Key role for Rac in the early transcriptional response to extracellular matrix stiffness and stiffness-dependent repression of ATF3

ABSTRACT The Rho family GTPases Rac and Rho play critical roles in transmitting mechanical information contained within the extracellular matrix (ECM) to the cell. Rac and Rho have well-described roles in regulating stiffness-dependent actin remodeling, proliferation and motility. However, much less is known about the relative roles of these GTPases in stiffness-dependent transcription, particularly at the genome-wide level. Here, we selectively inhibited Rac and Rho in mouse embryonic fibroblasts cultured on deformable substrata and used RNA sequencing to elucidate and compare the contribution of these GTPases to the early transcriptional response to ECM stiffness. Surprisingly, we found that the stiffness-dependent activation of Rac was dominant over Rho in the initial transcriptional response to ECM stiffness. We also identified activating transcription factor 3 (ATF3) as a major target of stiffness- and Rac-mediated signaling and show that ATF3 repression by ECM stiffness helps to explain how the stiffness-dependent activation of Rac results in the induction of cyclin D1.

), and a similar number of ROSA26 (control) clones were serum-starved, pre-treated with 10 µM EHT1864 in suspension, and plated on FNcoated tissue culture dishes with 10% FBS-DMEM in the continued presence of EHT1864 for 5 hr before determination of ATF3 protein levels by immunoblot.GAPDH was the loading control.Table S2.Identification of putative ATF3-depleted MEF clones.MEF clones from the Crispr transfection were lysed in QuickExtract™ DNA Extraction buffer (Lucigen).Genomic DNA was extracted by two cycles of vortexing (15 sec) and heating (first at 68°C for 15 min and then at 95°C for 10 min).The DNA was subjected to PCR amplification, the purified PCR products were sequenced using a reverse ATF3 primer for gRNA-1 (GGTGCACACTATACCTGCTC), and the sequences were analyzed using the ICE CRISPR Analysis Tool (Synthego).A Knockout (KO) Score was generated for each clone.
The KO score represents the proportion of cells that have either a frameshift or 21+ bp indel.We selected several clones with KO scores >90 for further analysis.Clone 1-5 was not used in this study.

Fig. S2 .
Fig. S2.Effects of ECM stiffness, EHT1864 and CT04 on ATF3 protein.Starved MEFs were plated on soft or stiff FN-coated hydrogels with 10% FBS and either DMSO (vehicle control; Cntl), EHT1864, or CT04 for 5 hr.(A) The cells were analyzed by immunoblot for ATF3 with GAPDH as the loading control.(B) Quantification of immunoblot results in A; the graph shows mean + SD with results normalized to GAPDH and plotted relative to the normalized ATF3 signal on the soft hydrogels; n=4-5.Statistical significance was determined by 2-tailed, paired t-tests.

Fig. S3 .
Fig. S3.Inverse effects of ECM stiffness on ATF3 and cyclin D1 protein.Starved MEFs were plated on soft or stiff FN-coated hydrogels with 10% FBS for 24 hr.(A) The cells were analyzed by immunoblot for ATF3 and cyclin D1 with GAPDH as the loading control.(B) Quantification of immunoblot results in A; the graph shows mean + SD with results normalized to GAPDH abundance and plotted relative to the normalized signal on the soft hydrogels; n=3.Statistical significance was determined by 2-tailed, paired t-tests.

Fig. S4 .
Fig. S4.Involvement of Arp2/3 in the regulation of ATF3 and cyclin D1 mRNAs.Starved MEFs were plated on soft or stiff FN-coated hydrogels with 10% FBS for 9 hr in the presence of DMSO, CK666 (Arp2/3 inhibitor) at 100 µM or 150 µM, or the inactive analog (CK689) at 150 µM.(A-B) mRNAs were extracted, and the levels of ATF3 and cyclin D1 mRNA were quantified by RT-qPCR.Results show mean + SE (n=4-6) and are normalized to the soft hydrogels.Statistical significance was determined by oneway ANOVA; asterisks show the results of Dunnett's post-tests performed relative to the soft hydrogel.

Fig. S5 .
Fig. S5.Identification and characterization of ATF3 KO MEFs.(A) Several putative ATF3 deletion clones, including those identified by the ICE Crispr Analysis Tool (Table S2), and a similar number of ROSA26 (control) clones were serum-starved, pre-treated with 10 µM EHT1864 in suspension, and plated on FNcoated tissue culture dishes with 10% FBS-DMEM in the continued presence of EHT1864 for 5 hr before determination of ATF3 protein levels by immunoblot.GAPDH was the loading control.Red arrowheads show several ROSA and ATF3 KO clones with the expected behavior for ATF3.(B) Images of control and ATF3 KO MEFs that had been serum-starved and plated on stiff FN-coated hydrogels in DMEM with 10% FBS for 9 hr before fixation and staining with Alexa Fluor™ 594 Phalloidin.(C-D) Quantitative analysis of cell area and phalloidin intensity per cell in the ROSA26 controls (R3, R11, R12, R15) and ATF3 KO clones (1-20, 1-29, 1-44, 1-49) as determined using Image J. Graphs show ~170-370 cells analyzed per condition, accrued from 2 independent experiments.Statistical significance is shown for Tukey post-tests of one-way ANOVAs.
Fig. S5.Identification and characterization of ATF3 KO MEFs.(A) Several putative ATF3 deletion clones, including those identified by the ICE Crispr Analysis Tool (Table S2), and a similar number of ROSA26 (control) clones were serum-starved, pre-treated with 10 µM EHT1864 in suspension, and plated on FNcoated tissue culture dishes with 10% FBS-DMEM in the continued presence of EHT1864 for 5 hr before determination of ATF3 protein levels by immunoblot.GAPDH was the loading control.Red arrowheads show several ROSA and ATF3 KO clones with the expected behavior for ATF3.(B) Images of control and ATF3 KO MEFs that had been serum-starved and plated on stiff FN-coated hydrogels in DMEM with 10% FBS for 9 hr before fixation and staining with Alexa Fluor™ 594 Phalloidin.(C-D) Quantitative analysis of cell area and phalloidin intensity per cell in the ROSA26 controls (R3, R11, R12, R15) and ATF3 KO clones (1-20, 1-29, 1-44, 1-49) as determined using Image J. Graphs show ~170-370 cells analyzed per condition, accrued from 2 independent experiments.Statistical significance is shown for Tukey post-tests of one-way ANOVAs.

Table S1 . Gene lists for the Venn diagrams shown in Figure 2 (continued) Panels E-F
Gene list, fold changes (FC) and adjusted p-values (padj) for the graphs shown in Figures2E-F.