Genome-Wide CRISPR Screen Identifies Phospholipid Scramblase 3 as the Biological Target of Mitoprotective Drug SS-31

Abstract

Key Points:

1. Szeto–Schiller-31–mediated mitoprotection is phospholipid scramblase 3–dependent.

2. Phospholipid scramblase 3 is required for recovery after AKI.

Background: The synthetic tetrapeptide Szeto–Schiller (SS)-31 shows promise in alleviating mitochondrial dysfunction associated with common diseases. However, the precise pharmacological basis of its mitoprotective effects remains unknown.

Methods: To uncover the biological targets of SS-31, we performed a genome-scale clustered regularly interspaced short palindromic repeats screen in human kidney-2, a cell culture model where SS-31 mitigates cisplatin-associated cell death and mitochondrial dysfunction. The identified hit candidate gene was functionally validated using knockout cell lines, small interfering RNA-mediated downregulation, and tubular epithelial–specific conditional knockout mice. Biochemical interaction studies were also performed to examine the interaction of SS-31 with the identified target protein.

Results: Our primary screen and validation studies in hexokinase 2 and primary murine tubular epithelial cells showed that phospholipid scramblase 3 (PLSCR3), an understudied inner mitochondrial membrane protein, was essential for the protective effects of SS-31. For in vivo validation, we generated tubular epithelial–specific knockout mice and found that Plscr3 gene ablation did not influence kidney function under normal conditions or affect the severity of cisplatin and rhabdomyolysis-associated AKI. However, Plscr3 gene deletion completely abrogated the protective effects of SS-31 during cisplatin and rhabdomyolysis-associated AKI. Biochemical studies showed that SS-31 directly binds to a previously uncharacterized N-terminal domain and stimulates PLSCR3 scramblase activity. Finally, PLSCR3 protein expression was found to be increased in the kidneys of patients with AKI.

Conclusions: PLSCR3 was identified as the essential biological target that facilitated the mitoprotective effects of SS-31 in vitro and in vivo.

Graphical abstract

Figure 1

CRISPR screen identifies PLSCR3 as a biological target of SS-31 in tubular epithelial cells. (A) Schematic representation of the screening approach. HK-2 cells were transduced with a pooled lentiviral library and 2 days later treated with PBS (control) or cisplatin or cisplatin+SS-31. At 24 hours, live cells were collected followed by gDNA extraction, DNA sequencing, and bioinformatic analysis to identify factors that drive the protective effects of SS-31. (B) Graphical representation of the data obtained in the primary screen. The plot shows the significance score (Cis+SS-31 versus Cis+Veh) of all gene depletion (minus log-transformed P values) as determined by the Kolmogorov–Smirnov test. The top hit was identified as PLSCR3 and is highlighted in blue. (C) CRISPR/Cas9-mediated PLSCR3 gene knockout was carried out in HK-2 cells. Representative immunoblot confirming gene knockout. (D and E) Control and PLSCR3 knockout HK-2 cells were treated with cisplatin or cisplatin+SS-31 followed by assessment of cellular viability (trypan blue) and caspase activity. The results showed that PLSCR3 knockout suppresses the protective effects of SS-31. In all the bar graphs (eight biologically independent samples), experimental values are presented as mean±SD. The height of error bar=1 SD, and P < 0.05 was indicated as statistically significant. One-way ANOVA followed by Tukey's multiple-comparison test was performed, and statistical significance is indicated by ***P < 0.001. Con, control; CRISPR, clustered regularly interspaced short palindromic repeats; FDR, false discovery rate; gDNA, genomic DNA; gRNA, guide RNA; HK-2, human kidney-2; KO, knockout; ns, not significant; PLSCR3, phospholipid scramblase 3; sgRNA, single guide RNA; SS, Szeto–Schiller; Veh, vehicle.

Figure 2

SS-31 mitigates cisplatin-associated mitochondrial dysfunction in a PLSCR3-dependent manner. (A) Whole cell, cytosolic, and mitochondrial fractions from parental HK-2 cells were used to examine the localization of PLSCR3. Representative blot showing that PLSCR3 is localized in the mitochondria. COX IV and USP10 were used as mitochondrial and cytosolic markers. (B) Characterization of the control and PLSCR3-KO HK-2 cells showed that PLSCR3 deficiency did not influence the viability (trypan blue), mitochondrial numbers (mitochondrial to nDNA ratio), or cardiolipin levels. (C–E) Control and PLSCR3-KO HK-2 cells were seeded in a SeahorseXF-24e analyzer and treated with vehicle, cisplatin, or cisplatin+SS-31 for 24 hours, and OCR was determined during sequential treatments with oligomycin A (A), trifluoromethoxy carbonylcyanide phenylhydrazone (FCCP) (B), and antimycin A/rotenone (C). The graph depicts the quantification of basal OCR and ATP production. SRC indicates spare respiratory capacity and MR indicates maximum respiration. In all the bar graphs (3–5 biologically independent samples), experimental values are presented as mean±SD. These experiments were repeated three times and yielded similar results. The height of error bar=1 SD, and P < 0.05 was indicated as statistically significant. One-way ANOVA followed by Dunnett's was performed, and statistical significance is indicated by ***P < 0.001. nDNA, nuclear DNA; OCR, oxygen consumption rate; USP10, ubiquitin-specific peptidase 10; WT, wild type.

Figure 3

Plscr3 protein expression is induced in tubular epithelial cells during cisplatin nephrotoxicity. Cisplatin nephrotoxicity was induced in 8- to 12-week-old male C57BL/6J mice by a single intraperitoneal injection (30 mg/kg). At 0–72 hours, serum and kidneys were collected for further analysis. (A–D) BUN, serum creatinine, kidney Ngal expression, and histological analysis showed a significant loss-of-kidney function and structural damage at 72 hours. In the representative histological images, the lower panel shows higher magnification as compared with the upper panel. (E and F) Western blot analysis of kidney cortical tissues showed that Plscr3 expression increases during the development of cisplatin-associated AKI. Short and long indicates the exposure time. The graph represents densitometric analysis from five independent samples. To label, isolate, and examine tubular epithelial cells, we crossed the Ggt1-Cre mice with ROSAmT/mG mice. The resulting transgenic mice express membrane-localized EGFP in renal tubular epithelial cells, whereas the other cell types express membrane-localized tdTomato. (G) A representative image showing EGFP expression in kidney tubular cells (arrow), whereas the cells within the glomerulus are tdTomato-positive (asterisks). (H) Cisplatin nephrotoxicity was induced in these transgenic mice, followed by isolation of GFP-positive and GFP-negative cells from the murine kidneys. Immunoblot showed that Plscr3 expression is induced specifically in renal tubular epithelial cells and not other cell types during cisplatin nephrotoxicity. The effectiveness of the cell isolation method was monitored by immunoblot analysis of renal tubular epithelial cell (GFP) and non–renal tubular epithelial cell (tdTomato) markers. In all the bar graphs (five biologically independent samples from three independent experiments), experimental values are presented as mean±SD. The height of error bar=1 SD, and P < 0.05 was indicated as statistically significant. One-way ANOVA followed by Tukey's multiple-comparison test was performed, and statistical significance is indicated by ***P < 0.001. Scale bar=100 µm. DAPI, 4′,6-diamidino-2-phenylindole; EGFP, enhanced green fluorescent protein; GFP, green fluorescent protein; Ngal, neutrophil gelatinase–associated lipocalin; ROS, reactive oxygen species; tDT, tdTomato.

Figure 4

In vivo siRNA-mediated Plscr3 knockdown abrogates the protective effects of SS-31 during AKI. Age-matched male (8–9 weeks) C57BL/6 mice were administered with intravenous injection (hydrodynamic) of control (nonspecific) or Plscr3-targeting siRNAs (25 μg in 0.5 ml of PBS). Two days after injection, kidney cortical tissues were collected for Western blot analysis of Plscr3 protein. (A and B) Representative blots showing more than 80% knockdown of Plscr3 by the targeted siRNA. Blots are representative of three independent experiments, all producing similar results. The graph represents results from the densitometric analysis of Plscr3 protein expression normalized to β-actin levels. (C) Schematic representation of experimental treatment strategy. Briefly, mice were injected with siRNA to knockdown Plscr3 gene, sequentially followed by SS-31, cisplatin, and SS-31 injections and examination of kidney injury at 72 hours. (D) BUN, (E) serum creatinine, (F) kidney cortical Ngal gene expression, and (F and G) kidney histological analysis show that Plscr3 knockdown did not influence cisplatin-associated kidney injury; however, it abrogated the protective effects of SS-31. In all the bar graphs (eight biologically independent samples from three independent experiments), experimental values are presented as mean±SD. The height of error bar=1 SD, and P < 0.05 was indicated as statistically significant. One-way ANOVA followed by Tukey's multiple-comparison test was performed, and statistical significance is indicated by ***P < 0.001. siRNA, small interfering RNA.

Figure 5

Renal tubular epithelial cell–specific Plscr3 gene ablation abrogates the protective effects of SS-31 during AKI. Ggt1-Cre mice were crossed with Plscr3-floxed mice to generate renal tubular epithelial cell–specific knockout mice (Plscr3^PT−/−). (A and B) Immunoblot analysis of Plscr3 protein in kidney cortical tissues showed successful gene knockout. Blots are representative of three independent experiments. The graph represents results from densitometric analysis of Plscr3 expression normalized to β-actin levels. (C) Eight- to 12-week-old male littermate control and Plscr3 conditional knockout mice were then challenged with cisplatin (30 mg/kg, single intraperitoneal injection) nephrotoxicity in the presence of SS-31 followed by examination of the severity of AKI using biochemical and histological analysis. (D–G) BUN, serum creatinine, and kidney Ngal gene expression (qPCR) showed that tubular Plscr3 deficiency results in abrogation of the protective effects of SS-31. In all the bar graphs (eight biologically independent samples from three independent experiments), experimental values are presented as mean±SD. The height of error bar=1 SD, and P < 0.05 was indicated as statistically significant. One-way ANOVA followed by Tukey's multiple-comparison test was performed, and statistical significance is indicated by ***P < 0.001. qPCR, quantitative PCR.

Figure 6

SS-31 binds and stimulates PLSCR3 activity in proteoliposome-based assays. (A) Schematic representation of SS-31 and PLSCR3 binding assay. Briefly, SS-31 conjugated beads were incubated with recombinant human PLSCR3 protein followed by pull-down experiments and subsequent immunoblot analysis. (B) Schematic representation of WT PLSCR3 protein and deletion mutants. (C) Recombinant WT and mutant PLSCR3 proteins were incubated with SS-31 conjugated beads, followed by pull-down, and immunoblot analysis with FLAG antibody. Representative blots showing that WT PLSCR3 protein can bind with SS-31. Deletion of the putative acetyl-CoA carboxylase biotin carboxyl carrier protein subunit domain abrogated the interaction between SS-31 and PLSCR3. The input blot shows the preincubation levels of the WT and mutant proteins. (D) Schematic representation of the proteoliposome-based PLSCR3 activity assay. (E) Proteoliposome-based PLSCR3 assay showed that SS-31 can increase PLSCR3 activity in a dose-dependent manner. Tat peptide had no effect on PLSCR3 activity. In all the bar graphs (eight biologically independent samples from three independent experiments), experimental values are presented as mean±SD. The height of error bar=1 SD, and P < 0.05 was indicated as statistically significant. One-way ANOVA followed by Tukey's multiple-comparison test was performed, and statistical significance is indicated by ***P < 0.001. Lipo, liposome; Mut, mutant; NBD, nitrobenzoxadiazole; PC, phosphatidylcholine.

Figure 7

Plscr3 plays a protective role during repeated low-dose cisplatin nephrotoxicity. (A) Schematic representation of the dosing strategy. For repeated low-dose cisplatin treatment, control and Plscr3 conditional knockout male mice were given four consecutive weekly injections of 9 mg/kg cisplatin, and blood samples and kidney tissues were collected for further analysis. (B and C) BUN and serum creatinine measurements showed that Plscr3 conditional knockout mice have more severe injury at 28 days. (D) Histological damage scoring of the kidneys from control and Plscr3 conditional knockout mice showed that Plscr3 plays a protective role under these conditions. (E and F) Western blot analysis of renal tissues at 28 days showed higher α-SMA accumulation in the Plscr3 conditional knockout mice. The image is a representative blot, and the graph represents densitometric analysis of α-SMA protein expression. In all the bar graphs (six biologically independent samples from three independent experiments), experimental values are presented as mean±SD. The height of error bar=1 SD, and P < 0.05 was indicated as statistically significant. One-way ANOVA followed by Tukey's multiple-comparison test was performed, and statistical significance is indicated by **P < 0.01, ***P < 0.001. α-SMA, alpha smooth actin; i.p., intraperitoneal injection.

Figure 8

PLSCR3 protein is increased in the kidneys of patients with AKI. Frozen biopsy samples from control (nephrectomy) and patients with AKI were used for Western blot analysis of PLSCR3 protein expression. (A and B) Representative Western blot results showing an upregulation in PLSCR3 protein expression in the kidneys from patients with AKI. The graph represents the densitometric analysis of PLSCR3 protein expression normalized to β-actin. (C) We propose that under stress conditions, SS-31–mediated PLSCR3 activation might result in favorable cardiolipin and/or phospholipid redistribution in the inner mitochondrial membrane, facilitating the stabilization of mitochondrial functions.