A natural biogenic nanozyme for scavenging superoxide radicals

Biominerals, the inorganic minerals of organisms, are known mainly for their physical property-related functions in modern living organisms. Our recent discovery of the enzyme-like activities of nanomaterials, coined as nanozyme, inspires the hypothesis that nano-biominerals might function as enzyme-like catalyzers in cells. Here we report that the iron cores of biogenic ferritins act as natural nanozymes to scavenge superoxide radicals. Through analyzing eighteen representative ferritins from three living kingdoms, we find that the iron core of prokaryote ferritin possesses higher superoxide-diminishing activity than that of eukaryotes. Further investigation reveals that the differences in catalytic capability result from the iron/phosphate ratio changes in the iron core, which is mainly determined by the structures of ferritins. The phosphate in the iron core switches the iron core from single crystalline to amorphous iron phosphate-like structure, resulting in decreased affinity to the hydrogen proton of the ferrihydrite-like core that facilitates its reaction with superoxide in a manner different from that of ferric ions. Furthermore, overexpression of ferritins with high superoxide-diminishing activities in E. coli increases the resistance to superoxide, whereas bacterioferritin knockout or human ferritin knock-in diminishes free radical tolerance, highlighting the physiological antioxidant role of this type of nanozymes.

, the ferritins with "0 mM Fe content" corresponds to the ferritins purified from the basal medium, but their "inhibition rate of SOD" are not consistent with the results in Fig S3a ; The OD562 of "Pi/Fe=1" in Fig S15d was lower than that of "Pi/Fe=0.2and 0.5".Please explain these discrepancies.3.What are the pH conditions when measuring SOD-like and CAT activity?Please add in the experimental methods.4.Please define the meaning of the abbreviations used in the text/figures where appropriate, such as apo-pfFn, -Fe, Fe-in vitro, Fe-in vivo, HFn-WT, HFn-3-Fold-M, HFn-3-Fold-M-Fe, etc. 5.Some figures in the text were cited in reverse order, e.g.Supplementary Figure 1b and 1c, d; Fig. 4d and 4e; Supplementary Figure 12f and 12g, h.6."However, the formation of H2O2 after adsorption of the second HOO• on ferrihydrite isendothermic with an increase in energy by 0.73 eV (Fig. 5h)": Please check if 0.73 eV is correct.7.There are some typing errors in this article, such as "pFfn" in Fig 2f , "3-pore" should be "3-Fold" in Supplementary Figure 7c and 7d.8.The article referred to Supplementary Figure 16f, which is not available in the SI.9.Supplementary Figure 2c: The unit of oxygen generation rate should be mg•L-1•min-1.10.Supplementary Figure 4: What are the red, blue, and gray represent respectively?11.Supplementary Figure 5a: "CkFn" is not mentioned in this article; "Thermococcus barophilus (TaFn)"should be"Thermococcus barophilus (TbFn)"; "TbFn" is misspelled as "TaFn" in Supplementary Figure 5b.12.Please revise the format of the references carefully, noting issues such as letter case, subscripts, and page numbers.
Reviewer #2: Remarks to the Author: Fan and coworkers have described mechanism behind antioxidant role of ferritins with high SOD like activity.They established a relationship between the strength of catalytic activity and the source of species.The results are incremental to the previous studies in the field, and most of the results in the manuscript are in agreement with what authors have claimed.I have listed some of my concerns below-: 1.The authors should express the bacterial viability in terms of colony forming units/mL instead of OD450.2. The TEM images need to be cleaner showing a higher magnification image in inset.3. Authors need to demonstrate the potential applications of their findings in order to make this study suitable for broad readership of Nature Communications.One application may be enhanced tumor therapy.Overall, the manuscript is well written and should be published in Nature Communications after major revision.
Editorial Note: Parts of this Peer Review File have been redacted as indicated to maintain the confidentiality of unpublished data.

Reviewer #3:
Remarks to the Author: This work on the SOD activity of ferritin is remarkable, both for the themes addressed, the angles taken and the subjects explored.The text is very well written and very convincing.I think it should be accepted in Nature Comm., but some of the experiments need to be better described.I give a list of what could be improved.I think this is key to the potential impact of the article.That is why I have indicated that these are "major revisions".
(a) From the outset, I wondered whether the observed activity might be associated with metal ion leakage from the inorganic core and ferritin, especially as it has been evidenced that this occurs under oxidative stress.Some answers are given at the end, but I think they should be mentioned from the start (around line 127, perhaps just to indicate that further experiments will be provided at the end to ensure that the activity is not associated with the metal ion released from ferritin).
(b) My most important concern is the description of SOD-type activity.First of all, the assay they use is based on Xanthine-Xanthine Oxidase (X/XO) with a UV-vis marker.This should be made clearer than simply referring to an assay kit with the name of a supplier.I had to download the supplier's data sheet to better understand what had been done.It is technically important to indicate (i) the nature of the UV-vis marker (WST-1?), (ii) the concentration used for the marker (iii) the pH.The SOD like activity is reported as a « % of inhibition of SOD » in the figures and with the exp.part line 578-593, but this label is incorrect: the assay does not measure an inhibition of SOD at all.Here how it goes: the enzyme xanthine oxidase (XOox) oxidizes the substrate xanthine leading to XOred and, when a proper pH is used, XOred re-oxidizes using dioxygen and leading to a flow of ca.75 % of superoxide (and some H2O2).This superoxide reacts with the UV-vis marker (here WST-1) and changes its color.When a SOD like agent is added to the medium, it competes with WST-1 and limits its reaction by superoxide (in a way, there is less superoxide available for WST-1 as it is partially used by the SOD-like agent).So, what is measured here is an "inhibition of the marker's reaction with superoxide".SOD-like activity is indicated as "% SOD inhibition" in the figures and on line 578-593 of the exp.section, but this label is misleading, as the assay does not measure SOD inhibition at all.Here's how it works: the oxidized form of the enzyme xanthine oxidase (XOox) oxidizes the substrate xanthine leading to XOred and, when an appropriate pH is used, XOred re-oxidizes using dioxygen and leading to a flux of around 75% superoxide (and some H2O2).This superoxide reacts with the UV-vis marker (here WST-1) and changes color.When a SOD-like agent is added to the medium, it competes with WST-1 and limits its reaction with superoxide (in a way, there is less superoxide available for WST-1 as it is partially utilized by the SOD-like agent).What is being measured here, therefore, is "inhibition of the marker's reaction with superoxide".As this is a competition between WST-1 and the SOD-agent, the measured "inhibition" will be dependent of the concentration of WST-1 and of the SOD-agent.They state "the sample with equal quality or equal iron content was mixed with 200 μL of WST-1 working solution".Can they provide this "equal quantity and equal iron content" and WST-1 concentration more explicitly?What is the actual metal content?I assume it was measured by ICP-MS: can they specify that?Also, from other figures/exp (see fig6), it looks like they can provide a ferritin amount in µg.Could this be used here?I mean indicate the amount of iron AND the µg/mL of ferritin used in the assay?Another concern is there is here no way to evaluate whether the activity is a good one or not.It should be possible by the kinetic method and a kcat could be recalculated provided the kcat of the marker is known.That would be a lot a work to reperform the analyses using the kinetic method for this paper, but I would advise to try and set-up the kinetic experiment in the future.With this set of experiments, what I would suggest is to measure the activity of a commercially available protein SOD.An idea could be to estimate a number of SOD units that would provide an inhibition of the WST-1 oxidation in a range similar to what they have obtained for the ferritin.There would be two ways to report the results: (a) give a result for a certain amount of Fe and just say it is eq. to ca. xx SOD unit (b) report the results per quantity of metal ion content, but I guess this would be less convincing as SOD contain one metal ion and ferritin a large core in which most probably only a few are catalytically active (those onto the surface and probably not those in the bulk of the particle?).SOD activities are not given in such a way that the experiment can be redone if we wish, or compared with another experiment.How the so-called "SOD inhibition %" is calculated is unclear and does not exactly match the supplier's description (see the equation, SOD activity (inhibition rate %) = [(Ablank 1 -Ablank 3) -(Asample -Ablank 2)]/ (Ablank 1 -Ablank 3) x 100, which is different from that given in the exp.part of the article).Could they describe more precisely what they have done?Note that this measurement should be sensitive to incubation time and that the most robust way to perform this type of measurement is the kinetic method also described by the supplier.Another problem is that there is no way of assessing whether the activity is good or not.This should be possible using the kinetic method, and a kcat could be recalculated provided that the kcat of the marker is known.Although it surely would be a lot of work to redo the analyses using the kinetic method for the present article, I would advise trying to set up the kinetic experiment in the future.With this series of experiments, I would suggest measuring the activity of a commercially available SOD protein.One idea might be to find a number of SOD units that would provide inhibition of WST-1 oxidation in a range similar to that obtained for ferritin.There would be two ways of presenting the results: (a) give a result for a certain amount of Fe and simply say that it is equivalent to about xx SOD units (b) present the results by amount of metal ions, but I suppose this would be less convincing as SOD contains one metal ion and ferritin a large core in which, most likely, only a few are catalytically active (those on the surface and probably not those in the bulk of the particle?).
(c) Another concern is that this SOD-assay can give false positive results.This would be the case for instance if the SOD-like agent inhibited the enzyme XO.In this case, the reduction in superoxide flow will not be due by the reaction of the agent with superoxide BUT to the reduction in superoxide production by XO.This can be checked by measuring, upon addition of the putative SOD agent, the activity of XO through the oxidation of xanthine into urate.This can be checked by UV-visible (see ( 1 -I like this idea that the protein core shapes the iron core.But can they be more specific and may be give a reference?-how do they prepare the "biomineralized ferritin"?I could not find any information about that, in the text, exp part or SI. (e) Paragraphe "Phosphate coordinated with the iron atom leading to high SOD-like activit" -They write: "It has been previously shown that phosphate exists on either the surface of the iron core in horse spleen ferritin or in the interior of the bacteria ferritin, which suggests that the phosphate inner the iron core might be the determining factor of the SOD-like activity."I do not see how this information from the literature suggest that.This is what follows that show it.Can they rephrase?-They indicate (line 311) that the iron oxidation state is (+III) ; I agree, but this is not necessarily its actual charge (due to delocalization onto the ligand).May be rephrase.
-Pre-edge: please provide a reference for the pre-edge indicating an Oh structure for Fe(III).Similarly, identification of FePO4 by EXAFS, please, give a preference for the comparison with a EXAFS spectrum for FePO4.
(f) Figure 5: I guess route (i) has more phosphate than (h).Can that they be indicated more clearly in the figure/text?(g) Figure 6g/h: (g) the ESR spectrum is not that of superoxide but that of superoxide trapped by DMPO.Please, change the caption.The amount indicated on the figure as xx µg/mL: are they in µg of enzyme or of iron?It seems it is Fe (from h) but specify that more clearly on the figure.(h) Can they specify which line(s) was used to estimate the intensity of the ESR? (h) The results with DCF are not very clear to me in the processing.Can they be more explicit about that?Do they look at the area below the curve, at the maximum absorbance?Sorry, this is just that I am not used to this kind of reporting, but I guess other readers will be like me.
Overall, this an excellent paper but that needs to be made more precise on the overall chemical and physico-chemical description.I hope my comments will help improving the impact of this nice study within the community of chemists.
Bibliography: the references are not homogeneously provided (titles in capital letters or not for instance).
Reviewer #1 (Remarks to the Author): In this work, the authors report that biogenic ferritins act as natural SOD enzymes and find that the biogenic iron cores of prokaryotic ferritins have a higher ability to scavenge superoxide radicals than that of eukaryotes.
In addition, it is revealed that the protein structure of ferritins determines the P/Fe ratio in the core and that different P/Fe content leads to different iron nucleus lattice structures and SOD-like activities.This work uses extensive methodology and provides lots of experimental data that can support their conclusions.Overall, it is a very interesting work.However, in my opinion, there are some issues that should be clarified and modified.

Response:
We really appreciate the reviewer's careful review and positive comments on our manuscript, we have addressed the issues according to the reviewer's suggestions.
1.The authors demonstrated that the SOD-like catalysis of ferritins is different from the Haber-Weiss reaction involving Fe ions, is it possible to show the specific process of the SOD-like reaction of ferritins by using the reaction equations or schematic diagram?
Response: As suggested, we have drawn schematic diagrams of the ferritin with different cores by DFT (Fig 5g & h), in which the ferrihydrite and P-doped ferrihydrite were used to mimic the ferritin core with lower and higher P/Fe ratios, respectively.As shown in Figure 5g-i, the detailed superoxide diminishing processes include the absorption and disproportionation of HOO • on the ferrihydrite core surface.In the ferritin with a lower P/Fe ratio, the first HOO • is absorbed on the ferrihydrite surface where the O-H splits without energy barriers.This process causes the O2 production and retains an H atom on the ferrihydrite surface.Subsequently, the second HOO • comes into the active site and experiences the same O-H split process.Due to its high affinity to the H proton, the ferrihydrite would form the stable H-binding steady state and could not catalyze the dismutase reaction efficiently (Fig 5h).Contrarily, on the ferritin core with a higher P/Fe ratio, the doped phosphate could apparently lower the affinity of ferrihydrite to H, which would lower the energy barrier of  Also, does ferritin produce hydroxyl radicals in exerting SOD-like activity?Please provide ESR data.
Response: Thanks very much for the expert question.Both in the Electron Spin Resonance experiments of ferritins and ferric ions, we used the DMPO as the trapper for the free radicals, which could efficiently trap O-, C-, N-, and S-centered radicals to form distinguishable ESR spectra [1][2][3][4] .As shown in Fig 6g and Supplementary 15h-j, with the substrate concentration increasing, the superoxide produced by the xanthinexanthine oxidase system was apparently transformed into hydroxyl radicals by ferric ions, however, there is no hydroxyl radical production in the ferritin groups with the same iron content.These data indicate that the ferritins diminished the superoxide without the hydroxyl radicals forming.2. There are inconsistencies in the results of SOD-like activity measurements of some ferritins in the paper.
For example, the "inhibition rate of SOD" of "pfFn" in  Revised Supplementary Figure 3. (f-h) SOD-like activity of biomineralized ferritins with gradient iron content.

Response:
We have noticed the discrepancies and confirmed with additional experiments (Figure only for reviewer 1).Of note, with similar SOD-like activities and superoxide diminishing efficiency in the same iron content, pyFn also exhibited apparently lower OD562 than the pfFn group.We believe that the OD562 only reflected the ferrous ions generation during the reaction of ferritin or ferrihydrite with superoxide, not the SOD activities.This discrepancy in the ferrous ion production and SOD activities also reflects that the ferritin and ferrihydrite decompose the superoxide in a manner different from that of the ferric ion.

Response:
We apologize for missing the information.The pH conditions during the SOD-like and CAT-like activities were both pH 7.4, which has been added to the experiment methods of the revised manuscript (page 27, paragraph 3 & Page 28, paragraph 2). 4. Please define the meaning of the abbreviations used in the text/figures where appropriate, such as apo-pfFn, -Fe, Fe-in vitro, Fe-in vivo, HFn-WT, HFn-3-Fold-M, HFn-3-Fold-M-Fe, etc.
Response: Thank you very much for pointing out these issues.We have defined the meaning of the abbreviations used in the revised manuscript.They are added in Figure 1f  5. Some figures in the text were cited in reverse order, e.g.Supplementary Figure 1b and 1c, d; Fig. 4d and   4e; Supplementary Figure 12f and 12g, h.
Response: Thank you for pointing out these problems.We have checked the citations of all the figures and revised their orders accordingly.Please refer to Page 5, line 3-5 & Page 13 line 27, and Page 14 in the revised version.
6. "However, the formation of H2O2 after adsorption of the second HOO• on ferrihydrite isendothermic with an increase in energy by 0.73 eV (Fig. 5h)": Please check if 0.73 eV is correct.

Response:
We apologize for the mistake and have revised the change of reaction energies to 0. 1.The authors should express the bacterial viability in terms of colony forming units/mL instead of OD450.
Response: Thank you very much for the suggestion.Accordingly, colony-forming experiments were conducted after different strains were treated with the paraquat (Figure 7d).Discussion: More importantly, our disclosing of the natural ferritin SOD nanozyme and catalytic mechanism would provide a new platform for biomedical applications.Superoxide could not only directly damage the enzyme with Fe-S center, like the aconitase and succinate dehydrogenase, but also transform into other highly toxic reactive oxygen radicals such as hydroxyl radicals and peroxynitrite, which are harmful to most cells and tissues.So superoxide plays an important role in many diseases, including radiation damage, stroke, neurodegenerative diseases, cancer, and so on 58 .Previous research has applied the natural superoxide dismutase to treat rheumatoid arthritis 59 , osteoarthritis 60 and mitigate the radiotherapy and chemotherapy damage 61 ; however, owing to the low cellular uptake, immunogenicity, and short half-life, their clinical translation was greatly restricted 58 .With the high stability and easy modification of ferritin structures, the finding of the SOD-like activity of natural ferritin nanozyme would not only contribute to the understanding of the biominerals but also offer a new strategy for the superoxide-related disease.For example, our previous works have proved that HFn itself could target the tumor cells with high specificity and sensitivity based on their TfR-1 binding ability 62 .Besides, the overexpression of superoxide dismutase has been found to inhibit the growth or invasion of the tumor cells 63,64 .Combined with these two characteristics, it was believed the ferritin nanozyme could be used to enhance tumor therapy.Moreover, the protein cages could not only be easily modified to endow them with different targeting characteristics but also act as a good drug carrier 65 , which further broadened their application scenarios.Besides tumor therapy, the ferritin nanozyme might be also used as an antioxidant for both in vitro and in vivo applications (e.g.anti-inflammation, anti-aging).
Nevertheless, our finding here just preliminarily resolved the catalytic mechanism of the ferritin nanozyme, much work was needed to further optimize their structures and activities before their application.
Overall, the manuscript is well written and should be published in Nature Communications after major revision.
Response: Thank you very much for the positive comments.
Reviewer #1: Remarks to the Author: The author has carefully explained the issues I have raised and has revised the article accordingly.I think this paper can be accepted with some editorial changes.
Reviewer #2: Remarks to the Author: This manuscript by Fan and co-workers shows the SOD like activity of ferritins.The authors have done a commendable job in addressing the reviewer's comments.I recommend this manuscript to be published in Nature Communications.
Reviewer #3: Remarks to the Author: I think the authors did a good job with this second version that should be accepted.Only one point still puzzles me, for the community and for their future work.Indeed, if the supplier of the SOD measurement kit does not provide the concentration in WST, this a difficulty for providing a kcat for comparison with other compounds from the literature.Most of us do not use a kit and that works perfectly while we can control the concentration in markor we put and thgius back-calculate the kcat.
Be careful in the added sections, the ref are given before the punctuation mark.Tehre are alsa dome places where liter is l and not L (capital letter).

Figure 5 .
Figure 5. DFT analysis of the reaction process of Ferrihydrite and superoxide.(g) Proposed two possible reaction pathways for superoxide with products of H 2 O 2 and O 2 (main reaction) or O2 and H* (side reaction) and catalytic centers in Ferrihydrite (side reaction) and P-doped Ferrihydrite (main reaction).(h) Reaction profiles with key intermediate structures and reaction energies (eV) for Ferrihydrite.(i) Reaction profiles with key intermediate structures and reaction energies (eV) for P-doped Ferrihydrite.Only important structural fragments were shown for clarity.

Fig 6 .
Fig 6. (g) ESR spectrum of the DMPO/ HOO• in the presence of XOD/Xan and gradient pyFn, the marked concentration referred to the iron content of ferritins.(h) The ESR intensity of DMPO/ HOO• of HFn, pfFn and pyFn groups.The line at 3488.125 G of DMPO/ HOO• spectra was used to represent the intensity.
Fig 1f was about 40%, however, in Fig S3a, the "inhibition rate of SOD" of "pfFn" was less than 20%; The data of "pyFn" in Fig S3a and Fig S3c were also inconsistent; In Fig S3e-g, the ferritins with "0 mM Fe content" corresponds to the ferritins purified from the basal medium, but their "inhibition rate of SOD" are not consistent with the results in Fig S3a; Response: We appreciate the insightful comments.These variations likely resulted from different preparations of the proteins for each of the experiments.To address these discrepancies, we have performed the above SOD activity tests again with indicated concentrations of ferritins (Revised Fig 1f, Revised Supplementary Fig 3c

Figure only for Reviewer 1 .
Figure only for Reviewer 1.The ferrous generation curve of different phosphate-doped ferrihydrite in the iron concentration at 44.8 μg/mL.All data are expressed as means ± S.D.
abbreviations used in the revised manuscript.They are added in Figure1flegend, Figure 2e & f legend, Page 10, line 19-20, and Supplementary Figure 7b legend, respectively.
7 eV, please refer to Page 16, line 26.Reviewer #2 (Remarks to the Author): Fan and coworkers have described mechanism behind antioxidant role of ferritins with high SOD like activity.They established a relationship between the strength of catalytic activity and the source of species.The results are incremental to the previous studies in the field, and most of the results in the manuscript are in agreement with what authors have claimed.I have listed some of my concerns below-:

Figure 7d .
Figure 7d.The colony formation of the different ferritin transgenic E. coli with or without the paraquat stimulation.All data are means ± S.D, One-way ANOVA with Tukey's test, *p < 0.05, n=3.

Figure only for Reviewer 2 .
Figure only for Reviewer 2. TEM images of ferritins from different organisms.(a) Stained and unstained TEM images of biomineralized ferritins from 8 different species, (b) Stained and unstained TEM images of ferritins of 8 different species purified from the basic medium.(c) Stained TEM images of ferritins from different species.(d) Stained TEM image of three-fold pore mutated HFn.(e) Stained and unstained TEM images of human heavy chain and light chain heteromeric ferritin and human mitochondrial ferritin.The scale bar referred to 50 nm, inset scale bar referred to 20 nm.

3 .
Authors need to demonstrate the potential applications of their findings in order to make this study suitable for broad readership of Nature Communications.One application may be enhanced tumor therapy.Response: Thank you very much for the instrumental suggestion.While this manuscript focuses on the characterization and mechanisms of the ability of natural ferritin nanozymes to scavenge superoxide radicals, our data also indicate that the activity is physiologically important.The potential applications of the activity are what we are actively exploring in the laboratory at this moment.Therefore, we discussed the potential biomedical applications of our findings in the Discussion part below (please also refer to Page 23-24 in the revised manuscript).