Regulating Plant Physiology with Organic Electronics

Regulating plant physiology with organic electronics

David J. Poxsona,1, Michal Karadyb,1, Roger Gabrielssona,c,1, Aziz Y. Alkattanc, Anna Gustavssond, Siamsa M. Doyleb, Stéphanie Robertb, Karin Ljungb, Markus Grebed,e, Daniel T. Simona,2, and Magnus Berggrena

aLaboratory of Organic Electronics, Department of Science and Technology, Linköping University, 601 74 Norrköping, Sweden; bUmeå Plant Science Centre, Department of Forest Genetics and Plant Physiology, Swedish University of Agricultural Sciences, 901 83 Umeå, Sweden; cDepartment of Physics, Chemistry and Biology, Linköping University, 581 83 Linköping, Sweden; dUmeå Plant Science Centre, Department of Plant Physiology, Umeå University, 901 87 Umeå, Sweden; and ePlant Physiology, Institute of Biochemistry and Biology, University of Potsdam, 14476 Potsdam, Golm, Germany

Edited by David C. Martin, University of Delaware, Newark, DE, and accepted by Editorial Board Member John A. Rogers March 8, 2017 (received for review October 26, 2016) The organic electronic ion pump (OEIP) provides flow-free and chemical compounds is routinely used for probing plant hor- accurate delivery of small signaling compounds at high spatio- mone biology (13, 14). The applied compounds passively diffuse temporal resolution. To date, the application of OEIPs has been and/or are actively imported by the plant into the target tissues, limited to delivery of nonaromatic molecules to mammalian sys- where their effects can be observed. Commonly used methods tems, particularly for neuroscience applications. However, many include spraying or soaking of the plant (15), as well as applying long-standing questions in plant biology remain unanswered due gels, paraffin, or polymer beads (10, 16) that have been soaked in to a lack of technology that precisely delivers plant hormones, known concentrations of compound or have been allowed to based on cyclic alkanes or aromatic structures, to regulate plant absorb compounds from the plants themselves. For more local- physiology. Here, we report the employment of OEIPs for the ized studies, application of hormone-containing microdroplets delivery of the plant hormone auxin to induce differential concen- via microscope-guided micromanipulators has been demonstrated tration gradients and modulate plant physiology. We fabricated (17). Others have used micro- or nanofluidic systems capable of OEIP devices based on a synthesized dendritic polyelectrolyte that fluidic transport and delivery of a variety of chemical species (18– enables electrophoretic transport of aromatic substances. Delivery 20). Finally, some have turned to nanoscale functional systems for Arabidopsis thaliana of auxin to transgenic seedlings in vivo was directed introduction of materials and molecules within plant cells monitored in real time via dynamic fluorescent auxin-response re- and tissues (21). As with similar techniques for in vitro and in vivo SCIENCES

porters and induced physiological responses in roots. Our results animal studies, these methods all suffer from poor dynamic con- APPLIED PHYSICAL provide a starting point for technologies enabling direct, rapid, and trol, for example in the case of bead or nanoparticle-based de- dynamic electronic interaction with the biochemical regulation sys- livery, or from cumbersome liquid transport that disrupts native tems of plants. concentration gradients or introduces undesirable stresses on cells and tissues. The shortcomings of currently available localized auxin | Arabidopsis thaliana | dendritic polymer | bioelectronics | delivery methods, combined with the cellular-scale effects of auxin polyelectrolyte in particular, point toward an unmet technological need. The

development of a method allowing controlled, localized delivery PLANT BIOLOGY ’ he coordination of plants physiological activity is regulated of hormones and other compounds at the tissue and cellular scale Tby a complex array of chemical signals within and between would thus represent a significant advance for the plant research their cells, tissues, and organs. Although plants do not possess a community. central nervous system, fluxes and gradients of chemical hormone compounds play a central role in the overall management Significance of growth, response to environment, and homeostasis (1, 2). Among the hormones that are generally conserved across the plant kingdom, auxin (indole-3-acetic acid, or IAA) was the first Hormones play a crucial role in the coordination of the physio- discovered, is perhaps the best characterized, and is certainly logical processes within and between the cells and tissues of one of the most crucial (3). Auxin plays an important role in a plants. However, due to a lack of capable technologies, direct and ’ multitude of physiological processes and is involved in many dynamic interactions with plants hormone-signaling systems re- mains limited. Here, we demonstrate the use of an organic elec- aspects of plant development from the single-cell level (endo- — — cytosis and morphogenesis) to macroscopic phenomena (em- tronic device the organic electronic ion pump to deliver the planthormoneauxintothelivingroottissuesofArabidopsis bryogenesis and organ formation). It is understood that the thaliana presence of tightly controlled auxin gradients within cells and seedlings, inducing differential concentration gradients and modulating plant physiology. Electronically regulated trans- tissues is essential for regulating physiology throughout the life port of aromatic structures such as auxin in an organic electronic of the plant (4). Precise regulation of cell-to-cell auxin gradients device was achieved by synthesis of a previously unidentified class and their role in plant development can be found in a variety of of dendritic polyelectrolyte. Such bioelectronic technology opens tissues, such as the base of the developing embryo (5, 6), the the door for precise, electronically mediated control of a plant’s inner apical hook of young seedlings (7), at the tips of the de- growth and development. veloping cotyledons (5, 8), at the primary root tip (9), and at the

primordia of organs such as lateral roots, leaves, and flowers (8). Author contributions: D.J.P., A.G., S.R., K.L., M.G., D.T.S., and M.B. designed research; D.J.P., The cellular scale of auxin activity is clearly demonstrated by the M.K., R.G., A.Y.A., A.G., and S.M.D. performed research; D.J.P., M.K., and S.M.D. analyzed isolated effects of its application on single cells or small cell data; and D.J.P. wrote the paper. groups in certain tissues. For example, auxin application affects The authors declare no conflict of interest. the emergence of root hairs from specific epidermal cells (10) This article is a PNAS Direct Submission. D.C.M. is a guest editor invited by the Editorial + and modulates K channel currents within individual stomatal Board. guard cells (11). As such, deciphering auxin’s molecular and Freely available online through the PNAS open access option. cellular modes of action is of fundamental importance for the 1D.J.P., M.K., and R.G. contributed equally to this work. elucidation of plant biology (12). 2To whom correspondence should be addressed. Email: [email protected]. Researchers have traditionally conducted studies of hormone This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. effects in plants via exogenous application. A wide range of 1073/pnas.1617758114/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1617758114 PNAS | May 2, 2017 | vol. 114 | no. 18 | 4597–4602 Downloaded by guest on September 30, 2021 Fig. 1. De novo design of an OEIP delivering IAA in vitro. Schematic diagrams of (A) OEIP device materials and geometries and (B) conceptualization of the cationic dendrolyte membrane. Anionic species such as IAA are selectively transported and migrate through the ion conducting channel in proportion to the applied potential gradient. (C) Photograph of the fully fabricated OEIP device. (D) Dendritic polyglycerol-based polyelectrolyte system (green) showing cross- linkages (black) and terminal groups (blue) with positive charge group (red). (E) OEIP mounted to a motorized micromanipulator and Arabidopsis seedlings positioned vertically on agar-growth plates. (F) OEIP positioned in proximity to the seedling root apical meristem (AM) and elongation zone (EZ). (G) OEIP delivery tip and root cross-section shown submerged in the agar-growth gel. Electrical current source, voltage meter (V), and electrode arrangement illustrated. Delivery of IAA is pictured as a diffusive concentration gradient from the OEIP delivery tip through the agar gel and exogenous to the root tissue.

In recent years, a range of organic electronic tools has been rate. These device characteristics allow for the precise control of developed (22) that enable precise dynamic delivery of small chemical concentration gradients with high spatial and temporal ionic molecules. The organic electronic ionic pump (OEIP) is resolution. one of these technologies and was developed primarily as an However, the materials used for all previous OEIP-based tech- application for mammalian systems to enable diffusive synapse- nologies pose a significant limitation. Controlled transport through like delivery of neurosignaling compounds (alkali ions and neu- OEIPs and similar “iontronic” devices has been demonstrated only rotransmitters) with high spatiotemporal resolution. Recently, for atomic ions or the smallest of linear molecules (23–26). How- OEIP devices have been demonstrated for a variety of in vitro ever, many biological processes—and bioelectronic application (23, 24) as well as in vivo applications (25), including therapy in scenarios—require transport of larger compounds. The number of awake animals (26). OEIPs are electrophoretic delivery devices available polyelectrolyte materials suitable for OEIP device tech- that leverage the unique ionic and electronic properties of con- nologies is limited. One class of materials—indeed, the ones used ducting polymers and polyelectrolytes to convert electronic sig- in all previous OEIPs—is cross-linked semirandom networks of nals into ionic fluxes. The OEIP’s polymer delivery channel (i.e., linear polyelectrolytes, such as poly(styrenesulfonate) or poly electrophoresis channel) is composed of a polycationic (or pol- (vinylbenzylchloride) (qPVBC) (27). However, such linear poly- yanionic) material with a high density of fixed charge groups that mers have not yet demonstrated the capability to transport larger allows for the selective transport of anions (or cations). The and more rigid molecular compounds, and there exist inherent electrophoretic transport used by OEIP devices is flow-free— challenges for further optimization. only the intended molecules are delivered to the target region, First, it is difficult to synthesize linear polyelectrolytes from not additional liquid or oppositely charged counter ions that may prefunctionalized monomers bearing both cross-linkable and be present in the source solution. The selective electrophoretic ionic groups because any cross-linkable groups also tend to in- transport of the desired molecular species through an OEIP advertently polymerize during polymerization. Second, linear device results in high concentration gradients localized at the polyelectrolytes are challenging to postfunctionalize to a high OEIP outlet (24), on the scale of ∼100 μm to 1 mm. Additionally, degree owing to their immiscibility to most postsynthetic meth- electronic addressing to the OEIP enables the molecular delivery ods. Thus, it is difficult to control key structural characteristics of to be rapidly switched on and off, and, importantly, the electrical linear polymer networks relevant to ion transport properties, driving current can be directly correlated with the ionic delivery namely the size and distribution of fixed charges and void

4598 | www.pnas.org/cgi/doi/10.1073/pnas.1617758114 Poxson et al. Downloaded by guest on September 30, 2021 − averaged IAA delivery rate of 0.45 ± 0.16 pmol·min 1. Using a finite element analysis method (24), based on this measured delivery rate and a basic diffusion model for IAA [that uses the diffusion rate of IAA (34) but neglects potentially biologically relevant parameters such as exogenous uptake and transport within the root], we calculated the expected concentration evo- lution of delivered IAA as a function of distance from the OEIP outlet. This calculation shows that micromolar [IAA] is rapidly established in close proximity to the OEIP delivery tip. Hypo- thetically, plant tissue located 50 μm away from the delivery tip would be exposed to 30 μM IAA after 60 s whereas tissue Fig. 2. OEIP-mediated delivery of IAA. (A) MS measurements of IAA delivered μ μ via OEIP operated continuously at 1 μA, total (summed) IAA vs. time ± SD, located at a distance of 200 m would be exposed to 5 M. − corresponding to an averaged IAA delivery rate of 0.45 ± 0.16 pmol·min 1.(B) Further, it can be observed from the calculation that a near- Calculated IAA delivery concentrations as functions of distance from the OEIP linear concentration gradient across the lateral position of the outlet (x axis) and time (various color lines) using the above measured delivery root is formed within 5–15 min of OEIP operation (Fig. 2B). rate and a basic diffusion model for IAA. The approximate (Approx.) hori- These results indicate that the cationic dendrolyte material sys- zontal position of the root is highlighted in gray. tem is capable of transporting IAA in biologically active quantities (35). Trace amounts of 2-oxindole-3-acetic acid (oxIAA), a known IAA catabolite (36), were also detected during mass fraction, the effective porosity of the bulk, and the degree of spectrometry measurements, typically in concentrations 100– – swelling of the polymer network during hydration (28 30). In- 1,000 times lower than the measured IAA (Fig. S1A). The oxIAA deed, the capability to transport IAA using OEIPs based on the detected was likely formed by nonenzymatic oxidation of IAA polyelectrolyte qPVBC was initially investigated. According to during the OEIP experiments. However, oxIAA has been reported mass spectroscopy analysis, qPVBC-based devices were found to to be inactive in bioassays (36). A deliver only negligible quantities of IAA (Fig. S1 ). Further, as We proceeded to use the dendrolyte-based OEIPs for in vivo described below, similar testing of qPVBC-based OEIPs to de- experiments on a highly accessible model plant system suitable liver IAA to Arabidopsis thaliana plant models was unsuccessful. for live-cell imaging in the intact organism. Specifically, the SCIENCES To address the need for OEIP technologies capable of trans- apical root meristem and early elongation zone of 5-d-old Ara- porting larger ionic compounds, we investigated hyperbranched bidopsis seedlings positioned on agar gel were targeted for de- APPLIED PHYSICAL polymers (31) as the foundation for a previously unidentified livery of IAA via the OEIP. Root tips were monitored using a class of polyelectrolyte materials. Hyperbranched polymers have horizontally oriented spectral macroconfocal laser-scanning mi- generally spherical or globular structures and possess a high num- croscope system schematically illustrated in Fig. 1E. In this ar- ber of terminal functional groups that define their customizable rangement, seedlings were positioned and imaged vertically. physiochemical properties (32). Here, we present a dendritic Using the OEIP devices we targeted the root apical meristem of polyelectrolyte material system using highly branched polyglycerols Arabidopsis seedlings with IAA (Fig. 1 F and G). It is known that

as the base unit, phosphonium chloride as the ionic charge com- IAA can either stimulate or suppress processes such as organ PLANT BIOLOGY “ ” ponent, and allylic groups for cross-linking. These dendrolyte growth in plants, depending on its concentration and the tissue in materials enable the density of ionic and cross-linking groups to be question (4). Root growth was used as a rapidly accessible pa- tuned during synthesis (33) instead of during postfunctionalization. rameter to demonstrate the physiological activity of OEIP- In this way, fundamental limitations of previous OEIPs can be delivered IAA, because it is well established that high IAA addressed: swelling and rigidity of the polymer network can be concentrations inhibit root elongation (35, 37). Additionally, as a controlled by cross-linking, and transport of “larger” or rigid aromatic substances can be facilitated by tuning the void fraction distribution and effective porosity of the bulk. Importantly, dendrolytes enable processing from a “one-pot” three-component miscible mixture of functionalized dendritic polyglycerols, cross- linker, and photoinitiator. One-pot mixtures enable a homoge- neous distribution of bulk charge and cross-linking in the membrane and further offer a high degree of compatibility with a variety of patterning processes such as printing or lithographic techniques (30). Results In this paper we report on the cross-over of molecular delivery technology to plant applications and the capability of transporting aromatic compounds by an OEIP device, enabled by the dendrolyte material system (Fig. 1 B and D). OEIP devices were prepared by photolithographic patterning of the cationically functionalized dendrolyte film (2 μm thick) on a flexible poly- Fig. 3. OEIP-mediated delivery of IAA. (A) Bright-field images of Arabi- ethylene terephthalate (PET) plastic substrate. The shape and dopsis root tips at different time intervals during continuous OEIP delivery of dimensions of the resulting OEIP device structure are illustrated IAA. The position of the OEIP’s 25-μm-wide polyelectrolyte delivery channel and pictured in Fig. 1 A and C. is highlighted in green. Reduction in growth rate is observed during delivery Mass spectrometry was used to quantify the capability of of IAA compared with benzoic acid negative control over the same time dendrolyte-based OEIPs to transport IAA. In this regard, IAA interval. Start and end root tip positions are indicated with blue circles, image area matching Fig. 4A highlighted. (Scale bar, 250 μm.) (B)The played the dual role of biologically relevant plant hormone and growth rate of A. thaliana root tips are plotted as a function of OEIP de- model aromatic substance. The OEIP was operated continuously livery time (averages ± SEM from n = 5 independent treatments are displayed at 1 μA and samples were collected at 15-min intervals for from a time interval of 15 min) of IAA, benzoic acid, and for nontargeted 135 min (Fig. 2A). Under these conditions, OEIPs achieved an control.

Poxson et al. PNAS | May 2, 2017 | vol. 114 | no. 18 | 4599 Downloaded by guest on September 30, 2021 brighter than those on the right. This lateral intensity variation is consistent with the [IAA] gradients calculated from the diffusion model (Fig. 2B) and was observed in most DR5 trials (Fig. S5). Roots targeted with the control molecule benzoic acid did not display alterations in fluorescent intensity of the DR5 reporter. Discussion Using a dendrolyte-based OEIP device, we were able to demonstrate delivery of an aromatic compound—the plant-signaling hormone IAA (auxin)—to a living plant model. Induction of dynamic auxin-response alterations was visualized in near real Fig. 4. Live imaging of auxin delivered via OEIP using Arabidopsis DII-Venus time using two different fluorescent auxin reporters in transgenic seedlings. (A) Confocal fluorescent image sequence of the root tip of DII- Arabidopsis seedlings. With this method, we elicited rapid phys- Venus reporter seedling at intervals 0, 30, and 60 min (benzoic acid control Arabidopsis can be found in Fig. S4). (B) Fluorescence intensity of DII-Venus reporter iological changes in the growth rate of developing seedlings plotted for OEIP delivery of IAA, benzoic acid, and nontargeted roots and observed induced differential lateral [IAA] gradients (Control). Averages ± SEM from n = 5 independent treatments are displayed. across root tissues. (Scale bar, 50 μm.) Images are representative of five roots treated. These results were made possible by the dendrolyte material, a hyperbranched dendritic core-shell polyelectrolyte system that ad- dresses many of the previous limitations of OEIPs and other negative control, benzoic acid (38) was delivered by the OEIP iontronic technologies. The hyperbranched polyglycerol dendrolyte device operated in the same configuration. system, used as the ion transport channel of the OEIP, enables A Fig. 3 shows bright-field images taken of the OEIP device the controlled transport of larger and more rigid ionic com- and seedling root tips at the beginning and after 60 min of de- pounds while overcoming the limited control of important livery of IAA or benzoic acid. Root tip position was measured at polyelectrolyte materials parameters such as porosity, swelling, 15-min intervals and averaged over five trials, and the growth and processability. Specifically, in addition to the natural auxin rate of roots targeted with IAA was compared with benzoic acid Arabidopsis hormone IAA, the capability of the dendrolyte-based OEIP negative control and nontargeted seedlings. For devices to deliver other aromatic substances was also verified seedlings targeted with IAA, a rapid decrease in growth rate was − using the synthetic auxin analog 1-NAA (Fig. S6). observed starting at 15 min of delivery, from 4.7 ± 1.0 μm·min 1 ± μ · −1 Although the dendrolyte system was rationally designed to to 2.4 0.7 m min after 60 min, whereas both benzoic acid address materials parameters such as porosity, swelling, and and nontargeted control seedlings maintained their growth rates B processability, in this study no systematic optimizations were (Fig. 3 ). The reduction in growth rate of plant seedlings by made of the many tailorable parameters of the polyelectrolyte delivery of IAA via the OEIP is consistent with previous findings (i.e., size of the dendritic core shell, length of the cross-linking on exogenous application of IAA (35, 37) (images of IAA and polymer, degree of cross-linking, and degree or type of func- benzoic acid growth rate trials are available in Figs. S2 and S3). tionalization). We fully expect that these parameters will have a To detect, visualize, and monitor IAA delivery in near real ’ Arabidopsis significant role in the polyelectrolyte s transport characteristics time we used two widely used engineered transgenic and corresponding influence on organic electronic device per- lines expressing the semiquantitative 35S::DII-Venus (39) reporter DR5rev::GFP formance. Still, given that the majority of plant hormones such as or (5) marker, both of which show a dynamic abscisic acid, brassinosteroids, gibberellins, and cytokinins are all fluorescence response in the presence of IAA. DII-Venus is a of comparable size and similar cyclic or aromatic molecular negative reporter; IAA causes quenching of the Venus yellow structure, this work provides the foundation for organic elec- fluorescent protein, leading to an inverse relationship between tronic devices that are capable of delivering a wide assortment fluorescence signal and IAA concentration. Conversely, in of biomolecules to directly interact with many fundamental DR5rev::GFP, IAA triggers transcription of new GFP, yielding a chemical signaling systems in plants. direct relationship between fluorescence signal and transcrip- Seedling root growth rate was used as an easily accessible tional response to IAA, which might be correlated to IAA levels. physiological parameter to demonstrate OEIP-mediated delivery The relative IAA abundance is therefore visualized faster and of IAA. Subsequent studies can leverage the larger ion transport more accurately by DII-Venus than by DR5 (38), because the capabilities afforded by the dendrolyte materials for more de- DII-Venus signal relies on a protein degradation mechanism in tailed investigations of the role that auxin plays during many direct correlation with IAA concentration rather than the slower growth and behavioral process such as cellular morphogenesis, transcriptional and translational production mechanisms of DR5. Using Arabidopsis DII-Venus seedlings, we monitored fluorescent signal intensity and observed onset of strong fluorescence reduction between 30–60 min (Fig. 4A). Similar roots targeted with the control molecule benzoic acid preserved their fluorescence (Fig. S4). Quantitative analyses comparing normalized fluorescent intensities of DII-Venus seedlings targeted with IAA or benzoic acid, as well as nontargeted controls, revealed a strong and significant decrease in fluorescence only after IAA delivery via the OEIP (Fig. 4B). In the second experiment we used the dendrolyte-based OEIP to target the elongation zone of DR5rev::GFP reporter seedlings Fig. 5. Live imaging of auxin delivered via OEIP using the Arabidopsis with IAA (Fig. 5A). Confocal images of the root elongation zone DR5rev::GFP reporter line. (A) Elongation zone of DR5rev::GFP seedlings targeted with OEIP, with image area matching B highlighted. (Scale bar, cells revealed the onset of fluorescence in plant tissues after 1 h μ B 250 m.) (B) Confocal fluorescent image sequence of the elongation zone of and the signal continued to increase between 2 and 3 h (Fig. 5 ). DR5rev::GFP reporter seedlings at intervals 0, 1, 2, and 3 h. Image intensities From the image sequence and lateral intensity profile, significant were summed from 16 z-stack layers with 3-μm spacing. Lateral fluorescent variation in the lateral fluorescent intensity of the roots can be intensity across the root elongation zone is summed vertically, normalized, observed—with cells on the left side (OEIP side) of the root being and superimposed. (Scale bar, 50 μm.)

4600 | www.pnas.org/cgi/doi/10.1073/pnas.1617758114 Poxson et al. Downloaded by guest on September 30, 2021 cell elongation and planar polarity, regulation of cytoskeletal were used to monitor the response to IAA delivered via OEIP. To this end, organization, vesicular trafficking, and cell wall formation. Further, seeds of both genotypes and wild-type Col-0 were surface-sterilized by coupling dendrolyte materials with other recent advancements with 70% ethanol for 1 min, incubated in pool cleaner (550 mg/g tri- in OEIP technology, such as multiple addressing points and chloroisocyanuric acid, one tablet per 2,000 mL H2O; Biltema) for 12 min and – washed four times with sterile, distilled water (dH2O). Seeds were plated on rapid on off speeds (40), it will become possible to create more 1/2 Murashige and Skoog (MS) (Duchefa Biochemie), 0.5 g/L MES (Sigma- sophisticated tools to produce complex, electronically controlled Aldrich), 1% sucrose, and 0.7% plant agar (Duchefa) growth medium, hormone concentration gradients with unprecedented spatial and pH 5.6 (120- × 120-mm square Petri-dishes; Gosselin), and vernalized for 3 d temporal resolution. at 4 °C in the dark. The plates were then placed in a growth chamber in OEIP-based technologies were envisioned and developed vertical orientation and the seedlings were grown at 23 °C with 16 h of light primarily for mammalian systems, ultimately as therapeutics for per day. Twelve hours before the start of the experiments, 5 d post- humans. We hope that this study serves as a reminder that chemical germination, seedlings were positioned onto fresh plates of identical MS signaling plays a fundamental role in all biological systems, and such media composition additionally supplemented with 0.01 M KCl. an anthropocentric focus has overlooked many complimentary and potentially important application areas for organic electronics. We Fluorescent Imaging Protocol. A custom reoriented macro confocal laser- scanning microscope with a vertical stage was used to acquire images. The anticipate this technology to be the starting point for precise reg- — — macro confocal consisted of a horizontally placed AZ100 macroscope (Nikon ulation of chemical signaling networks in and between plants and Bergman-Labora) adapted with a specially built XYZ motorized stage and other living systems. (Prior Scientific fitted by Bergman-Labora) and supplied with diascopic white light and episcopic fluorescence light (Nikon). A climate enclosure with Materials and Methods passive humidification was designed to surround the stage area to keep Hyperbranched Dendritic Polyglycerol (Dendrolyte) Synthesis. HyPG [molecular plants in a humid, dark environment. The AZ100 macroscope was connected − weight (Mw) 10,000 g·mol 1, 135 hydroxyl groups per molecule] was pur- to a C2+ confocal laser scanning system (Nikon) equipped with lasers for chased from Nanopartica GmbH. All other chemicals were obtained from 405-nm, 457/488/514-nm, and 561-nm excitation and a transmitted light de- Sigma-Aldrich and used as received. Dimethylformamide (DMF) was dried tector. A coarse manipulator, MM-89 (Narishige), was attached to the stage over 4-Å molecular sieves before use. Reactions were run at room tem- for placing and keeping the OEIP at a specific position. An agar-growth plate perature unless otherwise specified. Equivalents of reagents means from which the lid was removed was placed vertically on the macroconfocal molar equivalents relative to number of –OH groups in the dendrolyte stage and an OEIP loaded with IAA was placed with the delivery outlet in the (10 kDa = 135 –OH groups). growth agar in close proximity of the seedlings. For imaging, a 2× AZ Plan

NMR spectra were recorded on a Varian 300-MHz instrument using Fluor objective (N.A. 0.2, WD 45 mm; Nikon) or a 5× AZ Plan Fluor objective SCIENCES

deuterium oxide (D2O), methanol-d4 (MeOD), or chloroform-d (CDCl3)as (N.A. 0.5, WD 16 mm; Nikon) was used. Excitation was at 488 nm and emission APPLIED PHYSICAL solvent [NMR spectra (41) are available in Fig. S7]. Internal solvent peaks detected with a 525/50 filter. DII-Venus fluorescence intensities from a single were used as reference. Concentrations were performed under diminished confocal image layer were normalized to initial intensity, averaged, and pressure (1–2 kPa) at bath temperatures of 40–60 °C. For purification by compared (standard deviation of the mean). DR5rev::GFP image intensities dialysis, Spectra/Por Regenerative Cellulose (RC) membranes with 3.5-kDa were summed from 16 z-stack layers with 3-μm spacing. The lateral fluores- molecular weight cutoff (MWCO) were used and purchased from Spectrum cent intensity, plotted at the bottom of each image (Fig. 5C), was summed Laboratories. along the y axis and normalized to the maximum intensity of the image se- A fully detailed description of the synthesis and chemical verification can quence (the image sequence for all root trials is shown in Fig. S5). We tested be found in Supporting Information. the OEIP’s ability to deliver the synthetic auxin 1-naphthalene acetic acid (1-NAA) and observed similar dynamic fluorescence quenching in DII-Venus PLANT BIOLOGY Dendrolyte Membrane and OEIP Device Preparation. Circular PET substrates reporter seedlings (Fig. S6). (Policrom screens) with diameter 101.6 mm (4 inch) were washed with acetone and water and subsequently dried at 110 °C for 10 min before they were OEIP Operation. Immediately before the experiments, the OEIPs were oper- −5 treated with 02 plasma (150 W for 60 s). The activated substrates were spin- ated in a target solution of 10 M KCl(aq). A Keithley 2612b SourceMeter coated with a 2-mL solution of 5% (3-glycidyloxypropyl)trimethoxysilane (Keithley Instruments Inc.) and custom LabVIEW (National Instruments Corp.) (GOPS; Alfa Aesar) in water at 500 rpm (Photo Resist Spinner Model 4000; software were used to source current and record voltage. The OEIP device Electronic Micro Systems) for 30 s and allowed to rest in open air for 15 min. was turned on immediately before the first imaging sequence and operated The surfaces were washed with ethanol (EtOH) and dried at 110 °C for at a constant electrical current of 1 μA. For these experiments, the OEIP 10 min. Treated surfaces were spin-coated with 2 mL MeOH stock solution delivery tip was submerged in the MS media in close proximity (100–200 μm) containing 264 mg of dendrolyte material [Compound C in Dendritic Polyglycerol- to the root epidermal tissue (Fig. 3 B and C) and was held at a fixed position Allyl-Butyltrimethylphosphonium Chloride (Compound C)], 18 mg Thiocure 1300 in the growth MS media for the duration of each trial. (Bruno Bock Chemische Fabrik GmbH & Co), and 18 mg Irgacure 2959 (Sigma- Aldrich). UV cross-linking was carried out under nitrogen atmosphere inside a Mass Spectrometry Measurement Protocol. The OEIP reservoir was loaded with −5 glove box and the films were exposed to UV light (254 nm) for 10 min. 80 μL of 10% methanol in dH2O containing IAA at 10 M concentration. The Ion channels were patterned using photolithography of Microposit OEIP outlet tip was submerged in target solution of 50 μL of 1/2 MS media, S1818 photoresist and developed for 60 s in MF319 (both supplied by Shipley). pH 5.7 (Plant Material and Growth Conditions) without agar and containing

Unpatterened, cross-linked dendrolyte material was removed using a CF4 + 0.01 M KCl. Poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT: O2 reactive ion etch (150 W for 90 s) and remaining photoresist was removed PSS) electrodes on a PET substrate (cut from Orgacon F-350 film; AGFA- with acetone. To facilitate ion exchange, patterned wafers were soaked in Gevaert) were used in the source reservoir and the target solution. The OEIP 1 M NaCl(aq) for 5 min. OEIPs were encapsulated with 2× 10-μm bar-coated pump was operated sourcing a constant current of 1 μA using a Keithley 2612b DuPont 5018 UV curing ink. Individual OEIP devices were cut out and SourceMeter and custom LabVIEW software. The target solution was collected packaged in ADW-400 heat-shrink tubing containing sealant glue (Kacab and replaced with fresh solution every 15 min, at time intervals of 15, 30, 45, Teknik AB). The OEIP delivery tips were shaped by hand using a scalpel. 60, 75, 90, 105, 120, and 135 min. In between the analyses, the pump was

To hydrate the dendrolyte channel, OEIP devices were soaked and washed with methanol/dH2O and stored in dH2O. This was repeated five times stored in deionized water before use. Additionally, to reduce the amount with the same OEIP. of unreacted polymers and chemical compounds remaining in the polyelec- To estimate the amount of IAA (and oxIAA) pumped into the target so- trolyte after the above processing steps, OEIP devices were preconditioned lution, IAA quantification was performed according to Novák et al. (42) with by operating the device with 0.1 M KCl(aq) in both the target and source minor modifications. Briefly, 20 μL from each target solution was mixed with

reservoirs. Following the KCl flushing, OEIP devices underwent a loading 500 μLofdH2O and purified by solid phase extraction using hydrophilic– − − phase to exchange the Cl ions in the polyelectrolyte with IAA ;OEIPswere lipophilic balance reversed-phase sorbent columns (Oasis HLB, 1 cc/30 mg; 13 13 operated continuously at 250 nA until steady voltage characteristics were Waters). Before purification, 4 pmol of [ C6]-labeled IAA and 4 pmol of [ C6]- observed (∼12 h). labeled oxIAA were added to each sample as internal standards to validate the quantification. Purified samples were analyzed using an LC-MRM-MS (liquid Plant Material and Growth Conditions. A. thaliana seedlings expressing the chromatography–multiple reaction monitoring–mass spectrometry) system. auxin-responsive fluorescent markers 35S::DII-Venus (39) or DR5rev::GFP (5) The LC-MS system consisted of 1290 Infinity Binary LC System coupled to

Poxson et al. PNAS | May 2, 2017 | vol. 114 | no. 18 | 4601 Downloaded by guest on September 30, 2021 6490 Triple Quad LC/MS System with Jet Stream and Dual Ion Funnel ACKNOWLEDGMENTS. We thank Rishikesh Bhalerao and Henrik Jönsson for technologies (Agilent Technologies). Chromatograms were analyzed using valuable discussions and Ove Nilsson for helping to initiate this collabora- MassHunter software version B.05.02 (Agilent Technologies). A Milli-Q de- tion. We thank Nanopartica GmbH for NMR spectra and analysis of the ionization unit (Millipore) was used for preparation of the purified water for dendritic polyglycerol and the Swedish Metabolomics Centre for the use 13 mobile phases and solutions. The 2-oxo-[indole- C6]-IAA was obtained from of instrumentation. This work was supported by Knut and Alice Wallenberg 13 Olchemim Ltd., and [indole- C6]-IAA was obtained from Cambridge Isotope Foundation ShapeSystems project Grant KAW 2012.0050, with additional Laboratories. All other chromatographic solvents and chemicals were of support from the Swedish Foundation for Strategic Research (Project Grant analytical grade or higher purity from Sigma-Aldrich Chemie GmbH. RMA-11:0104).

+ 1. Jaillais Y, Chory J (2010) Unraveling the paradoxes of plant hormone signaling in- 23. Isaksson J, et al. (2007) Electronic control of Ca2 signalling in neuronal cells using an tegration. Nat Struct Mol Biol 17:642–645. organic electronic ion pump. Nat Mater 6:673–679. 2. Santner A, Estelle M (2009) Recent advances and emerging trends in plant hormone 24. Tybrandt K, et al. (2009) Translating electronic currents to precise acetylcholine- signalling. Nature 459:1071–1078. induced neuronal signaling using an organic electrophoretic delivery device. Adv 3. Zazímalová E, Petrášek J, Benková E, eds (2014) Auxin and Its Role in Plant Mater 21:4442–4446. Development (Springer, Vienna). 25. Simon DT, et al. (2009) Organic electronics for precise delivery of neurotransmitters to 4. Vanneste S, Friml J (2009) Auxin: A trigger for change in plant development. Cell 136: modulate mammalian sensory function. Nat Mater 8:742–746. 1005–1016. 26. Jonsson A, et al. (2015) Therapy using implanted organic bioelectronics. Sci Adv 1: 5. Friml J, et al. (2003) Efflux-dependent auxin gradients establish the apical-basal axis of e1500039. Arabidopsis. Nature 426:147–153. 27. Tybrandt K, Forchheimer R, Berggren M (2012) Logic gates based on ion transistors. 6. Robert HS, et al. (2013) Local auxin sources orient the apical-basal axis in Arabidopsis Nat Commun 3:871. embryos. Curr Biol 23:2506–2512. 28. Gao C, Yan D (2004) Hyperbranched polymers: From synthesis to applications. Prog 7. Friml J, Wisniewska J, Benková E, Mendgen K, Palme K (2002) Lateral relocation of Polym Sci 29:183–275. auxin efflux regulator PIN3 mediates tropism in Arabidopsis. Nature 415:806–809. 29. Geise GM, Cassady HJ, Paul DR, Logan BE, Hickner MA (2014) Specific ion effects on 8. Benková E, et al. (2003) Local, efflux-dependent auxin gradients as a common module membrane potential and the permselectivity of ion exchange membranes. Phys Chem for plant organ formation. Cell 115:591–602. Chem Phys 16:21673–21681. 9. Sabatini S, et al. (1999) An auxin-dependent distal organizer of pattern and polarity in 30. Kariduraganavar MY, Nagarale RK, Kittur AA, Kulkarni SS (2006) Ion-exchange the Arabidopsis root. Cell 99:463–472. membranes: Preparative methods for electrodialysis and fuel cell applications. 10. Fischer U, et al. (2006) Vectorial information for Arabidopsis planar polarity is me- Desalination 197:225–246. diated by combined AUX1, EIN2, and GNOM activity. Curr Biol 16:2143–2149. 31. Schüll C, Frey H (2013) Grafting of hyperbranched polymers: From unusual complex + 11. Blatt MR, Thiel G (1994) K channels of stomatal guard cells: Bimodal control of the polymer topologies to multivalent surface functionalization. Polymer 54:5443–5455. + K inward-rectifier evoked by auxin. Plant J 5:55–68. 32. Astruc D, Boisselier E, Ornelas C (2010) Dendrimers designed for functions: From physical, 12. Sauer M, Robert S, Kleine-Vehn J (2013) Auxin: Simply complicated. J Exp Bot 64: photophysical, and supramolecular properties to applications in sensing, catalysis, mo- 2565–2577. lecular electronics, photonics, and nanomedicine. Chem Rev 110:1857–1959. 13. Hicks GR, Raikhel NV (2012) Small molecules present large opportunities in plant 33. Sunder A, Hanselmann R, Frey H, Mülhaupt R (1999) Controlled synthesis of hyper- biology. Annu Rev Plant Biol 63:261–282. branched polyglycerols by ring-opening multibranching polymerization. Macromolecules 14. Doyle SM, Vain T, Robert S (2015) Small molecules unravel complex interplay between 32:4240–4246. auxin biology and endomembrane trafficking. J Exp Bot 66:4971–4982. 34. Mitchison GJ (1980) The dynamics of auxin transport. Proc R Soc Lond B Biol Sci 209: 15. Cheng Y, Dai X, Zhao Y (2006) Auxin biosynthesis by the YUCCA flavin mono- 489–511. oxygenases controls the formation of floral organs and vascular tissues in Arabi- 35. Rahman A, et al. (2007) Auxin, actin and growth of the Arabidopsis thaliana primary dopsis. Genes Dev 20:1790–1799. root. Plant J 50:514–528. 16. Ikeda Y, et al. (2009) Local auxin biosynthesis modulates gradient-directed planar 36. Pencík A, et al. (2013) Regulation of auxin homeostasis and gradients in Arabidopsis polarity in Arabidopsis. Nat Cell Biol 11:731–738. roots through the formation of the indole-3-acetic acid catabolite 2-oxindole-3-acetic 17. Geisler M, et al. (2003) TWISTED DWARF1, a unique plasma membrane-anchored acid. Plant Cell 25:3858–3870. immunophilin-like protein, interacts with Arabidopsis multidrug resistance-like 37. Grieneisen VA, Xu J, Marée AF, Hogeweg P, Scheres B (2007) Auxin transport is suf- transporters AtPGP1 and AtPGP19. Mol Biol Cell 14:4238–4249. ficient to generate a maximum and gradient guiding root growth. Nature 449: 18. Meier M, Lucchetta EM, Ismagilov RF (2010) Chemical stimulation of the Arabidopsis 1008–1013. thaliana root using multi-laminar flow on a microfluidic chip. Lab Chip 10:2147–2153. 38. Geisler M, Wang B, Zhu J (2014) Auxin transport during root gravitropism: Trans- 19. Grossmann G, et al. (2011) The RootChip: An integrated microfluidic chip for plant porters and techniques. Plant Biol (Stuttg) 16:50–57. science. Plant Cell 23:4234–4240. 39. Brunoud G, et al. (2012) A novel sensor to map auxin response and distribution at 20. Busch W, et al. (2012) A microfluidic device and computational platform for high- high spatio-temporal resolution. Nature 482:103–106. throughput live imaging of gene expression. Nat Methods 9:1101–1106. 40. Jonsson A, Sjöström TA, Tybrandt K, Berggren M, Simon DT (2016) Chemical delivery 21. González-Melendi P, et al. (2008) Nanoparticles as smart treatment-delivery systems array with millisecond neurotransmitter release. Sci Adv 2:e1601340. in plants: Assessment of different techniques of microscopy for their visualization in 41. Gottlieb HE, Kotlyar V, Nudelman A (1997) NMR chemical shifts of common labora- plant tissues. Ann Bot (Lond) 101:187–195. tory solvents as trace impurities. J Org Chem 62:7512–7515. 22. Simon DT, Gabrielsson EO, Tybrandt K, Berggren M (2016) Organic bioelectronics: 42. Novák O, et al. (2012) Tissue-specific profiling of the Arabidopsis thaliana auxin Bridging the signaling gap between biology and technology. Chem Rev 116:13009–13041. metabolome. Plant J 72:523–536.

4602 | www.pnas.org/cgi/doi/10.1073/pnas.1617758114 Poxson et al. Downloaded by guest on September 30, 2021