Versatile non-luminescent color palette based on guest exchange dynamics in paramagnetic cavitands

Multicolor luminescent portrayal of complexed arrays is indispensable for many aspects of science and technology. Nevertheless, challenges such as inaccessible readouts from opaque objects, a limited visible-light spectrum and restricted spectral resolution call for alternative approaches for multicolor representation. Here, we present a strategy for spatial COlor Display by Exploiting Host-guest Dynamics (CODE-HD), comprising a paramagnetic cavitand library and various guests. First, a set of lanthanide-cradled α-cyclodextrins (Ln-CDs) is designed to induce pseudo-contact shifts in the 19F-NMR spectrum of Ln-CD-bound guest. Then, capitalizing on reversible host-guest binding dynamics and using magnetization-transfer 19F-MRI, pseudo-colored maps of complexed arrays are acquired and applied in molecular-steganography scenarios, showing CODE-HD’s ability to generate versatile outputs for information encoding. By exploiting the widely shifted resonances induced by Ln-CDs, the guest versatility and supramolecular systems' reversibility, CODE-HD provides a switchable, polychromatic palette, as an advanced strategy for light-free, multicolor-mapping.

The authors describe the induction of pseudo-contact shifts in the 19F spectrum of guests that are bound to lanthanide-cradled cyclodextrin hosts. The approach enabled the creation of highly differentiated signals that depend on the lanthanide and the guest species, based on paramagnetic guest exchange saturation transfer (paraGEST). The responses were assigned a pseudo-color and thereby color maps were created that do not depend on optical phenomena such as light absorption or emission. This was used for steganography, including code manipulation and tomographic analysis of objects. The described implication of paraGEST in 19F NMR and the developed applications are interesting. However, the enabling technology is known in the literature and the advantages of host-guest complexes for designing reversible stimuli-responsive systems have been exemplified in numerous examples. The herein reported expansion is an interesting case study and the linkage to pseudo-color maps is a smart move (albeit such protocols are rather usual in imaging). Every system that is capable of generating signals with a high degree of diversification has potential value for steganography. Without surprise the herein described library is not an exception. I think this is a nice work, well described and competently performed. However, the demonstrated conceptual advance appears too limited for publication in Nature Communications.
Reviewer #2 (Remarks to the Author): The chemical exchange saturation transfer method is applied to F19 nuclei in guests held within a CD cavity by various paramagnetic ions embedded in the base of the host. A limitation is that the guests need to be outfitted with a primary amine functionality in order to be well-recognized by this technique. Still, a range of guests and lanthanide ions become members of the library. This is a nice extension of a general MRI technique to samples in multi-well plates, even though false-colouring of segments of the range of any variable is common across science. The application to steganography fits well with currently available molecular systems and goes beyond them in terms of the multi-layer capabilities offered. Erasing the data by displacing the fluorinated guest and the tomographic capability are extra positive features which are missing in many current molecular systems. However, the overall weakness of the method is the need for an MRI instrument. But this is not a fault of the method.
CD-DTPA and its lanthanide complexes are known (J. Am. Chem. Soc. 1996, 118, 7414). This should be acknowledged in the citations.  Figure 4; A relevant reference for molecular generation of graphics displays is ACS Synth. Biol. 2020Biol. , 9, 1490 Reviewer #3 (Remarks to the Author): I quite like the idea of this manuscript, which is to use a supramolecular chemistry approach to access a color-based spatial display, and I recommend that the manuscript be accepted for publication after the following, relatively minor issues are addressed: 1. In the introduction, the authors use the phrase "light-based color." They should clarify the intended meaning of this phrase. 2. In general, the introduction has language that is out of place in a technical scientific paper. The authors refer to an "endless number of innovative supramolecular platforms," as well as the fact that "synthetic supramolecular systems are at the core of advanced fields in science." These and similar statements are hyperbolic and should be modified to more accurately reflect the scientific reality. 3. The supporting information does not include any NMR information for the various cyclodextrin functionalized materials. This omission is curious and should be addressed; while the mass spectra shown certainly match with the masses of the target compounds, the NMR spectra of the compounds before they participate in host-guest inclusion and the associated GEST phenomena should be provided.
Reviewer #4 (Remarks to the Author): Thig highly focused communication present's a supramolecular chemistry based technique named CODE-HD for the development of paramagnetic host-guest complexes that through the use of MRI can be used to develop advanced (artificial) colour patterns that can be use to develop codes and patterns that are both readable and erasable. This is a very nice idea and offers the generation of a large number of patters that can be employed in various research areas as well as in communication of encrypted information and data, etc. I really enjoyed this communication, the idea is clearly outlined and the results seem to support the authors design very nicely. The use of the lanthanides allows the authors to generate the library of the CD-hosts, that they then employed to form the Host-Guest complexes with 4 different guest, three of which are able to demonstrate the necessary shift, from which that colour patterns, using magnetization-transfer 19F-MRI, results were generated from; the authors using guest 2 to demonstrate this in a very nice way. I think this paper is quite nice, focused as I mentioned, but there are few points I have that the authors need to address before this is published. I would include more description on the synthesis and the charactersation of the host and the Ln-complexes in the main document. The Ln-CDs have been characterized by HRMS but no CHN is given, or other in-depth analysis. A statement that the HRMS for these LN-CD matched the calculated isotopic patterns for the relevant complexes might be good to include.
A comment on the stability of the Ln-complexes for the various Ln-CDs should be included. If binding constants are known these should be included too. LN-CDs are well known and some of that work dates back till the 1990s. Some ref. to that should be included, as these types of complexes are normally not employed in the clinic due to low stability.
Were Ce-CD and Sm-CD), that showed relatively small Δω values, tested in the experiences carried out using the MRI scanner (as was proposed they could be employed for)?

R2.1.
A limitation is that the guests need to be outfitted with a primary amine functionality in order to be wellrecognized by this technique. Still, a range of guests and lanthanide ions become members of the library.
We thank the Reviewer for highlighting this point. To elaborate on this issue, we examined an additional guest, i.e., 4-fluorobenzoic acid, in order to emphasize the potential of CODE-HD to be furthered extended with additional guests. This additional demonstration, and the fact that both the primary-amine and the carboxylate functionalities of the fluorinated guests were found to be applicable for CODE-HD, highlight the necessity of functional groups with Ln 3+ -coordination capabilities for obtaining the reported paraGEST phenomena. This interpretation implies that additional guests with Ln 3+ -coordination capabilities, such as phosphate, sulfonate, thiolates and even pyridine, should be considered in the future in order to extend the putative library shown in our manuscript.
The following discussion was added to the text: "Importantly, an analog of guest 1, which is substituted with a carboxylic acid instead of a primary amine group, also yielded a pronounced paraGEST effect ( Figure S41a), while other analogs, substituted with hydroxyl or amino-methyl functional groups, did not generate any noted effect ( Figure S41b-d). These observations reflect the cruciality of a functional group with Ln 3+ -coordination capabilities (primary amine or a carboxylic acid) for a successful magnetization transfer effect (Figure 2 and S41)."

R2.2.
The overall weakness of the method is the need for an MRI instrument. But this is not a fault of the method.
We agree with the Reviewer that the need for an MRI is an intrinsic limitation of the method. However, as emphasized in the conclusions, CODE-HD can be also applied with a conventional NMR setup, which is available at any research institute. Thus, it can be widely implemented by many researchers, but with the limitation of 1D encoding feature (NMR) and without the 2D display capabilities, which, indeed, require an MRI instrument, as mentioned by the Reviewer. "In summary, we show here the design, development, principles and implementation of a non-luminescent supramolecular system, CODE-HD, which enables 1D (NMR), 2D, and 3D displays of artificial colors." R2.3. CD-DTPA and its lanthanide complexes are known (J. Am. Chem. Soc. 1996, 118, 7414). This should be acknowledged in the citations.
We thank the Reviewer for highlighting the previously reported CD-DTPA lanthanide complexes, which should have been acknowledged in our first submission. We have now added two new references (i.e., references 44 and 45) to highlight the originally proposed Ln-CDs. Figure 2 line 3; from an aqueous.

R2.4.
We thank the Reviewer for this comment. This has been corrected. Figure 4; A relevant reference for molecular generation of graphics displays is ACS Synth. Biol. 2020, 9, 1490.

R2.5.
We thank the Reviewer for highlighting this very important example of a molecular generation of graphic displays. The relevant reference was added as the new reference 55.

Reviewer #3
R3.1. In the introduction, the authors use the phrase "light-based color." They should clarify the intended meaning of this phrase.
We thank the Reviewer for highlighting this point. This phrase was replaced with "luminescent color" throughout the text.

R3.2.
. In general, the introduction has language that is out of place in a technical scientific paper. The authors refer to an "endless number of innovative supramolecular platforms," as well as the fact that "synthetic supramolecular systems are at the core of advanced fields in science." These and similar statements are hyperbolic and should be modified to more accurately reflect the scientific reality.
We thank the Reviewer for pointing this out. These phrases were modified to more accurately reflect the scientific reality, as requested.
(a) "…establishing synthetic supramolecular systems as accessible tools for both fundamental research and advanced applications" (b) "…created the opportunity to design a large number of innovative supramolecular platforms" R3.3. The supporting information does not include any NMR information for the various cyclodextrin functionalized materials. This omission is curious and should be addressed; while the mass spectra shown certainly match with the masses of the target compounds, the NMR spectra of the compounds before they participate in host-guest inclusion and the associated GEST phenomena should be provided.
We thank the Reviewer for highlighting this very important point, which was not clearly shown in the first version of the manuscript. As requested by the Reviewer, we performed extensive 1D ( 1 H-NMR, 13 C-NMR) and 2D ( 1 H-1 H COSY, 1 H-13 C HSQC and 1 H-1 H ROESY) NMR experiments to generate a new set of NMR spectra for the studied cavitand (CD-DTPA) that participate in the host-guest inclusion and the associated GEST (new Figures  S2-S19).
The following sentence was added to the text: "The purity of the obtained CD-DTPA product was evaluated by an analytical HPLC ( Figure S1), followed by its full characterization using a set of 1D ( 1 H-, 13 C-) and 2D ( 1 H-1 H COSY, 1 H-13 C HSQC and 1 H-1 H ROESY) NMR experiments (Figures S4-S19 and Appendix A)…" A relevant explanation was added to the SI file as Appendix A: "While native CDs exhibit high symmetry, which allows a relatively easy characterization by NMR, partial modifications on their glucose units can leads to the splitting of the spin systems due to symmetry breaks. This results in a crowded spectrum with an overly-splitted, overlapped, broad aliphatic area, as evident only from Diamino-CD's NMR spectra (Figures S2 and S3). The additional modification with DTPA had caused splitting to three spin-systems in CD-DTPA ( Figures S4-S8), which can be mainly evident by the location of three representing peaks for the anomeric H-1 protons (i.e., H-1, H-1' and H-1'', Figure S4, 4.99-5.05 ppm), indicating three types of CD's sugar units. To complete the characterization, advanced 2D-NMR experiments, such as COSY, HSQC and ROESY ( Figures S9-S18), were performed. 2D COSY NMR ( Figures S9-S12) had allowed us to first identify the CD-DTPA's inner protons (H-1 to H-6) and assign their "ranges". Following that, 2D HSQC NMR (Figures S13-S16) had enabled to distinguish between most of CD-DTPA's inner protons/carbons (H/C-1 to H/C-5) to the DTPA-bridge's protons/carbons (H/C-7 to H/C-11) and H-6, based only on the correlations' phases (i.e., green and blue, representing primary and secondary carbons respectively). Moreover, this experiment indicates the existence of inequivalent protons on several carbons (C-6, C-8 and C-9, indicated by "a" and "b"). The two H-6 protons undergo dramatic changes in their chemical shifts due to the DTPA modification. 2D ROESY NMR ( Figure S17 and S18) had enabled the identification of an H-4 peak underlying H-2's multiplet, due to a "through-space" interaction between H-4 and H-1. Considering the intricate NMR spectra and broadside chemical shifts (summarized at Figure S19), characterization was mainly based on MS analysis ( Figures S21-S42)."

Reviewer #4
R4.1 I would include more description on the synthesis and the charactersation of the host and the Ln-complexes in the main document. The Ln-CDs have been characterized by HRMS but no CHN is given, or other in-depth analysis. A statement that the HRMS for these LN-CD matched the calculated isotopic patterns for the relevant complexes might be good to include.
We thank the Reviewer for this suggestion. As recommended by the Reviewer, we have added the description of the synthesis to the main document: "Following 16 hours of stirring in anhydrous dimethyl sulfoxide and triethylamine, the resultant white powder was collected after centrifugation and purified using preparative reversed-phase high-pressure liquid chromatography (HPLC). The purity of the obtained CD-DTPA product was evaluated by an analytical HPLC ( Figure S1), followed by its full characterization using a set of 1D ( 1 H-NMR, 13 C-NMR) and 2D ( 1 H-1 H COSY, 1 H-13 C HSQC and 1 H-1 H ROESY) NMR experiments (Figures S4-S19 and Appendix A) and high-resolution mass spectrometry ( Figures S20 and S21). Then, based on the procedure applied to obtain lanthanide-cradled b-cyclodextrins, 44, 45 pure CD-DTPA was refluxed in water…" In addition, as requested by the Reviewer, we have added an in-depth analysis of the products, including: (i) Elemental CHN analysis (Table S1 For Reviewer).
The content of C, H and N atoms in the sample (36.67%, 4.11% and 4.11% respectively) was found to be smaller than theoretically calculated for C50H81N5O36 (45.21%, 6.15% and 5.27% respectively). A 19 F-NMR spectrum was acquired and revealed a 1:1 ratio between CD-DTPA and TFA (from the HPLC eluent based on the integration of the representing peak for TFA, -76 ppm, and a previously calibrated Sodium fluoride capillary). Fitting the data into a JavaScript Percentage Elemental Results Calculator (JASPER), while limiting the number of TFA molecules to one, resulted a chemical formulation of C50H81N5O36•1TFA•9.2H2O (38.84%, 6.29% and 4.36% respectively).
To the best of our understanding, elemental analysis of macrocycles might not be accurate, due to the presence of bound water molecules, and in particular for CDs, which includes water inside their cavity, and are sensitive to moisture. Thus, we examined the C/N ratio in the measured sample (8.92) and compared it to the calculated formula C50H81N5O36•1TFA•9.2H2O (8.91) to characterize the compound. For this reason, the new results were not reported in the text and are added for the Reviewer revision. Nevertheless, it should be emphasized here that we have performed an extensive NMR studies (see point ii next) to characterize our synthetic host in full.

(ii)
Extensive NMR experiments, including 1D ( 1 H-NMR, 13 C-NMR) and 2D ( 1 H-1 H COSY, 1 H-13 C HSQC and 1 H-1 H ROESY) NMR experiments to further characterize the newly proposed compounds (New Figures S4-S19 and Appendix A). The following sentence was added to the text: "The purity of the obtained CD-DTPA product was evaluated by an analytical HPLC ( Figure S1), followed by its full characterization using a set of 1D ( 1 H-, 13 C-) and 2D ( 1 H-1 H COSY, 1 H-13 C HSQC and 1 H-1 H ROESY) NMR experiments (Figures S4-S19 and Appendix A)…" A relevant explanation was added to the SI file as Appendix A: "While native CDs exhibit high symmetry, which allows a relatively easy characterization by NMR, partial modifications on their glucose units can leads to the splitting of the spin systems due to symmetry breaks. This results in a crowded spectrum with an overly-splitted, overlapped, broad aliphatic area, as evident only from Diamino-CD's NMR spectra (Figures S2 and S3). The additional modification with DTPA had caused splitting to three spin-systems in CD-DTPA ( Figures S4-S8), which can be mainly evident by the location of three representing peaks for the anomeric H-1 protons (i.e., H-1, H-1' and H-1'', Figure S4, 4.99-5.05 ppm), indicating three types of CD's sugar units. To complete the characterization, advanced 2D-NMR experiments, such as COSY, HSQC and ROESY ( Figures S9-S18), were performed. 2D COSY NMR (Figures S9-S12) had allowed us to first identify the CD-DTPA's inner protons (H-1 to H-6) and assign their "ranges". Following that, 2D HSQC NMR (Figures S13-S16) had enabled to distinguish between most of CD-DTPA's inner protons/carbons (H/C-1 to H/C-5) to the DTPA-bridge's protons/carbons (H/C-7 to H/C-11) and H-6, based only on the correlations' phases (i.e., green and blue, representing primary and secondary carbons respectively). Moreover, this experiment indicates the existence of inequivalent protons on several carbons (C-6, C-8 and C-9, indicated by "a" and "b"). The two H-6 protons undergo dramatic changes in their chemical shifts due to the DTPA modification. 2D ROESY NMR ( Figure S17 and S18) had enabled the identification of an H-4 peak underlying H-2's multiplet, due to a "through-space" interaction between H-4 and H-1.
Considering the intricate NMR spectra and broadside chemical shifts (summarized at Figure S19), characterization was mainly based on MS analysis ( Figures S21-S42)."

(iii)
As requested by the Reviewer, the HR-MS data is now accompanied by calculated isotopic patterns, which show good correlation with the experimental spectra obtained for the Ln-CDs (New Figures  S21, S22-S40).

R4.2.
A comment on the stability of the Ln-complexes for the various Ln-CDs should be included. If binding constants are known these should be included too.
As recommended by the Reviewer, we have added a comment on the stability of the obtained Ln-complexes: "Note here that the obtained octadentate aminopolycarboxylate CD-DTPA should have a strong binding affinity towards lanthanides in the resultant Ln-CDs, as those obtained for the clinically-used Gd-DTPA-BMA contrast agent 46 . Nevertheless, as the binding affinities may slightly deviate for different Ln 3+ , as reported for different Ln-DTPA-BMAs 47, 48 , further stability studies and safety profiles should be obtained prior to the use of Ln-CDs in any biological application in the future." New relevant references (46-48) were added as well.

R4.3.
Were Ce-CD and Sm-CD), that showed relatively small Δω values, tested in the experiences carried out using the MRI scanner (as was proposed they could be employed for)?
We thank the Reviewer for this comment. To elaborate on the potential use of both Ce-CD and Sm-CD in CODE-HD and their function as additional artificial colors in future applications, both hosts were examined using a 15.2 T MRI scanner, the results of which now appear as new Figure S45. An additional comment was added to the text as well: "Note here that future applications including Ce-CD and Sm-CD, which induced relatively small Δω values in 2 ( Figure S43), can add additional colors to CODE-HD, enabling even more color combinations (although limited for strong-field MRI scanners, as the one used here, Figure S45). By this, CODE-HD is able to offer, in principle, eleven factorial (11! = 39,916,800) different combinations in dedicated setups."

R4.4. The ref. in SI are not given in full.
This was corrected.

New Figures
New Figure S1. CD-DTPA absorbance at 220 nm. Measured at analytical HPLC.