Controlling the Subcellular Localization of Signaling Proteins Using Chemically Induced Dimerization and Optogenetics

Abstract

A given protein can perform numerous roles in a cell with its participation in protein complexes and distinct localization within the cell playing a critical role in its diverse functions. Thus, the ability to artificially dimerize proteins and recruit proteins to specific locations in a cell has become a powerful tool for the investigation of protein function and the understanding of cell biology. Here, we discuss two systems that have been used to activate signal transduction pathways, a chemically inducible dimerization (CID) and a light-inducible (LI) system to control signaling and cytoskeletal regulation in a spatial and temporal manner.

1. Introduction

Having the capacity to localize or activate molecules in a spatially and temporally defined manner has revolutionized our ability to unmask the function of many components in cell biology. One method, the chemically inducible dimerization (CID) system, has been used in a variety of systems and has been successful in elucidating and studying various proteins and signaling pathways (reviewed in [1–4]). The dimerization field has been steered by a system that takes advantage of the mechanism of TORC1 inhibition by rapamycin, an immunosuppressive drug, with its use first demonstrated in the control of gene expression [5]. Rapamycin controls the dimerization of two components; it binds first to FKBP, and then the complex can bind to FRB, which is the binding domain from the TOR protein (Fig. 1). One protein (either FKBP or FRB) can be anchored at a specific cellular localization, while the other protein component is fused to a protein of interest (POI) and is free in the cytoplasm. Both components can be color coded by genetically fused proteins, and the readout is typically the translocation of the cytoplasmic molecule to the anchored protein. The addition of rapamycin, which is membrane permeable, forms this irreversible ternary complex in a rapid manner, taking place on the order of seconds to minutes.

Fig. 1.

Two methods for triggering signal transduction using synthetic biology. (Top) The FKBP/FRB system lowers plasma membrane PI(4,5)P2 levels. Diagram of FKBP/FRB system. (Top, Left) FKBP and FRB form a heterodimer in the presence of rapamycin (rapa). (Top, Right) Illustration of an experimental setup for recruiting a protein to the membrane with the FKBP/FRB system in Dictyostelium. FKBP domains are anchored to membrane by myristoylation (myr), and the actuator attached to an FRB domain is diffusing in the cytosol. When rapamycin is added, the actuator is recruited to membrane by binding to FKBP. The actuator is fused to the lipid phosphatase Inp54p which lowers PI(4,5)P2 locally and results in PI 3 kinase activity and the synthesis of PI(3,4,5)P3. (Bottom) Optogenetic system for controlling protein localization in Dictyostelium. (Bottom, Left) Diagram of blue light dimerization systems. In the dark, a photosensitive protein (iLID) has low affinity for its ligand (SSBP). When exposed to 488 nm light (optimally, it would be 440 nm light), a conformational change in the photosensitive domain allows it to bind to the ligand with affinity. (Bottom, right) Diagram of iLID-based systems for recruiting proteins to the membrane in Dictyostelium. The actuator (blue), fused to PI5 Kinase, is recruited to the uniformly distributed membrane anchor fused to cAR1 (red) only at the area exposed to blue light. Levels of PI(4,5)P2 should be elevated

The second technique (Fig. 1) discussed here uses optogenetics, which has several advantages [6]. Unlike CID, optogenetics can be reversible, and because specific wavelengths of light are used to activate the dimerization, it is rapid (msec to second timescales) and can be localized to specific subdomains of a cell. With optogenetics, light-induced conformational changes can expose binding domains or separate protein domains so that one can perform dimerization experiments similar to those seen with CID [7, 8]. In Dictyostelium, protein-protein interactions have been studied using the Crytochrome CRY2-mediated approaches [7, 9], but more recently, studies have also taken advantage of improvement in the light-oxygen voltage (LOV) domains used for optogenetic control, which will be discussed here [8, 10]. The newer system relies on an improved light-induced dimer (iLID) which has an added benefit of a quicker turnover of the light-induced state (half-life of 30 seconds vs. about 6 minutes for Cry2). This means that a response can be recorded over and over again on the same sample more readily. In addition, for systems where neither component is anchored, the cellular response is more limited to the area that was targeted, whereas in the Cry2 system, the activated molecules could diffuse away and provide a response in areas of the cell outside of your targeted zone. The main disadvantage for the optogenetics systems is that for targeted activation, one needs to use a scanning confocal system or a digital micromirror device (DMD), which adds a significant cost to the technique. Experiments using optogenetics investigating the small Rho GTPases have already made important contributions to our understanding of cellular mechanics during cell migration [11–13]. This approach is now being rapidly incorporated for use in the model system Dictyostelium discoideum. Here we demonstrate that CID can be used to get robust signaling responses critical for control of cell morphology and also show that iLID can be used to spatially control the recruitment of an enzyme that will be used for future cell migration studies. The ability to use these techniques in D. discoideum will permit use of this system to continue to play a critical role in furthering our understanding of signal transduction, cell polarity, and cell migration.

2. Materials

2.1. Cell Lines Used

1. Wild-type Dictyostelium discoideum cells of the AX2 strain were obtained from the R. Kay laboratory (MRC Laboratory of Molecular Biology, UK). Cells were cultured at 22 °C in HL5 medium supplemented with 300 μg/mL streptomycin. Cells were passaged on cell culture dishes every 3–4 days, and experiments were performed within 7–8 weeks of thawing cells from frozen stocks. Stable cell lines were grown similarly with selection antibiotics.

2. Heat-killed Klebsiella aerogenes (KA), from Dictybase (http://dictybase.org).

2.2. Media and Solutions

1. HL5 axenic medium for CID experiments: Formedium liquid media; 22 g of Formedium HL-5 powder and 10 g of dextrose were added to 1 L of distilled water and then autoclaved. HL5 media for iLID experiments: Recipe was obtained from DictyBase and includes 5 g proteose peptone, 5 g thiotone E peptone, 10 g glucose, 5 g yeast extract, 0.35 g Na2HPO4*7H2O, 0.35 g KH2PO4, and 0.05 g dihydrostreptomycin-sulfate (GoldBio, #S-150–100) were added to distilled water, and the volume was brought to 1 liter. pH was adjusted with HCl to pH 6.4–6.7. Medium was autoclaved for 20 mins to sterilize.

2. Development buffer (DB): Recipe was obtained from DictyBase. 5 mM Na2HPO4, 5 mM KH2PO4, 1 mM CaCl2, and 2 mM MgCl2 were added to distilled water. pH was adjusted to 6.5and filter sterilized.

3. H-50 buffer: Recipe was obtained from DictyBase. Briefly, 20 mM HEPES, 1 mM MgSO4, 50 mMKCl, 5 mM NaHCO3, 10 mM NaCl, and 1 mM NaH2PO4 were added to distilled water. pH was adjusted to 7 and filter sterilized.

4. SM agar plate (peptone 10 g/L; yeast extract 1 g/L; glucose 10 g/L; KH₂PO₄ 1.9 g/L; K₂HPO₄ × 3 H₂O, 1.3 g/L; MgSO₄ anhydrous 0.49 g/L; 1.7% agar) for KA growth.

5. LB-medium (lysogeny broth) for bacterial cell growth and plasmid preparation.

2.3. Recombinant DNA Used

1. Both the sequences encoding the first 150 amino acids at the N-terminus of PKBR1 (denoted as “myr”) for the membrane-anchored FKBP constructs, as well as tandem FKBPs from the ARGENT^™ Regulated Heterodimerization Kit (Version 2.0), were joined with a linker sequence of AGTGCTGGTGGT and cloned into pCV5 and also pDM358 [14].

2. Sequences encoding mCherry, FRB, and 1–331 amino acids of Inp54p (or inactivate form Inp54pD281A) were linked and cloned into pCV5 as well as pB18. The linker sequence between mCherry and FRB is GGAGCAAGTGGA, and two repeats of AGTGCTGGTGGT were used between FRB and Inp54p. PH_crac-YFP was also cloned in pCV5 [14].

3. pDM358 plasmid (DictyBase #534) expressing cAR1-CIBN or CAAX-deleted Venus-iLID fused to N150 gene fragment (N150-Venus-iLID; Addgene #201763) [15].

4. pCV5 plasmid (DictyBase #23) expressing CRY2PHR-mCherry (Addgene #26866) or tgRFPt-SspB R73Q fused to the N-terminal of the catalytic domain of RacGEF1 (gift from Firtel lab, UCSD).

5. PI5K Truncation 367–719aa ORF was amplified from PI5K 367–719aa-GFP/pCV5 and then subcloned into mRFPmars-SspB-R73Q/pCV5 at the NheI and NotI site to generate mRFPmars-SspB-PI5K 367–719aa/pCV5.

2.4. Cell Transformation

1. Gene Pulser Electroporation Cuvettes, 0.1 cm gap (BIO-RAD #1652089).

2. Gene Pulser Xcell Electroporation System (BIO-RAD #1652660).

2.5. Microscopy

1. CID experiments were performed on an Intelligent Imaging and Innovations (3I) Marianas system outfitted with a Yoko-gawa spinning disk confocal system. The microscope base was an inverted Zeiss-Axiovert 200 M epifluorescence microscope (Zeiss Inc., Oberkochen, Germany) with a Photometrix Evolve EMCCD camera. The lens used was a 40X PlanNeofluar 1.3NA oil immersion objective.

2. iLID experiments were performed on a Zeiss LSM800 GaAsP single-point, laser scanning confocal microscope with a wide-field camera using ZEN Blue software. All images were acquired with 63X/1.40 PlanApo oil DIC objective, along with digital zoom.

3. 8-well chamber (Lab-Tek, #155409 PK).

3. Methods

All procedures were carried out at 22 °C unless otherwise noted.

3.1. Stable Cell Transformations

1. Dictyostelium stable cell lines were generated by electroporation (see Note 1). For CID, cells were first transformed with mCherry-FRB-Inp54p in pB18 and PH_crac-YFP in CV5 by electroporation and selected with G418 (20 μg/mL). Then the selected cells were transformed again with myr-FKBP in pDM358 and cultured with G418 (20 μg/mL) and hygromycin (50 μg/mL) (see Note 2). For iLID, cells were electroporated simultaneously with both the N150-Venus-iLID/pDM358 and mRFPmars-SspB-PI5K 367–719aa/pCV5 constructs. To perform the transformation, 1 × 10⁷ cells were harvested, washed twice, and resuspended in 100 μL ice-cold H-50 buffer. For each transformation, 2 μg of plasmid DNA was mixed with the cell suspension and then moved to an ice-cold 0.1 cm-gap cuvette (see Note 3). Electroporation was performed with two pulses at 0.85 kV/25 μF with a 5 s interval (see Note 4). Cuvettes were then incubated on ice for an additional 10 min. Next, cells were transferred to a 10 cm culture dish containing 12 mL HL5 culture medium supplemented with heat-killed bacteria. On the next day, 10–20 μg/mL G418 sulphate and when appropriate 30–40 μg/mL hygromycin B were added to cells and selected over 1–3 weeks (see Note 5).

2. Once cells have reached confluency, cells are maintained in the 10 cm culture dishes with HL5 and streptomycin and are passed every few days as needed (see Note 6).

3.2. CID Microscopy

1. Vegetative cells were placed in an eight-well cover slip chamber and allowed to adhere for 10–15 min. Next, the media was replaced with 200 μL DB (see Note 7).

2. 10 μM rapamycin was prepared fresh from a 10 mM DMSO stock solution. This was diluted with nuclease-free water (NFW) to reach a final concentration of 100 μM. Then, 25 μL of 100 μM rapamycin were mixed with 25 μL of NFW in a separate tube. 50 μL of this 50 μM solution is added to a well containing 200 μL of HL5 with adhering vegetative cells (see Note 8).

3. Time-lapse microscopy was performed on an inverted Zeiss Axiovert 200 M using spinning disk confocal microscopy. Cells were imaged with a 40X lens and images were acquired every 5 s using 488 nm and 561 nm illumination to acquire GFP and RFP images depicted in Fig. 2. As can be seen, the Inp54P-RFP was recruited to the plasma membrane after addition of rapamycin. Interestingly, there was a simultaneous burst of PI3 Kinase activity, demonstrated by the near simultaneous recruitment of PH_crac-GFP to the plasma membrane, a biosensor for PI(3,4)P2 and PI(3,4,5)P3. This occurs after the PI (4,5)P2 is depleted upon recruitment of the Inp54p, a 5′ lipid phosphatase. These studies provide further evidence that a PI (4,5)P2 threshold regulates signaling responses [14, 16].

Fig. 2.

Synthetic lowering of the PM PI(4,5)P2 levels triggers cell spreading and PI3 Kinase activity. (TOP) Time-lapse spinning disk confocal images showing the recruitment of FRB-Inp54p-RFP after 10 μM rapamycin addition. (Bottom) The biosensor PH_crac-GFP translocates to the plasma membrane after PI(4,5)P2 is depleted

3.3. iLID Microscopy

1. Vegetative cells were seeded in an eight-well cover slip chamber and allowed to adhere for 10–15 min (see Note 9). Next, the media was replaced with 250 μL DB. Cells were then imaged. At time zero, the 488 nm laser is not yet on, so the Venus channel is blank, whereas 561 nm excitation was used for mRFPMars and the mRFPmars-SspB-PI5K 367–719aa is still in the cytosol (Fig. 3). For uniform recruitment, a 0.4% 488 nm argon laser was turned on to acquire a full-field image. The membrane anchor was uniformly distributed on the plasma membrane, as previously described [17]. Once the 488 nm laser is turned on, the mRFPmars-SspB-PI5K 367–719aa is recruited uniformly to the plasma membrane. Images were acquired ~8 s.

2. For local recruitment, the interactive photobleaching module on Zeiss LSM800 was used to perform local recruitment experiments (Fig. 4). A 514 nm (Ar laser) was used to excite YFP and mVenus (not shown). A small region of interest was placed in front or back of migrating cells (box) using the region module in the Zeiss Black software and bleached with 488 nm laser (laser power of ~0.4%) for one iteration. Since Dictyostelium are light sensitive, the laser intensity is low, ~0.06–0.08 W/cm² at the objective. Time interval of photo-activation and image acquisition was ~8 s. Two examples of the mRFPMars are shown. In the first cell, the 488 box is placed near the right side of the cell, and a localized recruitment of the mRFPmars-SspB-PI5K 367–719aa is observed. In the second cell (bottom panel), a box is placed below the cell, and a similar recruitment of mRFPmars-SspB-PI5K 367–719aa takes place. Images were acquired ~8 s (see Note 10).

Fig. 3.

Uniform recruitment of PI5 Kinase to the plasma membrane in response to illumination. At time zero, the 488 nm laser is not yet on, so the Venus channel is blank, whereas 561 nm excitation was used for mRFPMars and the mRFPmars-SspB-PI5K 367–719aa is still in the cytosol. For uniform recruitment, a 0.4% 488 nm argon laser was turned on to acquire a full-field image, and the mRFPmars-SspB-PI5K 367–719aa is recruited uniformly to the plasma membrane

Fig. 4.

Two examples of localized recruitment of PI5 kinase in response to targeted illumination. For local recruitment, the interactive photobleaching module on Zeiss LSM800 was used to perform local recruitment experiments. A small region of interest (box) was placed in front or back of migrating cells and bleached with 488 nm laser (laser power of ~0.4%) for one iteration. In the first example (top), the 488 box is placed near the right side of the cell, and a localized recruitment of the mRFPmars-SspB-PI5K 367–719aa is observed. In the second cell (bottom), a box is placed below the cell, and a similar recruitment of mRFPmars-SspB-PI5K 367–719aa takes place

4. Notes

1. Cells should be frozen immediately once transformants reach confluency. Cells can often change their phenotype when expressing a protein of interest, so it is critical that they be frozen so that the experimenter can thaw new cells out. The selection process for extrachromosomal plasmids is typically faster due to high copy numbers present in D. discoideum cells. Transfectants can sometimes be there in just a few days. Check your 10 cm plates regularly under a phase contrast microscope when transfectants come up so that you can freeze them quickly [18]. To obtain consistent results, new cultures should be started every 3–4 weeks, and cultures should never be allowed to grow to densities higher than 4 × 10⁶ cells/mL. If overgrowth occurs, a new culture should be thawed out.

2. While many motifs (e.g., myristoylation membrane localization) have been used for membrane localization of signaling regulators, it is recommended that one use a receptor that should be even more uniformly distributed across the membrane, like the g protein coupled receptor cAR1, which is used here for the iLID experiments. We have found that including two FKBP domains on the C-terminus tail increases recruitment of the POI. The POI should have any localization domains removed so that targeting is controlled by the CID or iLID systems.

3. Going higher than 2 μg of DNA can be toxic to cells and decrease transfection efficiency. Keep the DNA volume below 5 μL.

4. When expressing an optogenetic pair, use pCV5 for your actuator and pDM358 for your anchor. Use a dual expression system for imaging your biosensors. Caution: The antibiotics used as selectable marker for your plasmids are toxic. Wear gloves.

5. When only expressing a CID pair in a cell, the actuator can be expressed off pCV5 or PB18 and the anchor pDM358. If including a biosensor, use pCV5 vectors for both CID vectors and pDM358 for the biosensor. Make sure to include a proper control. For enzymes, a catalytically inactive protein should be used as an actuator. The best controls are done under blinded conditions.

6. If one is having difficulty expressing two different fluorescently colored probes in the same line, the creation of giant cells through cell fusion can be used to overcome this problem. This just requires two individual cell lines each expressing a given fluorescent protein. To fuse cells, start with a total of 1.5 × 10⁸ axenically grown cells, washed two times and resuspended in 10 mL SB (17 mM Soerensen buffer, 15 mM KH2PO4 and 2 mM Na2HPO4, pH 6.0). Next, the cells were put inside a conical tube and gently rolled for 30–40 min to induce cluster formation. Subsequently, 800 μL of rolled cells were transferred to a 0.4 cm gap Bio-Rad cuvette, using pipette tips with cutoff tips (to ensure aggregated clusters remain intact). Using a BioRad Gene Pulser, the electroporation was performed at 1 kV, 3 μF once, and then with 1 kV, 1 μF twice more to induce cell fusion. Cells were allowed to recover, settle, and adhere to substrate for ~1 hour before starting the image acquisition. One could alter the proportion of cells to increase the concentration of one fluorescent cell line over the other. This technique also allows you to fuse mutant and wild-type backgrounds and is not limited to just two cell lines.

7. D. discoideum can be very sensitive to the illumination laser. Incubating the cells in DB for at least 30 minutes helps with viability prior to microscopic observation. Moreover, we have found that growing cells on an SM lawn seeded with K. aerogenes overnight can make cells extremely resistant to photodamage [19]. The majority of bacteria should be removed prior to imaging. K. aerogenes bacteria can be removed by a low speed spin (1200 g). The D. discoideum will pellet, while most of the bacteria will remain in the supernatant. Similarly, we found that allowing the D. discoideum to feed on a low density of K. aerogenes for an hour or two, once in DB, also helps with protection from photodamage. Phototoxicity is a big problem and can lead to artifacts, but keep in mind that growing cells in the presence of bacteria will change their phenotype and may influence your results.

8. We have found that gently adding 50 μL dropwise to an 8 well chamber containing 200 μL of DB results in quick mixing of the compound with the DB, with minimal loss of cells from dispersion. The final concentration of rapamycin used for CID experiments is 5–10 μM, depending on the CID pair. Allow sufficient time (minimal, 1–2 minutes) for membrane recruitment during the imaging, but keep in mind that rapamycin is a toxin, and experiments beyond an hour may begin to show phenotypes as a result of cell death.

9. Always record the temperature in the microscopy room. Many microscope rooms are notorious for overheating from running equipment. Temperature fluctuations can influence cell physiology.

10. For statistical analysis, Fiji ImageJ and Microsoft Excel with the XLSTAT add-on can be used to analyze collected data. Fiji ImageJ is typically used to analyze videos taken from all time lapses and isolate desired frames. Microsoft Excel with XLSTAT can be used in order to organize numerical data, run student t-tests, perform norming, derive standard deviation, and create graphs.