English

• Among biomedical imaging technologies, ¹⁹F MRI stands out as a valuable tracer technology for in vivo targets [1–4]. The absence of detectable in vivo ¹⁹F signals in biological systems allows quantitative and selective "hot spot" imaging of targets without background interference, ionizing radiation, and tissue depth limit [5–7]. Despite its great potential, ¹⁹F MRI has not yet reached clinical application due to the lack of sensitive ¹⁹F MRI agents [6]. Since ¹⁹F signals originate only from ¹⁹F MRI agents, a relatively high local ¹⁹F concentration of about 5 mmol/L is usually required to generate "hot spot" images within a few minutes of data acquisition. For regular ¹⁹F MRI agents with ¹⁹F signal splitting and low longitudinal relaxation rates (R₁), even higher ¹⁹F concentration and longer data acquisition time are required. For in vivo studies, it is very challenging to achieve such high ¹⁹F concentrations and to keep the animals alive and still for a long time. Therefore, there is an urgent need for highly sensitive imaging agents to reduce the ¹⁹F MRI agent dosage and shorten the data acquisition time.

  Two major strategies have been developed to improve the sensitivity of ¹⁹F MRI agents. First, the assembly of multiple equivalent fluorines can avoid ¹⁹F signal splitting and generate a unified strong signal for ¹⁹F MRI. Perfluoro–tert-butanol derivatives [8], perfluoro-15-crown-5 [9], and perfluoropolyethers [10] are among the most successful ones. However, the more equivalent fluorines require the more complicated synthesis and the higher cost. Second, extending the ¹⁹F R₁ can significantly shorten the data acquisition time and acquire more ¹⁹F signals in a given time. Many methods have been established to extend the ¹⁹F R₁, such as the introduction of paramagnetic metal ions [11,12], conjugation to large molecules [13], self-assembly into aggregates [14,15]. However, these strategies are often hampered by the delicate design and tedious synthesis, the toxicity of paramagnetic metal ions, the ¹⁹F signal quenching by the severely extended transverse relaxation rate (R₂), etc. Thus, developing novel strategies to extend the ¹⁹F R₁ of metal-free agents with multiple equivalent fluorines may yield highly sensitive ¹⁹F MRI agents.

  [2]Rotaxanes [16–19], a class of interlocked molecules, have recently attracted our attention. Compared to the relatively free motion of wheel molecules, their motion in [2]rotaxanes is restricted to mechanical bond-guided shuttling and tumbling [20–22]. Although, to our knowledge, the relaxation behavior of rotaxanes has never been investigated, it is reasonable to predict that the R₁ and R₂ of the wheel fluorines can be significantly extended after threading into [2]rotaxanes, resulting in a significant improvement in ¹⁹F MRI sensitivity. To validate this novel strategy for improving ¹⁹F MRI agent sensitivity, we have herein designed 2-blade pinwheel [2]rotaxanes Rx-1 and Rx-2 as sensitive ¹⁹F MRI agents and investigated the role of the threading in modulating the motion of wheel fluorines (Scheme 1). With 36 equivalent fluorines to generate a unified strong ¹⁹F signal for ¹⁹F MRI, fluorinated macrocycle 1 was used both as a regular ¹⁹F MRI agent and as the wheel of [2]rotaxanes. Since the ¹⁹F R₁ and R₂ are strongly influenced by molecular dynamics, the fluorines act not only as ¹⁹F MRI signal sources but also as sensitive status reporters, providing an unprecedented strategy to investigate the mechanical bond and molecular dynamics of [2]rotaxanes, which is a very important and challenging task for rotaxanes [23–25].

  Scheme 1

  

  Scheme 1.  The strategy of improving ¹⁹F MRI agent sensitivity by threading the macrocycle 1 into [2]rotaxanes Rx-1 and Rx-2.

  DownLoad: Full-Size Img PowerPoint

  First, we prepared macrocycle 1 through a bromomethylation-perfluoro–tert-butoxylation strategy, which was then supramolecularly assembled into Rx-1 and Rx-2 through a one-pot thread-cap strategy on multi-hundred-milligram scales (see Supporting information for details) [26]. Despite its bulky size, the 4 perfluoro–tert–butyl groups at the far end of 1 did not hinder the threading and the [2]rotaxanes were conveniently prepared in good yields. Meanwhile, capped axles A₁ and A₂ were also prepared as reference standards.

  Subsequently, 1D and 2D ¹H NMR were used to verify the formation of Rx-1 and Rx-2. By comparing the ¹H NMR spectra, the typical chemical shift changes (Δδ) of the corresponding protons confirmed the formation of the [2]rotaxanes (Figs. 1a-e). Compared with axle A₁, the downfield shift of proton H_h in Rx-1 indicated the formation of hydrogen bonds between the crown ether and the positively charged amine, while the upfield shifts of protons H_A, H_e, and H_g indicated π-π stacking of the wheel and axle aromatic groups. Similar chemical shift changes were also observed in the ¹H NMR spectrum of Rx-2. Furthermore, the stationing of the wheel in both [2]rotaxanes was verified by the appearance of cross-peaks between wheel protons H_C, H_D, H_E and axle protons H_d, H_h, H_c’, H_h’ in their 2D ROESY ¹H NMR spectra (Fig. 1g and Fig. S1 in Supporting information). To investigate the "shuttling" motion of the wheel, we removed the hydrogen bonds in the [2]rotaxanes with sodium hydroxide. However, considerable free 1 was detected in the reaction mixture of Rx-1, suggesting that the 3,5-bis(trifluoromethyl) benzyl group was not bulky enough to lock 1. The bulkier 3,5-di–tert-butylbenzyl group successfully locked 1 and provided Rx-2′. The complicated ¹H NMR spectrum of Rx-2′showed that the wheel "shuttled" from the central amine to one of the triazole groups and the corresponding protons on either side became non-equivalent (Figs. 1d and f). For example, the tert–butyl protons H_i’ split into two equivalent peaks at 1.27 ppm (H_ia’) and 1.24 ppm (H_ib’, see Supporting information for details). The upfield shifts of protons H_E, H_c’, H_h’ and the downfield shifts of protons H_a’, H_g’ suggest that the "shuttling" is probably driven by the π-π interaction between the wheel benzene and the axle triazole.

  Figure 1

  

  Figure 1.  Partial ¹H NMR spectra of the axles, macrocycle, and [2]rotaxanes (a-f); 2D ROESY ¹H NMR spectrum of Rx-1 (g); single-crystal X-ray structure of 1 (h) and Rx-1 (i); partial ¹⁹F NMR spectra of 1 (j), A₁ (k), Rx-1 (l), Rx-2 (m), Rx-2′ (n). NMR conditions: 500 MHz, 1 mmol/L, 298 K, CD₃CN. The labeling of ¹H and ¹⁹F can be found in Scheme 1 and Scheme S2 (Supporting information).

  DownLoad: Full-Size Img PowerPoint

  Furthermore, we obtained the single-crystal X-ray structure of 1 and discovered a "wide open" conformation (Fig. 1h), providing enough space for the threading. Consistent with the ¹H NMR results, the single-crystal X-ray structure of Rx-1 showed the hydrogen bonds between the wheel and the axle (dotted lines in Fig. 1i) and an "open" conformation of the wheel to accommodate the axle. The X-ray data also showed the π-π stacking between the wheel and the axle, with the distances between adjacent aromatic groups about 0.36–0.40 nm. Notably, the mechanical bond, hydrogen bonds, and π-π stacking in Rx-1 resulted in a twisted "Z" conformation of the wheel and a broad "W" conformation of the axle.

  Meanwhile, ¹⁹F NMR was employed to investigate the structure of the [2]rotaxanes (Figs. 1j-n). As designed, the 36 wheel fluorines in Rx-1 and Rx-2 gave a unified sharp ¹⁹F peak at −71.3 ppm (F_A, Figs. 1l and m), while the 12 axle fluorines in Rx-1 gave another unified sharp ¹⁹F peak at −64.0 ppm (F_B, Fig. 1l). The unified ¹⁹F peaks indicate the rapidly interchangeable and equivalent positions of the fluorines, even though the X-ray structure of Rx-1 shows an unsymmetrical conformation with a crowded π-π stacking side and a loose side. Furthermore, the similar ¹⁹F diffusion coefficients (D) of the wheel (F_A: D = 8.75 × 10⁻¹⁰ m²/s) and the axle (F_B: D = 8.74 × 10⁻¹⁰ m²/s) in Rx-1 suggest their synchronous motion as a whole molecule (Table S1 in Supporting information). Compared to 1 (F_A: D = 1.17 × 10⁻⁹ m²/s) and axle A₁ (F_B: D = 1.14 × 10⁻⁹ m²/s), the threading was accompanied by a significant decrease in diffusion coefficients (F_A: 25% reduction, F_B: 23% reduction) as a result of forming a larger molecule.

  After elucidating the [2]rotaxane structures, we comparatively studied the effect of the threading on relaxation rates (R). Compared to 1, the [2]rotaxanes showed similar relaxation rate changes (∆R) in wheel fluorines F_A (Rx-1: ∆R₁ = 30%, ∆R₂ = 40%; Rx-2: ∆R₁ = 28%, ∆R₂ = 38%, Fig. 2a), indicating that the motion of F_A was significantly restricted by the threading. Meanwhile, compared to A₁ and A₂, the [2]rotaxanes showed similar ∆R in axle fluorines F_B (Rx-1: ∆R₁ = 8%, ∆R₂ = 9%) and protons H_i’ (Rx-2: ∆R₁ = 13%, ∆R₂ = 11%, Fig. 2b). Notably, the threading resulted in less ∆R in axle fluorines, suggesting that the axle motion is less constrained by the threading. We also investigated the effect of hydrogen bonds and molecular weight on the F_A R. Surprisingly, Rx-2′with no hydrogen bond and lower molecular weight gave almost identical F_A R as Rx-2 (Fig. 2a), illustrating their minor influence on the wheel motion. Interestingly, opposite ∆R was observed in the tert–butyl protons of Rx-2′, i.e., substantial increases on the wheel side (H_ib’: ∆R₁ = 14%, ∆R₂ = 28%) and slight changes on the other side (H_ia’: ∆R₁ = −3%, ∆R₂ = 2%, Fig. 2b). These observations show that the threading can efficiently improve the ¹⁹F relaxation rates of the wheel, in which the mechanical bond dominantly constrains the motion of the fluorines and protons in a distance-dependent manner. Since molecular motion is closely related to temperature, the impact of temperature on the relaxation rates was investigated (Figs. 2c and d). Upon heating the [2]rotaxanes from 274 K to 318 K, the wheel fluorines F_A R decreased significantly in a similar trend (∆R₁ ≈ −35%, ∆R₂ ≈ −43%), whereas those of axle fluorines F_B decreased only slightly (∆R₁ = −6%, ∆R₂ = −9%), showing that heating promotes the motion of the wheel much more than that of the axle, probably due to the distance-dependent effect of the mechanical bond. It is noteworthy that the perfect proportional relationship between wheel ¹⁹F R and temperature makes the [2]rotaxanes valuable temperature probes.

  Figure 2

  

  Figure 2.  R₁/R₂ of F_A (a), F_B and H_i (b) in the axles, macrocycle, and [2]rotaxanes; temperature-dependent ¹⁹F R₁ (c) and R₂ (d) of the [2]rotaxanes. Drmsd of Rx-1 (upper) and 1 (lower) over time (e), (f) ¹⁹F MRI phantom images (9.4 T, 298 K, CH₃CN) and the plot of logSI versus logC(¹⁹F) (g) of 1, Rx-2 and Rx-2′. Statistical significance: ^*P < 0.1 and ^***P < 0.001.

  DownLoad: Full-Size Img PowerPoint

  To investigate the origin of the ∆R, molecular dynamics simulations were performed on 1 and Rx-1. The simulations showed that the distance root-mean-square deviation (Drmsd) of wheel fluorines F_A relative to the X-ray structure of 1 changed rapidly and significantly over time (Fig. 2e), indicating that F_A underwent rapid and intense motion. In contrast, the Drmsd of F in Rx-1 changed much slower and smaller, suggesting that the motion of F_A was significantly constrained after the threading. Therefore, the ¹⁹F ∆R is probably a result of the wheel motion restricted by the mechanical bond.

  In addition, solid-state ¹⁹F magic angle spinning (MAS) NMR was performed on 1 and Rx-2 (Fig. S7 in Supporting information) to obtain the rotational correlation time (τ_c), which is an important parameter characterizing the internal rotations of a given group [27]. It was found that the τ_c of Rx-2 is about 2.2-fold longer than that of 1 (Fig. S8 in Supporting information), indicating that the rotation of F_A in Rx-2 is significantly restricted by the threading. In addition, the broader ¹³C NMR linewidth in Rx-2 than that of 1 further confirms the slower rotational motion of F_A in Rx-2 (Fig. S9 in Supporting information). Thus, the rotational motions of wheel fluorines F_A were severely restricted by the threading, resulting in a significant ∆R.

  With the significantly increased R₁ and a unified ¹⁹F signal from 36 fluorines, the ¹⁹F MRI capability of Rx-2 and Rx-2′ was evaluated. For ¹⁹F MRI agents, short longitudinal relaxation times (T₁, T₁ = 1/R₁) and high T₂/T₁ ratios (T₂ = 1/R₂) are highly preferred to improve sensitivity by providing an intense ¹⁹F peak for rapid data acquisition [28]. The significantly reduced T₁ and high T₂/T₁ ratios of 0.69 establish Rx-2 and Rx-2′ as valuable ¹⁹F MRI agents. Using the T₁-weighted ¹⁹F MRI, Rx-2 and Rx-2′ generated high contrast "hot spot" images with a short data acquisition time of 307 s at a low concentration of 62.5 µmol/L (Fig. 2f), which is beyond the reach of most ¹⁹F MRI agents. For instance, previous studies have reported the ¹⁹F MRI of perfluorinated erlotinib and gefitinib analogues at concentrations as low as 10 mmol/L [29]. Additionally, a fluorinated peptide with the lowest detectable concentration for ¹⁹F MRI was reported to be 0.13 mmol/L [30]. As expected, threading 1 into Rx-2 and Rx-2′ significantly improved the ¹⁹F MRI sensitivity with 77%−79% signal intensity (SI) enhancement. Notably, the SI enhancement is more significant over the range of in vivo drug concentrations, making the [2]rotaxanes highly sensitive ¹⁹F MRI agents for potential in vivo applications. In each case, the logarithm of SI is proportional to the logarithm of fluorine concentration (Fig. 2g), facilitating quantitative ¹⁹F MRI. Since mechanical bond and hydrogen bond are formed during the threading, their effects were detected by the SI enhancement, i.e., mechanical bond and hydrogen bond induced 33% SI enhancement in Rx-2 and mechanical bond induced 29% SI enhancement in Rx-2′. Therefore, introducing mechanical bonds is an effective and robust strategy to improve ¹⁹F MRI sensitivity.

  In conclusion, we have developed a novel strategy to improve the sensitivity of ¹⁹F MRI agents. On the one hand, the introduction of mechanical bond into ¹⁹F MRI agents has been demonstrated to restrict the motion of the fluorines, shorten the T₁, maintain high T₂/T₁ ratios, and thus dramatically improve the sensitivity, providing a novel strategy to address the sensitivity issue of ¹⁹F MRI. On the other hand, fluorinated rotaxanes are rare and their potential in ¹⁹F MRI has never been explored. We have unprecedentedly synthesized a series of fluorinated 2-blade pinwheel [2]rotaxanes and employed ¹⁹F NMR/MRI to investigate the structure and molecular dynamics of [2]rotaxanes, establishing ¹⁹F NMR/MRI as a valuable technology for sensing mechanical bonds in molecular devices. This study successfully integrated ¹⁹F MRI and rotaxanes, which may promote the development of highly sensitive and stimuli-responsive ¹⁹F MRI agents, such as temperature probes, and self-status-reporting high-performance molecular devices.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  Acknowledgments

  This work was supported by the Strategic Priority Research Program of the Chinese Academy of Sciences (No. XDB0540000), the National Key R & D Program of China (No. 2018YFA0704000), the National Natural Science Foundation of China (Nos. 22327901, 22077098, U21A20392, 21921004, and 82127802), and the Knowledge Innovation Program of Wuhan-Basic Research (No. 2022020801010137). Shizhen Chen acknowledges the support from the Youth Innovation Promotion Association and the Young Top-notch Talent Cultivation Program.

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.109512.

  
  
• 1. [1]

     M. Higuchi, N. Iwata, Y. Matsuba, et al., Nat. Neurosci. 8 (2005) 527–533. doi: 10.1038/nn1422

  2. [2]

     M.J. Couch, I.K. Ball, T. Li, et al., Magn. Reson. Imaging 49 (2019) 343–354. doi: 10.1002/jmri.26292

  3. [3]

     P. Bouvain, S. Temme, U. Flögel, WIREs Nanomed. Nanobiotechnol. 12 (2020) e1639. doi: 10.1002/wnan.1639

  4. [4]

     H. Lin, X. Tang, A. Li, J. Gao, Adv. Mater. 33 (2021) 2005657. doi: 10.1002/adma.202005657

  5. [5]

     S. Kunjachan, J. Ehling, G. Storm, F. Kiessling, T. Lammers, Chem. Rev. 115 (2015) 10907–10937. doi: 10.1021/cr500314d

  6. [6]

     I. Tirotta, V. Dichiarante, C. Pigliacelli, et al., Chem. Rev. 115 (2015) 1106–1129. doi: 10.1021/cr500286d

  7. [7]

     J. Ruiz-Cabello, B.P. Barnett, P.A. Bottomley, J.W.M. Bulte, NMR Biomed. 24 (2011) 114–129. doi: 10.1002/nbm.1570

  8. [8]

     T. Wu, A. Li, K. Chen, et al., Chem. Commun. 57 (2021) 7743–7757. doi: 10.1039/D1CC02133H

  9. [9]

     L. Mignion, J. Magat, O. Schakman, et al., Magn. Reson. Med. 69 (2013) 248–254. doi: 10.1002/mrm.24245

  10. [10]

      E.T. Ahrens, R. Flores, H. Xu, P.A. Morel, Nat. Biotechnol. 23 (2005) 983–987. doi: 10.1038/nbt1121

  11. [11]

      D. Xie, M. Yu, R.T. Kadakia, E.L. Que, Acc. Chem. Res. 53 (2020) 2–10. doi: 10.1021/acs.accounts.9b00352

  12. [12]

      K.H. Chalmers, E. DeLuca, N.H.M. Hogg, et al., Chem. Eur. J. 16 (2010) 134–148. doi: 10.1002/chem.200902300

  13. [13]

      L. Zhu, Y. Li, M. Jiang, et al., ACS Appl. Mater. Interfaces 15 (2023) 2665–2678. doi: 10.1021/acsami.2c19161

  14. [14]

      A.T. Preslar, F. Tantakitti, K. Park, et al., ACS Nano 10 (2016) 7376–7384. doi: 10.1021/acsnano.6b00267

  15. [15]

      Y. Takaoka, T. Sakamoto, S. Tsukiji, et al., Nat. Chem. 1 (2009) 557–561. doi: 10.1038/nchem.365

  16. [16]

      P.L. Anelli, P.R. Ashton, R. Ballardini, et al., J. Am. Chem. Soc. 114 (1992) 193–218. doi: 10.1021/ja00027a027

  17. [17]

      C.A. Schalley, K. Beizai, F. Vögtle, Acc. Chem. Res. 34 (2001) 465–476. doi: 10.1021/ar000179i

  18. [18]

      B. Zheng, M. Zhang, S. Dong, X. Yan, M. Xue, Org. Lett. 15 (2013) 3538–3541. doi: 10.1021/ol401241w

  19. [19]

      B.L. Feringa, Angew. Chem. Int. Ed. 56 (2017) 11060–11078. doi: 10.1002/anie.201702979

  20. [20]

      R.E. Fadler, A.H. Flood, Front. Chem. 10 (2022) 856173. doi: 10.3389/fchem.2022.856173

  21. [21]

      G. Gholami, K. Zhu, G. Baggi, et al., Chem. Sci. 8 (2017) 7718–7723. doi: 10.1039/C7SC03736H

  22. [22]

      K. Yamauchi, A. Miyawaki, Y. Takashima, H. Yamaguchi, A. Harada, J. Org. Chem. 75 (2010) 1040–1046. doi: 10.1021/jo902393n

  23. [23]

      Y.F. Yin, M.Y. Yun, L. Wu, et al., Angew. Chem. Int. Ed. 58 (2019) 12705–12710. doi: 10.1002/anie.201906761

  24. [24]

      A. Garci, Y. Beldjoudi, M.S. Kodaimati, et al., J. Am. Chem. Soc. 142 (2020) 7956–7967. doi: 10.1021/jacs.0c02128

  25. [25]

      P. Rajamalli, F. Rizzi, W. Li, et al., Angew. Chem. Int. Ed. 60 (2021) 12066–12073. doi: 10.1002/anie.202101870

  26. [26]

      V. Blanco, A. Carlone, K.D. Hänni, D.A. Leigh, B. Lewandowski, Angew. Chem. Int. Ed. 51 (2012) 5166–5169. doi: 10.1002/anie.201201364

  27. [27]

      X. Lu, C. Huang, M. Li, et al., J. Phys. Chem. B 124 (2020) 5271–5283.

  28. [28]

      Y. Mo, C. Huang, C. Liu, et al., Macromol. Rapid Commun. 44 (2023) 2200744. doi: 10.1002/marc.202200744

  29. [29]

      S.E. Kirberger, S.D. Maltseva, J.C. Manulik, et al., Angew. Chem. Int. Ed. 56 (2017) 6440–6444. doi: 10.1002/anie.201700426

  30. [30]

      H. Shi, B. Lai, S. Chen, et al., Chin. J. Chem. 35 (2017) 1693–1700. doi: 10.1002/cjoc.201700240

• 

  Scheme 1  The strategy of improving ¹⁹F MRI agent sensitivity by threading the macrocycle 1 into [2]rotaxanes Rx-1 and Rx-2.

  下载: 全尺寸图片 幻灯片
  

  Figure 1  Partial ¹H NMR spectra of the axles, macrocycle, and [2]rotaxanes (a-f); 2D ROESY ¹H NMR spectrum of Rx-1 (g); single-crystal X-ray structure of 1 (h) and Rx-1 (i); partial ¹⁹F NMR spectra of 1 (j), A₁ (k), Rx-1 (l), Rx-2 (m), Rx-2′ (n). NMR conditions: 500 MHz, 1 mmol/L, 298 K, CD₃CN. The labeling of ¹H and ¹⁹F can be found in Scheme 1 and Scheme S2 (Supporting information).

  下载: 全尺寸图片 幻灯片
  

  Figure 2  R₁/R₂ of F_A (a), F_B and H_i (b) in the axles, macrocycle, and [2]rotaxanes; temperature-dependent ¹⁹F R₁ (c) and R₂ (d) of the [2]rotaxanes. Drmsd of Rx-1 (upper) and 1 (lower) over time (e), (f) ¹⁹F MRI phantom images (9.4 T, 298 K, CH₃CN) and the plot of logSI versus logC(¹⁹F) (g) of 1, Rx-2 and Rx-2′. Statistical significance: ^*P < 0.1 and ^***P < 0.001.

  下载: 全尺寸图片 幻灯片
•