Beyond Microcystins: Cyanobacterial Extracts Induce Cytoskeletal Alterations in Rice Root Cells

Microcystins (MCs) are cyanobacterial toxins and potent inhibitors of protein phosphatases 1 (PP1) and 2A (PP2A), which are involved in plant cytoskeleton (microtubules and F-actin) organization. Therefore, studies on the toxicity of cyanobacterial products on plant cells have so far been focused on MCs. In this study, we investigated the effects of extracts from 16 (4 MC-producing and 12 non-MC-producing) cyanobacterial strains from several habitats, on various enzymes (PP1, trypsin, elastase), on the plant cytoskeleton and H2O2 levels in Oryza sativa (rice) root cells. Seedling roots were treated for various time periods (1, 12, and 24 h) with aqueous cyanobacterial extracts and underwent either immunostaining for α-tubulin or staining of F-actin with fluorescent phalloidin. 2,7-dichlorofluorescein diacetate (DCF-DA) staining was performed for H2O2 imaging. The enzyme assays confirmed the bioactivity of the extracts of not only MC-rich (MC+), but also MC-devoid (MC−) extracts, which induced major time-dependent alterations on both components of the plant cytoskeleton. These findings suggest that a broad spectrum of bioactive cyanobacterial compounds, apart from MCs or other known cyanotoxins (such as cylindrospermopsin), can affect plants by disrupting the cytoskeleton.


Introduction
Cyanobacteria are an ancient group of oxygenic photosynthetic prokaryotes, thriving in both aquatic and terrestrial habitats, even under the harshest conditions [1]. Their ability to inhabit numerous diverse environments is also reflected by the plethora of bioactive compounds, which they are able to produce [2]. Among these compounds, a multitude are known to be toxic to other organisms, especially eukaryotes [2], and their potential drug-attributes are an emerging research field (for a review, see [3]).
The most notoriously harmful and well-studied cyanobacterial toxins (also referred to as cyanotoxins), are the microcystins (MCs), harmful to animal cell systems [4], by affecting the cytoskeleton [5]. MCs are monocyclic heptapeptides with two variable amino acids and more than 250 different MC variants exist [2], the most common of which are MC-LR, MC-RR, and MC-YR. MCs are produced by several cyanobacteria, not only of freshwater, but also of marine or terrestrial environments [2], and seem to have functional roles within cyanobacterial cells, such as modulation of specific proteins and protection against oxidative stress [6]. MCs also inhibit serine/threonine protein phosphatases 1 (PP1) and 2A (PP2A) [7,8], which are involved in protein complexes that orchestrate Table 1. Inhibitory effects of the cyanobacterial extracts on PP1, trypsin, and elastase, with reference to each strain. "+": inhibition; "−": no effect.

Morphological Alterations
Microfilaments were abundant in control root cells (Figure 1), of the apical meristem ( Figure 1A) and the elongation zone ( Figure 1B), while cells treated with MC+ extracts (from MC-producing strains; 1410, 2410, 1414, 1614) exhibited time-dependent alterations of the F-actin cytoskeleton (Figures 2 and 3). After just 1 h of treatment, cortical actin filaments in meristematic root cells appeared to be disoriented and branched (Figure 2A,J) or even bundled (arrows in Figure 2D,G), sometimes exhibiting ring-shaped F-actin conformations (arrowheads in Figure 2G). After 12 h, cortical microfilaments, were heavily disoriented ( Figure 2E,H,K) and bundled (arrows in Figure 2B,E,H). In 1410-treated root cells, except for bundles, no discernable actin filaments could be observed after 12 h ( Figure 2B), while F-actin was almost absent after 24 h ( Figure 2C). Treatment with the other MC+ extracts for 24 h led to various F-actin defects, ranging from disorientation ( Figure 2I) and the appearance of actin circular conformations/rings (arrowheads in Figure 2F,I), to severe deterioration of the F-actin network integrity ( Figure 2L).  Table 1. Inhibitory effects of the cyanobacterial extracts on PP1, trypsin, and elastase, with reference to each strain. "+": inhibition; "−": no effect.      Figure 1A) and tended to form bundles (arrows in D,G) or even rings (arrowheads in G). After 12 h, bundling (arrows in B,E,H) and disorientation (E,H,K) appeared intensified. After 24 h, the F-actin network deteriorated significantly (F,L) and appeared disoriented (I), or even collapsed (C). Circular actin aggregates (arrowhead in F) and ring-shaped conformations (arrowhead in I) could also be detected. Scale bars: 10 µm.  Figure 1B). After 12 h, actin cables appeared highly affected, being either fragmented (B), disoriented (arrows in E,H), or scarce (K, note also the actin ring indicated by arrowhead). After 24 h, F-actin either disappeared (C,L) or was fragmented (F). Actin aggregates could also be detected (arrows in I). Scale bars: 10 μm.  Figure 1B). After 12 h, actin cables appeared highly affected, being either fragmented (B), disoriented (arrows in E,H), or scarce (K, note also the actin ring indicated by arrowhead). After 24 h, F-actin either disappeared (C,L) or was fragmented (F). Actin aggregates could also be detected (arrows in I). Scale bars: 10 µm.

MCs Production
Cells in the root elongation zone were also affected ( Figure 3). Longitudinal F-actin cables, a common feature of untreated cells in the elongation zone ( Figure 1B), could still be observed after 1 h of treatment ( Figure 3A,D,G,J). However, after 12 h, these cables appeared significantly altered in various ways, being either disoriented ( Figure 3E,H), scarce ( Figure 3K), or even absent, in which case F-actin was severely damaged and fragmented ( Figure 3B). Eventually, after 24 h, actin filaments either disappeared completely ( Figure 3C,L), or were scarce and disoriented, wherever present ( Figure 3F,I). Actin aggregates were also detected (arrows in Figure 3I).
In addition, F-actin was adversely affected in root meristematic cells after treatment with MC− extracts (from strains not producing MCs; . Disorientation/branching of cortical actin filaments ( Figures 4A, 5G and 6D,J), bundling ( Figure 4D,G and Figure 5A), or a combination of these effects ( Figure 5D,J and Figure 6A,G) were commonly observed in affected cells after only 1 h of treatment. Actin rings were also detected in some cases (arrowheads in Figure 5J). These effects persisted or were intensified after 12 h ( Figure 4B,E, Figure 5B,E,H,K and Figure 6E,H,K), while loss of actin network integrity ( Figures 4H and 6B), ring-shaped conformations (arrowhead in Figure 6B), and even cells devoid of F-actin (arrowheads in Figures 4B and 5E,K) could be observed. After 24 h, actin filaments were either weak and scarce ( Figure 4C, Figure 5F,I and Figure 6L) occasionally forming rings (arrowheads in Figure 5C), heavily bundled (Figures 4F, 5L and 6F,I,L), or eventually disappearing ( Figure 4I, arrowheads in Figure 5F,I).
As for the elongation zone (Figures 7-9), short treatments (1 h) with MC− extracts (from non-MC-producing strains) did not greatly affect F-actin in elongating cells ( Figure 7A,D and Figure 9D,J), except for some bundling/aggregates (arrows in Figure 7G, Figure 8D,G,J and Figure 9A,G,J) or disorientation effects (transverse actin cables instead of longitudinal) noticed (arrows in Figure 8A, arrowheads in Figure 9J). After 12 h, disorientation of actin cables ( Figure 7B,E and Figure 9E,H) and F-actin bundling/aggregates (arrows in Figure 8H,K and Figure 9B,K) were common effects, along with F-actin diminishing ( Figure 8B,E) and, in some cases, actin rings (arrowheads in Figure 7E) and cells devoid of F-actin (arrowheads in Figure 7B,H). After 24 h, the F-actin network was heavily disoriented, bundled, or degraded ( Figure 7C,F,I, Figure 8C,F,I,L and Figure 9C,F,I,L). Actin rings were also detected (arrowheads in Figure 7I).

F-actin Fluorescence Intensity Measurements
The detrimental effects of the cyanobacterial extracts on F-actin in rice root cells were further confirmed by measurements of the corrected total cell fluorescence (CTCF; Figure 10). In both meristematic ( Figure 10A) and elongation zone root cells ( Figure 10B), CTCF dropped upon treatment with each cyanobacterial extract, readily from 1 h of the extract application. The exposed elongating root cells exhibited a pronounced fluorescence intensity drop, compared to the meristematic cells ( Figure 10B; cf. Figure 10A). CTCF measurements drop observed in all extract treatments, showed a statistical significance, set at p < 0.05.

Effects on Microtubules and Chromatin
Microtubules were severely affected by MC+ extracts (Figure 11), compared to the control ( Figure 11A-F). Microcystis (1410 and 2410) extracts exhibited their effects on root meristematic cells starting from 1 h of exposure. Misoriented ( Figure 11G,N,S) or even fractured ( Figure 11M,Q) microtubules, deformed mitotic spindles ( Figure 11H-J,P,R; cf. Figure 11D) and phragmoplasts ( Figure 11K; cf. Figure 11F), absence of perinuclear microtubules in preprophase cells ( Figure 11O; cf. Figure 11C), and abnormal condensation ( Figure 11J,N) or dispersal ( Figure 11R,S) of chromatin were observed in affected cells after only 1 h of treatment. After 12 h, root cells affected by the 1410 extract exhibited no microtubules, as well as abnormally condensed chromatin ( Figure 11L), while cells treated with the 2410 extract exhibited scarce cortical microtubules ( Figure 11T; cf. Figure 11A), which eventually disappeared after 24 h ( Figure 11U). Extracts from other MC+ strains (1414 and 1614) exhibited slighter effects on rice root cells than the Microcystis extracts. After short-term treatments (1 h), 1414-treated root cells exhibited mainly microtubule disorientation in various cell cycle stages, such as cortical microtubules in interphase cells ( Figure 11V; cf. Figure 11A), perinuclear microtubules in preprophase cells ( Figure 11W; cf. Figure 11C) and phragmoplast microtubules ( Figure 11X; cf. Figure 11E), without further disturbance after longer treatments. Short-term (1 h) treatment with the 1614 extract also disturbed the microtubule network, leading to the prevalence of endoplasmic microtubules in affected cells ( Figure 11Y; cf. Figure 11B) and, in some cases, chromatin dispersal ( Figure 11Z). After 12 h, disoriented cortical ( Figure 11AA; cf. Figure 11A) and endoplasmic microtubules ( Figure 11AB; cf. Figure 11B), along with preprophase cells lacking preprophase band (PPB) ( Figure 11AC; cf. Figure 11C), were observed. Eventually, after 24 h treatment with the 1614 extract, microtubules disappeared almost totally ( Figure 11AD).         Figure 8B,E) and, in some cases, actin rings (arrowheads in Figure 7E) and cells devoid of F-actin (arrowheads in Figure 7B,H). After 24 h, the F-actin network was heavily disoriented, bundled, or degraded ( Figures 7C,F,I, 8C,F,I,L, and 9C,F,I,L). Actin rings were also detected (arrowheads in Figure 7I).   Actin cables were not significantly affected after 1 h of treatment (A,D), except for some actin aggregates (arrows in G). After 12 h, F-actin disorientation was visible (arrows in B,E), along with actin rings (arrowheads in E) and cells devoid of F-actin (arrowheads in B,H). After 24 h, actin rings were still detectable (arrowheads in I), as well as disorientation (C) and actin aggregates (arrows in F,I). Scale bars: 10 μm.  confirmed by measurements of the corrected total cell fluorescence (CTCF; Figure 10). In both meristematic ( Figure 10A) and elongation zone root cells ( Figure 10B), CTCF dropped upon treatment with each cyanobacterial extract, readily from 1 h of the extract application. The exposed elongating root cells exhibited a pronounced fluorescence intensity drop, compared to the meristematic cells ( Figure 10B; cf. Figure 10A). CTCF measurements drop observed in all extract treatments, showed a statistical significance, set at p < 0.05. Microtubules appeared to be also affected by MC− extracts from certain non-MC-producing strains ( Figure 12). Disorientation of cortical microtubules ( Figure 12A,C,H,Q; cf. Figure 11A) and the formation of excess endoplasmic microtubules ( Figure 12B,D,I,M,R, left cell in V; cf. Figure 11B) were common effects observed after treatment. Defects of the mitotic spindles ( Figure 12E,K, right cell in V,W; cf. Figure 11D), lack of perinuclear microtubules in preprophase cells ( Figure 12J,N,S; cf. Figure 11C, preprophase bands of microtubules are defined by brackets), and anomalies in phragmoplasts during cytokinesis ( Figure 12F,G,L,O,T; cf. Figure 11F) were reported as well. For root cells treated with the 0499 and 0599 extracts, these effects were rather minor, compared to the control, detectable at all time points of exposure (1-24 h). However, the Jaaginema (0110 and 0210) extracts and the Microcystis 1810 extract induced such alterations after only 1 h of treatment, along with even harsher effects, including abnormal chromatin condensation ( Figure 12V median cell, Figure 12X,Y). After longer exposure, affected cells appeared devoid of tubulin polymers, with abnormal chromatin condensation ( Figure 12P,U,Z).  Preprophase cells lacking perinuclear microtubules (O) were spotted, along with distorted mitotic spindles (note the misaligned chromosome outside the spindle, pointed with arrow in P). Affected cells of undefined chromatin state exhibited extremely short (Q) or misplaced and disoriented microtubules (R,S). After 12 h, only interphase cells with a sparse microtubule network were encountered (T), which eventually collapsed after 24 h (U). 1414: After 1 h of treatment, disorientation was common for both interphase (V) and perinuclear microtubules in preprophase cells (W), as well as phragmoplast microtubules in affected cytokinetic cells (X). Harsher effects were not recorded after longer treatments. 1614: After 1 h of treatment, affected cells exhibited scarce and misoriented microtubules (Y) and, in some cases, abnormally condensed and scattered chromatin (Z). After 12 h, disorientation of the microtubule network could still be observed in interphase (AA), also with abundant short perinuclear microtubules (AB). Preprophase cells without a PPB were recorded (AC). After 24 h, almost no tubulin polymers could be detected (AD). Scale bars: 5 µm. cells (X). Harsher effects were not recorded after longer treatments. 1614: After 1 h of treatment, affected cells exhibited scarce and misoriented microtubules (Y) and, in some cases, abnormally condensed and scattered chromatin (Z). After 12 h, disorientation of the microtubule network could still be observed in interphase (AA), also with abundant short perinuclear microtubules (AB). Preprophase cells without a PPB were recorded (AC). After 24 h, almost no tubulin polymers could be detected (AD). Scale bars: 5 μm.

Induction of Oxidative Stress
Six of the extracts induced oxidative stress in rice roots, compared to control ( Figure 13A,H), initiating at 12 h of exposure. All Microcystis extracts, from both MC-producing (1410, 2410) and non-MC-producing (1810) strains, produced increased levels of H 2 O 2 after 12 h (Figure 13B,C,G). After 24 h, fluorescence was only visible in cells of the root apical meristem ( Figure 13I,J,N). MC− extracts from both Jaaginema strains (0110, 0210) also produced elevated H 2 O 2 levels in treated roots after 12 h ( Figure 13E,F) and 24 h ( Figure 13L,M), while similar effects were observed after treatment with the MC+ extract of Trichormus variabilis TAU-MAC 1614 ( Figure 13D,K). detected, and chromatin state was also abnormal (U). 1810: After 1 h of treatment, affected cells exhibited numerous endoplasmic microtubules (left cell in V), abnormally short microtubules attached to chromosomes (middle cell in V) and malformed mitotic spindles (right cell in V). Misaligned chromosomes, outside the equator plate (arrows in W) could also be observed, as well as masses of chromosomes attached to aberrant spindle-like microtubules (X,Y). After 12 h, no tubulin polymers could be detected, while chromatin was abnormally condensed (Z). Scale bars: 5 μm.

Induction of Oxidative Stress
Six of the extracts induced oxidative stress in rice roots, compared to control ( Figure 13A,H), initiating at 12 h of exposure. All Microcystis extracts, from both MC-producing (1410, 2410) and non-MC-producing (1810) strains, produced increased levels of H2O2 after 12 h (Figure 13B,C,G). After 24 h, fluorescence was only visible in cells of the root apical meristem ( Figure 13I,J,N). MC− extracts from both Jaaginema strains (0110, 0210) also produced elevated H2O2 levels in treated roots after 12 h ( Figure 13E,F) and 24 h ( Figure 13L,M), while similar effects were observed after treatment with the MC+ extract of Trichormus variabilis TAU-MAC 1614 ( Figure 13D,K).

Discussion
Almost all of the cyanobacterial extracts applied for treatments affected rice root cells. More specifically, all the MC+ extracts affected both F-actin (Figures 2, 3 and 10) and microtubules (Figure 11), and increased the levels of H 2 O 2 (Figure 13), except Raphidiopsis raciborskii TAU-MAC 1414 extract, which did not induce an H 2 O 2 increase. MC− extracts (from non-MC-producing strains), except Nodularia sp. TAU-MAC 0717 extract, caused F-actin disorders (Figures 4-10). Jaaginema sp. TAU-MAC 0110 and 0210 and Microcystis viridis TAU-MAC 1810 extracts also affected microtubules and induced an increased H 2 O 2 production (Figures 12 and 13), while Synechococcus sp. TAU-MAC 0499 and Chlorogloeopsis fritschii TAU-MAC 0599 affected microtubules but did not increase H 2 O 2 ( Figure 12; for an overview, see Table 2). Therefore, it is further consolidated that extracts from various cyanobacterial strains (deriving from a multitude of environments; Table 3) target both plant microtubules and F-actin and are capable of inducing H 2 O 2 production. Table 2. Effects observed in rice root cells after exposure to cyanobacterial extracts, with reference to each strain. "": affected; "": no effect.
The inhibitory effects of several cyanopeptides-beyond MCs-on various proteases [26] provide a useful tool for correlating the microscopically observed cytoskeletal changes with the activity of such compounds. Therefore, we demonstrated the effects of the extracts on hydrolytic enzymes. The inhibition of the activity of proteolytic enzymes, such as chymotrypsin, trypsin, elastase, and thrombin, by cyanobacterial extracts has been frequently reported [36,37]. Inhibition of both elastase and trypsin was also assigned to peptides such as micropeptins, cyanopeptolins, microviridins, and banyasides [37][38][39][40][41][42]. The production of potent inhibitors has also been found in several cyanobacteria, like Microcystis, Planktothrix, Anabaena, Nostoc, Lyngbya, and Symploca [43][44][45]. Inhibition of hydrolytic enzymes might not be a threat for extracellular enzymes, diluted in bulk water. However, the probable accumulation of cyanopeptides by aquatic organisms, including plants, may result in reaching an intracellular concentration high enough to inhibit intracellular enzymes, as plants cannot regulate their endogenous peptidase activity in combination with serine peptidases of cyanobacterial origin [46,47].
Cyanotoxins, such as MCs and cylindrospermopsin, are known to affect plant growth [48,49], cause chromatin defects [12], and induce disorganization of microtubules in plant cells [50,51]. MC-LR was recently found to induce F-actin alterations in Oryza sativa root cells [13]. However, even MC− extracts which did not affect microtubules, appeared to disrupt the F-actin network, implying that each cytoskeletal component is affected by independent mechanism of toxicity and suggesting that F-actin is a primary target of cyanobacterial toxicity, beyond MCs, in plant cells. To our best knowledge, this is the first report of cytoskeletal alterations in plant cells induced by extracts from cyanobacterial strains not producing MCs or cylindrospermopsin, underlining that several more cyanobacterial bioactive compounds are able to disrupt the plant cytoskeleton. The exact identity and mode of action of these compounds (which may also exert their effects synergistically) remain to be further studied.
Oxidative stress, detected in roots treated with extracts from certain strains (Figure 13), could also play a role in the induction of cytoskeletal defects. Elevated ROS levels have been associated with reorganization of microtubules in plant cells [52]. ROS-induced F-actin remodeling has also been reported in innate immunity responses of Arabidopsis thaliana pavement cells [53], suggesting that the increase in ROS levels due to extract treatment could affect F-actin as well. Nevertheless, cytoskeletal alterations were also observed after treatment with extracts that did not induce oxidative stress. In addition, increased ROS production was detected at 12 h, while disorders of the cytoskeleton appeared even after 1 h of treatment. This is possibly a hint that cyanobacterial toxicity against the plant cytoskeleton may not always involve ROS.
Cyanobacterial extracts induced a multitude of alterations in rice root cells and these findings could also be of ecological significance. Indeed, cyanobacteria produce a wide range of bioactive compounds [54,55] and cyanobacterial blooms often consist of several species [56][57][58]. An emerging challenge is to identify the above compounds and analyze their specific effects on plant cells. This would be the target of further research.

Culture of Cyanobacteria, Biomass Collection, and Preparation of Extracts
Sixteen cyanobacterial strains of the TAU-MAC culture collection [59], representatives of various taxonomic orders and habitats, were used for experimental purposes (Table 3).  Table A1). Extracts from the cyanobacterial strains were prepared according to [13]. Wet biomass from each culture was harvested by centrifugation at the exponential growth phase (about 30 days) and freeze-dried. Lyophilized biomass (150 mg dry weight) from each strain was dissolved thrice in a total of 21 mL of 75% (v/v) methanol. All samples were sonicated during the first extraction step for 10 min. Methanol was finally evaporated and each pellet was resuspended in 5 mL of double-distilled water. Aqueous extracts were filtered through Whatman Polydisc TF filters (Whatman plc, Little Chalfont, UK) with a pore size of 0.2 µm.

Enzyme Inhibition Assays
Dilutions of crude cyanobacterial extracts (1:25 and 1:50) in double-distilled water (ddH 2 O) were used. All enzyme inhibitors were also diluted in ddH 2 O, at various concentrations. All assays were performed in 96-well microplates.

Protein Phosphatase Inhibition Assay
Potential bioactivity of the cyanobacterial extracts against PP1 activity was tested using the colorimetric method assay protocol described in [62], modified according to [28,36].

Trypsin Inhibition Assay
Extracts were tested for trypsin inhibition using the assay described by [42], with modifications. Porcine trypsin (1 mg mL −1 ) and its substrate, N α -benzoyl-DL-arginine 4-nitroanilide hydrochloride (BAPNA, Santa Cruz Biotechnology, 2 mM), were diluted in buffer containing 50 mM Tris-HCl (pH 7.5), 100 mM NaCl, and 1 mM CaCl 2 . Aprotinin was used as enzyme inhibitor. A total of 10 µL of sample/inhibitor solution/ddH 2 O + 10 µL of enzyme + 100 µL of buffer were added in each well and preincubated for 5 min at 36 • C. Afterwards, 100 µL of substrate were loaded and the mixture was incubated for 20 min at the same temperature. Absorbance was measured at 405 nm.

Elastase Inhibition Assay
Elastase inhibition was tested using the protocol by [63], modified according to [36]. Porcine elastase (75 µg mL −1 ) and its substrate, N-succinyl-Ala-Ala-Ala-p-nitroanilide (2 mM) were diluted in 0.2 M Tris-HCl (pH 8) buffer. Elastatinal was used as enzyme inhibitor. Then, 10 µL of sample/inhibitor solution/ddH 2 O + 10 µL of enzyme + 150 µL of buffer were added in each well and preincubated for 15 min. The reaction was started by adding 30 µL of substrate and the mixture was incubated for another 10 min. Absorbance was measured at 405 nm.

Plant Material and Treatments
Rice (Oryza sativa cv. Axios), generously provided by the National Cereal Institute (Thessaloniki, Greece), was germinated on moistened filter paper at 24 ± 1 • C in the dark. Four-to five-day-old seedlings were transferred in Eppendorf tubes containing either aqueous cyanobacterial extracts or double-distilled water (control) and placed with their roots submerged for various time periods (1, 12, or 24 h) under the same conditions as during germination. After treatment, root tips were prepared for fluorescence microscopy. All chemicals and reagents were purchased from Applichem (Darmstadt, Germany), Sigma-Aldrich (Taufkirchen, Germany), and Merck (Darmstadt, Germany) and all the following experimental procedures were performed at room temperature, unless otherwise stated.

Confocal Fluorescence Microscopy
Cytoskeletal elements in fluorescent specimens were observed under a Zeiss Observer.Z1 (Carl Zeiss AG, Munich, Germany) microscope, equipped with the LSM780 confocal laser scanning (CLSM) module and the appropriate filters for each fluorophore. Imaging was achieved with ZEN2011 software, according to the manufacturer's instructions.

Fluorescence Intensity Measurements
Fluorescence intensity measurements of F-actin in control and extract-treated root tip cells (from the meristematic and differentiation zone) were performed in maximum intensity projections of serial CLSM sections with ImageJ (https://imagej.net/Fiji), according to [65]. The corrected total cell fluorescence (CTCF; [66]), was calculated with the formula: CTCF = integrated density − (area of selected cell × mean fluorescence of background readings). Thirty individual cells from three different roots per treatment were measured for fluorescence intensity and results were statistically analyzed (ANOVA with Dunnett's test) with GraphPad Prism, at a significance of p < 0.05.

Detection of Hydrogen Peroxide Production
H 2 O 2 was detected using 2,7-dichlorofluorescein diacetate (DCF-DA), according to [67,68]. Extract-treated rice seedlings were incubated with their roots submerged in 25 µM DCF-DA aqueous solution for 30 min, washed with double-distilled water, and examined under a Zeiss AxioImager.Z2 light microscope (Carl Zeiss AG, Munich, Germany), equipped with an AxioCam MRc5 camera (Carl Zeiss AG). Seedlings treated with double-distilled water +5% DMSO for 24 h or 10 mM H 2 O 2 for 1 h, were used as positive and negative control, respectively. Images were captured using AxioVision Rel. 4.8.2 software (Carl Zeiss AG).

Conclusions
According to our observations, not only MC+ extracts, but also MC− extracts from various non-MC-producing cyanobacterial strains (isolated from a multitude of environments) were able to induce alterations in both plant cytoskeletal components, i.e., F-actin and microtubules, in Oryza sativa (rice) root cells ( Table 2). This is the first report of cyanobacterial extracts, not containing any known cyanotoxins, affecting the plant cytoskeleton. Therefore, it is supported that MCs or cylindrospermopsin are not the only cyanobacterial compounds able to induce cytoskeletal alterations in plant cells. Certain, but not all, MC− extracts also raised H 2 O 2 levels in treated roots, which is an implication that oxidative stress is not necessarily involved in the cytoskeletal alterations observed. Obviously, several bioactive compounds may be present in the extracts of non-MC-producing cyanobacterial strains, the identity and mode of action of which remains to be revealed by future studies.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.