Modulating organelle distribution using light-inducible heterodimerization in C. elegans

Summary The relative positioning of organelles underlies fundamental cellular processes, including signaling, polarization, and cellular growth. Here, we describe the usage of a light-dependent heterodimerization system, LOVpep-ePDZ, to alter organelle positioning locally and reversibly in order to study the functional consequences of organelle positioning. The protocol gives details on how to accomplish expression of fusion proteins encoding this system, describes the imaging parameters to achieve subcellular activation in C. elegans, and may be adapted for use in other model systems. For complete details on the use and execution of this protocol, please refer to De Henau et al. (2020).


SUMMARY
The relative positioning of organelles underlies fundamental cellular processes, including signaling, polarization, and cellular growth. Here, we describe the usage of a light-dependent heterodimerization system, LOVpep-ePDZ, to alter organelle positioning locally and reversibly in order to study the functional consequences of organelle positioning. The protocol gives details on how to accomplish expression of fusion proteins encoding this system, describes the imaging parameters to achieve subcellular activation in C. elegans, and may be adapted for use in other model systems. For complete details on the use and execution of this protocol, please refer to De Henau et al. (2020).

BEFORE YOU BEGIN
Manipulation of subcellular organelle positioning can be achieved using the light-dependent heterodimerization system LOVpep-ePDZ, an example of an optogenetic approach, in which a photosensitive LOVpep domain binds an engineered PDZ domain (ePDZ) after exposure to blue light (<500 nm) (Strickland et al., 2012). In the following, we describe how to implement and use this system. While we focus on the application of this system in the Caenorhabditis elegans zygote, most principles described here are likely adaptable to any transparent biological model system or organism.
To set up this system, plasmids encoding LOVpep and ePDZ need to be generated, where one is fused to an organelle targeting sequence and the other to a protein domain that localizes to the desired site for organelle relocation (Figure 1). Subsequent stable integration of these constructs is preferred: this allows the relative concentrations between dimerizing proteins to be more consistent, which improves the efficiency of light-induced activation (Krishnamurthy et al., 2016). Stable integration is also essential for expression in the C. elegans germline and early embryo. However, transient expression can be useful for rapid screening of functional constructs. In the following, we describe how to design and generate the constructs encoding the LOVpep-ePDZ system.
Note: Besides LOVpep-ePDZ, a number of other light-inducible protein dimerization techniques have been developed. The binding affinity and activation and reversion kinetics of these light-inducible protein dimerization systems are important parameters to consider: systems with a tighter affinity are more suited to induce a fast functional response. However, it also makes them more sensitive to residual dark-state binding. For multiday experiments, minimal dark-state binding seems crucial to avoid unwanted background activation and perturbation of the system. The activation and reversion kinetics in turn determine how frequently the system needs to be exposed to maintain dimerization, and faster kinetics allow for more temporal resolution Pathak et al., 2014;Strickland et al., 2012).
A third example of a blue-light-inducible dimer is the CRY2-CIB1 pair. Rather than an intrinsic change in affinity upon blue light stimulation, it appears it is the light-induced homo-oligomerization of CRY2 that drives colocalization with CIB1 (Bugaj et al., 2013;Hallett et al., 2016;Kennedy et al., 2010;Lee et al., 2014). Because of this, orientation-specific tagging effects of either component need to be considered when implementing this system. Compared with the previous two methods, CRY2-CIB1 shows slower activation and reversion kinetics . (B) Scheme for the induced trapping of mitochondria with TOMM20::HALO::ePDZ to membrane EGFR-TM::mTagBFP2::LOV and underlying construct design. (C) Scheme for the induced transport of mitochondria with TOMM20::HALO::LOVpep via dynein heavy chain fused to ePDZ and mCherry, together with underlying construct design.
A final example of a light-sensitive heterodimerization system is the Phy/Pif pair, which uses a different light spectrum than the three previous methods: this pair forms under red light and dissociates in darkness or under far-red illumination. Compared to the previous systems, it shows higher fold levels of activation and low background binding. This method is limited in that it requires the external addition of the cofactor phycocyanobilin, which is not always straightforward (Adrian et al., 2017;Levskaya et al., 2009;Pathak et al., 2014).
A direct comparison of these systems, which allows for a better understanding of their unique advantages and limitations, can for example be found in Hallett et al. (2016) and Pathak et al. (2014). We advise the reader to carefully assess the requirements of their own light-dependent heterodimerization experiments before deciding on which system to use.

Design the plasmids containing the LOVpep and ePDZ constructs
Timing: 1-4 h 1. Choose sequences encoding proteins or protein domains that connect LOVpep and ePDZ to the organelle of interest and to the protein that will be used to relocalize the organelle to the desired site respectively (Figure 1). To achieve local activation, it is critical that diffusion of activated LOVpep outside of the activated region is limited as much as possible. To this end, LOVpep is tagged to the least dynamic structure of the two. For example, to transport mitochondria along microtubules we fused LOVpep to the relatively immobile mitochondria using TOMM-20 and fused dynein heavy chain DHC-1 to ePDZ ( Figure 1C) (Fan et al., 2019). On the other hand, to trap mitochondria at the cell membrane we fused LOVpep to a transmembrane domain to localize it to the plasma membrane, while ePDZ was now fused to the relatively more mobile mitochondria using TOMM-20 ( Figure 1B) (De Henau et al., 2020). 2. ePDZ can be placed either N-or C-terminal of the protein (domain) of choice in the final construct.
On the other hand, LOVpep can only be placed at the C terminus of the final construct. LOVpep interacts with ePDZ via a C-terminal Ja-helix, which needs to remain free to maintain its functionality Strickland et al., 2012). 3. Ideally, label both constructs with fluorophores to ensure proper expression and localization.
Keep in mind here that LOVpep is traditionally activated using 470-500 nm light (Harper et al., 2003;Strickland et al., 2012) but is also significantly activated by shorter and slightly longer wavelengths. For example, we observed light-induced LOVpep-ePDZ heterodimerization using an excitation power of 4.8 mW 405 nm laser light, 0.002 mW 458 nm laser light and 0.005 mW 514 nm laser light (all 5 iterations and pixel dwell time of 8 ms, exposure every 2 s). These settings are considerably lower than what we normally use to excite blue, green, and yellow fluorophores. Therefore, in case the LOVpep or ePDZ constructs need to be visualized and followed during light-activation experiments, one cannot use fluorophores that require excitation with wavelengths between 400 and 520 nm, such as mTAGBFP2, eGFP and Venus . Excitation of fluorophores with an excitation spectrum that is distinct from the excitation wavelength of the LOVpep domain, such as mCherry, mScarlet and HaloTag combined with the JF 646 Halo-Tag ligand (Grimm et al., 2015), do not induce LOVpep-ePDZ dimerization. When it is not required to visualize LOVpep or ePDZ during light-activation experiments, you can tag them with mTAGBFP2 or eGFP to ensure proper expression. This way, higher wavelengths remain free for visualization of other proteins or biological processes. 4. It is recommended to use suitable linkers (such as GGSGGSGGS or GAGAGAGAGAGA) between the LOVpep, ePDZ, and the protein domains of choice to allow for proper protein folding. 5. LOVpep-ePDZ is sensitive to differences in expression levels because it is restricted to a 6-fold increase in dimerization affinity upon illumination. It is therefore recommended to use identical or comparably strong regulatory regions (promoter, 3 0 UTR) to drive expression. When going for single-copy integration techniques, aim for regulatory regions that cause moderate-strong protein expression, in order to achieve LOVpep and ePDZ levels that are sufficiently high to ll OPEN ACCESS STAR Protocols 2, 100273, March 19, 2021 permit light-depending LOVpep-ePDZ interactions. We successfully used mex-5 promoter or fbf-1 promoter in combination with tbb-2 3 0 UTR (sequences and expression patterns detailed in Merritt et al. (2008)) to achieve LOVpep and ePDZ germline expression that is compatible with lightdepending activation (De Henau et al., 2020;Fan et al., 2019). 6. Once the transgene is designed, codon optimization for expression in C. elegans (Redemann et al., 2011), or the model organism of choice, is advised. To specifically increase expression in the C. elegans germline and avoid germline silencing, germline-specific codon optimization is available and highly recommended (Fielmich et al., 2018). Additional approaches to avoid germline silencing are the removal of homology to piRNAs (Batista et al., 2008) and introduction of PATC introns (Frokjaer-Jensen et al., 2016).
CRITICAL: Make sure that the final ePDZ and LOV constructs show an overlapping subcellular expression pattern, so that upon light activation LOVpep is capable of binding ePDZ. When using motor proteins, make sure that they are active in the cell of interest. Alternatively, use constitutively active or inducible active motor proteins (Nijenhuis et al., 2020). Also make sure the cell of interest has the cytoskeletal structures that are required for the motor proteins to relocate the organelle of interest.

Assemble and integrate the plasmids carrying LOVpep and ePDZ constructs
Timing: 2 weeks Assembling the LOVpep and ePDZ constructs requires combining multiple genetic elements. Modular cloning techniques such as Gateway (Walhout et al., 2000), Gibson (Gibson et al., 2009) and Golden Gate (Engler et al., 2009) are preferred, given that they are faster and hence more favorable for screening for functional constructs. New constructs for photo-inducible heterodimerization do not always result in the expected organelle manipulation and might for example suffer from high levels of dark-state heterodimerization or low efficiency light-induced heterodimerization. It might therefore be necessary to test alternative organelle adaptors, relocalization proteins, mutants of LOV-ePDZ with different affinity properties (Strickland et al., 2012) or even different light-inducible heterodimerization systems (Adrian et al., 2017;Hallett et al., 2016;Nijenhuis et al., 2020) before the desired experimental setup is achieved.
To stably introduce transgenes in C. elegans, we recommend generating plasmids that are compatible with the Mos1-mediated Single-Copy Insertion (MosSCI) method (Frokjaer-Jensen et al., 2012;Frokjaer-Jensen et al., 2008;Frokjaer-Jensen et al., 2014). This method creates single-copy transgenes at defined positions within the genome, which facilitates subsequent crossing of the transgenes into a single line, generates comparable expression levels of the transgenes and decreases the probability of germline silencing. Alternatively, CRISPR/Cas9 mediated gene editing (Nance and Frokjaer-Jensen, 2019) can be used when endogenous proteins need to be tagged.
With the above in mind, we recommend the recently developed Golden Gate cloning technique SapTrap to assemble C. elegans MosSCI transgene vectors (Fan et al., 2019) carrying LOVpep and ePDZ constructs in a fast, efficient, inexpensive, and scar-free manner ( Figure 2). The MosSCI backbone and donor plasmids carrying LOVpep and ePDZ that are compatible with this method are available at Addgene (Fan et al., 2019). In the following we detail how MosSCI transgene vectors using the SapTrap method can be generated.
7. Find templates encoding the parts you want to combine and order primers to produce gene fragments flanked by SapI restriction sites. Key in the SapTrap method is that SapI cuts DNA at defined positions adjacent to its recognition sequence to generate three-base 5 0 overhangs ( Figure 2, red sequence/bars=recognition site, blue sequence/bars= three-base 5 0 overhangs). By designing SapI restriction fragments with complementary overhangs, multiple fragments can ll OPEN ACCESS be assembled together in a defined order in a single digestion and ligation reaction ( Figure 2B, the defined order is illustrated by the fragments ranging from light gray to dark gray). Primer design and examples of primers can be found in Figure 2C and in the study by Fan et al. (2019). If templates are not available, codon-optimized sequences can be ordered. We order sequences as gBlocks (Integrated DNA Technologies). 8. When you amplify your gene fragments with your primers and template DNA, use a high-fidelity PCR kit as per the manufacturer's instructions. 9. Generate donor plasmids by cloning the PCR products or gBlocks into the pCR BluntII vector backbone using the Zero Blunt Topo system (Thermo Fisher Scientific), as per the manufacturer's instructions ( Figure 2A). 10. Transform the Zero Blunt Topo assembly reaction into chemically competent cells: Thaw 50 mL competent cells on ice. Still on ice, add 5 mL of assembly reaction to the cells and incubate for 20-30 min. Heat shock the cells at 42 C for 30 s. After transformation and before plating, add 0.2 mL of 18 C-24 C SOC Medium and shake at 225 rpm 37 C for 1 h. Spread the cells onto pre-warmed kanamycin-selective plates and incubate 16-20 h at 37 C.
Pause point: Bacterial colonies can be stored at 4 C for up to 2-3 weeks.
11. Pick colonies, grow 16-20 h at 37 C and extract the plasmid DNA using a miniprep kit, as per the manufacturer's instructions. Carry out diagnostic restriction enzyme digests followed by sequencing to confirm that the donor plasmid has the correct, mutation-free insert. 12. Adjust the volume of donor plasmids to bring to a final concentration of 50 nM. 13. MosSCI targeting vectors are assembled using the SapTrap method in a single tube ( Figure 2B) using the following protocol: add 1 mL 50 nM pXF87 (MosSCI backbone, available at Addgene), 1 mL 50 nM of each donor cassette plasmid, 5 mL 103 NEB cutsmart buffer, 5 mL 10 mM ATP (not dATP), 1 mL SapI enzyme (10 units), 1 mL T4 DNA ligase (400 units) and ddH 2 O to a final volume of 50 mL. Incubate this reaction mixture 5 min at 37 C (=SapI digestion) and 5 min at 16 C (ligation), repeating this for a total of 35 cycles. This is followed by a final SapI digestion step of at least 1 h and up to 16 h at 37 C to cut any remaining, unligated pXF87 backbone. After this final step, put immediately on ice and use 5 mL for transformation into chemically competent cells as described above. Spread the cells onto pre-warmed ampicillin-selective plates and incubate 16-20 h at 37 C (Troubleshooting 1, Troubleshooting 2). 14. Isolate plasmid DNA and screen for correct assembly using diagnostic restriction and sequencing. In our hands, between 10%-90% of colonies have the correctly assembled plasmids. 15. Once the MosSCI transgene vectors are made, these can be integrated at defined landing sites in the chromosome of choice by injecting the vectors into universal MosSCI strains (Frokjaer-Jensen et al., 2014). We co-inject the MosSCI transgene vector (50 ng/mL) together with a helper plasmid encoding the Mos1 transposase (50 ng/mL pCFJ601), and with three negative selection markers to select against extrachromosomal array-bearing transgenic animals (10 ng/mL pMA122, 2.5 ng/mL pCFJ90 and 5 ng/mL pCFJ104) (Frokjaer-Jensen et al., 2014). For a detailed protocol describing the microinjection procedure, please refer to Berkowitz et al. (2008). 16. Injection of 20-30 animals is in general sufficient to obtain at least one stable transgenic line.
Place 3-4 injected animals per NGM plates with OP50 E. coli and keep them at 25 C until they are starved (7-10 days). 17. Heat-shock animals for 2 h at 34 C to activate the peel-1 toxin (encoded by pMA122) and selectively kill animals that carry extrachromosomal arrays. Using an air incubator with a fan and spreading out plates evenly ensures that plates warm up relatively fast to 34 C to efficiently induce the heat shock. 18. 4-24 h following the heat shock, screen for plates that contain animals that are alive, move well, and lack the fluorescent co-injection markers. Chunk positive plates to new seeded NGM plates. Two days later, transfer a single adult animal to a new seeded NGM plates. If possible, screen for expression of the inserted transgene (germline expression, for example) before picking a worm to set up a clonal culture. 19. In the isolated lines, verify that the transgene is expressed and localizes to the expected location. For germline expression, note that it can take 1-3 generations before germline silencing of the transgene disappears and the transgene becomes expressed (Troubleshooting 3). 20. Verify that the transgene has been correctly inserted at the chromosome of choice using the oligos described on www.

STEP-BY-STEP METHOD DETAILS Prepare imaging setup
Timing: 10 min CRITICAL: Very low levels of blue light are sufficient to activate LOVpep, so all environmental blue light that could reach the animals during culturing and sample preparation needs to be eliminated as much as practically feasible.
1. Store the animals in a box that blocks environmental light, e.g., by wrapping the box in aluminum foil. 2. If the room used for sample preparation and imaging needs to be illuminated, use a light source filtered for blue light. If necessary, aluminum foil can be used to cover the microscope setup and blue light of monitors can be omitted by adjusting the hardware display settings. 3. Insert optical (orange) filters in the light paths of the dissection scope and of the imaging microscope to remove LOVpep-activating wavelengths from the transmitted light used to handle samples.

Prepare microscope settings for global and local light-induced heterodimerization
Timing: 3-4 h CRITICAL: The conditions used for light-induced activation need to be thoroughly optimized to ensure minimal activation before the experiment or outside the region of interest, especially since LOVpep activation requires very low light exposure.

4.
To setup the microscope, use a strain in which heterodimerization is straightforward to score and where ePDZ is tagged with a red or far-red fluorophore so its distribution can be visualized without activating LOVpep. a. For example, a strain expressing cytosolic ePDZ combined with a membrane anchored LOVpep allows for easy scoring of membrane recruitment of ePDZ-mCherry (De Henau et al., 2020; Fielmich et al., 2018) ( Figure 3A, Methods videos S1 and S2). A cytosolic ePDZ combined with an organelle anchored LOVpep works equally well . 5. Assess for unintended light-induced heterodimerization during sample preparation ( Figure 4).
a. Prepare the sample for imaging, while preventing exposure to environmental blue light as described above. b. Locate the sample of interest using transmitted light that has been filtered for blue light and acquire a snapshot of the ePDZ fusion protein.  parameters for which you can still see ePDZ relocation and which are compatible with your experiments. c. Take along a negative control that only expresses the ePDZ fusion construct. No relocation or phenotype (e.g., light-induced toxicity) should be observable using the settings from (b).
Optional: After acquisition, process the time-lapse images to reduce background noise and segment activated organelle/structure if needed for signal quantification. d. Calculate the ratio of fluorescence intensity inside the activated ROI to the intensity in a ROI of similar size outside the area of activation. This analysis produces a maximum intensity ratio as well as the activation half-life  ( Figure 3B). e. With the results from the previous step, determine the minimal amount of power needed to achieve light-induced activation. Note: The sensitivity and activation and reversion kinetics of LOVpep-ePDZ dimerization following light illumination are important parameters to consider. They determine how often the proteins need to be exposed to blue light in order to induce, maintain, and turn off dimerization to get the desired experimental conditions. Light-induced dimerization of LOVpep-ePDZ occurs with a half-life of approximately 40 s and will revert to its original state in the dark with a reversion half-life of approximately 50 s (Fielmich et al., 2018;Hallett et al., 2016) ( Figure 3B). This is relatively fast and in practice means that photoactivation is carried out in between every image acquisition to maintain dimerization. h. Also here, take along a negative control that only expresses the ePDZ fusion construct. No relocation should be observable using the settings used in (g).
CRITICAL: When designing and interpreting experiments keep in mind that also in the dark-state ePDZ will bind to LOVpep, although with an approximately 6-fold lower affinity compared to the photo-activated state .
Note: Fusing LOVpep to different protein domains might influence activation efficiency or prevent ePDZ binding altogether . For example, we needed approximately 10 times more energy to activate LOVpep anchored to the membrane using a transmembrane region (Figure 6), compared to when anchoring it using a pleckstrin homology domain ( Figure 3B) (De Henau et al., 2020).
7. Perform light-induced heterodimerization experiments a. Using the settings for 488 nm light-induced activation determined in step 6, set up your imaging parameters. b. Locate the sample of interest using transmitted light that has been filtered for blue light. c. As noted above, global light-induced heterodimerization can be achieved similar to acquiring GFP images, while local light-induced heterodimerization is achieved by exposing only the region of interest (ROI) with a 488 nm laser. d. Afterwards, determine if the light-induced heterodimerization of LOVpep and ePDZ has the desired effect on organelle relocation. For example, to determine if local activation of membrane LOVpep effectively trapped mitochondria labeled with ePDZ in the activated area, we used particle tracking analysis of the mitochondrial signal ( Figure 5C) (De Henau et al., 2020). Alternatively, organelle redistribution dynamics can be quantified by comparing changes in total intensity in the activated and non-activated area (Nijenhuis et al., 2020)  Note: Trapping of an organelle, for example trapping of mitochondria near the cell membrane ( Figure 5) (De Henau et al., 2020), will most likely result in only moderate organelle enrichment at the site of interest. Trapping also requires the organelle of interest to occasionally be present at the site of interest in order for LOV and ePDZ dimerization to be able to occur upon light activation. We therefore prefer to refer to this method as local trapping of an organelle rather than relocation.

EXPECTED OUTCOMES
Using a strain in which heterodimerization is straightforward to score, such as cytosolic ePDZ and membrane anchored LOVpep, there will be no to minimal dimerization observable in dark state (Figure 6). In such a strain, global light activation will cause a clear and fast dimerization of ePDZ and LOVpep (Methods video S1). Local light activation will cause clear ePDZ-LOVpep dimerization in the activated region, with an expected 2-to 6-fold increase in relative fluorescence intensity of ePDZ in the subcellular location of LOVpep, with lower levels of dimerization observable in the surrounding region (Methods video S2, Figures 3 and 6). Exposure to laser light of 561 nm or above will not cause dimerization, nor exposure to transmitted light passed through an orange optical filter. LOVpep and ePDZ fusion constructs that are expressed individually will not change subcellular distribution upon light activation.
Using LOVpep-ePDZ for organelle relocation will cause moderate to strong redistribution of the tagged organelle, depending on the cellular machinery that is used (and available) for relocation (Figure 7, compare panel C and E, Methods videos S3, S4, and S5).

LIMITATIONS
Blue-light-induced LOVpep-ePDZ dimerization is a powerful technique that allows to address the importance of protein and organelle positioning. This technique is applicable to many organisms, and the single wavelength of light necessary to manipulate their dimerization makes for a simple experimental setup. However, newly designed constructs for photo-inducible heterodimerization do not always result in the desired organelle manipulation and/or can show unwanted dark-state

OPEN ACCESS
binding, and it is likely that alternative organelle adaptors, proteins used for relocalization, LOV-ePDZ variants with altered affinity (Strickland et al., 2012) or even other light-inducible heterodimerization systems need to be tested. In addition, LOVpep does not tolerate C-terminal fusions, posing a problem to directly label a number of organelle adaptors (van Bergeijk et al., 2015). It is therefore recommended to combine the technique of light-inducible heterodimerization with an efficient and fast cloning approach to be able to rapidly test alternative fusion constructs.
Given the high sensitivity of the LOVpep-ePDZ system, preventing unwanted activation by environmental light or light scattering during culturing and during imaging can be difficult (see for example Figure 3B, asterisks). Activation is also caused by imaging fluorescent proteins that have spectral overlap with the LOVpep domain, such as mTagBFP2 and Venus. Combining the LOVpep-ePDZ system with multiple fluorescent proteins can therefore become challenging and requires careful controls for aberrant activation. In addition, also in the dark-state ePDZ will bind to LOVpep, and while this is with an approximately 6-fold lower affinity compared to the photo-activated state LOVpep-ePDZ dimerization following light illumination has a relatively high activation and reversion half-life. While these properties allow to control organelle position with high spatial and temporal resolution, they also impose that the system needs to be continuously illuminated with blue light if stable activation is desired. This could be technically challenging to combine with fast live imaging and might potentially induce phototoxicity in long-term imaging sessions.
Finally, as discussed earlier, several light-inducible protein dimerization techniques have been developed, each with their own sensitivity and dynamic range. The choice of system to use will ultimately depend on the requirements for speed of activation, reversibility, and depth of tissue to be accessed. It is therefore recommended to understand the advantages of each system before deciding which one to use Nijenhuis et al., 2020;van Bergeijk et al., 2015).

Problem 1
No colonies after transformation of the SapTrap assembly

Potential solution
Make sure all the SapI cleavage sites are compatible. Make sure to resuspend SapI before pipetting, SapI is prone to precipitation. Make sure to use ATP and not dATP. Upon heterodimerization of LOVpep-ePDZ, dynein is expected to transport mitochondria along these microtubules in retrograde direction, toward the centrosomes and (pro)nuclei.
(C-E) (C) Mitochondrial dynamics in WT embryos that do not express LOVpep and ePDZ, and (D-E) in embryos that express dynein heavy chain fused to ePDZ (epdz::mcherry::dhc-1) and mitochondria labeled with LOVpep. Embryos in (C) and (E) were exposed to 488 nm laser light, with laser intensity at 0.001% and pixel dwell time of 8 ms, applied in between each time point (5 s interval). The embryo in (D) was imaged under the same conditions but with the 488nm laser shut off and shielded from all environmental blue light. Note that mitochondria relocate toward the center of the cell (= toward the (pro) nuclei), in a moderate manner in (D) and an even more pronounced manner in (E). Anterior side of the embryos is to the left, t(0) = contact between maternal and paternal pronuclei. Mitochondria were visualized using Mitotracker Deep Red FM (Thermo Fisher Scientific).

Problem 2 Many colonies with no insert after transformation of the SapTrap assembly
Potential solution Make sure the last 37 C cutting step is at least 1 h. In addition, place the reaction tube immediately on ice after the SapTrap assembly protocol is finished. Leaving it at room temperature (18 C-25 C) allows re-ligation of the cut empty pXF87 backbone.

Problem 3
No expression of a transgene construct designed for germline expression

Potential solution
Verify that the construct has no internal stop codons or design flaws, for example by expressing it in non-germline tissues. If the construct is properly expressed in non-germline tissues, it likely suffers from germline silencing when targeted to the germline. To circumvent silencing, optimize the sequence for germline expression (Fielmich et al., 2018), remove homology to piRNAs (Batista et al., 2008) and/or introduce PATC introns (Frokjaer-Jensen et al., 2016).

Problem 4 Significant unintended light-induced heterodimerization during sample preparation
Potential solution Make sure to work in a closed room so all environmental light that might reach the sample can be controlled. Ideally, perform sample preparation and imaging in the same room to avoid light exposure in between these two steps.

Problem 5
Significant light-induced heterodimerization outside the activated ROI

Potential solution
Lower the exposure to the activating 488 nm laser to reduce light scattering. In addition, consider if the mobility of the tags used to anchor LOVpep might explain diffusion of activated heterodimers outside the ROI and potentially replace these with tags with lower mobility.

Problem 6
No light-induced heterodimerization or organelle translocation

Potential solution
Analyze if LOVpep and ePDZ in the fusion constructs are still functional, by combining them with characterized and suitable complementary constructs. For example, to know if an organelle targeted LOVPEP fusion construct is functional, combine it with a cytoplasmic ePDZ and determine if ePDZ relocates to the organelle after light activation.
If both constructs are functional but no organelle translocation is observed, make sure that the constructs effectively overlap in subcellular location and are able to interact. When motor proteins are used, make sure that they are active at the time of light activation and that the cytoskeletal network that is needed for translocation is available.

Problem 7
Heterodimerization or organelle translocation is observed in dark-state conditions. Potential solution Firstly, make sure that the dark-state condition lacks all forms of blue light, by using filters that block blue light in the dissection scope and microscope, illuminating the working room with red light instead of white light, turning off blue light emission from computer monitors, etc.
Secondly, the binding affinity of LOVpep and ePDZ in dark-state can be sufficient to cause constitutive and unwanted organelle translocation even in the absence of all forms of blue light. As mentioned above, we observed a clear example of dark-state mitochondrial translocation when using a combination of mitochondrial ePDZ and LOVpep fused to dynein heavy chain DHC-1 ( Figure 7D).
Reducing dark-state activation might be achieved by lowering expression levels of both constructs or of the LOVpep containing construct in case overexpression constructs are used (Nijenhuis et al., 2020). Fusing alternative motor proteins to LOVpep could also help to reduce dark-state activation.
A noteworthy example here is the development of a photosensitive kinesin that is activated upon blue light. This adds a second layer of light-sensitive control and effectively reduced dark-state activation (Nijenhuis et al., 2020).
Finally, mutants of LOVpep or ePDZ with lower dark-state binding affinity (Strickland et al., 2012), or other light-inducible heterodimerization systems with lower dark-state binding affinity, such as the milli variant of the iLID-SspB system , could help reduce dark-state organelle translocation.

RESOURCE AVAILABILITY
Lead contact Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Sasha De Henau (sasha.dehenau@gmail.com).

Materials availability
This study did not generate any unique materials or reagents.

Data and code availability
This study did not generate any unique datasets or code.