Caenorhabditis elegans (C. elegans) have proved to be a valuable model system for studying developmental and cell biological processes. Understanding these biological processes often requires long-term and repeated imaging of the same animal. Long recovery times associated with conventional immobilization methods done on agar pads have detrimental effects on animal health making it inappropriate to repeatedly image the same animal over long periods of time. This paper describes a microfluidic chip design, fabrication method, on-chip C. elegans culturing protocol, and three examples of long-term imaging to study developmental processes in individual animals. The chip, fabricated with polydimethylsiloxane and bonded on a cover glass, immobilizes animals on a glass substrate using an elastomeric membrane that is deflected using nitrogen gas. Complete immobilization of C. elegans enables robust time-lapse imaging of cellular and sub-cellular events in an anesthetic-free manner. A channel geometry with a large cross-section allows the animal to move freely within two partially sealed isolation membranes permitting growth in the channel with a continuous food supply. Using this simple chip, imaging of developmental phenomena such as neuronal process growth, vulval development, and dendritic arborization in the PVD sensory neurons, as the animal grows inside the channel, can be performed. The long-term growth and imaging chip operates with a single pressure line, no external valves, inexpensive fluidic consumables, and utilizes standard worm handling protocols that can easily be adapted by other laboratories using C. elegans.
Caenorhabditis elegans has proved to be a powerful model organism to study cell biology, aging, development biology, and neurobiology. Advantages such as its transparent body, short life cycle, easy maintenance, a defined number of cells, homology with several human genes, and well-studied genetics have led to C. elegans becoming a popular model both for fundamental biology discoveries and applied research1,2. Understanding cell's biological and developmental processes from repeated long-term observation of individual animals can prove to be beneficial. Conventionally, C. elegans is anesthetized on agar pads and imaged under the microscope. Adverse effects of anesthetics on the health of animals limit the use of anesthetized animals for long-term and repeated intermittent imaging of the same animal3,4. Recent advances in microfluidic technologies and their adaptation for anesthetic-free trapping of C. elegans with negligible health hazards enable high-resolution imaging of the same animal over a short and long period of time.
Microfluidic chips have been designed for C. elegans'5 high throughput screening6,7,8, trapping and dispensing9, drug screening10,11, neuron stimulation with high-resolution imaging12, and high-resolution imaging of the animal12,13,14. Ultra-thin microfluidic sheets for immobilization on slides have also been developed15. Long-term studies of C. elegans have been performed using low-resolution images of animals growing in liquid culture to observe growth, calcium dynamics, drug effects on their behavior16,17,18,19, their longevity, and aging20. Long-term studies using high-resolution microscopy have been carried out to assess synaptic development21, neuronal regeneration22, and mitochondrial addition23. Long-term high-resolution imaging and tracing of cell fate and differentiation have been done in multichannel devices24,25. Several cellular and sub-cellular events occur over the time scales of several hours and require trapping the same individual at different time points during their development to characterize all intermediate steps in the process to understand cellular dynamics in vivo. To image biological process such as organogenesis, neuronal development, and cell migration, the animal needs to be immobilized in the same orientation at multiple time points. We have previously published a protocol for high-resolution imaging of C. elegans for over 36 h to determine where mitochondria are added along the touch receptor neurons (TRNs)23.
This paper provides a protocol for establishing a microfluidics-based methodology for repeated high-resolution imaging. This device, with a single flow channel, is best suited for repeated imaging of a single animal per device. To improve throughput and image many animals at once, multiple devices could be connected to the same pressure line but with separate three-way connectors controlling a single animal in each device. The design is useful for studies that demand high-resolution time-lapse images such as post-embryonic developmental processes, cell migration, organelle transport, gene expression studies, etc. The technology could be limiting for some applications such as lifespan and aging studies that require parallel growth and imaging of many late-stage animals. Polydimethylsiloxane (PDMS) elastomer was used for fabricating this device due to its biostability26, biocompatibility27,28, gas permeablility29,30, and tunable elastic modulus31. This two-layer device allows the growth of animals with continuous food supply in a microfluidic channel and the trapping of individual C. elegans via PDMS membrane compression using nitrogen gas. This device is an extension of the previously published device with the advantage of growing and imaging the same animal in the microchannel under a continuous food supply3. The additional isolation membrane network and a 2 mm wide trapping membrane enable efficient immobilization of developing animals. The device has been used to observe neuronal development, vulval development, and dendritic arborization in sensory PVD neurons. The animals grow without adverse health effects in the device and can be repeatedly immobilized to facilitate imaging sub-cellular events in the same animal during its development.
The entire protocol is divided into five parts. Part 1 describes device fabrication for the growth and imaging chip. Part 2 describes how to set up a pressure system for the PDMS membrane deflection to immobilize and isolate individual C. elegans. Part 3 describes how to synchronize C. elegans on a nematode growth medium (NGM) plate for device imaging. Part 4 describes how to load a single animal in the device and grow the animal inside the microfluidic device for several days. Part 5 describes how to immobilize an individual animal at multiple time points, capture high-resolution images using different objectives, and analyze the images using Fiji.
1. Fabrication of growth and imaging device
- SU8 mold fabrication
- Design patterns 1 (flow layer) and 2 (control layer) using rectangular shapes in a word processing software (or a computer-aided design CAD software) and print the photomasks with the help of a laser plotter with a minimum feature size of 8 µm on polyester-based film (Figure 1).
- Cut silicon wafers in 2.5 cm × 2.5 cm pieces and clean them with 20% KOH for 1 min. Rinse the wafers in deionized (DI) water. Use one wafer each for the flow and the control layer.
CAUTION: KOH is corrosive and should be handled with care.
- Dry the pieces with 14 psi compressed nitrogen gas followed by dehydration on a hot plate at 120 °C for 4 h. Before proceeding to the next step, cool the two pieces down to room temperature.
- Take one of the silicon pieces and put it on the chuck of a spin coater and turn on the vacuum to hold the wafer in place. On the silicon piece, put ~20 µL of hexamethyldisilane (HMDS) and coat it using the spin coater at 500 rotation per min (rpm) for 5 s followed by 3,000 rpm for 30 s.
CAUTION: This step should be performed in yellow light. Do not use white light in the room.
- To get a uniform photoresist thickness of ~40 µm (specific to the flow layer; suitable for imaging early larval stage 1 to stage 3 (L1 - L3) animals), coat the silicon wafer with ~1.5 mL of negative photoresist-1 using a spin coater at 500 rpm for 5 s followed by 2,000 rpm for 30 s.
- Repeat steps 1.1.4 and 1.1.5 with the second wafer to obtain a uniform photoresist thickness of ~40 µm specific to the control layer.
- Alternatively, to increase the thickness of the flow layer to ~80 µm for older animals, coat silicon wafers with ~1.5 mL of negative photoresist-2 using the spin coater at 500 rpm for 5 s followed by 2,000 rpm for 30 s. This thickness is suitable for L3 stage to adult animals.
CAUTION: Hold the silicon wafers by their sides to avoid any damage to the spin-coated layers. Keep white lights turned off during this step.
- Bake the photoresist-coated silicon pieces (for flow and control layers) on a hot plate at 65 °C for 1 min followed by 95 °C for 10 min. Cool the baked pieces to room temperature.
NOTE: The baked silicon pieces can be stored for one day before proceeding to the next step. Store in the dark and do not expose to white light.
- Put soft-baked silicon pieces on the exposure stage of the UV illuminator with the photoresist-coated surface facing the UV lamp. Expose the two pieces separately to UV for 15 s, using a 200 W lamp, through a photomask with patterns 1 and 2 to get flow and control layers, respectively.
CAUTION: Wear safety goggles and avoid direct exposure to UV light. Do not turn on white light in the room during this stage.
- Bake the two exposed silicon pieces with coated layer facing up, at 65 °C followed by 95 °C for 1 min and 10 min respectively. Cool the pieces to room temperature before proceeding to the next step.
- Develop the patterns by soaking the silicon pieces in the photoresist developer solution (1:3 dilution of the developer in isopropanol) for 20 min. Once the pattern is visible, rinse the pieces with pure iso-propyl alcohol (IPA) and gently blow dry using nitrogen gas (14 psi).
CAUTION: Use a well-ventilated environment for this chemical treatment to avoid human exposure. Use white light only after features are developed and rinsed with IPA.
- Keep the silicon pieces in a desiccator with the coated surface facing up. Expose the pieces to silane vapors by pouring 50 µL of pure trichloro (1H, 1H, 2H, 2H-perfluorooctyl) silane on a small plastic cup or a glass slide. Place the cup/slide inside a desiccator and incubate for 2 h.
CAUTION: Avoid direct exposure to silane vapor. Always use a sealed chamber for silane vapor treatment.
NOTE: If necessary, the developed silicon pieces can be stored for 1-2 days before proceeding to the next step.
- PDMS chip fabrication
- Make PDMS in a plastic cup by mixing the elastomer base with the curing agent in a 10:1 ratio. Mix the contents well by stirring constantly for 3 min. The mixing will create a lot of air bubbles in the PDMS mix.
- Degas the PDMS mix in a desiccator for 30 min to remove all air bubbles.
CAUTION: Ensure air bubbles are removed from the PDMS mix before it is poured on the features since bubbles can cause defective and non-functional devices.
- Place the silicon wafers with the control layer (pattern 2) in a Petri dish. Gently pour a 5 mm thick PDMS mix layer on the silicon piece avoiding any bubble formation.
- Degas the PDMS mix in a desiccator to remove additional bubbles that are formed during PDMS pouring process.
- Place the silicon wafer with flow layer (pattern 1) on the spinner chuck applying 200-500 mTorr vacuum pressure to hold the wafer. Pour ~1 mL of PDMS on the silicon wafer and coat it using a spin coater at 500 rpm for 5 s followed by 1,000 rpm for 30 s to get an ~80 µm thick layer.
- Bake the two silicon wafers with the spin coated PDMS and poured PDMS layers at 50 °C in a hot air convection oven for 6 h. After baking, wait for the pieces to cool down at room temperature.
- Cut the 5 mm thick PDMS layer from the silicon piece around the control layer (pattern 2) using a sharp blade and peel it off from the silicon substrate.
- Punch two holes of ~1 mm diameter using a Harris puncher at the reservoir of the PDMS block to connect the immobilization channel and isolation channel inlets to the gas lines for PDMS membrane deflections.
- Place the silicon piece with the spin coated PDMS layer on the pattern 1 (flow layer), with the PDMS-coated surface facing up, on a plastic tray. Keep the punched PDMS block with pattern 2 (control layer) on the tray with molded side facing up.
- Keep the plastic tray inside a plasma cleaner and expose the two PDMS surfaces to 18 W air plasma for 2 min under low vacuum (200-600 mTorr). Apply vacuum until the chamber turns bright violet. Perform this step under low light to see the plasma color change.
- Take out the two plasma-treated blocks and gently bind the blocks by pressing the plasma-treated surfaces of patterns 1 and 2 together. Bake the bonded patterns at 50 °C for 2 h in a hot air convection oven.
- Take the bonded device out of the oven. Cut the bonded device out of silicon wafer with pattern 1 and pattern 2, and punch holes in the inlet and outlet reservoirs of the flow layer using the Harris puncher.
- Place the bonded PDMS block with the flow layer facing up on a plastic tray. Keep a clean cover glass (#1.5) on the same tray. Expose the blocks and the cover glass to 18 W air plasma for 2 min. Adjust vacuum pressure to see a violet chamber.
NOTE: This step should be done in low light to see the plasma color change. To clean the cover glass, wash it with IPA and blow dry using nitrogen gas at 14 psi.
- Place the plasma exposed PDMS block on top of the cover glass and bake the bonded structure in an oven at 50 °C for 2 h. Store the device in a clean chamber for any future experiment.
2. PDMS membrane priming
- Take the device and put it on a stereomicroscope and attach the tubings. Connect micro flex tube (inner diameter ~5 mm, outer diameter ~8 mm) to a compressed nitrogen gas line on one end. Connect a three-way connector on the other end. Tubes 1 and 2 of the three-way connectors will be connected to the trap and isolating membranes, respectively.
- Connect two micro flex tubes (inner diameter ~1.6 mm, outer diameter ~5 mm) to the two outlet ports of the three-way stopcock. Connect the other end of the two tubes to an 8 mm long 18 G needle.
- Fill the flow layer with M9 buffer using a micropipette through the inlet port. Fill both tubes with DI water through the end connected to the needle. Insert the two needles into the punched holes, connecting the isolating and trapping membrane, respectively.
- Open the nitrogen gas regulator at 14 psi and turn the three-way valve from tube 1 to push the water into the device through the microfluidic channels in the control layers namely, the trap and isolation membranes.
- Wait until water fills the channel without the presence of any air droplet in both channels. Once the channels are completely filled with water, the channels are considered to be primed.
- Release the pressure using the three-way stopcock once the channels are filled with water and primed. Priming can lead to bubbles in the flow layer, remove the bubbles by flowing additional media through the flow channel.
3. C. elegans maintenance and synchronization
NOTE: C. elegans strains: The study used following transgenes PS3239 (dpy-20(e1282) syIs49 IV [MH86p(dpy-20(+) + pJB100(ZMP-1::GFP)]) for vulval development32, jsIs609 (mec7p::MLS(mitochondrial matrix localization signal)::GFP)33 for touch receptor neuron (TRN) development and mitochondria transport imaging, and wdIs51(F49H12.4::GFP + unc-119(+)) to track PVD development34. Standard C. elegans culture and maintenance protocol was followed35.
- Grow C. elegans on the nematode growth medium (NGM) Petri plates with E. coli OP50 as the food source at 22 °C. Maintain the C. elegans strains by repeatedly transferring a few hermaphrodites or chunking a small amount of agar with a few animals to a new NGM plate with an OP50 lawn.
- After 3-5 days, check the NGM plate for animal growth and C. elegans eggs. Collect and transfer approximately 30 eggs from a plate to a fresh NGM plate with OP50 lawn.
- To synchronize animals for imaging, transfer all unhatched eggs from the plate every 2 h to a fresh plate and maintain them at 22 °C. Approximately 15-20 eggs will hatch on each plate.
- Pick the animals between 14-16 h and 28-30 h after hatching for larval 2 (L2) and larval 3 (L3) stages, respectively. Transfer the animals to the microfluidic device for imaging. Add food supply as described in the next section to maintain the animal inside the device for long-term growth and imaging experiments.
4. C. elegans growth inside the growth and imaging microfluidic device
- Mount a growth and imaging microfluidic device on an inverted microscope and view pattern 2, after connecting the isolation membranes and immobilization membrane, at low magnifications (4x or 10x). Ensure that the channels are filled with clean distilled water.
- Prepare a fresh 1 L solution of S medium using 10 mL of 1 M potassium citrate pH 6.0, 10 mL of trace metals solution, 3 mL of 1 M CaCl2, 3 mL of 1 M MgSO4. Prepare the solution under sterile conditions. Do not autoclave the S medium.
- Fill the flow channel with growth media (S medium) 10 min before the experiment. Avoid any air bubbles in the flow channel. Flow additional medium if needed to remove air bubbles.
- Pick a single animal from the required developmental stage from an NGM plate in 10 µL of S medium using a micropipette and push the animal into the flow channel through the inlet hole.
- Monitor the animal position in the flow channel using a low magnification objective. Flow additional medium through the inlet or outlet to push the animal and position it within the flow channel restricted between the two isolation membranes.
- Open the three-way stopcock to apply 14 psi pressure in the isolation channels and push the membranes down into the flow channel.
NOTE: The membrane partially seals the flow channel and restricts animal movement to the region between the two isolation membranes.
- Use a single colony of E. coli OP50 from a streaked plate to inoculate 250 mL of L Broth (2.5 g bacto-tryptone, 1.25 g bacto-yeast, 1.25 g NaCl in H2O). Grow the inoculated culture overnight at 37 °C.
- Aliquot 500 µL of OP50 culture into 1.5 mL sterile centrifuge tubes and store the stock and the aliquot for 2 weeks at 4 °C.
- Pellet down OP50 culture by centrifugation at 1.3 x g for 5 min. Dissolve the pellet with 1 mL of fresh S medium (0.5x dilution) and store it at room temperature for 3-4 days. Use this diluted OP50 to feed C. elegans inside microfluidic devices.
- Leave a drop of S medium on top of the inlet and outlet reservoirs to reduce evaporation of the S medium in the flow channel.
- Take diluted OP50 solution in a 10 µL micropipette. Remove the micro tip filled with the OP50 solution from the pipette and press-fit the tip into the inlet reservoir. Then place another micropipette tip with 10 µL of food solution and insert it into the outlet reservoir.
- Seal the tip head applying pressure using a finger to ensure continuity of food solutions without any air gap. Change the micropipette tip with OP50 solution every day that is not older than 3-4 days.
- Fill an additional 20-30 µL of food solution in both the tips. Add or remove food solution to and from the micropipette tip to adjust the gradient to push the animal in the flow channel and to adjust the position of the animal under the trapping membrane for imaging.
- Monitor the flow of bacteria to ensure it is continuous through the partially closed isolation membranes in the flow channel using bright-field images.
NOTE: In absence of bacteria flow, animals might eat the available bacteria in the flow channel between the two isolation membranes. In case of no bacteria or no flow present in the channel, replace the two micropipette tips at the inlet and outlet with new pipette tips filled up with freshly prepared food supplies.
5. C. elegans immobilization and imaging
- C. elegans immobilization under the trapping membrane
- Locate a single animal in the growth channel at low magnifications. Adjust the food solution heights in the two micropipette tips connected to the inlet and outlet reservoirs. Push the animal in the required direction using the hydrostatic pressure differences in the flow channel.
- Position the animal at the center of the trapping membrane and monitor its swimming behavior using a low magnification objective (4x).
- Turn the three-way stopcock to increase the pressure in the trap channel slowly. Immobilize the animal under the trap membrane in a straight posture along the growth channel boundary wall.
CAUTION: Avoid trapping the animal across the flow channel in a Z or U bend position. This causes the animal body to be squeezed with greater pressure and causes permanent damage to its health.
- C. elegans imaging and release from the membrane
- Carry and load the device on a microscope stage. Set up an inverted microscope at the desired imaging settings with all the necessary optical components (objective, light source, fluorescence filters, and a detector) for high-resolution bright field, differential imaging contrast (DIC), or fluorescence imaging (Supplementary Figure 1).
- Acquire a single or several time-lapse fluorescence images to capture cellular and sub-cellular events.
- Acquire fluorescence images of C. elegans neurons as a function of their development using a 60x, 1.4 numerical aperture (NA) objective, a 488 nm wavelength laser with 7%-10% laser power, a CCD camera, on a spinning disc microscope. Perform time-lapse imaging using the software provided with the microscope and acquire images at a speed of 4 frames per second (Supplementary Figure 1).
- After the acquisition of images, release the trapping pressure and monitor the locomotion of the animal at low magnifications - 4x and 10x. Keep the animal restricted within the region defined by isolating membranes (kept under 14 psi all through the experiment).
- Adjust the volume of food solution in the two micropipette tips to continue a slow gravity-driven food flow in the growth channel. This food flow is obtained by adjusting the levels of the food solutions in micropipettes at the inlet and the outlet (typically 1-5 mm height difference).
- Visualize the flow under a bright-field from the flow pattern of the bacteria in the growth channel.
NOTE: In the absence of flow, the animal eats the available OP50 bacteria, the channel appears clear, and the animal starves over a few hours. A starved animal typically develops high autofluorescence, easily observable when acquiring time-lapse fluorescence images. Such events have detrimental effects on animal physiology and were avoided in the experiments.
- Repeat steps 5.2.1-5.2.3 after a predetermined time interval to acquire fluorescence/ DIC/bright-field images of the same individual at multiple time points.
- After the images are acquired at multiple desired time points from the same individual animal, release the isolation and trapping pressure. Flush the channel with M9 buffer and push the animal through the inlet reservoir. Recover the animal from the reservoir and place the animal on a fresh NGM plate for further health monitoring.
- Flush the channel with M9 buffer a few times to remove bacteria. Rinse the flow channel with 70% ethyl alcohol (diluted in distilled water). Dry the channels by pushing air using a syringe. Store the device in a dry dust-free place for repeated use in the future.
- Image analysis and statistics
- Use FIJI ImageJ software for tiff image analysis. Open tiff images of the PVD time-lapse images from the wdIs51 animals in Fiji ImageJ. Extract the best planes from the stack of z-planes showing the primary processes by selecting Image tab > Stacks > Z project.
- Screen the entire series of images and find specific image frames that successfully cover the entire neuronal processes using overlapping sections of the animal image frames. Select Plugin > Analyze > Cell Counter from the FIJI toolbar. This will open a window for numbering different branches and keeping a count of each selected neuron.
- Select Initialize > Counters > Type1, then mark the secondary branches. Then select Type 2 and mark Quaternary, click Results window showing counts of all branches (Supplementary Figure 2). This method will display the branches that are already counted.
- Open the next overlapping image and count the remaining branches. This process will prevent double counting and ensure every process is counted. Identify and count the total number of primary branches present in the early larval stages of C. elegans. Perform similar analysis for the images for secondary and quaternary processes in older animals.
NOTE: Adults show many processes that innervate in body wall muscles and can be difficult to count. Ensure that every process is counted only once.
- Calculate the distance between the PVD cell bodies in the wdIs51 animals by drawing a segmented line from the PVD cell body (CB) to PVC CB. For larval 4 (L4) stage animals, the cell bodies are far apart and are not in the same frame when imaged with a 60x objective.
- Select overlapping images from stack to cover the entire animal length from head-to-tail using Image tab > Stacks > Z project from ImageJ. Draw segmented lines along the neuronal process between the PVD and PVC CBs.
- Measure the lengths for all the segmented lines using ImageJ > Analyze > Measure. Add the lengths for all the segments to calculate the total distance between PVD and PVC CBs for each time point.
- For neuron length of TRNs in jsIs609 animals, load the images in Fiji. Draw a segmented line along the posterior lateral microtubules (PLMs; visible with soluble GFP) from the cell body to the centroid of the first mitochondrion. Calculate the length of the line for the first distance value.
- Draw a segmented line from the center of the first to the second mitochondrion. Repeat this process for each pair of mitochondria till the end of neuronal process length. Add all the lengths to calculate the total neuronal process length.
- For the cell lineage analysis, load all the vulval images in Fiji and extract the best focus image of the cells from the time-lapse images of multiple z-planes.
- Represent the data as mean ± standard error of the mean (SEM). Calculate statistical significance using one-way ANOVA for more than two samples or a two-sample t-test for a pair of samples. Denote significance by p-value < 0.05 (*), p-value < 0.005 (**), and p-value > 0.05 (ns, not significant).
Device characterization: The growth and imaging device consists of two PDMS layers bonded together (Figure 1) using irreversible plasma bonding. The flow layer (pattern 1) which is 10 mm in length and 40 µm or 80 µm in height allows us to grow the animal in liquid culture (Figure 1A). The trapping layer (pattern 2) has a 2 mm wide membrane (Figure 1B) for immobilizing the animal for high-resolution imaging. The mask for the trapping layer also creates a pair of isolation membranes for restricting animal movement within a section of the flow channel between successive imaging time points. The trapping membrane of the device is adapted from our previous imaging device3. The current device (Figure 1C,L) includes the addition of an isolation membrane and an increase in the width of the trapping membrane to 2 mm for the efficient immobilization of the same animal over multiple time points.
To test the device, clean distilled water was filled in the microfluidic channels connecting the trapping membranes (Figure 1D,E, and Supplementary Figure 3) and pressurized using 14 psi nitrogen gas, also used for trapping an animal under a 2 mm thick PDMS membrane (Figure 1H,K,L). A single animal was picked from an NGM plate using a micropipette and loaded into the flow channel (Figure 1F). The growth channel was filled with bacteria resuspended in S media (Figure 1F,G,I,J). A constant flow was maintained throughout the duration of imaging by adjusting the media heights in the pipette tips connected to the device inlet and outlet while monitoring the animal in the flow channel between the two isolation membranes. Freshly prepared OP50 solution was filled in the micropipette tips daily to ensure a healthy food source for the animal inside the microchannel. Fresh food source and gas permeable PDMS material ensure sufficient oxygen supply inside the microchannel30,29. The growth of animals in the device was tracked by using either cell-specific markers or markers that show cell lineage or variable cellular expression patterns during development. The animals were immobilized under the trapping membrane using 14 psi nitrogen gas and holding worms in a straight position along the channel wall. The animal grew slower inside the microfluidic channel compared to on an NGM plate23.
Cell lineage study using the long-term imaging device: To assess C. elegans development inside the microfluidic device, we grew the PS3239 (ZMP-1::GFP) strain32 with constant food supply to track vulva development at different time points of their growth. ZMP-1 codes for a zinc metalloprotease and expresses in the anchor cell in the L3 stage, in vulD and vulE cells in the L4 stage, and in vulA in day 1 (1D) adult animals (Figure 2A). The changes in expression pattern represents an example of temporal gene regulation where the same gene is expressed in different cells at different stages of development. To observe vulval development, the animal was first immobilized and the vulval region was then imaged across multiple z planes every 8-10 h from L3 onwards until adult stages. ZMP-1::GFP is expressed in different vulval cells according to the developmental stage (Figure 2B). The high-resolution fluorescence images of the vulval cells from the animals growing inside the microfluidic device demonstrate normal vulval development and ZMP-1::GFP expression and localization are similar to prior reports32.
Tracking neuron development using long-term growth and imaging device: To demonstrate the use of the device for sub-cellular imaging from an individual animal, the development of two mechanosensory neurons PVD and TRNs was monitored. The NC1686 strain expressing wdIs51 that expresses GFP in the two PVD neurons was used34,36. Each PVD neuron shows menorah-like branched dendritic architecture comprising of primary (PP), secondary (SP), tertiary (TP), and quaternary (QP) processes at L2 (14-16 h), late L2 (20-22 h), L3 (24-26 h), and L4 (36-42 h) developmental stages, respectively (Figure 3). The PVD cell body is present posterior to the vulva. It sends out one axon and two primary dendritic processes that give rise to SP and TP, which in turn give rise to QP dendritic branches that innervate the body wall muscles. Late L3 and L4 stages show high arborization37,34. We grew single NC1686 animals inside the microfluidic device and fed them continuously with bacterial food. The animals were immobilized during L2 up to 1D stage of development and the PVD neurons were repeatedly imaged every 8-12 h using 60x, 1.3 NA oil objective to count the number of SP, TP, and QP branches in three dimensions. The extent of branching increased during development as shown in Figure 3B. The numbers of SP and QP at different stages of the worm development increased with age (Figure 4A) as has been seen in prior studies37. L2 animals showed 10 ± 3.8 (n = 9) SPs but no QP, when measured using the device (Figure 4A). We placed C. elegans in a drop of 3 mM levamisole, an anesthetic that causes contraction of body wall muscle and paralysis, to minimize adverse effects on organelle transport while reducing animal movements and causing sufficient paralysis, required for high-resolution imaging3,4. The SP values (4 ± 1.6, n = 25, p = 0.15) were not significantly different when measured from similar staged animals grown on NGM plates and imaged using 3 mM levamisole on an agar slide (Figure 4B). The data suggest that the device immobilization enables high-resolution imaging of complex neuronal architectures and does not affect their development. L3 stage animals have a small number of QP that increases in number with development. 1D adults showed 67 ± 2.8 QPs (n = 5) in animals grown in the device. Previous studies have shown the formation as well as a retraction of PVD processes during development known as self-avoidance, where they reduce their branch numbers38. The reduction in SP at 51 h (1D) after hatching could be the result of such retractions. Animals growing on NGM plates and imaged using 3 mM levamisole showed a similar branching number trend for comparable stages of development (Figure 4A,B). Further, the distance between the two PVC cell bodies, one present in the tail and the second present near the vulva, increases as they move apart in animals grown in the device or those grown on NGM plates (Figure 4C,D). Throughout the imaging process, the animals remained healthy and could lay eggs even after the animals were immobilized under the membrane repeatedly over this long period.
Using this device, we were able to immobilize the animal in the same orientation and image the identical neuron and its architecture with high resolution even with faint GFP expression. To demonstrate the utility of our growth and immobilization technology, the development of the TRNs from L3 to adult was imaged using jsIs609 (mec-7p::MLS::GFP) animals. The posterior TRNs are present at the lateral side of the body and are named PLMR and PLML which corresponds to the right and left sides of the animal. We imaged the same animals growing in the microfluidic chip and immobilized them at 12-14 h intervals. The montage represents the entire PLMR neuronal process at successive time points highlighting both mitochondria and the neuronal process (Figure 5A). The total neuronal process length was calculated and found to increase with a slope of 10.4 µm per h (Figure 5B). This rate of increase in the neuronal process length agrees with a previous report of ~10 µm/h18,31,39,40,41. We also observed the addition of new mitochondria in the growing process and the change in the synaptic branch point position with respect to the cell body as reported earlier23.
Figure 1: A simple microfluidic chip for long-term growth and imaging of C. elegans. (A) The flow layer consists of a 10 mm long microfluidic channel (pattern 1) in a thin PDMS layer. (B) The trapping layer consists of a 2 mm thick immobilization channel and a pair of thin isolation channels in the bulk PDMS layer. (C) The two PDMS layers are bonded together along with a thin cover glass to make the device. (D, E) The trapping and isolation control channels are filled with deionized (DI) water, under compressed nitrogen (N2) gas. (F) Age synchronized C. elegans are picked from NGM plates and loaded inside the flow layer of the device. (G) Food is provided using two micropipette tips at the inlet and outlet reservoirs. (H) The animals are free to move within the two isolation channels and are trapped under the immobilization membrane. (I) Image of the growth and imaging device with food supply through two micropipette tips. (J, K) Images of a freely moving and immobilized C. elegans inside the microfluidic channel. Scale bar is 1 mm (I) and 200 µm (J, K). (L) Schematic of the device cross-section showing the heights of the flow channel (FC), PDMS membrane (M), and control channel (CC). Please click here to view a larger version of this figure.
Figure 2: Tracking the expression of a vulval marker in C. elegans developing inside the device. (A) Schematic of vulval cells expressing ZMP-1::GFP in PS3239 animals during development. The GFP signal appears in the anchor cell at the L3 stage, in the vulD and vulE cells at the L4 stage, and in vulA cells at the adult 1D stage. (B) Images of PS3239 animal growing inside the microfluidic chip and immobilized every 8-10 h to capture fluorescence images of ZMP-1::GFP expression during vulval development from L3, L4, and 1 day (1D) adult. Abbreviations: AC = anchor cell, A = vulA, D = vulD, E = vulE. Scale bar is 10 µm. Please click here to view a larger version of this figure.
Figure 3: Imaging of individual wdIs51 C. elegans to track PVD neuronal development. (A) Schematic of PVD neuron growth and dendritic arborization from L2 to L4 stage. Green circle shows PVD cell body (CB, yellow arrow). The dendrites show the emergence of primary processes (PP) at L2 from both the anterior and posterior side of the CB, secondary processes (SP) during late L2, tertiary processes (TP) in L3, and quaternary processes (QP) during L4 stages. (B) Images of PVD neurons from animals growing inside a microfluidic device. (C) Images of PVD neurons from animals grown on an NGM plate and immobilized with 3 mM levamisole (Lev). Scale bar is 10 µm. Please click here to view a larger version of this figure.
Figure 4: Comparison of PVD development in device grown animals and animals grown on NGM plates. (A) The average number of secondary processes (SP) and quaternary processes (QP) from the same animals growing inside the microfluidic device. The values are calculated at 16 h (L2), 24 h (L3), 36 h (early L4), 42 h (late L4), and 51 h (1D adult). (B) The average number of SP and QP from a different batch of animals growing on NGM plates and anesthetized with 3 mM levamisole. (C) Average separation between the two PVD neuronal cell bodies from the same animal growing in the microfluidic device. (D) Average separation from different sets of animals growing on NGM plates and anesthetized with 3 mM levamisole. Data represented as mean ± SEM (n ≥ 10 for 3 mM levamisole and n ≥ 6 for device immobilized animals). The statistical significances are calculated by one-way ANOVA and denoted as p-value < 0.05 (*), p-value < 0.005 (**), and p-value > 0.05 (ns). Please click here to view a larger version of this figure.
Figure 5: High-resolution imaging of touch receptor neurons (TRNs) from animals growing inside the microfluidic device. (A) Montage of neuronal processes from a single animal imaged at different times. The cell body (CB, arrow), the synaptic branch point (BP, arrowhead), and the neuron tip (Tip, up arrow) are labeled. The neuronal process and position of each mitochondrion are manually traced using Fiji. Scale bar is 10 µm. (B) Average neuronal process length at different imaging time points. Data represented as mean ± SEM (n = 8). The statistical significance is calculated using one-way ANOVA and denoted as p-value < 0.005 (**) and p-value > 0.05 (ns). A regression equation is fit with a slope of 10.4 µm neuronal process elongation per hour, R2=0.9425. Please click here to view a larger version of this figure.
Supplementary Figure 1: Screenshots of microscope settings for fluorescence imaging of C. elegans. (A) A panel of the image analysis software shows the highlighted sections for setting up the scan speed, filter set, and objective for imaging. (B) The software is then used to choose the 488 nm laser with 7% of the full laser power. (C) The image analysis software can take a single image frame or a series of time-lapse images and store them for future analysis. Please click here to download this File.
Supplementary Figure 2: Screenshots of FIJI software showing the analysis steps for counting PVD branches in wdIs51 animals. (A) Shows PVD image opened in Fiji ImageJ software and link to the cell counter plugin. (B) Cell counter window to initialize the counter for an image. (C) Selection of counters to mark branches that are included and to avoid multiple counting. (D) Results window showing the total number of branches under each category. Please click here to download this File.
Supplementary Figure 3: Schematic of the growth and imaging device. Two micropipette tips, filled with different volumes of bacteria food solution (yellow), are inserted in the inlet and outlet punches of the flow channel at the bottom layer. The difference in the heights of food solutions in the tips causes the solution to flow inside the channel which in turn supplies bacteria to the animal for its growth. The trapping and isolation channels (blue) are filled with distilled water and compressed using a nitrogen gas (set at 14 psi) supply. The isolation channel is always connected to 14 psi. The pressure on the trapping membrane is switched between 0 (animal is free to move and feed) and 14 psi (animal is immobilized for high-resolution imaging) using a three-way connector. Please click here to download this File.
In this paper, a protocol for fabrication and use of a simple microfluidic device for growing C. elegans with constant food supply and high-resolution imaging of a single animal during its development has been described. This fabrication process is simple and can be done in a non-sterile environment. A dust-free environment is critical during fabrication steps. The presence of dust particles would lead to improper contact between the two bonding surfaces, resulting in poor bonding and leakage of the device during high-pressure application while C. elegans are immobilized. Out of all the devices fabricated, 95% of the devices are suitable for experiments. A small number of devices fail (5%) due to inappropriate bonding during fabrication or use. Failure in bonding can be reduced by carrying out the fabrication process inside a dust-free (> 1,000-grade clean room) available at several academic institutions around the globe. To clean bacterial food from the flow channel and enable device reuse, flush the device by flowing alcohol through it after every experiment.
The protocol demonstrates a lithography fabrication process with a simple design that can easily be modified for different developmental stages and sizes of C. elegans. Further, this device design can be adapted for other model organisms like zebrafish and Drosophila, by increasing the dimensions of the flow and trapping layers3 to observe a variety of developmental and sub-cellular processes. In this protocol, two different heights of the flow layer channel - 40 µm and 80 µm - have been used for complete immobilization and tracking of cellular and sub-cellular features in different developmental stages of C. elegans as appropriate for their body sizes. For example, the dendritic arborization in the PVD sensory neurons begins in the early L2 stages and continues up to the L4 stage. This was imaged in a device with a 40 µm high flow layer. As the PVD neurons form dendrites that innervate body wall muscles it requires optical sectioning to quantify the processes in 3D. The quantitative analysis of dendritic arbors requires complete immobilization of the animals within the device that matches the body diameter. However, for monitoring vulval development, 80 µm height of flow layer was used to track the vulval lineages. For TRNs length and mitochondria imaging, we imaged L2 to L4 stages using devices with 40 µm flow layer thickness. Using a device with the wrong height will cause difficulty in immobilization. For example, small animals (L2 or early L3 stages) inside an 80 µm flow layer height device will permit animal flipping inside the device and show incomplete immobilization even after the trapping membrane is under 14 psi pressure. On the other hand, animals with large body diameters (beyond late L4) do not easily enter inside a device with 40 µm flow layer height. Trapping large animals in small devices under high pressure can damage their body. The application of the current device with a 40 µm flow layer is limited and not suitable for high-resolution imaging of very young early larval stage animals (such as L1 stage animals younger than 10 h after hatching) as they are not completely immobilized under the membrane. Due to small body sizes, the small size larval animals keep moving and occasionally escape the trap when they are excited with laser light. For L1 stage animals, one can perform low-resolution bright field or fluorescence imaging using 4x or 10x objectives and short camera exposure times. As an alternative approach, a new flow layer device with < 20 µm height will be necessary to completely immobilize young larval stage C. elegans.
Tracking developmental phenomena in the same animal over a long timescale using high-resolution fluorescence imaging can result in increased autofluorescence. Every time-lapse imaging assay requires optimization of excitation intensity, exposure time, imaging time interval, animal recovery time, and food quality for physiologically relevant developmental information from C. elegans studies. We have selected > 3 h time interval for developmental studies when using animals from L4 stages onwards. The device is capable of acquiring more frequent images (every 5-10 min for a few hours) or multiple frames per second (for < 1 h) to study other dynamic processes such as organelle transport and distribution in C. elegans neurons3,23. Trapping and imaging of early developmental stages require more careful observation as repeated trapping with short time intervals without complete recovery can affect their health. During experiments, it was found that animals that required high pressure to move them into the flow layer show poor health and more autofluorescence over time. Animals with high body fluorescence, known to be associated with a high-stress level, were avoided for imaging studies. To maintain good physiology, animals were trapped in a straight position adjacent to the channel wall. Immobilizing the animals with their body across the channel width was avoided since animals become sick after their body is squeezed under 14 psi pressure in this position. To maintain a good signal-to-noise ratio during repeated imaging with oil objectives, the coverslip should be cleaned properly after every time point. Following special care during the fabrication process of the device and its application, the same devices could be reused over several imaging sessions.
This device could be useful for studies using mutants that could be sensitive to anesthetics. The approach presented can eliminate adverse anesthetic effects on growth and physiology to facilitate observation of morphological, functional, and behavioral defects in animals over a long period. The device is easy to fabricate and can be set up in any laboratory to address long-term developmental/cell biological questions in C. elegans that require intermittent high-resolution imaging.
S.M. and S.P.K. are authors of a pending patent on the microfluidic growth and imaging device (Patent application number 640/CHE/2011).
We thank CIFF imaging facility, NCBS for use of the confocal microscopes supported by the DST - Centre for Nanotechnology (No. SR/55/NM-36-2005). We thank research funding from DBT (SPK), CSIR-UGC (JD), DST (SM), DBT (SM), spinning disc supported by DAE-PRISM 12-R&D-IMS-5.02.0202 (SPK and Gautam Menon), and HHMI-IECS grant number 55007425 (SPK). HB101, PS3239, and wdIs51 strains were provided by the Caenorhabditis Genetics Center (CGC), which is funded by the NIH Office of Research Infrastructure Programs (P40 OD010440). S.P.K. made jsIs609 in Mike Nonet's Laboratory.
|18 G needles||Sigma-Aldrich, Bangalore, India||Gauge 18|
|3-way stopcock||Cole-Parmer||WW-30600-02||Masterflex fitting with luer lock|
|CCD camera||Andor Technology||EMCCD C9100-13no|
|Circuit board film||Fine Line Imaging, Colorado, USA||The designs are printed with 65,024 dots per inch (DPI)|
|Convection Oven||Meta-Lab Scientific Industries, India||MSI-5|
|Coverslips||Blue stat microscopic cover glass||22mm x 10Gms|
|Harris uni-core puncher 1mm||Qiagen||Z708801|
|Hexamethyldisilazane||Sigma-Aldrich, Bangalore, India||440191|
|Hot plate||IKA||RCT B S 22|
|Laser Scanning Microscope||ZEISS||LSM 5 LIVE|
|Micropipette tips||Tarsons||0.5-10 µL micropipette tips are used for food supply|
|Nitrogen gas||Local Supplier||Commercial nitrogen gas||Cylinder volume of 7 cubic meter|
|PDMS (Curing solution)||Dow Corning Corporation, MI, USA||Sylgard curing solution||curing agent|
|Petri plates||Praveen Scientific Corporation|
|Plasma cleaner||Harrick Plasma, NY, USA||PDC-32G|
|Razor and blades||Lister surgical Blade|
|Silicon Elastomer (Base)||Dow Corning Corporation, MI, USA||Sylgard 184 base||elastomer base|
|Silicon tubes||Fisher Scientific||Plastic tubes with the inner diameter 1.59 mm and the outer diameter 3.18 mm|
|Silicon wafer||University Wafer, MA, USA|| orientation, 4-inch diameter||Small pieces (2 mm × 2 mm) were cut from 100 mm diameter wafer|
|Spin Coater||SPS-Europe B.V., The Netherlands||SPIN 150|
|Spinning Disk microscope||Perkin Elmer ultra-view VOX system||CSU-X1-A3 N||The system was equipped with four (405/488/561/640 nm) lasers and controlled with the Volocity software package.|
|SU8 developer||Microchem, MA, USA||SU8 Developer|
|Trichloro (1H, 1H, 2H, 2H-perfluorooctyl) silane||Sigma-Aldrich, Bangalore, India||448931||Trichloro (1H, 1H, 2H, 2H-perfluorooctyl) silane vapor is toxic|
|UV lamp||Oriel Instruments, Bangalore, India||200 Watt and collimated UV light source|
|Volocity software||Perkin-Elmer||Image analysis|
- Doitsidou, M., Poole, R. J., Sarin, S., Bigelow, H., Hobert, O. C. elegans Mutant Identification with a One-Step Whole-Genome-Sequencing and SNP Mapping Strategy. PloS One. 5, (11), 15435 (2010).
- Hobert, O. Neurogenesis in the nematode Caenorhabditis elegans. WormBook. The C. elegans. Research Community, WormBook. (2010).
- Mondal, S., Ahlawat, S., Rau, K., Venkataraman, V., Koushika, S. P. Imaging in vivo neuronal transport in genetic model organisms using microfluidic devices. Traffic. 12, (4), Copenhagen, Denmark. 372-385 (2011).
- Steele, L. M., Sedensky, M. M. Approaches to Anesthetic Mechanisms: The C. elegans Model. Methods in Enzymology. 602, 133-151 (2018).
- San-Miguel, A., Lu, H. Microfluidics as a tool for C. elegans research. WormBook. The C. elegans. Research Community, WormBook. (2013).
- Cáceres, I. C., Valmas, N., Hilliard, M. A., Lu, H. Laterally orienting C. elegans using geometry at microscale for high-throughput visual screens in neurodegeneration and neuronal development studies. PloS One. 7, (4), 35037 (2012).
- Ai, X., Zhuo, W., Liang, Q., McGrath, P. T., Lu, H. A high-throughput device for size based separation of C. elegans developmental stages. Lab on a Chip. 14, (10), 1746-1752 (2014).
- Chung, K., Crane, M. M., Lu, H. Automated on-chip rapid microscopy, phenotyping and sorting of C. elegans. Nature Methods. 5, (7), 637-643 (2008).
- Desta, I. T., et al. Detecting and Trapping of a Single C. elegans Worm in a Microfluidic Chip for Automated Microplate Dispensing. SLAS Technology. 22, (4), 431-436 (2017).
- Ben-Yakar, A. High-Content and High-Throughput In Vivo Drug Screening Platforms Using Microfluidics. Assay and Drug Development Technologies. 17, (1), 8-13 (2019).
- Mondal, S., et al. Large-scale microfluidics providing high-resolution and high-throughput screening of Caenorhabditis elegans poly-glutamine aggregation model. Nature Communications. 7, 13023 (2016).
- Fehlauer, H., et al. Using a Microfluidics Device for Mechanical Stimulation and High Resolution Imaging of C. Elegant. Journal of Visualized Experiments: JoVE. (132), e56530 (2018).
- Saberi-Bosari, S., Huayta, J., San-Miguel, A. A microfluidic platform for lifelong high-resolution and high throughput imaging of subtle aging phenotypes in C. elegans. Lab on a Chip. 18, (20), 3090-3100 (2018).
- Cornaglia, M., et al. An automated microfluidic platform for C. elegans embryo arraying, phenotyping, and long-term live imaging. Scientific Reports. 5, 10192 (2015).
- Suzuki, M., et al. Development of ultra-thin chips for immobilization of Caenorhabditis elegans in microfluidic channels during irradiation and selection of buffer solution to prevent dehydration. Journal of Neuroscience Methods. 306, 32-37 (2018).
- Hulme, S. E., et al. Lifespan-on-a-chip: microfluidic chambers for performing lifelong observation of C. elegans. Lab on a Chip. 10, (5), 589-597 (2010).
- Chronis, N., Zimmer, M., Bargmann, C. I. Microfluidics for in vivo imaging of neuronal and behavioral activity in Caenorhabditis elegans. Nature Methods. 4, (9), 727-731 (2007).
- Lagoy, R. C., Albrecht, D. R. Microfluidic Devices for Behavioral Analysis, Microscopy, and Neuronal Imaging in Caenorhabditis Elegans. Methods in Molecular Biology. 1327, Clifton, N.J. 159-179 (2015).
- Levine, E., Lee, K. S. Microfluidic approaches for Caenorhabditis elegans research. Animal Cells and Systems. 24, (6), 311-320 (2020).
- Rahman, M., et al. NemaLife chip: a micropillar-based microfluidic culture device optimized for aging studies in crawling C. elegans. Scientific Reports. 10, (1), 16190 (2020).
- Allen, P. B., et al. Single-synapse ablation and long-term imaging in live C. elegans. Journal of Neuroscience Methods. 173, (1), 20-26 (2008).
- Guo, S. X., et al. Femtosecond laser nanoaxotomy lab-on-a-chip for in vivo nerve regeneration studies. Nature Methods. 5, (6), 531-533 (2008).
- Mondal, S., et al. Tracking Mitochondrial Density and Positioning along a Growing Neuronal Process in Individual C. elegans Neuron Using a Long-Term Growth and Imaging Microfluidic Device. eNeuro. 8, (4), (2021).
- Keil, W., Kutscher, L. M., Shaham, S., Siggia, E. D. Long-Term High-Resolution Imaging of Developing C. elegans Larvae with Microfluidics. Developmental Cell. 40, (2), 202-214 (2017).
- Gritti, N., Kienle, S., Filina, O., Van Zon, J. S. Long-term time-lapse microscopy of C. Elegans post-embryonic development. Nature Communications. 7, 12500 (2016).
- Kim, S., et al. A biostable, anti-fouling zwitterionic polyurethane-urea based on PDMS for use in blood-contacting medical devices. Journal of materials chemistry B. 8, (36), 8305-8314 (2020).
- Peterson, S. L., McDonald, A., Gourley, P. L., Sasaki, D. Y. Poly(dimethylsiloxane) thin films as biocompatible coatings for microfluidic devices: cell culture and flow studies with glial cells. Journal of Biomedical Materials ResearchPart A. 72, (1), 10-18 (2005).
- Folch, A., Toner, M. Cellular micropatterns on biocompatible materials. Biotechnology Progress. 14, (3), 388-392 (1998).
- Leclerc, E., Sakai, Y., Fujii, T. Microfluidic PDMS (polydimethylsiloxane) bioreactor for large-scale culture of hepatocytes. Biotechnology Progress. 20, (3), 750-755 (2004).
- Mehta, G., et al. Quantitative measurement and control of oxygen levels in microfluidic poly(dimethylsiloxane) bioreactors during cell culture. Biomedical Microdevices. 9, (2), 123-134 (2007).
- Palchesko, R. N., Zhang, L., Sun, Y., Feinberg, A. W. Development of polydimethylsiloxane substrates with tunable elastic modulus to study cell mechanobiology in muscle and nerve. PloS One. 7, (12), 51499 (2012).
- Inoue, T., et al. Gene expression markers for Caenorhabditis elegans vulval cells. Mechanisms of Development. 119, Suppl 1 203-209 (2002).
- Fatouros, C., et al. Inhibition of Tau aggregation in a novel caenorhabditis elegans model of tauopathy mitigates proteotoxicity. Human Molecular Genetics. 21, (16), 3587-3603 (2012).
- Smith, C. J., et al. Time-lapse imaging and cell-speci fi c expression pro fi ling reveal dynamic branching and molecular determinants of a multi-dendritic nociceptor in C. elegans. Developmental Biology. 345, (1), 18-33 (2010).
- Brenner, S. The genetics of Caenorhabditis elegans. Genetics. 77, (1), 71-94 (1974).
- Oren-Suissa, M., Hall, D. H., Treinin, M., Shemer, G., Podbilewicz, B. The fusogen EFF-I controls sculpting of mechanosensory dendrites. Science. 328, (5983), New York, N.Y. 1285-1288 (2010).
- Smith, C. J., et al. Sensory neuron fates are distinguished by a transcriptional switch that regulates dendrite branch stabilization. Neuron. 79, (2), 266-280 (2013).
- Shrestha, B. R., Grueber, W. B. Neuronal morphogenesis: worms get an EFF in dendritic arborization. Current Biology: CB. 20, (16), 673-675 (2010).
- Rao, G. N., Kulkarni, S. S., Koushika, S. P., Rau, K. R. In vivo nanosecond laser axotomy: cavitation dynamics and vesicle transport. Optics Express. 16, (13), 9884-9894 (2008).
- Sure, G. R., et al. UNC-16/JIP3 and UNC-76/FEZ1 limit the density of mitochondria in C. elegans neurons by maintaining the balance of anterograde and retrograde mitochondrial transport. Scientific Reports. 8, (1), 8938 (2018).
- Awasthi, A., et al. Regulated distribution of mitochondria in touch receptor neurons of C. elegans influences touch response. bioRxiv. (2020).