Unveiling nutrient flow-mediated stress in plant roots using an on-chip phytofluidic device

Abstract

The initial emergence of the primary root from a germinating seed is a pivotal phase that influences a plant's survival. Abiotic factors such as pH, nutrient availability, and soil composition significantly affect root morphology and architecture. Of particular interest is the impact of nutrient flow on thigmomorphogenesis, a response to mechanical stimulation in early root growth, which remains largely unexplored. This study explores the intricate factors influencing early root system development, with a focus on the cooperative correlation between nutrient uptake and its flow dynamics. Using a physiologically as well as ecologically relevant, portable, and cost-effective microfluidic system for the controlled fluid environments offering hydraulic conductivity comparable to that of the soil, this study analyzes the interplay between nutrient flow and root growth post-germination. Emphasizing the relationship between root growth and nitrogen uptake, the findings reveal that nutrient flow significantly influences early root morphology, leading to increased length and improved nutrient uptake, varying with the flow rate. The experimental findings are supported by mechanical and plant stress-related fluid flow–root interaction simulations and quantitative determination of nitrogen uptake using the total Kjeldahl nitrogen (TKN) method. The microfluidic approach offers novel insights into plant root dynamics under controlled flow conditions, filling a critical research gap. By providing a high-resolution platform, this study contributes to the understanding of how fluid-flow-assisted nutrient uptake and pressure affect root cell behavior, which, in turn, induces mechanical stress leading to thigmomorphogenesis. The findings hold implications for comprehending root responses to changing environmental conditions, paving the way for innovative agricultural and environmental management strategies.

---

Introduction

The primary root that emerges first from the germinating seed functions as the physical anchor of the plant.¹ It is responsible for absorbing water and nutrients from its microhabitat and facilitating the transport of these resources to other parts of the plant.² The emergent primary root encounters challenges as it must grow and penetrate the heterogeneous soil substrate. Thus, the early stages of root growth are a critical window for the plant's survival.¹ Early growth and development of the root system are influenced by various abiotic factors such as nutrient supply, pH, total cationic concentration, aeration and soil temperature, humidity, strength, composition, and water content.^3–6 Besides these, external hydrostatic and hydrodynamic pressure are also known to have a non-trivial impact on root architecture.^7,8 Nutrient supply can potently affect root growth, morphology, and distribution in the substrate.⁹ The flow of nutrient solution or water provides plant roots with a suitable level of mechanical stimulation that subsequently promotes root growth.^10,11 The root growth and its branching are adjusted to optimize the nutrient provision to the plant. Among the nutrients required by plants, the availability of the macronutrient, nitrogen, plays a crucial role in plant growth and development. Nitrogen is known to play a vital role in the modulation of root architecture, root biomass, and growth. It may be mentioned here that a localized nutrient supply can stimulate root growth in an optimized fashion.^12–14 However, the experimental setup for studies involving roots typically utilizes macroscopic vessels and containers, which necessitates the development of methods for large-scale phenotyping.^15,16 This, in turn, results in challenges when studying the dynamics of plant root systems.

The advent of microfluidics, which refers to the study of flow dynamics, characterization of fluid properties, and its control within micrometer-sized structures,¹⁵ has enabled the recreation of in vitro environments for cell studies.¹⁶ In the realm of plant studies, however, only a limited number of microdevices have been developed so far, among which the majority of the literature deals with root–bacteria interactions,^17,18 hormonal signaling,¹⁹ and the growth of pollen tubes.^2,20–23 Nevertheless, as witnessed from the referred literature,^2,16,19,24 the exploration of in vitro investigations into plant root dynamics and its real-time analysis using on-chip microfluidic platforms has remained unexplored until recently.

While the influence of abiotic factors such as pH and nutrient availability on root morphology is well-documented,^25–34 the novel aspect of our study is the examination of the specific effects of mechanical stimuli resulting from nutrient flow on thigmomorphogenesis by using an on-chip microfluidic device, which has not been previously explored. Thigmomorphogenesis is a very real phenomenon in plants which affects many aspects of plant growth and development.³⁵ Thigmomorphogenesis is a phenomenon observed in plants where mechanical stimuli, such as touch or wind, induce changes in plant growth and development. These changes often result in alterations in plant morphology, physiology, and gene expression patterns. Thigmomorphogenesis can lead to modifications in stem elongation, leaf size, branching patterns, and root architecture. It is considered a form of plant response to mechanical stress and plays a role in adaptation to environmental conditions, including wind, soil texture, and competition with neighboring plants.^36–39 It is important to note that the nature or extent of the response depends on the species or variety as well as the physiological stage of the plant when it is stimulated. No literature has been found yet which investigates the influence of nutrient flow and its dynamics through microfluidics in plant roots. Hence, there is a need for perfusion devices that can be used to study the root dynamics and the interplay of nutrient solution at a high resolution, eliminating the challenges associated with specimen handling.

Taking into account the constraints related to current plant root studies, herein we describe and present the design of a physiologically and ecologically relevant, portable, easy-to-use, and low-cost microfluidic setup. The fluidic configuration considered in this endeavor offers hydraulic conductivity of equal order to that of the soil in real scenarios.^40,41 In this study, we employ Brassica juncea L. cv. Pusa Jaikisan, a fast growing, non-competitive, oilseed, dicot crop having an effective root diameter within the micrometer range.^42–44 The goal is to explore how nutrient flow conditions impact the growth of the developing root during post-germination stages. Also, we demonstrate the uptake of one of the vital macronutrients, nitrogen, and its correlation with plant root growth. This work provides a comprehensive description of the experimental setup and the reasoning behind its design choices. The presented experimental findings are validated through three key proofs of concept: (I) the estimation of generated mechanical stress through the numerical simulations of nutrient flow–root interaction within the microfluidic channel using a mathematical framework, (II) determination of nitrogen uptake using the total Kjeldahl nitrogen method,⁴⁵ and (III) the influence of fluid-flow-assisted nutrient uptake and fluid pressure on root cells using freehand anatomical sections.

Materials and methods

Plant material: seed sources and surface sterilization

Mature seeds of Brassica juncea (Pusa Jaikisan) were collected from ICAR-IARI, Regional Station, Karnal, Haryana, India. The seeds were stored in airtight falcon tubes at room temperature. To remove dirt and microbial contaminants present on the seed surface, the seeds were surface sterilized sequentially with 70% (v/v) ethyl alcohol for 30 seconds and 4% (w/v) sodium hypochlorite solution for 5 min. They were then thoroughly rinsed with autoclaved double-distilled water for 1 min (5 times). To surface-dry, the seeds were placed on UV-sterilized blotting paper in a Petri plate. The sterilization process was carried out inside a laminar-airflow cabinet, after which the seeds were used for in vitro germination.

Germination setup and conditions

Following the surface drying process, the sterilized seeds were finally placed within a sterile 90 mm Petri plate (Tarsons™) containing filter paper (Grade 1, Whatman®) wetted with sterile distilled water (cf.Fig. 1a-i). Seeds were arranged in a 5 × 5 pattern, forming columns and rows across the plate and sealed with laboratory film (Parafilm M®). Subsequently, the Petri plate containing the seeds was placed in a germination chamber maintained at a temperature of 26 °C and a relative humidity of 75% under a 16/8 hour (light/dark) photoperiod with a light intensity of 3000 lux provided by white fluorescent tubes.^{15,16,19,46–50}

Fig. 1 (a) Schematic representation of the experimental methods in the microfluidic study conducted. (i) Seeds were germinated under sterile conditions (ii) then the seeds were carefully transferred to the microfluidic setup. (iii) Nutrient flow within the channel was introduced using a mechanical syringe pump with a definite flow rate (Q). Syringes containing the nutrient were connected to the channels using sterile baby-feeding tubes, serving as the inlet for the channels. (iv) Upon successful integration of the root system with the microfluidic system, the root was observed under a microscope for 12 hours; micrographs were taken at 2 hour intervals without disturbing the plant system. (b) Experimental images of the setup showing the nutrient being pumped into the microfluidic setup in side view (left), single microfluidic channel as an inset of the figure; upper side, top view of the plant-integrated microfluidic setup (right). Note: images are not to scale.

Fabrication of PRFD and PRFS assembly

For observing and analyzing the real-time root growth dynamics of Brassica juncea, we fabricated a plant root fluidic device (PRFD), as shown in the inset of Fig. 1b. Each PRFD was fabricated with dimensions of 50 mm (length) × 5 mm (breadth) × 30 mm (height), with a channel of effective length 45 mm. For the study, eight such PRFDs, collectively known as Plant Root Fluidic System (PRFS), were used in experiments. The master mold of the device was designed using AutoCAD® software (Autodesk™, USA) and then 3D printed with acrylonitrile butadiene styrene (ABS) thermoplastics. The device was fabricated using a vinyl-terminated polydimethylsiloxane (SYLGARD™ 184 Silicone Elastomer, Sigma Aldrich®, USA) substrate, following an inexpensive wire-drawing method with adaptations from published methods.^51–53 It was then soft-baked at 45 °C for 10 hours in a hot air oven (IKON™ Instruments, India). Kindly refer to the ESI† for the 3D-printed ABS mold and casting setup of the wire-drawing method. The silicon polymer formed a negative mold of the device, which was carefully separated from the master mold to create the final device of interest. Locally purchased brass wire with a square-shaped cross section of 0.8 mm × 0.8 mm was used to mimic the horizontal channel for root growth in the PRFD, which enabled enhanced visualization and alignment stability. Pipette tips with a definite length and diameter were used to hold the seedling vertically and create inlet and outlet connections for fluid exchange. Here, it should be mentioned that the PDMS-based channels are biocompatible^54,55 (chemical inertness between root and PDMS) for the growth of the biological tissues.

Seedling integration in the PRFD

At 2 days after germination (DAG), under sterile conditions, the seedlings were taken from the Petri plate and placed inside the vertically positioned pipette tip connected with the horizontal channel (also known as plant inlet in Fig. 1a-ii), which was 15 mm away from the inlet port of the PRFD. The hydrodynamic head (= (u_avg)²/2g) for the highest flow rate (1.20 mL h⁻¹) was found to be 1.37 × 10⁻⁵ mm, much lower than the height of the top plant root inlet port, to prevent spillage of the nutrient solution. u_avg and g are the average velocity inside the PRFD and the gravitational constant, respectively. The procedural steps of seedling transfer from the Petri plate to each PRFD are schematically demonstrated in Fig. 1a-ii and iii. This placement allowed the growing roots to conveniently enter the horizontal channel underneath. Each channel in the PRFD, configured with a plant, was supplied with full-strength MS medium for 12 hours via the inlet port, ensuring definite and distinct fluid-flow conditions while maintaining consistent environmental conditions for all.

Real-time micrography of the root dynamics

For real-time micrography of the root growing inside each PRFD, the device was placed under a Leica® DMI3000 M inverted microscope coupled with a high-speed camera (Phantom, Vision Research Inc., USA; model: Miro LAB320) with a resolution of 1900 × 1200 pixels, as represented schematically in Fig. 1a-iv. The objectives used were HC FL PLAN 2.5×/0.07, 5×/0.12 N PLAN EPI, and 10×/0.25 N PLAN (Leica). Micrographs for each case were imaged every 2 hours using the aforementioned objectives over a 12 hour period.

Image acquisition and data analysis

Image analysis was performed using ImageJ/Fiji (open source) software to measure the lengths of the roots under different flow rate conditions. Subsequently, data analysis was carried out using MATLAB® R2022a (The MathWorks Inc.).

AFM-assisted imaging and measurement of Young's modulus

To determine the Young's modulus of the growing root (soft tissue), the procedure followed in this study is briefly outlined here for the sake of completeness. The root sample was removed from the PRFD and then the root was detached from the shoot portion. The detached root was surface dried with the help of blotting paper and split longitudinally. The sample was then air-dried (about 10 min) to remove excess water from the surface of the root to allow it to be placed on carbon tape for imaging. The root samples, containing plant roots under all the flow rate conditions in the PRFD set overnight (12 hours), were placed on carbon tape stuck on thoroughly cleaned glass slides using sterilized forceps. Subsequently, the air-dried root sample was imaged using an Asylum Research MFP-3D-BIO (Asylum Research of Oxford Instruments®, UK) atomic force microscope (AFM). The slide was fixed temporarily onto a metal disc and then placed in the AFM for scanning. AFM imaging was conducted in nanoindentation mode using an AC160TS-R3 probe (Oxford Instruments®, UK) with a setpoint of 50 nN and a speed of 4 μm s⁻¹. It is to be mentioned here that wet root samples are more adhesive to the cantilever tip of the AFM, which can result in erroneous data with a larger negative deviation in the approach curve.⁵⁶ The Young's modulus was determined by fitting curves using the Hertz model as depicted in Fig. 4a, while the sample preparation steps, from fixation to imaging, together with the sample used for this part are shown in the inset of Fig. 4a. Likewise, the pointwise modulus and indentation curves are shown in Fig. 4b and its inset, respectively. The analysis of AFM images and the calculation for Young's modulus were performed using both AtomicJ (open source) and NanoScope Analysis (Bruker, Billerica, MA, USA) software.

Anatomical analysis of the root sections: staining, imaging, and analysis

After 12 hours of observation in the microfluidic channel, the plant was carefully taken out and freehand sections of the delicate root were done by embedding the root in agar medium. Sections were taken within a distance of 1 to 2 mm from the root tip for 3 different samples of the respective flow rates. Kindly refer to the ESI† for the region of interest (RoI) containing the elongation and differentiation zone of root selected for anatomical sectioning. The thin sections were then stained with safranin, followed by a series of washes with graded alcohol and water. The thin sections (stained samples) were observed under a microscope, and images of the sections were analysed using ImageJ/Fiji software (open source).

Estimation of nitrogen uptake

Whole plant samples from each experimental case were collected and oven-dried for 24 hours. Subsequently, 200 mg of each sample was used for the estimation of nitrogen by the total Kjeldahl nitrogen (TKN) method⁴⁵ using a TKN analyzer (Kelpus-Distyl EM VA, Pelican Equipment, India).

Statistical analysis of root growth dynamics

The data in this study were expressed as mean values ± standard deviation (S.D.). Data were analyzed using IBM® SPSS® version 29.0 (IBM Corp., Armonk, NY) for significance using one-way analysis of variance (ANOVA), followed by Duncan's multiple range test (DMRT) for contrasting differences. A significance level of p_r < 0.05 was considered significant, unless otherwise specified, where p_r defines the probability and determines the likelihood that any observed variation between groups is a result of random occurrences.

Results and discussion

Plant root fluidic system (PRFS) and experimental setup

The seedlings of Brassica juncea, after germination (Fig. 1a-i), were placed in the plant root fluidic device (PRFD) as depicted schematically in Fig. 1a-ii and iii. The actual experimental setup consisted of eight PRFDs, each having a designated microfluidic channel with effective dimensions of 45 mm (length) × 0.8 mm (breadth) × 0.8 mm (height), collectively termed as plant root fluidic system (PRFS). Each vertically standing PRFD (precisely, each microfluidic channel) in the PRFS, shown in Fig. 1b as top and side view, was connected to individual 10 mL syringes (NIPRO, India) containing MS medium solution (Murashige & Skoog medium, PT021, HIMEDIA®, India) through individual inlet ports using sterile feeding tubes with an effective length and diameter of 500 mm and 2.70 mm, respectively. It is to be noted that the MS medium, used as nutrient solution in the present study, was found to be a Newtonian liquid, exhibiting properties similar to that of water. The syringes were, in turn, connected to a syringe pump (New Era Pump, India) to supply an effective and continuous flow of the nutrient medium/solution through the channel of each PRFD. Henceforth, for the sake of convenience, we will simply write channel instead of PRFD. The outlet ports of each channel were extended to the collection sump using feeding tubes and straight connectors as seen in the side view of Fig. 1b. To avoid contamination during the plant growth period, the entire experimental setup was placed in a UV-sterilized, closed plant growth chamber maintained at a temperature of 26 °C and 75% relative humidity, with a 16 hour light/8 hour dark cycle.

Fluid-flow configuration

Following the integration of the seedling into the PRFDs, as described in the Materials and methods section, each seedling was allowed to grow in a no-flow condition until its root entered the horizontal channel from the vertically placed pipette tip. Once the root entered the channel, syringe pumps were used to infuse nutrient solution to nourish the growing roots. Fluid-flow inside the channel was set at various rates: 0.05 mL h⁻¹, 0.1 mL h⁻¹, 0.2 mL h⁻¹, 0.4 mL h⁻¹, 0.6 mL h⁻¹, 0.8 mL h⁻¹, 1.0 mL h⁻¹ and 1.2 mL h⁻¹. The chosen window of flow rates complies with the physically permissible range typically considered in this paradigm.^57–59 For each flow rate, the flow direction was maintained in the positive direction of the root growth (towards the channel outlet). An additional no-flow condition was maintained as a control for each flow rate configuration. The setup remained inside the sterile growth chamber and was observed for 12 hours after the root entered the horizontal channel. Each experiment was replicated three times on three separate instances, with eight root samples for each case.

Hydraulic conductivity of PRFD and soil

We attempted to estimate the hydraulic conductivity of the flow inside the developed PRFD and compared it with the hydraulic conductivity of soil available in the referred literature.^57,58 We undertook this endeavour to ascertain whether the flow rate considered through the PRFD during experiments in this study mimics practical scenarios. The instantaneous flux, or flow rate per unit cross-sectional area, can be expressed using Darcy's law: q = (KΔp)/μL. Here, K, Δp, μ and L stand for the soil's permeability (measured in m²), pressure drop, dynamic viscosity, and length of PRFD, respectively. Expressing it as q = (KρgΔh)/μL or q = (Kρg/μ)(Δh/L) using p = ρgh, where h is the pressure head, the term Kρg/μ is denoted by the letter κ and stands for hydraulic conductivity (in m s⁻¹).

It should be noted that the hydraulic conductivity of the PRFD is expected to be in the order of O(κ) ∼10⁻¹ m s⁻¹ based on the pressure drop and flow rate per unit cross-sectional area. For a flow rate of 0.05 mL h⁻¹ (equivalent to 1.38 × 10⁻¹¹ m³ s⁻¹), with a rectangular cross section of the PRFD of 64 × 10⁻⁸ m², the flux is calculated as q = 2.17 × 10⁻⁵ m s⁻¹. Similarly, the value of q is obtained as 5.208 × 10⁻⁴ m s⁻¹ for a flow rate of 1.2 mL h⁻¹. Moreover, using μ = 10⁻³ Pa s, ρ = 1000 kg m⁻³, L = 10 × 10⁻³ m, and the range of pressure drop was estimated from numerical simulation for channels with and without root as 0.028635 Pa to 0.80188 Pa and 0.010109 Pa to 0.28323 Pa, respectively, for the considered range of flow rates (0.05 mL h⁻¹ to 1.20 mL h⁻¹). Accordingly, the hydraulic conductivity κ (= Kρg/μ) was found to be in the range of 0.06365 m s⁻¹ to 0.07427 m s⁻¹ and 0.18021 m s⁻¹ to 0.21038 m s⁻¹, for channels with and without root for the flow rates 0.05 mL h⁻¹ to 1.20 mL h⁻¹ with g = 9.8 m s⁻².

Notably, available research in the referred literature indicates that the biochar-rich agricultural soil utilized in agriculture has hydraulic conductivity of up to O(κ) ∼10⁻² m s⁻¹.⁵⁸ Also, for the sustainable growing medium from waste used in hydroponics, the hydraulic conductivity is up to 0.073 m s⁻¹ (ref. 57) or in the order of O(κ) ∼10⁻¹ m s⁻¹. A similar order of hydraulic conductivity approaches for biochar-based soil was found in the review presented by Yan et al. (2021).⁵⁹

Mathematical model for nutrient flow–root interaction

Here, we attempt to estimate the internal mechanical stress developed inside the root resulting from stationary flow loading at the nutrient solution–root interface. In order to do this, we solve for the flow field and deformation inside the root under the impact of stationary flow loading, consistent with the equivalent fluid–root interactions using the finite element framework of COMSOL Multiphysics™.

Governing equations for flow field

The following continuity and momentum equations, compatible with the current experimental setup, are employed to simulate the flow field. We assume a steady state and incompressible flow of Newtonian nutrient solution in this endeavor:

∇·(u) = 0 (1)

ρ(u·∇)u = −∇p + μ∇²u (2)

Here, u, p, μ and ρ denote the velocity vector, local pressure, dynamic viscosity, and density of nutrient solution, respectively. The density and viscosity of the nutrient solution, which exhibits properties similar to water, are taken as ρ = 1000 kg m⁻³ and μ = 0.001 Pa s, respectively. We used a no-slip and no-penetration condition (u = 0) at the PRFD wall as well as at the root–nutrient liquid interface in order to calculate the flow field. The average velocity determined from the flow rate measured through experiments is imposed as the uniform velocity normal to the inlet wall.

Governing equation for deformation field in the root and coupling with flow

The equations below are capable of estimating the internal mechanical stress that develops in the root under stationary flow loading:^60,61

∇(FS)^T = 0 (3)

Here, F and S are the displacement gradient and the second Piola–Kirchhoff stress. Note that F can be represented in terms of the displacement field (V_s) as follows:⁶¹

F = I + ∇V_s (4)

where I is the identity matrix. Moreover, the expression of S is denoted as:⁶¹

S = 2μ_L[small epsilon, Greek, macron] + λ_Ltr([small epsilon, Greek, macron])I (5)

Here, [small epsilon, Greek, macron] is the Lagrange–Green strain and expressed as [small epsilon, Greek, macron] = 0.5(F(F)^T − I). The first and second Lame parameters appearing in eqn (5) are as follows:⁶¹

λ_L = νE/(1 + ν)(2ν − 1), μ_L = E/2(1 + ν) (6)

where E and ν are the modulus of elasticity and Poisson's ratio, respectively. We represent the Cauchy stress tensor inside the root by the following relation:In eqn (7), J stands for the Jacobian of F.

The following interfacial boundary condition is imposed at the fluid and root interface to couple the flow field and deformation in the root:

[small sigma, Greek, macron]·n_s,Γ = −(pI + μ∇u)·n_f,Γ (8)

where n is the normal unit vector to the interface, Γ; the suffixes s and f denote solid root and nutrient fluid domains, respectively.

As evident from eqn (3), the Young's modulus of the growing root is a critical parameter to estimate the deformation of the root under external forces⁶² (flow loading). Unfortunately, no prior literature reporting the mechanical properties, including the Young's modulus, of the same plant is available. To address this gap, we attempt to experimentally determine the Young's modulus of our sample plant root (Brassica juncea) using atomic force microscopy (AFM), in accordance with the Hertz fit model, as described in the Materials and methods section. The estimated values of average Young's modulus and Poisson's ratio for the growing root are obtained as 4.1625 MPa and 0.49, respectively. For the sake of conciseness in the presentation and to maintain the focus of the present analysis, we do not explicitly discuss here the measurement method. However, interested readers may refer to the seminal studies available in this paradigm for this part.^63–71 For the simulations of the underlying flow field, we apply a no-slip boundary condition at the liquid–solid interface and assume a fully developed velocity field at the inlet to solve eqn (1) and (2). It is worth mentioning here that the effect of root deformation on fluid flow is neglected (i.e., one-way coupling) due to the weaker flow loading.

PRFS facilitates real-time imaging of root micrograph

Morphological analysis of the data obtained from micrography of the roots revealed that the variations in the flow rate of the nutrient solution in the microfluidic channel affected root length, as shown schematically in Fig. 2a and as experimental micrographs (inset) in Fig. 2b. As evident from regime-I of Fig. 2b and c, the root length, in contrast to the no-flow scenario (control), increased significantly with the increase in the flow rate from 0.05 mL h⁻¹ (root length = 5.87 ± 0.373 mm) to 0.4 mL h⁻¹ (root length = 11.16 ± 0.378 mm). This is attributed to the cellular longitudinal elongation¹² facilitated by the flow rate-modulated enhancement of nitrogen uptake as discussed in the forthcoming section. Interestingly, as observed in the analyzed results, any further increase in flow rate beyond 0.4 mL h⁻¹ (i.e., 0.6 mL h⁻¹, 0.8 mL h⁻¹, 1.0 mL h⁻¹ and 1.2 mL h⁻¹) resulted in a declining trend in root length (regime-II of Fig. 2b and d). This reduction in root length in regime II is attributed to the augmented mechanical stress, which leads to excess accumulation of auxin.¹⁴ This excess auxin accretion inhibits the elongation of the root despite the higher nitrogen uptake (as evident in the forthcoming section). Notably, the mean values marked with the same letter(s) did not differ significantly at p_r ≤ 0.05 according to Duncan's multiple range test, where p_r defines the probability and determines the likelihood that any observed variation between groups is a result of random occurrences (Fig. 2e). We refer to the flow rate of 0.4 mL h⁻¹, at which the root length is maximum, as the “optimum flow rate”. The root length obtained at a flow rate of 0.6 mL h⁻¹ (root length = 9.29 ± 0.392 mm) shows a steep decrease compared to the root length at the optimum flow rate, i.e., 0.4 mL h⁻¹. The intricate competition between hydrodynamic stress (mechanical stress) inevitably associated with the nutrient flow and corresponding nutrient uptake is deemed pertinent to reveal the basic morphogenesis of root length architecture with a change in flow rate. The observed reduction in root length seems to be associated with heightened mechanical stress on growing roots induced by hydrodynamic stimulation, as indicated by our simulated data. However, it is important to note that in flow conditions, the observed root length consistently exceeded that in no-flow conditions. This can be primarily attributed to progressive nitrogen uptake by the growing root with nutrient flow rate, which indeed counteracts the mechanical stress experienced by them. We will discuss these aspects in greater detail in the later part of this article from the perspective of histology, biochemical analysis, and simulations.

Fig. 2 (a) Schematic representation of channel configuration for Q = 0.05 mL h⁻¹, Q = 0.4 mL h⁻¹ and Q = 1.2 mL h⁻¹ and no flow state. The red arrow () represents the direction of the flow of the nutrient through the channel. (b) Variation in the length of the root growing inside the microfluidic channel with respect to different flow rates (Q) at t = 12 h. The two regimes, namely, regime-I and regime-II, depict the change in root length in mm. The pink circle () denotes the length of the root (4.63 ± 0.34 mm) in the no-flow condition. The red marker () denotes the root length at optimum flow rate, 0.4 mL h⁻¹. Regime-I denotes the increase in root length as Q increases from 0.05 mL h⁻¹ to 0.4 mL h⁻¹. Regime-II denotes the decrease in root length when Q is increased beyond 0.4 mL h⁻¹ up to 1.2 mL h⁻¹. Micrographs in regime-II show the root length variation for the no-flow condition, Q = 0.05 mL h⁻¹, Q = 0.4 mL h⁻¹ and Q = 1.2 mL h⁻¹ at t = 12 h. Different markers are used to denote the ascending trend (Q = 0.05 mL h⁻¹ to 0.2 mL h⁻¹), descending trend (Q = 0.6 mL h⁻¹ to 1.2 mL h⁻¹) and optimum growth condition (Q = 0.4 mL h⁻¹). (c) Increasing variation in the length of the root growing inside the microfluidic channel for different Q with respect to time. (d) Decreasing variation in the length of the root growing inside the microfluidic channel for different Q with respect to time. In both variations, increasing as well as decreasing, the length of the root for different Q is greater than that in the no-flow condition. (e) Variations in the average root length at different flow rates. The graphs were plotted using MATLAB® R2022a, and the statistical significance of the data was assessed through one-way ANOVA using IBM® SPSS® software. The mean values marked with the same letter(s) do not exhibit significant differences at a significance level of p_r ≤ 0.05, as determined by Duncan's multiple range test, where p_r defines the probability and determines the likelihood that any observed variation between groups is a result of random occurrences.

Anatomical section of root and cellular structure

The analysis of micrographs from the freehand anatomical cross section of roots under different flow rates (0.05 mL h⁻¹, 0.4 mL h⁻¹, and 1.2 mL h⁻¹) reveals a peculiar change in the count and size of the cortical cells beneath the epidermal layer of the root sections. Note that Fig. 3a demonstrates anatomical sections of root depicting the cellular arrangement with the cortex cell for three different flow rates. The cortical cell area is the region marked within the blue lines, represented by the symbol “Ω” (cf.Fig. 3a). Specifically, the average cortical cell count (C_c) increases with the enhancement in flow rate. For the roots under flow conditions consistent with the rate of 0.05 mL h⁻¹ (Q₁), 0.4 mL h⁻¹ (Q₂), and 1.2 mL h⁻¹ (Q₃), C_c is found to be 68 ± 3^a, 73 ± 2^a and 85 ± 3^b, respectively (Fig. 3a). The mean values for C_c(Q₁) and C_c(Q₂), marked with the same letter, do not exhibit significant differences between themselves, whereas C_c(Q₃) shows a marked difference in C_c compared to both C_c(Q₁) and C_c(Q₂), at a significance level of p_r ≤ 0.05 as determined by Duncan's multiple range test (Fig. 3b). Notably, the higher flow rate (Q₃) exhibits a significant increase in C_c compared to the lower flow rates (Q₁) and (Q₃). In contrast, C_c for Q₁ and Q₂ shows a smaller variation in the number of cells, suggesting a change in C_c with an increase in flow rate. These observations can be attributed to cellular dedifferentiation^72,73 resulting from varying levels of auxin accumulation due to changes in mechanical stress in different flow regimes. For a change in flow rate from Q₁ to Q₂, relatively lower magnitudes of flow rate allow for a relatively lesser mechanical stress; hence the lower amount of auxin accumulation inside the root may lead to minimal dedifferentiation of cells, whereas the substantially higher mechanical stress at larger flow rates (Q₂ to Q₃) leads to augmented dedifferentiation of cells by amplified auxin accumulation. Because of this flow rate-modulated enhancement of auxin accumulation, a significant increase in cortical cell count is predicted from the increase in flow rate from Q₂ to Q₃. Furthermore, we calculated the average area of cortical cells (A_c) for the flow rates Q₁, Q₂ and Q₃ using ImageJ as 0.1164 ± 0.0286 mm², 0.1115 ± 0.0296 mm² and 0.0912 ± 0.0269 mm², respectively, a portion of which is shown as zoomed-in-view images in Fig. 3a. The mean values for A_c(Q₁) and A_c(Q₂) do not differ significantly between themselves, whereas A_c(Q₃) exhibits a difference in A_c compared to both A_c(Q₁) and A_c(Q₂), at p_r ≤ 0.01 according to Duncan's multiple range test as seen in Fig. 3c. The data were used to deduce the reduction (R) percentage in the area of cells in the cortex for flow rates Q₂ and Q₃ relative to the lower flow rate Q₁ (as graphically shown in the inset of Fig. 6a in the upcoming section). The cortical cell distribution count and cell size reduction parameters are indicative of the adaptive nature of the roots to the mechanical stress induced by the varying fluid flow rate inside the fluidic channel. As seen in Fig. 2b, the change in root length from no flow condition to flow rate 0.4 mL h⁻¹ is higher than the change in length from 0.4 mL h⁻¹ to 1.2 mL h⁻¹, yet there is a negligible change in the cortical cell area with the window of flow rate considered (cf.Fig. 3b and c). The negligible change in cortical cell area from no-flow condition to flow rate 0.4 mL h⁻¹ is attributed to the minimal dedifferentiation due to the miniscule auxin accumulation at lower mechanical stress provided by lower flow rate. Moreover, at higher flow rates (0.4 mL h⁻¹ to 1.2 mL h⁻¹), the higher compressive hydrodynamic stress on the growing root and large auxin accumulation¹⁴ shows a combined effect leading to the dedifferentiation of cortical cells which further tends to compress the root. The eventual consequence of these two effects leads to a change in the cortical cell area with an increase in the flow rate as witnessed in Fig. 3c. Quite notably, we see that the morphological changes of the root are accompanied by anatomical (cellular) changes, such as increased number of cortical cells and decreased size of the cortical cells in the root.^72,73 We have elaborated on this aspect in the next section.

Fig. 3 (a) Anatomical sections of the root depicting cell distribution and average cortical cell count (C_c) for (i) Q = 0.05 mL h⁻¹, (ii) Q = 0.4 mL h⁻¹ and (iii) Q = 1.2 mL h⁻¹. The zoomed view depicts detailed information about a portion of cortex beneath the epidermal layer, and the average area of cortical cells for each root section (A_c) is denoted below the zoomed view boxes. The cortical cell region is marked within the blue lines, represented by the symbol “Ω”. (b) Variation in the average cortical cell count (C_c) of the flow rates Q = 0.05 mL h⁻¹, Q = 0.4 mL h⁻¹ and Q = 1.2 mL h⁻¹. (c) Variations in the average area (A_c) of cortical cells at different flow rates Q = 0.05 mL h⁻¹, Q = 0.4 mL h⁻¹ and Q = 1.2 mL h⁻¹. The graphs were plotted using MATLAB® R2022a and the significance of the data was verified by one-way ANOVA using IBM® SPSS® software. The mean values marked with the same letter(s) do not differ significantly at p_r ≤ 0.05 according to Duncan's multiple range test.

Simulating nutrient flow–root interaction and uptake

The effective growth of the plant root in the nutrient-saturated microenvironment depends on the intricate competition between nutrient uptake, together with its flow rate, and abiotic stress: here, flow induced hydrodynamic stress resulting in thigmomorphogenesis. Understanding this complex interplay between nutrient flow and root structure interaction, which is typical in the growth of plant roots in nutrient-mediated flow environments, is of particular importance. To delve deeper into this competition, we developed a mathematical modeling framework and performed numerical simulations on nutrient flow–root interaction using real-time images of the root captured from our experiments at a given temporal instant. Young's modulus, a critical parameter for simulations of the growing root inside the channel, was determined by curve fitting (Fig. 4a and b) obtained from atomic force microscopy (AFM), as elaborately discussed in the Materials and methods section. The simulated results illustrate contours representing effective internal mechanical stress, normal loading at the interface, and the magnitude of loading, all of which indicate average hydrodynamic stress levels (Fig. 5a). Here, loading implies the combined imposition of the pressure normal to the root surface and tangential resistance inside the root against the shear stress of liquid at the root–nutrient solution interface (−(pI + μ∇u)·n_f,Γ)). It is to be noted that the magnitude of loading at the root–nutrient interface ollows the decreasing trend in the longitudinal direction. It is attributed to the reduction in normal load because of the longitudinal decrease in pressure inside the PRFD for all the flow rates considered, as shown in the third column of Fig. 5a. The simulated results illustrate contours representing effective internal mechanical stress, normal loading at the interface, and the magnitude of loading, all of which indicate average hydrodynamic stress levels (Fig. 5a). These simulated results are associated with the anatomical sections of the root focusing on the vascular bundles depicted in the last column of Fig. 5a. Vascular bundles are composed of two major tissues, xylem and phloem, which are the main conduits for the transport of water and solutes throughout the plant.⁷⁴ Mechanical stress (hydrodynamic stress) can alter the size, shape and arrangement of the tissues of the vascular bundle.^72,74 The anatomical sections of the roots reveal minimal differences in both the structure and area (A(Q₁) = 0.2299 ± 0.0025^a mm², A(Q₂) = 0.2311 ± 0.0021^a mm², A(Q₃) = 0.2287 ± 0.0026^a mm²) of the vascular bundles within the root. The superscript letter ‘a’ signifies that there is no significant difference in the values of A. These results indicate that the outer layers of the cells act as a shield against the hydrodynamic stress, which are substantiated with the simulated results elaborated in the forthcoming discussion. A schematic representation of the anatomical section is shown in Fig. 5b, depicting the cellular architecture. The mean values, A(Q₁), A(Q₂) and A(Q₃), do not differ significantly at p_r ≤ 0.01 according to Duncan's multiple range test (Fig. 5c).

Fig. 4 (a) Force–distance curve obtained from AFM analysis and then fitted to the Hertz model to extract the Young's modulus of the root sample. (i–iii) Roots of Brassica juncea in post-germination stage were cut into pieces (1 cm), air-dried and then placed on a microscopic glass slide to minimize spacing between them. The air-dried root sample was imaged using an Asylum Research MFP-3D-BIO (Asylum Research of Oxford Instruments) atomic force microscope (AFM). AFM imaging was conducted in nanoindentation mode using a silicon cantilever probe with a radius of 7 nm and a spring constant of 26 N m⁻¹ (model: AC160TS-R3, Oxford Instruments). (b) And the inset within represent the pointwise modulus and indentation curves obtained from the analysis of AFM data using AtomicJ software. The Young's modulus values obtained were subsequently utilized in simulations conducted in COMSOL Multiphysics Software™.

Fig. 5 (a) Contours of effective internal mechanical stress under flow loading (Pa), normal loading at the interface, and magnitude of loading (Pa) obtained from numerical simulations; anatomical sections (extreme right) represent the area of vascular bundles (VB) with no significant structural and area difference. (b) Schematic representation of the cellular architecture shown in Fig. 4a. (c) Variations in the average area of vascular bundles at different flow rates Q = 0.05 mL h⁻¹, Q = 0.4 mL h⁻¹ and Q = 1.2 mL h⁻¹. The graphs were plotted using MATLAB® R2022a and the significance of the data was verified by one-way ANOVA using IBM® SPSS® software. The mean values marked with the same letter(s) do not differ significantly at p_r ≤ 0.05 according to Duncan's multiple range test.

The average mechanical stress (inside the root, σ_avg) experienced by the growing root in the PRFD varies from 8.05 × 10⁻³ N m⁻² to 166.42 × 10⁻³ N m⁻² for the flow rates 0.05 mL h⁻¹ to 1.2 mL h⁻¹ (Fig. 6a). The gradual increase in flow rates induces greater fluid flow-induced mechanical stress on the soft tissues of the growing root. In regime-I of Fig. 6a, which maps the flow rates from Q = 0.05 mL h⁻¹ to 0.04 mL h⁻¹, the average mechanical stress exhibited a steeper slope of 0.1659 N m⁻² mL⁻¹ h. In regime-II of Fig. 6a, mapping flow rates beyond 0.4 mL h⁻¹, specifically, Q = 0.6 mL h⁻¹ to 1.2 mL h⁻¹, the average mechanical stress exhibited a relatively lesser steep slope of 0.1288 N m⁻² mL⁻¹ h than that of regime-I. These variations in mechanical stress are reflected in the anatomical changes of the root, i.e., percentage reduction (R) of the average area of cortical cells for flow rates Q₂ and Q₃ relative to the lower flow rate Q₁, as shown in the inset of Fig. 6a. We observed a 4.159% reduction (R_Q2,Q1) in the area between cortical cells of roots subjected to flow rate Q₂ and a 21.580% reduction (R_Q3,Q1) in area for roots subjected to flow rate Q₃. These changes, along with the change in the number of cortical cells (Fig. 3a) indicate thigmomorphogenesis pertaining to the hydrodynamic stress experienced by the growing root with an increase in flow rates from Q₁ to Q₃. It is to be noted here that although changes in cortical cell count and area were observed with the change in flow rates (Fig. 3a), no significant change in the structure and area of vascular bundles within the roots, critical for water uptake and transport, was observed (Fig. 5a, last column). These observations indicate the cumulative shielding effect provided by cortical cells to the inner vascular bundles against flow-induced mechanical stress.

Fig. 6 (a) Variation in average mechanical stress (N m⁻²), obtained from numerical simulations, exerted by the nutrient solution on the root growing inside the microfluidic channel with respect to different flow rates (Q) at t = 12 h. Q₁, Q₂ and Q₃ denote the flow rate: 0.05 mL h⁻¹, 0.4 mL h⁻¹ and 1.2 mL h⁻¹, respectively. The anatomical sections for roots subjected to flow rates Q₁, Q₂ and Q₃ are provided in the inset. The percentage reduction (R) in the area of individual cortical cells is calculated with respect to the flow rate, Q₁, for Q₂ and Q₃. A 4.159% reduction in the area is observed between cortical cells of the root subjected to flow rates Q₁ and Q₂, whereas a 21.580% reduction in area is observed for root subjected to flow rates Q₁ and Q₃, the change in trend of the area of cortical cells is represented in inset. The two regimes, namely, regime-I and regime-II, are based on the slope of the average mechanical stress with flow rate (Q). The equations for the slope of the corresponding flow rates obtained from the best curve fitting are obtained as average mechanical stress inside the root (σ_avg) = 0.1659Q + 0.0023 in N m⁻² and σ_avg = 0.1288Q + 0.0098 in N m⁻², with Q in mL h⁻¹, for regime-I (R² value for fitting: 0.98006) and regime-II (R² value for fitting: 0.98098). (b) Variations in the percentage of N₂ extracted from root samples with an equal mass (0.2 g) in relation to changes in flow rate. Total nitrogen percentage was determined using the total Kjeldahl nitrogen (TKN) method. The two regimes, namely, regime-I and regime-II, are based on the slope of the percentage of nitrogen with flow rate (Q). The slope of the corresponding flow rates obtained from the best curve fitting is depicted in the inset, and the corresponding equations are obtained as % N₂ = 2.765Q + 0.9433 and % N₂ = 0.665Q + 1.862, with Q in mL h⁻¹, for regime-I and regime-II, respectively.

Flow-modulated mechanical stress versus nutrient uptake

The macronutrient nitrogen is crucial for plant growth, and its uptake plays a vital role in stimulating plant root growth.^75,76 Plants have developed adaptive mechanisms to cope with fluctuating soil nitrogen levels, particularly nitrate (NO₃⁻) and ammonium (NH₄⁺). They regulate hydraulic properties in specific root sections based on local nitrogen conditions as discussed next: as NO₃⁻ depletes in an area, reduced hydraulic conductance redirects water to regions with higher NO₃⁻ concentrations. Roots proliferate rapidly in NO₃⁻-rich patches due to its higher mobility and signaling role, while growth in NH₄⁺-rich areas is slow due to NH₄⁺ ion's lower mobility and different signaling properties. Structurally, plants develop thinner, longer roots when NH₄⁺ is the main nitrogen source, with NH₄⁺ absorption concentrated at the root apex and NO₃⁻ absorption highest in the root hair zone, where hydraulic conductivity is enhanced by aquaporins. Initially, root tips absorb more NH₄⁺, while extended roots absorb NO₃⁻ in the root hair zone. This nitrogen-regulated water flux controls the mass flow of other nutrients to the rhizosphere, influenced by water flux, soil nutrient concentration, and nutrient mobility.⁷⁷ Thus, the analysis of the nitrogen uptake can provide valuable insights into root growth dynamics under different flow conditions. The analysis of macronutrient nitrogen was performed using the TKN method. The overall data from the Kjeldahl analysis revealed a trend where nitrogen uptake increased with an increase in the flow rate. In regime-I of Fig. 6b, which covers flow rates from Q = 0.05 mL h⁻¹ to 0.04 mL h⁻¹, including no-flow condition, the nitrogen percentage exhibited a steep slope (R² value for fitting = 0.98006) of 2.765% N₂ per mL per h. This steep slope signifies a gradual and linear increase in nutrient uptake induced by the flow. Interestingly, this linear increase continues until the optimum flow rate, Q = 0.4 mL h⁻¹, beyond which the slope becomes less steep. In regime-II of Fig. 6b, encompassing flow rates beyond 0.4 mL h⁻¹, specifically Q = 0.6 mL h⁻¹ to 1.2 mL h⁻¹, the nitrogen percentage exhibited a relatively lesser steep slope (R² value for fitting = 0.98098) of 0.665% N₂ per mL per h. The lesser steepness of the slope in regime-II of Fig. 6b is attributed to the reduction in effective surface area of roots (see Fig. 2) for nitrogen absorption from nutrient solution. At the same time, the enhanced flow rate reduced the residence time for nitrogen-rich species (NO₃⁻ and NH₄⁺) to interact with the root surface from the nutrient solution. These combined effects led to a reduction in the slope of nitrogen uptake for the higher flow rates considered in this endeavor, as seen in regime II of Fig. 6b. It is indeed fascinating to observe that the mechanical stress induced by the increase in flow rate is mitigated by the cortical cells, thereby leaving the vascular bundle region almost unaffected (Fig. 5a shows that the vascular bundle area, shown in the last column, remains almost unchanged with increasing flow rate). However, during this process, the cortical cells experience a reduction in average cell area (Fig. 3c and inset of Fig. 6a). This reduction in cell area is accompanied by an increase in the number of cortical cells (Fig. 3a and b, showing that the C_c increases with flow rate). This structural adaptation, discussed in the reported studies as well,^78,79 may contribute to the overall increase in nitrogen uptake observed in regime-II, although the uptake rate remains lower compared to regime-I (Fig. 6b).

In 1973, Jaffe³⁵ conducted a study involving a variety of plant species, wherein he applied gentle mechanical stimulation by rubbing the internodes and observed a noteworthy reduction in stem elongation as a direct response to the mechanical stimulus. This investigation sheds light on the suppressive impact of mechanical stimulation, particularly through rubbing, on stem elongation within specific plant species. This phenomenon is characterized as thigmomorphogenesis and is perceived as an adaptive mechanism aimed at mitigating environmental stresses, with potential involvement of the hormone ethylene. Thus, thigmomorphogenesis exerts a significant influence on various facets of plant growth and development. It is worth emphasizing that the character and extent of this response hinge on factors such as the plant species or variety and the specific physiological stage at which the plant undergoes stimulation. Our study advances the understanding of root cell behavior in Brassica juncea plant using a microfluidic platform, with a focus on the cooperative correlative influence of fluid-flow-assisted nutrient uptake and corresponding mechanical pressure as stimuli, resulting in what we term ‘flow-induced thigmomorphogenesis’.

The water and nutrients absorbed by the root epidermis are transported by the cortex, located between the root epidermis and the endodermis, into the vascular bundle by apoplastic and symplastic routes, and are involved in determining the fate of epidermal cells.⁸⁰ The cortex in plants is a relatively undifferentiated cell type called ground tissue. Although the cortex layers vary among species, they remain almost unchanged at a particular developmental stage for a specific species.⁸¹ However, under stressed conditions, such as mechanical stress,⁸² cortical cell division and root developmental plasticity⁸² are regulated by auxin, a phytohormone that primarily controls growth and development patterns in plant.^83,84 Auxin is known to accumulate at the site of stress through long- and short-distance auxin flow,⁸³ resulting in cortical cell division. Our experimental findings and simulated data corroborate the enhancement of stress with an increase in flow rates (cf.Fig. 6a). Remarkably, as evident from Fig. 3a, an increase in flow rate (Q₂ to Q₃) results in a significant increase in the number of cortical cells. This observation aligns with the concept of stress-stimulated auxin accumulation in roots as discussed in the reported studies.^14,83 In our research, we have, for the first time, investigated the impact of mechanical stimuli induced by variations in flow rates on the morphological and anatomical characteristics of Brassica juncea plant roots. Our findings align with the well-established concept of thigmomorphogenesis, a term frequently employed to elucidate plant responses to mechanical stimuli. The changes observed in our study, which encompass modifications in root structure and the outcomes of stress-related fluid–root interaction simulations, underscore the dynamic and adaptive nature of thigmomorphogenesis in plants undergoing alterations in mechanical forces.

Conclusion

This study introduces a novel plant root fluidic device designed for studying flow-induced thigmomorphogenesis in germinating roots. The developed fluidic device, whose hydraulic conductivity conforms to that of the soil, serves as an inert, easy-to-use, and cost-effective platform to analyze individual root samples and enables seamless real-time monitoring of root morphology. Findings reveal that increasing flow rates boost root length and nitrogen uptake until reaching an optimum flow rate, beyond which flow-induced stress reduces root length as aptly confirmed by the results of this endeavor. However, roots in flow conditions consistently outperform those in no-flow conditions due to enhanced nitrogen uptake. The study concludes that nutrient flow induces morphological changes, offering insights into root adaptation and cellular responses. The novelty of this research lies in its exploration of flow-induced thigmomorphogenesis in early root growth, the development of a novel microfluidic system, the investigation of the nitrogen–root growth relationship, the integration of simulation and quantitative analysis, and the high-resolution insights it provides. These contributions collectively advance our understanding of underlying nutrient flow–plant root interactions and have practical implications for agriculture and environmental science. Future studies could explore the molecular mechanisms underlying phenotypic changes related to different flow rates. A deeper understanding of the cellular and molecular processes involved in flow-induced thigmomorphogenesis could inform the design of resilient hydroponic systems and support soil-less crop production.

Data availability

The data that support the findings of this study are available in the manuscript.

Author contributions

K. A.: conceptualization, data curation, visualization, formal analysis methodology, investigation, experimentation, writing – original draft. S. K. M.: conceptualization, formal analysis, visualization, data curation, methodology, writing – original draft, software. P. K. M.: conceptualization, methodology, supervision, writing – review & editing, project administration, funding acquisition, resources.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work is supported by the Science and Engineering Research Board (SERB), Govt. of India, under the project no. CRG/2022/000762. The Ph.D. fellowship of Kaushal Agarwal is supported by the Ministry of Education (MoE), Govt. of India. The authors thank the TIDF (TIH) Indian Institute of Technology, Guwahati, for their active support. The authors are grateful to CIF, IIT Guwahati, for the instrumentation facility. The authors greatly acknowledge Dr. Sudha Gupta, Department of Botany, University of Kalyani, India, for her help with the discussion pertaining to the plant biology. The authors thank the anonymous reviewers for their insightful comments to improve the quality of the manuscript.

References

1. I. Taylor, K. Lehner, E. McCaskey, N. Nirmal, Y. Ozkan-Aydin, M. Murray-Cooper, R. Jain, E. W. Hawkes, P. C. Ronald, D. I. Goldman and P. N. Benfey, Proc. Natl. Acad. Sci. U. S. A., 2021, 118, e2018940118.
2. G. Grossmann, M. Meier, H. N. Cartwright, D. Sosso, S. R. Quake, D. W. Ehrhardt and W. B. Frommer, J. Visualized Exp., 2012, 65, e4290.
3. M. A. Suslov and A. V. Anisimov, Environ. Exp. Bot., 2019, 162, 239–246.
4. J. Lynch, P. Marschner and Z. Rengel, Effect of Internal and External Factors on Root Growth and Development, Elsevier Ltd, 2011.
5. S. Kumari, D. Patra, N. Mishra and K. C. S. Panigrahi, Chapter 16 - Impact of abiotic stress on the root growth and development, ed. A. Kumar and S. B. T.-M. M. of P. S. Droby, Woodhead Publishing, 2021, pp. 235–250.
6. B. L. McMichael and J. E. Quisenberry, Environ. Exp. Bot., 1993, 33, 53–61.
7. M. U. Gul, A. Paul, S. Manimurugan and A. Chehri, Water, 2023, 15, 567.
8. F. A. Abdrakhimov, M. A. Suslov and A. V. Anisimov, Cell Tissue Biol., 2013, 7, 479–486.
9. N. K. Fageria and A. Moreira, Chapter Four - The Role of Mineral Nutrition on Root Growth of Crop Plants, ed. D. L. B. T.-A. in A. Sparks, Academic Press, 2011, vol. 110, pp. 251–331.
10. B. Baiyin, Y. Xiang, J. Hu, K. Tagawa, J. E. Son, S. Yamada and Q. Yang, Int. J. Mol. Sci., 2023, 24, 16616.
11. B. Baiyin, K. Tagawa, M. Yamada, X. Wang, S. Yamada, Y. Shao, P. An, S. Yamamoto and Y. Ibaraki, Plants, 2021, 10, 1840.
12. K. Ötvös, M. Marconi, A. Vega, J. O'Brien, A. Johnson, R. Abualia, L. Antonielli, J. C. Montesinos, Y. Zhang, S. Tan, C. Cuesta, C. Artner, E. Bouguyon, A. Gojon, J. Friml, R. A. Gutiérrez, K. Wabnik and E. Benková, EMBO J., 2021, 40, e106862.
13. D. Qi, T. Hu, X. Song and M. Zhang, Sci. Rep., 2019, 9, 490.
14. R. A. Korver, I. T. Koevoets and C. Testerink, Trends Plant Sci., 2018, 23, 783–793.
15. G. Grossmann, W. J. Guo, D. W. Ehrhardt, W. B. Frommer, R. V. Sit, S. R. Quake and M. Meier, Plant Cell, 2011, 23, 4234–4240.
16. K. Sakai, F. Charlot, T. Le Saux, S. Bonhomme, F. Nogué, J. C. Palauqui and J. Fattaccioli, Plant Methods, 2019, 15, 1–12.
17. H. Massalha, E. Korenblum, S. Malitsky, O. H. Shapiro and A. Aharoni, Proc. Natl. Acad. Sci. U. S. A., 2017, 114, 4549–4554.
18. M. F. Noirot-Gros, S. V. Shinde, C. Akins, J. L. Johnson, S. Zerbs, R. Wilton, K. M. Kemner, P. Noirot and G. Babnigg, Front. Plant Sci., 2020, 11, 505518.
19. D. Mandal, S. Datta, G. Raveendar, P. K. Mondal and R. Nag Chaudhuri, Plant J., 2023, 113, 106–126.
20. M. Ghanbari, A. S. Nezhad, C. G. Agudelo, M. Packirisamy and A. Geitmann, J. Biosci. Bioeng., 2014, 117, 504–511.
21. M. Horade, N. Yanagisawa, Y. Mizuta, T. Higashiyama and H. Arata, Microelectron. Eng., 2014, 118, 25–28.
22. A. Sanati Nezhad, M. Ghanbari, C. G. Agudelo, M. Naghavi, M. Packirisamy, R. B. Bhat and A. Geitmann, Biomed. Microdevices, 2014, 16, 23–33.
23. A. K. Yetisen, L. Jiang, J. R. Cooper, Y. Qin, R. Palanivelu and Y. Zohar, J. Micromech. Microeng., 2011, 21, 054018.
24. M. Meier, E. M. Lucchetta and R. F. Ismagilov, Lab Chip, 2010, 10, 2147–2153.
25. M. L. Dotaniya and V. D. Meena, Proc. Natl. Acad. Sci., India, Sect. B, 2015, 85, 1–12.
26. A. A. Robles-Aguilar, J. Pang, J. A. Postma, S. D. Schrey, H. Lambers and N. D. Jablonowski, Plant Soil, 2019, 434, 65–78.
27. J. López-Bucio, A. Cruz-Ramírez and L. Herrera-Estrella, Curr. Opin. Plant Biol., 2003, 6, 280–287.
28. M. I. Ciochina, UCLA Electronic Theses and Dissertations, Master of Arts, University of California, 2017.
29. A. Cruz-Ramírez, C. Calderón-Vázquez and L. Herrera-Estrella, Root Dev., 2009, 37, 288–324.
30. L. A. Msimbira and D. L. Smith, Front. Sustain. Food Syst., 2020, 4, 106.
31. I. Arduini, C. Kettner, D. L. Godbold, A. Onnis and A. Stefani, Chemosphere, 1998, 36, 733–738.
32. G. G. McNickle and J. F. Cahill, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 4747–4751.
33. W. Chen, R. T. Koide, T. S. Adams, J. L. DeForest, L. Cheng and D. M. Eissenstat, Proc. Natl. Acad. Sci. U. S. A., 2016, 113, 8741–8746.
34. C. Paungfoo-Lonhienne, T. G. A. Lonhienne, D. Rentsch, N. Robinson, M. Christie, R. I. Webb, H. K. Gamage, B. J. Carroll, P. M. Schenk and S. Schmidt, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 4524–4529.
35. M. J. Jaffe, Planta, 1973, 114, 143–157.
36. Q. Wu, Y. Li, M. Lyu, Y. Luo, H. Shi and S. Zhong, Sci. Adv., 2020, 48, eabc9294.
37. I. Saidi, S. Ammar, N. Demont-Caulet, J. Thévenin, C. Lapierre, S. Bouzid and L. Jouanin, Plant Sci., 2009, 177, 1–6.
38. C. Coutand, Plant Sci., 2010, 179, 168–182.
39. E. W. Chehab, E. Eich and J. Braam, J. Exp. Bot., 2009, 60, 43–56.
40. T. P. A. Ferré and A. W. Warrick, in Encyclopedia of Soils in the Environment, Elsevier Inc., 2004, pp. 254–260.
41. K. Mitchell, L. Beesley, V. Šípek and L. Trakal, in Biochar in Agriculture for Achieving Sustainable Development Goals, Elsevier, 2022, pp. 25–33.
42. U. Batool, R. Nawaz, S. Ahmad, M. A. Irshad, A. Irfan, A. R. Z. Gaafar, M. Arshad, G. F. Wondmie, M. M. N. Qayyum and M. Bourhia, Sci. Rep., 2024, 14, 2185.
43. Y. Tian and F. Deng, CYTA J. Food, 2020, 18, 704–718.
44. C. Bassegio, M. A. Campagnolo, D. Schwantes, A. C. Gonçalves Junior, J. Manfrin, A. da P. Schiller and D. Bassegio, Int. J. Phytorem., 2020, 22, 134–139.
45. K. C. Maturi, I. Haq and A. S. Kalamdhad, Environ. Sci. Pollut. Res., 2022, 29, 84600–84615.
46. M. O. Viji, M. M. Mathew and R. Parvatham, Int. J. Curr. Microbiol. Appl. Sci., 2013, 2, 354–360.
47. J. Louvieaux, M. Spanoghe and C. Hermans, Front. Plant Sci., 2020, 11, 568009.
48. M. Kafi and Z. Rahimi, Soil Sci. Plant Nutr., 2011, 57, 341–347.
49. X. Quan, Y. Qiao, M. Chen, Z. Duan and H. Shi, Evol. Ecol., 2021, 11, 12389–12400.
50. L. Bacete, J. Schulz, T. Engelsdorf, Z. Bartosova, L. Vaahtera, G. Yan, J. M. Gerhold, T. Ticha, C. Øvstebø, N. Gigli-Bisceglia, S. Johannessen-Starheim, J. Margueritat, H. Kollist, T. Dehoux, S. A. M. McAdam and T. Hamann, Proc. Natl. Acad. Sci. U. S. A., 2022, 119, e2119258119.
51. S. Choudhury, S. Dutta and S. Chatterjee, Micro Nano Lett., 2019, 14, 860–864.
52. P. Padhi, S. K. Mehta, K. Agarwal and P. K. Mondal, Phys. Fluids, 2024, 36, 043602.
53. S. H. Song, C. K. Lee, T. J. Kim, I. C. Shin, S. C. Jun and H. Il Jung, Microfluid. Nanofluid., 2010, 9, 533–540.
54. A. Victor, J. Ribeiro and F. F. Araújo, J. Mech. Eng. Biomech., 2019, 4, 1–9.
55. A. Raj, P. P. A. Suthanthiraraj and A. K. Sen, Microfluid. Nanofluid., 2018, 22, 128.
56. N. Gavara, Microsc. Res. Tech., 2017, 80, 75–84.
57. N. Kennard, R. Stirling, A. Prashar and E. Lopez-Capel, Agronomy, 2020, 10, 1092.
58. D. L. Gelardi, I. H. Ainuddin, D. A. Rippner, J. E. Patiño, M. A. Najm and S. J. Parikh, Soil, 2021, 7, 811–825.
59. Y. Yan, S. A. Akbar Nakhli, J. Jin, G. Mills, C. S. Willson, D. R. Legates, K. N. Manahiloh and P. T. Imhoff, J. Environ. Manage., 2021, 295, 113143.
60. S. Pabi, M. K. Khan, S. K. Jain, A. K. Sen and A. Raj, Phys. Fluids, 2023, 35, 101908.
61. H. G. Matthies and J. Steindorf, Comput. Struct., 2003, 81, 805–812.
62. A. Raj, M. Dixit, M. Doble and A. K. Sen, Lab Chip, 2017, 17, 3704–3716.
63. A. Peaucelle, J. Visualized Exp., 2014, 89, e51317.
64. T. Zhou, Y. Hua, B. Zhang, X. Zhang, Y. Zhou, L. Shi and F. Xu, Plant Cell Physiol., 2017, 58, 1991–2005.
65. L. Yoo, J. Reed, A. Shin and J. L. Demer, J. Biomech., 2014, 47, 1899–1903.
66. Y. M. Efremov, W. H. Wang, S. D. Hardy, R. L. Geahlen and A. Raman, Sci. Rep., 2017, 7, 1541.
67. S. V. Kontomaris and A. Malamou, Mater. Res. Express, 2020, 3, 033001.
68. S. Bovio, Y. Long and F. Monéger, J. Visualized Exp., 2019, 149, e59674.
69. B. Cappella and G. Dietler, Surf. Sci. Rep., 1999, 34, 1–104.
70. S. Kim, Y. Lee, M. Lee, S. An and S. J. Cho, Nanomaterials, 2021, 11, 1593.
71. J. J. Roa, G. Oncins, J. Diaz, F. Sanz and M. Segarra, Recent Pat. Nanotechnol., 2011, 5, 27–36.
72. I. Potocka and J. Szymanowska-Pulka, Ann. Bot., 2018, 122, 711–723.
73. G. Grafi and S. Barak, Biochim. Biophys. Acta, Gene Regul. Mech., 2015, 1849, 378–384.
74. J. R. Dinneny and M. F. Yanofsky, Curr. Biol., 2004, 14, R112–R114.
75. S. J. Leghari, N. A. Wahocho, G. M. Laghari, A. H. Laghari, G. M. Bhabhan, K. H. Talpur and A. A. Lashari, Adv. Environ. Biol., 2016, 10, 209–218.
76. L. Wang, L. Yang, F. Xiong, X. Nie, C. Li, Y. Xiao and G. Zhou, Molecules, 2020, 25, 381.
77. M. D. Cramer, H. J. Hawkins and G. A. Verboom, Oecologia, 2009, 161, 15–24.
78. J. P. Lynch, Ann. Bot., 2013, 112, 347–357.
79. Z. Ma, D. G. Bielenberg, K. M. Brown and J. P. Lynch, Plant, Cell Environ., 2001, 24, 459–467.
80. H. Kim, J. Jang, S. Seomun, Y. Yoon and G. Jang, Front. Plant Sci., 2022, 13, 953225.
81. H. Cui and P. N. Benfey, Plant Signaling Behav., 2009, 4, 551–553.
82. J. D. Julien, A. Pumir and A. Boudaoud, Bull. Math. Biol., 2019, 81, 3342–3361.
83. O. Leyser, Plant Physiol., 2018, 176, 465–479.
84. N. Nakayama, R. S. Smith, T. Mandel, S. Robinson, S. Kimura, A. Boudaoud and C. Kuhlemeier, Curr. Biol., 2012, 22, 1468–1476.

---

Footnote

† Electronic supplementary information (ESI) available: (1) Details of region of interest (RoI) for anatomical sectioning. (2) 3D printed acrylonitrile butadiene styrene (ABS) mold and casting setup. See DOI: https://doi.org/10.1039/d4lc00180j

---