Allyl-isothiocyanate (AITC) with a chemical formula of CH₂ = CH-CH₂-N = C = S is a main component of the volatile oil of several Brassicaceae species. The main AITC source is the root system of horseradish (Armoracia sp.). It is a simple lipophylic molecule, produced from degradation by myrosinase of sinigrin, a glucosinolate after wounding of plant tissues (Zhang, 2010; Nguyen et al., 2013).

There are multiple biochemical and cellular/ physiological effects of isothiocyanates and among them, AITC in eukaryotes. In isothiocyanate producing higher plants, it is thought to play an essential role in the defense against herbivores and pathogens, since it has bactericidal and fungicidal activity (Åsberg et al., 2015). For vertebrates, its effects were observed mainly in mammalian tumor cell lines (see the review of Nguyen et al., 2013) and here we list those related to alterations of cell cycle and cell viability:

Other effects of isothiocyanates include anti-inflammatory and antimicrobial properties.

Even though the permeability of plant cell membranes for AITC is not known, its chemical structure is suitable for penetrating plant membranes and several studies demonstrate that it can be transported easily through bacterial, mammalian, and probably fungal membranes (Purhonen et al., 2008; Zhang, 2010; Bertóti et al., 2016). Moreover, AITC affects plant intracellular events like modulation of glutathione-S-transferase (GST) gene expression (Øverby et al., 2015).

Microcystin-LR (MCY-LR) is a cyclic heptapeptide produced by several cyanobacterial genera/species containing two unusual amino acids: N-methyldehydroalanine (Mdha) and the hydrophobic D-amino acid, 3-amino-9-methoxy-2-6,8-trimethyl-10-phenyldeca-4, 6-dienoic acid (Adda). Its amino acid sequence is cyclo (-D-Ala¹-L-Leu²-Derythro-β-methylAsp³-L-Arg⁴-Adda⁵-D-isoGlu⁶-N-methyldehydro-Ala⁷) (Botes et al., 1985; Chen and Xie, 2016). MCY-LR is a potent inhibitor of type 1 and 2A phosphoserine-threonine specific protein phosphatases (PP1 and PP2A), with IC₅₀ values at the subnanomolar range in in vitro assays (MacKintosh et al., 1990). The Adda residue enters into the hydrophobic groove of phosphatases near the active site, thereby it blocks irreversibly the binding of phosphorylated serine or threonine residues of substrate proteins (Goldberg et al., 1995). PP1 and PP2A are involved in nearly every aspect of intracellular biochemical processes, including cell cycle regulation, signal transduction etc., in all eukaryotes including plants (Luan, 2003). It is not surprising therefore, that the inhibition of their activity alters mitosis and the organization of subcellular structures involved in cell division (Máthé et al., 2016). Besides protein phosphatase inhibition, MCY-LR has a strong reactive oxygen species (ROS) inducing effect in many eukaryotes, that may also contribute to subcellular structure alterations induced by this compound (Máthé et al., 2016; Pflugmacher, 2016).

As seen above, both AITC and MCY-LR may alter mitotic division in eukaryotes. In a preliminary study, we have observed that crude extracts of horseradish roots arrested Vicia faba (broad bean, a model eukaryote for the study of mitotic events, Olszewska et al., 1990; Doležel et al., 1999) meristematic cells in metaphase and induced abnormal mitotic spindle organization, when combined with MCY-LR. Our hypothesis was that a compound from horseradish could alter mitotic activity and this could be related to alterations in protein phosphatase activities. Thus, we aimed to study (i) the cytological effects of the main compound of horseradish volatile oil, AITC alone or in combination with MCY-LR in order to see whether it can be used as a tool for studying the regulation of eukaryotic subcellular events; (ii) to investigate whether PP1 and PP2A are modulated during the subcellular changes detected and to answer the question: is a certain type of phosphatases involved in the regulation of certain mitotic events, e.g., metaphase-anaphase transition and mitotic spindle organization? In this study, we provide novel contributions for the understanding of the mechanisms of metaphase-anaphase transition in plants and in eukaryotes in general.

AITC is known to bind to the thiol group of cysteine as a free amino acid or in polypeptides (Zhang, 2010). We have predicted that since MCY-LR does not have any sulfur-containing amino acids, it is not likely to form any complex with AITC. Even though during LC-MS analysis, a theoretical dithiocarbamate adduct that is formed on the Arg sidechain in the reaction of MCY-LR with AITC was sought at [M+H]⁺ m/z 1094.5, this theoretical situation occurs only at long-term (several days) of co-existence of AITC and MCY-LR in the same solution. Thus, LC-MS analysis did not reveal any significant reaction product or adduct between AITC and MCY-LR for short-term (12 h) that was used in our experiments (Supplementary Figure S1). As a consequence, the subcellular changes at combined treatments with AITC and MCY-LR detected in the present study are not the effect of a reaction product of these two compounds.

The most important cytological effects of AITC alone or in combination with MCY-LR can be summarized as follows: (i) hyperphosphorylation of histone H3 at Ser10 in metaphase and anaphase chromosome arms; (ii) the mitotic stimulatory effects of low concentration AITC and MCY-LR combinations, and most importantly, (iii) combinations of high AITC and MCY-LR concentrations induced not only histone H3 hyperphosphorylation, but abnormal mitotic spindle organization and metaphase arrest as well. Metaphase arrest was observed both in synchronized and non-synchronized cells, and was comparable to the effects of colchicine (Pan et al., 1993) and the combined treatment was suitable for high quality chromosome preparation for karyotyping (see Figures 1–3A, 7 and Supplementary Figures S3, S6, S7).

Both AITC and MCY-LR are known for inducing oxidative stress in all eukaryotes including plants, although for AITC this was described only at very high doses (Nguyen et al., 2013; Máthé et al., 2016; Pflugmacher, 2016). However, we proved that in V. faba roots, none of these compounds increases ROS levels (Supplementary Figure S4). V. faba has been proven before to be resistant to oxidative stress, since H₂O₂ and MCY-LR did not increase ROS levels, but increased the activity of antioxidant enzymes (Garda et al., 2016). Since both compounds had significant effects on PP1 and PP2A activities in vivo and/or in vitro, we will discuss their cytological effects in this context.

It was previously known that both PP1 and PP2A are involved in the dephosphorylation of histone H3 at diverse serine and threonine residues in yeast and animal cells and the inhibition of their activities by diverse drugs will increase the overall level of histone H3 phosphorylation or its hyperphosphorylation at specific residues like Ser10 (Guo et al., 1995; Hsu et al., 2000; Manzanero et al., 2002). Similar events were suggested for plant cells as well (Manzanero et al., 2002; Beyer et al., 2012). For eukaryotic cells other than plants, it is known that Aurora B kinase phosphorylates histone H3, leading to the increase of chromosome condensation. PP1 has a contrary effect, since it dephosphorylates Aurora B substrates involved in chromosome condensation and cohesion, allowing metaphase-anaphase transition (Afonso et al., 2017). Thus, PP1 has an important, if not primary, role in histone H3 dephosphorylation. In animal and human cells, PP2A has also a role in pH3 Ser10 dephosphorylation (Li et al., 1996). For plants, an interaction between PP2A and histone H3 is likely to occur (Bíró et al., 2012), but to date, little was known on the role of PP1 in histone H3 dephosphorylation. In this study, we have observed that AITC used alone or in combinations with MCY-LR induced pH3 Ser10 hyperphosphorylation both in metaphase and anaphase chromosomes at concentrations that inhibited PP1 activity in lateral root tips. This was true even for treatments where PP1 was inhibited, but PP2A activities were not changed significantly in vivo (Table 1). For metaphase chromosomes, the increase in the pool of Ser10 phosphorylated histone H3 occurred at 0.2 and 10 μM MCY-LR used in the absence of AITC as well (Figures 6, 7, Supplementary Figure 6, and Table 1). Moreover, OA known to inhibit primarily PP2A (for in vivo assays, proven in this study as well, Supplementary Figure S5B), did not induce significant changes in the phosphorylation state of histone H3, while Tm (known to inhibit primarily PP1 activity and proven in vivo in this study as well, Supplementary Figure S5C) did induce a significant increase in the phosphorylation of histone H3 in metaphase chromosome arms (Supplementary Figure S7C and Table 1). These data suggest that although the role of PP2A in histone dephosphorylation in plants cannot be ruled out, PP1 seems to play a crucial role in this process.

Histone H3 hyperphosphorylation contributes to hypercondensation of metaphase chromosomes and the delay of metaphase-anaphase transition (Manzanero et al., 2002; Beyer et al., 2012). However, in our study total metaphase arrest occurred only at (1)10 μM AITC + 10 μM MCY-LR treatments. Other AITC (with or without MCY-LR) treatments did induce histone H3 hyperphosphorylation without total metaphase arrest (Table 1). We drew the conclusion that hyperphosphorylation of histone H3 at Ser10 by itself is not enough for arrest, other factors must contribute to induce this blocking of metaphase-anaphase transition (see below).

For the induction of abnormal spindle development, this could not be observed in cases when one of the protein phosphatases (PP1 or PP2A) was not inhibited in vivo (Table 1). Inhibition of both PP1 and PP2A is needed for the development of such anomalies. In general, this seems to be true not only for the combination of 10 μM AITC and 10 μM MCY-LR, but for long-term treatment with 10 μM MCY-LR applied as the sole inhibitor as well. The latter treatment inhibits both phosphatases, but induces only a slight increase of abnormal spindle frequency after 12 h treatment (Table 1 and Supplementary Figure S3). However, 48 h treatments with 1 and 10 μM MCY-LR increased significantly the formation of abnormal metaphase spindles and induced abnormal sister chromatid segregation, and this could be correlated with inhibition of total PP1 + PP2A activities (Garda et al., 2016). Moreover, long-term (6 days) of MCY-LR (as single inhibitor) treatment, that induces severe protein phosphatase inhibition, high frequency of spindle anomalies and metaphase histone H3 hyperphosphorylation with chromosome hypercondensation- leads to severe inhibition of metaphase/anaphase transition in non-synchronized V. faba cells (Beyer et al., 2012; Garda et al., 2016). Abnormal chromatid segregation like the appearance of lagging chromosomes may be induced by MCY-LR solely by histone H3 hyperphosphorylation (without abnormal spindle organization) as shown in this study (Figure 1c and Supplementary Figure S2) and in our previous work (Beyer et al., 2012). Even though treatments with MCY-LR as the sole inhibitor induced such anomalies, the short-term treatments with 1 μM or 10 μM AITC + 10 μM MCY-LR were the only that induced nearly complete metaphase arrest.

To sum up, out of the treatments tested (1 μM AITC + 10 μM MCY-LR treatment was not checked for microtubule organization) only a combination of 10 μM AITC and 10 μM MCY-LR will lead to a substantial increase of the frequency of metaphase spindle anomalies at short-term (12 h) treatments (Figure 1 and Supplementary Figure S3). In this study, the inhibition of both PP1 and PP2A activities induced the development of abnormal spindles (Table 1). This confirms previous studies showing that both PP1 and PP2A are required for normal spindle assembly (Axton et al., 1990; Ayaydin et al., 2000). Thus, metaphase arrest was induced, when the following events occurred concomitantly in root tip meristems (i) histone H3 hyperphosphorylation; (ii) abnormal spindle development; (iii) severe inhibition of both PP1 and PP2A activities (Table 1).

There are various effects of protein phosphatases (and among them, PP1) on metaphase-anaphase transition in eukaryotes. Inhibition of PP1 by neutralizing antibodies leads to metaphase arrest in mammalian cells, but this report did not provide information on alterations of spindle organization or histone H3 phosphorylation (Fernandez et al., 1992). Protein phosphatases affect the functioning of the APC/C complex and their inhibition may lead to the delay of metaphase/anaphase transition in Xenopus (Tunquist and Maller, 2003). In contrast, Pinsky et al. (2009) reported that PP1 activity prevents spindle checkpoint activation in budding yeast. In plants, knowledge on the effects of PP1 on the onset of late mitosis is limited. For tobacco BY-2 cells, OA induces prophase arrest, however, this drug inhibits mainly PP2A, not PP1 (Zhang et al., 1992). As a consequence, the present report might give an important contribution to the understanding of the role of PP1 in metaphase-anaphase transition, given the significant histone H3 hyperphosphorylation induced by its inhibition.

In the following sections, we attempt to explain the effects of AITC treatments (with or without MCY-LR) on protein phosphatase activities, mainly in vivo.

MCY-LR has been known for a long time as a potent inhibitor of PP1 and PP2A and it also inhibits a number of other protein phosphatases of the PPP family-PP4, PP5, PP6 (MacKintosh et al., 1990; Fontanillo and Köhn, 2018). It should be noted that there is a controversy in literature regarding the preference of this heptapeptide to a given type of phosphatase. While several authors state that it inhibits PP1 and PP2A with equal potency (MacKintosh and Diplexcito, 2003), others claim that it has a preference to PP2A over PP1 (Honkanen et al., 1990, 1994). Here we demonstrate that in the in vitro assay for the activity of catalytic subunits, the inhibition of PP1 and PP2A by microcystin is similar (Figure 4). In contrast, for V.faba root extracts, the inhibitory effect is tissue-specific, but in principle PP2A is preferentially inhibited (Figures 5, 6). However, it should be noted that the in vitro study is not in accordance with the early finding of Honkanen et al. (1990), who found that even during the in vitro assay with purified PP1 or PP2A, MCY-LR inhibits PP2A more potently than PP1. Thus, the degree of phosphatase inhibition induced by MCY-LR depends on the assay conditions as well. The discrepancy between in vitro and in vivo assays may be caused by the differences in the nature and activity of protein phosphatase regulatory subunits in different tissues (see below).

As concerning the effects of AITC on protein phosphatase activities in vivo, one should take into consideration, that even though it is an unstable compound, it is probable that at least a fraction remains intact after entering cells during the 12 h exposure period used in this study. On the other hand it is known that a given fraction of compounds of horseradish essential oil can be detoxified by GST in Candida albicans (Bertóti et al., 2016) and this was described for pure AITC in mammals (Zhang, 2010). AITC increases GST gene expression in the model plant Arabidopsis (Øverby et al., 2015). Thus, it is likely that intracellular AITC is partly detoxified by eukaryotic cells. If we admit that the remaining free AITC interacts directly with PP1/PP2A in vivo, the chemical structure of this isothiocyanate makes highly probable its ability to exert nucleophilic attack against several amino acids including cysteine (e.g., the probable mechanism of its conjugation to glutathione). Indeed, AITC interacts with cysteine residues of proteins and this is probably a covalent interaction (Zhang, 2010). An early study showed that PP1 and PP2A activity is inhibited by reagents that modify sulfhydryl groups (Nemani and Lee, 1993). Later on, Choi et al. (2017) proved that binding of tautomycetin to Cys127 of PP1 is the key mechanism of its inhibition of PP1 activity. This reactive Cys residue is also involved in the inhibition of PP1 by a small inorganic molecule: H₂S, which induces persulfidation (Yadav et al., 2017). MCY-LR binds covalently to Cys273 of PP1c, also important for PP1 inhibition (Fontanillo and Köhn, 2018). Even if we accept a possible binding to cysteine residues, AITC does not affect significantly PP1c activity in vitro (Figure 4). This means that AITC probably does not bind to Cys273. The PP1 inhibitory (that is, opposite) effect of 10 μM AITC in vivo might be related to its complex interaction with regulatory and catalytic subunits (see below). Concerning PP2A activities, drugs like phoslactomycin A that bind to the Cys269 residue of PP2Ac important for its activity are inhibiting PP2A (Teruya et al., 2005). A mutation characterized by the change of this Cys residue to Gly leads to OA resistance (Shima et al., 1994). Interestingly, MCY-LR also binds to Cys269 of PP2A by its Mdha residue, and although the Adda residue is responsible mainly for PP1 and PP2A inhibition, the Cys binding also influences the effect of MCY-LR on PP2A activity (Pereira et al., 2011; Fontanillo and Köhn, 2018). AITC is not likely to interact with Cys269, because it does not produce adducts with MCY-LR, but preincubation with 1 nM MCY-LR firstly, then with 100 nM AITC induces more severe inhibition of PP2A, than MCY-LR alone (see Figure 4D and Supplementary Figure S1). To sum up our in vitro protein phosphatase inhibition assays, at least for PP1, AITC does not interact significantly with the catalytic subunit (Figure 4).

On the other hand, we have two reasons to presume that in vivo, AITC interacts with regulatory subunits of PP1 and PP2A: (i) the differences between in vitro and in vivo AITC effects on protein phosphatase activities: no significant effects (alone or in combinations with MCY-LR) on PP1c activities in vitro and the significant inhibition of PP1 activities by AITC both in whole roots and root tips in vivo (compare Figures 4–6); (ii) the tissue specific modulation of protein phosphatase activities by AITC (and MCY-LR) (see Figures 5, 6). Moreover, even though both 10 μM MCY-LR and the 10 μM AITC + 10 μM MCY-LR combination induce severe PP1 and PP2A inhibition in vivo, the degree of inhibition at the combined treatment is significantly lower, than for the single-MCY-LR treatment (Figures 6B,C), but this is valid for root tips and not for whole root extracts (Figures 5B,C). In the next section, we present the second argument in detail.

For the in vivo protein phosphatase assays, AITC modulates their activities in a tissue-specific manner. 1 μM AITC inhibits only PP1 in whole lateral roots, but its effects are non-significant in lateral root tips where the meristematic tissue- subject of our study- occurs (compare Figures 5, 6). In eukaryotes, PP1 and PP2A are in fact complexes containing catalytic as well as a various number of regulatory subunits. Regulatory subunits are responsible for the diversity and tissue-specificity of functions of the catalytic subunits (Shi, 2009). This is valid for plants, too, where, as in other multicellular organisms, regulation of PP1 activity is diverse in different tissues and/or subcellular compartments (for example, RSS1 is involved in the regulation of cell cycle by this phosphatase). An interesting example is Inhibitor 3 (INH3), a regulator of PP1 found in V. faba and Arabidopsis. It is expressed in diverse organs and tissues, with particularly strong expression in Arabidopsis root tips (Takemiya et al., 2009). Diverse PP1 regulatory subunits were proposed for contributing to control of the spindle assembly checkpoint in eukaryotes (Bollen et al., 2009). Regulatory subunits of PP2A are even more diverse, than for PP1, being classified into classes B, B’, B,” and B”’ (Shi, 2009; Uhrig et al., 2013). The Ton2/Fass gene family encodes for a B” regulatory subunit, involved in the regulation of the onset of mitosis and the assembly of mitotic microtubules (Spinner et al., 2013).

To sum up the biochemical mechanisms of AITC-protein phosphatase interaction(s), at present we do not have enough data to construct a model for this interaction. However, as shown above, AITC may also affect the interaction of core enzymes with or act directly on their regulatory subunits in a plant tissue-specific manner. An important future task is to reveal these molecular interactions. The effects of AITC on PP1 and PP2A activities raise the possibility that this volatile oil, alone or in combination with naturally occurring plant protein phosphatase inhibitors (like INH3-like proteins mentioned above) may modulate phosphatase mediated signal transduction pathways involved in defense responses upon wounding in glucosinolate-producing plants.

The main conclusions of the present study are (i) AITC alone or in combination with MCY-LR influences protein phosphatase PP1 and PP2A activities in vivo, but generally, does not have significant influences on the catalytic subunits in vitro; (ii) we provide experimental evidence that PP1 is a protein phosphatase of crucial importance in pH3 Ser10 dephosphorylation in V. faba and (iii) together with the proper organization of spindle assembly, that is dependent on both PP1 and PP2A activities, PP1-modulated histone H3 phosphorylation regulates metaphase-anaphase transition. Concomitant and rapid alterations of these two cellular processes lead to complete metaphase arrest. We have proved that the combinations of AITC with MCY-LR are useful for the more detailed elucidation of protein phosphatase mediated regulation of mitosis in plant cells. It is a future task to investigate whether similar effects could be induced in other eukaryotic, like mammalian cells and to identify the molecular events/ signaling pathways that connect histone hyperphosphorylation, altered spindle organization and protein phosphatase activities to the arrest of cells in metaphase. This report shows for the first time that AITC modulates serine-threonine protein phosphatase activities and that acute PP1 (and PP2A) inhibition induces metaphase arrest in plants.

The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.

TG performed a significant part of experiments including preparation of extracts for in vivo protein phosphatase assays and contributed to data analysis. ZK and FE performed protein phosphatase assays and contributed to data analysis related to these measurements. CF and BS contributed essentially to immunohistochemistry and histochemistry experiments. SG and AK-S performed the LC-MS measurements. GáV purified MCY-LR. MM-H made the statistical analysis. GyV made essential contributions to confocal microscopy observations. CM made microscopical observations, designed experiments, and wrote the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.