Tebuconazole derivatives are potent inhibitors of strigolactone biosynthesis

Strigolactones (SLs) are terpenoid lactones produced in various plant species and were originally found as seed germination stimulants for root parasitic weeds such as Orobanche and Striga, in the 1960s.¹⁾ Later, SLs were also shown as root-derived signal molecules that induce hyphal branching of arbuscular mycorrhizal fungi and as a plant hormone that regulates shoot branching and root morphology in land plants.^2–4) SL biosynthesis and signaling mutants have been identified by genetic analysis of branching mutants in Arabidopsis, peas, petunias and rice. Analysis of these mutants revealed that the biosynthesis of SLs is mediated by one β-carotene isomerase (D27), two carotenoid cleavage dioxygenases, CCD7 (MAX3, RMS5 D17/HTD1, DAD3) and CCD8 (MAX4, RMS1, D10, DAD1) and one cytochrome P450 monooxygenase (MAX1).^5,6)

Controlling phytohormone biosynthesis in plant tissues is important to uncover its physiological role. Specific biosynthetic inhibitors can control the endogenous levels of phytohormones in tissues and organs in various plants and can be useful tools for unveiling new functions of phytohormones. In addition, though a targeted knockout of an individual gene in sets of paralogous genes could affect phenotypes only to a small extent, biosynthetic inhibitors can overcome such gene redundancy in many cases. Therefore, the use of specific biosynthesis inhibitors is a valuable alternative in determining the physiological functions of endogenous substances. As shown by previous studies on gibberellin and brassinosteroid biosynthesis inhibitors,^7–11) specific SL biosynthesis inhibitors would be useful tools for functional studies of SLs in plants. Moreover, SL biosynthesis inhibitors have novel potential for controlling the germination and infestation of root parasitic weeds.

Previously, we screened a chemical library of triazole derivatives to discover novel SL biosynthesis inhibitors because of several triazole-containing chemicals, such as uniconazole-P and propiconazole. These chemicals have previously been shown to act as efficient inhibitors of cytochrome P450 monooxygenases which are involved in gibberellin and brassinosteroid biosyntheses and there is at least one cytochrome P450 (CYP711A) in the proposed SL biosynthesis pathway. In that screening process, we discovered one SL biosynthesis inhibitor, 2,2-dimethyl-7-phenoxy-4-(1H-1,2,4-triazol-1-yl)heptan-3-ol (TIS13).¹²⁾ TIS13-treated rice seedlings showed reduction both in levels of endogenous SLs and in the germination rate of Striga seeds as compared to the untreated controls. However, TIS13 treatment strongly reduced plant height. This growth retardation was not recovered by applying SL, which suggests that TIS13 also inhibits biosynthesis of other growth-promoting phytohormones, such as gibberellin and/or brassinosteroid.

In this report, we screened known cytochrome P450 inhibitors to find new lead chemicals for SL biosynthesis inhibitors (Fig. 1). Among the compounds examined, tebuconazole and bitertanol inhibited epi-5-DS production in rice. We selected tebuconazole as a lead compound and performed a structure-activity relationship (SAR) study.

Fig. 1. Structures of known azole derivatives assayed in this report.

Materials and Methods

1. General

Chemicals for synthesis were purchased from Kanto Chemical Co., Inc. (Tokyo, Japan) and Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). ¹H-NMR spectra were recorded with a JEOL JNM-A500 spectrometer, with chemical shifts being expressed in ppm downfield of TMS as an internal standard. High-resolution mass spectra were recorded on an AB SCIEX TripleTOF 5600 system.

2. Chemistry

1-Phenyl-2-(1H-1,2,4-triazol-1-yl)ethanone (3a) and its derivatives (3b–3m) were prepared as described previously.¹³⁾

4-(4-Chlorophenyl)-2-phenyl-1-(1H-1,2,4-triazol-1-yl)butan-2-ol (4a) was prepared and the Grignard reaction performed as described previously.¹⁴⁾

A solution of 1-phenyl-2-(1H-1,2,4-triazol-1-yl)ethanone (3a) (2.0 mmol) in dichloromethane (4 mL) was added to a suspension of magnesium bromide diethyl etherate (4.0 mmol) in dichloromethane (6 mL) at room temperature and stirred for 0.5 hr. A solution of [3-(4-chlorophenyl)propyl]magnesium bromide in THF was added to ice. After stirring for 1 hr, the mixture was quenched with an aqueous solution of ammonium chloride and extracted with ethyl acetate. The organic layer was dried over anhydrous Na₂SO₄, concentrated and purified by column chromatography.

¹H NMR δ_H (CDCl₃): 1.97–2.04 (1H, m), 2.13–2.19 (1H, m), 2.27–2.33 (1H, m), 2.65–2.71 (1H, m), 4.37 (1H, s), 4.44 (2H, dd J=14.0, 22.5 Hz), 6.99 (2H, d J=8.5 Hz), 7.19 (2H, d J=8.5 Hz), 7.26 (1H, m), 7.33–7.38 (4H, m), 7.76 (1H, s), 7.85 (1H, s); ESI-HRMS m/z (M+H⁺): Calcd. for C₁₇H₂₄Cl₂NO: 328.1229, Found: 328.1230.

2-(2-Chlorophenyl)-4-(4-chlorophenyl)-1-(1H-1,2,4-triazol-1-yl)butan-2-ol (4b) ¹H NMR δ_H (CDCl₃): 2.00–2.05 (1H, m), 2.16–2.22 (1H, m), 2.65–2.78 (2H, m), 4.52 (1H, d J=14.0 Hz), 4.96 (1H, s), 5.24 (1H, d J=14.0 Hz), 7.01–7.03 (2H, m), 7.15–7.20 (4H, m), 7.31–7.34 (1H, m), 7.80 (1H, s), 7.87 (1H, s); ESI-HRMS m/z (M+H+): Calcd. for C₁₈H₁₈Cl₂N₃O: 362.0821, Found: 362.0807.

2-(3-Chlorophenyl)-4-(4-chlorophenyl)-1-(1H-1,2,4-triazol-1-yl)butan-2-ol (4c) ¹H NMR δ_H (CDCl₃): 1.95–2.02 (1H, m), 2.10–2.17 (1H, m), 2.25–2.31 (1H, m), 2.64–2.70 (1H, m), 4.41 (2H, dd J=14.0, 17.5 Hz), 4.97 (1H, s), 6.97 (2H, m), 7.17–7.19 (2H, m), 7.24–7.29 (3H, m), 7.45 (1H, t J=1.5 Hz), 7.77 (1H, s), 7.84 (1H, s); ESI-HRMS m/z (M+H⁺): Calcd. for C₁₈H₁₈Cl₂N₃O: 362.0821, Found: 362.0834.

2,4-bis(4-Chlorophenyl)-1-(1H-1,2,4-triazol-1-yl)butan-2-ol (4d) ¹H NMR δ_H (CDCl₃): 1.96–2.02 (1H, m), 2.10–2.16 (1H, m), 2.24–2.30 (1H, m), 2.63–2.69 (1H, m), 4.41 (2H, dd J=14.0, 20.0 Hz), 4.86 (1H s), 6.97 (2H, d J=8.5 Hz), 7.18 (2H, d J=8.5 Hz), 7.30–7.36 (4H, m), 7.77 (1H, s), 7.82 (1H, s); ESI-HRMS m/z (M+H⁺): Calcd. For C₁₈H₁₈Cl₂N₃O: 362.0821, Found: 362.0832.

4-(4-Chlorophenyl)-2-(4-phenoxyphenyl)-1-(1H-1,2,4-triazol-1-yl)butan-2-ol (4e) ¹H NMR δ_H (CDCl₃): 1.95–2.01 (1H, m), 2.12–2.18 (1H, m), 2.33–2.39 (1H, m), 2.65–2.72 (1H, m), 4.42 (2H, dd J=14.0, 22.0 Hz), 4.54 (1H, s), 6.94–6.97 (2H, m), 6.99–7.01 (4H, m), 7.11–7.14 (1H, m), 7.18–7.20 (2H, m), 7.31–7.37 (4H, m), 7.82 (1H, s), 7.83 (1H, s); ESI-HRMS m/z (M+H⁺): Calcd. for C₂₄H₂₃Cl₂N₃O₂: 420.1473, Found: 420.1461. 4-(4-Chlorophenyl)-2-p-tolyl-1-(1H-1,2,4-triazol-1-yl)butan-2-ol (4f) ¹H NMR δ_H (CDCl₃): 1.94–2.00 (1H, m), 2.11–2.17 (1H, m), 2.28–2.34 (4H, m), 2.63–2.69 (1H, m), 4.37–4.44 (3H, m), 6.98 (2H, dd J=2.0, 6.5 Hz), 7.14 (2H, d J=8.0 Hz), 7.16–7.19 (2H, m), 7.24 (2H, dd J=2.0, 6.5 Hz), 7.76 (1H, s), 7,81 (1H, s); ESI-HRMS m/z (M+H⁺): Calcd. for C₁₉H₂₁ClN₃O: 342.1368, Found: 342.1377.

4-(4-Chlorophenyl)-2-(2,4-dichlorophenyl)-1-(1H-1,2,4-triazol-1-yl)butan-2-ol (4g) ¹H NMR δ_H (CDCl₃): 1.97–2.23 (2H, m), 2.69 (2H, dd J=12.5, 21.5 Hz), 4.50 (1H, d J=14.0 Hz), 5.02 (1H, s), 5.22 (1H, d J=14.0 Hz), 7.01–7.03 (2H, m), 7.16–7.20 (3H, m), 7.33 (1H, d J=2.0 Hz), 7.64 (1H, d J=8.5 Hz), 7.84 (1H, s), 7.90 (1H, s); ESI-HRMS m/z (M+H⁺): Calcd. for C₁₈H₁₇Cl₃N₃O: 396.0432, Found: 396.0437.

4-(4-Chlorophenyl)-2-(4-methoxyphenyl)-1-(1H-1,2,4-triazol-1-yl)butan-2-ol (4h) ¹H NMR δ_H (CDCl₃): 1.94–2.00 (1H, m), 2.10–2.16 (1H, m), 2.29–2.35 (1H, m), 2.63–2.68 (1H, m), 3.77 (3H, s), 4.39 (2H, dd J=14.5, 19.0 Hz), 4.69 (1H, s), 6.86 (2H, d J=9.0 Hz), 6.97 (2H, d J=8.5 Hz), 7.17 (2H, d J=8.5 Hz), 7.28 (2H, d J=9.0 Hz), 7.76 (1H, s), 7.78 (1H, s); ESI-HRMS m/z (M+H⁺): Calcd. for C₁₉H₂₁ClN₃O₂: 358.1317, Found: 358.1334.

4-(4-Chlorophenyl)-2-(2,4-difluorophenyl)-1-(1H-1,2,4-triazol-1-yl)butan-2-ol (4i) ¹H NMR δ_H (CDCl₃): 1.98–2.06 (1H, m), 2.23–2.32 (2H, m), 2.69–2.75 (1H, m), 4.47 (1H, d J=14.0 Hz), 4.74 (1H, d J=14.0 Hz), 4.98 (1H, s), 6.74–6.82 (2H, m), 7.00 (2H, d J=8.0 Hz), 7.19 (2H, d J=8.0 Hz), 7.48–7.53 (1H, m), 7.81 (1H, s), 7.89 (1H, s); ESI-HRMS m/z (M+H⁺): Calcd. for C₁₈H₁₇ClF₂N₃O: 364.1028, Found: 364.1037.

4-(4-Chlorophenyl)-2-m-tolyl-1-(1H-1,2,4-triazol-1-yl)butan-2-ol (4j) ¹H NMR δ_H (CDCl₃): 1.93–2.00 (1H, m), 2.13–2.19 (1H, m), 2.28–2.34 (4H, m), 2.62–2.68 (1H, m), 4.41 (2H, dd J=14.0, 19.0 Hz), 4.81 (1H, s), 6.96 (2H, d J=8.0 Hz), 7.07 (1H, d J=7.0 Hz), 7.14–7.18 (3H, m), 7.22 (2H,d J=7.0 Hz), 7.75 (1H, s), 7.78 (1H, s); ESI-HRMS m/z (M+H⁺): Calcd. for C₁₉H₂₁ClN₃O: 342.1368, Found: 342.1379.

4-(4-Chlorophenyl)-2-(4-fluorophenyl)-1-(1H-1,2,4-triazol-1-yl)butan-2-ol (4k) ¹H NMR δ_H (CDCl₃): 1.96–2.03 (1H, m), 2.11–2.17 (1H, m), 2.26–2.32 (1H, m), 2.63–2.69 (1H, m), 4.41 (2H, dd J=14.0, 24.0 Hz), 4.79 (1H, s), 6.96–6.99 (2H, m), 7.00–7.05 (2H, m), 7.17–7.20 (2H, m), 7.34–7.38 (2H, m), 7.78 (1H, s), 7.81 (1H, s); ESI-HRMS m/z (M+H⁺): Calcd. for C₁₈H₁₈ClFN₃O: 346.1117, Found: 346.1125.

2-(4-Bromophenyl)-4-(4-chlorophenyl)-1-(1H-1,2,4-triazol-1-yl)butan-2-ol (4l) ¹H NMR δ_H (CDCl₃): 1.95–2.01 (1H, m), 2.09–2.15 (1H, m), 2.24–2.30 (1H, m), 2.64–2.70 (1H, m), 4.41 (2H, dd J=14.0, 21.0 Hz), 4.73 (1H, s), 6.96–6.99 (2H, m), 7.17–7.20 (2H, m), 7.25–7.28 (2H, m), 7.45–7.48 (2H, m), 7.80 (1H, s), 7.81 (1H, s); ESI-HRMS m/z (M+H⁺): Calcd. for C₁₈H₁₈BrClN₃O: 406.0316, Found: 406.0309. 4-(4-Chlorophenyl)-1-(1H-1,2,4-triazol-1-yl)-2-(4-(trifluoromethyl)phenyl)butan-2-ol (4m) ¹H NMR δ_H (CDCl₃): 1.99–2.05 (1H, m), 2.15–2.28 (2H, m), 2.65–2.71 (1H, m), 4.45 (2H, dd J=14.5, 24.0 Hz), 4.93 (1H, s), 6.96–6.98 (2H, m), 7.17–7.20 (2H, m), 7.54 (2H, d J=8.0 Hz), 7.61 (2H, J=8.0 Hz), 7.79 (1H, s), 7.83 (1H, s); ESI-HRMS m/z (M+H⁺): Calcd. for C₁₉H₁₈ClF₃N₃O: 396.1085, Found: 396.1091.

3. SL analysis

A rice culture and SL analysis were performed as described in the previous paper.¹³⁾

Results and Discussion

Analysis of the level of epi-5DS in rice root exudates is a convenient method for evaluating SL biosynthesis inhibitors.¹³⁾ Therefore, we estimated the inhibitory activity of SL biosynthesis by analyzing the level of epi-5DS in rice root exudates.

Extensive studies on various cytochrome P450 inhibitors strongly suggest that the azole moiety of the inhibitors should act as a ligand to bind to the iron atom of the heme prosthetic group of the cytochrome P450 enzyme.¹⁵⁾ In many cases, azole derivatives inhibit various P450s. For example, uniconazole-P, which is known as a gibberellin biosynthesis inhibitor, was reported to inhibit brassinosteroid biosynthesis and abscisic acid metabolism.^16,17) Thus, chemical modifications of azole derivatives should be important for increasing the specificity of P450 inhibitors to the target cytochrome P450. To find new lead chemicals for SL biosynthesis inhibitors, we investigated the inhibitory potency of several triazole and imidazole derivatives that are on the market. Among the tested chemicals, tebuconazole and bitertanol reduced epi-5DS levels in root exudates more effectively than TIS13 (Fig. 2). Tebuconazole and bitertanol are fungicides that target lanosterol 14α-demethylase in the ergosterol biosynthesis pathway.^18,19) Tebuconazole is also reported to show plant growth retardation activity in cress.²⁰⁾ Tebuconazole derivatives were not found in the in house triazole library that we previously constructed,¹²⁾ but the activity of some bitertanol derivatives was estimated in our previous report.¹²⁾ Therefore, we selected tebuconazole as a lead chemical for SL biosynthesis inhibitors and performed a SAR study in this report. Chemicals were synthesized as described in Fig. 3. Since we could not detect epi-5DS in root exudates of rice seedlings treated with 10 µM tebuconazole (Fig. 2), we evaluated the inhibitory activity of tebuconazole and its derivatives against epi-5DS production in rice at 5 µM (Fig. 4A and B). Substitution of the tert-butyl group of tebuconazole with a phenyl group (4a) did not affect the inhibitory activity against epi-5DS production, but several 4-substituted phenyl derivatives, such as chloro (4d), phenoxy (4e), methyl (4f), fluoro (4k), and bromo (4l) derivatives, exhibited higher inhibitory activity against epi-5DS production. However, except for 4e, the data are not statistically significant. Substitution at position 2 of the phenyl group by a chlorine atom (4b) strongly increased the inhibitory activity against epi-5DS production and caused slight dwarfism in rice. Introduction of two chlorines to the phenyl ring both at positions 2 and 4 made compound 4g completely ineffective in the inhibition of epi-5DS production. As a result, among the 13 analogues synthesized in this study, only 4b and 4e showed a significant inhibitory effect on epi-5DS production.

Fig. 2. Effects of azole derivatives on epi-5DS production. Inhibitors were added at 10 µM. The data are means±SD of three biological replicates. All replicates include three seedlings. (Abbreviations and their meanings: PPI, propiconazole; DIF, difenoconazole; BIT, bitertanol; TAF, triadimefon; KETO, ketoconazole; TEB, tebuconazole; TAN, triadimenol; ND, not detected.)

Fig. 3. Synthesis of designed chemicals.

Fig. 4. Effects of chemicals on rice seedlings. (A) and (B) indicate epi-5DS levels in root exudates of 5 µM chemically treated rice seedlings determined by LC-MS/MS. The data are means±SD of three biological replicates. All replicates include three seedlings. The symbol * indicates means that are statistically different from the mean of the control treatment as determined by one-way ANOVA with the Dunnett test (p<0.05). (C) The length of the 2nd leaf sheath in 50 µM chemically treated rice seedlings. The data are means±SD of nine samples. The symbols * and ** indicate means that are statistically different from the mean of the tebuconazole treatment as determined by one-way ANOVA with the Dunnett test (p<0.05 and p<0.01, respectively).

There are many P450s involved in the biosynthesis and metabolism of phytohormones. Therefore there is a possibility that azole derivatives examined in this study may affect some of the P450s involved in these hormone biosynthesis pathways and reduce plant height. We evaluated the effect of these derivatives on plant height by measuring the 2nd leaf sheath in rice. Tebuconazole strongly reduced the length of the 2nd leaf sheath at 50 µM (Fig. 4 bottom). Most of the other azole derivatives examined in this study also reduced the length of the 2nd leaf sheath at 50 µM. The inhibitory activity of 4-substituted phenyl derivatives decreased in order as follows: chloro (4d)=bromo (4l)>fluoro (4k)>trifluoromethyl (4m)=methyl (4f)=methoxy (4h)=phenoxy (4e). These results suggest the possibility that the electron withdrawing lipophilic substituents at the 4-position of the phenyl group may be important for the dwarfism inducing activity. Among the 13 analogues synthesized and examined for their dwarfism-inducing activity in this study, compound 4e exhibited the weakest activity. From these results (Fig. 4), we conclude that 4e is the most specific SL biosynthesis inhibitor among the chemicals tested here.

In this report, we showed that tebuconazole and some of its derivatives reduce strigolactone production, although the target site(s) of the azole derivatives is currently unclear. Several azole-type chemicals, such as uniconazole-P, target various P450s, and CYP711A1 (MAX1) can be a good candidate as a target of these chemicals. Recently, Alder et al. reported that three enzymes, D27, CCD7, and CCD8, produce an SL-like chemical, carlactone from β-carotene in vitro and proposed an SL biosynthesis pathway.²¹⁾ According to this report, carlactone may be subsequently converted to strigolactone by MAX1 and/or an unknown enzyme or enzymes. Carlactone analysis in chemically treated plants will be of great help in identifying the target site(s).

The data obtained in this research suggest strongly that various triazole and imidazole derivatives can be a new scaffold for developing potent inhibitors of SL biosynthesis because all azole derivatives tested in Fig. 2 inhibited the production of epi-5DS at 10 µM. Further studies on structure-activity relationships will lead to the discovery of specific SL biosynthesis inhibitors with greater potency. The results thus obtained here will play an important role in the design of new SL biosynthesis inhibitors.

References

• 1)  C. E. Cook, L. P. Whichard, B. Turner, M. E. Wall and G. H. Egley: Science 154, 1189–1190 (1966).
• 2)  K. Akiyama, K. Matsuzaki and H. Hayashi: Nature 435, 824–827 (2005).
• 3)  V. Gomez-Roldan, S. Fermas, P. B. Brewer, V. Puech-Pagès, E. A. Dun, J. P. Pillot, F. Letisse, R. Matusova, S. Danoun, J. C. Portais, H. Bouwmeester, G. Bécard, C. A. Beveridge, C. Rameau and S. F. Rochange: Nature (London) 455, 189–194 (2008).
• 4)  M. Umehara, A. Hanada, S. Yoshida, K. Akiyama, T. Arite, N. Takeda-Kamiya, H. Magome, Y. Kamiya, K. Shirasu, K. Yoneyama, J. Kyozuka and S. Yamaguchi: Nature 455, 195–200 (2008).
• 5)  C. A. Beveridge and J. Kyozuka: Curr. Opin. Plant Biol. 13, 34–39 (2010).
• 6)  R. S. Drummond, N. M. Martínez-Sánchez, B. J. Janssen, K. R. Templeton, J. L. Simons, B. D. Quinn, S. Karunairetnam and K. C. Snowden: Plant Physiol. 151, 1867–1877 (2009).
• 7)  E. Nambara, K. Keith, P. McCourt and S. Naito: Plant Cell Physiol. 35, 509–513 (1994).
• 8)  T. Asami, Y. K. Min, N. Nagata, K. Yamagishi, S. Takatsuto, S. Fujioka, N. Murofushi, I. Yamaguchi and S. Yoshida: Plant Physiol. 123, 93–100 (2000).
• 9)  Y. Yin, Z. Y. Wang, S. Mora-Garcia, J. Li, S. Yoshida, T. Asami and J. Chory: Cell 109, 181–191 (2002).
• 10)  T. Asami, T. Nakano, H. Nakashita, K. Sekimata, Y. Shimada and S. Yoshida: J. Plant Growth Regul. 22, 336–349 (2003).
• 11)  T. Komatsu, H. Kawaide, C. Saito, A. Yamagami, S. Shimada, M. Nakazawa, M. Matsui, A. Nakano, M. Tsujimoto, M. Natsume, H. Abe, T. Asami and T. Nakano: Plant J. 61, 409–422 (2010).
• 12)  S. Ito, N. Kitahata, M. Umehara, A. Hanada, A. Kato, K. Ueno, K. Mashiguchi, J. Kyozuka, K. Yoneyama, S. Yamaguchi and T. Asami: Plant Cell Physiol. 51, 1143–1150 (2010).
• 13)  S. Ito, M. Umehara, A. Hanada, N. Kitahata, H. Hayase, S. Yamaguchi and T. Asami: PLoS ONE 6, e21723 (2011).
• 14)  H. Itoh, H. Kajino, T. Tsukiyama, J. Tobitsuka, H. Ohta, Y. Takahi, M. Tsuda and H. Takeshiba: Bioorg. Med. Chem. 10, 4029–4034 (2002).
• 15)  Y. Yoshida and Y. Aoyama: Biochem. Pharmacol. 36, 229–235 (1987).
• 16)  Y. K. Min, T. Asami, S. Fujioka, N. Murofushi, I. Yamaguchi and S. Yoshida: Bioorg. Med. Chem. Lett. 9, 425–430 (1999).
• 17)  S. Saito, M. Okamoto, S. Shinoda, T. Kushiro, T. Koshiba, Y. Kamiya, N. Hirai, Y. Todoroki, K. Sakata, E. Nambara and M. Mizutani: Biosci. Biotechnol. Biochem. 70, 1731–1739 (2006).
• 18)  H. Ishii, H. Takeda, Y. Nagamatsu and H. Nakashima: Pestic. Sci. 30, 405–413 (1990).
• 19)  R. Akallal, D. Debieu, C. Lanen, M. Daboussi, R. Fritz, C. Malosse, J. Bach and P. Leroux: Pestic. Biochem. Physiol. 60, 147–166 (1998).
• 20)  K. Sekimata, S. Y. Han, K. Yoneyama, Y. Takeuchi, S. Yoshida and T. Asami: J. Agric. Food Chem. 50, 3486–3490 (2002).
• 21)  A. Alder, M. Jamil, M. Marzorati, M. Bruno, M. Vermathen, P. Bigler, S. Ghisla, H. Bouwmeester, P. Beyer and S. Al-Babili: Science 335, 1348–1351 (2012).