Optogenetic modulation of guanine nucleotide exchange factors of Ras superfamily proteins directly controls cell shape and movement

In this article, we provide detailed protocols on using optogenetic dimerizers to acutely perturb activities of guanine nucleotide exchange factors (GEFs) specific to Ras, Rac or Rho small GTPases of the migratory networks in various mammalian and amoeba cell lines. These GEFs are crucial components of signal transduction networks which link upstream G-protein coupled receptors to downstream cytoskeletal components and help cells migrate through their dynamic microenvironment. Conventional approaches to perturb and examine these signaling and cytoskeletal networks, such as gene knockout or overexpression, are protracted which allows networks to readjust through gene expression changes. Moreover, these tools lack spatial resolution to probe the effects of local network activations. To overcome these challenges, blue light-inducible cryptochrome- and LOV domain-based dimerization systems have been recently developed to control signaling or cytoskeletal events in a spatiotemporally precise manner. We illustrate that, within minutes of global membrane recruitment of full-length GEFs or their catalytic domains only, widespread increases or decreases in F-actin rich protrusions and cell size occur, depending on the particular node in the networks targeted. Additionally, we demonstrate localized GEF recruitment as a robust assay system to study local network activation-driven changes in polarity and directed migration. Altogether, these optical tools confirmed GEFs of Ras superfamily GTPases as regulators of cell shape, actin dynamics, and polarity. Furthermore, this optogenetic toolbox may be exploited in perturbing complex signaling interactions in varied physiological contexts including mammalian embryogenesis.


Introduction
Directed cell migration is a highly orchestrated phenomenon fundamental to various physiological functions, including embryogenesis, wound healing, immune response, and cancer metastasis (Luster et al., 2005;Kruger et al., 2015;Devreotes et al., 2017;SenGupta et al., 2021). Dynamic, directed migratory events require precise spatiotemporal regulation of small GTPases, namely, Ras, Rac, and Rho. Guanine nucleotide exchange factors (GEFs) and GTPase-activating proteins (GAPs) perform this crucial function; GEFs catalyze the exchange of GDP to GTP to activate GTPases while GAPs accelerate GTP hydrolysis rate to turn them off. Since Ras GTPases transduce signals from chemoattractant-stimulated G-protein coupled receptors to downstream Rac and Rho GTPases to coordinate cytoskeletal activities, their respective GEFs and GAPs are vital network components of directed migration (Bos et al., 2007;Cherfils and Zeghouf, 2013;Artemenko et al., 2014;Devreotes et al., 2017;Pal et al., 2019;Gray et al., 2020).
Although GEFs have been implicated in migration through knockout, knockdown, or overexpression studies, it has been difficult to assign specific roles for them (Insall et al., 1996;Uhlenbrock et al., 2004;Francis et al., 2006;Nalbant et al., 2009;Pakes et al., 2012;Suire et al., 2012;Ding et al., 2018). This is largely because phenotypes observed from conventional genetic and biochemical assays require multiple days, or even months, to develop allowing sufficient time for signaling and cytoskeletal networks to re-adjust through differential gene expression or protein rearrangement (Kok et al., 2015;Rossi et al., 2015;Stainier et al., 2015;Morgens et al., 2016;El-Brolosy and Stainier, 2017;Housden et al., 2017;El-Brolosy et al., 2019). Moreover, these tools lack the spatial resolution to locally activate GEFs and examine their effects on confined actin organization and directed migration.
In recent times, light-induced dimerization systems enabled local and reversible perturbation of upstream G-protein coupled receptor and Ras signaling, or the activities of downstream cytoskeletal components, such as Rho, Rac, or Cdc42, and steered migration. For acutely modulating the activities of these small GTPases, the catalytic domain of their respective GEFs was optically recruited to the plasma membrane (Yoo et al., 2010;O'Neill et al., 2018;Wu et al., 2009;O'Neill et al., 2016;Wu et al., 2011;Pal et al., 2023;Kato et al., 2014;de Beco et al., 2018;Bell et al., 2021;Valon et al., 2015;O'Neill and Gautam, 2014;Karunarathne et al., 2013;Krueger et al., 2019;Zhou et al., 2012;Wang et al., 2010). In this article, we demonstrate how to use these cryptochrome-and lightoxygen-voltage-sensing (LOV) domain-based optical tools to acutely clamp activities of Ras, Rac, and Rho GEFs on the cell membrane of different cell types (Figure 1). Our optical perturbations have brought forth a multitude of cytoskeletal and migratory responses which would have been challenging with previous methods evaluating GEF function.

Imaging system and reagents
Zeiss LSM800 GaAsP single-point, laser scanning confocal microscope with a wide-field camera.

Immunoblotting system
For more details, see Supplementary Materials and Methods.
All constructs were sequenced and verified at the JHMI Synthesis and Sequencing Facility.

Stable cell line construction
For stable protein expression in HL-60 cells, a combination of lentiviral and transposon approaches was used sequentially Hadjitheodorou et al., 2021;Li et al., 2021;Pal et al., 2023). The procedure involved three steps. In step 1, virus was made using HEK293T cells at~80% confluency. For each transfection, 2 µg pMDLg/pRRE, 4.64 µg pRSV-Rev, 3.32 µg pMD2. G, and 10 µg CIBN-CAAX/pLJM1 constructs were mixed with Lipofectamine 3000 transfection reagent as per manufacturer's instructions (Dull et al., 1998). After 96 h, virus-rich culture medium was collected at 3000 rpm for 20 min at 4°C. Next, viral supernatant was added to 4 × 10 6 HL-60 cells, seeded at a density of 0.25 × 10 6 cells/mL, in a 6-well plate. 10 μg/mL polybrene was added to this cell-virus mixture. Post 24 hour-incubation, virus was aspirated, and infected cells were added to a mixture of conditioned and fresh (mixed) medium. To select CIBN-CAAX-expressors, 1 μg/mL puromycin was added to infected cells after a day of recovery. Infected cells were kept in presence of puromycin for 5 days till only resistant cells were alive. Next, cells were harvested and transferred to a 48-well cell culture plate without puromycin for 3 weeks till resistant cells grew to confluency. These CIBN-CAAX expressing cells were next infected with LifeAct-miRFP703-expressing virus as explained above. To select LifeAct-miRFP703 expressors, cells were sorted on the fifth day post-infection, and grown to confluency. Finally, cells expressing CIBN-CAAX and LifeAct-miRFP703 were maintained in puromycin. Next, recruitable RasGRP4 was introduced in CIBN-CAAX and LifeAct-miRFP703 dual expressing HL-60 cells using PiggyBac transposon system. 5 μg CRY2PHR-mCherry-RasGRP4/pPB plasmid was co-electroporated with an equal amount of transposase expression plasmid (pCMV-hyPBase) into 2 × 10 6 cells using Neon transfection kit. DNA-cell mixture was resuspended in buffer "R" before single-pulse electroporation was carried out in a 100 µL pipette at 1350 V for 35 m sec using Neon electroporation system. Cells were resuspended in mixed culture medium in a 6-well plate and selected in presence of 10 μg/mL blasticidine S, as described for puromycin. Finally, these triple expressors were cultured throughout in puromycin and blasticidine S.

Optogenetic recruitment
All optogenetic experiments were done in the absence of any chemoattractant. Photoactivation was done with argon laser (488 nm excitation). All recruitable effectors, tagged with mCherry, RFP, or tdTomato were visualized with solid-state laser (561 nm excitation and 579-632 nm emission). miRFP703 or iRFP670 was excited with 633 nm diode laser and emission was   (Kuhn et al., 2021;Banerjee et al., 2022;Banerjee et al., 2023;Pal et al., 2023). For photoactivation of differentiated HL-60 cells, we discovered that their pre-treatment with heat-killed Klebsiella aerogenes greatly improved efficiency of cryptochrome system. Briefly, 10 7 differentiated neutrophils, grown on 10 cm culture dish, were incubated with 13 μg/mL heat-killed bacteria for 7 h. Next, cells were allowed to adhere to chambered coverglass, coated with fibronectin at a density of 35-40 μg/cm 2 , for 40 min. Dead bacteria and unattached neutrophils were rinsed off from fibronectin-coated surface before imaging. We performed global recruitment experiments on Zeiss LSM780 microscope. The 488 nm excitation laser was turned on after imaging for at least 5 min. Image acquisition and subsequent photoactivation were carried out once every 7 s. Laser intensity during image capture was maintained at a low level (laser power of 1.7%-2% or 0.14-0.17 W/cm 2 at the objective) and exposure time was set at~2 s. This ensured RasGRP4 recruitment over the cell boundary without causing light damage. All local recruitment experiments were done using the Zeiss LSM800 microscope. A small region of interest was drawn near the cell (shown as dashed white boxes in the images), which was illuminated with the 488 nm laser (power of 5%-7% or 0.6-0.8 W/ cm 2 at the objective) in multiple iteration. All imaging was completed within 5 h . Before imaging RAW 264.7 cells, culture medium was removed and 450 µL pre-warmed HBSS buffer was added to cells. For MCF-10CA1h cell imaging, culture medium was aspirated and replaced with 500 µL pre-warmed imaging medium. Image acquisition and photoactivation were carried out once every 7 s or 8 s for macrophage or MCF-10CA1h cells, respectively. Other parameters for recruitment studies were similar to HL-60 cells.
Vegetative and developed Dictyostelium cells were allowed to adhere on chambered coverglass for 30 min before imaging. For recruitment, 488 nm laser was switched on after imaging for at least 5-10 min. Photoactivation during single plane imaging was carried out once every 5-15 s since half-life of iLID-SspB was~30 s. Since Dictyostelium are light sensitive, very low laser intensity (0.017-0.030 W/cm 2 or 0.06-0.08 W/cm 2 at the objective on the Zeiss LSM780 or LSM800, respectively) was used to stably recruit SspB-RacGEF over the cell boundary without any light damage.

Cell sorting
HL-60 cells, 5 days after infection, were harvested, washed twice and resuspended in sorting buffer at a density of 15 × 10 6 cells/mL. Cells were declumped by passing suspension through 40 μm cell strainer once and collected in 5 mL round-bottom tubes. Wild type HL-60 cells (2 × 10 5 cells resuspended in 300 µL sorting buffer) was used as unstained control. Sorting was done at the Ross Flow Cytometry Core and Centre for Cell Dynamics, JHU using 100 μm microfluidic sorting chip. 561 nm excitation laser was used to sort and collect RFP expressing cells, whereas 633 nm excitation was used to sort miRFP703 expressors. The detector has 735LP and 780/60BP. High expressors (top 1%-10%) were taken in 0.5-5 mL collection medium, spun down, and medium was discarded. Sorted cells were resuspended in fresh collection medium and grown to confluency.

Data analysis
For Figure 1, Figure 2D, Figure 2I, Figure 3E, Figure 5D, and Supplementary Figure S6B, linescans were created on Fiji/ImageJ 1.52i software (https://imagej.nih.gov/ij/) (Schneider et al., 2012). On the red channel, a "straight line" segment (12-pixel width) was drawn across the cell using the 'line tool' option. Next, we obtained the average intensity value along that line using the "Plot Profile" option. Values were normalized and graphed in Microsoft Excel.
For Supplementary Figure S3C, turn-off rate (membranecytosol cycling) for the mutants was calculated by counting the number of frames from the last flash of the laser till the signal in the red channel was no longer present on the cell boundary. The frame with the last flash of the 488 nm laser was considered as "0 s" and frame rate was 7 s. For Supplementary Figure S3E, red signal intensity in the cytosol and nucleus was measured immediately before or after turning on the laser. This allowed us to calculate percentage decrease in signal intensity in the cytosol/nucleus with photoactivation. We used these values to obtain the percentage of total recruitable protein translocated to membrane for each cell. GraphPad Prism 8 was used to prepare box-and-whisker plots.
For Supplementary Figure S4B, acquired images from different experimental conditions were adjusted to the same contrast by ImageJ software. Cells were outlined by "polygon tool" in ImageJ and the mean intensity for each was recorded. GraphPad Prism 8 (https://www. graphpad.com/) was next used to prepare the scatter plot.
For Figure 4B, cell was first segmented against the background with help of a custom code written in MATLAB 2019b (https://www. mathworks.com/). Next, membrane kymograph was created from the segmented cell as described previously (Banerjee et al., 2022;Banerjee et al., 2023;Pal et al., 2023). A linear color map for normalized intensities was used; blue denoted the lowest intensity whereas yellow denoted the highest.
For Figures 4C,D, and Figures 6B-D, cells were segmented in Fiji/ImageJ 1.52i software . This was done in a stepwise manner. First, using the "Threshold" option, the image stack was thresholded. We made sure that the "Calculate threshold for each image" box was kept unchecked, and the range was not reset. Second, using the "Analyzed Particles" option, we created cell masks by size-based thresholding. Third, these binary masks were optimized by performing "Fill holes", "Dilate", and "Erode" multiple times. Fourth, "Area", "Shape descriptors", "Centroid", "Min and max gray value" and "Mean gray value" boxes in the "Set Measurements" tab under "Analyze" were checked. This allowed us to obtain values for aspect ratio and centroid coordinates. Using GraphPad Prism 8, mean and SEM obtained from aspect ratio value replicates were plotted. We calculated velocity by quantifying displacement between two consecutive frames. Subsequently, cell speed was acquired by dividing displacement with time interval. Using GraphPad Prism 8, average cell speed values, obtained from time-averaging cell speeds over all the frames, were plotted as box-and-whisker plots.
For Figure 4F and Figure 6, local protrusion formation was analyzed in a stepwise manner (Banerjee et al., 2022;Pal et al., 2023). First, using the "segmented line" tool in Fiji/ImageJ software, GEF recruitment region in the red channel was marked. Second, the midpoint of the recruited region was determined with the sequential use of customized macros, "Fit Spline" and "Straighten". Third, after having determined the centroid with assistance from another Fiji/ImageJ macro, the angle between the nascent protrusion and recruitment region midpoint was calculated using the "angle" tool, holding the centroid as vertex. Using "polarhistogram" in MATLAB, these values were plotted. Atleast 35 fresh protrusions were considered for each histogram, and the minimum number of bins for each plot was calculated by Sturges' formula.
In Figures 5E,F, cell was binarized using ImageJ based on the LifeAct channel. One typical Life-act labelled actin patch, as indicated by pink arrows in Figure 5B, was quantified throughout  Figure S4B, paired 2tailed non-parametric test (Wilcoxon matched-pairs signed rank test) was used. Results are expressed as mean ± SD from at least 3 independent experiments. ns denotes p > 0.05, ** denotes p ≤ 0.01, *** denotes p ≤ 0.001, **** denotes p ≤ 0.0001.

Establishment of different cryptochrome and iLID optogenetic systems in multiple cell lines
In this study, we developed and utilized blue light-inducible dimerization systems to acutely perturb Ras superfamily GEFs in migrating cells. As discussed in the methods section, we generated stable cell lines, co-expressing cryptochrome optogenetic system with F-actin biosensor, LifeAct-miRFP703, in differentiated HL-60 neutrophils. Interestingly, western blot and confocal imaging results showed that treating these triple expressors with heat-killed bacteria, before experimentation, improved expression levels of CRY2PHR recruitable component as well as LifeAct, in a time-dependent manner (Supplementary Figures S1A,B). We presume that this pretreatment would have a similar effect on expression of the membrane anchor, CIBN-CAAX, but we could not test that in our study. When we turned on the 488 nm laser on the entire cell, it caused cytosolic CRY2PHR to recruit globally on the plasma membrane within seconds. LifeAct expression strongly defined the front of these migrating cells (Figures 2A,B).
Could we spatially confine cryptochrome recruitment to a certain region on the neutrophil membrane? To answer this, we selectively turned on the blue light at the back of the cell, as denoted by the dashed white box. Time lapse imaging and linescan analysis clearly showed that nearly half of the total cytosolic CRY2PHR protein translocated to the illuminated region of the membrane within 20 s (Figures 2C-E). This result led us to ask whether the distance of the illuminated region (dashed white box) from the cell boundary would have any effect on localized recruitment. When the laser was directed far from the cell boundary, we did not observe any appreciable recruitment (Supplementary Figure S2A). Once the blue light was applied near the boundary, CRY2PHR recruitment was strongly Frontiers in Cell and Developmental Biology frontiersin.org localized near the illumination region ( Figure 2C; Supplementary Figure S2B). However, when we shined the laser within the cell boundary, it led to global recruitment of the CRY2PHR protein (Supplementary Figure S2C). These data suggest that blue light is travelling past the edges of the region of illumination, upto a certain distance, leading to unwanted recruitment. Hence, we performed all localized illumination experiments by applying blue light just outside the cell boundary. With our success in neutrophils, we attempted to establish the CRY2-CIBN system in Dictyostelium amoeba. We were successfully able to express CRY2PHR (red) and CIBN-CAAX (untagged) in developed and vegetative Dictyostelium as demonstrated by our global and local recruitment experiments, respectively ( Figures  2F,H). CRY2PHR recruitment was fairly uniform upon global illumination with blue light, whereas it could be spatially constricted to a specific part of the cell membrane by localized irradiation (Figures 2F-J). Moreover, light-sensitive Dictyostelium seemed to be unaffected with the blue laser as suggested by their polarized morphology and motility ( Figures 2F,H).
To aid our investigation, we characterized several fast-cycling cryptochrome mutants. In RAW 264.7 macrophages, both CRY2PHR (W349R)-mCherry and CRY2PHR (R489E, A491D)-tdTomato mutants translocated back to the cytosol within 3 min and 4.5 min of switching off the blue laser, respectively. This is a significant improvement from the turn-off rate of~9.5 min for CRY2PHR-mCherry ( Supplementary Figures S3A-D). Moreover, 40% of total CRY2PHR (R489E, A491D)-tdTomato mutant translocated to the membrane after irradiation, compared to 21% or 23% for CRY2PHR-mCherry or CRY2PHR (W349R)-mCherry, respectively (Supplementary Figure S3E). The primary reason for this is that a major portion of total CRY2PHR-mCherry or CRY2PHR (W349R)-mCherry localized within the nucleus upon expression, whereas CRY2PHR (R489E, A491D)-tdTomato expressed entirely in the cytosol and was excluded from the nucleus (Supplementary Figures S3A-C). Thus, we identified CRY2PHR (R489E, A491D)-tdTomato mutant to be the most recruitable cryptochrome with a relatively fast membrane-cytosol turnover, making it ideal for studying time-critical biological reactions.
In addition to the CRY2-CIBN system, we established an improved Light Induced Dimerization (iLID) optogenetic system in mammalian and amoeba cells to examine GEF function. As shown in Figures 3A,F, SspB component was optically recruited uniformly on the macrophage cell membrane, and soon after switching off the light source, translocated back to the cytosol. SspB could be subsequently recruited after switching on the laser again. With localized illumination, as shown with the dashed white box, we noticed a distinct crescent of recruited SspB on the membrane, which gradually diminished within a minute of taking away the light ( Figures 3B,G). In Dictyostelium, we integrated SspB with a doxycycline-inducible expression system under the inducible promoter, TRE-P min , which consists of seven repeats of the TetO operator fused to a small fragment of the act15 promoter (Veltman et al., 2009). This allowed us to tightly control expression of recruitable SspB in a time-dependent manner upto 14 h, making this a useful tool for preventing build-up of potentially toxic effector proteins for optical studies (Supplementary Figures   S4A,B). Irradiating 488 nm laser on these doxycycline-treated cells elicited uniform membrane recruitment within 10 s (middle panel; Figure 3C). With prolonged exposure to light, we observed recruitment in more cells in the population (right panel; Figure 3C). When we restricted the laser at a certain location, it induced a gradient of recruited SspB near the illuminated membrane. Once this was turned off, SspB again translocated back to the cytosol (Figures 3D,G). Linescan analysis clearly underlines SspB recruitment and shows that 50% of total cytosolic SspB was recruited to the membrane ( Figure 3E). Moreover, upon testing SspB recruitability with varying blue laser strengths (0.2%-5% of total laser intensity or 0.024-0.639 W/cm 2 laser power density), we identified a laser power density of~0.061 W/cm 2 (0.5%) to be ideal (Supplementary Figure S5C); a weaker laser strength was unable to efficiently recruit SspB (Supplementary Figures S5A, B) whereas higher intensities compelled Dictyostelium to move away from the light source (Supplementary Figures S5D,E). These results suggested easy recruitability and high membrane-cytosol turnover of the iLID system. Overall, both CRY2-CIBN and iLID-SspB systems provided us with tightly-regulated, spatiotemporal control necessary to investigate GEF function in different cell types.
Optical recruitment of different GEF proteins has strong effect on cytoskeletal dynamics, cell shape and motility We first examined the effects of an opto-RasGEF in neutrophil morphology and migration . Blue light-driven global recruitment of full-length RasGRP4 resulted in increased spreading and F-actin patches in these cells ( Figure 4A). A representative membrane kymograph showed increase in cell area with a concomitant increase in LifeAct patches after RasGRP4 recruitment ( Figure 4B). Across the population, RasGRP4 recruitment induced a~90% or~20% increase in neutrophil cell area or aspect ratio (which serves as a proxy for cell polarity), respectively ( Figures 4C,D). Next, selectively recruiting RasGRP4 to the back of the cell, as shown by dashed white box, caused new protrusions to arise locally with a concomitant disappearance of mature protrusions from the other end ( Figure 4E). Angular histogram analysis showed that probability of fresh protrusion formation is highest at or near the RasGRP4 recruitment site ( Figure 4F). Thus, opto-RasGRP4 increased cell area, F-actin polymerization, and reversed preexiting polarity. Since RasGEFs activate small Ras GTPase by exchanging GDP to GTP, we hypothesized that cytoskeletal effects induced by opto-RasGRP4 should be similar to an optically-recruited constitutively active Ras, KRas4B G12V. Upon locally recruiting CAAX-deleted KRas4B G12V to the Dictyostelium membrane, we noticed new protrusions appearing at the recruitment site (denoted by pink arrow in Supplementary Figure S6).
We next investigated how optically recruiting the catalytic Dbl homology (DH) domain of LARG would affect cell shape or size (O'Neill et al., 2018;Wagner and Glotzer, 2016;Rich et al., 2020;Inaba et al., 2021;Valon et al., 2017;Ridley, 2015). During imaging,  (Figures 5A,D). This reduced the number of distinct F-actin patches at the cell cortex as denoted by pink arrows in Figure 5B. Overall, there was a >50% reduction in cortical F-actin intensity upon LARG (DH) recruitment ( Figure 5E). This caused increased blebbing (shown with pink arrows in Figure 5C) and cell shrinkage ( Figure 5F). We next focused on opto-RacGEF mediated Rac GTPase activation and its role in regulating actin cytoskeleton and cell movement (Park et al., 2004;Pankov et al., 2005;Wu et al., 2009;Wang et al., 2010;Yoo et al., 2010;Kato et al., 2014). Upon global recruitment of the GEF domain of RacGEF1, the Dictyostelium cell triggered more protrusions around its periphery, as shown by pink arrows, and demonstrated appreciable spreading ( Figure 6A). Across the population, RacGEF1 recruitment resulted in~20% increase in cell area ( Figure 6B). There was a concomitant~10% or~40% reduction in cell polarity or average speed, respectively ( Figures 6C,D). Next, we looked at the local effects of RacGEF1 recruitment on cell migration. Since RacGEFs trigger actin polymerization usually at the front of migrating cells, we anticipated that locally recruiting RacGEF1 catalytic domain would either enforce or oppose the pre-existing direction of migration. Indeed, when we locally recruited RacGEF1 to the back of the migrating cell, it first induced a protrusion at the recruitment site (denoted by pink arrows near the dashed white box), then changed the direction of migration by generating bigger membrane ruffles, and finally made the cell move towards the light source ( Figure 6E). Angular histogram analysis confirmed that the probability of new protrusion generation was greatest at or near the site of GEF recruitment ( Figure 6F).

Discussion
In the last decade, we have witnessed a meteoric increase in size of the optogenetic toolkit for inducing specific and acute perturbations in signal transduction and cytoskeletal networks. This technology has enabled scientists to inactivate or activate particular biochemical activities and interactions with increasingly superior spatiotemporal resolution in live-cell experiments. Wider applications of these tools in complex tissue and animal systems will require technological improvements, specifically towards fine tuning key parameters of binding affinity and kinetics, and developing new proteins with redshifted excitation spectra which offer compatibility with two-photon imaging. Moreover, current optical systems depend on overexpression of individual proteins for their functioning, which makes it hard to control signaling activity at endogenous expression levels. Despite these shortcomings, optogenetics has become the most useful tool in basic and biomedical research, as highlighted by Nature Methods as 'Method of the Year' in 2010 (Ross et al., 2016;Repina et al., 2017;Kuhn et al., 2021).
Here, we have successfully utilized these optical tools and demonstrated that multiple light-induced dimerization systems can be transferred to various cell types, ranging from immune and epithelial cells to soil amoeba. We have specifically validated two blue light-triggered optogenetic systems, CRY2-CIBN and iLID, in these cells. Of these two, the cryptochrome-based system was compatible in all cell lines we tested; however, we were unable to establish a functional iLID system in developed Dictyostelium and HL-60 neutrophil-like cells. This was primarily due to poor expression and mislocalization of iLID protein components in these cells. However, instantaneous uncoupling of iLID dimers upon light removal allowed a finer temporal control than cryptochrome. Coupled with a tendency for low cytosolic clustering in the light state, iLID was superior to the cryptochrome system which displayed high light-dependent CRY2-CRY2 homo-oligomerization. To improve upon these shortcomings, we tested various available CRY2 mutants of which CRY2(W349R) fused to mCherry and CRY2 (R489E, A491D) fused to tandem dimeric Tomato (tdTomato) were the most promising (Taslimi et al., 2016;Duan et al., 2017). Both mutants showed significantly faster dark-state membrane-cytosol cycling rates, comparable to the iLID system. On account of low lightdependent cytosolic CRY2 clustering, the CRY2 (R489E, A491D)-tdTomato mutant was more recruitable on the plasma membrane than all other CRY2 fusion proteins. It would be interesting to check if a combination of these two mutants, i.e., a CRY2(W349R, R489E, A491D)-tdTomato, would display a faster dark-state cycling rate with a higher membrane recruitability.
We tested the physiological potential of our optogenetic systems by acutely perturbing membrane activity of GEFs specific for Ras, Rho, and Rac GTPases, which are important regulatory components of growth, metabolic, and migration signaling (Pylayeva-Gupta et al., 2011;Goicoechea et al., 2014;Devreotes et al., 2017;Lawson and Ridley, 2018;Pal et al., 2019;Li et al., 2020). We observed that opto-GEFs could globally or locally stimulate GTPase signaling, resulting in cytoskeletal reorganization and profound changes in cell shape and size. Global recruitment of opto-RasGEF and -RacGEF resulted in cell spreading whereas their local recruitment to the cell back caused neutrophil and Dictyostelium cells to form new protrusions at the recruitment site. With opto-RacGEF, we were even able to induce directional migration in the amoeba. On the other hand, opto-RhoGEF caused epithelial cells to shrink in size and form blebs around the periphery (Yoo et al., 2010;O'Neill et al., 2018;Wu et al., 2009;Wu et al., 2011;Pal et al., 2023;Kato et al., 2014;Wang et al., 2010;Ridley, 2015;Pankov et al., 2005;Coleman and Olson, 2002). All of these alterations could be directly induced, by recruiting either full length GEF proteins or only their GEF domains, within a matter of minutes without allowing signaling or cytoskeletal networks to readjust. Thus, optical control of signaling is an effective approach for spatiotemporal control of cellular signaling (Schwechter et al., 2013;Toettcher et al., 2013;Guntas et al., 2015;Valon et al., 2017;Yang et al., 2018;Lamas et al., 2020;Inaba et al., 2021;De Belly et al., 2023). Opto-SOS, a recruitable RasGEF, induced ERK signaling in Drosophila and uncovered the role of this pathway in promoting endodermal differentiation (Toettcher et al., 2013;Krueger et al., 2019;McFann et al., 2021). An opto-Rho1 (DH domain of LARG) induced ectopic deformations in the ventral and dorsal epithelia of Drosophila embryos (Rich et al., 2020). In light of these elegant studies, our opto-GEFs will be a beneficial addition to the optogenetic toolbox to probe mechanisms determining cell fate during embryogenesis.
Frontiers in Cell and Developmental Biology frontiersin.org