Sec61 inhibition alters endothelial cell morphology and adhesion.

HDMECs were exposed to 10 ng/mL mycolactone (Myco), 0.02% DMSO, 400 nM Ipomoeassin F or 20 nM ZIF-80. A-C. Mycolactone-treated cells were imaged at indicated times in (A) and cell numbers of each phenotype (i.e. normal, elongated or rounded) were counted and presented as a percentage of total cell number per field in (B) (mean ± SEM of 3 independent experiments, **, p<0.01, *, p<0.05. (C) Length and width of each cell exposed to mycolactone for 16 and 24 hours or DMSO for 24 hours per image were measured and presented as a ratio. Data is representative of 3 independent experiments. ****, p<0.0001 (D) A scratch was introduced to a HDMEC monolayer prior to the treatment and visualised at 0, 16, 24 hours. The scratch area is presented as a percentage of the value obtained at 0 hour (mean ± SEM of 3 independent experiments) ns, not significant, *, p<0.05. Cells exposed to an alternative Sec61 inhibitor ipomoeassin F (IpomF) (E) or ZIF-80 (F) were imaged at indicated times. Images are representative of 3 independent experiments. Scale bar = 100 μm. G-H. Live cell imaging was performed with the zenCELL Owl incubator microscope every 30 minutes over 48 hours. Algorithms of cell coverage (G), detached cell numbers (H) per time point from 3 independent experiments are summarised as mean ± SEM. Data presented as cell coverage relative to the value obtained from initial time point (G) or a % of detached cells to total cell number (H) of each condition. I. Confluent HDMECs were incubated on a rotary shaker (SHEAR) or not (STATIC) for 24hr before addition of 0.02% DMSO or 10ng/ml mycolactone and phase-contrast images were taken at 0, 24 and 48hr, within the shear-stress area where appropriate. Data is representative of 3 independent experiments. J. Cell numbers of each phenotype at different time points, presented as a percentage of total cell number per field in (I), showing mean of 3 independent experiments ± SEM, * p< 0.05. K. Length and width measurements of cells exposed to mycolactone or DMSO under shear stress conditions for 24 and 48 hours presented as a ratio. Data is representative of 3 independent experiments. ****, p<0.0001. Statistical analysis was performed by 2-way ANOVA (panel B and I) or 1-way ANOVA with Dunnett’s (panel C) or Tukey’s (panel D) correction in GraphPad Prism Version 9.4.1. Panel K was analysed using a mixed-effects model with Tukey’s correction for multiple comparisons.

Fibrinogen penetrates the vascular wall at early stages of infection.

Immunohistochemistry for fibrin(ogen) in the feet of C57BL/6 mice that received vehicle control (PBS) (A1-2) or intradermal injection of 1 x105 colony forming units M. ulcerans at 21 (Grade 1; B1-3) or 28 days (Grade 3; C1-3) post-infection. Positive fibrin(ogen) staining is brown in colour, the haematoxylin counterstain is purple. Scale bars in A1, B1 and C1: 2 mm; all others: 20 µm. Data representative of three mice at each infection grade, in two independent infections.

Proteomic analysis reveals loss of proteins associated with glycosylation, adhesion and migration.

(A) Workflow for isolation and proteomic analysis of HDMEC membrane proteins. In 3 independent replicates, HDMEC were exposed to 10ng/ml mycolactone or DMSO for 24 hours, lysed by hypotonic lysis and membrane fractions enriched by differential centrifugation as described in Methods. Acetone precipitated proteins were reduced, alkylated and trypsinised then subjected to TMT labelling for quantitative proteomic analysis by LC-MS/MS. (B) Volcano Plot of differential expression between DMSO and mycolactone treated samples, plotting mean fold change against false discovery rate adjusted p-values; orange= downregulated, p<0.05; blue= upregulated, p<0.05; black=p>0.05 (C) Pie charts showing subcellular localisation of proteins in total, >2 fold upregulated or downregulated (p<0.05) fractions. (D) Quantitation of membrane or secreted proteins according to type: blue = upregulated; white = unchanged; orange = downregulated. (E) Percentage of downregulated, unchanged and upregulated multi-pass membrane proteins possessing a signal peptide. (F) Overlap between mycolactone downregulated endothelial membrane proteome and Sec61-dependent proteome. Venn diagram created using JVenn [41], showing overlap in significantly downregulated proteome between the dataset presented here and those obtained in Hela cells treated with siRNA for Sec61α (Nguyen et al., 2018). (G) Top significantly over-represented (p<0.05) GO groups in downregulated and upregulated data sets, compared to whole genome. Data generated with WebGestalt. (H) Quantitation of numbers of up and down regulated proteins in GO groups identified in (G).

Endothelial glycosaminoglycan chain synthesis is blocked by mycolactone.

HDMECs exposed to 10 ng/mL of mycolactone (Myco) or 0.02% DMSO for 24 hours or indicated times were subjected to proteomic analysis (A-D), surface immunostaining (E-F, H) or immunoblotting (G). (A) Volcano Plot of differential expression between DMSO and mycolactone treated samples, plotting mean fold change against false discovery rate adjusted p-values. Pale blue=total detected proteins; dark blue=Type II membrane proteins; yellow=Golgi-localised Type II membrane proteins. (B) Violin plot showing fold change in protein levels for Type II membrane proteins grouped according to subcellular location. ns, not significant; **, p < 0.01; ****, p < 0.0001. (C) Heat map showing fold change in Golgi-localised O- and N-glycosylation enzymes in mycolactone exposed HDMEC. Dual-colour coding is shown., only one unique peptide detected in asterisks, and significantly downregulated (p < 0.05) or not (p ≥ 0.05) in bold or Italic respectively. (D) Genes in GAG biosynthesis categorised according to function and side chains of chondroitin sulphate/ dermatan sulphate (CS/DS), heparan sulphate (HS) or keratan sulphate (KS). Heatmap showing Log2 fold change of these genes in response to mycolactone in three independent experiments. Dual-colour coding is shown. Genes undetected are indicated as crossed, only one unique peptide detected in asterisks, and significantly downregulated (p < 0.05) or not (p ≥ 0.05) in bold or Italic respectively. (E-F) Cells were treated with or without chondroitinase ABC (ABC) or heparinase III (HepIII), immunostained with anti-chondroitin sulphate (CS), anti-Δ-heparan sulphate (dHS) antibodies or the isotype controls for flow cytometry analysis. Histogram plot for single cell population of CS (E) and dHS (F) and the respective mean fluorescence intensity (MFI) are shown. Unstained untreated cells filled grey; isotype control of untreated cells, dashed line in black; Untreated cells incubated with chondroitinase ABC prior to CS staining or without HepIII prior to dHS staining, grey line; untreated cells with CS-PE or dHS-PE, black line; cells exposed to DMSO stained with antibodies, blue line; cells exposed to mycolactone stained with antibodies, red line. MFI is presented as a % of untreated control (mean ± SEM of 3 independent experiments). **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. (G) Cells were lysed, treated with heparinase III and analysed by immunoblotting. HS neoepitopes were visualised with anti-Δ-heparan sulphate (dHS) antibody with the approximate migration of molecular weight markers in kDa. GAPDH as loading control. Images are representative of 3 independent experiments. (H) Cells were incubated with HepIII, fixed and immunostained with anti-dHS antibody (green), permeabilised and labelled with TRITC-conjugated phalloidin (magenta). Nuclei were stained with DAPI (blue). Images are representative of 2 independent experiments. Scale bar = 50 μm. Statistical analysis was performed 1-way ANOVA with Tukey’s (panel B) or Dunnett’s (panel E&F) correction for multiple comparisons in GraphPad Prism Version 9.4.1.

Loss of B3GALT6 affects endothelial cell morphology and monolayer permeability.

A. HDMECs exposed to 10 ng/mL mycolactone (Myco) or 0.02% DMSO for indicated times. B-G. HUVECs transfected with si-B3GALT6 or si-negative control (ctrl) oligos for 48 hours. (A-B) Cells were fixed, permeabilised and immunostained with anti-B3GALT6 and anti-GOLGB1 antibodies. B3GALT6 (green) and the Golgi apparatus (magenta) were visualised and nuclei stained with DAPI (blue). Scale bar = 50 μm (20 μm in the crop panels of A). Corrected total cell fluorescence of B3GALT6 in Golgi apparatus per cell was measured and presented as a value normalised to the mean value obtained from untreated control of each experiment. More than 30 cells per condition were measured per experiment. Images and quantification are representative of 3 independent experiments. (C-E) HUVECS exposed to 10 ng/mL mycolactone (Myco) or 0.02% DMSO for 24 hours one day post-transfection were imaged by an inverted microscope. (D) Length and width of each cell presented as a ratio. At least 100 cells were measured for each treatment. Values are representative of 3 independent experiments. (E) Rounded cell number per image presented as a % of total cell number per condition (values represent the mean ± SEM of 3 independent experiments). (F) Permeability of transfected HUVEC monolayers on inserts with 1 μm pores treated with 100 ng/mL IL-1β, 10 ng/mL mycolactone (Myco) or 0.02% DMSO for 24 hours was quantified. Fluorescence intensity of FITC-dextran in the receiver wells was measured and presented as a % where 100% is the value obtained from transwell lacking a cell monolayer, and 0% is untreated control wells (mean ± SEM of 3 independent experiments). (G) HUVECs were transfected with si-B3GALT6 or si-negative control (si-ctrl) oligos. A scratch was introduced to the monolayer prior to the treatment (10 ng/mL mycolactone (Myco) or DMSO) and live cell imaging was performed with the zenCELL Owl incubator microscope every 15 minutes for 30 hours. Migration time in hours (hrs) to reform the monolayer is presented as mean ± SEM (n = 3); wells with no visible monolayer at the end point were given a maximum value = 30. ns, not significant; *, P < 0.05; ****, P < 0.0001. Statistical analysis was performed using 1-way ANOVA with Dunnett’s correction for multiple comparisons in GraphPad Prism Version 9.4.1 (all analysed panels).

Mycolactone causes a rapid loss of multiple proteoglycans.

HDMECs exposed to 10 ng/mL mycolactone (Myco), 0.02% DMSO or 20 nM ZIF-80 for 24 hours or indicated times. (A) Heatmap showing representative data for genes encoding proteoglycans. Dual-colour coding for log2 fold change in response to Myco is shown. Possible attached glycosaminoglycan chains such as heparan sulphate (HS), chondroitin sulphate (CS) or dermatan sulphate (DS) Candidates with one unique peptide detected indicated with asterisks, significantly downregulated (p < 0.05) or not (p ≥ 0.05) in bold or Italic respectively. (B) Cells were harvested for flow cytometry analysis. Histogram plots for single cell population of perlecan, glypican-1 and biglycan. Unstained untreated cells, filled grey; isotype control of untreated cells, dashed black line. untreated cells stained with antibodies, black line; cells exposed to DMSO stained with antibodies, blue line; cells exposed to MYC stained with antibodies, red line. MFI is presented as a % of untreated control (mean ± SEM of 3 independent experiments). **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. (C) Cells were fixed and immunostained with anti-perlecan antibody (green), permeabilised and labelled with TRITC-conjugated phalloidin (magenta). Nuclei were stained with DAPI (blue). Images are representative of 3 independent experiments. Scale bar = 50 μm. Statistical analysis was performed using 1-way ANOVA with Dunnett’s correction for multiple comparisons in GraphPad Prism Version 9.4.1 (all analysed panels).

Mycolactone interferes with the interaction between endothelial surface receptors and the basement membrane.

HDMECs exposed to 10 ng/mL mycolactone (Myco), 0.02% DMSO or 20 nM ZIF-80 for 24 hours or indicated times. (A) Heatmap showing representative data for genes encoding junctional or adhesion molecules, basement membrane components and proteins involved in platelet adhesion. Dual-colour coding for log2 fold change in response to Myco is shown. Candidate with one unique peptide detected is indicated with asterisks, significantly downregulated (p < 0.05) or not (p ≥ 0.05) in bold or Italic respectively. (B) Cells were fixed and immunostained with anti-laminin α4 antibody (green), permeabilised and labelled with TRITC-conjugated phalloidin (magenta). Nuclei were stained with DAPI (blue). Images are representative of 2 independent experiments. Scale bar = 50 μm. (C) Cells were harvested for flow cytometry analysis. Histogram plots for single cell population of integrin β1, integrin β4 and laminin α5. Unstained, untreated cells, filled grey; isotype control of untreated cells, dashed black line. untreated cells stained with antibodies, black line; cells exposed to DMSO stained with antibodies, blue line; cells exposed to Myco stained with antibodies, red line. MFI is presented as a % of untreated control (mean ± SEM of 3 independent experiments). ns, not significant; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. (D) Alcian blue-periodic acid Schiff stain in the feet of C57BL/6 mice that received vehicle control (PBS) (D1, D4) or intradermal injection of 1*105 colony forming units M. ulcerans at 21 (Grade 1; D2, D5) or 28 days (Grade 2/3;I3, D3, D6-7) post-infection. Neutral glycans are indicated by purple staining and acidic glycans by light blue Blood vessel lumens are indicated by an “L”. Scale bars in D1-3: 2 mm; all others: 20 µm. E Representative transmission electron microscopy scan of grade 1 M. ulcerans infected murine footpad. The ultrathin section of glabrous skin shows the pericapillary interstitium slightly expanded, collagen (c) fibrils variably disaggregated and infiltrating macrophages (m). The endothelial cells (e) are reactive and exhibit cytoplasmic projections (*) whilst the basement membrane is multifocally disrupted (arrows). A thin layer of pericytes (p) is variably expanded by oedema. Scale bar: 2μm. Statistical analysis was performed using 1-way ANOVA with Dunnett’s correction for multiple comparisons in GraphPad Prism Version 9.4.1 (Panel C).

Laminin α5 ameliorates mycolactone-driven endothelial cell detachment and impaired migration.

Endothelial cells exposed to 10 ng/mL mycolactone (Myco) or 0.02% DMSO for 24 hours or indicated times. A. Treated HDMECs were harvested and re-seeded to laminin-511, 411, 111 or uncoated plates. After an hour, unbound cells were washed away and attached cells were imaged and cell numbers per field are presented as mean ± SEM of 3 independent experiments. B-D. Endothelial cells seeded onto laminin-511 or uncoated plates were exposed to mycolactone (Myco) or DMSO. (B-C) HDMECs were imaged every 30 minutes over 48 hours. Rounded or attached cells per condition were counted at 0, 24, 36, 48 hours. Data are presented as a % of total cell number of each condition (B) or normalised to the attached cell number counted at 0 hours (C) (mean ± SEM of 3 independent experiments). (D) A scratch was introduced to a HUVEC monolayer prior to treatment. The wounded area was imaged every 15 minutes for 24 hours. Cells migrating into the original scratch area were counted at 0, 8 and 16 hours. Data are presented as cell count per scratch area (mean ± SEM of 3 independent experiments). ns, not significant; *, P < 0.05; **, P < 0.01. Statistical analysis was performed in GraphPad Prism Version 9.4.1 using 2-way ANOVA with Tukey’s correction for multiple comparisons (all analysed panels; panels B&C also included the Geisser Greenhouse correction for sphericity).

Sec61 inhibition alters endothelial cell morphology.

A. Dose response of HDMEC to mycolactone produced synthetically (Myco-Syn) or purified from natural sources (Myco-Bio). Cells were exposed in triplicate to serial dilutions of compound for 96hr and viability measured with resazurin. Results are expressed as a percentage of the DMSO control. Data represents the mean of 3 independent experiments ± SEM. B. Cells treated with 10ng/ml synthetic (Syn) or natural (Nat) or 0.02% DMSO were imaged using the ZenCELL owl at indicated times. Scale bar =50μ. Images are representative of three independent experiments C.Cell numbers of each phenotype were counted and presented as a percentage of total cell number per field in (B) (mean ± SD for triplicate wells) *, p<0.05; **,p<0.01, n=1. D. Length and width of cells exposed to mycolactone for 16 and 24 hours or DMSO for 24 hours measured and presented as a ratio. **,p<0.01; ****, p<0.0001, Data is representative of duplicate independent experiments. E-F. Analysis determined by zenCell Owl inbuilt algorithms for images collected over 48hrs of cell coverage (E) and detached cells (F). Data represents the mean ± SD for triplicate wells, Data is representative of duplicate independent experiments.

Profile of Buruli ulcer mouse footpad model.

C57BL/6 mice receiving 1*105 colony forming units M. ulcerans Mu_1082 strain injected intradermally into the footpad were sacrificed at 21or 28 days. Injected feet were fixed, decalcified and embedded in paraffin. (A) Sections stained for acid fast bacilli with Ziehl Neelson stain from mice infected for 21 days (A1) or 28 days (A2). At 21 days immune cell infiltration as well as apparently intracellular bacteria can be clearly seen in proximity to mycobacterial clusters (B) M. ulcerans mouse feet infected stained with isotype control antibody. Scale bars: 20 µm.

Proteomic analysis reveals loss of proteins associated with glycosylation, adhesion and migration.

(A) Heat map showing fold change detected in this dataset for previously validated endothelial cell mycolactone targets (Ogbechi et al, 2015, Hsieh et al 2022). Dual-colour coding is shown., significantly downregulated (p < 0.05) or not (p ≥ 0.05) in bold or Italic respectively. (B) Significant association between multipass protein membrane protein signal peptide (SP) ΔG values and level of downregulation by mycolactone (p<0.05). Membrane protein type and SP sequence based on Uniprot data, ΔG quantified using ΔG Prediction Server V1.0 (https://dgpred.cbr.su.se). (C). Impact of distance between signal peptide and first transmembrane domain on susceptibility of multipass proteins to downregulation by mycolactone. Statistical analysis was performed by one-way ANOVA with Tukey’s correction for multiple comparisons in GraphPad Prism Version 9.4.1. *, p<0.05 (D) Model depicting assisted and unassisted channel opening by signal peptides. (i) Binding of a nascent chain signal peptide to Sec61α opens a lateral gate into the membrane and causes a shift in the position of the plug domain, allowing access to the ER lumen. (ii) TRAP increases translocation of proteins whose signal peptides bear a high GP content. TRAP is a heterotetrameric complex which interacts with the ribosome on the cytosolic side of the ER membrane and with Sec61α on the luminal side, binding to a hinge region between the N- and C-terminal halves of the protein facilitating channel opening [1, 2] (iii) The Sec62/63 complex is involved in post-translational translocation but can also assist opening of the channel for proteins with signal peptides that gate slowly due to the presence relatively long but less hydrophobic “H-regions” and lower carboxy terminal polarity [3]. This complex also interacts with the ribosome and the translocon and may also interact directly with the nascent peptide chain [4]. In addition, Sec63 recruits BiP to the translocon to further assist channel opening on the luminal side [3]. (E) Overlap between mycolactone-downregulated endothelial membrane proteome and translocon-dependent proteome. Venn diagram created using JVenn [5] showing overlap in significantly downregulated proteome between the dataset presented here and those obtained in Hela cells treated with siRNA for Sec61α and translocon associated protein TRAP or HEK293 cells with Sec62 or Sec623 knocked out [1, 3]. (F) Top significantly over-represented (p<0.05) Gene Ontology groups in upregulated data set, compared to whole genome. Data generated with WebGestalt.

Mycolactone targets Golgi proteins involved in glycosaminoglycan chain synthesis initiation.

(A) Downregulated intracellular proteins in HDMECs following 24 hours exposure to 10 ng/mL mycolactone, according to subcellular location, presented as percentage of total. (B) Impact of membrane anchor ΔG values on fold change in expression induced by mycolactone for all identified Type II membrane proteins. (C) Cartoon representing mycolactone-downregulated steps of GAG enzymatic synthesis. Each intermediate product is shown in blue and alongside with its responsible enzyme(s). Candidates significantly downregulated (p < 0.05) by mycolactone are shown in red, unchanged in black and undetected in grey. Unique peptide numbers detected in proteome are in purple (underlined).

Sec61 blockade suppresses B3GALT6 expression in endothelial cells.

A. HDMECs exposed to 10 ng/mL synthetic (Myco-Syn) or natural mycolactone (Myco-Bio) or 0.02% DMSO for 24hr were fixed, permeabilised and immunostained with anti-B3GALT6 and anti-giantin antibodies. B3GALT6 (green) and the Golgi apparatus (magenta) were visualised and nuclei stained with DAPI (blue). Scale bar = 50 μm (20 μm in the crop panels). Corrected total cell fluorescence of B3GALT6 in Golgi apparatus per cell measured and presented as a value normalised to the mean value obtained from untreated control of each experiment. At least 60 cells per condition were measured. Data is representative of duplicate experiments. ****, p<0.0001. (B) HDMECs exposed to 10 ng/mL of mycolactone (Myco), 0.02% DMSO, 20 nM ZIF-80 or untreated for 24 hours were fixed, permeabilised and immunostained with anti-B3Galt6 and anti-giantin antibodies and nuclei stained with DAPI. Images are representative of 2 independent experiments. Scale bar = 50 μm. Statistical analysis was performed by one-way ANOVA with Dunnett’s correction for multiple comparisons in GraphPad Prism Version 9.4.1

Mycolactone’s effect on endothelial surface proteoglycans.

HUVECs exposed to 10 ng/mL of mycolactone (Myco), 0.02% DMSO or remained untreated for indicated times. Cells were harvested for flow cytometry analysis. Histogram plots for single cell population of perlecan and glypican-1. Unstained, untreated cells, filled grey; isotype control of untreated cells, dashed black line. untreated cells stained with antibodies, black line; cells exposed to DMSO stained with antibodies, blue line; cells exposed to mycolactone for 2, 6, and 24 hours stained with antibodies, pink, orange and red line, respectively. MFI is presented as a % of untreated control (mean ± SEM of 3 independent experiments). B. HDMECs exposed to 10 ng/mL synthetic (Syn) or natural mycolactone (Nat) or 0.02% DMSO for 24 hours. Cells were fixed and immunostained with anti-perlecan antibody (green), permeabilised and labelled with TRITC-conjugated phalloidin (magenta). Nuclei were stained with DAPI (blue). Scale bar = 20 μm, n=1. C. Confluent HDMEC incubated on a rotary shaker for 24hr before addition of 0.02% DMSO or 10ng/ml mycolactone (Myco) for 48hr were fixed and stained with anti-perlecan antibody (green) and counterstained with DAPI (blue). Scale bar = 10 μm, n=1.. Statistical analysis was performed by one-way ANOVA with Dunnett’s correction for multiple comparisons in GraphPad Prism Version 9.4.1. ns, not significant; **, P < 0.01; ***, P < 0.001.

Mycolactone impacts endothelial cell adhesion molecules.

A. HDMECs exposed to 10 ng/mL synthetic (Myco-Syn) or natural mycolactone (Myco-Bio) or 0.02% DMSO for 24 hours. Cells were fixed and immunostained with anti-LAMA4 antibody (green), permeabilised and labelled with TRITC-conjugated phalloidin (magenta). Nuclei were stained with DAPI (blue). Scale bar = 20 μm, n=1. B. Confluent HDMEC incubated on a rotary shaker for 24hr before addition of 0.02% DMSO or 10ng/ml mycolactone (Myco) for 48hr were fixed and stained with anti-LAMA4 antibody (green) and counterstained with DAPI (blue). Scale bar = 10 μm, n=1. C-D. HDMECs exposed to 10 ng/mL of mycolactone (Myco), 0.02% DMSO or remained untreated for indicated times. Cells were lysed and subjected to immunoblotting with anti-fibronectin (C) and anti-integrin α5 D) antibodies. Each immunoblot intensity was normalised according to GAPDH and untreated controls. Statistical analysis was performed by one-way ANOVA with Dunnett’s correction for multiple comparisons in GraphPad Prism Version 9.4.1Data from 3 independent experiments are presented (mean ± SEM). ns, not significant; **, P < 0.01; ****, P < 0.0001.

Endothelial cell adhesion to various laminin isoforms.

(A) HDMECs were harvested and layered to laminin-511, 411, 111 or uncoated wells for one hour. Non-adherent cells were washed away and attached cells per field were counted. mean ± SEM (n = 3). (B) HDMECs seeded onto different laminin isoforms were untreated or exposed to 0.02% DMSO or 10 ng/mL mycolactone (Myco) for 48 hrs. Their viability was assayed using CellEvent detection kit as described in [6]. The number of live cells (negative for both active caspase 3/7 and PI) in three fields was determined and expressed as a proportion of total cells (Mean ± SEM, n = 3 independent experiments). Statistical analysis was carried out using GraphPad Prism Version 9.4.1. Panels A&D used a repeated measures one-way ANOVA with Dunnett’s correction for multiple comparisons (D also included the Geisser Greenhouse correction for sphericity). Panels B,C&E used a two-way ANOVA with Tukey’s correction for multiple comparisons (panels B&C also included the Geisser Greenhouse correction for sphericity). ns, not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001.