Spatial and temporal control of protein secretion with light

Abstract

How newly synthesized integral membrane proteins and secreted factors are sorted and trafficked to the appropriate location in different cell types remains an important problem in cell biology. One powerful approach for elucidating the trafficking route of a specific protein is to sequester it following synthesis in the endoplasmic reticulum and trigger its release with an externally applied cue. Combined with fluorescent probes, this approach can be used to directly visualize each trafficking step as cargo molecules progress through the different organelles of the secretory network. Here we discuss design strategies and practical implementation of an inducible protein secretion system we recently developed (zapalog mediated ER trap: zapERtrap) that allows one to use light to initiate secretory trafficking from targeted cells or subcellular domains. We provide detailed protocols for experiments using this approach to visualize protein trafficking from the endoplasmic reticulum to the plasma membrane in fibroblast cell lines and neurons.

1. Unit Introduction

Nearly all integral membrane and secreted proteins are processed through the cellular secretory network. In Eukaryotic cells, these proteins are synthesized into the endoplasmic reticulum (ER) where they are folded, assembled, and undergo various post-translational modifications including glycosylation and disulfide bond formation. Following maturation, proteins leave the ER at specialized exit sites and accumulate in the Golgi apparatus (GA) where they are further biochemically modified and ultimately sorted into mobile carrier vesicles that shuttle them to the appropriate subcellular domain. Different protein cargoes utilize distinct trafficking routes, with some routes poorly characterized, highlighting the importance of new tools and approaches for investigating secretory protein trafficking.

Classical experiments tracking proteins as they navigate the secretory network were performed by fractionation of newly synthesized proteins pulse labeled with radioactive amino acids and biochemical analysis of post-translational modifications performed by enzymes residing in different secretory organelles. However because this method requires cell lysis and solubilization, it yields no information about the spatial organization of secretory trafficking. The first efforts to directly visualize a protein on its journey through the secretory network in live cells utilized the vesicular stomatitis viral glycoprotein (VSV-G) ts 0 45 temperature sensitive mutant, which is retained in the ER at elevated temperatures (~40°C) but traffics normally when cells are cooled (<37°C) [1–3]. Thus, simply by decreasing the temperature, one could liberate VSV-G (tagged with a fluorescent protein) from the ER and simultaneously visualize its progression through each organelle of the secretory network. While this powerful approach provided the first glimpse into the cellular organization of protein secretion, it was not generalizable to different proteins and lacked fine temporal control. To address these issues, several different approaches emerged that use small molecules (rather than a temperature shift) to trigger forward trafficking of user-defined proteins in cell lines and diverse model organisms [4–9]. While these “trap and release” approaches (i.e. they rely on trapping a protein in the ER and releasing it by addition of a chemical) have proven powerful for studying secretory trafficking in diverse cell types, they lack the spatial regulation required for precision control at the level of individual cells or subcellular domains. The ability to manipulate the surface levels of specific receptors, or secreted proteins within user-defined cell populations would be powerful for studying cell signaling, especially in intact organisms. Furthermore, in large cells of the nervous system such as neurons or glia, whose secretory organelles are broadly dispersed, the ability to initiate secretory trafficking from specific subcellular domains (e.g. dendrites, cell body or axons) would enable one to address previously intractable questions concerning how (or whether) proteins are spatially compartmentalized in complex cellular environments [10]. Thus, there remains an unmet need for tools that enable control of secretory trafficking in both space and time.

1.1. Why use light to control cellular function?

New approaches for precise activation, inhibition and observation of cellular processes are continuously deepening and refining our understanding of cell biology. The confluence of chemical genetics (the ability to control the function of an engineered protein with a small molecule that is otherwise biologically inert) and live cell fluorescence microscopy has yielded a powerful toolkit to directly manipulate and observe diverse cellular functions [11–13]. While traditional chemical genetic approaches using small molecule actuators are powerful, they generally lack spatial control for targeting specific groups of cells, individual cells or subcellular domains. This issue is being addressed through development of diverse photosensitive actuator proteins (phytochromes, cryptochromes, and LOV domains, among others) and small molecules that can be activated with a pulse of UV, visible or infrared light [6, 14, 23–28, 15–22]. In many ways, light is an ideal actuator for controlling cellular biochemistry since it can be delivered for precise durations at user-defined times over spatial scales ranging from entire organisms to sub-micrometer subcellular domains. Pioneering approaches used light-sensitive “caged” small molecule second messengers or neurotransmitters that could be photoconverted to a bioactive state [29, 30]. Coupled with real time readouts of cellular function, this “photo-uncaging” approach has had immense impact for studying compartmentalized signaling pathways. More recent efforts have culminated in photosensitive chemical-induced protein dimerization strategies to leverage the spatial control afforded by light [31–40]. One such approach is based on a compound called zapalog, which is composed of a rapamycin homologue (which binds FK506 binding protein; FKBP), linked to trimethoprim (which binds bacterial dihydrofolate reductase; DHFR) via a photocleavable chemical linker (Fig. 1A) [41]. Zapalog therefore heterodimerizes two target proteins, one tagged with DHFR and the other tagged with FKBP, and this interaction can be subsequently disrupted with a brief pulse of near UV light (405nm) (Fig. 1B). This approach was originally developed for toggling organelles on and off microtubule-based motor proteins for precise control of organelle localization [41]. Here we describe a detailed method for a new zapalog-based application (zapERtrap) for “uncaging” proteins from the ER, allowing precise spatial and temporal control of secretory trafficking from targeted cells or subcellular domains [42].

Figure 1.

Photo-inducible protein secretion with zapERtrap.

A. Zapalog chemical structure. Zapalog is comprised of the antibiotic trimethoprim (TMP), a synthetic ligand of FKBP (SLF) and a dialkoxynitrobenzyl (DANB) moiety susceptible to photolysis by 405 nm light. The purple dashed line marks the location of photocleavage.

B. Photo-release of DHFR/FKBP linked through zapalog. In the presence of zapalog, the proteins DHFR and FKBP become tightly associated through their respective binding to the TMP and SLF moieties of zapalog. Photoexcitation with 405 nm light cleaves the DANB linker and disrupts the association between DHFR and FKBP.

C. ZapERtrap control of protein secretion. Following synthesis in the ER membrane, DHFR-fused secretory cargo proteins are retained via zapalog-mediated, association with FKBP-KDEL present in the ER lumen. Photolysis of zapalog with 405 nm light rapidly disrupts the link between the secretory cargo and the FKBP-KDEL ER anchor, allowing forward trafficking to proceed. The DHFR fusion construct is engineered with a fluorescent protein (FP) tag for visualization of subsequent forward trafficking steps. DHFR, dihydrofolate reductase; FKBP, FK506-binding protein.

1.2. General considerations for zapERtrap regulated protein secretion:

1.2a. Construct design:

We have adapted zapalog to control protein secretion using FKBP fused to a fluorescent protein containing an ER retention motif (the C-terminal amino acid sequence: KDEL) and a target protein fused to DHFR (Fig. 1C). It is critical to understand the membrane topology of the target protein such that the engineered DHFR tag is presented in the lumen of the ER to make it available for zapalog-mediated binding to the FKBP-KDEL “anchor” which binds to and retains the target protein in the ER (Fig. 2). Brief exposure to 405nm light, applied either globally (e.g. using full field illumination with a 405nm laser or CFP filter cube coupled to a full spectrum light source), or locally (using a galvanometer-steered 405nm laser or digital micromirror device, details below) photolyzes zapalog, liberating the DHFR-tagged target protein from the FKBP-KDEL retention module, allowing forward trafficking to proceed. This general strategy has worked well for both type I membrane proteins where the DHFR is fused to the N-terminus (e.g. neuroligin 1 and the glutamate receptor subunit, GluA1) and for type II membrane proteins where DHFR is fused to the C-terminus (e.g. transferrin receptor). For N-terminally tagged proteins it is important to preserve the signal peptide that generally directs nascent peptides to the ER (Fig. 2B). In these cases, we have placed the DHFR tag immediately following the signal peptide with good success. While the DHFR tag appears to work well for FKBP-mediated ER retention when it is placed at the C- or N-terminus, we have not yet attempted to integrate DHFR into the middle of a target protein, but this may be an important consideration in cases where N- or C-terminal fusion is known to disrupt protein function and/or localization. Plasmids we have designed that express the FKBP-KDEL ER anchor fused to different fluorescent proteins should be generally effective for retaining diverse targets and will be available at the Addgene plasmid repository (along with several DHFR-tagged secretory proteins we have generated).

Figure 2.

ZapERtrap construct design.

A. Anchor construct design. The essential components of the FKBP anchor construct include an N-terminal signal peptide sequence from mouse immunoglobulin kappa light chain (Igκ signal peptide) for proper ER targeting, an epitope tag to confirm expression/ER localization, and a C-terminal KDEL sequence to retain it within the ER lumen (top). Constructs with tandem copies of FKBP (middle) enhance ER retention of DHFR-fused cargoes with no apparent effect on the kinetics of cargo release; however, this may not be the case for all cargoes and therefore must be empirically determined. Care should be taken to preserve the N- and C-terminal position of the Igκ signal peptide and KDEL sequence, respectively. Effective anchor constructs include those with an epitope tag (e.g. hemagglutinin) positioned in between the Igκ signal peptide and FKBP (top and middle) or with a fluorescent protein (FP) positioned in between FKBP and the KDEL sequence (bottom).

B. Cargo construct design. N- or C-terminal placement of DHFR depends on the orientation of the cargo of interest in the ER membrane. Transmembrane proteins of types I, III, and IV-B are oriented with the N-terminus on the luminal (extracellular/exoplasmic) face and the C-terminus on the cytosolic face (top right). Type II and IV-A transmembrane proteins orient with cytoplasmic N- and luminal (extracellular/exoplasmic) C-terminus (bottom right). For cargo retention, DHFR is fused to the terminus within the ER lumen. If the cargo of interest does not contain an internal ER targeting sequence (present in most type II–IV transmembrane proteins), a suitable N-terminal signal peptide sequence must be included for proper targeting to the secretory pathway. Additional considerations are the inclusion of an epitope/FP tag and/or a protease (e.g. thrombin) cleavage site to remove the epitope if necessary. The N- or C-terminal placement of an epitope/FP tag for surface detection depends on the orientation of the cargo in the plasma membrane, and is therefore the same terminus that DHFR is fused to. DHFR should be positioned downstream of the N-terminal signal peptide sequence (as is the case for type I/III/IV-B transmembrane proteins that do not have an internal ER targeting sequence; top left) or at the C-terminal end of type II/IV-A transmembrane proteins (bottom left). The protease cleavage site is positioned between the epitope/FP tag and the cargo sequence.

C. Example of an IRES-containing zapERtrap construct for co-expression of cargo and anchor. IRES-mediated (5’ cap-independent) translation results in a relatively lower expression of the second gene, therefore it is recommended to prioritize expression of the anchor by positioning it upstream of the IRES. Use of a biscistronic zapERtrap construct increases the likelihood for a given transfected cell to express both anchor and cargo and may result in more uniform expression ratios across cells than co-transfection of multiple plasmids. β-act, β-actin; CMV, cytomegalovirus; DHFR, dihydrofolate reductase; ER, endoplasmic reticulum; FKBP, FK506-binding protein; FP, fluorescent protein; IRES, internal ribosomal entry site; PM, plasma membrane; C-term., C-terminus; N-term., N-terminus; sig. pep., signal peptide.

1.2b. Light sources for triggering zapERtrap:

Zapalog is photosensitive in the near UV spectrum and we have found that 405 nm light delivered either from a laser source or light emitting diode (LED) is effective for photolyzing zapalog and triggering protein trafficking in live cells. A metal halide, mercury, or xenon light source commonly used for fluorescence microscopy should be equally suitable. Importantly, common red and green fluorescent proteins (e.g. EGFP, mNeon, mCherry, etc.) can be simultaneously imaged with negligible zapalog photolysis using common laser lines (e.g. 488 and 561 nm) or filter cubes (e.g. Cy2/EGFP, Cy3/TRITC, etc.). However, we wish to stress that ambient room light could photolyze zapalog, depending on the duration and intensity of exposure. We perform all handling of zapalog-treated samples (and the zapalog compound itself) under dim red light and filter the transmitted light on our microscope (we use a 45mm Schott RG610 long pass light filter) when we set up samples for imaging.

For local zapERtrap experiments (e.g. releasing proteins from the ER only in targeted cells or subcellular domains) we generally use galvanometer-steered mirrors (FRAPPA device, Andor Technologies) to scan a diffraction limited 405 nm excitation spot over the region of interest. Typical power settings are 6% laser power from a fiber-coupled 100 mW 405 nm laser (which yields 23.0 μW/μm² through a 60x Plan Apochromat NA 1.42 objective) with a scanning dwell time of 1 msec. We have also successfully used an LED source coupled to a digital micromirror device (Mosaic, Andor Technologies) for illumination of a region of interest. Finally, for global ER release (e.g from all cells within an imaging field), we use full field excitation with a 405 nm laser source directed through a spinning disk scan head with a 50 msec exposure at 80% laser power (100 mW fiber coupled laser), which yields 912 μW at the objective (60x, NA 1.42). One powerful feature of zapERtrap is the ability to titrate the level of protein released simply by adjusting the light intensity. Here we use an acousto-optic tunable filter (AOTF) to control the intensity of 405 nm laser excitation (focused to a diffraction limited spot). Typical powers for triggering ER release are 23.0 μW/μm² for 100% release and 6.83 μW/μm² triggering ~50% release (measured through a Plan Apochromat NA 1.42 objective; 1 msec dwell time). For titrating the amount of cargo release with full field illumination, we deliver a 50 msec pulse of 405nm excitation light through a spinning disk scan head. Typical excitation powers (measured at the objective are: 912 μW (80% laser power) of full-field excitation for ~100% release; 194 μW (30% laser power) triggers ~50% release and 97 μW (15% laser power) triggers ~25% of release. An intensity-controllable LED or constant power mercury/xenon/metal halide light source should be similarly effective with neutral density filters to achieve the appropriate power. While the light intensities above are a good starting point, a careful light titration is recommended to optimize exposure duration and intensity for effective ER release. At the light exposure times and intensities noted above we have not observed signs of cell toxicity.

2. Materials:

2.1. Cell transfection

1. Zapalog compound (MedChemExpress, cat. no. HY-126316; see Note 1)

2. Lipofectamine 2000

3. Aluminum foil

4. Tissue culture biosafety cabinet

5. Tissue culture incubator

6. Plasmids encoding FKBP “anchor” (see Note 2) and DHFR target cargo (see Note 3)

   For experiments with heterologous cells:

7. COS-7, HeLa or other suitable adherent cell line grown on #1 or #1.5 cover glass (see Notes 4 and 5)

8. Dulbecco’s Modified Eagle Medium (DMEM)

9. DMEM supplemented with 10% Fetal Bovine Serum (FBS), optional: 50 units/mL penicillin, 50 μg/mL streptomycin

   For experiments with primary cultured neurons:

10. Primary rat hippocampal or cortical cultured neurons plated on poly-d-lysine-coated #1 or #1.5 cover glass (see Note 5)

11. Neurobasal-A medium

12. Neurobasal-A medium supplemented with GlutaMAX and B27

2.2. Live-cell imaging

1. HEPES-based imaging solution: 130 mM NaCl, 5 mM KCl, 10 mM HEPES, 30 mM D-glucose, 2 mM CaCl₂, 1 mM MgCl₂ (pH to 7.4 using 1 M NaOH)

2. Cover glass chamber for live cell imaging (Ludin chamber type 1, Life Imaging Services; see Note 5)

3. Light source: Any of the following are suitable for activating zapalog: laser light source from confocal microscope, LED/DMD, xenon/mercury/LED lamp with near UV excitation filter (e.g. CFP)

4. Red safelight (LED bicycle tail lights or headlamps with wavelengths ~620 nm work well)

5. Alexa Fluor 647 conjugated polyclonal anti-GFP antibody if simultaneously imaging surface accumulation (Invitrogen, cat. no. A31852; RRID:AB_162553)

3. Methods:

3.1. Cell transfection (for one 18 mm coverslip in a 12 well plate)

3.1a. Transfection of heterologous cells (split/plate cells the day before so they are ~50% confluent at the time of transfection; see Note 4):

1. Add 50 μL of serum-free, antibiotic-free DMEM to a 1.5 mL Eppendorf tube containing 1.5 μL of Lipofectamine 2000 and incubate at room temperature for 5 min.

2. Add 0.5–1 μg of plasmid DNA (for each CMV-driven expression construct) and/or 100–300 ng of plasmid DNA (for each CAG-driven expression construct) to a 1.5 mL Eppendorf tube containing 50 μL serum-free DMEM (see Note 6). If necessary, adjust the volume of Lipofectamine 2000 to maintain a 1:0.5 to 1:5 DNA (μg):Lipofectamine 2000 (μL) ratio and scale according to manufacturer’s recommendations.

3. Add the diluted Lipofectamine 2000 to the diluted DNA and gently mix by pipetting. Incubate at room temperature for 20 min.

4. During the incubation, remove medium from the well such that only ~800 μL remains, and save this conditioned media. Return cells to incubator.

5. In a tissue culture biosafety cabinet illuminated with red LED safelight, make a 40 μM working stock of zapalog diluted in the saved conditioned media (see Note 1). Wrap in aluminum foil until ready to use.

6. Add DNA-Lipofectamine transfection mixture to cells dropwise.

7. In a tissue culture biosafety cabinet illuminated with red LED safelight, add 10–20 μL of a 40 μM working stock of zapalog (for a final concentration of 0.5–1 μM) directly to cell media and gently swirl plate to mix (see Note 7). Wrap plate in aluminum foil.

8. Incubate cells in a 5% CO₂ incubator at 37°C for 12–24 hours to allow expression of the constructs (see Note 8).

3.1b. Transfection of cultured rat hippocampal/cortical neurons (see Note 9):

1. For each well (using standard 12 well dish with cells grown on 18mm round coverglass) to be transfected, add 50 μL of serum-free Neurobasal-A media to a 1.5 mL Eppendorf tube containing 1.5 μL of Lipofectamine 2000 and incubate at room temperature for 5 min.

2. Add 0.5–1 μg of plasmid DNA (for each CMV- or Synapsin-driven expression construct) and/or 100–300 ng of plasmid DNA (for each CAG-driven expression construct) to a 1.5 mL Eppendorf tube containing 50 μL serum-free DMEM (see Note 6). If necessary, adjust the volume of Lipofectamine 2000 to maintain a 1:0.5 to 1:5 DNA (μg):Lipofectamine 2000 (μL) ratio and scale according to manufacturer’s recommendations.

3. Add the diluted Lipofectamine 2000 to the diluted DNA and gently mix by pipetting. Incubate at room temperature for 20 min.

4. During the incubation, remove medium from the well such that 0.8–1 mL remains, and save this conditioned media. Return cells to incubator.

5. In a tissue culture biosafety cabinet illuminated with red LED safelight, make a 40 μM working stock of zapalog diluted in conditioned media (see Note 1). Wrap in aluminum foil until ready to use.

6. Add an equal volume of fresh Neurobasal-A supplemented with B27 and GlutaMAX to the saved conditioned media to obtain a 1:1 ratio. Warm in 37°C water bath or cell culture incubator.

7. Add DNA-Lipofectamine transfection mixture to cells dropwise. Gently swirl plate to mix and return to incubator.

8. After 1.5 hours, remove medium containing DNA-Lipofectamine transfection mixture and replace with 800 μL warmed 1:1 conditioned:fresh media.

9. In a tissue culture biosafety cabinet illuminated with red LED safelight, add 10–20 μL of a 40 μM working stock of zapalog (for a final concentration of 0.5–1 μM) directly to cell media and gently swirl plate to mix. Wrap plate in aluminum foil.

10. Incubate cells in a 5% CO₂ incubator at 37°C for 12–24 hours (see Note 8).

3.2. Live cell imaging: global ER release

1. We use a full enclosure live cell incubator on our spinning disk confocal microscope that requires prewarming to 34–37°C for at least 1 hour prior to use to ensure focal stability during time lapse imaging.

2. Optional: If your cargo constructs have incorporated a thrombin protease site to remove extracellular epitopes from cargo that may have escaped zapalog-mediated ER trapping, add 1 unit/mL of thrombin directly to cell medium 10–30 min prior to imaging to eliminate background surface signal (see Note 10).

3. In a tissue culture biosafety cabinet illuminated with red LED safelight, remove coverslip from the 12 well plate and place in an appropriate live cell imaging chamber (see Note 6) and add 0.7 mL prewarmed HEPES imaging solution with or without 0.5–1 μM zapalog (see Note 11). Gently clean the bottom of the coverslip with a Kimwipe dampened with 70% ethanol. Transport imaging chamber in a light proof container (an empty tip box wrapped in aluminum foil works well).

4. Place imaging chamber on microscope, taking care to keep cells shielded from light during transfer. Use a red LED safelight for placement. Avoid the use of unfiltered transmitted light. Our full enclosure incubation chamber is made of black plexiglass to avoid ambient light exposure of our samples. We also suggest altering the display mode of computer screens in the imaging area with built-in or third-party blue light filtering programs (we use the f.lux app in “dark mode”).

   1. For experiments involving live, real time surface labeling of released cargoes as they appear at the cell surface, add 50 μL of diluted dye-conjugated antibody (typically 1:750 final dilution in the imaging chamber). Gently mix by pipetting. We recommend using fluorescent protein or epitope tag Alexa Fluor 647-conjugated antibodies. See Note 12.

5. Locate transfected cells with 561 nm excitation light, if possible. Brief exposure to 488 nm excitation light does not appear to appreciably photolyze zapalog; however, it is unknown whether long-term exposure of high intensity 488 nm excitation light causes unwanted photolysis.

6. Establish a 5–10 minute baseline by acquiring a series of images obtained with 488, 561 and 640 nm excitation light. We take confocal z-stacks with 0.4 μm spacing and capture 12 optical sections per time point. For purposes of measuring kinetics of cargo accumulation at the Golgi apparatus and/or plasma membrane, it is sufficient to acquire a z-stack image every 1–2 minutes. Take single-channel images at a frame rate of 2 to 5 Hz to capture fast dynamics of post-ER and/or post-Golgi vesicles as these structures move on the order of 0.5–1 μm/second. Imaging at this frequency should be limited to ~3–5 minutes as cells can become photodamaged; however this is highly dependent on the wavelength and intensity of excitation light (see Note 13).

7. After acquiring the baseline images, confirm the absence of background surface signal (640 nm channel) arising from recycling or surface receptors that may have escaped the ER. If appreciable signal is observed, a new cell in a different region of the coverslip should be located and step 6 repeated (see Note 12).

8. Initiate global ER release by full-field illumination with 405 nm light of appropriate intensity and exposure time (see section 1.2b above). We typically acquire 2 z-stacks with light from a 100 mW fiber-coupled 405 nm laser at 80% laser power (912 μW measured from the objective) delivered from an acousto-optic tunable filter (AOTF) controlled laser launch.

9. Continue imaging to visualize cargo as it progresses from the ER to the Golgi apparatus, to the plasma membrane. It will be obvious if efficient ER release occurred as cargo should begin to accumulate in the Golgi apparatus within 5 min of the 405nm light exposure, although kinetics of ER trafficking can widely vary for different cargoes. Subsequent surface signal (detected with dye conjugated anti-extracellular epitope antibody) should be apparent approximately 30 min following 405nm illumination, although some cargoes may take longer. See Fig. 3B for an example.

Figure 3.

Inducible and local control of secretory trafficking in COS-7 cells.

A. TfR-GFP-DHFR (green) colocalizes with ER marker mCh-Sec61 (red) in COS-7 cells expressing FKBP-KDEL in the presence of zapalog. The white arrowhead denotes the nuclear envelope, which is contiguous with the ER and contains both red and green signal. The bottom panels show enlarged images of mCh-Sec61 and TfR-GFP-DHFR at the cell periphery (denoted by dashed box in top panel) where ER morphology is distinct. A plot of pixel intensities along the red line (in image inset) is plotted on the right. Scale bar, 10 μm; magnified images are 6×6 μm.

B. Shown is a COS-7 cell expressing TfR-mCh-DHFR, the GA marker mEmerald-GaIT and FKBP-KDEL in the presence of zapalog before (left) and 15 min following 405 nm light exposure (right). The magenta arrowheads denote the GA. The insets show colocalization between mEmerald-GaIT and TfR-mCh-DHFR. Note the display settings were adjusted for the TfR-mCh-DHFR channel 15 min post image to avoid apparent saturation of the signal as it concentrated in the GA. Pixel intensities along the red line shown in the upper left panel are plotted for mEmerald-GaIT (green line), TfR-mCh-DHFR before ER release (red dashed line) and 15 min after ER release (solid red line). Scale bar, 10 μm; magnified images are 5×5 μm.

C. Live-cell antibody surface labelling reveals the spatiotemporal dynamics of cargo surface presentation following ER release. Top Left: Schematic of strategy for real-time visualization of surface accumulation using Alexa647-anti-GFP, which rapidly binds to TfR-GFP-DHFR as it appears on the cell surface. Right: Image time series of TfR-GFP-DHFR accumulation in the GA (top panels; arrowhead) and on the cell surface (middle panels) following light-triggered ER release at time 0. Scale bar, 15 μm. Bottom Left: Kinetics of TfR-GFP-DHFR accumulation in the GA (blue) and on the cell surface (red) following release; mean ± SEM (n=6 cells from 2 independent experiments).

D. Local 405 nm excitation triggers ER release of TfR-GFP-DHFR selectively in targeted cells. Left: Focal illumination with 405 nm light (region marked by the pink dashed line) triggers TfR-GFP-DHFR (green) trafficking to the GA (arrowhead) and the surface (right, Alexa647-anti-GFP puncta shown in red) only in the photoactivated cell. Magnified images (taken from the regions marked by the white boxes) to the right show TfR-GFP-DHFR surface label (arrowheads) in the photoactivated cell (“stim.”; top) and neighboring control cell (bottom). Scale bar, 20 μm. Inset scale bar, 10 μm. The plot to the right shows the average time course of TfR-GFP-DHFR surface accumulation for photoactivated cells (pink line) and neighboring control cells that were not photoactivated (gray line); mean ± SEM (n=6 photoactivated cells, 8 control cells from 2 independent experiments).

Data reproduced with permission from ref. 42.

3.3. Live cell imaging: Local ER release from single cells and subcellular domains

3.3a. Local release in COS-7 cells

1. Performs steps 1–7 from section 3.2.

2. Identify a single cell or cells to target

3. Using a light source capable of targeted illumination (we use the FRAPPA system from ANDOR to steer a diffraction limited spot from a 405nm laser source), locally stimulate the region of interest. With the FRAPPA, we typically scan the region of interest with a pixel dwell time of 1msec with 405nm power =23.0 μW/μm² measured through a 60x Plan Apochromat NA 1.42 objective). An example of a local release experiment from a single targeted COS7 cell is shown in Fig. 3D.

3.3b. Local release in neurons

1. Performs steps 1–7 from section 3.2.

2. Initiate local ER release by targeted illumination with 405 nm light of appropriate intensity. We trigger local zapalog photolysis with 23.0 μW/μm² 405 nm illumination (6% total laser power from a 100 mW fiber-coupled laser) using a FRAPPA unit to steer a diffraction limited 405 nm spot with a 0.5–1 msec dwell time. Examples of local release experiments where cargo is release either from the soma, or from a defined section of neuronal dendrite is shown in Fig. 4. Once more, efficacy of light-triggered ER release can be assessed by redistribution of cargo to the Golgi apparatus and/or mobile intracellular vesicles following ER exit (Fig. 4).

Figure 4.

Local control of secretory trafficking in neurons.

A. A hippocampal neuron expressing DHFR-GFP-Neuroligin1 (NL1) along with FKBP-mCh-KDEL (not shown) was imaged before (top) and 14 min following (bottom) local illumination of the cell body with 405 nm excitation light (purple circle). Note the robust redistribution of GFP signal to the Golgi apparatus in the cell body.

B. Magnified images of the cell body from (A) are shown in grey scale to better visualize the difference in DHFR-GFP-Neuroligin1 localization pre/post illumination.

C. A segment of a dendritic branch on a hippocampal neuron expressing TfR-GFP-DHFR along with FKBP-mCh-KDEL (not shown) was locally illuminated with 405 nm light (purple box). The magnified images to the right demonstrate no detectable accumulation TfR-GFP-DHFR in the cell body Golgi apparatus following local dendritic ER release.

D. Magnified images of the targeted (left) and control (right) dendritic branches from (C) show intracellular trafficking organelles accumulating TfR-GFP-DHFR in the illuminated branch but not in the neighboring control branch, confirming selective initiation of secretory trafficking within the illuminated region.

Data reproduced with permission from ref. 42.

4. Notes:

1. Zapalog is sensitive to ambient light. Zapalog and zapalog-treated cells should be handled in the dark at all times. To prepare 1 M zapalog stocks, dilute in DMSO and make 10 μL aliquots in a tissue culture biosafety cabinet illuminated with red LED safelight. We wrap each aliquot in aluminum foil and store at −20°C. Use these 1 M stocks to prepare the 40 μM working stock on the day of transfection. The 1 M stocks can undergo multiple freeze/thaw cycles.

2. We use the signal peptide sequence from mouse immunoglobulin kappa light chain (METDTLLLWVLLLWVPGSTG) to direct FKBP to the ER. See Figure 2A for more information on anchor construct design.

3. It is critical to know the classification of your protein of interest (e.g. types I–IV) for proper construct design. It may be that your protein of interest contains an internal ER targeting signal (as is the case with the majority of type II–IV transmembrane proteins). The classification also refers to the orientation of the N- and C-termini across the ER membrane. For zapERtrap experiments, DHFR must be fused to the cargo at the terminus that faces the ER lumen, which may not be feasible if this is known to disrupt proper localization of the cargo. In all cases, it is critical to confirm proper subcellular targeting of a non-retained DHFR-fused cargo by comparing it to other fusion constructs (preferably to those that have been previously validated/published). See Figure 2B for more information on cargo construct design.

4. For imaging experiments using heterologous cells, we prefer to use COS-7 cells, which have a flat, “spread out” morphology on the coverslip, allowing clear visualization of trafficking organelles.

5. Instead of using 18 mm coverslips and a Ludin chamber, one can use 35 mm MatTek glass bottom dishes (#1 or #1.5 cover glass, 14 mm glass diameter, poly-d-lysine-coated).

6. The exact amounts of cargo and anchor plasmid DNA ratios must be empirically determined. Bicistronic expression of cargo and anchor is a possibility (see Figure 2C); however, keep in mind that excess anchor can still lead to background leakiness caused by saturation of endogenous KDEL receptor. Thus, it is critical to assess the efficiency of cargo retention using various cargo/anchor DNA ratios (by surface staining for an extracellular epitope tag in dark vs light treated cells).

7. To conserve zapalog compound, it is possible to “recycle” zapalog on sequential days by applying zapalog-conditioned media from a previous experiment to newly transfected cells.

8. Background “leakiness” of cargo to the plasma membrane increases with expression duration. We find it is best to perform imaging experiments 12–18 hours after transfection, especially when expressing proteins driven by CAG and Synapsin promoters.

9. We plate rat hippocampal neurons at a density of 80,000–100,000 cells per well from a 12-well plate. We plate rat cortical neurons at a density of 175,000–200,000 cells per well from a 12-well plate.

10. Incubate cells with 1 unit/mL thrombin 10–30 minutes prior to imaging in order to cleave GFP (or other affinity tag) and thereby prevent surface signal detection of DHFR-fused cargoes already present on the plasma membrane prior to 405 nm light illumination. Background “leakiness” of cargo to the plasma membrane varies with the target cargo under investigation and from cell to cell in the same experiment, and must be empirically determined. We find thrombin protease cleavage to be faster and more complete than Tobacco Etch Virus (TEV) protease cleavage.

11. We find robust ER retention for at least an hour after washout of zapalog, so inclusion of zapalog in the imaging solution may be unnecessary for short-term imaging experiments. We typically include 0.5–1 μM zapalog in our imaging solution as we typically image for ~2 hours after release.

12. Alexa Fluor 647-anti-GFP (Invitrogen, cat. no. A31852) binding to surface cargoes occurs within tens of seconds; however, we recommend waiting 10–15 minutes after antibody addition before baseline image acquisition to confirm that there is no baseline “leakiness” present in the cell(s) under investigation. Note that presence of dye-conjugated antibody will result in background fluorescence that may obscure weak surface delivery. If this is the case, the antibody can be washed off for imaging at different time points.

13. To minimize cell phototoxicity and maximize detection sensitivity of intracellular cargo trafficking, we perform live-cell labeling of HaloTag-fused cargo constructs using the cell permanent Janelia Fluor 646 HaloTag ligand. Imaging with 640 nm excitation light is less phototoxic to cells than 488 nm excitation light. For combined intracellular trafficking and surface detection, we use DHFR-fused cargoes tagged with both HaloTag and GFP, and Alexa Fluor 555-conjugated GFP antibody for surface detection (GFP antibody binding yields brighter, punctate signaling compared to halo-dye).