Real-time In Vivo Recording of Arabidopsis Calcium Signals During Insect Feeding Using a Fluorescent Biosensor

Calcium ions are predicted to be key signaling entities during biotic interactions, with calcium signaling forming an established part of the plant defense response to microbial elicitors and to wounding caused by chewing insects, eliciting systemic calcium signals in plants. However, the role of calcium in vivo during biotic stress is still unclear. This protocol describes the use of a genetically-encoded calcium sensor to detect calcium signals in plants during feeding by a hemipteran pest. Hemipterans such as aphids pierce a small number of cells with specialized, elongated sucking mouthparts, making them the ideal tool to study calcium dynamics when a plant is faced with a biotic stress, which is distinct from a wounding response. In addition, fluorescent biosensors are revolutionizing the measurement of signaling molecules in vivo in both animals and plants. Expressing a GFP-based calcium biosensor, GCaMP3, in the model plant Arabidopsis thaliana allows for the real-time imaging of plant calcium dynamics during insect feeding, with a high spatial and temporal resolution. A repeatable and robust assay has been developed using the fluorescence microscopy of detached GCaMP3 leaves, allowing for the continuous measurement of cytosolic calcium dynamics before, during, and after insect feeding. This reveals a highly-localized rapid calcium elevation around the aphid feeding site that occurs within a few minutes. The protocol can be adapted to other biotic stresses, such as additional insect species, while the use of Arabidopsis thaliana allows for the rapid generation of mutants to facilitate the molecular analysis of the phenomenon.


Introduction
Calcium (Ca 2+ ) is one of the most ubiquitous signaling elements in plants. A  ] cyt ) is decoded by a complex network of downstream components and is involved in the response to both abiotic and biotic stresses 1,2 . A rise in [Ca 2+ ] cyt is one of the first responses to microbial elicitors, forming a common part of the plant defense response 3,4,5 . Rises in [Ca 2+ ] cyt have also been observed in response to wounding caused by chewing insects, such as lepidopterans 6,7 . However, the potential role of plant Ca 2+ signals in response to live biotic threats that cause damage to only a few cells has not been explored. The green peach aphid Myzus persicae is a hemipteran insect that represents a significant threat to world agriculture 8,9 , and Ca 2+ efflux from the extracellular space has been observed in leaves infested with M. persicae 10 .This protocol outlines a robust and repeatable method for measuring plant Ca 2+ signals while M. persicae feed from leaves using a fluorescent Ca 2+ biosensor, with both aphids and GCaMP3 offering novel tools with which to dissect the role of Ca 2+ during biotic interactions.
Ca 2+ -selective microelectrodes were formerly used to measure [Ca 2+ ] in plants 11,12 . More recently, bioluminescent and fluorescent approaches have become standardized. These biosensors bind Ca 2+ and emit light, allowing for un-paralleled opportunities to study Ca 2+ dynamics in both cells and whole tissues. Ca 2+ biosensors can be injected as dyes or stably produced upon the introduction of the biosensor coding sequence into the genome of the organism via transformation (i.e., genetically encoded biosensors). The latter offers the major advantages of being easily expressed in live tissue and capable of subcellular localization 13 . The aequorin protein, isolated from Aequorea victoria (jellyfish) was the first genetically encoded Ca 2+ biosensor deployed in plants 14 . As a bioluminescent protein, aequorin does not require excitation by external light, which avoids chromophore bleaching and autofluorescence 15 . Aequorin has been successfully used to measure [Ca 2+ ] fluxes in response to various stimuli, including temperature 16 , pathogens 17,18,19 , salt stress 20,21 , and wounding 7 . However, it is disadvantaged by the relatively low signal intensity, making the detection of [Ca ] through the calculation of the ratio of the fluorescence signals from the two fluorophores 22 . FRET Cameleons are superior to aequorin and non-ratiometric florescent dyes, as they are less affected by the expression level of the protein 23 and often have a greater fluorescent yield, allowing for cellular and subcellular imaging 23 . For example, FRET Cameleons have been recently used to identify long-distance Ca 2+ signals in plants and to resolve these to the cellular level 24,25,26 .
A recent breakthrough with fluorescent GFP-based Ca 2+ biosensors has been the development of highly sensitive single-fluorophore (single-FP) biosensors. Single-FP biosensors consist of a single circularly permutated GFP linked to a calmodulin and M13 peptide, with Ca 2+ binding to calmodulin, resulting in a water-mediated reaction between calmodulin and GFP so as to protonate GFP and increase fluorescent yield 27,28,29 . Single-FP sensors offer several advantages over FRET Cameleons, including simpler experimental design and a potentially higher temporal resolution of imaging 30 . Although single-FP sensors cannot quantify absolute [Ca 2+ ] as simply as FRET sensors, they are superior for the analysis of the temporal and spatial dynamics of Ca 2+ signals 5,23 . GCaMPs are one of the best-established single-FP sensors 28 and have undergone several revisions to enhance their fluorescent yield, dynamic range, Ca 2+ affinity, and signal-to-noise ratios 31,32,33,34 . The GCaMPs have been successfully used in animal systems, such as zebrafish motor neurons 35 and fruit fly neuromuscular junctions 34 . Random mutagenesis of GCaMP3 has resulted in additional classes of single-FP sensors, including the ultrasensitive GCaMP6 36 and the GECOs 29 . The GECOs were recently used in Arabidopsis thaliana (henceforth referred to as Arabidopsis) to measure Ca 2+ fluxes in response to ATP, chitin, and the bacterial elicitor flg22. This study also demonstrated that the R-GECO biosensor outperformed the FRET Cameleon YC3.6 in terms of maximal signal change and signal-to-noise ratio 5 .
Because of the ease of use, high fluorescent yield, and high temporal resolution that can be achieved with GCaMP biosensors, GCaMP3 was genetically encoded in Arabidopsis under the Cauliflower mosaic virus 35S promoter. The genetic tools available for Arabidopsis research allow for the detailed molecular analysis of the Ca 2+ signals measured by GCaMP3. In addition, the GCaMP3 biosensor can be visualized under a fluorescence microscope rather than a costlier confocal system. This protocol allows for whole-tissue imaging, essential when conducting experiments with live biotic stresses. The experiment is designed such that detached leaves from 35S::GCaMP3 plants are floated in water, to prevent insect escape and to restrict feeding to a specific tissue. The method outlined in this paper therefore allows for the analysis of leaf Ca 2+ dynamics during feeding by M. persicae, resulting in the characterization of a novel plant signaling response. This method can also be adapted to work with other biotic stresses, such as additional insect species and microbial pathogens, and with other plant tissues, such as roots.

Plant Preparation (Days 1 -4)
1. On day 1, sterilize 35S::GCaMP seeds using three 75% ethanol washes, 1 min per wash, and plate them on 100 mm 2 square plastic plates with ¼-strength Murashige and Skoog (MS) medium (recipe: 1.1 g of Murashige and Skoog medium, 7.5 g of sucrose, 10 g of Formedium agar, and 1 L of de-ionized water) 37 . 2. Stratify the seedlings in the dark for three days at 8 °C to obtain synchronous germination. 3. On day 4 of the experiment, move the GCaMP seedlings to a controlled environment room (CER) at 23 °C, with a 16 h light and 8 h dark photoperiod.

Fluorescence Microscopy (Days 20 -22)
1. On day 20 of the experiment, remove the 96-well plate from the aluminum foil and transfer it to a fluorescence stereomicroscope. Configure the stereo fluorescence microscope to excite GFP with a 450-490 nm light and to capture the fluorescent emission between 500 and 550 nm. magnification and a focus of -127.833 mm was used. 5. Transfer one aphid to a detached leaf under the microscope using a moist paint brush. Leave an adjacent leaf untreated as a control, but lightly touch it with the paintbrush to mimic the touching that occurs during aphid transfer. Remember to place the plastic wrap back on top of the 96-well plate during microscopy to prevent the insect from escaping. 6. Begin the fluorescence recording of the leaves in pairs (1 aphid-treated and 1 untreated) by clicking "start experiment" in the built-in microscope software (Figure 2). Record measurements for 50 min. NOTE: For the current protocol, a time interval of 5 s between measurements was used. 7. After 50 min, stop recording by clicking on "stop experiment" and remove the aphid from the leaf. Save the fluorescence measurements as image files (e.g., tagged image file format, TIFF). 8. Repeat the experiment with further pairs of leaves; imaging can be extended to image 2 pairs of leaves at once, allowing for the simultaneous imaging of 2 genotypes.

Data Collection
1. Import the image files into Fiji (Image J) and convert them to 32 bits by clicking on "Image" > "Type" > "32-bit;" this allows for the conversion of the images into heatmap videos (see step 7). 2. Discard the samples in which the aphids do not settle in one location for more than 5 min by viewing the insect movement under the microscope. NOTE: This is often the majority of samples, and up to 100 treatments may be required to find 30 samples exhibiting successful settling. 3. Set the measurement scale to pixels (or convert to mm, if known) and the time frame to the same time interval as used during the microscopy by clicking on "Image" > "Properties." 4. Using the cursor, place a region of interest (ROI) around the area of tissue for GFP analysis.
NOTE: In the current protocol, a circular ROI 50 pixels (0.65 mm) in diameter was used at 3 locations of interest: the aphid feeding site (Fs), a systemic region on the midrib of the leaf (Sm), and a systemic region adjacent to the midrib ("lateral tissue," Sl) (Figure 2). 1. Create the ROI by drawing an oval (use the "oval tool"), and edit the size using "Edit" > "Selection" > "Specify." See  6. Use the Time Series Analyzer plugin by clicking "Plugins" > "Time Series Analyzer" to analyze the raw fluorescence values (F) in the ROI over time. Add the ROI of interest ("Add [t]"). Making sure that the ROI is selected, select "get average;" this will display a table of F values for each frame in that ROI. 7. Copy this data into a spreadsheet. 8. Calculate the area of the feeding site signal by first selecting the region using the "freehand selection tool." Outline the maximal GFP signal and then calculate the area of this shape by clicking "Analyze" > "Measure." 9. Calculate the speed of the feeding site signal using the MTrackJ plugin by clicking "Plugins" > "Tracking" > "MTrackJ." 1. Click on the "Add" button and then click the cursor at the center of the signal when it is first visible. Click again on the edge of the signal at its point of furthest spread. Click "measure" to calculate the speed of the signal.

Time-course Video Creation
Representative Results Figure 3 and Figure 4 are representative results from an experiment comparing an aphid-treated leaf with an untreated control. A highly localized increase in GFP fluorescence can be seen around the feeding site within a few minutes in the majority of samples, whilst the Ca 2+ dynamics in the untreated control leaf stay relatively stable (Figure 3A and 3B). It is also possible to observe secondary increases in GFP fluorescence after the initial peak in some experiments ( Figure 3B). In up to 50% of treated leaves, the aphids do not settle and the samples should be discarded. Of the samples in which settling occurs, 27% of samples do not exhibit clear increases in GFP fluorescence around the feeding site ( Figure 3C and Table 1 ] cyt elevations should be detected systemically within the leaf upon aphid treatment, either in the systemic midrib (Figure 4A) or the systemic lateral tissue regions (Figure 4B). A representative sample of [Ca 2+ ] cyt dynamics over time is shown in Video 1. It is also possible to analyze aphid settling behavior by tracking the number and duration of individual settling events under the microscope. Representative results for these behaviors are shown in Table 1, showing that the aphids take around 10 min before settling, and when they do settle successfully, this lasts for 20 min on average. Therefore, the insects are settled in a single location for the entirety of the [Ca 2+ ] cyt elevation.

Discussion
The method described in this paper allows for the real-time analysis of plant Ca 2+ signaling during a biotic stress such as insect feeding. It demonstrates that one of the first plant responses to such threats is a localized [Ca 2+ ] cyt elevation around the feeding site of the insect. Through the use of mutants, this method will allow for the the molecular and physiological characterization of such signals, which was not previously possible. A critical step in this protocol is to ensure that the detached leaves are not excessively disturbed during the detachment process (step 3.2) or when transferring insects to the leaves (step 4.5). Given that the current protocol provides a relative measurement of [Ca 2+ ] cyt rather than an absolute concentration, it is vital that the microscope settings are kept constant throughout the experiment. There is also the potential for human bias during the selection of ROIs and the analysis of the data, and as such, it is recommended that the experiments are conducted double-blind.
There are several significant advantages of measuring [Ca 2+ ] cyt during biotic stress with this protocol. First, the use of a single fluorophore with a high fluorescent yield allows the imaging to be conducted on a stereomicroscope, which is less costly than using a confocal microscope. The use of a single fluorophore also makes data collection and analysis simple, as there is just one measurement to record. In addition, the use of a stereomicroscope allows for the imaging of entire leaves, which is essential given that many biotic interactions, including plant-aphid interactions, occur on a large spatial scale. The high temporal resolution of image capture possible with GCaMP3, based on the rapid disassociation of Ca 2+ from the sensor after binding 23,30 and the high florescent yield, allows for measurements to be taken up to every 5 s. Furthermore, the leaf assay prevents the escape of the insect, a key limiting step to conducting such experiments on whole plants (in preparation) . The detached leaves also ensure that the insect feeds from a pre-defined location, allowing for the analysis of Ca 2+ dynamics before, during, and after feeding. This protocol also ensures that leaves of similar developmental stages are used for analysis.
The main disadvantage of this protocol originates from the use of a non-ratiometric biosensor. With single-FP biosensors, variation in GFP emission may result from experimental variables other than [ Ca   2+ ] cyt , such as changes in cellular pH, motion, or the expression level of the biosensor. These issues are not encountered with FRET Cameleons during FRET, as the transfer of energy from CFP to YFP only occurs upon Ca 2+ binding. Other conditions that alter the fluorescent properties of the individual fluorophores are unlikely to mimic the opposing changes in intensity of CFP and YFP, and the ratiometric calculation that is used inherently normalizes the measurements for many of these other optical artifacts 23,30 . . Importantly, single-FP biosensors typically display a greater fluorescent yield and greater dynamic range (i.e., an increase in florescence upon Ca 2+ binding) than FRET Cameleons 23 , which makes GCaMP more suited to tissue-level imaging, while FRET Cameleons are a useful tool for cellular imaging with a confocal microscope 5,25 .
During the execution of this protocol, it is possible that some issues will arise that require troubleshooting. For example, it is recommended that samples in which the control (untreated) leaf displays large [Ca 2+ ] cyt elevations are discarded (step 6.3). Such transients are most likely the result of stress induced by the microscopy. Indeed, blue light is known to elicit Ca 2+ signals 38,39,40,41 , and the high-intensity light might also result in temperature and osmotic stresses, both of which also elicit [Ca 2+ ] cyt elevations 21,25,42 . Issues may also be encountered with insect settling. With M. persicae, the insects do not settle on the leaves in several samples. This could be a result of wound-elicited defense in the detached leaves 46,47 , or the disturbance of the insects by the blue light. Indeed, vision in M. persicae is governed by three photoreceptors, including one with a peak sensitivity of 490 nm 48 . Reducing the microscopy exposure and handling the aphids with care might reduce distress and encourage settling.
The protocol outlined in the current paper gives new insights onto the molecular understanding of plant-insect interactions and the plant response to biotic stress. It allows for the visualization of one of the first plant responses to insect feeding and facilitates further investigations through the use of the considerable Arabidopsis genetic resources available. In addition, this protocol allows for the use of live organisms, as opposed to extracts 49 or elicitors 50 . In the future, this technique could be applied to other biotic stresses, such as additional insect species, nematodes, or microbial pathogens, as well as to abiotic stresses. The GCaMP3 microscopy can also be modified to image other plant tissues, alternative ROIs on the leaf, or even whole plants. Furthermore, there is the potential for the biosensor to be genetically encoded in additional plant species. Consequently, the protocol outlined in this paper has the potential to undercover the molecular basis of Ca 2+ signaling in a range of novel biotic interactions between plants and other species.

Disclosures
The authors have no conflicts of interest to declare.