L. pneumophila targets mitochondria using T4SS effectors with tail anchors

To identify L. pneumophila T4SS effectors with characteristics of mitochondrial TA proteins, we conducted a bioinformatic screen of 302 experimentally confirmed or putative L. pneumophila T4SS effector proteins (Fig 1A and S1 Text) [4,35,36]. Using TMHMM and TOPCONS, 22 effectors were predicted to contain a single TMD [37,38]. Of these, five had a TMD within 30 residues of the stop codon, and four (Lpg1625, Lpg1803, Lpg2344, and Lpg2444) exhibited a positively charged C-terminal tail, a hallmark of mitochondrial TA proteins (Fig 1B and S1 Table) [33]. To validate the mitochondrial localization of the four predicted TA proteins, N-terminally FLAG-tagged effectors were transiently expressed in HeLa cells and analyzed by immunofluorescence microscopy. All four proteins showed colocalization with the mitochondrial marker NDUFAF2, confirming mitochondrial localization (Fig 1C). Lpg1625 and Lpg2444 have been shown to localize to mitochondria previously, with Lpg1625 suggested to induce caspase 3 activation [39], and Lpg2444 suggested to prevent mitochondrial fragmentation [17]. However, detailed functional characterization of these effectors has not been undertaken. Lpg2344 was reported to have a vesicular localization [40], and Lpg1803 has not been functionally characterized. In this study, we selected Lpg1625 as the first candidate for further functional analysis.

Download:

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

Fig 1. L. pneumophila targets mitochondria using T4SS effectors with tail anchors (TA).

(A) Schematic representation of the bioinformatics workflow used to screen for L. pneumophila T4SS effectors (see S1 Text) that contain a TA with net positive tail charges (see S1 Table). (B) Features of the four effectors predicted to contain a TA with a net positive charge in their C-tail. TMHMM and TOPCONS were used to predict the number and location of transmembrane domains (TMDs) present within the 302 effectors screened. A black dot indicates the prediction of one TMD, while a white dot indicates no predicted TMD. The length of the predicted TMD and C-tail, along with the net C-tail charge, is indicated. (C) Representative images of HeLa cells transiently transfected with ^FLAGLpg1625, ^FLAGLpg1803, ^FLAGLpg2344, or ^FLAGLpg2444. Transfected cells were fixed and processed for immunofluorescence with antibodies against FLAG (green) and NDUFAF2 (mitochondria; magenta). Nuclei were stained with Hoechst 33258 (blue). Scale bar represents 10 μm. (D) Mitochondrial sub-fractionation and carbonate extraction using mitochondria isolated from HeLa cells transiently transfected with ^FLAGLpg1625. Mitochondria were either left intact (lanes 1 and 2), subjected to hypotonic swelling (lanes 3 and 4), or solubilized with 0.5% Triton X-100 (lanes 5 and 6). Samples treated with Proteinase K (PK) are indicated (+). Mitochondria were also subjected to sodium carbonate extraction, with pellet (P) and supernatant (S) fractions (lanes 7 and 8) separated via ultracentrifugation. Samples were analyzed using SDS-PAGE and immunoblotting using the indicated antibodies. Corresponding raw images are available in the Supporting information (S1 Raw Images). (E) HeLa cells transiently transfected with ^FLAGLpg1625, ^FLAGLpg1625^1-108, ^FLAGLpg1625^85-130, ^FLAGLpg1625^K128A, and ^FLAGLpg1625^AAA (K97A/R100A/K128A) were processed for immunofluorescence with antibodies against FLAG (green) and NDUFAF2 (mitochondria; magenta). Nuclei were stained with Hoechst 33258 (blue). Scale bar represents 10 μm.

https://doi.org/10.1371/journal.pbio.3003261.g001

Lpg1625 is a mitochondrial tail-anchored (TA) effector

Deletion of lpg1625 did not influence the ability of L. pneumophila to replicate in THP-1 macrophage-like cells (S1 Fig and S2 Table). However, it may contribute to the intracellular success of L. pneumophila under specific circumstances and/or in a manner that is functionally redundant with other T4SS effectors [6]. We ectopically expressed ^FLAGLpg1625 in HeLa or HEK cells to facilitate downstream biochemical characterization. First, we confirmed that ^FLAGLpg1625 localizes exclusively to mitochondria and not to other organelles typically targeted by TA proteins (S2A–S2D Fig) [32]. We then biochemically validated its localization to the mitochondrial outer membrane using sub-fractionation analysis and carbonate extraction (Fig 1D). Mitochondria isolated from HeLa cells transfected with ^FLAGLpg1625 were either left untreated, subjected to hypo-osmotic swelling, or solubilized in Triton X-100, followed by treatment with Proteinase K (PK) (Fig 1D). ^FLAGLpg1625 was accessible to external protease in intact mitochondria, mirroring the outer membrane marker Mfn2, consistent with an outer membrane localization (Fig 1D, lanes 1–2). Sodium carbonate extraction recovered ^FLAGLpg1625 in the pellet fraction, suggesting the protein is membrane integrated (Fig 1D, lanes 7–8). Based on this profile, we concluded Lpg1625 is integrated into the mitochondrial outer membrane via a C-terminal TMD.

To confirm the TA characteristics of Lpg1625, we generated truncation and charge mutants and examined their subcellular localization by immunofluorescence analysis (Fig 1E). Constructs included: ^FLAGLpg1625^1-108 (cytosolic domain), ^FLAGLpg1625^85-130 (TMD and C-tail), ^FLAGLpg1625^K128A (single charge mutant), and ^FLAGLpg1625^AAA (K97A/R100A/K128A triple charge mutant) (Fig 1E). These experiments demonstrated that the TMD and C-tail of Lpg1625 are sufficient for mitochondrial targeting of Lpg1625 (Fig 1E). The K128A and K97A/R100A/K128A mutants, designed to remove positive charge flanking the TMD, showed mislocalization to the endoplasmic reticulum (ER) (Figs 1E and S2E), consistent with previous findings [41]. We confirmed ^FLAGLpg1625^85-130 (TMD and C-tail) was also sufficient for membrane integration using carbonate extraction (S2F Fig). Taken together, this suggests that the L. pneumophila effector repertoire contains TA proteins, and we have biochemically confirmed that Lpg1625 is a bona fide mitochondrial outer membrane TA effector protein.

Lpg1625 interacts with protein phosphatase 1 (PP1)

We sought to elucidate the functional role of Lpg1625 by identifying its interacting partners using co-immunoprecipitation and mass spectrometry. This revealed the three isoforms of PP1 (PPP1CA, PPP1CB, and PPP1CC), as the top interactors of Lpg1625, detected with 4, 7, and 5 unique peptides, respectively (Fig 2A and S3 Table). This suggests a novel and functional role for all three PP1 isoforms in conjunction with Lpg1625 at the mitochondrial surface. Additional co-immunoprecipitated proteins included MTCH1 and MTCH2 which may facilitate membrane integration of Lpg1625 [31,42]. The VDAC proteins were also co-immunoprecipitated, likely due to their abundance and physical proximity to Lpg1625 in the mitochondrial outer membrane (Fig 2A). To rule out potential artifacts, we generated a stable tetracycline-inducible Flp-In T-REx 293 cell line expressing ^FLAGLpg1625, allowing regulation of expression to low levels (S3A Fig). Sub-fractionation analysis of mitochondria isolated from these cells confirmed the mitochondrial outer membrane profile of ^FLAGLpg1625 (S3B Fig). Co-immunoprecipitation using this stable cell line confirmed PP1 as the primary interactor of Lpg1625, even at low expression levels (S3C Fig and S4 Table).

Download:

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

Fig 2. Lpg1625 recruits PP1 to mitochondria via an RVxF motif.

(A) Mitochondria were isolated from wild-type HeLa cells or cells transiently transfected with ^FLAGLpg1625 and solubilized in 1% digitonin prior to immunoprecipitation with anti-FLAG conjugated agarose beads. Eluates were analyzed by mass spectrometry. The Log₂ fold change of the mean LFQ intensity is plotted against −Log₁₀ p-value (n = 3 technical replicates). The curve indicates significantly enriched proteins (FDR = 0.05, s0 = 1). Mitochondrial outer membrane proteins are labeled in orange with the annotation from MitoCarta 3.0. Corresponding data are available in the Supporting information (S3 Table). (B) AlphaFold [43] was used to predict the structure of Lpg1625, which was labeled with the RVxF motif and transmembrane domain (TMD) using ChimeraX [44]. The model has an average per-residue model confidence score (pLDDT) of 65.94, indicating low model confidence. (C) Mitochondria sub-fractionation and carbonate extraction were performed using mitochondria isolated from HeLa cells transiently transfected with ^FLAGRVxF_dead. Mitochondria were either left intact (lanes 1 and 2), subjected to hypotonic swelling (lanes 3 and 4), or solubilized with 0.5% Triton X-100 (lanes 5 and 6). Samples treated with Proteinase K (PK) are indicated (+). For carbonate extraction, mitochondria were washed in a fresh solution of 0.1 M sodium carbonate and integral (pellet, P) and soluble (supernatant, S) proteins were separated using ultracentrifugation. Samples were analyzed using SDS-PAGE and immunoblotting using the indicated antibodies. Corresponding raw images are available in the Supporting information (S1 Raw Images). (D) Mitochondria isolated from wild-type HeLa cells or cells transiently transfected with ^FLAGRVxF_dead were solubilized in 1% digitonin buffer prior to immunoprecipitation with anti-FLAG conjugated agarose beads. Eluates were processed for mass spectrometry. The Log₂ fold change of mean LFQ intensity is plotted against −Log₁₀ p-value (n = 3 technical replicates). The curve indicates significantly enriched proteins (FDR = 0.05, s0 = 1). Mitochondrial outer membrane proteins are labeled in orange with the annotation from MitoCarta3.0. Corresponding data are available in the Supporting Information (S3 Table). (E) HeLa cells expressing ^FLAGLpPIP, or ^FLAGRVxF_dead were stained with 100 nM MitoTracker Deep Red FM (mitochondria; MTDR) and, following fixation, processed for immunofluorescence using antibodies against FLAG (green) and PPP1CB (magenta). Nuclei were stained with Hoechst 33258 (blue). Scale bar represents 10 μm. Fluorescence signal profiles of FLAG and PPP1CB were measured across the indicated line of interest. Corresponding raw data are available in the Supporting information (S1 Data). (F) The correlation between PPP1CB and ^FLAGLpPIP was assessed using the Pearson correlation coefficient (r) in HeLa cells expressing ^FLAGLpPIP or ^FLAGRVxF_dead (n = 3 independent experiments, 45 cells in total). Data represent mean ± SEM of the three experiments represented in three different colors. Larger circles represent the mean values from each experiment, while smaller circles are values from individual cells. An unpaired t test on the means revealed a significant difference of 0.59 ± 0.02 between groups (p-value < 0.0001). Corresponding raw data are available in the Supporting information (S1 Data). (G) Proportion of PPP1CB localized to mitochondria in untransfected and HeLa cells transiently transfected with ^FLAGLpPIP or ^FLAGRVxF_dead, was calculated using raw integrated density in Fiji. Representative images are shown in Fig 2E. Data are presented as mean ± SEM (n = 3 independent experiments, 5 cells each). Ordinary one-way ANOVA showed no significant difference between the groups. Corresponding raw data are available in the Supporting information (S1 Data). (H) AlphaFold3 predicted structure of LpPIP-PPP1CB complex (ipTM = 0.86; pTM = 0.85) labeled with the RVxF motif on LpPIP (green) and known RVxF binding pocket on PPP1CB (red). Image generated using ChimeraX [44]. (I) AlphaFold3 structure of LpPIP-PPP1CB complex labeled with the active site (purple) and three substrate binding grooves (yellow) of PPP1CB.

https://doi.org/10.1371/journal.pbio.3003261.g002

PP1 is a ubiquitously expressed serine/threonine (Ser/Thr) phosphatase, which has a broad substrate specificity and is responsible for a significant proportion of dephosphorylation events in eukaryotic cells [45]. It contains three independent catalytic subunits encoded by three different genes, namely, PPP1CA, PPP1CB, and PPP1CC [46], with 85% sequence identity among them. These proteins contain highly conserved catalytic domains but diverge at the N- and C-termini [46,47]. PP1 functions as a holoenzyme through interaction with PIPs, a diverse group of regulatory proteins that modulate PP1 activity, substrate specificity, and subcellular localization [45,48]. We hypothesized that Lpg1625 acts as a PIP by interacting with all three catalytic subunits of PP1 and recruiting them to the mitochondrial outer membrane, thereby modulating the host mitochondrial phosphoproteome.

Lpg1625 recruits PP1 to mitochondria via its RVxF motif

PIPs bind PP1 typically through short linear motifs such as RVxF, which is found in approximately 70% of known PIPs [45,49,50]. Examining the Lpg1625 protein sequence revealed the presence of a RVxF motif within its N-terminal cytosolic domain (amino acids 17–21), with two lysine residues upstream of the motif (at position 15 and 16), which can increase affinity for PP1 (Fig 2B) [51]. To investigate the functional importance of this motif in the interaction between Lpg1625 and PP1, we introduced alanine substitutions at two key residues of the RVxF motif of Lpg1625 (K17A and V19A) to create the mutant construct ^FLAGRVxF_dead [51]. While ^FLAGRVxF_dead localized correctly to the mitochondrial outer membrane in both transient (Fig 2C) and stable expression systems (S3D Fig), co-immunoprecipitation experiments demonstrated a complete loss of interaction with PP1 (Figs 2D and S3E). These data confirm that Lpg1625 is functioning as a PIP at the mitochondrial outer membrane for the recruitment of PP1 via the RVxF motif, which is likely the only PP1-binding motif on Lpg1625. Given this, we henceforth refer to Lpg1625 as LpPIP.

Under basal conditions, the PP1 catalytic subunits PPP1CA, PPP1CB, and PPP1CC are distributed throughout the cytosol and nucleus [46]. Their distribution is dynamically regulated by various PIPs, cellular events, and cell cycle stages [46]. Since the RVxF motif is commonly used as an anchor for PP1 recruitment, we assessed the localization of the PP1 catalytic subunit, PPP1CB in HeLa cells expressing either ^FLAGLpPIP or ^FLAGRVxF_dead using immunofluorescence (Fig 2E) [47,52]. Untransfected HeLa cells showed cytosolic and nuclear localization of PPP1CB (Fig 2E, top panel). In contrast, cells expressing ^FLAGLpPIP showed a proportion of PPP1CB now localized to mitochondria (Fig 2E, middle panel), whereas PPP1CB localization remained cytosolic/nuclear in cells expressing ^FLAGRVxF_dead (Fig 2E, bottom panel). This supports a critical role for the Lpg1625 RVxF motif in PP1 recruitment to mitochondria. Pearson’s correlation analysis quantitatively demonstrated a significant positive colocalization between ^FLAGLpPIP and PPP1CB, whereas ^FLAGRVxF_dead showed no detectable colocalization (Fig 2F). Using Fiji quantification, we calculated the proportion of PPP1CB at mitochondria in HeLa cells expressing ^FLAGLpPIP, and found that approximately 17% of the protein is at mitochondria (Fig 2G).

To gain structural insight into the interaction, AlphaFold 3 was used to predict the structure of the LpPIP-PPP1CB holoenzyme [53]. The high-quality prediction revealed the RVxF motif of LpPIP is in contact with the known RVxF binding pocket on PPP1CB (Fig 2H) [47,52]. The model suggests that the PPP1CB active site, which is located at the intersection of the potential substrate binding grooves, remains accessible in the complex (Fig 2I) [54]. This suggests that PP1 retains its catalytic activity when bound to LpPIP, although potential allosteric effects cannot be excluded. Collectively, these findings show that a Legionella-specific PIP effector, LpPIP, is localized to the mitochondrial outer membrane to recruit PP1 to the organelle via its RVxF motif.

Recruitment of PP1 to mitochondria promotes dephosphorylation of resident mitochondrial proteins

To explore the function of PP1 at mitochondria following recruitment by LpPIP, we compared the mitochondrial phosphoproteomes of wild-type HEK293 cells to stable cell lines expressing either ^FLAGLpPIP or ^FLAGRVxF_dead (Fig 3A). The steady-state proteomes of these cell lines remained unchanged, eliminating the need for normalization of phosphopeptide abundance based on protein levels (Fig 3B and S5 Table). A total of 8031 proteins were detected, of which 982 were mitochondrial proteins, representing ~80% coverage of the mitochondrial proteome (Fig 3B and S5 Table) [55]. Phosphoproteomic analysis detected 18,420 peptides, of which 15,151 were phosphopeptides, corresponding to an 82% phosphopeptides enrichment. Four hundred twelve of these were phosphopeptides belonged to known mitochondrial proteins. Cells expressing ^FLAGLpPIP displayed a significant decrease in the abundance of phosphopeptides derived from mitochondrial outer membrane proteins (Fig 3C and S5 Table). This trend was consistently observed across experiments using wild-type and ^FLAGRVxF_dead expressing cells as controls (Fig 3D). These results demonstrate that LpPIP-mediated recruitment of PP1 to mitochondria leads to the dephosphorylation of specific mitochondrial outer membrane proteins, indicating a targeted mechanism by which L. pneumophila manipulates the host mitochondrial phosphoproteome.

Download:

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

Fig 3. Recruitment of PP1 by LpPIP dephosphorylates mitochondrial proteins.

(A) Schematic of the phosphoproteomics workflow used in this study. Wild-type Flp-In T-REx 293 cells and stable cell lines expressing ^FLAGLpPIP or ^FLAGRVxF_dead were induced with 1 µg/mL tetracycline for 24 h. Proteins from isolated mitochondria were trypsinized and cleaned by solid phase extraction before phosphopeptide enrichment using titanium dioxide (TiO₂) beads. Global proteomics and phosphopeptide-enriched samples were analyzed by quantitative mass spectrometry and processed using Spectronaut and Perseus. (B) Scatterplot comparing mitochondrial protein abundance between Flp-In T-REx 293 stable cells expressing ^FLAGLpPIP or ^FLAGRVxF_dead (n = 3 technical replicates). Log₂ fold change of mean LFQ intensity is plotted against −Log₁₀ p-value. Significance thresholds were set at ± 1.5-fold change and p-value = 0.01. Corresponding data are available in the Supporting information (S5 Table). (C) Scatterplot comparing mitochondrial phosphopeptide abundances between Flp-In T-REx 293 cells expressing ^FLAGLpPIP and ^FLAGRVxF_dead (n = 3 technical replicates). Data was filtered for mitochondrial and mitochondrial-interacting proteins using the MitoCarta3.0 annotation and restricted to serine/threonine phosphopeptides. Significance thresholds were set at ± 1.5-fold change and p-value of 0.01 using Student t test. Phosphopeptides from mitochondrial outer membrane proteins are labeled in orange. Hollow circles indicate phosphopeptides with imputed values; solid circles represent those without. Corresponding data are available in the Supporting information (S5 Table). (D) Heatmap of the mitochondrial phosphopeptides significantly reduced in Flp-In T-REx 293 stable cells expressing ^FLAGLpPIP relative to wild-type (n = 4 technical replicates) and cells expressing ^FLAGRVxF_dead (n = 3 technical replicates). Average of the Log₂ abundance of phosphopeptide for each group was used to generate the heatmap in Prism 10. Missing values are indicated by crosses. Corresponding data are available in the Supporting information (S5 Table). (E) pLogo [56] sequence motif analysis of the seven residues flanking 39 phosphoserine/phosphothreonine (pSer/pThr) sites significantly dephosphorylated in ^FLAGLpPIP expressing cells (Fig 3C). Arginine (R) at position −3 relative to pSer/pThr was significantly enriched (log-odd probability of 6.213 and p-value of 1.715 e−4). (F) Gene ontology (GO) biological process enrichment analysis of mitochondrial proteins with significantly downregulated phosphopeptides (Fig 3D), performed using ShinyGO 0.80 [57]. The top 10 enriched GO terms are shown. Corresponding data are available in the Supporting information (S2 Data).

https://doi.org/10.1371/journal.pbio.3003261.g003

To confirm the dephosphorylated proteins are specific targets of PP1 activity, seven residues flanking the target phosphoserine/phosphothreonine (pSer/pThr) residue were aligned for motif assessment using pLogo [56]. A high frequency (35.29%) of arginine at position −3 relative to the pSer/pThr (Fig 3E) aligns with the intrinsic affinity of the PP1 catalytic subunit to dephosphorylate serine with N-terminal basic residues, in particular arginine at position −3 (RXXpS motif) [58]. This supports the hypothesis that the mitochondrial outer membrane proteins dephosphorylated by PP1 are targeted specifically due to the activity of PP1 [58]. Gene ontology (GO) enrichment analysis of the dephosphorylated mitochondrial peptides/proteins revealed enrichment in processes related to mitochondrial organization and morphology and protein biogenesis (Fig 3F) [57], which align with the functions of mitochondrial outer membrane proteins and with pathways known to be targeted by L. pneumophila during infection [16].

Consequences on mitochondrial function of LpPIP-PP1 dephosphorylation events

We compared the mitochondrial proteins dephosphorylated in our datasets (Fig 3D and S5 Table) with an existing phosphoproteomics dataset from HEK293 FcγR cells infected with L. pneumophila wild-type or a T4SS-deficient strain [17]. Notably, the phosphosites FUNDC2_S151, TOMM70_S91_S110, and MFF_S155_S157, which were among the top dephosphorylated sites in the presence of LpPIP (Fig 3D and S5 Table), were significantly dephosphorylated in HEK293 FcγR cells infected with wild-type L. pneumophila for 8 h compared to T4SS-deficient strain [17]. This overlap suggests that LpPIP contributes to the dephosphorylation of these mitochondrial phosphosites during L. pneumophila infection. We wanted to explore the potential functional implications on mitochondrial morphology, protein biogenesis, and mitochondrial health, given the readout of our phosphoproteomics.

First, we assessed whether cellular bioenergetics (a measure of mitochondrial health) was affected by LpPIP expression. We measured the oxygen consumption rate (mitochondrial oxidative phosphorylation) and extracellular acidification rate (glycolysis) in wild-type and stable cell lines expressing ^FLAGLpPIP or ^FLAGRVxF_dead, but did not observe any significant difference (Fig 4A and 4B). FUN14 domain containing 2 (FUNDC2) is an outer membrane protein implicated in various biological functions, including apoptosis signaling and regulation of mitochondrial dynamics through inhibition of Mitofusin 1 (Mfn1) [59,60]. The functional implications of FUNDC2_pS151 dephosphorylation have not been determined [60]. FUNDC2 has been reported to inhibit caspase-3 cleavage and apoptosis [59] and ectopic expression of ^GFPLpPIP has been shown to induce caspase-3 cleavage [39]. In line with this, stable cell lines expressing ^FLAGLpPIP showed increased caspase-3 cleavage upon staurosporine treatment compared to wild-type cells; however, this effect was independent of PP1 recruitment as it was also observed with ^FLAGRVxF_dead expression (Fig 4C).

Download:

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

Fig 4. Downstream consequences of LpPIP-PP1 dephosphorylation events.

(A) Time course trace of oxygen consumption rate (OCR) in wild-type Flp-In T-REx 293 cells (16 technical replicates) and stable cells expressing ^FLAGLpPIP (15 technical replicates) or ^FLAGRVxF_dead (15 technical replicates), following injections of oligomycin, FCCP, rotenone, and antimycin A. Data represents mean ± SD. Corresponding raw data are available in the Supporting Information (S3 Data). (B) Time course trace of extracellular acidification rate (ECAR) for the cells described in Fig 4A, following incubation with oligomycin and 2-deoxy-d-glucose. Data represents mean ± SD. Corresponding raw data are available in the Supporting Information (S3 Data). (C) Wild-type Flp-In T-REx 293 cells and stable cells expressing ^FLAGLpPIP or ^FLAGRVxF_dead were induced with 1 µg/mL tetracycline for 24 h, followed by treatment with 1 µM staurosporine (STS) for 24 h. Whole cell samples were processed for SDS-PAGE and immunoblots. Corresponding raw images are available in the Supporting Information (S1 Raw Images). (D) Quantification of mitochondrial mean aspect ratio in wild-type HeLa cells and those transfected with ^FLAGLpPIP or ^FLAGRVxF_dead (n = 3 independent experiments, 10 cells per experiment). Data are presented as mean ± SEM of the three experiments. Ordinary one-way ANOVA was performed on the means, showing a significant difference of 0.2273 between wild-type and ^FLAGLpPIP-transfected cells, with a p-value of 0.0309. Corresponding raw data are available in the Supporting information (S1 Data). (E) Quantification of mitochondrial mean form factor in HeLa cells and those transfected with ^FLAGLpPIP or ^FLAGRVxF_dead (n = 3 independent experiments, 10 cells in each experiment). Data are presented as mean ± SEM of the three experiments. Ordinary one-way ANOVA was performed on the means, showing no significant differences among the groups. Corresponding raw data are available in the Supporting information (S1 Data). (F) HeLa cells transiently transfected with ^FLAGLpPIP, or ^FLAGRVxF_dead were stained with 100 nM MitoTracker Deep Red FM (mitochondria; MTDR) and immunostained with antibodies against FLAG (green) and Drp1 (magenta). Nuclei were stained with Hoechst 33258 (blue). Scale bar represents 10 μm. (G) Proportion of Drp1 localized to mitochondria in HeLa cells transiently transfected with ^FLAGLpPIP or ^FLAGRVxF_dead, calculated using raw integrated density in Fiji. Representative images are shown in Fig 3F. Data are presented as mean ± SD from 13 cells in one experiment. Unpaired t test showed no significant difference between the groups. Corresponding raw data are available in the Supporting information (S1 Data). (H) [³⁵S]-Tim23 was imported into mitochondria isolated from wild-type Flp-In T-REx 293 cells and stable cells expressing ^FLAGLpPIP or ^FLAGRVxF_dead. Following import mitochondria were isolated and treated with PK. Isolated mitochondria were solubilized in 1% [w/v] digitonin and the separated by BN-PAGE and observed by phosphor image analysis. Densitometric quantification was calculated as the percentage of import at 60 min in mitochondria from wild-type Flp-In T-REx 293 cells, normalized to SDHA loading control. Data are presented as mean ± SD of three experiments. Ordinary one-way ANOVA showed no significant difference among the three cell lines. Corresponding raw data are available in the Supporting Information (S1 Raw Images and S4 Data). (I) [³⁵S]-GC1 was imported into mitochondria isolated from wild-type Flp-In T-REx 293 cells and stable cells expressing ^FLAGLpPIP or ^FLAGRVxF_dead as described in Fig 4H. Densitometric quantification and ordinary one-way ANOVA showed no significant difference among the three cell lines. Corresponding raw data are available in the Supporting information (S1 Raw Images and S4 Data). (J) Quantification of mitophagy from live-cell imaging of Flp-In T-REx 293 cell lines constitutively expressing mt-Keima. Experimental conditions include wild-type Flp-In T-REx 293 cells cultured in normal media or media adjusted to pH 4 (positive control), and stable cells induced to express either ^FLAGLpPIP or ^FLAGRVxF_dead. The 561/488 mt-Keima signal (red-to-green fluorescence ratio) was quantified using the Ratio Plus plugin in Fiji and normalized to that of wild-type cells in normal media. Data represent mean ± SEM from two to three independent experiments (>30 images per experiment). Statistical analysis by ordinary one-way ANOVA showed no significant differences among the cell lines, except for the pH 4 control. Corresponding raw data are available in the Supporting information (S1 Data).

https://doi.org/10.1371/journal.pbio.3003261.g004

Mitochondrial fission factor (MFF) functions to recruit the mitochondrial fission protein Drp1 [61]. Given this and the reported function of FUNDC2 in mitochondrial morphology, we quantified organelle morphology in ^FLAGLpPIP or ^FLAGRVxF_dead expressing cells, but found no changes to tubule length or network complexity (Fig 4D and 4E). Phosphorylation of MFF S155 and S172 promotes Drp1 recruitment to mitochondria, leading to mitochondrial fragmentation [62]. In the presence of ^FLAGLpPIP, S155 and S157 of MFF were dephosphorylated (Fig 3D), but the basal proportion of Drp1 at mitochondria was not influenced (Fig 4F and 4G).

TOMM70 is an import receptor of the outer membrane TOM complex [63] that recognises carrier proteins as part of the TIM22 pathway for the import of these hydrophobic substrates into the inner membrane [63,64]. Phosphorylation of S91 on TOMM70 has been shown to enhance the import of mitochondrial carrier proteins [63]. Thus, S91 and S110 dephosphorylated in the presence of ^FLAGLpPIP, could impair import of TIM22 substrates. However, in vitro import and assembly of the TIM22 substrates, [³⁵S]-Tim23 and [³⁵S]-GC1, was comparable across mitochondria isolated from wild-type and stable cell lines expressing ^FLAGLpPIP or ^FLAGRVxF_dead (Fig 4H and 4I). We also assessed BNIP3L, which contains a LC3-interacting region that facilitates autophagosome recruitment to mitochondria [65], and was dephosphorylated at S166 (Fig 3D). Previous studies have shown that both PP1 and PP2A can dephosphorylate the mitophagy receptor, BNIP3, leading to its degradation and suppression of mitophagy [66]. mt-Keima, a mitochondrial-targeted, pH-sensitive fluorescent protein that enables quantification of mitochondria undergoing lysosomal degradation [67], revealed no significant differences in basal mitophagy levels due to LpPIP expression (Fig 4J).

Taken together, these data suggest that Legionella effectors can interact with mitochondrial components in ways that may not produce immediate or easily detectable changes in cellular phenotype, possibly due to factors such as compensatory mechanisms within the host cell, or the need for specific environmental conditions, in relevant cell lineages to reveal their effects.

LpPIP-PP1 targets substrates based on cellular location

As a final question, we wanted to investigate if LpPIP has any role in determining PP1 substrate specificity at the mitochondrial outer membrane. To do this, we redirected LpPIP-PP1 to the ER and examined the dephosphorylation substrate profile when LpPIP was no longer mitochondrially-localized. Although the K128A and K97A/R100A/K128A charge mutants of LpPIP partially localized to the ER, they still colocalized with mitochondria (S2E and S4A Figs). To circumvent this, we generated a chimera consisting of the cytosolic domain of LpPIP (amino acids 1–103) fused to the tail-anchor region (amino acids 109–134) of the ER TA protein CYB5A (Fig 5A) [41]. The resulting construct, ^FLAGLpPIP-CYB5A^TA, colocalized with the ER marker PDI in transfected HeLa cells (Figs 5A and S4A). ^FLAGLpPIP-CYB5A^TA recruited PPP1CB to the ER in transfected HeLa cells, as shown quantitatively with the Pearson’s correlation analysis (Figs 5B and S4B).

Download:

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

Fig 5. LpPIP does not specify substrates selectivity for PP1.

(A) HeLa cells transiently transfected with ^FLAGLpPIP or ^FLAGLpPIP-CYB5A^TA were stained with 100 nM MitoTracker Deep Red FM (mitochondria; MTDR) and immunostained with antibodies against FLAG (green) and PDI (ER; magenta). Nuclei were stained with Hoechst 33258 (blue). Scale bar represents 10 μm. (B) Pearson correlation analysis of FLAG or PDI and PPP1CB fluorescence signals was performed on ^FLAGLpPIP-CYB5A^TA expressing HeLa cells and untransfected wild-type HeLa cells (n = 3 independent experiments, 16 cells per experiment). Data are shown as the mean ± SEM of three independent experiments. Larger circles represent mean values from each experiment, and smaller circles represent individual cell values. Unpaired t test was performed on the means, revealing a significant difference of 0.35 ± 0.03 between the two groups, with a p-value of 0.0005. Representative images are shown in S4B Fig. Corresponding raw data are available in the Supporting information (S1 Data). (C) Whole cell FLAG co-immunoprecipitation was performed using untransfected HeLa cells, HeLa cells expressing ^FLAGLpPIP, or ^FLAGLpPIP-CYB5A^TA (n = 3 technical replicates). The Log₂ difference of proteins enriched in ^FLAGLpPIP and ^FLAGLpPIP-CYB5A^TA relative to untransfected cells is shown. Proteins with significant fold change (± 1.5-fold change with a p-value ≤ 0.05) are represented as filled circles, while non-significant differences are shown as hollow circles. Mitochondrial proteins based on MitoCarta3.0 are indicated in orange, and ER proteins based on the Human Protein Atlas are indicated in blue. Corresponding data are available in the Supporting information (S6 Table). (D) Volcano plot showing proteins enriched following FLAG co-immunoprecipitation from HeLa cells expressing ^FLAGLpPIP-CYB5A^TA compared to untransfected cells (n = 3 technical replicates). Mitochondrial proteins based on MitoCarta3.0 are indicated in orange, and ER proteins based on the Human Protein Atlas are indicated in blue. Corresponding data are available in the Supporting information (S6 Table). (E) Scatterplot comparing phosphopeptides abundances in whole cell samples from HeLa cells transfected with ^FLAGLpPIP to empty vector pCDNA5 (n = 4 technical replicates). Data were filtered for serine and threonine phosphopeptides. The significance thresholds were set to a ±1.5-fold change with a p-value of 0.01 using Student t test. Phosphopeptides derived from mitochondrial outer membrane proteins (MitoCarta3.0) and ER proteins (Human Protein Atlas) are labeled in orange and blue, respectively. Non-imputed phosphopeptides are shown as filled circles, while phosphopeptides with imputed values are shown as hollow circles. Corresponding data are available in the Supporting information (S7 Table). (F) Scatterplot comparing phosphopeptides abundances in whole cell samples from HeLa cells transfected with ^FLAGLpPIP-CYB5A^TA to empty vector pCDNA5 (n = 4 technical replicates), processed as described in Fig 5E. Corresponding data are available in the Supporting information (S7 Table).

https://doi.org/10.1371/journal.pbio.3003261.g005

Co-immunoprecipitation of ^FLAGLpPIP or ^FLAGLpPIP-CYB5A^TA from whole cell lysates of transfected HeLa cells revealed that the ER-localized LpPIP chimera interacted with fewer mitochondrial proteins and more ER proteins and it continued to robustly interact with PP1 (Fig 5C and 5D and S6 Table). Whole cell phosphoproteomics (S7 Table) to determine the substrates of the ER-localized LpPIP-CYB5A^TA-PP1 revealed an increase in ER proteins and reduction in mitochondrial proteins being dephosphorylated in cells expressing ^FLAGLpPIP-CYB5A^TA (Figs 5E, 5F, S4C and S4D and S7 Table). These data suggested that PP1’s substrate specificity is largely determined by its localization and its proximity to organelle-specific proteins. Sequestering PP1 to either mitochondria or ER, thereby reducing its abundance from the typical cytosolic and nuclear locations, is expected to broadly disrupt the cellular phosphorylation landscape. This effect was observed in cells overexpressing ^FLAGLpPIP or ^FLAGLpPIP-CYB5A^TA, which showed increased phosphorylation of proteins predominantly localized to the nucleus (S4G and S4H Fig).

Cell culture and cell lines

Flp-In T-REx 293 (ThermoFisher Scientific #R75007; RRID: CVCL_U427) and HeLa cells (RRID: CVCL_0030) were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco # 11995073) supplemented with 5% [v/v] heat-inactivated fetal bovine serum (FCS; Gibco #A3382001) and 1% [v/v] penicillin-streptomycin (pen/strep; Gibco #15-140-122). THP-1 cells (ATCC # TIB-202) were cultured in Roswell Park Memorial Institute 1640 (RPMI) medium (Gibco, # 61870127) supplemented with 10% [v/v] heat-inactivated FCS (Bovagen # FBSAU-2207C). All were incubated at 37 °C with 5% CO₂ in a humidified incubator. ^FLAGLpg1625 and ^FLAGLpg1625^K17A/V19A stable cell lines were all generated from Flp-In T-REx 293 cells.

Bacterial cell cultures and strains

Escherichia coli strains DH5α and Pir2 were used for the propagation of genetic material and were cultured at 37 °C on Luria–Bertani (LB) agar or broth supplemented with 50 µg/mL ampicillin, 25 µg/mL chloramphenicol, or 15 µg/mL kanamycin where required. Legionella pneumophila Philadelphia-1 strain JR32 was grown on Buffered-Charcoal Yeast Extract (BCYE) agar supplemented with 135 µg/mL iron (III) nitrate nonahydrate and 400 µg/mL L-cysteine at 37 °C. Where appropriate, the bacteria were selected using 6 µg/mL chloramphenicol, 15 µg/mL kanamycin, or 100 µg/mL streptomycin. L. pneumophila was transformed via triparental mating [89] or electroporation [90]. E. coli strain HB101 carrying a pRK600 mobilization plasmid was used in triparental mating to facilitate the uptake of the donor DNA by the L. pneumophila recipients.

Molecular biology

L. pneumophila gene sequence was obtained from NCBI. Genomic DNA isolated from wild-type L. pneumophila Philadelphia-1 strain JR32 was used as template for the amplification of lpg1625, lpg1803, lpg2444, and lpg2344. These were cloned into the pCDNA5 vector with a FLAG tag encoded upstream. Site-directed mutagenesis was performed using Q5 site-directed mutagenesis kit (NEB #E0552S). Oligonucleotides were designed with the online NEB primer design software (NEBaseChanger.neb.com).

Genetic manipulation of L. pneumophila

The ∆lpg1625 strain was generated using an established method [90]. Briefly, the up- and downstream regions flanking the lpg1625 open reading frame were amplified from genomic DNA and ligated by overlap extension PCR. The resulting DNA fragment was inserted into the R6K vector pSR47s and introduced into L. pneumophila by triparental mating. The vector was integrated into the genome by homologous recombination and the target gene was subsequently removed upon resolution of pSR47s by sucrose counter-selection. Knock-out (KO) clones were identified using MyTaq HS Red polymerase as per the manufacturer’s instructions, using primers screening across the region of interest and primers targeting lpg1625.

The ∆lpg1625 strain was complemented with constitutively expressed gene variants. The previously generated constructs, ^FLAGLpg1625 and ^FLAGRVxF_dead, were inserted into a non-integrative pMMB207 plasmid and introduced into L. pneumophila ∆lpg1625 by electroporation. Expression of Lpg1625 derivatives was induced with 1 mM isopropyl ß-D-1-thiogalactopyranoside (IPTG), which was validated by western blot analysis.

Transfection and stable cell lines generation

Lipofectamine 3000 (ThermoFisher Scientific # L3000008) was used for transfection as per the manufacturer’s protocol. Briefly, cells were seeded to be approximately 70% confluent on the day of transfection. For generation of tetracycline-inducible stable cell lines, Flp-In T-REx 293 cells were co-transfected with pOG44 and pCDNA5 plasmids encoding the gene of interest. Following 72 h, cells were placed on selection with 200 µg/mL Hygromycin B (Gibco #10687010). Selection media was replaced as required until all non-transfected control cells were dead. Cells that survived the hygromycin selection were pooled and screened for expression of the protein of interest after 24 h induction with 1 µg/mL tetracycline (Sigma-Aldrich #T7660).

For generation of stable cell lines constitutively expressing mt-Keima, retrovirus was packaged in HEK293T cells following co-transfection with 9 µg pCHAC-mt-mKeima (Addgene #72342), 6 µg pUMVC3, and 3 µg pCMV-VSVG in a 10 cm tissue culture dish. After 48 h, the viral supernatant was collected, filtered through a 0.45 µm Millex-HV syringe filter unit (Millipore), and used to transduce Flp-In T-REx 293, ^FLAGLpg1625 Flp-In T-REx 293, and ^FLAGLpg1625^K17A/V19A Flp-In T-REx 293 cells in a 6-well plate for 24 h in the presence of 8 µg/mL polybrene (Sigma-Aldrich #107689). mt-Keima-positive cells were subsequently acquired via fluorescence-activated cell sorting.

Infection of THP-1 cells with L. pneumophila

Intracellular replication of wild-type and mutant strains of L. pneumophila in THP-1 cells was assessed by counting colony-forming units (CFU/mL). THP-1 cells were seeded at a density of 5 × 10⁵ cells and differentiated into macrophage-like cells with 10 nM phorbol 12-myristate 13-acetate (PMA) 72 h prior to infection. Stationary-phase L. pneumophila was resuspended in PBS and diluted to a multiplicity of infection (MOI) of 5 in RPMI containing 1 mM of IPTG, where appropriate. The L. pneumophila inoculum was incubated with THP-1 cells for 2 h at 37 °C and 5% CO₂ to allow bacterial internalization before the remaining extracellular bacteria were killed with 100 µg/mL gentamicin in RPMI for 1 h at 37 °C and 5% CO₂. At the appropriate time points post-infection, the THP-1 cells were lysed with 0.05% digitonin in PBS. The harvested samples were serially diluted and plated onto BCYE agar and incubated for 4 days at 37 °C before quantifying CFU/mL. Five biological replicates were used for the quantification. Statistical test was carried out with Prism 10 using two-way ANOVA.

Oxygen consumption rate and extracellular acidification rate measurement

Mitochondrial stress and glycolytic stress tests were performed using a Seahorse Bioscience XFe96 Analyzer according to manufacturer protocols. Forty-eight hours before the assay, 1.5 × 10⁴ of wild-type Flp-In T-REx 293 (16 technical replicates), ^FLAGLpg1625 Flp-In T-REx 293 (15 technical replicates), and ^FLAGLpg1625^K17A/V19A Flp-In T-REx 293 stable cells (15 technical replicates) were plated per well in XFe96 culture plates pre-coated with 100 µg/mL poly-d-Lysine and grown overnight under standard culture conditions. Twenty-four hours before the assay, cells were induced with 1 µg/mL tetracycline (Sigma-Aldrich #T7660). Cells were assayed in non-buffered DMEM media (Agilent #102353-100) containing either 10 mM glucose, 1 mM sodium pyruvate, and 2 mM glutamine (mitochondrial stress test) or 2 mM glutamine only (glycolytic stress test). For the mitochondrial stress test, rates were measured following sequential incubations with 2 μM oligomycin, 0.5 μM carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP), and 0.5 μM rotenone and 0.3 μM antimycin A. For the glycolytic stress test, these incubations instead contained 10 mM glucose, 2 μM oligomycin, and 50 mM 2-deoxy-d-glucose. For both tests, measurements consisted of 3 cycles each of 3 min mix and 3 min measure. Following the assay, cell numbers per well were normalized using CyQuant (Thermo Fisher Scientific #C7026).

Immunofluorescence and confocal microscopy

Twenty-two mm glass coverslips were sterilized with 75% [v/v] ethanol and placed in 6-well plates. HeLa cells were seeded onto the coverslips the day before transfection. When required, cells were stained with 100 nM MitoTracker Deep Red FM (ThermoFisher Scientific #M22426) in serum-free DMEM for 40 min at 37 °C. Cells were then fixed in 4% [w/v] paraformaldehyde (PFA; ProSci Tech #C004) diluted in 5% [w/v] sucrose (Chem-Supply #SA030) for 10 min at room temperature. For ER staining, 0.1% [w/v] glutaraldehyde (Sigma-Aldrich #G7526) was included for the fixation. The cells were washed with phosphate buffered saline (PBS; 137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 1.8 mM KH₂OP₄, pH 7.4) three times before being permeabilised by 0.1% [v/v] Triton X-100 (Sigma-Aldrich #T9284-500ML) in PBS (3× over 10 min). Samples were blocked with 3% [w/v] bovine serum albumin (BSA; Sigma-Aldrich #A7030) in PBS for 30 min, followed with three washes of PBS. Primary and secondary antibodies were diluted appropriately with 3% BSA in PBS. The coverslips were incubated with the primary antibodies for 1 h at room temperature. Coverslips were washed with PBS (4× over 10 min) before incubating with secondary antibodies for 1 h at room temperature. Three washes with PBS followed, with approximately 0.01 µg/µL Hoechst 33258 (Sigma-Aldrich #14530) added to the second wash for DNA staining. The coverslips were then mounted to microscope slides using approximately 50 µL of mounting medium (0.2 M 1,4-diazabicyclo[2.2.2]octane (Sigma-Aldrich #D27802), 80 mM Tris-Cl pH 8.0, 90% [v/v] glycerol) and sealed with nail polish. The slides were imaged using a Leica SP8 confocal microscope and processed with Fiji software in ImageJ.

Primary antibodies used for immunofluorescence included: anti-FLAG (Sigma #F1804, dilution 1:500), anti-NDUFAF2 (Ryan Lab, Monash University, made in-house, dilution 1:500), anti-PPP1CB (Abcam #ab53315, dilution 1:500), anti-PDI (Enzo # ADI-SPA-891-D, dilution 1:300), anti-PEX14 (Proteintech #10594-1-AP, dilution 1:500), anti-GM130 (Abcam #ab52649, dilution 1:600; gifted from Gleeson Lab, University of Melbourne), and anti-Drp1 (BD Biosciences #611113, dilution 1:500). Secondary antibodies used were AlexaFluor488 (Invitrogen #A11001, dilution 1:500), and AlexaFluor568 (Invitrogen #A11011, dilution 1:500).

Live-cell imaging for mt-Keima

Prior to live-cell imaging, cells were seeded onto FluoroDish cell culture dishes (World Precision Instruments) and grown for 1 day (37 °C, 5% CO₂) in the presence of 1 µg/mL tetracycline for 24 h to induce protein expression. The following day, confocal imaging was conducted using a Zeiss Elyra LSM880 (Carl Zeiss MicroImaging) equipped with a Plan-Apochromat 63×/1.40-NA oil immersion objective lens. mt-Keima fluorescence was imaged by sequential excitation with Argon 25 mW 458 and 561 nm 20 mW lasers, and emission was detected with BP 495-550 + LP570 filter. Control pH 4 samples were prepared as described previously [91]. Briefly, cells were permeabilised with 0.01% Triton X-100 in PBS and equilibrated in DMEM adjusted to pH 4.0. Imaging parameters, including laser power and gain, were kept consistent across all samples and conditions. Images were analyzed using the Ratio Plus plugin for Fiji image software [92] to determine the 561/488 mt-Keima ratio, which was normalized to the 561/488 ratio in wild-type cells.

Caspase-3 activation assay

3 × 10⁵ wild-type Flp-In T-REx 293, ^FLAGLpg1625 Flp-In T-REx 293, and ^FLAGLpg1625^K17A/V19A Flp-In T-REx 293 stable cells were plated in a 6-well plate and grown overnight under standard culture conditions. Cells were induced with 1 µg/mL tetracycline (Sigma-Aldrich #T7660) for 24 h (37 °C, 5% CO₂) before treating with 1 µM staurosporine (Sigma-Aldrich #GA8507) or an equivalent volume of DMSO (vehicle control) and grown for a further 24 h. Whole cell lysates were harvested for analysis by SDS-PAGE.

Image analyses using Fiji

For fluorescence profile, a line was first drawn across a region of interest (ROI) using ROI Manager. The intensities of pixels along the line for each channel were obtained through the Plot Profile in Fiji. The data points from different channels were copied and a graph generated in Prism 10. Colocalization analysis was performed with the plugin Coloc2 [93] using Costes threshold regression. Pearson’s R value (per cell) was quantified (number of cells indicated in figure legends for Figs 2F, 5B, S2B-D, and S4A) analyzed per experiment. Mean Pearson’s R values were compared using unpaired t test (for two groups, one independent variable) or ordinary one-way ANOVA (more than two groups, one independent variable) or ordinary two-way ANOVA (more than two groups, two independent variables. Statistical analysis was performed using Prism 10.

To quantify the proportion of PPP1CB or Drp1 signal at mitochondria, whole cell and mitochondria ROIs were defined. Whole cell ROIs were defined manually, while mitochondrial ROIs were defined using the plugin Mitochondrial Analyzer [94]. The raw integrated density of each ROI was measured with the function Integrated density. The proportion of the signal at mitochondria was then calculated by dividing the raw integrated density at mitochondria with that from the whole cell (number of cells indicated in figure legends for Fig 2G and 4G). Statistical analysis was performed using unpaired t test, Prism 10.

For quantification of mitochondrial morphology, the Fiji plugin Mitochondrial Analyzer [94] was used for mitochondrial thresholding and subsequent determination of mean aspect ratio and mean form factor quantified per cell. The means of these parameters across three independent experiments were compared statistically using ordinary one-way ANOVA in Prism 10.

Mitochondria isolation and mitochondrial treatments

Mitochondria were isolated from tissue culture cells using differential centrifugation [95]. Briefly, cells were harvested in cold PBS using a cell scraper and pelleted by centrifugation at 600g (5 min, 4 °C). Cells were homogenized in isolation buffer (20 mM HEPES-KOH pH 7.6, 220 mM mannitol 70 mM sucrose, 1 mM EDTA, and 0.5 mM phenylmethanesulfonyl fluoride (PMSF)) using a handheld glass Dounce homogenizer. The homogenate was centrifuged at 800g (10 min, 4 °C), with the resultant supernatant further centrifuged at 10,000g (10 mins, 4 °C). The pellets containing crude mitochondria were resuspended in isolation buffer (without PMSF). Protein concentration of the suspension was estimated using Pierce BCA protein assay kits (ThermoFisher Scientific #23225) before downstream applications/treatments. For phosphoproteomics, instead of PMSF, protease inhibitor (Roche #11873580001) and PhosSTOP (Roche #4906837001) were added into the isolation buffer and were included in every steps post homogenization.

For sub-mitochondrial fractionation, mitochondrial pellets were resuspended in isolation buffer, HEPES buffer (10 mM HEPES (Chem-Supply #HEPES05-500G), pH 7.6), or 0.5% Triton X-100 at 1 mg/mL. Mitochondrial lysates were treated with 50 µg/mL Proteinase K (PK; Sigma-Aldrich #P6556) for 10 min on ice, followed by a further 5 min incubation on ice after addition of final 1 mM PMSF in all samples. For carbonate extraction, mitochondrial pellet was resuspended in freshly made 100 mM sodium carbonate at 1 mg/mL and incubated on ice for 30 min. Centrifugation at 100,000g (30 min, 4 °C) then separated the pellet (membrane proteins) and the supernatant (soluble and peripheral membrane proteins) fraction. All samples were precipitated with 12.5% [w/v] trichloroacetic acid (TCA; Chem-Supply #TA030) before analysis with SDS-PAGE.

Mitochondrial in vitro protein import assays

Isolated mitochondrial pellets were resuspended to 1 mg/mL in import buffer (250 mM sucrose, 5 mM magnesium acetate, 80 mM potassium acetate, 10 mM sodium succinate, 1mM DTT, 5 mM ATP, 20 mM HEPES-KOH pH 7.4) with or without the addition of 10 µM carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP; Sigma-Aldrich #C2920) to dissipate the membrane potential. Samples were incubated at 37 °C for 1 min prior to the addition of radiolabeled protein and further incubation at 37 °C with radiolabeled protein to allow for its import. Import reactions were stopped by placing samples on ice. Fifty µg/mL PK was added to samples immediately following import for 10 min on ice, followed by the addition of 1 mM PMSF for 5 min on ice. The mitochondria were isolated at 12,000g for 5 min at 4 °C and washed once with import buffer. Mitochondrial pellets were then analyzed using BN-PAGE. After western transfer, radioactive signal on the PVDF membrane was exposed to a storage phosphor screen in a cassette (GE Healthcare; from 1 h up to 2 weeks). The radioactivity on the storage phosphor screen was then detected using phosphor imaging with Amersham Typhoon (GE Healthcare). If required, densitometric quantification of the radioactive signals was performed on Image Lab 6.1, normalized against SDHA loading controls.

Tricine SDS polyacrylamide gel electrophoresis and western blotting

Tris-tricine SDS-PAGE gel was used for electrophoresis of denatured samples before immunoblot analysis [95]. 16% [v/v], 10% [v/v], and 4% [v/v] acrylamide gel mixes which constitute the gels were made with acrylamide solution (49.5% [w/v] acrylamide, 3% [w/v] bis-acrylamide) and tricine gel buffer (1 M tris, 0.1% [w/v] SDS, pH 8.45), with an addition of 13% [w/v] glycerol in the 16% [v/v] acrylamide gel mix. 10%–16% acrylamide gradient gels were poured using a gradient mixer mixing 16% [v/v] and 10% [v/v] acrylamide gel mixes. The gels were polymerized with the addition of tetramethylethylenediamine (TEMED; Sigma-Aldrich #T22500) and 10% [w/v] ammonium persulfate (APS; Sigma-Aldrich #A3678) right before pouring. A 4% [v/v] acrylamide stacking gel was poured on top of the polymerized 10%–16% gradient gel. Samples resuspended in loading dye (50 mM Tris-Cl pH 6.8, 100 mM dithiothreitol (DTT), 2% [w/v] sodium dodecyl sulfate (SDS), 10% [v/v] glycerol, 0.05% [w/v] bromophenol blue) were boiled at 65 °C for 15 min before loaded onto the gel along with PageRuler Prestained Protein Ladder (ThermoFisher Scientific #26616) as the molecular weight marker. Electrophoresis of the gels was performed with anode buffer (0.2 M Tris-Cl, pH 8.9) and cathode buffer (0.1 M Tris, 0.1 M tricine, 0.1% [w/v] SDS, pH 8.45).

After electrophoresis, SDS-PAGE gels were transferred to 0.45 µm PVDF membrane (Millipore #IPVH00010) using Owl HEP Series Semidry Electroblotting System (ThermoFisher Scientific #HEP-1). After transfer, PVDF membranes were blocked with 5% [w/v] non-fat milk (Nestle #12428935) in PBST (PBS with 0.05% Tween20 (Sigma-Aldrich #P1397)) for 1 h at room temperature. Membranes were then incubated with primary antibodies at 4 °C overnight before incubated with horseradish peroxidase-conjugated secondary antibodies (Sigma-Aldrich #A9044 and #A0545, 1:5000 dilution in 5% [w/v] non-fat milk) for 2 h at room temperature. After that, the blots were developed using Clarity Western ECL Substrate (Bio-Rad #1705061) and imaged with ChemiDoc imaging system (Bio-Rad #12003153).

Primary antibodies used for western blotting include anti-FLAG (Sigma #F1804, dilution 1:1000), anti-MFN2 (Ryan Laboratory, dilution 1:250), anti-TIMM29 (Sigma-Aldrich #HPA041858, dilution 1:1000), anti-TIMM44 (Proteintech #13859-1-AP, dilution 1:500), anti-SDHA (Abcam #AB14715, dilution 1:1000), and anti-caspase-3 (Cell Signaling Technology #9662, dilution 1:500).

Blue native polyacrylamide gel electrophoresis (BN-PAGE)

BN-PAGE was used for the separation of native protein complexes. Acrylamide gradient gels were poured with 4% [v/v] and 16% [v/v] acrylamide gel mixes made in blue native gel buffer (66.7 mM ε-amino-n-caproic acid (Sigma-Aldrich #D7754), 50 mM Bis-Tris pH 7.0). 4% [v/v] acrylamide stacking gel was poured on top of the gradient gel. Mitochondrial samples were solubilized in 1% [w/v] digitonin (Merck #300410) in solubilization buffer (20 mM Tris-Cl pH 7, 50 mM NaCl, 10% [v/v] glycerol) on ice for 30 min. Insoluble material was removed by centrifugation at 16,000g for 20 min at 4 °C. One-tenth the volume of 10× BN loading dye (5% [w/v] Coomassie brilliant blue G-250 (MP Biomedicals #04808274), 0.5 M ε-amino-n-caproic acid, 100 mM Bis-Tris pH 7.0) was added to the clarified supernatants. Samples were loaded alongside BN molecular weight marker (0.3 mg/mL thyroglobulin, 0.3 mg/mL ferritin, and 0.6 mg/mL BSA in solubilization buffer). Electrophoresis was performed at 4 °C, with BN anode buffer (50 mM Bis-Tris, pH 7.0) and BN cathode buffer (50 mM Tricine, 15 mM Bis-Tris, with 0.02% [w/v] Coomassie brilliant blue G-250 for a third of the gel before changing to cathode buffer without it).

Co-immunoprecipitation and mass spectrometry

Isolated mitochondria or whole cell samples were solubilized in solubilization buffer (20 mM Bis-Tris pH 7, 50 mM NaCl, 10% [v/v] glycerol, 1× protease inhibitor (Roche #11873580001)) containing 1% [w/v] digitonin (Merck #300410) at 2 mg/mL (45 min, 4 °C). After centrifugation at 16,000g (20 min, 4 °C) to pellet insoluble debris, supernatant was diluted with the solubilization buffer to a final concentration of 0.1% [w/v] digitonin before addition to pre-equilibrated anti-FLAG affinity gel (Millipore #A2220) for 60 min at 4 °C. Four washes of the anti-FLAG affinity gel with solubilization buffer containing 0.1% [w/v] digitonin were performed before elution of bound proteins with 0.2 M glycine (pH 2.5). The eluates were precipitated with 100% acetone at −20 °C overnight before downstream processing for mass spectrometry.

The acetone pellets were solubilized in urea/ABC buffer (8 M urea, 50 mM ammonium bicarbonate (ABC)) with 15 min sonication at RT. Solubilized samples were reduced and alkylated using 50 mM TCEP and 500 mM CAA, with 30 min incubation at 37 °C with 500 RPM shaking. The samples were then diluted to 2 M urea with 50 mM ABC before digestion with 1 µg trypsin (37 °C overnight). Trypsinized samples were acidified to 1% [v/v] TFA before clarification (21,000g, 10 min, RT). In-house stage-tips made with 3 M Empore SDB-XC extraction disks (Sigma-Aldrich #66884-U) were activated with 100 µL ACN and washed once with 0.1% [v/v] TFA, 2% [v/v] ACN before addition of clarified samples (2,500g, RT). Stage-tips were washed once with 0.1% [v/v] TFA, 2% [v/v] ACN before eluted in 0.1% [v/v] TFA, 80% [v/v] ACN. Eluates were dried with CentriVap Benchtop Vacuum Concentrator (Labconco) and reconstituted in 0.1% [v/v] TFA, 2% [v/v] ACN for analysis.

Phosphopeptides enrichment for mass spectrometry

Twenty-four hours transfected HeLa cells or 24 h tetracycline (1 µg/mL) induced Flp-In T-REx 293 stable cells lines were harvested. Whole cell pellets or mitochondria pellets were solubilized, alkylated, and reduced at 2 µg/µL in SDC buffer (4% [w/v] SDC, 100 mM Tris pH 8.1, 40 mM CAA, 10 mM TCEP). Samples were boiled at 99 °C (5 min, 1,500 RPM), followed by 15 min sonication. Centrifugation at 21,000g (5 min, 4 °C) clarified the samples and the resultant supernatant was precipitated in final 65% [v/v] acetone, 20% [v/v] methanol at −20 °C overnight. The precipitates were pelleted with 16,100g (10 min, 4 °C) and washed once with 100% [v/v] acetone. The pellets were then solubilized at 5 µg/µL in 6 M guanidine HCl (Sigma-Aldrich), 50 mM Tris, 5 mM TCEP with sonication (20 min, RT). The samples were diluted to 1 M guanidine HCl with 50 mM Tris (pH 8) and all replicates are normalized to the same concentration and volume before digested with trypsin (protein:trypsin = 50:1) overnight at 37 °C.

SPE cartridges (Oasis # WAT094226) were washed once with 80% [v/v] ACN, 0.1% [v/v] TFA, followed by two washes with 0.1% [v/v] TFA. Trypsinized samples clarified with 21,000g (10 min, RT) were acidified to 1% [v/v] formic acid before loaded onto the pre-equilibrated cartridges. The cartridges were then washed two times with 0.1% [v/v] TFA before eluted with 80% [v/v] ACN, 0.1% [v/v] TFA. Twenty micrograms equivalent of each eluate were dried with CentriVap Benchtop Vacuum Concentrator (Labconco) and reconstituted in 0.1% [v/v] TFA, 2% [v/v] ACN for proteomics analysis. The rest of the eluates were freeze-dried with Christ Alpha 3–4 LSCbasic prior to phosphopeptides enrichment.

Titanium dioxide beads (TiO₂; GL Sciences #5210-21315) were washed three times with washing buffer (50% [v/v] ACN, 5% [v/v] TFA) at 1,000 rpm (10 min, RT), followed by two washes (1,000 rpm, 5 min, RT) with loading buffer (2 M lactic acid, 5% [v/v] TFA, 50% [v/v] ACN). TiO₂ beads were then equilibrated with loading buffer for at least 10 min (1,000 rpm, RT). Freeze-dried peptides were reconstituted at 1 µg/µL with loading buffer (15 min sonication, RT) and clarified with 21,000g (5 min, RT). Pre-equilibrated beads were added to the samples with TiO₂:peptides ratio of 6:1 for a 1 h incubation (1,000 rpm, RT). The suspensions were passed through stage-tips made in-house with 3 M Empore C8 extraction disks with centrifugation at 1,000 g. The flowthroughs were loaded back to the stage-tips once more to maximize binding of phosphopeptides binding to TiO₂. The TiO₂ columns were washed twice with loading buffer, followed by three washes with washing buffer, before three elutions with elution buffer (1% [v/v] ammonium hydroxide) and a final elution with 30% [v/v] ACN. The eluates were acidified to 10% [v/v] formic acid before freeze-dried and reconstituted in 2% [v/v] ACN, 0.05% [v/v] TFA for phosphopeptides analysis.

LC MS/MS

LC MS/MS analysis of were performed either on an Orbitrap Eclipse, Orbitrap Ascend or Exploris 480 Mass Spectrometer (ThermoFisher Scientific). The LC systems were equipped with an Acclaim Pepmap nano-trap column (Dinoex-C18, 100 Å, 75 µm × 2 cm) and an Acclaim Pepmap RSLC analytical column (Dinoex-C18, 100 Å, 75 µm × 50 cm). The tryptic peptides were injected into the enrichment column at an isocratic flow of 5 μL/min of 2% [v/v] CH3CN containing 0.05% [v/v] TFA for 6 min, applied before the enrichment column was switched in-line with the analytical column. The eluents were 0.1% [v/v] FA (solvent A) in water and 100% [v/v] CH3CN in 0.1% [v/v] FA (solvent B), both supplemented with 5% DMSO. The gradient was at 300 nL/min from (i) 0–6 min, 3% B; (ii) 6–7 min, 3%–4% B; (iii) 7–82 min, 4%–25% B; (iv) 82–86 min, 25%–40% B; (v) 86–87 min, 40%–80% B; (vi) 87–90 min, 80%–3% B; (vii) 90–90.1 min, 80%–3% B and equilibrated at 3% B for 10 min before injecting the next sample.

For DDA experiments on an Orbitrap Eclipse, the mass spectrometer was operated in the data-dependent acquisition mode, whereby full MS1 spectra were acquired in a positive mode at 60,000 resolution. The ‘top speed’ acquisition mode with 3 s cycle time on the most intense precursor ion was used, whereby ions with charge states of 2–7 were selected. AGC target was set to standard with auto maximum injection mode. The ‘top speed’ acquisition mode with 3 s cycle time on the most intense precursor ion was used, whereby ions with charge states of 2–7 were selected. MS/MS analyses were performed by 1.6 m/z isolation with the quadrupole, fragmented by HCD with collision energy of 30%. MS2 resolution was at 15,000. Dynamic exclusion was activated for 30 s. AGC target was set to standard with auto maximum injection mode. Dynamic exclusion was activated for 30 s.

For DDA experiments on an Orbitrap Exploris 480, the mass spectrometer was operated in the data-dependent acquisition mode, whereby full MS1 spectra were acquired in a positive mode at 60,000 resolution. The ‘top speed’ acquisition mode with 3 s cycle time on the most intense precursor ion was used, whereby ions with charge states of 2–7 were selected. AGC target was set to standard with auto maximum injection mode. MS/MS analyses were performed by 1.6 m/z isolation with the quadrupole, fragmented by HCD with collision energy of 30%. MS2 resolution was at 15,000 Dynamic exclusion was activated for 30 s. AGC target was set to standard with auto maximum injection mode. Dynamic exclusion was activated for 30 s.

For DIA experiments on an Orbitrap Ascend, full MS resolutions were set to 120,000 at m/z 200 and scanning from 350 to 1400 m/z in the profile mode. Full MS AGC target was 250% with an IT of 50 ms. AGC target value for fragment spectra was set at 2000%. Fifty windows of 13.7 Da were used with an overlap of 1 Da. Resolution was set to 30,000 and maximum IT to 55 ms. Normalized collision energy was set at 30%. All data were acquired in centroid mode using positive polarity.

LC MS/MS data analysis

Raw mass spectra from co-immunoprecipitation were searched against a concatenated Uniprot human proteome database (canonical and reviewed, downloaded March 2021) and Uniprot fasta sequence of Lpg1625 (downloaded August 2023) using MaxQuant (version 2.1.4.0 and 2.5.2.0) [96]. Default label-free quantification (LFQ) parameters and match between runs were applied for the searches. The output “ProteinGroups.txt” was processed statistically with Perseus (version 1.6.10.43 and 2.0.5.0). Proteins that were “only identified by site”, “reverse”, and “contaminant” were removed from the dataset. The LFQ intensities were Log₂-transformed and grouped according to the experimental groups (n = 3). The data was filtered for ≥2 valid values in the experimental group only (e.g., ^FLAGLpg1625), and missing values in the control group (e.g., WT) were imputed from normal distribution (width = 0.3, and down shift = 1.8). Unpaired t tests with FDR 0.05 were performed and data was displayed as volcano plots. Mitochondrial proteins and/or ER proteins were annotated with MitoCarta3.0 database (downloaded October 2023) [55] and Human Protein Atlas (proteins with main subcellular localization to ER, downloaded July 2024) [97].

For proteome and phosphoproteome analyses, their raw mass spectra were searched by the directDIA algorithm on Spectronaut 19 (version 19.0.240604) and against the Uniprot human proteome database (canonical and reviewed, downloaded March 2021). Default BGS Factory Settings and BGS Phospho PTM Workflow were used for directDIA searches of global proteome and phosphoproteome respectively, except that “Precursor PEP Cutoff”, “Protein Qvalue Cutoff (Run)”, and “Protein PEP Cutoff” were set to 0.01. The exported reports from Spectronaut were processed on Perseus (version 1.6.10.43 and 2.0.5.0) similar as above. PG.Quantity and PTM.Quantity from the proteomics and phosphoproteomics datasets respectively were Log₂-transformed and grouped according to the experimental groups (n = 3 or 4). Proteomics data were filtered for ≥ n − 1 valid values in each group before unpaired two-sample t test with FDR 0.05 were performed and data presented as a scatterplot. Phosphoproteomics data was filtered to include only phosphopeptides with serine/threonine modified, and for ≥ n − 1 valid values in at least one group. Missing values in each group were imputed from normal distribution (width = 0.3, and down shift = 1.8) before unpaired two-sample t test with FDR 0.05 were performed and data presented as a scatterplot.

pLogo sequence motif analysis

“PTM.FlankingRegion” derived from the Spectronaut searches for phosphopeptides of interest was used as the input for the “foreground” (fg) and human protein was selected as the “background” (bg) for the analyses on https://plogo.uconn.edu/ (accessed October 2024).