A mitochondrial contribution to anti-inflammatory shear stress signaling in vascular endothelial cells

Figures

• 

  Figure 1.. CRISPR screen to identify regulators of Klf2 expression. (a) The lentiviral Klf2:GFP reporter expresses d2EGFP (half-life of ∼2 h) driven by the Klf2 promoter. The vector also contains mCHERRY controlled by the PGK promoter as an internal control. MAECs expressing the reporter showed a uniform, ∼10-fold increase in green/red ratio after 16 h LSS. A distribution plot of the GFP/mCherry ratio (normalized GFP intensity) indicates approximate location of the FACS-sorting gates used for CRISPR screening. (b) Lentiviral vectors express Cas9 and one of 130,000 sgRNAs, with 6 sgRNAs targeting each gene and 1,000 nontargeting control sgRNAs via EFS and U6 promoters, respectively. (c) FACS gates were positioned according to GFP/mCherry percentiles to achieve a normal distribution. (d) Mek5 sgRNAs induce a leftward shift in the distribution among gates compared to nontargeting sgRNAs, indicating reduced reporter expression. The shaded areas indicate the difference from control for each guide. These values were weighted by the relative abundance of each sgRNA. (e) Plot of Z scores for all genes ≤0. Candidate genes among the lower fifth percentile are labeled in blue (complete list in Table…

• 

  Figure S1.. CRISPR screening and production of sequencing libraries. (a) Time-lapse images for EGFP in MAEC Klf2:d2EGFP reporter cells stimulated with 18 dynes/cm2 LSS. Images presented in inverted grayscale for enhanced contrast. n = 2. (b) Induction of MAEC Klf2:d2EGFP reporter cells compared to endogenous KLF2 in MAECs stimulated with LSS, assayed by qPCR. n = 3. (c) Endogenous Klf2 induction by LSS in MAECs assayed by RNAscope. Images are presented in inverted grayscale for enhanced contrast. n = 2. Scale = 50 μm. (d) Klf2 reporter cells’ fluorescence intensity distribution across seven FACS gate boundaries was graphed from three replicate experiments to reveal reproducibility. Dashed lines are means of n = 3, and blue error bars represent SD. (e) Strategy utilized to isolate the seven gates described in Fig. 1 c. (f) Quantification of the total gated and recovered fractions of FACS-sorted cells. (g) Lentiviral cassettes were extracted from genomic DNA and resolved by agarose gel electrophoresis. (h) Adaptors used for second and third round amplification of sgRNA cassettes. Round 2 amplification used a nested PCR strategy that labels the sgRNA with a random barcode and FACS gate ID…

• 

  Figure S2.. The MAPK cascade regulating Klf2 induction. (a) Domain architecture and interactions in the MEKK2/3–MEK5–ERK5 complex. MEKK2/3 and MEK5 PB1 domains’ basic and acidic patches dimerize to promote signaling. (b) Representative shear-induced gel-shifts for MEKK2, MEKK3, and ERK5. The lower and gel-shifted upper bands are indicated with black and red arrows, respectively. (b) Lysates from MAECs with or without 18 dynes/cm2 shear for 0.3 h were treated with alkaline phosphatase or phosphatase inhibitor as indicated. Gel shifts were quantified by densitometry. n = 3. (c) Time course of MEKK3 and ERK5 activation and Klf2 reporter induction in shear-stimulated MAECs. (d) Quantification of MEKK2, MEKK3, and ERK5 activation in shear-stimulated MAECs at various times by gel shift assay. Data are representative of n = 3. (e) Reproducibility of shear-stimulated ERK5 activation dynamics determined by gel shift assays. n = 5. (f) Validation of MEKK2 and MEKK3 KO. (g and h) Representative blots and quantification of ERK5 activation from control, MEKK2, and/or MEKK3 KOs stimulated with LSS for the indicated times. n = 3. (i) ERK5 staining showing nuclear translocation after shear. (j)…

• 

  Figure S3.. p62 assembles a p62–MAPK complex. (a) HUVECs were transduced with GFP-p62 lentivirus, subjected to LSS for 3 h, fixed, and stained for endogenous MEKK2 and MEK5. Arrows: three-channel colocalization in dense p62 + punctae. (b) HUVECs with or without p62 knockdown and expressing MEK5-Flag were subjected to anti-Flag immunoprecipitation and Flag peptide elution. Eluted proteins were subjected to anti-Flag and anti-p62 immunoblotting. (c) Schematic for isolation of MER5–p62 complex by dual IP. HUVECs expressing MEK5-Flag and GFP-p62 were subjected to LSS for 3 h then IPed with anti-FLAG or anti-GFP resin. One half of the Flag IP was eluted with Flag peptide and IPed with anti-GFP. The resultant MEK5–p62 complex was analyzed by mass spectrometry to identify and quantify interaction partners. The number of unique peptides and percent protein coverage from each group are shown. (d) HUVECs expressing MEK5-Flag and WT or mutant GFP-p62 constructs were IPed with anti-Flag, eluted with Flag peptide, and analyzed by immunoblotting. (e) MAEC were exposed to LSS for the indicated times and lysates immunoblotted for p62 and actin as a loading control. Graph: quantified p62 levels from…

• 

  Figure 2.. Laminar shear-induced mitochondrial remodeling. (a) MAECs treated with 2 mM NAC or 2 μM MitoQ were subject to 18 dynes/cm2 LSS for the indicated times and ERK5 activation assayed by the Western blot and gel-shift assay. Values are means ± SEM, n = 3. (b) HUVECs expressing the roGFP2 oxidative stress biosensor localized to the indicated compartments were subjected to 18 dynes/cm2 LSS and fluorescence monitored. Values are means ± SEM, n ≥ 6. (c) HUVECs expressing the mito-Grx1-Rogfp2 biosensor were treated with or without 5 μM BAPTA-AM and subjected to 18 dynes/cm2 LSS. Mitochondrial-roGFP fluorescence was monitored as in panel b. N = 2. (d) HUVECs expressing the mito-Grx1-Rogfp2 biosensor with or without 10 μM KB-R7943 were subjected to 18 dynes/cm2 LSS and mitochondrial-roGFP fluorescence monitored as in panel b. n = 2. (e) LSS-stimulated HUVECs were fixed and stained for the mitochondrial marker TOMM20. Mitochondrial morphology was then scored in ∼500 cells/experiment. N = 3 independent experiments. (f) HUVECs were exposed to LSS for the indicated times, fixed, and stained for LC3b and TOMM20. The fraction of cells exhibiting mitochondrial-LC3 colocalization was quantified by…

• 

  Figure S4.. Role of autophagy and mitochondria in regulating the ERK5 pathway. (a) HUVECs were subjected to 18 dynes/cm2 LSS for the indicated times, stained for TOMM20, imaged using confocal microscopy at 63× magnification, and mitochondrial length quantified. Values are means ± SD; N = 110–200 cells. (b) MAEC mitochondrial DNA content was assayed by qPCR at the indicated times. (c) To assay mitophagic flux, MAECs were subjected to LSS for the indicated times ± BafA1 for the final 2 h. Lysates were then immunoblotted for full length PINK1 (∼62 kD) and actin as a loading control. BafA1-dependent accumulation serves as a metric for mitophagy. LC3b-II accumulation is shown as a positive control for BafA1 treatment. (d) LC3b-II accumulation after PINK1-KO, representative of three experiments. (e) HUVECs expressing mito-Keima were subjected to 12 h LSS with or without KB-R9743. The mito-lysosomal fraction was then quantified as in Fig. 2 i. n = 3. (f–k) MAECs treated with inhibitors of ULK1 (MRT68921, 1 μM), PIKFYVE (YM201636, 200 nM), or vATPase (BafA1, 100 nM) were exposed to LSS for the indicated times, then lysates analyzed for ERK5 activation or Klf2 reporter induction. (l) MAECs were…

• 

  Figure 3.. The mitochondrial pathway enhances ERK5 coupling. (a and d) Klf2:GFP reporter induction after 16 h LSS stimulation of control, EMRE-KO, or PINK1-KO cells assayed by immunoblot. (b and e) ERK5 activation index of control, EMRE-KO, and PINK1-KO cells stimulated with LSS for the indicated times. (c and f) MEKK2 activation index of control, EMRE-KO, and PINK1-KO cells stimulated with LSS for the indicated times, assayed by Western blot gel-shift. (g) MAECs treated with the indicated inhibitors were subjected to 16 h shear and assayed for activation of MEKK2 and ERK5. NAC, an antioxidant; MitoQ, a mitochondrial-targeted antioxidant; BAPTA-AM, a calcium chelator; KB-R9743, an inhibitor of the mitochondrial calcium uniporter. n = 3 or 5 for KB-R9743. (h) Mouse arterial endothelium was examined in the IC of the ascending aorta and the straight portion of the TA. (i and j) Aortas from WT and PINK1-KO mice were stained for ERK5 (i) and the endothelium was imaged en face (j) before quantifying mean nuclear translocation in the IC and TA. n = 5. (k and l) Klf2 mRNA within TAs of WT and PINK1-KO mice was labeled using RNAscope, counterstained with DAPI, and imaged en face. Klf2 mRNA labeling…

• 

  Figure 4.. p62 is activated by LSS and enhances Klf2 induction. (a) Domain organization of MEKK2/3, MEK5, and p62. MEKK2/3 has Type II PB1 domains whereas MEK5 has a Type I PB1 domain. p62 has a Type I/II PB1 domain, KEAP1 and LC3 binding sites (KIR and LIR), and a ubiquitin binding domain (UBA). (b) Control or p62-KO MAECs were reconstituted with sgRNA-resistant p62WT-GFP or p62 with a mutant PB1 domain basic patch (p62KRR-AAA-GFP) and assayed for Klf2:GFP induction by LSS. Data points represent summary data from biological replicates. (c) MAECs were transduced with control or Nrf2 sgRNAs and stimulated with 16 h LSS. Lysates were then immunoblotted for p62, NRF2 and actin. Quantified values are means ± SEM; n = 3. (d) MAECs were treated with LSS and/or NAC as indicated. Active p62 was then purified with an anti-p62-UBA antibody and quantified by immunoblotting. Representative blots and quantification from three independent experiments are shown. (e) HUVEC transduced with the mitochondrial marker, mCHERRY-FIS1TMD, were subjected to shear, fixed, and stained for p62 and MEK5. (f and g) The Mander’s overlap coefficients of MEK5 colocalizing with p62 (f) and p62 colocalizing with…

• 

  Figure S5.. PIEZO1 and the Klf2 pathway. (a) CRISPR KO of PIEZO1 demonstrated by Sanger sequencing of genomic DNA. Analysis using the Synthego ICE toolbox reveals 95% of the alleles have indels. (b) MEKK2 and ERK5 activation after PIEZO1-KO quantified from gel shift assay. Values are means ± SEM, n = 3–5. (c) Quantification of Klf2:GFP induction after 16 h LSS, in control and PIEZO1-KO. Values are means ± SEM; n = 3. (d) MAECs treated with 1 μM Yoda1, a PIEZO1 agonist, for the indicated times, were analyzed by immunoblotting for ERK5 activation and Klf2:GFP induction. n = 3. *, Ρ < 0.05 by Student's t test or one-way ANOVA (b and c). Source data are available for this figure: SourceData FS5.

• 

  Figure 5.. The mitochondrial pathway distinguishes high laminar from low/disturbed shear stress. (a) Klf2:GFP reporter MAECs were exposed to either 18 dynes/cm2 LSS or 1 ± 3 dynes/cm2 OSS for 16 h. (b–d) We then assayed GFP reporter levels, MEKK2 and ERK5 gel-shift (b and c), and ERK5 nuclear translocation (d). Values are means ± SEM; n = 7–11. (e) HUVECs were exposed to LSS or OSS for 16 h and mitochondrial fragmentation was assayed as in Fig. 2 e. (f) HUVECs were exposed to LSS or OSS for 16 h and mitochondrial DNA assayed as in Fig. 2 h. (g) HUVECs were transduced with mito-Keima and stimulated with LSS or OSS for 12 h before imaging as in Fig. 2 i. The mean mito-lysosomal fraction was quantified and graphed as means ± SEM with OSS normalized to 1.0; n = 5 independent experiments. (h) Aortas from adult GFP-LC3b transgenic mice were fixed, stained for GFP and TOMM20, and imaged en face to visualize ECs within the IC and TA. Confocal stacks were obtained and discrete LC3b+/TOMM20+ objects were counted. Images show representative 63× fields; quantified values (below) are means ± SEM, n = 4. (i and j) MAECs were exposed to either 18 dynes/cm2 LSS or 1 ± 3 dynes/cm2 OSS for 16 h before…

• 

  Figure 6.. Arterial remodeling. (a) WT and PINK1-KO male mice were subjected to femoral arterial ligation and superficial blood flow measured by laser Doppler imaging at −1, +1, +3, +7, and +14 d. Representative images are shown. (b) Quantification of blood flow; values are means ± SEM. N = 5 WT and 7 PINK1-KO mice. *, Ρ = 0.05 by paired Student's t test. (c) 14 d after femoral arterial ligation, hindlimbs were subjected to microCT imaging. Representative images are shown. (d) Vessel segments intersecting with Z-planes running through the thigh and calf regions were counted and binned according to their cross-sectional diameters. N = 5 WT and 7 PINK1-KO mice. *, Ρ < 0.05 by Students' t test.