Mode of Action of the Natural Product Allicin in a Plant Model: Influence on the Cytoskeleton and Subsequent Shift in Auxin Localization

Abstract

Allicin is a defense substance produced by garlic cells when they are injured. It is a redox-active thiosulfinate showing redox-activity with a broad range of dose-dependent antimicrobial and biocidal activity. It is known that allicin efficiently oxidizes thiol-groups, and it has been described as a redox toxin because it alters the redox homeostasis in cells and triggers oxidative stress responses. Allicin can therefore be used as a model substance to investigate the action of thiol-specific prooxidants. In order to learn more about the effect of allicin on plants, we used pure synthetized allicin, and studied the influence of allicin on organelle movement in Tradescantia fluminensis as a cytoskeleton-dependent process. Furthermore, we investigated cytoplasmic streaming in sterile filaments of Tradescantia fluminensis, organelle movement using transgenic Arabidopsis with organelle-specifics GFP-tags, and effects on actin and tubulin in the cytoskeleton using GFP-tagged lines. Tubulin and actin were visualized by GFP-tagging in transgenic lines of Arabidopsis thaliana to visualize the influence of allicin on the cytoskeleton. Since auxin transport depends on recycling and turnover of the PIN protein involving cytoskeletal transport to and from the membrane localization sites, auxin distribution in roots was investigated using of transgenic PIN1–GFP, PIN3–GFP, DR5–GFP and DII–VENUS Arabidopsis reporter lines. Allicin inhibited cytoplasmic streaming in T. fluminensis, organelle movement of peroxi-somesperoxisomes, and the Golgi apparatus in a concentration-dependent manner. It also destroyed the correct root tip distribution of auxin, which probably contributed to the observed inhibition of root growth. These observations of the disruption of cytoskeleton-dependent transport processes in plant cells add a new facet to the mechanism of action of allicin as a redox toxin in cells.

1. Introduction

Actin and tubulin are major structural components of the cytoskeleton, and their polymerization and depolymerization are important, amongst other things, for cytoplasmic streaming and intracellular trafficking [1]. Soon after the discovery of actin, it was shown that oxidizing agents not only inhibited the polymerization of globular actin, but also depolymerized actin filaments [2]. The importance of sulfhydryl groups for protoplasmic streaming in plant cells was reported in 1964 by Abe [3] who showed the effect of the specific thiol-trapping reagent p-chloromercuribenzoate in bringing cytoplasmic streaming to a standstill in the internodal cells of Nitella flexilis, the leaf cells of Elodea densa, the root hair cells of Hydrocharis morsus ranae, and the stamen hair cells of Tradescantia reflexa. Elements of the cytoskeleton are known to be highly sensitive to oxidation [4,5], and redox-mediated posttranslational modifications of the cytoskeleton affect polymerization [6,7]. The redox regulation of actin and tubulin dynamics is now a well characterized phenomenon [8,9,10].

Allicin from garlic is the most prominent sulfur-containing bioactive compound in freshly damaged garlic tissue [11,12]. It oxidizes thiol-groups in a thiol-disulfide-exchange-like manner, leading to S-thioallylated adducts [13,14,15,16,17]. Since allicin is a strongly antimicrobial compound with a broad-spectrum of activity, it has the potential for application in agricultural pest control [18,19,20,21,22]. Allicin‘s degradation products, primarily polysulfanes, ajoenes, and vinyldithiins, can also show biological activity and can be active as antimicrobial components. The effect of these sulfur-containing components in allelopathic interactions with other organisms can be of importance, as is observed in particular for wild garlic (Allium ursinum L.), to which is attributed the exudation of sulfur-containing secondary metabolites [23].

In animal cells, it was shown that allicin, which acts as a “redox-toxin” [13], affects the integrity and function of the cytoskeleton [24]. It was reported that 0.5 µM of allicin preferentially affected the tubulin cytoskeleton, while the actin cytoskeleton was a minor target (ibid.). However, at a higher 25 µM concentration of allicin, the actin cytoskeleton was also reported to be disrupted [25]. Furthermore, 16 cytoskeletal proteins were shown to be S-thioallylated by 100 µM of allicin in human Jurkat cells within 10 min of exposure [26].

The actin cytoskeleton is of particular importance for the correct polar transport of auxin in relation to PIN efflux carrier recycling [27,28,29]. Auxin also exerts an influence on the regulation of actin turnover [30]. The high redox sensitivity of actin is probably significant for the observation that auxin transport is related to the state of the total cellular glutathione pool, which is the most important thiol-based redox buffer of the cell [31]. Thus, treatment of Arabidopsis thaliana seedlings with buthionine sulphoximine (BSO), a specific inhibitor of glutathione biosynthesis, led to a clear concentration-dependent inhibition of root growth and a loss of the auxin efflux carriers PIN1, PIN2, and PIN7, as well as the consequent destruction of the correct auxin distribution in the root tip (ibid.). This observation correlates with the finding that mutants in glutathione biosynthesis formed shortened roots, such as was seen in the weak allele mutants pad2 and cad2 of the Arabidopsis thaliana GSH1 gene, which had reduced glutathione levels [32]. Mutants of the strong rml1 (root meristemless) allele of GSH1 were almost unable to produce glutathione, could not maintain the root apical meristem, and consequently could not form a root [33,34].

Due to the need to understand the phytotoxicity of allicin with regard to its possible use in organic farming, as well as the allelopathic effects of these substances, we are interested in understanding the molecular causes of the effects of allicin on plants. It has already been reported that allicin inhibits Arabidopsis thaliana primary and lateral root growth in a concentration-dependent manner [35,36]. In the work reported here, we analyzed the effect of allicin, applied as a solution of chemically synthesized pure allicin, on the cytoskeleton and consequent cytoskeleton-dependent cellular processes like cytoplasmic streaming, organelle movement, and auxin distribution, which may have led to the root growth-inhibiting effects of allicin.

2. Materials and Methods

2.1. Plant material and cultivation

Tradescantia fluminensis VELL was cultivated in the greenhouse. For the induction of flowering, the plants were grown in nutrient-poor soil.

Mutants and different GFP-tagged Arabidopsis thaliana lines used in this study are listed in

Table 1. Mutants and transgenic Arabidopsis thaliana lines used in this study.

Mutant	Phenotype	Reference
ABD2–GFP	GFP-fusion protein with actin binding protein ABD2 allows visualization of actin cytoskeleton	[37]
Tubulin–GFP	GFP-fusion with beta-tubulin allows visualization of tubulin cytoskeleton	[38]
GFP–PTS	Fusion of GFP with peroxisomal targeting sequence allows visualization of peroxisomes	[39]
pad2	weak GSH1 allele, homozygote has ~30% GSH content of the wild type	[40]
axr1-12	No auxin perception	[41]
aux1-2	Lack of auxin importer AUX1-2.	[41]
DR5–GFP	GFP expression dependent on the auxin-responsive DR5 promoter.	[42]
DII–VENUS	Degradation of the fluorescent protein depending on the presence of auxin	[43]

. The plants were grown in solid Hoagland-medium (1 mM CaNO₃; 5.1 mM KNO₃; 0.5 mM MgSO₄; 0.5 mM MgCl₂; 0.13 mM (NH₄)H₂PO₄; 30 µM NH₃NO₃; 31 µM NaOH; 22 µM EDTA-Iron-Sodium Salt; 9.7 µM BH₃O₃; 22 µM FeSO₄; 2 µM MnSO₄; 0.31 µM ZnCl₂, 0.21 µM CuSO₄; 0.14 µM Na₂MoO₄; 86 nM Co(NO₃)₂, containing 8 g/L of plant agar (Duchefa, Haarlem, The Netherlands). After stratification for two days at 4 °C, the seedlings were grown under long day conditions (16 h light, 8 h dark; 22 °C) for two days before treatment.

2.2. Synthesis of Allicin and Quantification of Allicin

The chemical synthesis of allicin was conducted as described previously [13]. The purity of the chemically synthesized allicin was >95% (ibid.).

2.3. Measurement of Protoplasmic Streaming

As a measure of the influence of allicin on protoplastic flow, the time was recorded for cytoplasmic flow to cease following exposure to allicin. For this purpose, individual sterile flower filaments were plucked off with tweezers and placed under a microscope as a water preparation. To prevent the specimen from being crushed, the cover glass was slightly raised by small amounts of plasticine at the corners. Treatment with allicin or garlic juice was carried out by using a filter paper to draw off the water under the coverslip and replace it with allicin solutions of the appropriate concentration. In the same way, the allicin solution was exchanged for distilled water for the regeneration experiments. Microscopy was carried out at 400× magnification using transmitted light microscopy. Separate staining was not necessary.

2.4. Confocal Laser Scan Microscopy

For GFP-visualization, a Confocal Laser Scan Microscope (TCS SP1, Leica GmbH, Wetzlar, Germany) was used with an excitation wavelength of 488 nm and an emission filter of 500–550 nm. For microscopy, a 63× water immersion objective was used and processed with ImageJ (http://rsbweb.nih.gov/ij, last accessed on 18 August 2018).

For analysis of the cytoskeleton, two-day-old seedlings were used; the seedlings were placed on a slide in water or in an allicin solution of the indicated concentration. The analysis was carried out after 30 min. For the PIN and auxin localization study, seedlings were placed in Hoagland medium containing a concentration of 2 mM. The seedlings were incubated for 24 h under short day conditions (22 °C, 16 h dark, 8 h light) before microscopic analysis.

2.5. Growth Assays

Arabidopsis thaliana seeds were surface sterilized with ethanol and spread with a toothpick onto solid Hoagland medium. Plates were stored at 4 °C for two days for stratification before incubation at 22 °C for an additional two days under long day conditions (16 h light/8 h dark). After this time, seedlings were carefully transferred with forceps to Hoagland medium plates containing allicin at a final concentration of 0.5 mM and further incubated under these conditions. Root length was measured at day zero and after three days.

2.6. Statistics

Statistical analyses were carried out using the program SigmaStat. Data were tested for statistical significance using a multivariate data analysis (ANOVA) with a significance level of p < 0.05. The data were first tested for normal distribution to allow for valid application of the statistical tests.

3. Results and Discussion

3.1. Protoplasmic Streaming in Tradescantia Fluminensis VELL

As a model system for the influence of allicin on the cytoskeleton, we chose the sterile hyaline filaments from the flower of T. fluminensis (shown in Figure 1A–C). The time that elapsed until cytoplasmic streaming ceased after exposure to allicin solutions of various concentrations was considered as a measure of the relative activity of allicin on the cytoskeleton. Particles carried in the cytoplasmic stream were observed to have locomotion (Figure 1D,E). Soon after the hyaline filaments of T. fluminensis were exposed, the flow slowed down and stopped entirely after a certain time, which was dependent on the allicin concentration. As the concentration of allicin increased, the time required for the protoplasmic flow to come to a standstill decreased (Figure 1D). While a concentration of 0.5 mM of allicin led to a cessation of protoplasmic flow after approx. 13 min, a treatment with 4 mM of allicin led to a complete cessation of protoplasmic flow after approx. 2 min (Figure 1D).

A further observation for 5, 10, and 30 min after treatment showed that, in the presence of allicin, no new onset of protoplasmic disturbance was observed.

3.2. The Effect of Allicin on the Protoplasmic Streaming Was Partially Reversible

In the following experiments, it was investigated as to whether the effect of allicin was reversible by washing out the allicin. Filaments of Tradescantia were treated with allicin with a final concentration of 0.25 mM. After cytoplasmic flow stopped, the allicin solution was washed by capillary suction with filter paper that drew in pure water at the coverslip edge to exchange the bathing solution. After about 12 min, protoplasmic flow began again in isolated cells, although at a slower rate than before allicin treatment (not shown). Thus, it can be concluded that the inhibition of protoplasmic flow by allicin can be reversible after a certain regeneration phase; at a concentration higher than 0.25 mM allicin, this reversal did not occur even after a long observation period (>30 min).

3.3. Influence of Allicin on the Movement of Peroxisomes in Leaves

Since the movement of organelles, such as peroxisomes, also takes place within the framework of protoplasmic flow, an Arabidopsis thaliana line was used whose peroxisomes were tagged with GFP [39]. This made their movement traceable. Leaves of this Arabidopsis thaliana line were infiltrated by vacuum infiltration, with either water as a control treatment or an allicin solution of the indicated concentration, and the movement of the organelles was observed under a confocal laser scanning microscope (Figure 2A,B). While in the control, it was clearly indicated by observing the curved trail of peroxisome movements between the arrows that the peroxisomes moved substantially over the 10 min observation period (Figure 2A), in the sample infiltrated with 4 mM of allicin, movement of the peroxisomes was not observed (Figure 2B).

3.4. Allicin Inhibits Primary- and Lateral Root Growth and Root Hair Development

The morphology of Arabidopsis thaliana wild type Col-0 seedlings germinated for two days and exposed to 0.5 mM of allicin was very reminiscent of the morphology of the rml1 mutant [33]. The roots of wild type Col-0 ceased growing in the presence of allicin, while the shoot axis appeared to be largely unaffected and developed normally (Figure 3A). Compared to the untreated control, the roots of plants exposed to allicin were generally much shorter and showed no lateral roots or root hairs. The prooxidant cumene hydroperoxide also led to a similarly shortened root (Figure 3B). Root growth in the pad2 mutant was also similarly inhibited by allicin and cumene hydroperoxide under the conditions tested (Figure 3B).

3.5. Effect of Allicin on Actin- and Tubulin Cytoskeleton in Arabidopsis thaliana

To understand in more detail how allicin affects root development and whether this is associated with the putative effect on the cytoskeleton that was observed in Tradescantia, Arabidopsis thaliana lines were used whose actin or tubulin filaments were visualized with by beta-tubulin–GFP fusion [38] and GFP–actin-binding protein expression [37].

Confocal microscopy of the actin cytoskeleton clearly showed typical filamentous structures in the root of the untreated control (Figure 4A). After exposure to 2 mM of allicin for 24 h, the filamentous structures were destroyed; instead, GFP aggregate structures were visible at the margins of the cells. Allicin thus destroyed the integrity of the actin cytoskeleton in the roots of Arabidopsis thaliana.

In addition to the actin cytoskeleton, the tubulin cytoskeleton was also examined in an analogous manner. Here, too, the typical filament pattern was seen in the untreated control; treatment with allicin resulted in the dissolution of these filament structures and the diffuse fine-particulate distribution of GFP fluorescence, in contrast to the aggregate formation observed for the actin.

3.6. Effect of Allicin on Auxin Distribution PIN Localization

The cytoskeleton fulfils a wide variety of tasks in the cell and is involved, among other things, in transport processes in the cell. In order to investigate the influence of cytoskeletal integrity—here primarily concerning actin—on important cellular transport processes, we chose the localization of PIN proteins, whose localization is crucial for regulated auxin transport in tissues. Again, GFP-tagged versions of the proteins, in this case PIN1 and PIN3, were used to track their localization in vivo. In addition, the expression of the constructs was under the respective native promoter, so that the localization of the expression could also be followed at the tissue level.

In the untreated control, the PIN1 showed expression mainly in the region of the central cylinder (Figure 5A); at the cellular level, the strongest fluorescence was observed on the basipetal side of the cell. The exposure to allicin with a concentration of 2 mM led to a complete loss of this clear localization pattern, and GFP fluorescence was weakly distributed throughout the cell (Figure 5B). This corresponded exactly to the observations made by Koprivova during treatment with BSO, which suggests that allicin was also active here via the glutathione pool [31].

In contrast to the PIN1, the localization of the PIN3 was also found in the area of the cortex, but it was more prominent in the area of the root apical meristem. Again, fluorescence was strongest at the basipetal end of the cell (Figure 5C).

As observed with the PIN1, complete disruption of the PIN localization had also occurred with PIN3 by treatment with 2 mM of allicin. The GFP signal, which indicated the localization of the PIN1 protein, was diffusely present at both the cellular and tissue levels (Figure 5D). These results exactly mirrored the effects of the GSH1 inhibitor buthione sulfoximine (BSO) on PIN distribution [31]. BSO led to a reduction in cellular GSH levels, and this mirrored the result of allicin titrating out and oxidizing the GSH pool [13].

3.7. Effect of Allicin on Auxin Distribution

Based on the observation that allicin destroyed the localization of PIN proteins, and because PIN proteins are essential for the directional transport of auxin, the final aim was to investigate whether the aberrant localization of auxin occurred after allicin treatment. Two independent methods based on fluorescent proteins were used to check the localization of allicin. The DR5–GFP reporter system was based on the fact that auxin induced the activity of the DR5 promoter and thus drove GFP expression. GFP fluorescence thus indexed the presence of auxin.

Untreated Arabidopsis thaliana roots showed a clear localization of auxin around the quiescent zone of the root apical meristem, in the central cylinder and in the epidermal layers of the root (Figure 5E). Treatment with allicin resulted in the localization of GFP fluorescence being restricted to the root apex; auxin-dependent expression of GFP was no longer found in the quiescent and division zone of the root apex (Figure 5F). This showed that allicin caused a clear redistribution of auxin localization at the root tip.

As an alternative method to detect the localization of auxin at the tissue level, the DII–VENUS construct was used, in which a “fast maturing” form of yellow-fluorescent protein (YFP) was expressed in frame with an IAA protein under the control of a constitutive promoter. Because IAA proteins are degraded in the presence of auxin, a lack of fluorescence of the VENUS protein would indicate a high concentration of auxin, whereas a strong fluorescence of the protein would indicate a low concentration of auxin. The results obtained by means of DII–VENUS corresponded to the statements we also perceived with DR5–GFP. While the control showed that auxin was present in the root meristem area, the distribution was diffuse after allicin exposure (Figure 5G).

3.8. Effect of Allicin on Mutants Impaired in Auxin Homeostasis and Signaling

Because we observed an effect on auxin distribution and PIN localization after allicin treatment (Figure 5), we investigated the effects of allicin on the auxin transport mutant aux1-2 and the auxin-resistant mutant axr1-12 in comparison to wild type Col-0. Whereas the PIN proteins are functional in auxin efflux at the basal cell pole, the AUX1 protein transports auxin into the cell from the apoplast at the apical pole.

The Col-0 wild type showed that its exposure to 0.5 mM of allicin led to a significantly shortened root. The aux1-2 mutant root length, which already had a significantly shortened root in the control compared to the Col-0, was not further affected by the presence of allicin, i.e., this mutant showed a degree of allicin resistance (Figure 6A) and this suggests that the root phenotype we observed with allicin may have been dependent upon auxin over-accumulating to supraoptimal levels and that inhibited root growth. The axr1-12 mutant, on the other hand, showed no significant change in root length phenotype compared to the wild type in the control, but treatment with allicin caused the roots to be shortened, but to a much lesser extent than in the Col-0 control. This suggests that the influence of allicin on auxin transport was partly responsible for the observation that the roots did not grow as well under allicin exposure (Figure 6B). Possibly, another auxin receptor could have partially complemented the phenotype of axr1-12. To test this hypothesis, Col-0 and axr1-12 mutants were exposed to 20 µM of cytochalasin, which led to disruption of the actin cytoskeleton. As a control, it was examined whether the solvent DMSO at the concentration used (1% v/v) would lead to an effect, which was not the case. While the Col-0 wild type showed significantly shortened roots after treatment with cytochalasin, the axr1-12 mutant did not, which suggested that it was disruption of the actin cytoskeleton which altered auxin localization. Accordingly, a mutant that was not an auxin receptor mutant showed no phenotype in terms of root length upon destruction of the actin cytoskeleton. These results suggest that the effect allicin has on root development correlates, at least in part, with its influence on auxin distribution in the root, which are dependent upon allicin’s effects on the cytoskeleton.

4. Conclusions

The thiosulfinate allicin directly affects the cytoskeleton through its thiol-oxidizing properties. In the model system Tradescantia fluminensis, it was shown that the protoplasm flow was stopped. GFP-tagged cytoskeletal components revealed that the direct disruption of both actin and tubulin filaments also occurred. Since the localization of the phytohormone auxin, which is crucial for root development, is regulated by the orientation of PIN channels and crucially depends on the functionality of the cytoskeleton, we tested the hypothesis of whether, on the one hand, PIN channels showed an altered localization and, on the other hand, auxin was localized in an altered manner as a result. Both could be shown by GFP probes. This resulted in a shortening of the roots due to auxin mislocalization in the growth zone. Thus, it was clear that oxidative stress in roots led to a shortening of the roots, which was due to an influence on the cytoskeleton integrity and subsequent mislocalization of the auxin. This has far reaching environmental relevance, far beyond the study of the effect of allicin as a model substance.