Stiffness-switchable DNA-based constitutional dynamic network hydrogels for self-healing and matrix-guided controlled chemical processes

Constitutional dynamic networks (CDNs) attract interest as signal-triggered reconfigurable systems mimicking natural networks. The application of CDNs to control material properties is, however, a major challenge. Here we report on the design of a CDN consisting of four toehold-modified constituents, two of which act as bidentate units for chain-elongating, while the other two form a tetradentate structure acting as a crosslinking unit. Their hybridization yields a hydrogel of medium stiffness controlled by the balance between bidentate and tetradentate units. Stabilization of the tetradentate constituent by an auxiliary effector up-regulates the crosslinking unit, yielding a high-stiffness hydrogel. Conversely, stabilization of one of the bidentate constituents by an orthogonal effector enriches the chain-elongation units leading to a low-stiffness hydrogel. Using appropriate counter effectors, the hydrogels are reversibly switched across low-, medium- and high-stiffness states. The hydrogels are used to develop self-healing and controlled drug-release matrices and functional materials for operating biocatalytic cascades.

the orthogonal effectors which stabilized the bidentate molecules. By fine tuning the concentration of the effectors which supports the tetradentate and bidenate molecules, hydrogels of various stiffness could be achieved with the property of reversibility. They also showed the self-healing property and release of small molecules such as the drug doxorubicin/QDs from the CDN X hydrogels through the trigger. Major Comments For the transition of the hydrogel crosslink network, would there be a limit in using the trigger E2? As E2 trigger up-regulates the AA' and BB' concentration of the cross-link network, an "overexposure" to E2 trigger may transform too much of the cross-link into AA' and BB', and render the hydrogel liquid and unable to remain the gel form. Thus, it would be interesting to learn how much E2 trigger is needed to dissociate the gels. In Figure 3. The authors show the reversible property over the stiffness measures of the CDN hydrogels through visual characteristics. They showed that the treatment of the CDN hydrogel Y with the counter trigger E'1 restores the original shape of the CDN hydrogel X. Whether the CDN hydrogel Z restores its original shape (CDN hydrogel X) when using the trigger E'2 from CDN hydrogel Y? Gaining the shape of the hydrogel from this point through trigger E'2 would be more interesting and add significant meaning to the shape restoration properties of the quasi-liquid CDN hydrogel. This also gives more clue about the shape memory of the CDN hydrogels. The transition of shape change contributing for the stiffness properties of these CDN hydrogels is undoubtedly interesting (Figure 3). However, the authors show the reversible property of the CDN hydrogels only from X to Y and then to X. Can the authors do an experiment where they could change the shape of the CDN hydrogel X to CDN hydrogel Y and then to CDN hydrogel Z (i.e., X -->Y -->Z)? For the reversibility of the stiffness switch, is there a limit to the switch? Say, will the gels dissociates or maintain stiffness after certain cycles? The stiffness may start to have some weird result if the triggers remain clung to the gels, and is more possible to occur as the cycle increases. Moreover, from studying the data of Supplementary Figure 1 d (and also d), the stiffness after one cycle shows a slight decrease. Thus, to better understand the reversibility of the process, data data of showing the switching stiffness more than one cycle is needed to understand if there would be a continuous decrease in stiffness for the gel. For self-healing experiments, there is an interesting questions: are hydrogels of different CDN able to heal with each others? Hydrogels with different CDN still share the same crosslink sequences (though different structure), and treating them with same trigger should ultimately push the crosslinks to certain structures. As long as there is some structures changing around the interface between two gels, they should be able to form bounds on the interface and heal with each other. However, the bonds formed may or may not be sufficient to hold the two gels. Still, this would be a great experiments in testing the limit of the self-healing property of the hydrogel. In figure 4a and b, the authors show the self-healing property of the CDN hydrogels through the DNA triggers. In the control they showed in the absence of the trigger the hydrogels could not able to join by themselves. The readers would better understand if they provide a video file for this particular experiment showing the self-healing property of the CDN hydrogel through the addition of the trigger. The authors show the self-healing mechanism of the hydrogels stimulates a biocatalytic mechanism thereby increasing the fluorescence of the produced dye Resorufin. In Figure 4a Yui et al, showed that the self-healing nature of the CDN hydrogel by having two hydrogels and then healing the two to one through the addition of trigger E1. They prepared 2 CDN X hydrogels one containing glucose oxidase and other containing horseradish peroxidase. The self-healing property of the hydrogels is triggered through E1 resulting in the newly assembled CDN Y hydrogel which is essential for the bio-catalytic cascade reaction. The control experiments were also appropriate determining the cascade reactions takes place only with the healed hydrogel.
Why did the authors use two separate CDN hydrogels for loading of the catalytics contents, namely glucose oxidase and horseradish peroxidase? What would be the result if both the glucose oxidase and horseradish peroxidase are inoculated into one CDN Hydrogel X? For controlled release experiments of releasing anticancer drug doxorubicin, inefficient release of the drug was observed when using hydrogels with CDN X (figure 5c). Hypothesis is that the inefficient release was caused by the large pore size of the gel. Thus, an simple experiment may be conducted to test the hypothesis, or, at least, the limit of the gel. As hydrogels with CDN Y have a higher degree of crosslinking, they should not only be stiffer than CDN X gels ut also have smaller pore size. The authors may conduct experiments using CDN Y gels with doxorubicin entrapped to see if the pore size of the gels are smaller enough to contain the drug. Experiments with doxorubicin and QD's needs more attention. I feel this is to early to give a positive notion about the release of the drug molecules from the hydrogels In Page 8 (Probing the triggered….): The authors say that the hydrogels were treated with the trigger E1 for the release of doxorubicin/QDs, but in Figures 5b and c they display trigger E2 for the release if doxorubicin/QDs. This is confusing for the reader. What type of release profile is exhibited for doxorubicin from the hydrogel? Is it controlled release? If so, can you explain this with a release-kinetics model such as first-order, zero-order, etc? How does doxorubicin binds to the hydrogel, could the authors explain about the mechanism of the same? How do you confirm that the doxorubicin is in the DNA hydrogel? Is it possible for the authors to do SEM analysis showing the binding of the drug into the gel matrix? The authors claim that the trigger E2 helps in the release profile of the drug doxorubicin within 1 hour. The large porous size of the hydrogel Z and also the shape change to quasi-liquid helps in the release of the drug. An effective control would be after loading the drug into CDN X hydrogels, instead of the trigger activation, could the authors cut the gel into two and check the release profiles? Did the authors did the same experiment with the medium density pores, of CDN Y hydrogel? The authors claim that the CDN X hydrogel stimulated no release because of the small pore size. However, there should be slow/steady increase in the release profile of the drug when the drug is loaded in CDN Y hydrogel.

Detailed specific response to the reviewers' comments
The following point-by-point listed corrections were introduced into the paper:

Reviewer #1:
We feel that the reviewer undervalued the significance and novelty of the present study. The reviewer is correct that many stimuli-responsive nucleic acid-based hydrogels of controlled stiffness were reported. Nonetheless, the present study introduces new concepts in designing stimuli-responsive DNA-based hydrogels. These include: (i) The demonstration that a constitutional dynamic network of nucleic acids can undergo triggered reconfiguration across three hydrogel stiffness states is unprecedented, and this concept may be extended to other materials and systems. (ii) The study has an important basic scientific significance as it provides an optical (fluorescence) means to correlate between the molecular compositions of the constituents in the different hydrogel states and the bulk stiffness properties of the respective hydrogels. (iii) The broad applications of DNA-based constitutional dynamic network hydrogels as self-healing materials, controlled release matrices, and their control over biocatalytic cascades are novel.
These unique contributions of the present study were further emphasized at the end of the "introduction" section.
The authors have satisfactorily answered the questions in the first version. I recommend this work for publication after the authors address the following questions and comments in the manuscript: Comments and questions for the authors For transition of the hydrogel from X → Y → Z, methods of transition from Y → Z should be different from X → Y, as the latter takes a single inducer E1, while the former requires two, E1' and E2. Thus, method clarifying such treatment should be added whether in the method part of the main text or under the description of figure 18, specifying the concentration of each trigger strands and whether both strands are added to the system simultaneously or sequentially.
The self-healing ability of different hydrogels is very interesting! However, similarly, for the selfhealing process, specific method should also be added, clarifying the concentration of the inducer strands and also whether the strands are added simultaneously or sequentially.
In the main text, p13, the authors mentioned that "The release from hydrogel Y in the absence of triggers is substantially lower than that from hydrogel X, consistent with the higher stiffness and smaller pore-size of hydrogel Y." Comparing the result of figure 5b and supplementary figure 26, for readers' better understanding, the authors should point out the quantity difference between two experiment to avoid confusion.
Also, comparing the result of figure 5b and supplementary figure 36, the readers can see that hydrogel Y has a lower efficiency for release as the final intensity of the released drugs is lower than that of hydrogel X. However, the two hydrogels, X and Y, should all transform into the same type of hydrogel (Z) in the end, and has the same final intensity. Can the authors provide an explanation for this?
Are there difference between the two hydrogel Zs as they come from different hosts? If so, where does the difference comes from? For hydrogel Z from Y does it requires more E2 or both inducers to have properties similar to hydrogel Z from X ?