β-Actin mRNA compartmentalization enhances focal adhesion stability and directs cell migration

Abstract

Directed cell motility is at the basis of biological phenomena such as development, wound healing, and metastasis. It has been shown that substrate attachments mediate motility by coupling the cell's cytoskeleton with force generation. However, it has been unclear how the persistence of cell directionality is facilitated. We show that mRNA localization plays an important role in this process, but the mechanism of action is still unknown. In this study, we show that the zipcode-binding protein 1 transports β-actin mRNA to the focal adhesion compartment, where it dwells for minutes, suggesting a means for associating its localization with motility through the formation of stable connections between adhesions and newly synthesized actin filaments. In order to demonstrate this, we developed an approach for assessing the functional consequences of β-actin mRNA and protein localization by tethering the mRNA to a specific location-in this case, the focal adhesion complex. This approach will have a significant impact on cell biology because it is now possible to forcibly direct any mRNA and its cognate protein to specific locations in the cell. This will reveal the importance of localized protein translation on various cellular processes.

Figure 1.

ZBP1 is necessary for β-actin mRNA localization and directionality in mouse embryonic fibroblasts (MEFs). (A) Live-cell TIRF images of wild-type and ZBP1 knockout fibroblasts with TagRFPt-labeled β-actin mRNA and free GFP as the cytoplasmic marker. Corresponding polarization indices are shown, based on a reported algorithm that assesses asymmetry by computing the intensity-weighted centroids of mRNA and cytoplasmic GFP (Park et al. 2012). (B) The average polarization index of β-actin mRNA distribution was significantly lower in cells without ZBP1. n = 60 for each condition, and bars represent polarization index averages ± SEM. (C) The directionality of fibroblasts (computed as the net path divided by the total path) was also found to be significantly reduced in fibroblasts without ZBP1. n = 50 for each condition ±SEM.

Figure 2.

ZBP1 localizes β-actin mRNA to the adhesion compartment, where it dwells for minutes. (A) Endogenous β-actin mRNA labeled with 2xMCP-GFP bound to 24 MS2 stem loops in the 3′ UTR was imaged in TIRF every 10 sec for 10 min. (B) For each frame, mRNA particles were identified within a chosen region of interest (ROI) (yellow square) in the cellular periphery with Localize software (Larson et al. 2005), as indicated by the red dots. (C) Identified mRNAs that persisted for >1 min (seven consecutive frames) with less than five pixels of displacement between each frame were tracked and categorized as dwelling mRNAs (identified by the blue boxes). (D,E) The first and last frame included acquisition of paxillin-mCherry labeling to coordinate dwelling mRNAs to the nearest adhesion. The centroid coordinates of each tracked dwelling mRNA was coordinated to the nearest adhesion identified by thresholding (shown in E). (F) Distances (shown for each dwelling mRNA in the figure) were calculated for each dwelling mRNA centroid to the nearest adhesion (masked in white) and used to analyze the significance of ZBP1 on β-actin mRNA compartmentalization within the leading edge. (G) We found that ZBP1 expression in MEFs significantly increased the probability of β-actin mRNA dwelling for >1 min in the cell periphery. Data shown represents dwelling mRNA tracks found within each ROI divided by the average number of mRNA particles found per minute ± SEM, with the following n-values: wild type (WT) = 60; ZBP1 knockout (KO) = 46; ZBP1 KO + mCh-ZBP1 = 33. β-actin mRNA particles near the perinuclear region were not found to dwell for >1 min for all samples (data not shown). (H) All dwelling mRNAs for each condition as a factor of dwell time versus distance to the nearest adhesion were plotted. More dwelling mRNA was observed as distance to the nearest adhesion decreased.

Figure 3.

Focal adhesion stability is directly related to ZBP1 expression in migrating MEFs. Paxillin-mCherry was expressed in wild-type MEFs (A–C) and ZBP1 knockout MEFs (D,E) and imaged every minute for 6 h in TIRF (Supplemental Movie 3). (B,E) Average projections of adhesion movement movies depict progressive waves of adhesions in the direction of migration in wild-type MEFs but a disarray of adhesion movement in ZBP1 knockout MEFs. (G) We measured a significant reduction in average adhesion lifetimes in ZBP1 knockout cells, although adhesion size was not significantly different (data not shown). The time series of individual protrusions in C and F depict a coordinated protrusion and adhesion maturation paradigm in wild-type MEFs, while no clear adhesion stability is seen in the ZBP1 knockout protrusion over the same time course (each montage frame is 15 min). Data shown for wild type and ZBP1 knockout represent averages with n-values of >18 fields over three independent experiments ±SEM.

Figure 4.

β-Actin mRNA tethering to focal adhesions specifically increases adhesion size and lifetimes. (A) Schematic for the mRNA tethering assay. mRNA with multiple MS2 stem–loops in the 3′ UTR binds with high affinity to the vinculin construct fused to the MCP dimer and GFP. As adhesions mature, more vinculin-2xMCP-GFP accumulates, and tethered mRNA increases in concentration at the adhesion. (B) Cy3 FISH probes were used to confirm that β-actin mRNA was tethered to the vinculin-2xMCP-GFP construct. (Top) Tethered cells show dense colocalizing signal in TIRF. (Bottom) Non-MBS ZBP1 knockout cells (nontethered) showed no overlay of FISH signal with vinculin-2xMCP-GFP. Tethering β-actin mRNA to adhesions in ZBP1 knockout cells increased adhesion lifetimes (C) to wild-type values but also significantly increased the adhesion size (D). CFP mRNA with MS2 stem–loops in the 3′ UTR was used as a control tethered mRNA in ZBP1 knockout cells and produced no increase in adhesion size or lifetime. All data shown represent averages of fields analyzed with n-values as follows: ZBP1 knockout (KO) = 30; ZBP1 KO + CFP-MS2 = 17; MBS-ZBP1 KO = 30; wild type (WT) = 24. (E) To test the influence of translation, adhesions were tracked before and after cycloheximide (100 ug/mL) addition. The percent change in adhesion lifetime after dosage was calculated for each experimental group. n-values are as follows: wild type = 19; MBS-ZBP1 KO = 18; ZBP1 KO = 17. Error bars represent SEM, and P-values above data bars are from a two-tailed Student's t-test compared between experimental groups.