Combining FRET and optical tweezers to study RhoGTPases spatio-temporal dynamics upon local stimulation

Local stimulation with optical tweezers has been used to mimic natural stimuli that occur in biological processes such as cell migration or differentiation. Carriers (beads and lipid vesicles) with sizes down to 30 nm can be manipulated with a high spatial and temporal resolution: they are positioned with a sub-micrometric precision on a specific cell compartment and the beginning of the stimulation can be triggered with millisecond precision. RhoGTPases are a Ras-related family of proteins that regulate many different functions including cell polarity, microtubule dynamics and membrane transport pathways. Here we combine local stimulation with FRET microscopy to study RhoGTPases spatial and temporal activation following guidance cue local stimulation. We used two different vectors for local delivery: silica micro-beads and micro-sized lipid vesicles. The experimental methods associated with neuronal growth cone local stimulation are discussed in detail, as well as the analysis methods. Here we present a protocol that enables to study neuronal growth cone cytoskeleton rearrangements in response to a gradient of molecules in a way that better mimics physiological conditions, and it can be similarly applied to each secreted molecule involved in cell signaling.


BACKGROUND
RhoGTPases are a Ras-related family of proteins that play a major role in many physiological processes including cell polarity, microtubule dynamics, membrane transport pathways and transcription factor activity. They act as molecular switches that cycle between two conformational states: one bound to GTP ("active" state); the other one bound to GDP ("inactive" state) [1]. In the activated form, they bind different effector proteins leading to the activation of myriad downstream signals [2]. The best-characterized function of Rho GTPases is in the regulation of F-actin dynamics in both non-neuronal cells and neuronal growth cones [3,4]. During axon guidance, RhoGTPases translate signals from extracellular cues into growth-cone movement and axon guidance: RhoA increases the acto-myosin contractility, whereas Rac1 and Cdc42 induce the polymerization of actin to form lamellipodial and filopodial protrusions, respectively [5,6]. RhoGTPases' ability to regulate so many different functions, in highly dynamic cellular contexts, implies that they are tightly regulated at the spatio-temporal level. This has led to the development of fluorescent probes for the direct monitoring of their activation state in living cells based on the principle of Förster resonance energy transfer (FRET) [7][8][9][10]. Genetically encoded biosensors using protein-based affinity reagents and readout modules have been designed as intra-molecular (donor and acceptor on a single chain) [8] or inter-molecular (dual chain) [9] probes. FRET is an excellent readout for biosensors because small changes in the distance or orientation between the two fluorophores can cause large changes in FRET efficiency, allowing sensitive detection of protein binding or conformational changes. Measurement of FRET efficiency via ratiometric analysis, relying on the differences between the localization and activation images, is effective using intra-molecular probes. On the other hand, the use of intermolecular FRET design enhances sensitivity because, unlike the single chain design, there is no FRET when the biosensor is in the How to cite this article: Iseppon F, Napolitano LMR, Torre V, Cojoc D. Combining FRET and optical tweezers to study RhoGTPases spatio-temporal dynamics upon local stimulation. J Biol Methods 2017;4(1):e65. DOI: 10.14440/jbm.2017.159

POL Scientific
Protocol "off" state. The main disadvantage of the inter-molecular probes is that the two chains do not distribute equally through the cell, necessitating careful bleed-through corrections.
Most of the in vitro neuron stimulation experiments employ bath or micropipette administration of molecules to change the cell environment [11]. Hence, new strategies to address specific compartments of neuronal cells with spatial and temporal gradients of molecules, which better reflect cellular physiological conditions, should be considered [12]. Focal stimulation of neurons using optical tweezers and single micro-beads functionalized with a secretory molecule, the neurotrophin brain-derived neurotrophic factor (BDNF), demonstrated to induce focal increase of calcium signaling in the stimulated dendrite, with a specific activation of the TrkB receptor pathway and influenced the development of growth cones [13]. An elegant demonstration of the role of molecular gradients for cell stimulation has been reported for neutrophils by using optically manipulated microsources releasing soluble molecules that act as chemoattractants or perturb the actin cytoskeleton [14]. Another type of optically manipulated carriers are the lipid micro-vesicles encapsulating active molecules that are released by photolysis of the vesicle membrane using a laser pulse [15]. An optimization of this technique allowed the study of the effect of a very small number of different guidance cues on hippocampal neurons [16]. In principle, the number of molecules can be scaled down to a single molecule for a micron size vesicle at 1 nM concentration. Stimulation can be triggered with very high temporal precision, providing steep gradients over length scales from sub-micrometer to tens of micrometers and at time-scales from milliseconds to seconds.
Here, we present an experimental protocol that combines FRET microscopy with optical tweezers manipulation to observe the spatial and temporal activation of RhoA and Cdc42 following localized cell stimulation with Sema3A, a negative secreted guidance cue that steers a growth cone by inducing partial or local collapse [17,18]. We used NG108-15 neuroblastoma cells and an intermolecular RhoA/Cdc42 FRET sensor [19] for FRET. Two different vectors were used for Sema3A local delivery to specific sites of the cell through optical tweezers manipulation: silica micro-beads and micron-sized lipid vesicles. Combining FRET with local stimulation provides a new tool for the study of the cytoskeleton rearrangements in response to guidance cues stimulation, underlying the dynamic spatial and temporal activation of RhoGTPases.

NOTES:
The coverslip preparation step is meant to clean, sterilize and coat the surface of glass coverslips for cell culture.

1.1.
Wash glass coverslips overnight in 0.5 N HCl in a dedicated glass jar or beaker. 1.2. Remove the coverslips from HCl and wash twice in deionized water, then put them in 100% ethanol for 1.5 h. 1.3. Separate the coverslips, put them in a glass petri dish and sterilize them in a stove for 3 h at 180°C.

1.4.
Place the coverslips in a 12-well plate and coat them with 2 µg/cm 2 of laminin for 2-3 h.

1.5.
Rinse them twice with 1 ml of pre-warmed cell culture medium. Leave the coverslips in 1 ml of prewarmed medium.

NOTES:
Grow and maintain NG108-15 cell line following standard protocols. All cell incubations are performed at 37°C, 5% CO 2 incubator. Equipment and reagents coming into contact with cells must be sterile.

NOTES:
This protocol has been optimized for the transfection of Cdc42 and RhoA inter-molecular FRET probes developed by Murakoshi et al. [19] using Metafectene Easy + reagent.  The inter-molecular FRET biosensors are composed of two different proteins which DNA must be added to a total of 1 µg (0.5 + 0.5) to the transfection solution. Apart from the experiment samples, at least two samples must be transfected with 1 µg of each single probe for later imaging as control and off-line bleedthrough and cross-talk corrections (Steps 5-6).

Beads functionalization
NOTES: This protocol can be used for any protein to couple and any bead dimension, provided that the beads are functionalized with carboxylic (-COOH) groups.

3.2.
Pipette 60 µl of the diluted beads in a polypropylene micro-centrifuge tube.

3.4.
Discard the supernatant and suspend the pellet in 100 µl of Coupling Buffer.

3.6.
Discard the supernatant and suspend the pellet in 50 µl of Coupling Buffer.

Prepare a 200 mg/ml 1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDAC) solution in Coupling
Buffer and add 10 µl of this solution to the beads suspension. 3.8. Mix gently and leave to rest at RT for 1 h.

3.9.
Add the protein of choice at saturation. For the right amount of protein calculate 20-500 ng of protein for each mg of particles. 3.10. Leave to rest for 1 h at RT with gentle mixing.

POL Scientific
Protocol 3.11. Centrifuge for 10 min at 1000 × g, then remove the supernatant and suspend the pellet in 400 µl of Wash/Storage Buffer. Repeat this washing procedure twice. 3.12. Store the particles at 4°C for up to two weeks. NOTES: (Step 3.11): Keep the supernatant for the calculation of the amount of protein linked to the bead with the BCA protein assay.

NOTES:
This protocol is optimized for the BioMag Plus Carboxyl Coupling Kit.

Lipid vesicles preparation
NOTES: This protocol is meant to produce mono-or multi-layered lipid vesicles with a diameter varying between 1-10 µm.  NOTES: These lipid vesicles ought to be used in experiments only in the day of production, since they tend to deteriorate quickly.

NOTES:
For these experiments an inverted optical microscope equipped with an InfraRed (IR) optical tweezers and an UltraViolet (UV) microsurgery laser for optical manipulation of the vectors loaded with SemA3 is necessary.
The setup includes also a fluorescence imaging path adapted for FRET imaging: Donor and raw (r)FRET images are recorded simultaneously on the two halves of the CMOS sensor (Fig.1). Two filtersets (1-Donor excitation filter + dichroic, 2-Acceptor excitation + dichroic) are mounted in the fluorescence filter wheel of the microscope, and an optical splitter with emission filters for Donor (D) and Acceptor (A) is mounted at the exit port of the microscope to record two channels.

5.2.
Carefully mount the coverslip with cells into the recording chamber, fill it with Ringer's solution and set the chamber in the temperature controlled stage (37°C) of the microscope.

5.3.
Focus the objective on the cells seeking for transfected cells. Choose an isolated cell with neurite-like protrusions and growth cones.

5.4.
Drop 5-8 µl of lipid vesicles or beads in the recording chamber and then turn the IR trapping laser on.

5.5.
To stimulate with lipid vesicles adopt the following procedure.

5.5.1.
Trap a lipid vesicle and gently move the stage and the objective to position the trapped vesicle in the proximity (10-20 µm) of the cell compartment to be stimulated. ( Fig. 2A-2C).

5.5.2.
Using D excitation acquire rFRET and D fluorescence images for 90 s, to assess its spontaneous activity and dynamics.

5.5.3.
Acquire a video of the lipid vesicle photolysis using UV laser nanosecond pulses. Use Differential Interference Contrast (DIC) for imaging.

5.5.4.
Switch off the UV, IR laser beams, and the DIC imaging. Start acquisition of rFRET and D fluorescence images.

5.6.
To stimulate with coated beads adopt the following procedure.

5.6.1.
Record rFRET and D fluorescence images of the cell for 90 s to assess its spontaneous activity and dynamics.

5.6.2.
Trap a bead and move the stage and the objective to position the trapped bead in contact with the cell membrane at the desired site of stimulation. (Fig. 2D-2F).

5.7.
Generate the control images required to correct the raw data in the image processing step (step 6).

5.7.1.
Prepare and mount a fresh coverslip without cells, just with identical media and mounting conditions as that used for cells. Acquire minimum 20 images for both channels, using D excitation filterset. The resulting image set for each will be averaged later in the image processing for shading (uneven illumination) correction. Switch the fluorescence lamp off and acquire another set of minimum 20 images for dark current noise correction.

5.7.2.
Mount into the recording chamber a control coverslip with cells transfected only with the D sensor molecule. Using D excitation, acquire several (3-5) images of transfected cells in both D and A channels. These images will be used later to calculate the bleedthrough correction coefficient.

5.7.3.
Mount into the recording chamber a control coverslip with cells transfected only with the A sensor molecule. Acquire several (3)(4)(5) images of transfected cells in A channel using D and A excitation, respectively. These images will be used to calculate the cross-talk correction coefficient.
NOTES: (Step 5.7.1): Averaging over a large number of images is important to reduce the stochastic camera noise present in each image frame.

CAUTION:
This protocol involves the use of class 3b lasers and will require proper training and safety guidelines to be followed. In this case, a repulsive cue mediates collapse and retraction of the stimulated region (red line depicts putative morphological change after stimulation).

6.
Image processing and ratio calculation.

NOTES:
The goal of this step is to calculate the ratio image indicating the activity of the GTPases, starting from the rFRET and D images. This requires that any potential differences between these images be removed prior to calculating the ratio image. Therefore, several corrections, which are specific to the imaging system used to collect the raw data, are required. These corrections are termed: 1) dark current and shading corrections to ensure that the measured spatial variations in image intensities are accurate within each image and comparable across the different image channels, 2) background subtraction and photobleaching corrections to ensure that the measured intensities are comparable over time and across experiments at the whole image level and finally, 3) spectral overlap and imperfect spectral filters cause bleedthrough and cross-talk between D and A channels, which should be corrected to produce fully independent raw FRET image (rFRET) and D images.
If not otherwise mentioned, the image processing in this protocol was performed using the Biosensors 2.1. Matlab package (Danuser laboratory: http://lccb.hms. harvard.edu/software.html) of which flowchart is described in reference [20].

POL Scientific
Protocol 6.1. Split and align the D and rFRET images recorded on the two halves of the CMOS sensor during the experiment (step 5). We used a custom Matlab code here. 6.2. Correct the D and rFRET images for the dark current noise and shading, using the corresponding control images collected at step 5.7.1. 6.3. Image segmentation for D and rFRET to find the cell edge, create a background mask and subtract background.

NOTES:
In the case of inter-molecular (dual chain) FRET biosensor such as the biosensors we used in this protocol, the Ratio calculation is defined by the following equation [7]: where, R is the Ratio, rFRET is the raw FRET image, DD is the Donor image (intensity in D channel upon D excitation), AA is the Acceptor image (intensity in A upon A excitation), B and C the correction coefficients. Notice that three images are usually required in the above equation: rFRET, DD and AA. However, if the cross-talk correction coefficient C, is negligibly small, the weighted contribution of the Acceptor AA, can be neglected and the equation becomes: In this case the acquisition of the third image can be avoided, having at least two experimental advantages: reduce the cell exposure to light and avoid mismatching between the third image AA, which should be acquired at a different moment of time, with the rFRET and DD images, which are acquired simultaneously. In order to reduce the cross-talk, we chose the D excitation filter (445/30) centered on the minimum of the Acceptor A excitation spectrum. The measured correction coefficient C = 0.018 (mean value), which we considered negligibly small. Equation (2) is equivalent with the ratiometric formula for intra-molecular biosensors, since the same constant B, is subtracted always for each image pair, rFRET-DD. Choosing properly the Acceptor emission filter one can reduce also the bleedthrough of the light emitted by Donor through the Acceptor channel. The measured coefficient in our setup is, B = 0.055 (mean value).

6.5.
Correct photobleaching and calculate the Ratio images R using the corrected images with Biosensors 2.1. The final result is a set of images in tiff format, indicating the spatio-temporal activation of the RhoGTPase.

ANTICIPATED RESULTS
Local stimulation based on optically manipulated carriers containing active molecules is a flexible technique easy to integrate with the optical microscope for live cell analysis. Different imaging techniques can thus be used in parallel with optical manipulation, as phase contrast, fluorescence, or FRET microscopy. Micro-carriers are generally employed since it is easier to observe and monitor them by bright field imaging inside the sample chamber. Both carriers used in this work (beads and lipid vesicles) can deliver any type of active proteins, which makes them practical for a very wide range of applications. By using the method described herein, we were able to stimulate neuronal growth cones by mimicking the natural stimuli that lead to neuronal cell migration. We observed morphological rearrangements within the growth cones upon local Sema3A release. Figure 3 shows an example of the morphological changes occurred in NG108-15 cells upon Sema3A local stimulation with beads ( Fig.  3C and 3D) and lipid vesicles ( Fig. 3E and 3F) in comparison with a cell that is not stimulated (Fig. 3A and 3B). The plot of the mean edge growth of stimulated cells shows that both local stimulation methods are effective ways to trigger growth cone collapse and retraction, since there is no significant difference in the morphological response to either POL Scientific Protocol stimulation methods (Fig. 3G).
The combination of FRET imaging with local stimulation allows the monitoring of the spatial and temporal activation of signaling mole-cules (e.g. RhoGTPases) in response to local extracellular stimuli (e.g. guidance cues). The dynamics of the ratio between Donor and FRET channels reflect the spatio-temporal activation of the proteins studied.    vs. edge retraction that shows the increase in RhoA activity as the growth cone retracts. Figure 4 shows that Sema3A repulsive stimulus triggers an activation of RhoA in both the center of the growth cone that diffuses to the leading edge and in filopodia causing growth cone collapse and retraction. The dynamic changes can be further investigated by using image montages (Fig. 5A-5C) or by performing Area Intensity Analysis to plot the activity changes versus the leading edge dynamics (Fig. 5D).
In summary, this methodology can be used to study the dynamic growth cone rearrangement in response to different guidance cues, and has been recently defined as a very promising method for guidance cues delivery [21,22]. Since a great variety of FRET probes have been created to study the dynamics of many proteins [23][24][25][26], this combination provides a very useful tool for investigating intra-cellular signaling

POL Scientific
Protocol pathways at the basis of major cell dynamics.
However, the development of new vectors -for which the molecules release can be timely tuned -and of more sensitive FRET probes will represent important optimizations of this technique in the future to broaden even more its potential applications. Table 1 presents issues and problems that are most likely to be encountered during this procedure and some hints and solutions to solve them. The setup configuration, alongside IR and UV relative alignment are the most delicate and should be always checked and addressed carefully, since a misalignment between the two beams brings to impossibility in trapping the beads and breaking the lipid vesicles.