the presence of specific ubiquitin chains. We could confirm the presence of K63- (S2B Fig), K48- (S2C Fig), K11- (S2D Fig) and K6-linked ubiquitin chains (S2E Fig). However, when we used an antibody against M1-linkages, which were not detected by UbiCRest assay and mass spectrometry, we only saw a signal in the positive control, but not in the SneRING reaction, confirming the previous results (S2F Fig).

SneRING is expressed during the infection of human cells and its possible natural host, Acanthamoeba castellanii

We next analyzed the expression pattern of SneRING during infection. Since Sne can infect and grow in a variety of cells [52], as well as survive and multiply within Acanthamoeba castellanii, we measured the SneRING mRNA levels during the infection cycle of various cell lines and amoeba by RT-qPCR analysis and normalized it against the levels of the bacterial 5S RNA. In comparison, we measured the mRNA levels of a highly expressed gene, Sne chaperone SnGroEL. SneRING was expressed in all host cells we tested, though to a lesser extent than SnGroEL (S3 Fig). Expression intensities of SneRING in different human cell lines were comparable. In HeLa229 cells, we observed somewhat higher SneRING expression on day 2 p.i. compared to day 4 p.i. (S3A and S3B Fig).

Sne can infect and grow in the monocyte/macrophage cell line U937 [3], as well as in the THP-1 cell line that mimics the macrophage-like state after differentiation by PMA [53], in contrast to C. trachomatis [54]. We analyzed and compared the expression of SneRING in the THP-1 cell line and primary human M2-like macrophages. Based on the observation that infectious Sne particles can be detected on day 6 p.i. in U937 cells [3], we included this time point in our experimental setup. RT-qPCR analysis showed slightly higher expression of SneRING in THP-1 cells than in M2-like macrophages. There were no notable differences in the mRNA levels between different days p.i. (S3C and S3D Fig).

By taking up a high number of bacteria, many free-living protozoa mimic professional phagocytes [55]. Sne can survive and grow within amoebae [7]. Since Sne establishes a long-term infection in amoebae, we measured the mRNA level of SneRING only on day 4 p.i. A. castellanii culture was cultivated at 30 °C instead of 37 °C, which we used for the cultivation of human cell lines (or 35 °C for Sne-infected). mRNA of SneRING was detectable, but the levels varied strongly between the samples, both for SneRING and SnGroEL (S3E Fig).

SneRING localizes to the host cell cytoplasm, mitochondria, and ER after overexpression

We next studied the localization of SneRING in the host cell. Since we were unsuccessful in producing an antibody against SneRING that would recognize the native protein in immunofluorescence studies, we assayed the FLAG-tagged protein expressed from a plasmid in U2OS cells. SneRING was successfully expressed and exhibited cytosolic distribution as observed by immunofluorescence microscopy (Fig 4A). The FLAG-tagged protein could also be identified by western blot after transfection (Fig 4B). Upon subcellular fractionation and separation of the cells into the crude mitochondrial (containing mitochondria and ER), the light membrane (containing ER, Golgi, endosomes, and plasma membrane), and the cytosolic fraction, SneRING signal was distributed between all fractions (Fig 4C). In comparison, Mic60, a mitochondrial inner membrane protein, localized almost exclusively to the mitochondrial fraction, while calnexin, the integral protein of the ER, was detected in both the light membrane and the crude mitochondrial fraction. GAPDH and tubulin were associated with every fraction but mainly localized in the cytosol, somewhat resembling SneRING distribution (Fig 4C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. SneRING localizes to the host cell cytosolic, light membranes, and inner mitochondrial membrane after overexpression.

(A) U2OS cells were transfected with a FLAG-tagged version of the SneRING using the transfection reagent PEI MAX. 72 h post-transfection and 2 days p.i., cells were fixed and stained using DAPI (blue channel), and a primary antibody against FLAG followed by the fluorophore-coupled secondary antibody (green channel). Images were taken using laser scanning confocal microscopy. The scale bar represents 30 µm. (B) U2OS cells were transfected as in A and, 36 h post-transfection, analyzed by SDS-PAGE and western blot using antibodies against the FLAG-tag and tubulin. (C) U2OS cells were transfected with FLAG-SneRING as in A, and 48 h post-transfection, the cells were separated by differential centrifugation into a crude mitochondrial (Mito), light membrane (LM), and cytosolic (Cyt) fraction. Total cell lysate of non-transfected cells (Control) and transfected cells (Total) served as controls. The samples were analyzed by SDS-PAGE and western blot using antibodies against Mic60, Calnexin, Tubulin, GAPDH, and FLAG. (D) Hek293T cells were transfected with FLAG-SneRING using the calcium-phosphate method. After 48 h of expression, mitochondria were isolated and then incubated in isotonic buffer (-Swelling) or in hypotonic buffer (+Swelling) to rupture the outer mitochondrial membrane, combined with the treatment using 50 µg/mL of protease K (PK). For complete solubilization, mitochondria were treated with 1% Triton X-100 (1% Triton) and PK and precipitated using 72% trichloroacetic acid. Samples were analyzed by SDS-PAGE and western blot, using antibodies against Calnexin, Tom70, CPOX, Mic60, Tim44, and FLAG-tag. (E) Mitochondria as in D were either sonicated in a high-salt buffer (500 mM NaCl pH 7.6) or subjected to carbonate extraction (100 mM Na₂CO₃ pH 10.8 or 100 mM Na₂CO₃ pH 11.5). Pellet (P) and supernatant (S) fractions were separated by ultracentrifugation, while untreated total samples (T) served as controls. Samples were analyzed by SDS-PAGE and western blot using antibodies against Hsp60, Sam50, Mic60, and FLAG. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; Tom70, translocase of the outer mitochondrial membrane 70; CPOX, coproporphyrinogen oxidase; Mic60, mitochondrial contact site and cristae organizing system 60; Tim44, translocase of the inner mitochondrial membrane 44; Sam50, sorting and assembly machinery 50.

https://doi.org/10.1371/journal.ppat.1013626.g004

Immunofluorescence microscopy of HeLa229 cells with GFP-labeled mitochondria (S4A Fig) or dsRED-labeled ER (S4B Fig) that were transfected with FLAG-tagged SneRING gave a similar picture, with some overlap of the signal, indicating possible association of SneRING with these organelles. This did not significantly change upon infection with Sne (S4A and S4B Fig, lower panels). Notably, Sne-infected cells show fragmentation of the mitochondria, which is absent in cells transfected only with SneRING. In both cell lines, the FLAG-tagged protein is pushed to the cell periphery with the cytosol and is not identified inside the SnCV when the cells are infected (S4 Fig).

The observed association of SneRING with mitochondria could be caused by the interaction with the mitochondrial surface or result from the insufficient purity of the mitochondrial fraction, which contains ER and contaminations from other cellular compartments. To analyze this further, we isolated mitochondria from the FLAG-SneRING-transfected Hek293T cells and subjected them to swelling in a hypotonic buffer in the absence or the presence of Protease K (PK). The accessibility of the proteins to PK demonstrates their submitochondrial localization. As a positive control, mitochondria were completely solubilized by 1% Triton X-100 and treated with PK (Fig 4D). Surprisingly, FLAG-tagged SneRING could be detected even after PK treatment of intact mitochondria, indicating that the portion of the protein is not localized on the mitochondrial surface. In contrast, Tom70, a mitochondrial outer membrane protein, is degraded under similar conditions, comparable to the ER protein calnexin. Opening of the outer membrane by swelling leads to a partial release of an intermembrane-space protein coproporphyrinogen oxidase (CPOX), but the inner membrane protein Mic60, the matrix protein Tim44, and SneRING remain undisturbed. PK treatment of the mitochondria with the ruptured outer membrane leads to a degradation of CPOX as well as Mic60, which is exposed to the intermembrane space, whereas Tim44 is degraded completely only after complete solubilization of mitochondria by a detergent. SneRING behaves in this respect like Mic60 since most of the protein is degraded upon PK treatment of swollen mitochondria. A small portion of the protein, however, remains undigested, indicating possible protection by the inner mitochondrial membrane (Fig 4D).

We performed carbonate extraction of isolated mitochondria using 100 mM Na₂CO₃, pH 11.5 or pH 10.8 to assess membrane integration of SneRING. In addition, the membrane association of the protein was tested by sonicating mitochondria in a high salt buffer. Hsp60, a soluble mitochondrial matrix protein, was found in both pellet and supernatant fractions after milder carbonate extraction and extraction in the high-salt buffer, but was fully extracted at pH 11.5. In contrast, Sam50, an integral protein of the outer mitochondrial membrane, was only detected in the pellet after treatment in all three conditions. Mic60, the transmembrane protein of the inner mitochondrial membrane, was partially extracted from membranes only after carbonate extraction at a higher pH of 11.5. FLAG-tagged SneRING was detected in both the pellet and the supernatant after incubation in all three conditions, suggesting that the protein is mostly membrane-associated. Its distribution resembles that of Mic60, which possesses a membrane anchor, so it would appear that a small portion of SneRING could be a membrane-integrated protein (Fig 4E), though this could also be an artifact resulting from overexpression. We conclude that a portion of SneRING localizes to the host cell mitochondria and ER after overexpression. The mitochondrial protein is most likely associated with the inner mitochondrial membrane.

Putative SneRING-ligase interacting partners include mitochondrial and ER proteins involved in the regulation of organelle morphology, protein degradation, and stress response

To determine possible host cell targets of SneRING, we expressed the FLAG-tagged protein in U2OS cells with and without Sne infection and performed a pull-down using the FLAG antibody, followed by mass spectrometry analysis. We compared infected samples with and without ectopically expressed FLAG-tagged SneRING to determine significantly enriched proteins (Fig 5A). Highly enriched proteins (log2 fold change > 2) included mitochondrial proteins involved in organelle morphology (prohibitins (PHB1 and 2), Mic60/IMMT), oxidative phosphorylation (OXPHOS) (subunits of ATPase and respiratory chain components), as well as metabolite transport (VDAC1 and 2). However, several ER proteins were also identified, and these included ER-microtubule anchoring protein CKAP4, BCAP31, a protein involved in ER-associated degradation (ERAD), and acting as an ER stress sensor that mediates communication with mitochondria [56], and Erlin1 and 2, which are prohibitin-domain containing proteins involved in ERAD [57] (Fig 5A and S1 Table).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. SneRING interacts with mitochondrial and ER proteins.

(A, B) A FLAG-tagged version of the SneRING was overexpressed in U2OS cells using PEI MAX transfection. 24 h later, cells were infected with Sne at an MOI of 1. Non-transfected cells served as a control. 48 h p.i., samples were collected and immunoprecipitation was performed using FLAG-magnetic beads, followed by mass spectrometry analysis. The graphs show identified proteins, with significance (-log₁₀p-value calculated by two-tailed T-test, n = 3) plotted against the log₂ fold change (log₂FC) of transfected/infected U2OS cells (U2OS+Sne + SneRING) relative to non-transfected/infected controls (U2OS+Sne) (A) or of transfected/infected U2OS cells (U2OS+Sne + SneRING) relative to transfected/non-infected controls (U2OS+SneRING) (B). Enriched proteins are labeled in red, reduced proteins are shown in blue, and grey represents unchanged proteins. Only host cell proteins are shown. (C) Samples were prepared, and immunoprecipitation was performed as in A. Input and elution fractions were analyzed by SDS-PAGE and western blot using antibodies against Sne heat shock protein SnGroEL (star indicates that the signal represents a cross-reactive band observed at 130 kDa), Mic60, CKAP4, Erlin2, PHB, Cox5a, Tom70, and FLAG. Mic60, mitochondrial contact site and cristae organizing system 60; CKAP4, cytoskeleton‐associated protein 4; Erlin2, ER Lipid Raft Associated 2; PHB, Prohibitin; Cox5a, cytochrome c oxidase subunit 5A; Tom70, translocase of the outer mitochondrial membrane 70.

https://doi.org/10.1371/journal.ppat.1013626.g005

When comparing the cellular SneRING-associated proteins of Sne-infected and non-infected cells, a very similar set of proteins was obtained. The top fourteen cellular infection-specific SneRING interactors were also found on the previous list. We again found mitochondrial and ER proteins among the hits. Interestingly, GOLIM4, a protein localized to the Golgi apparatus and playing a role in endosome to Golgi protein transport [58], was also found as one of the most highly enriched proteins in both lists (Fig 5B and S1 Table).

Western blot analysis of the pull-down samples confirmed the presence of Mic60/IMMT, CKAP4, and Erlin2, but PHB or Cox5a could not be detected despite being identified by mass spectrometry. Interestingly, Mic60/IMMT, CKAP4, and Erlin2 were primarily present in the pull-downs of Sne-infected samples expressing FLAG-tagged SneRING, indicating that active infection is necessary for the interactions with SneRING to take place (Fig 5C).

To further confirm that SneRING is capable of ubiquitinating some of the identified putative interactors, we performed an in vitro ubiquitination assay of CKAP4 in the presence of SneRING. In a control reaction, we used GST as an SneRING substrate. Whereas SneRING did not ubiquitinate GST (Fig 6A), we observed a generation of several high-molecular weight CKAP4 products upon incubation with SneRING (Fig 6B, asterisks). In both instances, SneRING was able to perform autoubiquitination; however, in the presence of CKAP4, this process was inhibited (Fig 6). These experiments confirm that SneRING can, at least in vitro, ubiquitinate one of the interaction partners identified in the pull-down assay.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. SneRING ubiquitinates CKAP4 in vitro.

(A, B) A reaction containing purified recombinant SneRING, recombinant E1 ubiquitin-activating enzyme, E2 ubiquitin-conjugating enzyme (UbcH5b), ubiquitin, and ATP was mixed with the respective substrate (purified recombinant GST (A) or His-CKAP4 (B)) for the indicated periods at 37 °C. Reactions without the ligase and/or substrate served as controls. Samples were analyzed by SDS-PAGE and western blot, using antibodies against SneRING, GST, ubiquitin, and CKAP4. Asterisks indicate ubiquitinated CKAP4. Ponceau staining of the western blot membranes is shown for the loading control.

https://doi.org/10.1371/journal.ppat.1013626.g006

Besides these cellular proteins, a number of Simkania-derived proteins were also strongly enriched in the SneRING-pulldowns (S5 Fig and S1 Table). These proteins might be bacterial SneRING targets or co-factors of SneRING. None of the top-enriched Sne proteins appears to be experimentally characterized, but several members of this set have informative sequence relationships. SNE_A02810, SNE_A02800, SNE_A02780, SNE_A18120, and SNE_A05090 are relatives of the Legionella major outer membrane protein mOMP, and SNE_08170 contains mOMP-associated POTRA domains. Other enriched proteins are related to components of bacterial secretion systems: A21850 is related to the SecDF secretion protein, SNE_A20110 is related to the T2SS secretin GspD, while SNE_A10070 is related to bacterial signal peptidases. Overall, the bacterial SneRING-associated proteins appear to be located at the bacterial surface.

Taken together, these results show clustering of putative cellular SneRING targets in groups of ER and mitochondrial proteins functioning in quality control, bioenergetics, and morphology. The localization of these targets is in agreement with the possible localization of SneRING that we previously determined (Fig 4). The observation that most of the prominent SneRING interactors are only present in the infected samples suggests that SneRING needs additional bacterial factors for its function, or that the infection-induced remodeling of cellular compartments fosters SneRING interactions.

Ethics statement

Primary cells used in this study originate from leukoreduction system cones made available by the Institut für Klinische Transfusionsmedizin und Hämotherapie of the University of Würzburg, Germany, for research purposes and are not subject to approval by the Ethics Committee of the University of Würzburg.

Sequence analysis

Sequence alignments were generated using the MAFFT package [88]. Generalized profiles were derived from multiple alignments using pftools [43] and searched against the Uniprot database (https://www.uniprot.org) and the NCBI microbial genome reference sequence database (https://www.ncbi.nlm.nih.gov/genome/microbes) using pfsearchV3 [89]. Structure predictions were run using a local installation of Alphafold 2.3 [44]. Structure comparisons were performed using the DALI software [45]. For modeling the missing Zn²⁺ ions into Alphafold structures, a superposition with the best-scoring Zn²⁺-containing DALI hit was used to copy the zinc ions into the model. Subsequently, an energy minimization was performed using YASARA [90] to improve the sidechain positioning of the Zn²⁺ ligands.

Cloning and mutagenesis

The SneRING coding region for protein purification was obtained by amplification of S. negevensis genomic DNA (Leibniz Institute DSMZ, German Collection of Microorganisms and Cell Cultures, DSM No. 27360) using the following primers: forward, 5’-AAGTTCTGTTTCAGGGCCCGGAAAGGGTAAATCCTAATCAAGTTCTA – 3’, and reverse, 5’-ATGGTCTAGAAAGCTTTATGAAAAAATTCTACGAGAAATATCTTCC – 3’. Amplification was performed using the Phusion High Fidelity DNA Polymerase Kit (Thermo Fisher Scientific, Massachusetts, USA), and the PCR product was inserted into the pOPIN-K vector [91] using the In-Fusion Snap Assembly Master Mix (Takara Clontech). Point mutations were generated using the QuikChange Lightning kit (Agilent Technologies, California, USA). To introduce the first mutation, the following primers were used: forward, 5’ – GCATGGAATTGTCGCAGGTGCTGGGAACCAATTCA – 3’, and reverse, 5’ – TGAATTGGTTCCCAGCACCTGCGACAATTCCATGC – 3’. The second mutation was added by using the following primers: forward, 5’ – AATTGTCGCAGGTGCTGGAGACCAATTCAGGATGGTCC – 3’, and reverse, 5’ – GGACCATCCTGAATTGGTCTCCAGCACCTGCGACAATT – 3’. SneRING coding region containing a FLAG-tag for transfection experiments was synthesized by Thermo GeneArt Gene synthesis (Thermo Fisher Scientific, Massachusetts, USA) and cloned into the pcDNA3 vector by restriction digestion according to the manufacturer’s protocol.

Protein expression and purification

Wildtype SneRING was expressed from the pOPIN-K vector containing an N-terminal 6His-GST-tag in E. coli Rosetta (DE3) pLysS. The bacteria culture was grown in LB medium at 37 °C until reaching an OD₆₀₀ of 0.8. Afterward, the culture was precooled at 18 °C before SneRING expression was induced with 500 µM ZnCl₂ and 0.1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) for 16 h when bacteria were collected by centrifugation at 5,000 x g for 15 min. The pellets were frozen at -80 °C, then thawed on ice and resuspended in binding buffer (300 mM NaCl, 20 mM Tris pH 7.5, 20 mM imidazole, 2 mM β-mercaptoethanol), containing DNase and lysozyme. Cells were sonicated for 10 min with 10 s pulses at 50 W and the lysate was clarified by centrifugation for 1 h at 50,000 x g and 4 °C. The supernatant was applied to a HisTrap FF column (Cytiva, Massachusetts, USA), and affinity purification was performed according to the manufacturer’s protocol. The subsequent incubation of the fractions with 3C protease during dialysis in binding buffer overnight resulted in the 6His-GST tag removal, which was removed together with the His-tagged 3C protease in a second affinity purification run. Size exclusion chromatography (HiLoad 16/600 Superdex 75 pg (Cytiva, Massachusetts, USA)) was performed in a buffer containing 20 mM Tris pH 7.5, 150 mM NaCl, and 2 mM dithiothreitol (DTT). After concentration using VIVASPIN 20 Columns (Sartorius, Göttingen, Germany), the purified protein was frozen in liquid nitrogen and stored at -80 °C. Due to the cloning procedure, the purified protein contained additional Gly and Pro at its amino terminus.

The expression of SneRING S19R_E22R was induced as already described for the wild-type SneRING. After harvesting the cells, the pellet was resuspended in lysis buffer (40 mM Tris pH 8.2, 500 mM NaCl, 10 mM β-mercaptoethanol, 10 mM imidazole, containing 1xcOmplete EDTA-free protease inhibitor cocktail (Roche Holding, Basel, Switzerland), DNase and lysozyme). Cells were lysed by sonication, the cleared lysate was applied to Ni-NTA agarose (Qiagen, Hilden Germany), and affinity purification was performed according to the manufacturer’s protocol. Mutated SneRING was eluted using 40 mM Tris pH 8.2, 500 mM NaCl, 10 mM β-Mercaptoethanol, and 250 mM imidazole. The removal of the 6His-GST tag (leaving additional Gly and Pro at its amino terminus) was performed similarly to the wildtype SneRING and using the Ni-NTA agarose (Qiagen, Hilden Germany) in a buffer containing 20 mM Tris pH 7.5 and 300 mM NaCl. After concentration on Amicon Ultra-15, PLGC Ultracel-PL membrane (Sigma/Merck, Darmstadt, Germany), the protein was frozen in liquid nitrogen and stored at -80 °C.

6His-GST, expressed from the pOPIN-K empty vector, and the two Sne DUBs SnVTD and SnOTU used in the study were purified as described before [28]. The plasmids pOPINB-OTUB1* (Addgene plasmid #65441; http://n2t.net/addgene:65441; RRID:Addgene_65441), pOPINB-AMSH* (Addgene plasmid #66712; http://n2t.net/addgene:66712; RRID:Addgene_66712) [92], pOPINK-Cezanne (OTU, aa 53-446), (Addgene plasmid #61581; http://n2t.net/addgene:61581; RRID:Addgene_61581) [93] and pOPINS-USP21 (USP, aa 196-565) (Addgene plasmid #61585; http://n2t.net/addgene:61585; RRID:Addgene_61585) [48] were kind gifts from David Komander (WEHI, Melbourne, Australia). Purification of these DUBs was performed as already described for the SneRING.

Protein concentration was determined by measuring the absorption at 280 nm (A₂₈₀) and using the SneRING extinction coefficient calculated from its sequence.

Recombinant human His-tagged CKAP4 was purchased from Biozol (Eching, Germany) (Cat. No. TGM-TMPK-00936).

In vitro autoubiquitination assay

10 µL of the 10x ligation buffer (800 mM Tris pH 7.5, 200 mM MgCl₂, 12 mM DTT) was mixed with 10 µM purified recombinant SneRING, 100 nM recombinant E1 ubiquitin-activating enzyme, and 2.3 µM recombinant E2 ubiquitin-conjugating enzyme, to which 50 µM recombinant ubiquitin and 10 mM ATP were added. ddH₂O was added to a final volume of 100 µL. The reactions were incubated at 37 °C for different time points (0 min, 30 min, 60 min, 120 min, 180 min, and 300 min) and stopped by adding Laemmli sample buffer (62.5 mM Tris pH 6.8, 2% SDS, 10% glycerol, 5% β-mercaptoethanol, and 0.002% Bromophenol Blue) and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and western blot.

The purified proteins of the E2 conjugating enzymes UbcH5b, UBE2T, UbcH7, UBE2A, UBE2B, and UBE2W were a kind gift from David Komander (WEHI, Melbourne, Australia).

In vitro substrate ubiquitination assay

MasterMix was prepared, containing 80 mM Tris pH 7.5, 20 mM MgCl₂, 1 mM DTT, 0.1 mM mouse E1, 2.3 µM UbcH5b, 5 µM SneRING, 1–5 µM substrate (recombinant His-CKAP4 or recombinant GST), 50 µM ubiquitin, and 10 mM ATP. Reaction was incubated at 37 °C. Per timepoint, reaction was stopped by combining 10 µL MasterMix with 10 µL 2x Laemmli buffer. 10 µL of the stopped sample were analyzed by SDS-PAGE and western blot.

UbiCRest assay

The recombinant DUBs were first diluted in a buffer containing 150 mM NaCl, 20 mM Tris pH 7.5, and 10 mM DTT. The 5 h in vitro autoubiquitination assay reaction was treated for 15 min at RT with 2 mU apyrase to remove the remaining ATP. For the assay, the DUBs were added at a 1.5 µM concentration except OTUB1, for which a 3 µM concentration was used. During the double UbiCRest assay, the DUBs were added at 0.75 µM or 1.5 µM concentrations. The mixture was incubated for 1 h at 37 °C, and the reaction was stopped by adding Laemmli sample buffer.

Immunoblotting

The lysates for western blot analysis were prepared with Laemmli sample buffer and denatured at 95 °C for 5 min. After SDS-PAGE, proteins were transferred onto a PVDF membrane. The membrane was blocked with 5% milk in Tris-based saline (TBS) and incubated with the primary antibody, followed by incubation with an HRP-coupled secondary antibody (1:3000) in blocking solution. The signal was detected using SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Fisher Scientific, Massachusetts, USA) and a Chemo Cam Imager (Intas). In the case of a K6 affimer, the membrane was incubated overnight at 4 °C in the affimer dilution in blocking solution. The next day, the membrane was washed with TBS and incubated with a 6xHis primary antibody, followed by the HRP-coupled secondary antibody as already described.

Cell culture

HeLa (ATCC CCL-2.1), HeLa229 KDEL-dsRed, Hela 229 Mito-GFP, and THP-1 (ATCC TIB-202) cells were grown in Roswell Park Memorial Institute (RPMI) 1640 medium (Thermo Fisher Scientific, Massachusetts, USA), Hek293T cells (ATCC CRL-3216) were grown in Dulbecco’s modified Eagle’s medium (DMEM) medium (Thermo Fisher Scientific, Massachusetts, USA) and U2OS cells (ATCC HTB-9.6) were grown in DMEM/F12 medium (Thermo Fisher Scientific, Massachusetts, USA). All media were supplemented with 10% v/v heat-inactivated (56 °C for 30 min) fetal calf serum (FCS) (Sigma/Merck, Darmstadt, Germany). Growth took place at 37 °C and 5% CO₂. To differentiate THP-1 cells into macrophages, 5x 10⁵ cells/well were seeded into 6-well plates and incubated with 20 ng/mL phorbol 12-myristate 13-acetate (PMA) (Sigma/Merck, Darmstadt, Germany) in RPMI1640 medium containing 10% v/v heat-inactivated FCS for 24 h. The medium was exchanged for fresh RPMI1640 medium containing 10% v/v heat-inactivated FCS and the cells were incubated for a further 48 h.

Primary human macrophages were derived from peripheral blood mononuclear cells (PBMCs) isolated from leukoreduction system (LRS) cones, using the SepMate-50 system (#85450, STEMCELL Technologies, Cologne, Germany) and Ficoll-Paque (#GE17-1440-03, Sigma/Merck, Darmstadt, Germany) gradient according to manufacturer’s instructions. Red blood cells were removed by hypo-osmotic shock: PBMC pellets were mixed and incubated with 16 mL sterile water for 30 s, then immediately 4 mL 5 x DPBS were added. Monocytes were purified from the PBMC fraction using the EasySep Human CD14 Positive Selection Kit II (# 17858, STEMCELL Technologies, Cologne, Germany) according to the manufacturer’s instructions and seeded in 6-well cell culture plates (1.5 x 106 cells/well) in medium supplemented with 50 ng/mL M-CSF (#78057, STEMCELL Technologies, Cologne, Germany). Medium exchanges, with fresh M-CSF supplementation, were done on days 1 and 4. On day 7, “M2”-like polarization, was induced by replacing M-CSF-containing medium with medium containing 100 ng/mL IL-4 (#78045, STEMCELL Technologies, Cologne, Germany) and further incubation for 48 h. All cultivation steps were done in RPMI1640 (Thermo Fisher Scientific, Massachusetts, USA) supplemented with 10% (v/v) heat-inactivated FCS (Sigma/Merck Darmstadt, Germany). All Sne infection experiments were performed on day 9, without cytokine supplementation.

Acanthamoeba castellanii (ATCC 30010D) were cultured in axenic medium (56 mM glucose, 2.6 mM KH2PO4, 1.3 mM Na2HPO4 x 7H2O, 0.5% proteose peptone 2, 0.5% thiotone E peptone, 0.5% yeast extract) in cell culture flasks at 30 °C.

Bacteria and infection

E. coli DH5α and Rosetta (DE3) pLysS were grown in an LB medium containing the respective antibiotics for selection (1% bacto-tryptone, 0.5% yeast extract, 1% NaCl, 1% agar).

To prepare S. negevensis (ATCC VR-1471), U2OS cells were grown to 60% confluency, and infected at a multiplicity of infection (MOI) 1 in RPMI1640 medium without HEPES containing 5% v/v heat-inactivated FCS (infection medium). The infected cells were incubated at 35 °C and 5% CO₂ for 6 h, when the medium was exchanged for fresh infection medium. After 3 days of growth under the same conditions, the cells were detached using a cell scraper and lysed by vortexing the cells with 2–5 mm glass beads (Carl Roth, Karlsruhe, Germany). Intact cells and larger cell debris were removed by low-speed centrifugation at 600 x g for 10 min and the resulting supernatant was centrifuged at 20,000 x g for 30 min to pellet the released bacteria. The pellet was resuspended in 5 mL SPG buffer (250 mM sucrose, 4 mM monopotassium phosphate, 10 mM disodium phosphate, and 5 mM glutamate pH 7.4), followed by centrifugation (20,000 x g, 30 min), with the resulting pellet resuspended in 2.5 mL SPG buffer. The bacteria were homogenized using a syringe with one G20 and one G18 needle, aliquoted, and stored at -80 °C. All the experiments with S. negevensis were performed in a laboratory with biosafety level 2, which is registered with the Lower Franconia Government under code 55.1-8791.1.30.

For infection experiments, 8 x 10⁴ or 5 x 10⁵ human cells/well were seeded into 12-well or 6 well plates, respectively, and at 60% confluency infected with S. negevensis at MOI 1 in the infection medium, at 35 °C and 5% CO₂. 6 h later, the medium was exchanged for fresh infection medium. The infection proceeded for the indicated time points. When the infection took longer than 3 days, the medium was exchanged for fresh infection medium after every 3 days.

A. castellanii were seeded into cell culture flasks and infected with S. negevensis in an axenic medium. Infected amoebae were incubated at 30 °C for 4 days.

Transfection

Hek293T cells were transfected using the calcium-phosphate transfection method as described previously [94]. For 6-well plates, 4 µg of the respective plasmid were used; for transfecting cells in 15 cm dishes, we used 50 µg of plasmid DNA. U2OS and HeLa229 cells were transfected using the commercially available transfection reagents jetOptimus (Polyplus, Illkirch-Graffenstaden, France) or PEI MAX (Polysciences, Inc., Warrington, United Kingdom) according to the manufacturer’s protocol. For 12-well plates, 1 µg (jetOptimus) or 1.5 µg (PEI MAX) of plasmid per well was used. For 10 cm dishes, initially seeded with 1.2x10⁶ cells, we used 10 µg of plasmid (PEI MAX). Depending on the experiment, cells were infected with S. negevensis at MOI 1, 24 h post-transfection, and the infection was allowed to proceed for 2 days before sample collection.

Immunofluorescence and Microscopy

U2OS and HeLa cell lines were seeded at a density of 8 x 10⁴ cells/well into 12-well plates on 15 mm glass coverslips (Paul Marienfeld, Lauda-Königshofen, Germany). Cells were transfected and/or infected as indicated in the corresponding experiments. The cells were fixed at respective time points using 4% paraformaldehyde (PFA) in PBS for 1 h at RT. For S. negevensis infection experiments, the cells were permeabilized with 0.2% Triton X-100 in PBS for 45 min at RT while shaking, washed with PBS, and blocked in 2% goat serum in PBS for 1 h, followed by staining with a SnGroEL and FLAG primary antibodies in the blocking solution overnight at 4 °C. The following day, samples were washed and incubated with the mix of the respective secondary antibodies (1:100) and DAPI (1:3000) for 1 h in 2% goat serum in PBS at RT. After washing with PBS and ddH₂O, coverslips were mounted using Mowiol 4–88 (Carl Roth GmbH + Co. KG, Karlsruhe, Germany).

For immunofluorescence stainings that did not include staining of bacterial proteins, after fixing, cells were simultaneously permeabilized and blocked for 30 min at RT in blocking solution (10% BSA, 0.2% Triton X-100, in PBS), washed with PBS and stained with the primary antibody diluted in antibody dilution solution (3% BSA in 0.05% Tween in PBS) for 1 h at RT. After washing with PBS and blocking for 10 min in blocking buffer at RT, samples were washed again and incubated with DAPI (1:3000) and the secondary antibodies (1:100) for 1 h at RT. The coverslips were then washed with PBS and ddH₂O and mounted using Mowiol 4–88 (Carl Roth GmbH + Co. KG, Karlsruhe, Germany). Microscopy was performed using a LEICA SP5 confocal laser scanning microscope.

RNA isolation, cDNA synthesis, and RT-qPCR

Infected macrophages were collected on days 2, 4, and 6 p.i. The cells were washed with DPBS at the desired time point, and 100 µL of fresh medium and 500 µL RNAprotect Cell Reagent (Qiagen, Hilden, Germany) were added per well. Detached cells were collected in RNase-free tubes and stored at 4 °C for a maximum of 1 week. After centrifugation at 6200 x g for 5 min at 4 °C, 600 µL RLT Buffer supplemented with β-mercaptoethanol was added to each pelleted sample, and RNA was purified using the RNeasy Mini kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. For other human cell lines and amoeba, which were collected on day 2 and 4 p.i., or just on day 4 p.i., respectively, and washed with DPBS at the desired time point, 600 µL RLT Buffer supplemented with β-mercaptoethanol was added to each sample and the RNA was purified using the RNeasy Mini (Qiagen, Hilden, Germany) according to manufacturer’s instructions. DNase treatment was performed using the TURBO DNA-free Kit (Invitrogen, California, USA). 100 ng of pure RNA was used for cDNA generation, using the Revert Aid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Massachusetts, USA). RT-qPCR was performed on a StepOnePlus instrument (Life Technologies, California, USA), using the GreenMasterMix (2X) High ROX (Gennaxon, Ulm, Germany). Data analysis was performed using the StepOne Software (v2.3) and Microsoft Excel, with a modified (2 ^-ΔΔCt) method [95]. ΔCt values were calculated relative to the host reference gene YWHAZ (human) or the 5S RNA (for amoeba). Fold changes (2^-ΔΔCt) of gene expression for each tested Sne gene, were calculated relative to the non-infected controls. The following primers were used: forward, 5’ – ATACTGCGCAGAAAGCAC – 3’, and reverse, 5’ – ACCCCAGTACTAACACCG – 3’ (5S RNA reference gene for amoeba) [28]; forward, 5’ – GCAGAGTCGAAGTTTCAAGGG – 3’, and reverse, 5’ – TCGCTTGCCTGTTGAGGATT – 3’ (SneRING); forward, 5’ – TTCCATCACCGGCAACATCA – 3’, and reverse, 5’ – GCAAAAGAGATCGCGCTGAA – 3’ (GroEL C (SNE_A22590)) [28]; forward, 5’ – CACCTGATCCCATCCCGAAC – 3’, and reverse, 5’ – GCGACCTACTCTCCCGTACT – 3’ (Sne 5S RNA) [28]; forward, 5’ – ACTTTTGGTACATTGTGGCTTCAA – 3’, and reverse, 5’ – CCGCCAGGACAAACCAGTAT – 3’(YWHAZ) [96].

Mitochondrial isolation and fractionation

Mitochondria were isolated from cells transfected for 48 h as described previously [97]. Cellular fractionation and determination of submitochondrial localization by swelling and PK treatment were performed as already described [98,99]. Carbonate extraction for determining protein membrane extractability was performed as previously described [100]. In brief, isolated mitochondria from transfected cells were either sonicated in 10 mM Tris pH 7.6 and 500 mM NaCl or treated with 100 mM Na₂CO₃ at pH 10.8 or 11.5 for 20 min on ice. After centrifugation at 100.000 g, pellets were resuspended in Laemmli sample buffer, while supernatant and non-centrifuged control samples were precipitated using trichloroacetic acid (TCA).

Antibodies, chemicals, and controls

K6-, K48-, and K63-linked ubiquitin chains used as a control in S2 Fig were components of the Di-ubiquitin Explorer Panel purchased from UbiQ (Amsterdam, Netherlands, Cat. No. UbiQ-L01). M1-linked ubiquitin chains (kindly provided by S. Meyer and A. Buchberger) were produced by incubating the following reaction mixture at 37 °C for 2 h: 6 µg bovine ubiquitin, 200 ng yeast Uba1, 600 ng human UbcH5c, 4 µg HOIP RBR-LDD [101] in a buffer containing 50 µL 20 mM Tris/HCl pH 7.5, 5 mM MgCl₂, 1 mM DTT, and 4 mM ATP.

Primary antibodies against ubiquitin (Cat. No. 05-944), FLAG-tag (Cat. No. F7425; Cat. No. F3165), CPOX (Cat. No. HPA015736), K63-linked ubiquitin chains (Cat. No. 05-1308), K48-linked ubiquitin chains (Cat. No. 05-1307), M1-linked ubiquitin chains (Cat. No. MABS451, kindly provided by A. Buchberger), and the anti-ubiquitin Lys6 specific Affimer (Cat. No. MABS1918, kindly provided by M. Eilers) were purchased from Sigma/Merck (Darmstadt, Germany). Primary antibodies against Tubulin (Cat. No. ab18251) and Mic60 (Cat. No. ab48139) were purchased from Abcam (Cambridge, United Kingdom). Primary antibodies against Calnexin (Cat. No. 2433 and Cat. No. sc-23954) were purchased from Cell Signaling (Massachusetts, USA) and Santa Cruz Biotechnology (Texas, USA), respectively. The Erlin2 antibody (Cat. No. 2959) was purchased from Cell Signaling (Massachusetts, USA). The Prohibitin antibody (Cat. No. sc-377037) and the antibody against GAPDH (Cat. No. sc-47724) were purchased from Santa Cruz Biotechnology (Texas, USA). The anti-Hsp60 antibody (Cat. No. ADI-SPA-806) was purchased from Enzo Life Sciences (New York, USA). The primary antibody against Tim44 (Cat. No. 612582) was purchased from BD Transduction Laboratories (California, USA). The antibody against 6xHis (Cat. No. GTX18184) was purchased from Biozol (Eching, Germany). The antibody against GST (Cat. No. MA4-004) was purchased from Invitrogen (Thermo Fisher Scientific, Massachusetts, USA). The antibody against K11-linked ubiquitin chains (Cat. No. NBP3-05681, kindly provided by A. Buchberger,) and the CoxVa antibody (Cat. No. NBPI-32550) were purchased from Novus Biologicals (Littleton, USA). The CKAP4 antibody (Cat. No. 16686-1-AP) was purchased from Proteintech (Rosemont, USA). The antibody against Sam50 was custom-generated against the full-length protein (Gramsch laboratories, Schwabhausen, Germany). The antibody against rat Tom70 (crossreactive with human Tom70) was a gift from K. Mihara. The antibody against SnGroEL was produced as described in a previous study [22]. The SneRING antibody was custom-made against the 189 carboxy-terminal amino acids of the protein (Davids Biotechnologie GmbH, Regensburg, Germany). Secondary antibodies anti-rabbit-Alexa555, Cat. No. A-21428, anti-mouse-Alexa488 Cat. No. A-11001, and anti-mouse-Alexa555, Cat. No. A-21422 were purchased from Thermo Fisher Scientific (Massachusetts, USA). Anti-rabbit-Atto647N secondary antibody, Cat No. 40839, was purchased from Sigma/Merck (Darmstadt, Germany). The HRP-coupled secondary antibodies for western blot were purchased from Biozol (Eching, Germany).

Immunoprecipitation

For immunoprecipitation (IP) assays, U2OS cells were seeded at a density of 1.2 x 10⁶ cells/well into 10 cm dishes. The next day, the cells were transiently transfected with 10 µg of plasmid carrying a FLAG-tagged SneRING gene and 24 h later infected with Sne at an MOI 1. Non-transfected and non-infected cells served as controls. On day 2 p.i., the cells were washed once with PBS, harvested in lysis buffer (20 mM Tris pH 7.5, 150 mM NaCl, 0.1% NP-40, proteinase inhibitor (Roche: cOmplete, EDTA-free Protease Inhibitor Cocktail), 1 mM PMSF, 5 mM Iodoacetamide)), incubated for 30 min at 4 °C and sonicated for 15 min using an ultrasonic water bath, followed by a 30 min incubation on a rotary shaker at 4 °C. After removing the cell debris by centrifugation (15 min, 14000 x g, 4 °C), the supernatant was added to equilibrated Anti-FLAG M2 magnetic beads (Sigma/Merck, Darmstadt, Germany) and incubated on a rotary shaker overnight at 4 °C. The beads were washed three times for 5 min at 4 °C with washing buffer (20 mM Tris pH 7.5, 150 mM NaCl, 0.1% NP-40) and once with a buffer containing 20 mM Tris pH 7.5, 150 mM NaCl. For mass spectrometry analysis, bound proteins were eluted with 200 µg/mL FLAG peptide (Sigma/Merck, Darmstadt, Germany) in elution buffer (6 M urea, 2 M thiourea).

Mass spectrometry analysis

For in-solution digest, the samples were reduced by adding 1 µL of 1 M DTT for 1 h at RT, alkylated with 1 µL of 400 mM chloroacetamide (CAA) for 45 min at RT in darkness, digested with 1 µL of the Lysyl endopeptidase (Lys-C, 0.5 µg/µL) for 3 h at RT, treated with 270 µL of 50 mM triethylammonium bicarbonate buffer and finally digested with 1 µL of trypsin enzyme (1 µg/µL) overnight at RT. The next day, samples were acidified by adding 1 µL of 100% formic acid. After in-solution digestion, peptides were bound to styrene-divinylbenzene reverse-phase stage tips (SDB-RP). One stage tip was used per sample. The stage tips were equilibrated by adding 20 µL of 100% methanol and centrifuging for 1 min at 700 x g. Afterward, 20 µL of buffer B (0.1% formic acid in 80% acetonitrile) was added, and the stage tips were again centrifuged for 1 min at 700 x g. In the next step, 20 µL of buffer A (0.1% formic acid in water) was added to each stage tip, the samples were centrifuged for 1 min at 700 x g, and an additional 20 µL of buffer A was added. Acidified samples were spun down for 5 min at 17,000 x g, the supernatant was loaded onto the SDB discs of the equilibrated stage tips, and they were again centrifuged for 5 min at 700 x g. The stage tips were washed with buffers A and B and completely dried with a syringe.

After binding peptides to SDB-RP Stage tips, they were analyzed by liquid chromatography-mass spectrometry/mass spectrometry (LC-MS/MS). LC-MS/MS analysis was performed by the Proteomics Facility of the CECAD, University of Cologne, as described previously [102]. Briefly, for LC-MS/MS analysis, an EASY-nLC 1000 chromatograph (Thermo Scientific) was coupled to the quadrupole-based Q Exactive Plus (Thermo Scientific) instrument by a nano-spray ionization source. Peptides were separated on a 50 cm in-house-packed column using a two-solvent buffer system: buffer A (0.1% formic acid) and buffer B (0.1% formic acid in acetonitrile). The amount of buffer B was increased from 7% to 23% within 40 min, followed by an increase to 45% in 5 min, and a washing and re-equilibration step before the next sample injection. The mass spectrometer operated in a Top 10 data-dependent mode, using the following settings: MS1: 70,000 (at 200 m/z) resolution, 3e6 AGC target, 20 ms maximum injection time, 300–1750 scan range; MS2: 35,000 (at 200 m/z) resolution, 5e5 AGC target, 120 ms maximum injection time, 1.8 Th isolation window, 25 normalized collision energy. Data analysis was performed by MaxQuant software, V 1.5.4.742, using the Andromeda search engine against the human proteome reference data (including splice variants) from UniProt. The default mass tolerance and modification settings were used. Re-quantify, label-free quantification, and match between runs were enabled. ‘Oxidation on M’, ‘Phosphorylation on S,T,Y’, and ‘GlyGly on K’ were allowed as variable modifications. Intensities were averaged over biological triplicates, and the log2 of the intensity ratio ‘sample average/control average’ was used for enrichment quantification. To account for missing values, pseudo counts corresponding to the minimal observed intensity were added to the sample and control averages. Statistical analysis was performed using a two-tailed Student’s T-test (n = 3).