Stomata are microscopic pores surrounded by a pair of guard cells on the surface of leaves and other aerial parts of plants. Under ever-changing environments, regulation of the stomatal aperture is central for plants to control the carbon dioxide uptake required for photosynthesis at the expense of water loss via transpiration. Thus, quantification of the stomatal aperture has been instrumental to understanding plant environmental adaptation. However, quantifying the stomatal aperture is inherently time-consuming and cumbersome as it requires human labor to spot and measure stomatal pores in a leaf image captured by a microscope. To circumvent these limitations, various methods have been developed to facilitate the quantification of stomatal aperture in Arabidopsis thaliana, a model plant extensively used to study stomatal biology ^{1 , 2 , 3 , 4 , 5 , 6} . For instance, a porometer can be used to measure transpiration rate as a metric of stomatal conductance. However, this method does not provide direct information on the stomatal number and aperture that determine stomatal conductance. Some studies have used confocal microscopy techniques highlighting stomatal pores using a fluorescent actin marker, a fluorescent dye, or cell wall autofluorescence ^{1 , 2 , 3 , 4 , 5} . While these approaches facilitate the detection of stomata, the cost of both operating a confocal microscopy facility and preparing microscopy samples can be an obstacle to routine application. In a ground-breaking work by Sai et al., a deep neural network model was developed to automatically measure stomatal aperture from bright-field microscopic images of A. thaliana epidermal peels ⁶ . Yet, this innovation does not exempt researchers from the task of preparing an epidermal peel for microscopic observation. Recently, this obstacle was overcome by developing a portable imaging device that can observe stomata by pinching a leaf of A. thaliana , together with a deep learning-based image analysis pipeline that automatically measures stomatal aperture from leaf images captured by the device ⁷ .

Stomata contribute to plant innate immunity against bacterial pathogens. The key to this immune response is stomatal closure that restricts bacterial entry through the microscopic pore into the leaf interior, where bacterial pathogens proliferate and cause diseases ⁸ . Stomatal closure is induced upon recognition of microbe-associated molecular patterns (MAMPs), immunogenic molecules that are often common to a class of microbes, by plasma membrane-localized pattern recognition receptors (PRRs) ⁹ . A 22 amino acid epitope of bacterial flagellin known as flg22 is a typical MAMP that induces stomatal closure through its recognition by the PRR FLS2 ¹⁰ . As a countermeasure, bacterial pathogens such as Pseudomonas syringae pv. tomato DC3000 ( Pto ) and Xanthomonas campestris pv. vesicatoria have evolved virulence mechanisms to reopen stomata ^{9 , 11 , 12} . These stomatal responses to bacterial pathogens have been conventionally analyzed in assays in which either leaf epidermal peels, leaf discs, or detached leaves are floated on bacterial suspension, and then stomata are observed under a microscope followed by manual measurement of stomatal aperture. However, these assays are cumbersome and may not reflect stomatal responses to natural bacterial invasion that occur in a leaf attached to the plant.

Here, a simple method is presented to investigate stomatal closure and reopening during Pto invasion under the condition that closely mimics the natural plant-bacteria interaction. This method leverages the portable imaging device for direct observation of A. thaliana stomata on a leaf attached to the plant inoculated with Pto , together with the image analysis pipeline for automated measurement of stomatal aperture.

Previous studies used epidermal peels, leaf discs, or detached leaves to investigate stomatal responses to bacterial invasions ^{9 , 11 , 12} . In contrast, the method proposed in this study leverages the portable stomatal imaging device to directly observe stomata on a leaf attached to the plant after spray inoculation of Pto , mimicking natural conditions of bacterial invasion. In addition, because this method does not involve …

We thank all the members of the research project, 'Co-creation of plant adaptive traits via assembly of plant-microbe holobiont', for fruitful discussions. This work was supported by Grant-in-Aid for Transformative Research Areas (21H05151 and 21H05149 to A.M. and 21H05152 to Y.T.) and Grant-in-Aid for Challenging Exploratory Research (22K19178 to A. M.).