Alkaloids, the largest class of nitrogen-containing secondary metabolites that occur mainly in plants, animals, bacteria and fungi, have a great diversity of structures and exhibit a wide range of pharmacological properties, such as antibacterial, antiviral, anti-inflammatory and antitumor effects [1
]. The biological activities of alkaloids are mainly due to their interactions with various molecular targets, such as biomembranes, proteins, DNA topoisomerase, nucleic acids, neuroreceptors and ion channels, among which the cytoskeleton represents one of the most important targets [1
Microtubules and actin filaments, the two essential elements of the cytoskeleton in eukaryotic cells, have multiple functions in cellular processes: microtubules decide the spatial distribution of membrane-enclosed organelles and form mitotic spindles during the cell cycle, while actin filaments determine the cell shape and together with microtubules are responsible for cell motility [7
]. The importance of microtubules and actin filaments in biological activities, such as vesicular and organelle transport, cell and nuclear migration, cell proliferation and division, makes them attractive targets for natural toxins in cancer research [9
Microtubule-binding agents (MBAs) disrupt the structure and dynamics of microtubules, which arrest the cell cycle and eventually lead to cell apoptosis [12
]. MBAs are important components in combination chemotherapy and widely used in clinics to treat many kinds of cancers, such as leukemia, lymphomas and solid tumors [13
]. Alkaloids constitute the most important group of MBAs; well-known examples are the diterpene alkaloids paclitaxel and docetaxel (toxins from Taxus
), which stabilize the polymerized microtubule, and the monoterpene indole alkaloids vinblastine, vincristine and vinorelbine (vinca alkaloids from Catharanthus roseus
), which depolymerize microtubules or prevent tubulin assembly [15
The structure and dynamics of actin filaments can be stabilized by the hexapeptide phalloidin (a toxic alkaloid from Amanita phalloides
) or destabilized by the alkaloids cytochalasin (fungal toxins) and latrunculin (toxins from marine sponges) [1
]. Actin filaments, like microtubules, should be a good target for toxins. However, only a few drugs have been developed to target the actin system [1
Due to their intricate ring systems carrying active groups, which have been modified during evolution, alkaloids usually interact with several molecular targets [1
]. Our previous study [18
] has provided evidence that the tryptophan-derived quinoline alkaloids camptothecin and topotecan, which are known DNA topoisomerase I inhibitors, additionally interact with cytoskeleton. Alkaloids with multitarget properties show great potential in the future development of anticancer therapies. However, their corresponding mechanisms are not yet fully or correctly understood.
In this study, six cytotoxic alkaloids (Figure 1
) affecting different targets were tested to explore their effects on microtubules and actin filaments in vivo (Hela: human cervical cancer, MCF-7: human breast cancer and U2OS: human osteosarcoma cancer cells) by use of GFP markers and immunostaining. Furthermore, we analyzed their interaction with tubulin in vitro employing isolated tubulin in a polymerization assay. Modes of action include DNA intercalation and suppression of NF-κB activation (sanguinarine, a benzophenanthridine alkaloid from Sanguinaria canadensis
], DNA intercalation and inhibition of protein kinase c (chelerythrine, a benzophenanthridine alkaloid from Chelidonium majus
], inhibition of telomerase and tubulin (chelidonine, a benzophenanthridine alkaloid from Chelidonium majus
], agonist of σ site and antagonist of BK receptor (noscapine, a phthalide isoquinoline alkaloid from Papaver somniferum
], reduction of intracellular calcium flux (protopine, a benzylisoquinoline alkaloid from Chelidonium majus
] and inhibition of protein synthesis (homoharringtonine, a cephalotaxine alkaloid from Cephalotaxus harringtonia
]. Other microtubule-binding agents, such as colchicine, vinblastine, paclitaxel and the actin-binding agent latrunculin B, were studied as positive controls for comparison. In this study we can provide evidence for partly unknown interactions between these alkaloids and the cytoskeleton.
The present study clarifies the interactions of six alkaloids and four drugs with the elements of the cytoskeleton, such as microtubules and actin filaments. Except homoharringtonine, all other alkaloids apparently affected the dynamics of microtubules, while sanguinarine, chelerythrine and chelidonine affected actin filaments in addition.
Colchicine and vinblastine are microtubule-binding agents (MBAs) that depolymerize microtubules or prevent tubulin assembly. MBAs can alter the dynamic of mitotic spindles during mitosis, which triggers the cell cycle checkpoint and thus arrests the cell cycle in the G2/M phase [10
]. These can explain the effects of colchicine and vinblastine on tubulin polymerization and mitotic spindles observed in our study (Figure 4
, Figure 5
and Figure 6
). Latrunculin B, the actin-binding agent that destabilizes actin filaments by binding to G-actin, did not affect tubulin assembly and mitotic spindles during the study; however, it blocked the cell cycle in the G2/M phase. Cdc25 has been reported to be involved in cell size monitoring via a checkpoint mechanism during mitosis [27
]. Latrunculin B can dramatically alter cell morphology (Figure 2
and Figure 6
), by activating the checkpoint linked to Cdc25, and thus, blocks the cell cycle in the G2/M phase.
Paclitaxel was shown to promote the polymerization of cellular microtubules in living cells and the nucleation of tubulin assembly in vitro (Figure 2
, Figure 3
, Figure 4
and Figure 5
), which agrees with the literature that paclitaxel stabilizes microtubules and inhibits depolymerization by binding along the polymerized microtubule [10
]. However, we found that paclitaxel also affected the growth phase and inhibited tubulin assembly in the presence of MAPs in the tubulin polymerization assay. Paclitaxel can promote tubulin nucleation and assembly in the absence of MAPs and GTP [30
]. In addition, paclitaxel does not replace the MAPs on tubulin and affect microtubule growth if MAPs firstly interact with tubulin during the polymerization [31
]. Hence, we assume that a certain interaction might exist between paclitaxel and free MAPs, which would decrease the concentration of free MAPs and subsequently affect tubulin assembly.
Consistent with recent studies [33
], the benzophenanthridine alkaloids sanguinarine, chelerythrine and chelidonine were found to depolymerize the microtubule network in living cells (Figure 2
, Figure 3
and Figure 4
) and inhibit tubulin polymerization in vitro (Table 2
, Figure 5
). As reported by Wolffand Knipling [33
], sanguinarine, chelerythrine and chelidonine possibly bind to the colchicine site as moderate inhibitors of colchicine binding. Compared to the data of colchicine and vinblastine from our study, however, sanguinarine and chelerythrine did not cause mitotic arrest, and the way by which they affected microtubules is unique. These results agree with the same findings from other authors [35
], which further demonstrates that sanguinarine and chelerythrine act in a different mechanism from other microtubule-binding agents, and cell cycle arrest is not their main mechanism to induce apoptosis. It should be recalled that sanguinarine and chelerythrine can intercalate DNA and via the DNA disturbance induce apoptosis [1
]. The microtubules are composed of tubulin heterodimers that contain 20 accessible cysteine residues [38
]. It is believed that the -SH groups play an important role during the polymerization, and the blocking of -SH groups can lead to the loss of polymerization competence [38
]. Depending on the nucleophilic attack, the iminium bond C=N+ of sanguinarine and chelerythrine is susceptible to interact with tubulin by forming a reversible adduct between the iminium ion and the -SH groups [33
]. This could be the mechanism by which sanguinarine and chelerythrine interact with tubulin and the reason why they differ from other microtubule-binding agents.
However, chelidonine cannot form a pseudobase with tubulin, like sanguinarine and chelerythrine. Chelidonine has been reported to inhibit colchicine binding to tubulin, but does not inhibit podophyllotoxin binding [33
]. Based on our results, chelidonine was shown to be partially similar to colchicine in that they both depolymerized and reduced the mass of cellular microtubules and arrested cell cycle in the G2/M phase (Figure 2
, Figure 3
, Figure 4
and Figure 5
). In addition, our data have revealed a direct relationship between cytotoxicity and cell cycle arrest in colchicine/chelidonine-treated cells (Table 1
, Figure 7
), and the cell cycle arrest may be due to the inhibition of tubulin polymerization. These results raise the possibility that chelidonine inhibits tubulin polymerization by binding to the domain, which is located at the colchicine site on tubulin. Generally, our findings suggest that chelidonine acts as a microtubule-destabilizing agent and that the ability of chelidonine to induce apoptosis mainly depends on its alteration on microtubules.
Besides microtubules, sanguinarine, chelerythrine and chelidonine were found to exhibit a certain effect on cellular actin filaments; in addition, the mass of actin filaments was reduced. Sanguinarine has been reported to alter protein expression of actin 87E and actin-3 [41
] and induced in vitro polymerization of globular actin from rabbit muscle [42
], while chelidonine has been reported to affect cell spreading and reorganization of the actin cytoskeleton in MDA-MB-231 cells [43
]. These published data also indicate the potential interactions between actin filaments and benzophenanthridine alkaloids. However, it is still not clear how they act on actin filaments.
Noscapine appears to act as a microtubule-modulating agent and an antimitotic agent in this study. Noscapine significantly affected the microtubule network and spindles in living cells (Figure 2
, Figure 3
and Figure 4
) and arrested the cell cycle in the G2/M phase (Figure 7
); however, in contrast to other MBAs, such as colchicine, it neither stabilizes nor destabilizes microtubules during polymerization in vitro. Noscapine has been reported to alter the steady state of microtubule dynamics by binding stoichiometrically to tubulin to extend the duration of the attenuated phase without changing the tubulin polymer/monomer ratio, in which the status of microtubules cannot be monitored [44
]. This could explain the observations of our study. Although the alteration of microtubule polymerization does not occur, a slight change of the microtubule dynamics caused by noscapine is enough to activate the spindle checkpoint and induce apoptosis [10
]. Due to its unique effect on microtubules, noscapine does not inhibit the functions of microtubules, such as transportation, and thus, cause low toxicity in normal cells [16
]. Taken together, noscapine can be referred to as an anti-mitotic agent with potential antitumor application in clinical therapies.
Protopine was found to enhance the assembly of microtubules and spindles in living cells (Figure 2
, Figure 3
and Figure 4
) and to block the cell cycle in the G2/M phase (Figure 7
). However, the way protopine acted on microtubules differs from that of paclitaxel. Chen, et al. [48
] also observed the enhancement of microtubule assembly caused by protopine in human hormone refractory prostate cancer (HRPC) cells and proved this promotion through the in vitro tubulin polymerization assay. However, protopine was shown to neither promote nor inhibit tubulin assembly in vitro during our study (Table 2
). How could we explain this? Compared to other alkaloids, we noticed that the mode of action of protopine is similar to that of noscapine. Protopine probably binds to tubulin without affecting its polymerization, which alters microtubule dynamics and activates the spindle checkpoint to block the cell cycle. However, further investigations are required to determine the interactions between protopine and tubulin. Therefore, we report that protopine acts as a novel anti-mitotic agent, which probably alters microtubule dynamics without affecting tubulin polymerization.
Homoharringtonine is a known protein synthesis inhibitor that has been used clinically to treat chronic myeloid leukemia [24
]. Homoharringtonine displayed the second strongest cytotoxicity among the alkaloids during the study. However, in the following experiments, no notable alterations were observed on microtubules, cell cycle and actin filaments after the treatment of homoharringtonine, which suggests that homoharringtonine exerts its cytotoxicity without affecting the pathways related to microtubules, actin filaments and cell cycle, probably by inducing apoptosis.
4. Materials and Methods
Hela human cervical cancer cells were purchased from ATCC (Wesel, Germany), and MCF-7 human breast cancer cells were provided by Prof. Dr. Stefan Wölfl (Institute of Pharmacy and Molecular Biotechnology, Heidelberg University, Heidelberg, Germany); U2OS human osteosarcoma cancer cells, which were stably transfected with an α-tubulin-GFP construct, were supplied by Prof. Dr. Thomas Efferth (Institute of Pharmacy and Biochemistry, Johannes Gutenberg University, Mainz, Germany). Chelerythrine, chelidonine, protopine and homoharringtonine were purchased from Baoji Herbest Bio-Tech Co., Ltd. (Baoji, Shannxi, China); sanguinarine came from Chengdu Biopurify Phytochemicals Ltd. (Chengdu, Sichuan, China); G418, Atto 390 phalloidin, fetal bovine serum (FBS), 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), dimethyl sulfoxide (DMSO), paraformaldehyde, propidium iodide and RNase A, bovine serum albumin (BSA), piperazine-N,N′-bis(2-ethanesulfonic acid) (PIPES), EDTA, EGTA, ATP, GTP and Coomassie blue were obtained from Sigma-Aldrich Chemie GmbH (Steinheim, Germany); and mowiol® 4-88 from Carl Roth GmbH & Co. KG (Karlsruhe, Germany); Dulbecco’s Modified Eagle’s Medium (DMEM), non-essential amino acids, penicillin-streptomycin, CellLight® actin-RFP BacMam 2.0 actin-RFP and trypsin-EDTA came from Life technologies (Paisley, Scotland, UK) and triton X-100 from Merck KgaA (Darmstadt, Germany); mouse anti-α-tubulin monoclonal IgG and goat anti-mouse IgM-FITC were obtained from Santa Cruz Biotechnology (Heidelberg, Germany); 96-well plates, 24-well plates and 6-well plates were purchased from Greiner (Frickenhausen, Germany); and circular glass coverslips from Thermo Scientific (Braunschweig, Germany).
4.2. Cell Culture
Hela and MCF-cells were grown in DMEM with 10% FBS, 1% non-essential amino acids and 1% penicillin-streptomycin at 37 °C and 5% CO2. U2OS human osteosarcoma cancer cells were grown in DMEM with 10% FBS, 1% penicillin streptomycin and continuously treated with 250 μg/mL geneticin at 37 °C and 5% CO2. All experiments were performed with the cells in their logarithmic growth phase.
4.3. Cytotoxicity Assay
The anti-proliferative effects of alkaloids were assessed using the MTT assay, as previously described [50
]. In brief, 1 × 104
cells were seeded in 96-well-plates and incubated with serial dilutions of alkaloids for 48 h (Hela, U2OS) and 72 h (MCF-7). MTT solution was then added and incubated for 2 h. After the addition of DMSO, the plate was read at 570 nm using the Tecan infinite M200 Pro (Tecan, Crailsheim, Germany).
4.4. Imaging of Tubulin-GFP Transfected U2OS Cells
α-Tubulin-GFP U2OS cells (1 × 105) were seeded in 24-well plates and grown for 24 h. Then, 200 μL different concentrations of alkaloids (IC80, IC50) were added into each well, and the cells were imaged under a Keyence BZ-9000 microscope (Keyence, Neu-Isenburg, Germany) after incubation for 2 h, 4 h, 24 h and 48 h. The images were analyzed using BZ-II Analyzer software (version 2.1, Keyence, Neu-Isenburg, Germany).
4.5. Immunofluorescence Staining
Hela cells (1 × 105) were seeded on sterile circular glass coverslips and grown in the absence and presence of different concentrations of alkaloids (IC80, IC50) for 24 h. Then, cells were fixed in 4% paraformaldehyde at 4 °C for 20 min and permeabilized by 0.3% (v/v) Triton X-100 for 10 min. Nonspecific binding sites on the coverslips were blocked by incubating them with 5% FBS for 1 h. To stain the microtubules, cells were firstly incubated with 1:100 mouse anti-α-tubulin monoclonal IgG for 1 h and then with 1:200 goat anti-mouse IgM-FITC for 1 h. Cells were then incubated with 20 nm/mL Atto 390 phalloidin for 45 min to stain the actin filaments and 20 μg/mL propidium iodide for 30 min to stain the cell nucleus/chromosomes. After washing steps, the coverslips were mounted by mowiol. Microtubules, actin filaments and cell nucleus/chromosomes were observed and analyzed using the Keyence BZ-9000 microscope as described before.
4.6. Imaging of Actin-RFP Transfected Hela Cells
Hela cells (2 × 104
) were seeded in 24-well plates and mixed with CellLight®
Actin-RFP BacMam 2.0, which is a fusion construct of human actin and TagRFP, providing an accurate and specific targeting to cellular actin filaments. After 16 h of incubation, 200 μL of different concentrations of alkaloids (IC80
) were added into each well, and the cells were analyzed as described above (Section 4.4
4.7. In-Vitro Tubulin Polymerization Assay
Tubulin plus microtubule-associated proteins (MAPs) were prepared from porcine brains (freshly obtained from a local slaughter house) by two cycles of polymerization and depolymerization according to a standard protocol [51
]. Tubulin polymerization assays were carried out in polymerization buffer (100 mM PIPES, 2 mM EGTA, 0.1 mM EDTA, 3 mM MgCl2
, 1 mM ATP and 1 mM GTP, pH 6.85) by mixing 5.6 mg/mL tubulin-MAPs with different concentrations of alkaloids and incubating at 37 °C for 40 min. The rate and extent of the polymerization reaction were monitored by light scattering at 360 nm using Tecan infinite M200 Pro.
4.8. Cell Cycle Analysis
Cell cycle analysis was carried out as established in our laboratory [52
]. Hela cells (5 × 105
) were seeded in 6-well plates and treated with different concentrations of alkaloids for 24 h. Cells were then collected, centrifuged and fixed in 70% ice-cold ethanol for at least 8 h. After the washing steps, cells were treated with 0.2 mg/mL RNase A for 30 min at 37 °C and then stained with 0.1 mg/mL propidium iodide. Samples were analyzed using a FACScan flow cytometer (Becton Dickinson, Heidelberg, Germany). Data were analyzed using Cell Quest™ Pro software (Becton Dickinson) and Microsoft Excel 2013 (Microsoft Corporation, Redmond, WA, USA).
4.9. Statistical Analysis
The IC50 was determined as the amount of the substances needed to reduce 50% cell viability/tubulin polymerization and calculated from concentration-response curves by Sigmaplot software (Systat Software Inc., San Jose, CA, USA). All experiments were performed at least three times. Data are presented as the mean ± standard deviation. Statistical comparison between controls and different treatments were performed by an unpaired Student’s t-test. Significance was considered at p < 0.05.