Extracellular matrix stiffness and cell contractility control RNA localization to promote cell migration - PubMed (original) (raw)
Extracellular matrix stiffness and cell contractility control RNA localization to promote cell migration
Tianhong Wang et al. Nat Commun. 2017.
Abstract
Numerous RNAs are enriched within cellular protrusions, but the underlying mechanisms are largely unknown. We had shown that the APC (adenomatous polyposis coli) protein controls localization of some RNAs at protrusions. Here, using protrusion-isolation schemes and RNA-Seq, we find that RNAs localized in protrusions of migrating fibroblasts can be distinguished in two groups, which are differentially enriched in distinct types of protrusions, and are additionally differentially dependent on APC. APC-dependent RNAs become enriched in high-contractility protrusions and, accordingly, their localization is promoted by increasing stiffness of the extracellular matrix. Dissecting the underlying mechanism, we show that actomyosin contractility activates a RhoA-mDia1 signaling pathway that leads to formation of a detyrosinated-microtubule network, which in turn is required for localization of APC-dependent RNAs. Importantly, a competition-based approach to specifically mislocalize APC-dependent RNAs suggests that localization of the APC-dependent RNA subgroup is functionally important for cell migration.Adenomatous polyposis coli (APC) regulates the localization of some mRNAs at cellular protrusions but the underlying mechanisms and functional roles are not known. Here the authors show that APC-dependent RNAs are enriched in contractile protrusions, via detyrosinated microtubules, and enhance cell migration.
Conflict of interest statement
The authors declare no competing financial interests.
Figures
Fig. 1
Distinct RNA groups are targeted at protrusions through APC-dependent and APC-independent pathways. a Pie chart of non-localized or protrusion (Ps)-enriched RNAs in control knockdown (KD) cells from 4 replicate experiments. Cutoffs for Ps-enrichment were set at FC > 2 and _p_-value < 0.05. b Cumulative fraction plot of log2 fold-change differences between control and APC knockdown cells, for RNAs non-localized or localized in protrusions of control cells. c For all Ps-localized RNAs of control cells, the log2 differences in FC values between control and APC knockdown cells were plotted against the corresponding _p_-values. Applying the cutoffs marked by the red lines (see Supplementary Data 1 for details), Ps-localized RNAs were distinguished into APC-dependent and APC-independent groups. d Categories of molecular functions, derived through IPA analysis, significantly represented in APC-dependent and APC-independent RNA groups. The identities of the RNAs included in the groups are listed. e Control or APC knockdown cells were fractionated into Ps and cell body (CB) fractions. The indicated RNAs were detected through nanoString analysis to calculate Ps/CB enrichment ratios (n = 3; error bars: standard error). _p_-value: **< 0.0001, *< 0.04 by analysis of variance with Bonferroni’s multiple comparisons test against the corresponding control
Fig. 2
APC-dependent and APC-independent RNAs are enriched in different types of protrusions, associated with high or low levels of contractility. a NIH/3T3 fibroblast cells were plated on microporous filters containing pores of 3 or 0.4 μm in diameter. The cells were induced to extend protrusions by addition of LPA in the bottom chamber and were subsequently fixed and stained to detect actin (green), tubulin (red), and nuclei (DAPI, blue). Confocal xy-slices through the cell body or protrusions, or xz-orthogonal views are shown. Scale bars: 4 μm. b Ps and CB fractions were isolated from cells extending protrusions through 3 or 0.4 μm pores. The indicated RNAs were detected through nanoString analysis to calculate Ps/CB enrichment ratios (n = 3; error bars: standard error). c Ps and CB fractions, of cells extending protrusions through 3 or 0.4 μm pores, were analyzed by western blot to detect the indicated proteins. Results are representative of two independent repeats
Fig. 3
Increased substrate stiffness and actomyosin contractility enhance APC-dependent RNA localization. a Left panel: schematic depicting changes in PDI index values for the indicated hypothetical RNA distributions. Higher PDI values reflect more peripheral RNA distributions. Middle panel: in situ hybridization of Ddr2 RNA (localized, cyan) and polyA RNA (control, magenta) in a representative cell on glass. Right panel: corresponding PDI values from a population of cells. Note that, in order to compare cytoplasmic distributions, signals within the nuclear area are subtracted prior to calculation of PDI values. This creates an apparent increase in PDI, which is disproportionately observed on 1kPa substrates (e.g., in panels b, h), because of the larger relative area covered by the nucleus. b Representative Ddr2 and polyA RNA distributions on substrates of varying stiffness, and corresponding PDI values. Arrows point to peripheral Ddr2 RNAs. c Traction force microscopy analysis on substrates of the indicated stiffness. d Representative Ddr2 RNA distributions and PDI values of cells treated with blebbistatin (Bleb) or Y27632, plated on 5 kPa substrates. e Traction force microscopy analysis of cells treated as in d. f Representative Ddr2 RNA distributions and PDI values of control cells or cells knocked-down for Myosin heavy chain IIA, Myh9, plated on 5kPa substrates. g Traction force microscopy analysis of cells treated as in f. h Representative Ddr2 RNA distributions and PDI values of control cells or cells treated with LPA or Rho activator II, plated on 1 kPa substrates. i Traction force microscopy analysis of cells treated as in h. For all PDI quantifications, bars represent the mean with 95% confidence interval. Points indicate individual cells analyzed in at least two or more independent experiments. For all traction force analyses, error bars represent standard error and _n_-values are indicated within each bar. _p_-values: ****<0.0001, ***<0.001, **<0.01, *<0.05 by analysis of variance (b–f, h, i) or Mann–Whitney test (g). Scale bars: 10 μm. Yellow lines: cell outline; white dashed line: nuclear outline
Fig. 4
Actomyosin contractility is required for formation of the detyrosinated-microtubule network. Representative immunofluorescence staining images of Glu-tubulin or total alpha-tubulin. Graphs show corresponding percentages of cells exhibiting a Glu-MT network. a Cells were plated on substrates of the indicated stiffness. b Cells on 5 or 280 kPa substrates were treated with Blebbistatin or Y27632. We note that cells treated with Blebbistatin mostly retain Glu-tubulin staining within their long, thin protrusions. Cells were scored for the presence of Glu-MTs within the main cell body. c Cells on 1 kPa substrates were treated with LPA or Rho activator II. Error bars: standard error. N: total number of cells observed in two or more biological replicates. _p_-values: ***< 0.001, **< 0.01 by analysis of variance with Bonferroni’s multiple comparisons test. Scale bars: 10 μm
Fig. 5
Detyrosinated microtubules are necessary and sufficient for the peripheral localization of Ddr2 RNA. a Schematic depicting the microtubule detyrosination cycle. Indicated is the C-terminal tyrosine residue removed by an unknown carboxypeptidase. The effects of parthenolide treatment or knockdown of tubulin tyrosine ligase (TTL) are also shown. b FISH analysis of control or parthenolide treated cells plated on 5 kPa substrates. Representative Ddr2 RNA distributions and PDI values of Ddr2 and polyA RNA are shown. c FISH analysis of control or TTL knockdown cells plated on 1 kPa substrates. Representative Ddr2 RNA distributions and PDI values of Ddr2 and polyA RNA are shown. For PDI quantifications, bars represent the mean with 95% confidence interval. Points indicate individual cells analyzed in at least two or more independent experiments. _p_-values: ***< 0.001 by analysis of variance with Bonferroni’s multiple comparisons test. Scale bars: 10 μm
Fig. 6
Detyrosinated microtubules mediate the effect of contractility on APC-dependent RNA localization at protrusions. a FISH analysis of cells on 1 kPa substrates treated with LPA alone or LPA together with parthenolide. PDI values of Ddr2 and polyA RNA are shown. b FISH analysis of control or TTL knockdown cells, treated or not with blebbistatin as indicated. Cells were plated on 5 kPa substrates. PDI values of Ddr2 and polyA RNA are shown. c Representative Glu-tubulin immunofluorescence staining of cells treated as in b. Graph shows corresponding percentages of cells exhibiting a Glu-MT network. d FISH analysis of cells transfected with GFP or GFP-mDia(ΔN3), treated or not with blebbistatin as indicated. Cells were plated on 5 kPa substrates. PDI values of Ddr2 and polyA RNA are shown. e Representative Glu-tubulin immunofluorescence staining of cells treated as in d. Graph shows corresponding percentages of cells exhibiting a Glu-MT network. For all PDI quantifications, bars represent the mean with 95% confidence interval. Points indicate individual cells analyzed in at least two or more independent experiments. For cell percentages, error bars show standard error and N is the total number of cells observed in two or more biological replicates. _p_-values: ***< 0.001, **< 0.01 by analysis of variance with Bonferroni’s multiple comparisons test. Scale bars: 10 μm
Fig. 7
A competition-based approach preferentially mislocalizes APC-dependent RNAs from protrusions. a Schematic of exogenously expressed constructs containing the β-globin coding sequence followed by the indicated 3′UTRs (β-globin UTR (HBB; spaced dashed line); Rab13 UTR (dashed line), Pkp4 UTR (solid line), or deletion fragments of Pkp4 UTR (PkpB, PkpA12, and PkpA12.1)). Representative in situ hybridization images detecting the β-globin RNA, and the ability of each construct to localize at protrusions are shown. Images focus on individual protrusions. Scale bar: 10 μm. Whole-cell images and PDI values are shown in Supplementary Fig. 7a, b. b Cell lines, stably expressing β-globin constructs with the indicated UTRs, were analyzed by in situ hybridization and RNA distributions were assessed through PDI quantitation. (Ddr2, Kank2: APC-dependent RNAs; Arpc3: control, diffuse RNA). _N_-values of observed cells, from at least 3 independent experiments, are indicated within each bar. Error bars: standard error. _p_-values: ***<0.001, **<0.01, *<0.05 by analysis of variance with Dunnett’s multiple comparisons test. **c** Pie chart of non-localized or Ps-enriched RNAs in control HBB cells from 4 replicate experiments. Cutoffs for Ps-enrichment were set at FC > 2 and _p_-value < 0.05. d Cumulative fraction plot of log2 fold-change differences between control HBB and Pkp4 cUTR-expressing cells, for RNAs non-localized or localized in protrusions of control HBB cells. e For all Ps-localized RNAs of control HBB cells, the log2 differences in FC values, between HBB control and Pkp4 cUTR-expressing cells, were plotted against the corresponding _p_-values. Applying the cutoffs marked by the red lines (see Supplementary Data 3 for details), Ps-localized RNAs were distinguished into groups less- or equally-enriched in protrusions upon Pkp4 cUTR expression. f Categories of molecular functions, derived through IPA analysis, significantly represented in RNA groups equally- or less-enriched in protrusions of Pkp4 cUTR-expressing cells. g Results of differential enrichment analysis showing significance of overlap between APC-dependent RNAs and RNAs mislocalized upon Pkp4 cUTR expression
Fig. 8
Localization of APC-dependent RNAs is required for efficient migration on 2D and 3D substrates. a, b Cell lines, stably expressing β-globin constructs with the indicated UTRs, were plated on top of collagen gel a or Matrigel b and were induced to migrate by addition of serum (FBS) in the bottom chamber. The amount of cells reaching the bottom was quantified from at least 5 independent experiments, (error bars: standard error). Lower panels show the amount of cells reaching the bottom in the absence of matrix, in parallel experiments, to control for cell viability during the assay. Note that because of the length of the assay, migration in the absence of matrix does not reflect migration speed. c, d Cells stably expressing constructs with the indicated UTRs c or transfected with the indicated siRNAs d, were monitored while randomly migrating. Average velocity values were derived from multiple individual cell tracks in two independent experiments observing 40–70 cells in each. e Migration through Matrigel of control or APC knockdown cells. _p_-values: ***< 0.0001, *< 0.05 against HBB or si-Control, by analysis of variance with Bonferroni’s multiple comparisons test
References
Publication types
MeSH terms
Substances
LinkOut - more resources
Full Text Sources
Other Literature Sources