Figure 1.
Comparison of HLA class I stimulated EC cytoskeleton changes with other agonists.
Human aortic ECs were treated with either mIgG 1 µg/ml (A), HLA class I antibody (Ab) 1 µg/ml (B), thrombin 1 U/ml (C) or bFGF 25 ng/ml (D) for 10 min. Cells were stained with Texas-Red–phalloidin and analyzed by fluorescence microscopy. (E) The fluorescence intensity and the total area of each cell in each field were measured using ImageJ software (http://rsb.info.nih.gov/ij/). Three independent experiments were performed and in each experiment 3 fields were measured. The ratio between the mean fluorescence intensity and the cell area was calculated to account for changes in cell shape due to contraction. The data was normalized to the mIgG group and the others groups were calculated to reflect the change compared to the mIgG group. A student's t-test was performed to determine significant changes between the groups, P<0.05.
Figure 2.
Validation of the EC cytoskeleton isolation by Western blot.
(A) Human aortic ECs were treated for 10 min with either mIgG 1 µg/ml (lane 1), HLA class I Ab 1 µg/ml (lane 2), thrombin 1 U/ml (lane 3), bFGF 25 ng/ml (lane 4) or treated with cytochalasin D 5 µM (lane 5) for 30 min. The cytoskeleton was isolated from the lysates using tosylactived magnet Dynal beads and the fraction was analyzed by Western blot for components of the cytoskeleton. (B) Human aortic ECs were treated for 10 min with either mIgG 1 µg/ml (lanes 1 and 6), HLA class I Ab 1 µg/ml (lanes 2 and 7), thrombin 1 U/ml (lanes 3 and 8), bFGF 25 ng/ml (lanes 4 and 9) or treated with cytochalasin D 5 µM (lanes 5 and 10) for 30 min. The cytoskeleton was isolated using tosylactived magnetic Dynal beads. The cytoskeleton fraction (lanes 6–10) and the depleted fraction after cytoskeleton isolation (lanes 1–5) were immunoblotted to detect the presence of ß1-integrin and ß-actin as a loading control. The data presented in panels A and B are representative of 5 independent experiments.
Figure 3.
Directed acyclic graph (DAG) representing the enriched GO categories in the HLA class I stimulated EC cytoskeleton isolation.
The Gene Ontology (GO) enrichment analysis tool called Gene Onology Tree Machine (GOTM) was used to create the DAG. The gene list was created using the uniprot accession numbers and thus the gene ID type selected was “hsapiens_uniprot_swissprot_accession.” Next the reference for GOTM was selected as “hsapiens_genome.” The statistical method was set to the default “hypergeometric,” the multiple test adjustment was set to be done by Benjamini & Hochberg, the significance level was set to the top 10 and the minimum number of genes for a category was 2. The DAG is a re-creation of the graph generated by GOTM. The shaded boxes indicate categories that were not significantly enriched. The significantly enriched categories found in the HLA class I treated group are specified by the white boxes and include the adjusted p-values and the number of proteins found in each category. The categories found to be uniquely enriched only by HLA class I stimulation are indicated in bold.
Table 1.
12 Candidate Proteins Representing HLA Class I Induced Cytoskeleton Changes.
Table 2.
Predicted Kinase Families and the Names of Distinct Kinases and Candidate Target Phosphoproteins in Each Family.
Figure 4.
GO categories represented by the 70 kinases.
The 70 predicted kinases thought to have the potential to phosphorylate the 11 candidate phosphoproteins were annotated using GOanna and the annotations were categorizied using CateGOrizer which mapped the kinases based on EGAD2GO. The GO terms mapped to 47 EGAD2GO ancestor terms and this graph represents the top 17 definitions and the rest are grouped into the “other” category.
Figure 5.
DDX3X, TPM4 and eIF4A1 in the EC cytoskeleton fraction.
(A) The EC cytoskeleton was isolated for the 4 treatment groups, mIgG 1 µg/ml (lane 1), HLA class I 1 µg/ml (lane 2), thrombin 1 U/ml (lane 3), bFGF 25 ng/ml (lane 4) and ECs that were treated with cytochalasin D (5 µM, 30 min) (lane 5). The total cell lysate and the cytoskeleton fraction were subjected to SDS-PAGE separation followed by Western blot to examine the presence of MLC, DDX3X and TPM4 proteins. (B) The EC cytoskeleton isolation was isolated for the 4 treatment groups, mIgG 1 µg/ml (lane 1), HLA class I 1 µg/ml (lane 2), thrombin 1 U/ml (lane 3) and bFGF 25 ng/ml (lane 4). The total cell lysates, the depleted fractions and the cytoskeleton isolated fractions were subjected to SDS-PAGE separation followed by Western blot to examine ß-actin, total MLC and eIF4A1 protein levels in the different cell fractions. (C) The cytoskeleton isolation was performed on ECs treated with mIgG 1 µg/ml for 10 min (lane 1) or with cytochalasin D (lane 2) 5 µM for 30 min. The total cell lysates, depleted fractions and cytoskeleton isolated fractions were subjected to SDS-PAGE separation followed by Western blot to examine eIF4A1, total MLC and ß1-integrin in the cell fractions. The results presented in panel A and B are representative of 5 and 3 independent experiments, respectively.
Figure 6.
Colocalization of eIF4A1 and F-actin.
(A) The localization of eIF4A1 and F-actin in the EC was determined by confocal microscopy. Cell were treated and prepared for staining and microscopy. The scale bar is equal to 10 µm. Data is representative of 4 independent experiments. (B) eIF4A1 and F-actin colocalization was determined by the the ImageJ plugin Colocalization Finder (http://rsbweb.nih.gov/ij/plugins/colocalization-finder.html). Manders' Overlay coefficients were 0.916 (mIgG), 0.926 (HLA class I), 0.942 (thrombin) and 0.921 (bFGF). The data represents 4 independent experiments. (C) Intensities of the colocalization of 3 images per group were determined for each independent experiment (Avg ± SD): mIgG (3.8±1.7), HLA class I (15.9±2.3), thrombin (6.4±3.2) and bFGF (1.9±0.9). The colocalization intensity of eIF4A1 and F-actin in the HLA class I treated group was significantly increased compared to the unstimulated and bFGF groups; mIgG (p = 0.006), thrombin (p = 0.07) and bFGF (p = 0.002) as determined by student t-test.
Figure 7.
Colocalization of eIF4A1 and Paxillin.
(A) The localization of eIF4A1 and paxillin in the EC was determined by confocal microscopy. Cell were treated and prepared for staining and microscopy. The scale bar is equal to 10 µm. Data is representative of 4 independent experiments. (B) eIF4A1 and paxillin colocalization was determined by the ImageJ plugin Colocalization Finder. Mander's Overlap coefficients were 0.977 (mIgG), 0.929 (HLA class I), 0.933 (thrombin) and 0.888 (bFGF). The data represents 4 independent experiments. (C) Intensities of the colocalization of 3 images per group were determined for each independent experiment (Avg ± SD): mIgG (14.3±4.3), HLA class I (30.2±4.1), thrombin (19.1±7.8) and bFGF (11.7±3.0). The colocalization intensity of eIF4A1 and paxillin in the HLA class I treated group was significantly increased compared to the unstimulated and bFGF groups; mIgG (p = 0.01), thrombin (p = 0.09) and bFGF (p = 0.003) as determined by student t-test.