OptoTAT Reveals Microtubule Acetylation as a Rapid Trigger for GEF-H1–Mediated Cell Migration

Abstract

Microtubule acetylation is implicated in regulating cell motility, yet its physiological role in directional migration and the underlying molecular mechanisms have remained unclear. This knowledge gap has persisted primarily due to a lack of tools capable of rapidly manipulating microtubule acetylation in actively migrating cells. To overcome this limitation and elucidate the causal relationship between microtubule acetylation and cell migration, we developed a novel optogenetic actuator, optoTAT, which enables precise induction of microtubule acetylation within minutes in live cells. Implementing optoTAT in migration assays, we observed striking and rapid responses at both molecular and cellular levels. First, microtubule acetylation triggers release of the RhoA activator GEF-H1 from sequestration on microtubules. This release subsequently enhances actomyosin contractility and drives focal adhesion maturation. These subcellular processes collectively promote sustained and directional movement. Our findings position GEF-H1 as a critical molecular responder to microtubule acetylation, enabling a dynamic crosstalk between the actin and microtubule cytoskeletal networks in the coordination of cellular motility.

Introduction:

Microtubules undergo at least nine different types of post-translational modifications, which independently, or in concert, modulate microtubule properties including its dynamics as well as their interaction with microtubule-associated proteins¹. Acetylation of the lysine-40 residue of α-tubulin^2–6, hereafter called microtubule acetylation for simplicity, is conserved throughout eukaryotes^7–9, and is one of the only few modifications known to take place inside microtubule lumen⁹. Despite little significant structural changes¹⁰, microtubule acetylation provides structural stability against bending forces^11–16. Microtubule acetylation has been implicated in cellular processes, including mechanosensing^12,17–21, adaptation to extracellular environment^22–25, intracellular transport via motor proteins^26–31, DNA damage response³², autophagy^33–35, and regulation of cell motility^{22,23,36–39}. Directionally persistent cell migration, a process critical for physiological functions as well as pathological events, heavily relies on microtubule dynamics. While the involvement of microtubules in migration is well-established, the specific contributions of microtubule post-translational modifications remain poorly understood^40,41. Microtubule acetylation, in particular, appears to exhibit differential roles in cell migration depending on cell type and environmental context. For example, it inhibits three-dimensional migration in human foreskin fibroblasts²³ and transwell migration in NIH3T3 fibroblasts⁴² while promoting motility in astrocytes^22,24 and breast cancer cells^36–38. In contrast, acetylation is dispensable for the motility of RPE1 epithelial cells⁴³. These conclusions have largely been drawn from studies employing genetic engineering or pharmacological interventions to alter microtubule acetylation.

Genetic approaches are invaluable for identifying genes responsible for these effects; however, they often lack the temporal precision required to investigate rapid cytoskeletal dynamics during cell migration, which can occur within minutes. Similarly, pharmacological interventions to modulate microtubule acetylation may have unintended non-specific effects^42,44–47, complicating the interpretation of results. The paucity of molecular tools capable of controlling microtubule acetylation with rapid temporal resolution and high molecular specificity has presented a significant challenge in elucidating its real-time roles in dynamic cell behavior such as directional cell migration. To address this, we developed a genetically encoded actuator, termed optoTAT, that is designed to induce microtubule acetylation within minutes upon light illumination. By leveraging this optogenetic actuator in combination with genetic knock-out models, migration assays, and live cell fluorescence imaging, we aimed to uncover the molecular interplay between microtubule acetylation, actin cytoskeleton remodeling, and directional cell migration in real time.

Results:

Microtubule acetylation mediates directional migration:

α-TAT1 is the only enzyme known to acetylate microtubules in mammals^12,48,49, whereas the deacetylation is catalyzed by HDAC6 and Sirt2^42,50. Mouse Embryonic Fibroblasts (MEFs) obtained from α-TAT1 knockout (KO) mice do not have detectable microtubule acetylation^12,17,51. In a random migration assay, α-TAT1 KO MEFs showed significantly greater motility but reduced directional persistence compared to wild-type (WT) MEFs (Fig. 1a, b, c). To test whether the observed decrease in directional persistence in migrating α-TAT1 KO MEFs was simply due to a greater path covered within the same duration, we measured directional persistence of WT and α-TAT1 KO MEFs within a fixed range of pathlengths (mean ± 95% CI, WT : 64.65 ± 0.93 μm, KO : 65.52 ± 1.19 μm). We found that even with comparable pathlengths, the α-TAT1 KO MEFs showed less directional persistence than WT MEFs (Supplementary Fig. S1a). In wound healing assays, α-TAT1 KO MEFs closed the wound more rapidly than WT MEFs (Fig. 1d, e). To examine the effects of microtubule acetylation on chemotaxis, we utilized an Ibidi chemotaxis chamber with 0–20% FBS gradient (Fig. 1f). α-TAT1 KO MEFs failed to efficiently migrate towards the chemoattractant compared to WT MEFs (Fig. 1g, Supplementary Fig. S1b). Unlike the WT MEFs, the α-TAT1 KO MEFs exhibited reduced directional bias towards the chemoattractant gradient, as indicated by the reduced shift in the center of mass from the origin (Supplementary Fig. S1c). The α-TAT1 KO MEFs showed significantly reduced directional persistence compared to WT MEFs, as shown by decreased forward migration index (FMI) for the KO MEFs along the chemoattractant gradient (FMI^II) (Fig. 1h), but not perpendicular to the gradient (FMI^⊥) (Fig. 1i). Exogenous expression of mVenus-α-TAT1 in KO MEFs (Supplementary Fig. S1d) was sufficient to rescue these chemotaxis defects (Fig. 1g, h, supplementary Fig. S1a, b).

Figure 1. α-TAT1 modulates directional cell migration.

a) Tracks, b) Speed (μm/hr) and c) Directionality of WT and α-TAT1 KO MEFs in a random migration assay, WT: 18, KO: 23 cells; d), e) Temporal changes in wound width in a wound healing assay with WT and α-TAT1 KO MEFs, n = 12 wound regions, mean ± 95% C.I.; f) Schematic for chemotaxis assay (adapted from Ibidi); g) Rose plots of WT, α-TAT1 KO or KO-rescue MEFs migrating in a chemotactic gradient, h) Forward migration indices along the chemotactic gradient and i) Forward migration indices perpendicular to the chemotactic gradient for WT, α-TAT1 KO or KO-rescue MEFs, n = 120 cells for each group; j) Temporal changes in morphology of WT or α-TAT1 KO MEFs undergoing random migration; k) Persistence of protrusions, l) Frequency of new protrusion formations in randomly migrating WT or α-TAT1 KO MEFs, WT: 23 and KO: 19 cells, m) circularity and n) convexity of WT and α-TAT1 KO MEFs (WT: 40 and KO: 54 cells). Scale bars: 10 μm ***: p<0.001

On examining the motility of WT and α-TAT1 KO MEFs, we observed that in contrast to WT MEFs, the α-TAT1 KO MEFs change their direction of motion repeatedly (Fig. 1j, Supplementary Movie S1). The WT MEFs had two groups of protrusions, one short-lived and another long-lived, as indicated by the bimodal distribution of protrusion lifetimes (Fig. 1k). α-TAT1 KO MEFs had very few such long-lived protrusions (Fig. 1k, Supplementary Movie S1). The α-TAT1 KO MEFs also produced new protrusions more frequently than WT MEFs, leading to changes in direction of movement (Fig. 1l, Supplementary Movie S1). Directional persistence requires a long-lasting front-back polarity^52,53, and the frequent protrusion formation in α-TAT1 KO MEFs suggest defects in maintenance of such front-back polarity. Consistent with this, morphological analyses showed that α-TAT1 KO MEFs have higher circularity and higher convexity (Fig. 1m, n), consistent with more protrusive phenotypes. The chemotaxis defects in α-TAT1 KO MEFs were not due to defects in sensing the chemoattractant since serum-starved WT and α-TAT1 KO MEFs showed comparable morphological changes: increased protrusions, on treatment with 10% FBS (Supplementary Fig. S1e).

Other groups examining the impact of microtubule acetylation on cell motility have reported contradictory findings. Microtubule acetylation is critical for extra-cellular matrix stiffness sensing and astrocyte motility, as RNAi-mediated knock-down of α-TAT1 and resulting loss of microtubule acetylation significantly inhibited astrocyte migration²². In contrast, in human foreskin fibroblasts (HFF), exogenous expressions of acetyl-null or acetyl-mimic α-Tubulin mutants respectively increased and decreased cell motility, without significantly affecting microtubule dynamics²³. Our observations of increased motility in α-TAT1 KO MEFs are consistent with the reports in HFF cells, supporting the idea that microtubule acetylation has context-dependent effects on cell migration, generally reducing mesenchymal cell speed while increasing directional persistence.

Microtubule acetylation mediates focal adhesion maturation:

In migrating cells, nascent protrusions are stabilized by integrin mediated adhesion complexes, which undergo maturation in response to actomyosin contractility^54,55. Microtubule acetylation has been previously reported to promote adhesion maturation and decrease adhesion disassembly^22–24. In astrocytes, loss of α-TAT1 also led to an increase in overall adhesion numbers²⁴. In HFFs, microtubule acetylation promoted surface density of α₅β₁ integrins through decreased internalization of these molecules²³. To characterize the impact of microtubule acetylation on focal adhesions in MEFs, we performed immunostaining for Vinculin, a protein that localizes to focal adhesions. We observed that α-TAT1 KO MEFs had only around half of the number of adhesions compared to WT MEFs (Fig. 2a, b). α-TAT1 KO MEFs also had fewer mature adhesions compared to WT MEFs as indicated by a decrease in adhesion sizes (Fig. 2a, Supplementary Fig. S2a), suggesting a defect in adhesion maturation pathways. Consistent with this decrease in focal adhesion numbers and sizes, α-TAT1 KO MEFs also had smaller cell spread area (Supplementary Fig. S2b). α-TAT1 KO MEFs have a higher proliferation rate¹⁷, which may also contribute to a decrease in overall cell size and cell-spread area. In WT MEFs, we observed a polarized distribution of nascent and maturing adhesions in the front, and large mature adhesions at retractions, whereas the α-TAT1 KO MEFs lacked any such front-and-back polarity (Fig. 2a). Vinculin preferentially localizes to mature adhesions over nascent adhesions in a tensile force dependent manner⁵⁶. Adhesions in the α-TAT1 KO MEFs showed lower levels of Vinculin accumulation compared to WT MEFs both at a cellular level and at individual adhesion level (Fig. 2a, Supplementary Fig. S2c, d), which is consistent with lower tensile forces acting on these adhesions. Exogenous expression of mVenus-α-TAT1, but not a catalytically dead mVenus-α-TAT1(D157N)¹², in α-TAT1 KO MEFs rescued these adhesion defects (Fig. 2a, b, Supplementary Fig. S2a, c, d). These adhesion defects were not due to decreased Vinculin expression since both WT and α-TAT1 KO MEFs showed comparable levels of Vinculin expression (Fig. 2c, d).

Figure 2. Microtubule acetylation promotes focal adhesion maturation and actomyosin contractility.

a) Vinculin distribution in WT, α-TAT1 KO MEFs, and KO-rescue with mVenus-α-TAT1 or catalytic dead mVenus-α-TAT1(D157N) as indicated; b) Number of adhesions per cell (WT:20, KO: 17, rescue-WT: 16, rescue-D157N: 22 cells); c) Western blot showing Vinculin and α-Tubulin expression in WT and α-TAT1 KO MEFs; d) Normalized Vinculin expression levels in WT and α-TAT1 KO MEFs by western blots (3 independent experiments, error bar: standard deviation); e) VinTS FRET index in WT and α-TAT1 KO MEFs, f) Average VinTS FRET index in WT and α-TAT1 KO MEFs (WT:: 18, KO: 16 cells); g) Phalloidin and phospho-MRLC distribution in WT and α-TAT1 KO MEFs, red arrowheads indicate bundled actin; h) Phospho-MRLC levels in WT, α-TAT1 KO, rescue-WT and rescue-D157N MEFs (WT: 54, KO: 64, rescue-WT: 53 and rescue-D157N: 55 cells); i) mCherry-MRLC distribution and optical flow magnitudes of mCherry-MRLC in WT and α-TAT1 KO MEFs; j) Mean mCherry-MRLC optical flow magnitudes in WT and α-TAT1 KO MEFs (WT: 11, KO: 12 cells). Scale bars: 10 μm. ***: p<0.001

Adhesion maturation is mediated by tensile forces experienced by focal adhesion components through the actin cytoskeleton. Collagen gel contraction and traction force imaging assays indicate that microtubule acetylation has opposite effects on cell contractility, reducing it in HFFs while enhancing it in astrocytes^22,23. To directly test the tensile forces exerted on these focal adhesions, we utilized the Vin-TS FRET-based tension sensor⁵⁷. The Vin-TS tension sensor uses a tension sensing module comprising of a FRET donor-acceptor pair separated by a spider-silk protein, and this tension sensing module was inserted between the Vinculin head and tail domain^57,58. Since FRET is highly sensitive to distances between the donor and acceptor fluorophore, tensile forces acting on Vin-TS will lead to decreased FRET. Therefore, high FRET signal will indicate low cell contractility and low FRET signal will indicate higher cell contractility. We observed increased FRET in the α-TAT1 KO MEFs compared to WT MEFs at whole cell as well as individual adhesion levels, indicating that focal adhesions in the α-TAT1 KO MEFs experienced significantly reduced tensile forces (Fig. 2e, f, Supplementary Fig. S2e). Furthermore, the Vin-TS FRET signal showed a polarized distribution in the WT MEFs, but not in the α-TAT1 KO cells (Fig. 2e). This would suggest a more front-back polarized state of cell contractility in WT MEFs compared to the α-TAT1 KO cells. Taken together, our observations suggest that microtubule acetylation promotes focal adhesion maturation, largely consistent with previous reports. Surprisingly, we also found that our observations suggest that microtubule acetylation enhances tensile forces on adhesions, which is consistent with previous reports in astrocytes but not in HFFs^22,23.

Microtubule acetylation mediates actomyosin contractility:

Focal adhesions experience tensile forces through contractility of the actin cytoskeleton⁵⁹. Since our observations suggested a decrease in contractility in α-TAT1 KO MEFs, we performed characterization of the actin cytoskeleton in WT and α-TAT1 KO MEFs. Phalloidin staining showed a significant reduction in bundled actin in α-TAT1 KO cells (Fig. 2g, red arrowheads), suggesting that these cells have defects in actin contractility. Contractility in the actin cytoskeleton is generated through Myosin motor proteins, which are activated through phosphorylation of the Myosin Regulatory Light Chain (MRLC) at Serine19 by Myosin Light Chain Kinase (MLCK)⁶⁰. Myosin activation is also involved in directional persistence of migrating cells as well as Vinculin recruitment to focal adhesions^61,62. Immunostaining of WT or α-TAT1 KO MEFs with an antibody against phospho-MRLC Serine19 showed a significantly lower levels of phospho-MRLC (Fig. 2g, h), indicating decreased activation levels of Myosin. The decrease in phospho-MRLC levels was not due to a decrease in expression levels since WT and KO cells showed comparable Myosin expression levels (Supplementary Fig. S2f, g). These defects in MRLC phosphorylation could be rescued with exogenous expression of mVenus-α-TAT1 but not catalytically dead mVenus-α-TAT1(D157N) mutant (Fig. 2h). Since MRLC phosphorylation leads to association of MRLC with the actin cytoskeleton, we measured the magnitude of the optical flow of mCherry-MRLC as a proxy for MRLC binding to the actin cytoskeleton downstream of myosin activation⁶¹. mCherry-MRLC flow magnitude was considerably lower in α-TAT1 KO MEFs compared to WT cells (Fig. 2i, j), indicating decreased MRLC association with the actin cytoskeleton, presumably because of lower levels of myosin activity in these cells. Treating the α-TAT1 KO cells with 10 μM ROCK inhibitor Y-27632 led to a decrease in phospho-Myosin levels, suggesting a residual amount of myosin activity, however diminutive (Supplementary Fig. S2h, i).

Previous studies have reported contradictory relationships between microtubule acetylation and actomyosin contractility across different cell types. In HFFs, reduced cell contractility correlates with increased microtubule acetylation, and elevated microtubule acetylation leads to decreased actomyosin contractility^23,63. On the other hand, in astrocytes, reduced actomyosin contractility correlates with decreased microtubule acetylation, and loss of microtubule acetylation results in reduced myosin activation²². Interestingly, while our observations of MEF motility are consistent with findings in HFFs but not astrocytes, our results support a role for microtubule acetylation in myosin activation that more closely resembles observations in astrocytes rather than in HFFs.

Inhibiting HDAC6 weakly promotes myosin activation:

Given our observations that align with some but not other results from previous studies related to microtubule acetylation, actomyosin contractility and cell motility, we wanted to assess whether microtubule acetylation played a direct and causal role in myosin activation, or whether these observations based on genetic perturbations are related to cellular adaptation. Inhibiting HDACs, including HDAC6, using Trichostatin-A in HFFs led to an increase in microtubule acetylation and a decrease in myosin activation²³. In astrocytes, treatment with Tubacin, a pharmacological inhibitor of HDAC6⁶⁴, led to increased microtubule acetylation along with increased myosin activation²². To test whether increasing microtubule acetylation levels through inhibition of HDAC6 in MEFs could increase myosin activation, we used TIRF microscopy to characterize changes in mCherry-MRLC association with actin cytoskeleton in WT MEFs treated with 2 μM Tubacin. Over 5 hours of Tubacin treatment, we observed a minor increase in mCherry-MRLC signal on the TIRF plane (Supplementary Fig. S2j, k, l), indicating increased activation and association with the actin cytoskeleton. We did not observe any significant increase in mCherry-MRLC signal in α-TAT1 KO MEFs (Supplementary Fig. S2l). Immunostaining for phospho-MRLC also showed a minor increase with 2 μM Tubacin treatment for 2 hours in WT, but not α-TAT1 KO MEFs (Supplementary Fig. S2m). However, this increase over the course of hours does not eliminate the possibility of cell adaptation through transcriptional regulation or non-specific effects. HDAC6 has many other substrates other than α-Tubulin^44,65,66, and it can also deacetylate additional acetylated lysine residues on α-Tubulin⁶⁷. Thus, HDAC6 inhibition is not sufficiently specific to determine causal relationships between acetylation of Lysine-40 in α-Tubulin and cellular or molecular responses.

Developing an optogenetic actuator to rapidly induce microtubule acetylation:

We reasoned that to directly test a causal relationship between microtubule acetylation and myosin activation, it would be necessary to control microtubule acetylation levels in living cells in a reasonably rapid manner and characterize its impact on myosin activation. To achieve this, we sought to develop an inducible molecular actuator to control microtubule acetylation within a reasonably short duration to minimize the possibility of cell adaptation. Initially we tested Z-lock-α-TAT1 in HeLa cells⁶⁸. However, we observed a significant increase in microtubule acetylation in cells expressing mCherry-Z-lock-α-TAT1 even in dark (Supplementary Fig. 3a, b). We have previously shown that cytoplasmic localization of α-TAT1 through its C-terminal spatial regulatory domain is critical for microtubule acetylation. Nuclear localization of α-TAT1 is sufficient to sequester it from catalyzing microtubule acetylation⁵¹. Based on this, we reasoned that inducing export of a nuclear-localized α-TAT1 may induce acetylation of microtubules (Fig. 3a). We initially implemented the light-inducible nuclear export system (LEXY)⁶⁹ to sequester full-length α-TAT1(M1-R323) in the nucleus in dark. We named this construct Optogenetic Tubulin Acetyl-Transferase version 0 (optoTATv0) (Fig. 3b). On blue-light stimulation, we observed a rapid nucleus-to-cytoplasm translocation of mCherry-optoTATv0 (Fig. 3c, d). However, we also observed significant levels of cytoplasmic presence even in the absence of blue-light stimulation (Fig. 3c, d), presumably due to the presence of nuclear export and cytoplasmic retention machinery in α-TAT1 C-terminus⁵¹. To improve upon this design, we tethered only the catalytic domain of α-TAT1(M1-S236)¹² to LEXY (optoTATv1) or to LEXY with two NLS (optoTATv2) (Fig. 3b). These versions showed increased nuclear sequestration (Fig. 3c, d), with rapid, robust and reversible cytoplasmic translocation on blue light stimulation (Fig. 3c, d, e, f, Supplementary Movie S2). To examine whether blue-light stimulation of these tools could acetylate microtubules, we exposed HeLa cells expressing mCherry-optoTATv1 or mCherry-optoTATv2 to blue light for 2 hours and performed immunostaining for acetylated α-Tubulin and total α-Tubulin. To account for the natural variation in basal levels of acetylated microtubules in HeLa cells, we measured the acetylated over total α-Tubulin signal in both transfected and non-transfected cells, kept in dark or exposed to blue light. We normalized the ratio of acetylated over total α-Tubulin for transfected cells with the averaged value for 20 non-transfected cells for the same dish (dashed line at value 1 in Fig. 3h). We observed that blue light stimulation of optoTATv1 or optoTATv2 significantly induced microtubule acetylation in HeLa cells (Fig. 3g, h). We also observed that the cells expressing optoTATv1, but not optoTATv2, showed increased levels of microtubule acetylation in dark when compared to non-transfected cells (Fig. 3h). The presence of two NLS in optoTATv2 reduced the background activation in dark but did not reduce the level of microtubule acetylation induced by blue light stimulation (Fig. 3h), thereby increasing the dynamic range of this tool. Therefore, the dynamic range as well as the background ‘leakiness’ of optoTAT may be tuned by altering the strength or numbers of NLS tethered to it. Since optoTATv2 showed low background activity in dark, we have used it for all further experiments, and for simplicity, we will refer to it as optoTAT here onwards.

Figure 3. Developing an optogenetic actuator to induce microtubule acetylation.

a) OptoTAT design; b) OptoTAT versions; c) Ratio of cytoplasmic over nuclear signal of optoTAT versions in dark and on 10 min blue light stimulation (V0: 14, V1: 17 and V2: 21 cells); d) Intracellular distribution of mCherry-optoTAT V0, V1 and V2 in dark and with 5 min blue light stimulation; e) Kymograph of mCherry-optoTAT V2 response to blue light, reference: red line in top panel of (d); f) Changes in average nuclear intensity of mCherry-optoTAT V2 on blue light stimulation indicated by blue lines, means ± 95% C.I., n= 21 cells; g) Microtubule acetylation levels in HeLa cells expressing mCherry-optoTAT V2, kept in dark or exposed to blue light for 2 hours, red arrowheads indicate transfected cells; h) Acetylated/total α-Tubulin in HeLa cells expressing mCherry-optoTAT V1 or V2 in dark or exposed to 2 hours blue light (V1 dark: 30, V1 light: 34, V2 dark: 27, V2 light: 33 cells), ratios were normalized against average value for 20 non-transfected cells (dashed line) in the same dish; i) Temporal changes in Acetylated/total α-Tubulin in HeLa cells stably expressing mVenus-optoTAT and continuously exposed to blue light stimulation for the duration indicated (0 min: 54, 5 min: 50, 10 min: 61, 30 min: 66, 60 min: 61, 120 min: 62, 180 min: 60 and 240 min: 61 cells), means ± 95% C.I.; j) PptoTAT localization and microtubule acetylation levels in α-TAT1 KO MEFs stably expressing mVenus-optoTAT V2, kept in dark or exposed to blue light for 30 min. Scale bars: 10 μm. ***: p<0.001. Blue lines: Blue light stimulation.

To assess the kinetics of microtubule acetylation by optoTAT, we used lentiviral transduction to generate a cell line of HeLa cells stably expressing mVenus-optoTAT. We used flow cytometric cell sorting to select cells with comparable levels of mVenus expression. These cells were incubated in dark for 24 hours and then exposed to blue-light for 0 min, 5 min, 10 min, 30 min, 60 min, 120 min and 240 min, followed by immunostaining for acetylated α-Tubulin and α-Tubulin. We observed a significant increase in microtubule acetylation within 10 minutes of blue-light stimulation, reaching saturation after about an hour of stimulation (Fig. 3i, Supplementary Fig. S3c). We also used lentiviral transduction to stably express mVenus-optoTAT in α-TAT1 KO MEFs. Blue light stimulation of these cells led to a translocation of mVenus-optoTAT from the nucleus to the cytoplasm concomitant with an increase in microtubule acetylation levels, with comparable kinetics (Fig. 3j, Supplementary Fig. S3d).

To test whether optoTAT targeted the same pool of α-Tubulin as α-TAT1, we treated HeLa cells expressing mVenus-optoTAT with 2 μM Tubacin for 2 hours followed by 1 hour of blue light stimulation. We observed that in cells treated with Tubacin, optoTAT did not lead to any additional increase in α-Tubulin acetylation levels (Supplementary Fig. S3e, f), thereby confirming that optoTAT acts of the same α-Tubulin population as α-TAT1. Stimulating mVenus-optoTAT for 1 hour in HeLa cells did not lead to any significant changes in microtubule detyrosination or polyglutamylation levels (Supplementary Fig. S3g, h). Our data demonstrate that we have developed an optogenetic molecular actuator to rapidly and specifically induce microtubule acetylation in living cells.

OptoTAT stimulation rapidly induces myosin activation:

Since α-TAT1 KO MEFs showed decreased levels of phospho-MRLC compared to WT MEFs, we reasoned that optoTAT stimulation may lead to increased MRLC phosphorylation. Blue light stimulation of mVenus-optoTAT in HeLa cells for 30 min led to increased levels of phospho-MRLC Serine19 (Fig. 4a, b). To assess the kinetics of Myosin activation downstream of optoTAT stimulation, we used blue light stimulation of α-TAT1 KO MEFs stably expressing mVenus-optoTAT for 0 min, 10 min, 30 min, 60 min and 120 min. We observed an increase in phospho-MRLC within 10 min, reaching saturation between 30 and 60 min (Fig. 4c, d). Our observations suggest a comparable kinetics of Myosin activation as α-Tubulin acetylation with optoTAT stimulation.

Figure 4. OptoTAT stimulation rapidly induces actomyosin contractility.

a), b) OptoTAT localization and phospho-MRLC levels in HeLa cells with 0min and 30 min blue light stimulation, 30 cells each; c), d) OptoTAT localization and phospho-MRLC levels in α-TAT1 KO MEFs stably expressing mVenus-optoTAT exposed to blue light for the indicated durations, 100 cells per timepoint; e), f) TIRF images showing mCherry-MRLC distribution on miRFP703-optoTAT stimulation in HeLa cells; f) Temporal changes in mCherry-MRLC intensity on miRFP703-optoTAT stimulation for 30 min, mean ± 95% C.I., n = 14 cells; g) Temporal changes in mCherry-MRLC intensity on miRFP703-optoTAT stimulation for 5 hours, mean ± 95% C.I., n = 26 cells, h), i) Changes in mCherry-MRLC distribution coherence index on miRFP703-optoTAT stimulation, n = 14 cells; j) Changes in mCherry-MRLC intensity in TIRF plane on 30 min blue light stimulation of miRFP703-optoTAT (14 cells), miRFP703-optoTAT(D157N) (12 cells), miRFP703-optoTAT with pre-treatment with 2 μM Tubacin (12 cells) or 10 μM Y27632 (12 cells); k) Changes in LifeAct-mCherry on miRFP703-optoTAT stimulation in HeLa cells, red arrowheads indicate bundled actin; l) Changes in LifeAct-mCherry intensity in TIRF plane on 30 min blue light stimulation of miRFP703-optoTAT (12 cells), miRFP703-optoTAT(D157N) (12 cells), miRFP703-optoTAT and pre-treatment with 2 μM Tubacin (12 cells) or 10 μM Y27632 (10 cells) in HeLa cells, m) Changes in mCherry-Paxillin on miRFP703-optoTAT stimulation in HeLa cells; n) Changes in average focal adhesion sizes and o) changes in average mCherry-Paxillin intensity on 30 min miRFP703-optoTAT stimulation, mean ± 95% C.I., n = 14 cells. Scale bars: 10 μm. ***: p<0.001. Blue line: Blue light stimulation.

Phosphorylation of MRLC at Serine-19, leads to Myosin (and MRLC) activation, resulting in association with F-actin⁶⁰. We stimulated miRFP703-optoTAT in HeLa cells co-expressing mCherry-MRLC and visualized changes in MRLC distribution using total internal reflection fluorescence (TIRF) microscopy. We reasoned that optogenetic Myosin activation will increase MRLC association with the actin cytoskeleton, leading to an increase in mCherry-MRLC intensity in the TIRF plane⁷⁰. Within 10-min optoTAT stimulation, we observed a significant increase in mCherry intensity in the TIRF plane, suggesting an increased association of mCherry-MRLC with actin cytoskeleton (Fig. 4e, f, Supplementary Movie S3). We also observed increased number of F-actin decorated with mCherry-MRLC in the TIRF plane within 30 min of optoTAT stimulation, suggesting an increase in bundled actin and stress fibers (Fig. 4e, Supplementary Movie S3).

Consistent with our observations with phospho-MRLC, we observed that prolonged stimulation (5 hours) of optoTAT caused an increase in mCherry-MRLC signal which saturated between 45–60 min and persisted for up to 5 hours (Fig. 4g). Myosin II activation leads to increased bundling of branched F-actin leading to an increase in symmetrical alignment, or coherence, of the actin cytoskeleton within subcellular regions⁷¹. We wondered whether acetylation of microtubules may promote a more coherent actin cytoskeleton through myosin activation. On measuring the coherence of the mCherry-MRLC distribution, we observed a significant increase in the coherence of mCherry-MRLC signal after 30 min of miRFP-optoTAT stimulation (Fig. 4h, i), suggesting a more symmetrically aligned actin and Myosin network. Catalytically dead miRFP-optoTAT(D157N) failed to elicit any increase in mCherry-MRLC signal. Additionally, any increase in mCherry-MRLC intensity was abrogated on treating the cells with ROCK inhibitor Y27632 for 10 minutes before optoTAT stimulation, or by pre-saturating microtubule acetylation by treating the cells with 2 μM Tubacin for 4 hours before optoTAT stimulation (Fig. 4j).

Consistent with an increase in myosin activity, miRFP703-optoTAT stimulation also led to increased levels of bundled actin (Fig. 4k, l), and maturation of focal adhesions as indicated by increased adhesion sizes and mCherry-Paxillin accumulation (Fig. 4m, n, o, Supplementary Movie S4), as well as induction of filopodia-like structures (Fig. 4k). Taken together, these data suggest that optoTAT stimulation induced Myosin activation through increased microtubule acetylation and ROCK kinase activation, suggesting a direct and causal role of microtubule acetylation in myosin activation.

Microtubule acetylation releases GEF-H1 from sequestration:

Myosin activation on MRLC phosphorylation through MLCK is often downstream of RhoA-ROCK signaling. Our observation that inhibiting ROCK abrogated optoTAT mediated Myosin activation suggested that optoTAT stimulation leads to RhoA activation. GEF-H1 is an activator for RhoA, which is sequestered on microtubules and is activated on disrupting microtubules using nocodazole^72,73. GEF-H1 was reported to mediate α-TAT1 mediated cellular mechano-sensing and microtubule acetylation was reported to inhibit GEF-H1 association with microtubules in astrocytes and endothelial cells, leading to RhoA activation^22,74. We performed immunostaining for GEF-H1 in WT and α-TAT1 KO MEFs and observed nearly two-fold higher GEF-H1 sequestration on the microtubules in α-TAT1 KO MEFs compared to WT MEFs (Fig. 5a, b). This was further confirmed using colocalization analysis, which also showed significantly higher colocalization of GEF-H1 signal with α-Tubulin signal (Fig. 5c). This higher microtubule associated GEF-H1 signal was not due to increased expression levels since both WT and α-TAT1 KO MEFs had comparable GEF-H1 expression levels (Fig. 5d, e). Since α-TAT1 KO cells showed relatively short-lived protrusions (Fig. 1k), we speculated if these cells have greater GEF-H1 activation in the protrusions. We did not observe any significant decrease in microtubule-bound GEF-H1 in the protrusions of α-TAT1 KO cells (Fig. 5a). However, we cannot rule out additional signaling pathways that may regulate activation of the soluble fraction of GEF-H1 in these cells.

Figure 5. Microtubule acetylation releases GEF-H1 sequestration.

a) α-Tubulin and GEF-H1 localization in WT and α-TAT1 KO MEFs, inset for GEF-H1 is magnified on right; b) linear density of GEF-H1 along microtubules in WT and α-TAT1 KO MEFs (5 microtubules from 30 cells each, total 150); c) Colocalization of GEF-H1 and α-Tubulin in WT and α-TAT1 KO MEFs, 20 cells each; d), e) GEF-H1 expression levels in WT and α-TAT1 KO MEFs measured using western blots (3 independent experiments, error bar: standard deviation); f) Distributions of acetylated/total α-Tubulin (top panel) and GEF-H1/total α-Tubulin (bottom panel) in WT MEFs, treated overnight with 100 nM Paclitaxel, full panel shown in supplementary Fig. S4e; g) Colocalization of acetylated microtubules and microtubule-bound GEF-H1 in Paclitaxel treated WT MEFs, n = 33 cells; h) GEF-H1 localization in HeLa cells expressing mCherry-α-Tubulin (WT) or acetylation deficient Cherry-α-Tubulin(K40A), Scale bar: 1 μm, full panel shown in supplementary Fig. S4f; i) GEF-H1 localization in α-TAT1 KO MEFs stably expressing mVenus-optoTAT kept in dark or with 30 min blue light stimulation; j) linear density of GEF-H1 along microtubules in α-TAT1 KO MEFs stably expressing mVenus-optoTAT kept in dark or exposed to 30 min blue light stimulation (5 microtubules from 30 cells each, total 150); k) Changes in linear density of GEF-H1 along microtubules in α-TAT1 KO MEFs stably expressing mVenus-optoTAT and exposed to blue light for the indicated durations, 5 microtubules from 12 cells, total 60 for each time point, mean ± 95% C.I.; l) Changes in mCherry-GEF-H1/mVenus-MAP4m signal in HeLa cells expressing miRFP703-optoTAT on blue light stimulation, inset is magnified in the right panels; m) Temporal changes in mCherry-GEF-H1/mVenus-MAP4m on miRFP703-optoTAT stimulation, mean ± 95% C.I. are shown, n = 33 cells; n) Colocalization of mCherry-GEF-H1 and mVenus-Map4m in Hela cells expressing miRFP703-optoTAT in dark or with 30 min blue light stimulation, n = 33 cells. Scale bars: 10 μm or as indicated. ***: p<0.001. Blue line: Blue light stimulation.

Acetylation modification is often associated with stable microtubules^6,8. To test if α-TAT1 KO MEFs had lower microtubule stability, we treated WT and α-TAT1 KO MEFs with 100 nM nocodazole for 1 hour, followed by fixation and immunostaining for α-Tubulin. Consistent with previous reports, a larger number of WT MEFs had nocodazole resistant microtubules compared to α-TAT1 KO MEFs (Supplementary Fig. S4a, b). Since acetylation has also been reported to stabilize microtubules by making them more flexible and resistant to bending forces¹⁵, we speculated whether release of GEF-H1 was due to increased stability or flexibility of microtubules in WT MEFs. Paclitaxel has been reported to increase both flexibility and mechanical stability of microtubules^75–77. We did not detect any significant loss of GEF-H1 sequestration in α-TAT1 KO cells on overnight treatment with 100 nM Paclitaxel (Supplementary Fig. S4c), suggesting that an increase in microtubule stability or flexibility did not significantly affect GEF-H1 sequestration.

To examine if the acetyl moiety on the Lysine-40 of acetylated α-Tubulin molecule specifically was responsible for GEF-H1 release, we co-immunostained for GEF-H1, acetylated microtubules and α-Tubulin in WT MEFs that were treated with 100-nM Paclitaxel overnight to eliminate any potential effects of microtubule stability. We observed a negative correlation between the spatial distribution of microtubule-bound GEF-H1 and acetylated microtubules (Fig. 5f, g, Supplementary Fig.S4d). To further test whether GEF-H1 specifically binds to non-acetylated microtubules, we exogenously expressed mCherry-α-Tubulin or acetylation deficient mCherry-α-Tubulin(K40A) mutant in HeLa cells and immunostained for GEF-H1. We observed increased microtubule-bound GEF-H1 in the cells expressing mCherry-α-Tubulin(K40A), compared to non-transfected cells, or those expressing WT mCherry-α-Tubulin (Fig. 5h, Supplementary Fig. S4e).

OptoTAT stimulation releases GEF-H1 from microtubules:

Our observations as well as previous reports²² suggest that microtubule acetylation leads to release of GEF-H1 from sequestration. However, these observations do not provide any information on the relative kinetics of microtubule acetylation and GEF-H1 release. To characterize this relationship, we stimulated α-TAT1 KO MEFs that were expressing mVenus-optoTAT with blue light for 30 min and performed immunostaining for GEF-H1. We observed significant reduction of GEF-H1 localization on microtubules in stimulated cells compared to those kept in dark. (Fig. 5i, j). To obtain greater temporal resolution of GEF-H1 release in response to optoTAT stimulation, we stimulated these cells for 0 min, 5 min, 10 min, 15, min, 20 min, 25 min, 30 min and 60 min with blue light and performed immunostaining for GEF-H1. We observed significant decrease in GEF-H1 signal within 5 min stimulation, reaching saturation around 30 min (Fig. 5k, Supplementary Fig. S5a). The decrease in GEF-H1 signal was not due to photo-toxicity since blue light stimulation of α-TAT1 KO MEFs not expressing optoTAT for 2 hours did not affect GEF-H1 association with microtubules (Supplementary Fig. S5b).

To examine the release kinetics of GEF-H1 on optoTAT stimulation in live cells, we exogenously expressed miRFP703-optoTAT with full length mCherry-GEF-H1 and EYFP-Map4m (amino acids 849-1115 of Map4a that bind to microtubules)⁷⁸, in HeLa cells. We stimulated these cells with blue light and measured the ratio of mCherry and EYFP signal, after correcting for photobleaching effects, to characterize changes in GEF-H1 localization on microtubules on optoTAT stimulation. Since YFP excitation was sufficient to activate optoTAT, we could not obtain any ‘before’ images except time zero. Nevertheless, we observed a persistent decrease in mCherry/EYFP signal, indicating a release of GEF-H1 from microtubules (Fig. 5l, m, Supplementary Fig. S5c, Supplementary Movie S5). To further account for any photobleaching artefacts, we performed cross-correlation analysis of mCherry and EYFP signal, which is less sensitive to signal magnitudes than ratiometric analysis. Cross-correlation analysis of mCherry and EYFP signal in dark or after 30 min stimulation also showed a decrease, further confirming release of GEF-H1 (Fig. 5n) from microtubules on optoTAT stimulation. Our observations demonstrate that the release of GEF-H1 followed a timeline comparable to that of microtubule acetylation as well as Myosin phosphorylation on optoTAT stimulation (Fig. 3i, 4d, 5k). These data suggest that microtubule acetylation rapidly releases sequestered GEF-H1 to activate RhoA and actomyosin contractility.

GEF-H1 release mediates microtubule acetylation dependent myosin activation:

To test whether GEF-H1 mediated optoTAT induced myosin activation, we used RNAi to deplete GEF-H1 in HeLa cells (Fig. 6a, b) and examined changes in mCherry-MRLC signal on miRFP703-optoTAT stimulation. Cells treated with siRNA against GEF-H1, but not the control siRNA, did not show significant increase in mCherry-MRLC signal on optoTAT activation (Fig. 6c). To examine whether release of GEF-H1 from microtubules was critical for microtubule acetylation mediated myosin activation, we used lentiviral transduction to stably express mCherry-GEF-H1(C53R) in α-TAT1 KO MEFs. GEF-H1(C53R) is a mutant that does not bind to microtubules (Fig. 6d) but retains its capability to activate RhoA⁷⁹. Exogenous expression of mCherry-GEF-H1(C53R) alone was sufficient to partially rescue the phospho-MRLC levels in α-TAT1 KO MEFs, which is dependent on ROCK kinase activity (Fig. 6e, f). The successful rescue by C53R, albeit partial, implies that microtubule acetylation predominantly affects GEF-H1 binding to microtubules, but perhaps no other pathways regulating GEF-H1 activity post-release from microtubule sequestration. Collectively, microtubule acetylation causes GEF-H1 release from microtubules, thereby inducing RhoA/ROCK-mediated actomyosin contraction.

Figure 6. GEF-H1 mediates microtubule acetylation dependent actomyosin contractility.

a), b) GEF-H1 knock-down in HeLa cells by siRNA tested with western blots (2 independent experiments, error bar: standard deviation); c) Changes in mCherry-MRLC intensity on miRFP703-optoTAT stimulation in HeLa cells with optoTAT (20 cells), scramble siRNA (21 cells), siRNA1 (25 cells) and siRNA3 (22 cells) against GEF-H1; d) TIRF images of HeLa cells expressing GFP-GEF-H1 (top panel) and mCherry-GEF-H1(C53R); e) Phospho-MRLC levels in WT (67 cells), α-TAT1 KO (69 cells), α-TAT1 KO MEFs expressing mCherry-GEF-H1(C53R) (67 cells) and same cells treated with 10 μM Y-27632 (60 cells); f) Phalloidin and phospho-MRLC distribution in WT, α-TAT1 KO MEFs and α-TAT1 KO MEFs expressing mCherry-GEF-H1(C53R); g) Rose plots of WT, α-TAT1 KO and KO-GEF-H1(C53R) MEFs migrating in a chemotactic gradient; g) Forward migration indices along the chemotactic gradient and h) Forward migration indices perpendicular to the chemotactic gradient for WT, α-TAT1 KO and KO-GEF-H1(C53R) MEFs, n = 120 cells (40 each from three independent experiments). Scale bars: 10 μm. ***: p<0.001

Microtubule binding deficient GEF-H1 rescues chemotaxis defects of α-TAT1 KO MEFs:

Our observations thus far suggest that microtubule acetylation promotes the release of GEF-H1, leading to increased actomyosin contractility and slower, but more directionally persistent, migration in MEFs. While these findings align with previous reports in astrocytes showing that microtubule acetylation promotes myosin activation, we observe an opposite motility phenotype: microtubule acetylation enhances astrocyte motility but inhibits MEF motility²². Notably, the effects on MEF motility more closely resemble those observed in HFFs, where microtubule acetylation reduces actomyosin contractility and leads to slower cell migration²³. Therefore, we sought to test whether the microtubule acetylation mediated release of GEF-H1 caused an increase in directional persistence in WT MEFs. If this is true, expressing GEF-H1(C53R) in α-TAT1 KO MEFs should rescue their defects in chemotaxis. We performed chemotaxis assay with WT, α-TAT1 KO MEFs and α-TAT1 KO MEFs expressing mCherry-GEF-H1(C53R) in 0–20% FBS gradient. α-TAT1 KO-GEF-H1(C53R) MEFs showed significantly improved chemotactic capability compared to α-TAT1 KO MEFs (Fig. 6f, g, h, Supplementary Fig. S5d, e). Although GEF-H1(C53R) was ubiquitously present in the cytoplasm (Fig. 6d), the α-TAT1 KO cells expressing mCherry-GEF-H1(C53R) were able to regain directional motility. This suggests that there are additional factors that spatially regulate GEF-H1 mediated actomyosin contractility to maintain front-and-back polarity. Altogether, these data demonstrate that microtubule acetylation drives directional migration in MEFs by modulating actomyosin contractility through the dynamic release of sequestered GEF-H1 in migrating cells. In light of previous reports showing that microtubule acetylation promotes astrocyte motility through myosin activation and inhibits HFF motility through myosin inhibition^22,23, our observations highlight the broad range of migration phenotypes that can arise from molecular crosstalk between microtubule acetylation and the actomyosin network.

Discussion:

Our data suggest that microtubule acetylation reduces overall motility in MEFs, but facilitates directional motility by promoting a dominant protrusion, whilst inhibiting nascent ones (Fig. 1). This coordination was achieved by modulation of actomyosin contractility and stabilizing adhesions in the dominant protrusion (Fig. 2). While microtubules have been implicated in focal adhesion turnover and actomyosin contractility through Rac1 and RhoA^80–86,73, the specific effects of microtubule acetylation on myosin activation is unclear. In astrocytes and HUVEC cells, microtubule acetylation promotes myosin activation through GEF-H1 release, whereas in HFFs microtubule acetylation inhibits myosin activity through MYPT1^22,23. Our data demonstrate that migration defects in α-TAT1 KO MEFs arise from decreased actomyosin contractility through sequestration of GEF-H1 in non-acetylated microtubules (Fig. 6). Consistently, we observed rapid myosin activation on optoTAT stimulation through GEF-H1 release (Fig. 4, 5) in MEFs and HeLa cells. Intriguingly, microtubule acetylation mediated increased actomyosin contractility promoted astrocyte migration²², whereas decreased myosin activation in HFFs due to microtubule acetylation inhibited migration²³. Our data suggest an overall decrease in MEF motility due to increased actomyosin contractility (Fig. 1, 2). One possibility behind these differing observations may be due to differences in experimental approaches used to perturb microtubule acetylation across these studies. We propose that optoTAT provides one of the most direct and rapid approaches to induce microtubule acetylation, thereby minimizing confounding factors related to HDAC6 inhibition or potential cell adaptation associated with genetic perturbations. While these variations in migratory phenotypes across different cell types suggest a context dependent role of actomyosin contractility in migrating cells, our findings using optoTAT provide evidence for a specific molecular coupling between microtubule acetylation in GEF-H1 release and myosin activation. It should be noted that optoTAT lacks the C-terminus of α-TAT1, which may prevent it from fully recapitulating all signaling pathways involving α-TAT1. Additionally, the nuclear sequestration of optoTAT may influence cell behavior. Nevertheless, the absence of myosin activation upon stimulation with catalytically inactive optoTAT(D157N) (Fig. 4) and the increased microtubule sequestration of GEF-H1 in cells expressing α-Tubulin(K40A) (Fig. 5, Supplementary Fig. S4) strongly support a specific role for microtubule acetylation in facilitating GEF-H1 release and myosin activation. We want to emphasize that our observations do not exclude the possibility that MYPT1 plays a role in microtubule acetylation-mediated regulation of myosin activity. Myosin activation is spatiotemporally modulated in directionally migrating cells. It is tempting to speculate that microtubule acetylation may control GEF-H1 and MYPT1 activity in distinct spatiotemporal manner to control myosin dynamics. Using optoTAT in these systems to test the impact of microtubule acetylation on myosin activation may help resolve these contradictory observations. The critical role of GEF-H1 in optoTAT mediated myosin activation and rescue of chemotaxis defects in α-TAT1 KO MEFs by microtubule non-binding GEF-H1(C53R) (Fig. 6) suggest that GEF-H1 mediates crosstalk between actin and acetylated microtubules. Additionally, our data suggest that GEF-H1 release is not mediated solely by microtubule stability, but also through the recognition of the acetyl moiety on α-Tubulin, directly or indirectly (Fig. 5, Supplementary Fig. S4). How GEF-H1 detects acetylated versus non-acetylated microtubules is an intriguing question. Since acetylation occurs in the microtubule lumen^6,9, one possibility is that GEF-H1 enters the lumen to read the acetylation state of microtubules. Although GEF-H1 (~100 kDa) is not a small molecule and its access to the narrow 15 nm diameter microtubule lumen appears difficult, larger molecules such as CSPP1 (~138 kDa) have been reported to exist in the microtubule lumen⁸⁷. The rapid release of GEF-H1 from microtubules, in such a case, would imply a somewhat permissive structure of a subset of microtubules, allowing for molecular exchange between the lumen and cytoplasm. Another possibility is that while GEF-H1 binds to the microtubule surface, it contains a domain which probes the lumen to detect the acetyl moiety, or the conformational changes in α-tubulin due to acetylation. It is also possible that GEF-H1 localization on microtubule is controlled by a third-party molecule that directly senses the acetylation state of microtubules, and relays that information to GEF-H1^79,88.

Spatiotemporal regulation of GEF-H1 activation by microtubule dynamics has been reported by many groups, directly coupling microtubule and actin cytoskeletal dynamics in migrating cells^73,89,90. In addition to dynamic instability of microtubules, various signaling pathways have also been reported to regulate GEF-H1 association with microtubules, and thereby its activation, including PAK, PKA, MARK2 and MARK3 kinases, as well as 14-3-3 adaptor proteins^79,91,92. In order to achieve directional persistence in migrating cells, RhoA and myosin activation must be spatiotemporally regulated at sites of dynamic actin remodeling⁹³, and GEF-H1 activation has been reported to be spatiotemporally regulated^90,94. How, or even if, microtubule acetylation mediates spatiotemporal control of GEF-H1 activation is an intriguing question. One possibility is that microtubule acetylation only releases GEF-H1 to increase the cytosolic pool, where GEF-H1 may be activated in subcellular regions through additional factors such as Src kinase, ERK1/2 kinases and Phosphatase 2A^91,94–97. Our data show comparable kinetics of optoTAT mediated microtubule acetylation, GEF-H1 release, myosin activation and adhesion maturation (Figs. 3, 4, 5), indicating a direct and causal coupling of these events, and not one of long-term cellular adaptation. α-TAT1 KO MEFs expressing GEF-H1(C53R) are capable of directional migration (Fig. 6), suggesting that microtubule acetylation releases GEF-H1 from sequestration, but does not regulate subsequent activation. Of course, we cannot rule out cellular adaptation in this instance. OptoTAT design does not allow subcellular activation of microtubule acetylation, thus limiting our capability of probing the effects of spatially restricted microtubule acetylation on cell behavior. α-TAT1 localizes to focal adhesions through Talin binding²², which may provide localized interaction with microtubules to facilitate spatially regulated GEF-H1 release, followed by activation through other pathways. Spatial distribution of GEF-H1 may be further fine-tuned by combination of microtubule assembly-disassembly, lattice flexibility and acetylation state. Further examination of the spatial regulation of microtubule dynamics, microtubule acetylation, MYPT1 and GEF-H1 activation will help us better understand their interplay in migrating cells.

Materials and Methods

Cell culture and transfection:

HeLa and HEK-293T cells were cultured in DMEM basal media and passaged every third day of culture. For optimal growth, the media were supplemented with 10% (v/v) fetal bovine serum, L-Glutamine, Penicillin/Streptomycin, Non-essential amino acids and 0.05 mM β-mercaptoethanol. WT and α-TAT1 KO MEFs were a generous gift from Dr. Maxence Nachury and were cultured in DMEM basal media supplemented with 10% (v/v) fetal bovine serum, L-Glutamine, Penicillin/Streptomycin, Non-essential amino acids and 0.05 mM β-mercaptoethanol. HeLa cells, WT MEFs and α-TAT1 KO MEFs stably transduced with mVenus-α-TAT1, mVenus-α-TAT1(D157N), mCherry-MRLC, mVenus-optoTATV2 or mCherry-GEF-H1(C53R) were sorted using the Sony SH800 cell sorter using manufacturer’s instructions to select cell populations with similar mVenus or mCherry fluorescence thresholds to ensure similar expression levels of the proteins of interest. The cells were maintained under standard cell culture conditions (37 °C and 5% CO2) and were checked for mycoplasma contamination prior to use in experiments. The stably transduced cells were cultured in medium containing 1 μg/ml of puromycin. Effective puromycin dosage was ascertained by testing on WT and α-TAT1 KO MEFs. FuGENE 6 reagent (Promega, Madison, WI) was used for transient transfection of HeLa cells according to the manufacturer’s instructions. For generation of lentiviral particles, HEK-293T cells were transfected using polyethyleneimine (Kyfora Bio PEI MAX, Fisher Scientific catalog # 50-255-9821). Electroporation with Lonza electroporator was performed for expression of VinTS or in WT and α-TAT1 KO MEFs, according to the manufacturer’s instructions.

DNA plasmids:

α-TAT1 plasmid construct was a gift from Dr. Antonina Roll-Mecak. VinTS was a gift from Dr. Martin Schwartz (Addgene plasmid # 26019). mCherry-MRLC and mCherry-Paxillin were a gift from Dr. Yi I. Wu. NLS-mCherry-LEXY was a gift from Dr. Barbara Di Ventura & Dr. Roland Eils (Addgene plasmid # 72655). Z-lock αTAT was a gift from Dr. Klaus Hahn (Addgene plasmid # 175290). GFP-GEF-H1 was a gift from Dr. Hiroaki Miki. As indicated in the results and figure legends, tags of compatible fluorescent proteins including mCerulean, mVenus, mCherry and miRFP703 were appended to facilitate detection of the proteins of interest and the plasmids were subcloned into C1 vector (Clontech) or pTriEx4 vector (Novagen). Unless specified otherwise, the termini of tagging were positioned as in the orders they are written. Lentiviral plasmids were generated based on a modified Puro-Cre vector (Addgene plasmid # 17408, mCMV promoter and no Cre encoding region). Point mutations or truncations of indicated plasmid constructs were generated by PCR. The open reading frames of all DNA plasmids were verified by Sanger sequencing.

Drug treatments:

Pharmacological drugs were purchased as indicated: Y-27632 (LC Laboratories, catalog # Y-5301), Tubacin (Selleck Chemicals, catalog # S2239), Taxol or Paclitaxel (Cell Signaling Technology, catalog # 9807S), Nocodazole (Cell Signaling Technology, catalog # 2190). Y-27632 was applied at 10 μM final concentration 60 min before fixing cells or initiation of microscopy. Tubacin was applied at 2 μM final concentration for indicated durations before initiation of microscopy. Taxol was applied at 100 nM final concentration overnight before fixing cells. Nocodazole was applied at 100 nM final concentration for 30 min before fixing cells.

Immunofluorescence assays:

For immunostaining of acetylated α-Tubulin, total α-Tubulin or GEF-H1, cells were fixed using ice-cold methanol for 10 minutes, washed thrice with cold PBS, blocked with 2% BSA in PBS for one hour and then incubated overnight at 4°C with antibodies against α-tubulin (rat, MilliporeSigma, MAB1864), acetylated α-Tubulin (mouse, MilliporeSigma, T7451) or GEF-H1 (rabbit, ThermoFisher, PA5-32213). Next day, the samples were washed thrice with cold PBS and incubated with secondary antibodies (Invitrogen) for one hour at room temperature, after which they were washed thrice with PBS and images were captured by microscopy. For immunostaining of Vinculin, phospho-MRLC, Myosin IIa, detyrosinated α-Tubulin, polyglutamylated tubulin and α-Tubulin, cells were fixed using freshly prepared 4% paraformaldehyde at room temperature for 10 minutes, washed twice with PBS, blocked and permeabilized in 1% BSA in PBS with 0.1% TritonX-100 at room temperature for an hour and then incubated with antibody against Vinculin (mouse, Sigma Aldrich, MAB3574), phosphor-MRLC (rabbit, Cell Signaling Technology, 3671T), Myosin IIa (rabbit, Cell Signaling Technology, 3403T), detyrosinated α-Tubulin (rabbit, Abcam, ab48389), polyglutamylation (rabbit, Adipogen, GT335) in the above blocking buffer at room temperature for 1 hour, washed three times in PBS and incubated with corresponding secondary antibodies (ThermoFisher A48268, A21052, A21206, A21202, A31571, A11057, A21058, A21096, A11032), Phalloidin (ThermoFisher A22286) and DAPI (Cell Signaling Technology, 4083S). After that, they were washed three times in PBS and the images were captured by microscopy. For HeLa cells transiently transfected with mCherry-α-Tubulin or mCherry-α-Tubulin(K40A), fixing and immunostaining were performed 24 hours post-transfection. WT or α-TAT1 KO MEFs were treated with 10 μM Y-27632 or equal volume of vehicle (water), incubated for 1 hour, followed by PFA fixation and immunostaining. WT or α-TAT1 KO MEFs were treated with Taxol (100 nM) or vehicle (DMSO) overnight followed by methanol fixation and immunostaining. WT or α-TAT1 KO MEFs were treated with Nocodazole (100 nM) for 30 min followed by PFA fixation and immunostaining.

Western blot assays:

Cell lysates were prepared by scraping cells using lysis buffer (RIPA buffer, Cell Signalling # 9806S), mixed with protease/phosphatase inhibitor cocktail (Cell Signaling # 5872S). Cell lysates were rotated on a wheel at 4°C for 15 min and centrifuged for 10 min at 15,000 g 4°C to pellet the cell debris, mixed with NuPAGE^™ LDS Sample Buffer (Thermo Fisher # NP0007) with protease and phosphatase inhibitors and boiled 5 min at 95°C before loading in polyacrylamide gels. Gels were transferred and membranes were blocked with TBST (0.1% Tween) and 5% BSA and incubated overnight with the primary antibody, and 1 h with LiCOR IR-dye conjugated secondary antibodies after which bands were revealed using Odyssey imaging system. Primary Antibodies used: anti-α-Tubulin (rat, MilliporeSigma, MAB1864), anti-Vinculin (rabbit, Cell Signaling Technology, 13901S), anti-GEF-H1 (rabbit, ThermoFisher, PA5-32213). Secondary IR-dye conjugated antibodies were purchased from LiCOR.

Microscopy and image analyses:

All epifluorescence imaging was performed with an Eclipse Ti microscope (Nikon) with PCO.Edge sCMOS camera (Excelitas) or ORCA-FusionBT sCMOS camera (Hamamatsu), driven by NIS Elements software (Nikon). All TIRF imaging was performed with an Eclipse Ti microscope (Nikon) with ORCA-FusionBT sCMOS camera (Hamamatsu), driven by NIS Elements software (Nikon). All confocal images were captured using a laser scanning confocal Eclipse Ti2 microscope (Nikon) equipped with a tunable GaAsp detector and 2k resonant scanner (AXR, Nikon) with camera, driven by NIS software (Nikon). All live cell imaging was conducted at 37°C, 5% CO₂ and 90% humidity with a stage top incubation system (Tokai Hit). Vitamin and phenol red-free media (US Biological) supplemented with 2% fetal bovine serum were used in imaging to reduce background. Inhibitors and vehicles were present in the imaging media during imaging. All image processing and analyses were performed using Metamorph (Molecular Devices, Sunnyvale, CA, USA) and FIJI software (NIH, Bethesda, MD, USA). OptoTAT stimulation was provided by epifluorescent 440 nm excitation 20 s apart, or continual exposure to blue LED light (Amazon, B08FQSFFJ60). Cells that were rounded up or showed a high degree of blebbing were excluded from analysis to minimize artifacts from mitotic, apoptotic or dead cells. For ratiometric or intensity analyses, background subtraction and ibased on a cell free area on each image and manual intensity based thresholding was performed prior to quantification. For immunostaining assays, individual cells were identified based on Phalloidin or α-Tubulin staining. For colocalization analysis, Coloc2 function in FIJI was used to calculate the Pearson’s correlation coefficient (also called Pearson’s R) value. Images containing any saturated pixels in any channel (65535 value) within the cell area were excluded. The ratio of acetylated α-Tubulin over α-Tubulin (Ac. α-Tub/α-Tub) for transiently transfected cells was normalized against that for non-transfected cells averaging over 20 non-transfected cells from the same dish, indicated at the dashed-line at value 1. For kinetics of microtubule acetylation in HeLa cells of α-TAT1 KO MEFs stably expressing mVenus-optoTAT, only ratio of acetylated α-Tubulin over α-Tubulin (Ac. α-Tub/α-Tub) for individual cells are shown. Changes in mCherry-MRLC coherence in HeLa cells before and after optoTAT stimulation was analyzed by OrientationJ plugin in FIJI⁹⁸. mCherry-MRLC flow in WT or α-TAT1 KO MEFs were characterized using the Optic Flow plugin in FIJI with Gaussian window motion structure element. The mCherry image was used to segment and mask the cell area in the resulting optical flow magnitude image, and the segmented image was used for quantifying the MRLC flow.

Cell migration assays:

Random migration assays were performed in 24-well polystyrene tissue culture-treated plates. 1.5 × 10⁴ WT MEFs or 1.2 × 10⁴ α-TAT1 KO MEFs were seeded and incubated for 4–5 hours. After attachment, cells were imaged using phase contrast microscopy with 10X objective every 10 minutes for 15 hours.

The wound healing assay was performed using Ibidi Culture-Insert 3 Well in 24-well plates (Ibidi, catalog # 80369). 3.5 × 10⁴ WT and α-TAT1 KO MEFs were seeded in each insert well and incubated for 4–5 hours or until a full monolayer was created. After cells were settled, the 3-well insert was removed, creating a 500 μm cell-free area between cell monolayers. The existing media was aspirated, and the wells were gently washed with PBS once to remove any floating cells, followed by addition of fresh medium prior to imaging. Cells were imaged using phase contrast microscopy with 10X objective every 5 minutes for 15 hours. The wound closure rate was calculated by determining the area of the cell-free area over time.

Chemotaxis assay was performed using μ-slide chemotaxis chambers coated with collagen IV (Ibidi, catalog # 80322) following the manufacturer’s protocol. In short, 2.5 × 10⁶ cells/mL αTAT1 KO MEFs, 3 × 10⁶ cells/mL WT MEFs or 3 × 10⁶ cells/mL KO MEFs rescued with mVenus-α-TAT1 or mCherry-GEF-H1 (C53R), were seeded with serum-free DMEM and incubated at 37°C, 95% humidity, and 5% CO2 for 4–5 hours. After cells were settled, serum-free DMEM was added to the right and left reservoirs of the chamber. Half of the volume of the left reservoir was replaced with DMEM supplemented with 20% FBS to generate the chemoattractant gradient. Cells were imaged using phase contrast with 10X objective every 10 minutes for 15 hours at 37°C and 5% CO2.

Tracking and analyses were performed using ImageJ plug-ins, MTrackJ, and Chemotaxis_tool, respectively. The total pathlength traversed over 15 hours was measured to calculate cell speed (Fig. 1b). Displacement of each cells was measured as the linear distance between the position at time 0 and at time 15 hours. Directionality (Fig. 1c) for each migrating cell was measured as the ratio of displacement and pathlength over 15 hours. To compute directional persistence over a fixed pathlength (Supplementary Fig. S1a), we measured displacement for a fixed cumulative pathlength of around 65 μm (~ 50 pixels) for WT and α-TAT1 KO MEFs.

Statistical analyses and reproducibility:

Microsoft Excel (Microsoft, Redmond, WA, USA) and R (R Foundation for Statistical Computing, Vienna, Austria) were used for statistical analyses. The exact number of samples for each data set is specified in the respective figure legends.

For live cell assays, at least three independent experiments were performed on different days, and data from these experiments were pooled for data analyses. Unless otherwise specified, data for live cell assays with optoTAT stimulation were normalized to the value at time zero just before blue light stimulation. For immunocytochemistry-based assays, at least two independent experiments were performed, from which one experiment set was used for data analyses. Each group of immunostaining experiments were performed with cells plated on the same 8 well chambers (Cellvis, C8-1.5H-N), on the same day with the same reagents and imaging performed under the same conditions to account of experimental variability. For western blot assays, at least 2 independent experiments were performed, and data were pooled for quantification. Sample sizes were chosen based on the commonly used range in the field without performing any statistical power analysis and assumed to follow normal distribution. P-values were obtained from two-tailed Student’s t-test assuming equal variance or paired t-test where applicable.

Data availability

All data and plasmid constructs will be made available on reasonable requests.