Engineered magnetic plant biobots for nerve agent removal

Biohybrid micro/nanorobots that integrate biological entities with artificial nanomaterials have shown great potential in the field of biotechnology. However, commonly used physical hybridization approaches can lead to blockages and damage to biological interfaces, impeding the optimal exploitation of natural abilities. Here, we show that magnetically propelled plant biobots (MPBs), employing tomato-callus cultivation engineering in the presence of Fe3O4 nanoparticles (NPs), are capable of active movement and directional guidance under a transversal rotating magnetic field. The Fe3O4 NPs were transported through the cell growth media and then taken up into the plant tissue cells (PTCs), imparting the plant biobot with magnetic function. Moreover, Fe ions support the growth of callus cells, resulting in nanoparticle incorporation and enabling faster growth and structurally compact texture. The magnetic plant biobots demonstrated rapid and efficient removal of chlorpyrifos (approximately 80%), a hazardous nerve gas agent that causes severe acute toxicity, and recovery using an external magnetic field. The eco-friendly plant biobots described here demonstrate their potential in biomedical and environmental applications. Biohybrid micro/nanorobots that integrate biological entities with artificial nanomaterials have shown great potential in the field of biotechnology. Here, the magnetic plant biobots (MPBs) were fabricated by employing tomato-callus cultivation engineering in the presence of Fe3O4 nanoparticles that are capable of active movement and directional guidance under a transversal rotating magnetic field. The MPBs enabled rapid and efficient removal of chlorpyrifos (approximately 80%), a hazardous nerve gas agent that causes severe acute toxicity, and recovery using an external magnetic field. The eco-friendly plant biobots described here demonstrate their potential in environmental applications.


Introduction
Artificial self-propelled micro/nanorobots are able to perform complex tasks in a wide range of fields 1 , such as drug delivery [2][3][4] , biosensing 3,5 , and cell manipulation 6,7 , as well as numerous water remediation tasks [8][9][10] . Recently, micro/nanorobots have been studied using various mechanisms and external stimuli, such as light, ultrasound, and magnetic fields 11 . Artificial magnetic micro/ nanorobots based on ferromagnetic materials can be propelled by magnetic field, thus providing enhanced functionality such as moving along a driven trajectory [12][13][14] . Combining self-propulsion with the intrinsic properties of materials, micro/nanorobots can be used for the rapid removal of contaminants such as heavy metal ions and toxic organic compounds 8,14,15 .
Organophosphorus (OP) pesticides such as chlorpyrifos, parathion, and malathion are widely used in agriculture to prevent farm products from being damaged by insects 16,17 . Only a small amount of these highly toxic OP pesticides reach their target 18 and the residues accumulate in the environment, thus creating a threat not only to the surrounding environment but also to human and animal health. If accumulated in the human body, chlorpyrifos, one of the most widely used pesticides, can cause dermal irritation, subtle nervous system effects, respiratory distress, and death 19,20 . In addition, the accumulation of pesticides affects the soil half-life and the metabolism of soil microorganisms 21 and has been revealed as one of the main causes of water pollution 22,23 . To reduce damage by chlorpyrifos, a variety of removal and detection 5 methods have been studied.
Magnetically driven micro/nanorobots have been used in biomedical applications and environmental remediation and present advantages such as biocompatibility, remote maneuverability, and versatility and they do not require fuel. Recently, based on these merits, biohybrid micro/nanorobots have been developed in combination with ferromagnetic nanoparticles (NPs)/structures and natural materials (e.g., cells, bacteria, sperm, pollen, plants, and spores) [24][25][26][27][28][29] . Biohybrid micro/nanorobots can be operated by diverse propulsion methods, including as chemical, magnetic, and ultrasound techniques 30 . Biohybrid micro/nanorobots that can be propelled by toxic fuels, such as hydrogen peroxide 31 , hydrazine 32 , and sodium borohydride 33 , are limited in their practical applications. Hence, the fabrication of plant-derived biobots combined with plant tissue and ferromagnetic materials is a new direction for biohybrid robots.
In this study, tomato leaf-derived calli were used to fabricate plant robots, and their efficiency of chlorpyrifos removal was investigated. Engineered plant calli were cultivated with Fe 3 O 4 NPs and induced magnetic plant biobots (Scheme 1 and Scheme S1). Because the Fe 3 O 4 NPs were internalized inside the magnetic plant biobots, it was possible to propel and control them under a transversal rotating magnetic field. The localization of Fe 3 O 4 NPs inside the magnetic plant biobots was assessed using scanning electron microscopy (SEM), elemental analysis by energy-dispersive X-ray (EDX), and dark-field microscopy with hyperspectral imaging to obtain further information on Fe 3 O 4 NPs within the callus. We demonstrated that the magnetic plant biobots are propelled under magnetic fields without needing toxic chemicals as fuel. In addition, the usability of the magnetic plant biobots was verified by effectively removing chlorpyrifos from the solution. This study focuses on the development of plant-derived biobots and suggests their application for pesticide (chlorpyrifos) removal. The plant-based robots reported here can be propelled under a magnetic field and can efficiently remove pesticides while in motion. In the same vein, magnetic plant biobots can open a new methodology for biohybrid active materials for biomedical and environmental applications.

Results and discussion
Initially, magnetic plant biobots were prepared and characterized. Tomato callus was cultivated on Murashige and Skoog (MS) media 34,35 supplemented with Fe 3 O 4 NPs (2 mg/ml) and growth nutrients (6-benzyl aminopurine (BAP) and α-naphthaleneacetic acid (NAA)) as detailed in the Materials and Methods. Tomato callus growth was examined visually for size, pigment, and different concentrations of agar. As shown in Fig. 1A, during 14 days of culture, there was no significant difference in pigmentation and viability of callus cultured with and without Fe 3 O 4 NPs. In addition, Fig. S1 shows a comparison of representative digital photographs of callus with and without treatment with Fe 3 O 4 NPs. The Fe 3 O 4 NP-free callus was brittle compared to the compact texture of the callus cultured with media including Fe 3 O 4 NPs 36,37 . These phenomena could be explained by Fe deficiency influencing callus growth during the cultivation period. Fe is an essential element for metabolic reactions in organisms. Fe deficiency in plant callus correlates with its weight, chlorophyll and carotenoid contents, antioxidant enzyme activities, and lipid peroxidation 37 . As proposed in numerous studies, Fe deficiency of the callus induces stress, making it brittle and inhibiting growth, indicating that Fe may be involved in the cell wall structure as observed in the texture of the exposed callus. Fe deficiency also leads to high plant mortality with significant growth inhibition (including yellowing and symptoms of necrosis) [37][38][39][40] .
In addition, individual callus growth is also affected by the concentration of agar. To optimize the solutions to create magnetic plant biobots, callus was cultivated with Fe 3 O 4 NPs and various concentrations of agar (0.4%, 0.6%, 0.8%, and 1.2%). The growth responses of the calli are presented in Fig. 1, Supplementary Fig. S2A-D, and Table  S1. As shown in Fig. 1, the average size of the plant calli showed no significant difference with or without Fe 3 O 4 NPs in the first 7 days, whereas a remarkable each callus size increase was observed in the Fe 3 O 4 NP-treated groups after 14 days. After 28 days of cultivation, the mean growth area of callus with Fe 3 O 4 NPs (460% ± 95.1 growth compared to the first day) was much larger than that without Fe 3 O 4 NPs (328% ± 53.2) (Fig. 1B). The difference in the average size of the calli with and without Fe 3 O 4 NPs was~132%. In addition, to examine the variation in callus growth according to the concentration of agar, calli were cultured on 0.4%, 0.6%, and 1.2% agar plates (Supplementary Fig. S2A-C). The sizes of these calli samples are reported in Table S1. The largest difference in the growth of calli with and without Fe 3 O 4 NPs were observed on the 0.6% agar plate, with a difference of 152%. However, the greatest calli growth were observed on the 0.8% agar plate with Fe 3 O 4 NPs (460%) after 28 days (Fig. S2D). Therefore, 0.8% was selected as the optimal agar plate concentration for this study. A similar trend of growth promotion is also observed in Supplementary Fig. S2, suggesting that callus growth was affected by the concentration of agar as well as by the Fe 3 O 4 NPs. The optimum concentration and type of agar are important factors in callus growth to avoid the vitrification of callus 41 and decrease the necrosis of plant tissue cells 42 , which occur during in vitro plant cell proliferation 43 . In our experiment, the largest and most rigid calli were cultivated on a 0.8% agar plate, whereas calli cultivated on 0.4% and 1.2% agar were unstable and physically brittle, which is in agreement with previous reports 43,44 . Therefore, 0.8% agar was chosen as the optimum concentration for further MPB fabrication. In this study, we confirmed that the appropriate concentration of agar with Fe 3 O 4 NPs allowed the magnetic plant biobots to maintain a compact structure and promote the growth of callus for a long-term cultivation period. After 28 days of callus culture in normal media and magnetic plant biobot induction media, distinct differences between plant callus and magnetic plant biobots were observed at both macroscopic and microscopic levels.
Tomato callus cells and magnetic plant biobots were analyzed using SEM. As shown in Fig. 2A, B, the magnetic plant biobots maintained a compact structure compared to the plant callus (without Fe 3 O 4 NPs), even after processing by the homogenizer. The existence of Fe 3 O 4 NPs in magnetic plant biobots was further confirmed by EDX spectroscopy. EDX elemental mapping confirmed the presence of carbon (C), oxygen (O), and iron (Fe) in the structure of the magnetic plant biobots (Fig. 2C, D and Supplementary Fig. S3). Furthermore, callus cultivated with Fe 3 O 4 showed more than 10 times the Fe content in EDX spectroscopy ( Supplementary Fig. S3). To further demonstrate the differences in Fe content, inductively coupled plasma-optical emission spectrometry (ICP-OES) was performed ( Supplementary Fig. S4). ICP-OES analysis demonstrated that the Fe content in the plant biobots was 0.135 mg/g. However, in plant callus, the content of Fe was 0.021 mg/g. These results indicate that Fe 3 O 4 NPs were internalized by plant callus. It should be noted that the Fe content in plant callus is related to the intrinsic Fe that plants need for many biochemical processes from their natural metabolism [36][37][38] .
Hyperspectral imaging combined with dark-field microscopy facilitates the detection and visualization of NPs in biological and environmental fields 45 . To establish the internalization of Fe 3 O 4 NPs in cultivated callus, darkfield microscopy with hyperspectral imaging was used to obtain images of callus grown for 28 days in the presence and absence of Fe 3 O 4 NPs (Fig. S5A, B). Using hyperspectral imaging, it is easy to distinguish plant cells (green), tissue (no signal), and Fe 3 O 4 NPs (yellow). Furthermore, as shown in spectral libraries, it was confirmed that cells and tissues show absorbance at 400-500 nm, and Fe 3 O 4 NPs show absorbance at 700 nm or less. As Fe 3 O 4 NPs absorb wavelengths higher than those absorbed by plant cells, this also confirms the presence of Fe 3 O 4 NPs. In addition, the morphology and size of the Fe 3 O 4 NPs used in callus culture were characterized by transmission electron microscopy (TEM) and EDX spectroscopy (Fig. S6A, B) (14 nm to 665 nm) were exposed to plant roots for 7 days, NPs of less than 140 nm were found in the wheat roots, and NPs of less than 36 nm were found in the parenchyma. Therefore, the uptake of NPs is dependent on their size and the direct interaction between the cell wall components and TiO 2 NPs 51 . In addition, T. Sabo-Attwood et al. also revealed the internalization of NPs within individual plant cells 52 . The results showed that the NPs migrated across the plant cell wall and biomembrane barrier through the presence of clusters in the cytoplasm. Thus, several factors that determine the interaction between NPs and plants have been reported; however, the To evaluate the effect of Fe 3 O 4 NPs on callus growth over the culture period, we double-stained cells with H 2 DCFDA and Hoechst 33342 and observed the emitted green and blue fluorescence, respectively (Fig. 3). Reactive oxygen species (ROS) are known to have an important role in intracellular mechanisms, such as the delivery of cell signals, apoptosis, and gene expression, in addition to being a stress marker. H 2 DCFDA penetrates the cell membrane and infiltrates the cell 53,54 . Here, to examine the nucleic acids, the plant cells were dyed with Hoechst 33342. We observed that both control and treated cells showed a similar pattern and that MPBs can be used for the removal of chlorpyrifos. Furthermore, the samples were analyzed spectrophotometrically for basic stress markers such as carotenoids, total polyphenolics, flavonoids, and antioxidant capacity expressed as EC50 (Fig.  S7A-D). In all performed analyses, no significant difference in the content of individual compounds was found. These compounds respond mainly to stress conditions, ROS levels, etc. Under stress conditions, there is an increase in ROS and, thus, an increase in antioxidant compounds. However, as the data show, the samples exposed to Fe 3 O 4 NPs did not demonstrate any increase in the abovementioned parameters. These results are consistent with the microscopic analysis, where we observed ROS after H 2 DCFDA treatment. Both analyses suggest steady-state and balanced cellular homeostasis in both control and treated cells. Data from samples exposed to NPs show a larger error. This is likely due to the individual NPs not being equally distributed in each cell.
These results indicate that the Fe 3 O 4 NPs have no negative impact on intra-and extracellular callus.
In addition, the internalization of Fe 3 O 4 NPs by MPBs was demonstrated through magnetic actuation and directional control under a transversal rotating magnetic field. Plants absorb Fe ions from the soil 55,56 . However, magnetic motors need ferromagnetic or paramagnetic material to be propelled by the magnetic field 29,57 . Propelled MPBs were evaluated under a transverssal rotating magnetic field using six electromagnetic coils 6 . Video S1 shows the motion of the plant callus and MPBs controlled by magnetic manipulation at a frequency of 2 Hz and a magnetic field intensity of 5 mT in an aqueous solution. Furthermore, Fig. 4 shows the time-lapse tracking line images of plant callus (Fig. 4A) and MPBs (Fig. 4B) in distilled water. The MPBs were sufficiently magnetized to allow rotation with a tumbling motion along the X-Z plane and precise maneuverability under the transversal rotating magnetic field (Video S1). In addition, the controlled guidance of MPBs under rotating magnetic fields was demonstrated by presenting a square propelling path (Video S2). The average velocity of MPBs was 6.023 μm s −1 at 2 Hz with 5 mT. However, beyond 2 Hz, MPBs exhibited an unstable state with shaking (Video S3). By integrating biocompatible iron oxides with plant tissue cells, the use of toxic chemical fuels, such as H 2 O 2 , that potentially can cause damage to weak biological barriers can be avoided.
The comparison of magnetic responsiveness under the transversal rotating magnetic field shown in Video S2 to S3 is direct evidence of Fe 3 O 4 NP internalization. When we consider the scale of the callus and NPs (500 µm: 150 nm), this movement likely requires the internalization of many Fe 3 O 4 NPs, not just a few. Electron/optical microscopic data support this observation (Figs. 2, S5, and S6). Moreover, the EDX elemental mapping from SEM images (Fig. 2) and ICP-OES analysis (Fig. S4) clearly show the uptake of many Fe 3 O 4 NPs. In EDX images, the green area that corresponds to the Fe element is more than 25 µm in length.
After the efficient magnetic propulsion of the biobots was demonstrated, nerve agent (chlorpyrifos) removal was performed via a magnetic manipulation, and the performance of the magnetic plant biobots was compared to that of the plant callus without Fe 3 O 4 NPs (static mode). We demonstrated chlorpyrifos removal by PTCs and MPBs using programmed random mode propulsion, and the amount of chlorpyrifos was quantified by the calibration curves shown in Fig. S8A-C. Figure 5 shows that chlorpyrifos was efficiently removed by the actively moved magnetic plant biobots as a function of time. According to Fig. 5A, C, the final residual chlorpyrifos after exposure to plant callus and magnetic plant biobots was measured to be 3 ppm and 0.3 ppm, respectively. With the activation of MPBs, the chlorpyrifos removal efficiency reached >60% within only 30 min (Fig. 5B). Most of the removal was achieved after 180 min, when the removal efficiency was 80% and 64% for the magnetic plant biobots and plant callus, respectively, suggesting that the performance of dynamic MPBs was~20% more efficient than that of static plant callus (Fig. 5D). The main advantage of our motors, apart from functioning as mobile microcleaners, is the possibility of extracting the pesticide by the retrieval of MPBs using an external magnet (see Video S4), which we cannot do with simple tomato cells. Moreover, the MPBs were readily separated by an external permanent magnet, which can be utilized for the recovery of the MPBs after chlorpyrifos removal (Video S4 and Fig. S9A, B). Consequently, by exploiting the nature of plant callus cells and dynamic motion along with magnetically guided control and facile recovery, we demonstrated for the first time the ability of high-performance plant biobots to successfully remove chlorpyrifos from plant cells.
The uptake of pesticides (i.e., organic chemicals) into plants could be facilitated through three major pathways: dispersal into various parts of plants through root uptake, absorption from the surrounding atmosphere, or diffusion through the plant surface due to contaminants deposited on the plant epidermis 58,59 . Pesticides absorbed from the soil can be localized within plant tissue or dispersed to other plant regions through vascular bundles. Previous studies of the uptake of organic chemicals into plants have shown that these chemicals flow into plants mainly through passive and diffusive processes 60 . Translocation of pesticides within the plants penetrates multiple layers, including the epidermis, cortex, endodermis, and pericycle, through the xylem. This flow is operated by the water potential gradient created throughout the plant 61 . Therefore, the plant uptake process can depend on the enzyme activity in specific plant tissues and the environmental variables.

Conclusions
In this work, the removal of the pesticide chlorpyrifos was demonstrated using plant-based robots. To produce magnetic plant biobots, we cultivated plant calli on media containing ferromagnetic material (Fe 3 O 4 NPs). The Fe 3 O 4 NPs were taken up inside the plant cells during their growth. Moreover, Fe 3 O 4 NPs were not toxic to plant calli, providing structural stability and enhancing callus growth. The plant-based robots had a growth rate of 460% and excellent removal efficiency of 80% for chlorpyrifos. The great advantage of these magnetic plant biobots is their simple preparation. This procedure can be easily scaled up. These results suggest that magnetic plant biobots can be effectively used not only for pesticide removal but also for heavy metal removal. Furthermore, magnetic plant biobots are excellent candidates to remediate polluted water.

Seed germination and callus induction
Several leaves from well-grown tomatoes were carefully cut to an approximately uniform shape using a sterile scalpel blade. For surface sterilization, leaf explants were immersed in 70% ethanol for 1 min and 2.5% (v/v) sodium hypochlorite for 5 min and then washed three times with distilled water. Seeds of tomato thus prepared were germinated in 1/2 MS media consisting of 1% sucrose and 0.7% agar on a petri dish at 22-24°C under dark conditions. For callus induction, germinated seeds were inoculated in semisolid MS media with phytohormones such as NAA and BAP, and callus was obtained after a few weeks. NPs (2 mg/mL) were mixed with MS media by an ultrasound probe. Both media were solidified with four different concentrations of agar (0.4%, 0.6%, 0.8%, and 1.2%). After the explants were cultivated in two different types of media (with and without Fe 3 O 4 NPs), their size and morphology were recorded weekly, and the data were analyzed using a digital camera and ImageJ software.

Characterization of PTC
Tomato callus cells and magnetic plant biobots were examined by SEM (MAIA3 Tescan) with affiliated EDX (Oxford) spectroscopy and hyperspectral microscopy (CytoViva Inc., USA) analysis. For analysis, the callus samples were fixed in 5% glutaraldehyde in 0.1 M PBS (pH 7.2) for 2 h at room temperature. Samples were washed with distilled water and dehydrated in a graded ethanol series (40%, 50%, 60%, 70%, 80%, and 90%) for 15 min each and then placed in 100% ethanol for 10 min. Finally, the samples were completely dried. To compare the Fe content of the samples, inductively coupled plasmaoptical emission spectrometry (ICP-OES, Spectro Arcos) was used.

Plant tissue cell compatibility with Fe 3 O 4 NPs
To evaluate the effects of Fe 3 O 4 NPs on plant callus and magnetic plant biobots, samples were treated with H 2 DCFDA and Hoechst 33342. Each sample was washed with PBS before adding the working solution. The ROS levels in the PTCs were evaluated by staining with the 10 μM fluorescent probe 2'7'-dichlorodihydrofluorescein diacetate (H 2 DCFDA) and incubating at 37°C in the dark. For nuclear staining, Hoechst 33342 (10 μg/mL) was added before the end of the reaction time of H 2 DCFDA. Samples were then washed twice before imaging by confocal microscopy (ZEISS LSM 880). H 2 DCFDA fluorescence was determined at Ex 492 /Em 525 . The resolution was set to 792 × 792. 40× magnification was used for visualization.

Spectrophotometric analysis
Determinations of phenolic compounds, flavonoids, carotenoids, and antioxidant activity were based on a previous experiment 62 . Samples were homogenized at 15,000 rcf (Precellys Evolution Homogenizer, Bertin-Instruments, France) and then measured by a spectrophotometer (Infinite M200 Pro, Tecan, Männedorf, Switzerland with Tecan i-control software, 1.9).

Motion studies of magnetic plant biobots
Motion studies of magnetic plant biobots were conducted on rotating magnetic manipulation systems. The magnetic plant biobots were mixed well with distilled water and dropcast onto a glass slide. The samples were recorded as 50-second videos at 30 frames per second.
The movement of the magnetic plant biobots was observed by optical microscopy using a high-resolution camera, and the videos were analyzed through a computer system.

Chlorpyrifos removal
To evaluate the pesticide removal efficiency of the magnetic plant biobots, chlorpyrifos was employed, which is the most commonly used pesticide in agriculture. Briefly, part of the isolated plant callus and magnetic plant biobots were added to 10 ppm chlorpyrifos solution under a rotating magnetic field at room temperature for 3 h. The supernatant of the working solution was collected at constant time intervals by centrifugation and measured by absorbance from 200 to 800 nm using UV-Vis spectroscopy. The removal efficiency of chlorpyrifos was calculated as follows: where C 0 is the initial concentration of the chlorpyrifos solution and C 1 is the concentration of the chlorpyrifos solution after removal. The concentrations of chlorpyrifos and callus were also determined by UV-Vis spectroscopy.

Statistical analysis
The software R 3.3.2 and RStudio Desktop for Windows (RStudio Team (2020)) were used in the analysis. RStudio: Integrated Development for R (RStudio, PBC, Boston, MA, URL http://www.rstudio.com/) was used for statistical calculations. Mean values for individual treatments were compared with mean values of unexposed controls using the post hoc Tukey test. Mean values (n = 3). The results are expressed as the mean ± standard deviation.