Previously titled: Evidence for the oxidant-mediated amino acid conversion, a naturally occurring protein engineering process, in human cells

Reactive oxygen species (ROS) play an important role in the development of various pathological conditions as well as aging. ROS oxidize DNA, proteins, lipids, and small molecules. Carbonylation is one mode of protein oxidation that occurs in response to the ironcatalyzed, hydrogen peroxide-dependent oxidation of amino acid side chains. Although carbonylated proteins are generally believed to be eliminated through degradation, we previously discovered the protein de-carbonylation mechanism, in which the formed carbonyl groups are chemically eliminated without proteins being degraded. Major amino acid residues that are susceptible to carbonylation include proline and arginine, both of which are oxidized to become glutamyl semialdehyde, which contains a carbonyl group. The further oxidation of glutamyl semialdehyde produces glutamic acid. Thus, we hypothesize that through the ROS-mediated formation of glutamyl semialdehyde, the proline, arginine, and glutamic acid residues within the protein structure can be converted to each other. Mass spectrometry provided results supporting that proline 45 (a wellconserved residue within the catalytic sequence) of the peroxiredoxin 6 molecule may be converted into glutamic acid in cultured human cells, opening up a revolutionizing concept that biological oxidation elicits the naturally occurring protein engineering process.


Introduction
Reactive oxygen species (ROS) are produced through the electron reduction of molecular oxygen and include superoxide anion radicals, hydrogen peroxide (H 2 O 2 ), and hydroxyl radicals (Freeman & Crapo, 1982;Halliwell & Gutteridge, 2007). ROS have been implicated in the pathogenesis of various diseases (Freeman & Crapo, 1982;Halliwell & Gutteridge, 2007), as well as in the aging process (Harman, 1956;Reeg & Grune, 2015). One electron reduction of molecular oxygen produces superoxide, which in turn reacts with each other to produce H 2 O 2 and reduces cellular iron ions. Reduced iron donates an electron to H 2 O 2 and produces highly reactive hydroxyl radicals. Hydroxyl radicals in turn react with virtually all biological molecules, including DNA, proteins, lipids and small molecules, damaging the biological system (Davies, 2016;Freeman & Crapo, 1982;Halliwell & Gutteridge, 2007).
One important event that occurs in response to the metal (iron)-catalyzed oxidation process is the formation of carbonyls in the protein structure. Protein carbonylation has been shown to be increased in various diseases and in aging (Berlett & Stadtman, 1997;Levine & Stadtman, 2001;Levine, 2002;Stadtman et al., 1988). Protein carbonylation occurs in response to the ironcatalyzed, H 2 O 2 -dependent oxidation of amino acid side chains (Stadtman, 1990;Suzuki et al., 2010). Protein carbonylation inactivates protein functions and marks damaged proteins for degradation (Grune et al., 1997;Levine, 1989). While carbonylated proteins are believed not to undergo electron reduction, we previously discovered the protein de-carbonylation mechanism, in which carbonyl groups can be eliminated without proteins being degraded (Wong et al., 2008). While a number of different amino acids can undergo carbonylation, major amino acid residues that are susceptible to iron-catalyzed oxidation include proline and arginine, both of which are oxidized to become glutamyl semialdehyde, which contains a carbonyl group (Amici et al., 1989). Glutamyl semialdehyde is further oxidized into glutamic acid ( Figure 1).
We previously demonstrated the role of protein carbonylation in ligand/receptor-mediated cell signaling (Wong et al., 2008). We further noted that the kinetics of ligand-mediated protein carbonylation is transient. Typically, in cultured cells, ligands activate the carbonylation of various proteins within 10 min and the activated protein carbonylation reverts to baseline by 30 min. These results suggest that there is a mechanism for the elimination of the formed carbonyls. We named this process "de-carbonylation" (Wong et al., 2008). To understand the mechanism of de-carbonylation, we tested the hypothesis that protein carbonyls may be reduced. We found that the addition of reductants to rat heart homogenates resulted in a decrease in the protein carbonyl content (Wong et al., 2013). By contrast, reductants had no effect on the carbonyl content in purified proteins, suggesting that protein carbonyls are not reduced in the absence of other cellular components. From these results, we hypothesized that cells contain catalysts for the reduction of protein carbonyls. This hypothesis is supported by our results demonstrating that the heating of heart homogenates to inactivate cellular enzymes inhibits the decrease in protein carbonyls in vitro, and that knocking down glutaredoxin 1 in the cells inhibits protein de-carbonylation (Wong et al., 2013). We used two-dimensional gel electrophoresis and mass spectrometry to identify proteins that can be de-carbonylated and found that peroxiredoxin 6 (Prx6) is one such protein (Wong et al., 2013).
Since both arginine and proline residues can be oxidized to form glutamyl semialdehyde that can further be oxidized to form glutamic acid, we speculated that arginine, proline, and glutamic acid residues may be converted to each other in the biological system, in a process that resembles site-directed mutagenesis. This article reports experimental results that support that the proline residue 45 of the human Prx6 protein molecule can be converted into glutamic acid in cells, suggesting the possible existence of a naturally occurring site-directed mutagenesis/ protein engineering-like process that may be regulated by ROS.
Peptide sample preparation Immunoprecipitation samples were treated with dithiothreitol for reduction, then with iodoacetamide for alkylation, and further digested with trypsin (12.5 ng/µl) followed by a C18 Zip-tip clean-up (EMD Millipore, Billerica, MA, USA). Tryptic peptide samples were reconstituted in 20 µl of 0.1% formic acid before nanospray liquid chromatography/mass spectrometry/ mass spectrometry (LC/MS/MS) analysis was performed.

Nanospray LC/MS/MS analysis
The tryptic peptides mixture from each sample was analyzed using a Thermo Scientific Q-Exactive Hybrid Quadrupole-Orbitrap Mass Spectrometer (Thermo Electron, Bremen, Germany) equipped with a Thermo Dionex UltiMate 3000 RSLCnano System (Thermo Dionex, Sunnyvale, CA, USA). Tryptic peptide samples were loaded onto a peptide trap cartridge at a flow rate of 5 µl/min. The trapped peptides were eluted onto a reversed-phase 20-cm C18 PicoFrit column (New Objective, Woburn, MA, USA) using a linear gradient of acetonitrile (3-36%) in 0.1% formic acid. The elution duration was 60 min at a flow rate of 0.3 µl/min. Eluted peptides from the PicoFrit column were ionized and sprayed into the mass spectrometer using a Nanospray Flex Ion Source ES071 (Thermo Scientific, Waltham, MA, USA) under the following settings: spray voltage 1.6 kV and capillary temperature 250°C. The Q Exactive instrument was operated in the data-dependent mode to automatically switch between full scan MS and MS/MS acquisition. Survey full scan MS spectra (m/z 300−2,000) were acquired in the Orbitrap with 70,000 resolution (m/z 200) after the accumulation of ions to a 3 × 10 6 target value based on predictive AGC from the previous full scan. Dynamic exclusion was set to 20 s. The 15 most intense multiply charged ions (z ≥ 2) were sequentially isolated and fragmented in the Axial Higher Energy Collision-induced Dissociation (HCD) cell using normalized HCD collision energy at 25% with an AGC target of 1e5 and a maximum injection time of 100 ms at 17,500 resolution. Two independent MS analyses in triplicate (a total of six cell samples) were performed.

LC/MS/MS data analysis
The raw MS files were analyzed using the Thermo Proteome Discoverer 1.4.1 platform (Thermo Scientific, Bremen, Germany) for peptide identification and protein assembly. The raw data files were searched against the human protein sequence database obtained from the NCBI website (https://www.ncbi.nlm.nih. gov) using the Proteome Discoverer software based on the SEQUEST algorithm. The carbamidomethylation of cysteines was set as a fixed modification, and Oxidation and Deamidation Q/N-deamidated (+0.98402 Da), and Pro>Glu (+31.990 Da) were set as dynamic modifications. The minimum peptide length was specified to be five amino acids. The precursor mass tolerance was set to 15 ppm, whereas fragment mass tolerance was set to 0.05 Da. The maximum false peptide discovery rate was specified as 0.01.

Results
Results supporting the conversion of proline residues into glutamic acid in Prx6 To identify protein carbonylation sites, we enriched Prx6 by immunoprecipitation from cultured human cells. The Prx6 immunoprecipitation samples were processed for digestion by trypsin and the tryptic peptides were analyzed by nanoLC-MS/MS analysis and protein sequence alignment to identify proline sites conversion into glutamic acid in Prx6. The conversion was identified based on a mass shift of + 31.990 Da at the proline residue (Figures 2A and B). The experiments led to the identification of one specific site at Pro 45 in human Prx6 protein ( Figure 2C). Extracted ion chromatograms of Prx6 peptide (DFT P +31.990 VCTTELGR, +2 charge, m/z=714.33) (top) and its non-conversion counterpart (DFTPVCTTELGR, +2 charge, m/z=698.33) (bottom). Both peptides were eluted at the same retention time and are from affinity-enriched cultured human cell extract using the anti-Prx6 antibody. (B) High resolution MS spectra of the co-elution of peptides (DFTP +31.990 VCTTELGR, +2 charge, m/z=714.33) (right) and its non-conversion counterpart (DFTPVCTTELGR, +2 charge, m/z=698.33) (left). (C) Illustration of the identified proline 45 conversion into glutamic acid in cultured human cells (shown in bold red). Sequence areas containing amino acid residues shown in green are detected by LC-MS/MS analysis after trypsin digestion.

Confirmation of Prx6 peptides containing proline 45 modification by MS/MS
The identified mass shift of + 31.990 Da can be caused by the conversion of proline into glutamic acid or dihydroxylated proline. Since the conversion of proline to glutamic acid or to dihydroxylated proline in Prx6 is a novel post-translational modification identified so far, it is desirable to confirm the structure of the identified peptides to ensure that the derived mass shifts of +31.99 Da are caused by the modification of proline 45. MS/MS and HPLC co-elution are gold standards for verifying peptide identification. As demonstrated in Figure 3, both peptides, DFTP+31.990VCTTELGR, +2 charge, m/z=714.33, and its non-conversion counterpart DFTPVCTTELGR, +2 charge, m/z=698.33 were co-eluted with a peak shift of less than 0.2 min. Our result showed that the high resolution MS/MS fragmentation patterns of DFTP+31.990VCTTELGR and its non-conversion counterpart DFTPVCTTELGR peptide were almost identical except the addition of +31.990 Da of fragments that contain the proline 45 residue ( Figures 3A and B).
Analysis of the ion intensity of MS spectra of DFTP+31.990VCTTELGR and its non-conversion counterpart DFTPVCTTELGR peptide ( Figure 3C) determined that the mass shift of + 31.990 Da on proline 45 occurs in 5-10% of the Prx6 molecule in our samples with a mean of 7.4 ± 1.8% (N=6). This conversion is formed post-translationally, but not due to DNA mutation, as it was promoted by treating cells with hydrogen peroxide (H 2 O 2 ) for 10 min (Figure 4). Similar results were obtained in both growing and serum-starved cells.
The consequence of proline 45 to glutamic acid conversion in the Prx6 protein molecule We found that human muscle cells treated with H 2 O 2 exhibit oxidation of one of the two cysteine residues in human Prx6 using Dojindo SulfoBiotics Protein Redox State Monitoring Kit. In this system, 15 kDa SHifter labels free sulfhydryls and Western blotting allows for the detection of sulfhydryl oxidation in protein molecules of interest. In human Prx6 protein, one cysteine is the conserved catalytic cysteine 47 that is involved in the donation of an electron during the peroxidase activity. The other is a non-conserved cysteine, which does not occur in other species such as the rat. Untreated human cells exhibit mostly the 55 kDa band ( Figure 5A). This depicts that both of the cysteine residues are reduced in the cell and were labeled with the SHifter, resulting in a 30 kDa shift of the 25 kDa Prx6 protein. The treatment of cells with H 2 O 2 caused the shift of this 55 kDa band to 40 kDa, suggesting that one of the two cysteines got oxidized by H 2 O 2 . To determine which cysteine may be the target of H 2 O 2 -mediated oxidation in human Prx6, rat cells in which Prx6 has only the catalytic cysteine were treated with  . Protein-SHifter that covalently binds to reduced protein thiols was added and samples were subjected to electrophoresis through a polyacrylamide gel. Each Protein SHifter causes ~15 kDa shift of the protein bands. After electrophoresis, the gel was exposed to UV irradiation to excise the Protein-SHifter Plus moiety, and then subjected to electrotransfer and Western blotting with the Prx6 antibody. In untreated cells, a 55 kDa band was primarily observed, indicating that two reduced cysteine residues in human Prx6 interacted with the SHifter. H 2 O 2 treatment converted this to a 40 kDa band that is consistent with the Prx6 molecule with one cysteine oxidized. Human Prx6 contains two cysteine residues, one being the conserved catalytic cysteine 47 essential for its peroxidase activity and the other cysteine that is not conserved. (B) Cultured rat cells were treated with H 2 O 2 and the same experiments were performed. Rat Prx6 contains only one cysteine that is the catalytic cysteine 47. Untreated rat cell lysates exhibited the 40 kDa band that is consistent with cysteine 47 being reduced. The treatment with H 2 O 2 caused the disappearance of the 40 kDa band, suggesting that cysteine 47 is susceptible to oxidation. (C) Cells were infected with adenovirus expressing wild type (WT) Prx6 or Prx6 mutant with proline 45 replaced with glutamic acid (P45E) Figure 5B shows that untreated rat cells exhibited the 40 kDa band, suggesting that the catalytic cysteine (cysteine 47) is reduced and labeled with the SHifter. H 2 O 2 caused the shift of this 40 kDa band to 25 kDa, indicating that cysteine 47 is the target of oxidation.
To provide information on the effects of the proline 45 to glutamic acid conversion, we constructed a human Prx6 mutant, in which proline 45 was mutated to glutamic acid, and expressed in human cells by adenovirus-mediated gene transfer. Cells expressing wild-type Prx6 and the Prx6 proline 45 to glutamic acid mutant were then treated with H 2 O 2 . Similar to the results in Figure 5A, the treatment of cells expressing wild-type human Prx6 caused the shift of the 55 kDa band to 40 kDa, indicating the oxidation of cysteine 47 by H 2 O 2 ( Figure 5C). By contrast, this oxidation did not occur in cells expressing Pro45Glu mutant ( Figure 5C). These results indicate that the conversion of proline 45 to glutamic acid results in the inhibition of cysteine 47 oxidation by H 2 O 2 .

Discussion
The present study introduces a revolutionizing concept that a protein engineering-like process could occur naturally in the biological system. Specifically, we provided data that may suggest that proline 45 of the Prx6 protein can be converted into glutamic acid. Proline 45 is in the peroxidase catalytic domain (Fisher, 2011;Fisher, 2017), thus this conversion should have functional significance. Our data suggest that the modification of proline 45 indeed seems to decrease the catalytic activity of Prx6. Thus, proteins with altered amino acid sequences through oxidant-mediated conversions may confer the diversity of the functional roles of proteins in the biological system.
The results from the present study also open up a new mechanism of ROS, indicating that the protein amino acid conversion, specifically the proline-glutamic acid conversion, may be a consequence of oxidative stress mediated by the formation of glutamyl semialdehyde in the process of protein carbonylation. Through glutamyl semialdehyde, other conversions among arginine, proline, and glutamic acid are possible. Since the caged and site-directed production of hydroxyl radicals and carbonyl formation can occur via metal binding to specific sites of the protein structure (Stadtman & Berlett, 1991;Wong et al., 2010), ROS-mediated protein amino acid conversion may be regulated through this mechanism. The conversion of free proline to free glutamic acid through the formation of glutamyl semialdehyde is known to occur (Johnson & Strecker, 1962), and enzymatic mechanisms of oxidation of free glutamyl semialdehyde to free glutamic acid have been identified (Cappelletti et al., 2018). The determination of whether such mechanisms of the conversions of free amino acids also regulate protein amino acid conversions needs further investigations.
If protein amino acid conversions occur in the biological system, this would define that the DNA sequences are not the sole determinant of primary protein structures, opening up a new concept of biology.
The limitation of this study, however, is that, while the present study obtained results that are consistent with our hypothesis of oxidant-mediated protein amino acid converion, further work is needed to prove this concept.
Author contributions YJS conceived the study and designed the experiments. JH and YJS carried out the research. JH and YJS prepared the first draft of the manuscript. Both authors were involved in the revision of the draft manuscript and have agreed on the final content.

Grant information
This work was supported by the National Institute on Aging and National Heart, Lung, and Blood Institute (NIH; grants R03 AG047824 and R01 HL72844, respectively) to YJS. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

General overview:
This manuscript presents an interesting aspect of the effect of ROS on proteins -the possibility of converting one type of amino acid into another one. It briefly describes the idea and presents results of a single mass spectrometry experiment that identifies two forms of a peptide obtained by trypsin digestion of Prx6 protein. One form contains proline residue and the other form contains modified form of the proline residue. The modification mass is +31.990 Da and based on that the authors conclude that oxidative stress can lead to carbonylation of proline and its further conversion to glutamic acid. No evidence is provided to proof the link between protein oxidation, ROS and the conversion of proline to glutamic acid.
Detailed comments: Abstract. Modified proteins are degraded by multiple enzymatic mechanisms not just the proteasome. Well established roles for lysosomes, LON protease and other proteases have been demonstrated. 1.
The authors suggest that the Pro and Arg conversion to Glu is "interchangeable". This statement suggest reversibility of the process, which is clearly not the case -it is a one-way reaction. 2.
The conversion of one amino acid into another via oxidative reactions is definitely not a "revolutionizing concept". It is very well established that His is converted to Asn and Asp, that Trp (and also other amino acids) can be converted to Gly (via side-chain elimination reactions), that Cys can be converted to Ala.

3.
Introduction section is written with a focus on authors own work and its relevance to this work is actually not 100% clear. At the same time a lot of information is missing: Have similar processes been observed before (in vitro and/or in vivo)? Can other amino acids undergo similar conversion processes? What is the state-of-art in this field?

4.
Carbonylation is not limited to Pro and Arg ! Unfortunately this is not clear from the abstract or the introduction.

5.
Methods are described in a very short form and a number of details are missing e.g. conditions for immunoprecipitation or protein digestion, amount of starting material and material used for LC-MS analysis. RAW files and processed files should be submitted to MS data repository like for example PRIDE archive.

6.
It is not clear at all why the cells where starved prior the experiment and how this is linked to oxidative stress and protein carbonylation. No comparison to non-starved cells had been made.

7.
It is unclear from the text that the Cys in the DFTPVCTTELGR peptide is modified. Although not stated in the methods section it seems that the samples have been reduced and alkylated because the Proteome Discoverer search parameters included carbamidomethylation of cysteines as a fixed modification.

8.
+31.990 is the modification mass of proline residue. Are there any other types of post 9.
translational modifications that would result in a similar mass change?
Are there any other modifications present in Prx6? 10.
The MS/MS spectra in figure 3 is the only evidence of the Pro to Glu conversion. Is this the only spectra that have been observed? According to methods section six samples have been analysed by LC-MS. How many times this peptide was fragmented in each sample?
Presentation of multiple spectra would increase the credibility of the observation. Additionally the quality of the figure is not very high and it is very difficult to read the masses of the ions present in the spectra. Therefore again submitting the results to a MS data repository would help to validate the quality of the obtained results.

11.
The basis of this selective oxidation is not addressed. 12.
It is unclear why the authors do not test the functional significance of this modification, if they have already purified the material -it is not a very difficult assay.

13.
Additional experiments where cells are collected at different conditions involving oxidative stress would help to provide the link between the carbonylation and conversion of Pro to Glu. For the moment is in not clear if this conversion is driven by oxidative stress or another unknown process. 14.
The authors should not quote amino acid conversion levels based on ion intensities to 2 decimal places (7.43 +/-1.78%). Unlikely to be this accurate.

15.
The comments in the Discussion about "altered amino acids having functional roles" betrays a lack of knowledge of the protein oxidation field -this is very well established (e.g. all the work on oxidised Cys residues).

16.
Many of the references cited are rather old. The most recent publications is authors own work. The field has moved on since many of these works were published.

17.
The authors final statement "ROS-mediated amino acid conversion may be a tightly regulated process" should be tempered (or completely omitted). "ROS" is a very generic term, and the vast majority of oxidants do not show marked residue and site specificity. They are not "tightly regulated" in the vast majority of cases.

18.
No discussion on the limitations of the presented results and conclusions is given. 19.

If applicable, is the statistical analysis and its interpretation appropriate? Not applicable
Are all the source data underlying the results available to ensure full reproducibility? Partly

Are the conclusions drawn adequately supported by the results? No
Competing Interests: No competing interests were disclosed.
We confirm that we have read this submission and believe that we have an appropriate level of expertise to state that we do not consider it to be of an acceptable scientific standard, for reasons outlined above.

Reviewer 1
The paper submitted by Suzuki and Hao highlights the importance of modifications occurring in proteins as a consequence of oxidative stress. In this particular case, the authors provide data showing that proline residue at position 45 in peroxiredoxin 6 can be converted into glutamic acid residues through glutamyl semialdehyde. The results shown in the paper are well designed, solved and clear. From technical point of view, the approach is precise and gives the information necessary to draw the conclusions. Nevertheless there are some minor details that this reviewer consider that should be added or corrected in the text: 1. The conversion of proline to glutamic semialdehyde is already known. Since the results show that the carbonyl group is further oxidized to glutamic acid, the authors should add a sentence about how they believe this last step of oxidation occurs.
[RESPONSE: The conversion of free proline to free glutamic acid through the formation of free glutamic semialdehyde has been reported to occur. While it is not yet known whether the protein proline residue conversion to protein glutamic acid residue through the formation of glutamic semialdehyde within the protein molecule utilize the same mechanism, we have included a discussion of these published papers on free amino acids in [RESPONSE: The reviewer makes a very important point. We did check, but in these particular samples from the present study, we only detected some proline 45 to be a structure that is consistent with glutamic acid, but not glutamic semialdehyde. Further work is needed to define the nature of protein carbonylation processes in the biological system.] 3. There is an exciting idea concerning the concept of "naturally occurring protein engineering". Being this true, do the authors believe that this could be a motor for evolution? Could they add a short comment on that?
[RESPONSE: While further work is needed to prove the occurrence of protein amino acid conversion in the biological system, the present study provided data that is consistent with this concept. If it were true that posttranslational modification mechanisms can convert one type of amino acid to the other, this would imply that the DNA sequences are not the sole determinant of protein sequences. To ensure that our observations of the presumed occurrence of proline-glutamic acid conversion is not due to mutation of DNA, we treated cells with hydrogen peroxide for 10 min. This short treatment, during which gene transcription and translation processes should not be completed, caused a robust modification of proline 45, confirming that this event is post-translationally regulated. This new data has been included in Fig. 4

of the new version.]
4. Finally, I disagree with the use of "interchangeable" in the text. This would induce to think that a protein could have a P or a E or a R in a given position without compromising its function. It is hard to believe that changing an E for an R would result in a neutral consequence. Since the consequences of such change (increase or decrease activity, stability,…) has not been proved in the case of Prx6, it seems reasonable that the term should be removed and simply say that this could be a driving force for evolution, for instance (as suggested above).
[RESPONSE: We do mean that P, E and R can be in a given position within the protein structure. We, however, do not imply that this would not alter the function. Our theory is that such alterations would make the protein with altered functions, contributing to the diverse nature of the biological mechanisms. In fact, we performed experiments to mutate Pro45 to Glu and found that the redox interactions between peroxiredoxin 6 and hydrogen peroxide was altered. We have included this new data in Fig. 5 of the new version. We have also deleted the term "interchangeable".] In the "Discussion" I would suggest the authors to change the sentence starting with ".Such studies would open up…". I think they can really say "…studies will open up…" and change "…have functional roles…" for "can acquire new functional roles…". Do the authors agree?
[RESPONSE: This has been modified in the new version.] Introduction section is written with a focus on authors own work and its relevance to this work is actually not 100% clear. At the same time a lot of information is missing: Have similar processes been observed before (in vitro and/or in vivo)? Can other amino acids undergo similar conversion processes? What is the state-of-art in this field?
[RESPONSE: The concept of oxidant-mediated protein amino acid conversion was generated while we were studying the role of protein carbonylation in the mechanism of redox signaling. We discovered that formed protein carbonyls can be decarbonylated in the biological system, but not in purified proteins. These studies generated the concept that, in the biological system, the formation of glutamic semialdehyde from protein proline residues is reversible. This generated the idea that protein arginine residues can become protein proline residues. We initially tested this hypothesis, however, we have not yet come across this event. Instead, as described in this paper, we generated data that are consistent with the occurrence of proline-to-glutamic acid conversion that is also a process of oxidantmediated protein amino acid conversion.] Carbonylation is not limited to Pro and Arg ! Unfortunately this is not clear from the abstract or the introduction.
[RESPONSE: In the new version, we have modified the text to make this clearer in the Introduction section.] Methods are described in a very short form and a number of details are missing e.g. conditions for immunoprecipitation or protein digestion, amount of starting material and material used for LC-MS analysis. RAW files and processed files should be submitted to MS data repository like for example PRIDE archive.
[RESPONSE: In the new version, we have expanded the Methods section. The raw MS files are publicly available as described in Data availability section.] It is not clear at all why the cells where starved prior the experiment and how this is linked to oxidative stress and protein carbonylation. No comparison to non-starved cells had been made.
[RESPONSE: We detected this modification in both starved cells and in non-starved cells. In the new version, we have made this statement.] It is unclear from the text that the Cys in the DFTPVCTTELGR peptide is modified. Although not stated in the methods section it seems that the samples have been reduced and alkylated because the Proteome Discoverer search parameters included carbamidomethylation of cysteines as a fixed modification.
[RESPONSE: The reviewer is correct that the samples are reduced for MS analysis, thus we cannot analyze cysteine redox status using this approach. In the new version, we have added this information in the Methods section.] +31.990 is the modification mass of proline residue. Are there any other types of post translational modifications that would result in a similar mass change?
[RESPONSE: As pointed by Reviewer #2, this mass shift could also be due to the formation of dihydroxylated proline. The new version of the manuscript have been modified, so that it is clear that this work presents data that are consistent with the idea of proline-to-glutamic acid conversion, but further work is needed to prove this concept.] Are there any other modifications present in Prx6?
[RESPONSE: There are a number of other modifications. In the present work, we focused on providing evidence for protein amino acid conversions.] The MS/MS spectra in figure 3 is the only evidence of the Pro to Glu conversion. Is this the only spectra that have been observed? According to methods section six samples have been analysed by LC-MS. How many times this peptide was fragmented in each sample?
[RESPONSE: Experiments have been performed more than 6 times and results have been reproducible.] Presentation of multiple spectra would increase the credibility of the observation. Additionally the quality of the figure is not very high and it is very difficult to read the masses of the ions present in the spectra. Therefore again submitting the results to a MS data repository would help to validate the quality of the obtained results.
[RESPONSE: The raw MS files are publicly available as described in Data availability section.] The basis of this selective oxidation is not addressed. The authors should not quote amino acid conversion levels based on ion intensities to 2 decimal places (7.43 +/-1.78%). Unlikely to be this accurate.
[RESPONSE: In the new version, we have modified these.] The comments in the Discussion about "altered amino acids having functional roles" betrays a lack of knowledge of the protein oxidation field -this is very well established (e.g. all the work on oxidised Cys residues).
[RESPONSE: We are strictly talking about the oxidant-mediated protein amino acid conversion process.] Many of the references cited are rather old. The most recent publications is authors own work. The field has moved on since many of these works were published.
[RESPONSE: In the new version, we have added more recent references.] The authors final statement "ROS-mediated amino acid conversion may be a tightly regulated process" should be tempered (or completely omitted). "ROS" is a very generic term, and the vast majority of oxidants do not show marked residue and site specificity. They are not "tightly regulated" in the vast majority of cases.
[RESPONSE: The reviewer is correct that ROS in general may not confer specificity. However, in the case of the oxidant-mediated protein amino acid conversion, such specificity may possibly regulate this process. In accordance with the reviewer's comment, in the new version, we have deleted the term "tightly regulated ".] No discussion on the limitations of the presented results and conclusions is given.
[RESPONSE: In the new version, we have added a discussion on the limitation of the present study.]

If applicable, is the statistical analysis and its interpretation appropriate? Partly
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? No
Competing Interests: No competing interests were disclosed.

Reviewer 2
In their manuscript, Suzuki and Hao report the finding of a peptide in Peroxiredoxin 6 that shows a mass increment of 32 in mass spectrometry analysis. NanoLC-MSMS analysis maps this increment at the site of a proline residue (P45 in the protein). This mass increment is found to affect approximately 7% of the peroxiredoxin 6 protein present in the samples. In view of these results the authors interpret that proline has suffered an oxidative modification leading to its conversion in glutamic acid. Our impression is that the information provided is not sufficient to establish this point. The mass increment of 32 Da could also be due to dihydroxylation of proline, which is a known posttranslational modification. Please see: http://web.expasy.org/findmod/findmod_masses.html Therefore, additional experimental evidence will be required to confirm the authors' conclusion. Specifically, we would suggest several of the following approaches: -Synthetize both the peptide with proline and with glutamic acid -Analyze the two peptides by HPLC. If they separate, do the same type of analysis with the peptides from their samples -Attempt to oxidize the proline-containing peptide in vitro (or the intact protein) to monitor the changes in proline -Perform amino acid analysis to confirm the presence of glutamic acid -Employ other derivatization or detection strategies to confirm the presence of glutamic acid.
[RESPONSE: We thank the reviewer for pointing out that the observed mass shift could be due to the conversion of the proline residue to dihydroxyproline. We will modified the manuscript, so that it is clear that, while we have provided data that is consistent with the idea of the proline-to-glutamic acid conversion, the present study has not proven this as it is also possible that Proline 45 is dihydroxylated. While the reviewer's suggestion is excellent for purified proteins, our thesis is that this protein amino acid conversion is driven by biological factors, thus we need to prove this in the cell systems. We will try to find alternative avenues to prove this concept in the biological system and hope to publish such results in the future papers.] Ideally, the modification described could be explored in other cell types under different oxidative conditions.
[RESPONSE: The reviewer is correct. Indeed, we are currently studying various systems including other cell types as well as tissues for patients. We hope to publish these results in the future papers.] If the authors cannot obtain an unequivocal confirmation of the presence of glutamic acid, the title of the manuscript and the main interpretations should be changed.
[RESPONSE: In the new version, we have changed the title to "Results supporting the concept of the oxidant-mediated protein amino acid conversion, a naturally occurring protein engineering process, in human cells" and modified the text to make it clear that, while the present study obtained results that are consistent with our hypothesis, further work is needed to prove this concept.]

Joaquim Ros
Department of Basic Medical Sciences, IRB-Lleida (Biomedical Research Institute of Lleida), University of Lleida, Lleida, Spain The paper submitted by Suzuki and Hao highlights the importance of modifications occurring in proteins as a consequence of oxidative stress. In this particular case, the authors provide data showing that proline residue at position 45 in peroxiredoxin 6 can be converted into glutamic acid residues through glutamyl semialdehyde. The results shown in the paper are well designed, solved and clear. From technical point of view, the approach is precise and gives the information necessary to draw the conclusions. Nevertheless there are some minor details that this reviewer consider that should be added or corrected in the text: The conversion of proline to glutamic semialdehyde is already known. Since the results show that the carbonyl group is further oxidized to glutamic acid, the authors should add a sentence about how they believe this last step of oxidation occurs.

2.
There is an exciting idea concerning the concept of "naturally occurring protein engineering". Being this true, do the authors believe that this could be a motor for evolution? Could they add a short comment on that? 3.
Finally, I disagree with the use of "interchangeable" in the text. This would induce to think that a protein could have a P or a E or a R in a given position without compromising its function. It is hard to believe that changing an E for an R would result in a neutral consequence. Since the consequences of such change (increase or decrease activity, stability,…) has not been proved in the case of Prx6, it seems reasonable that the term should be removed and simply say that this could be a driving force for evolution, for instance (as suggested above).

4.
In the "Discussion" I would suggest the authors to change the sentence starting with ".Such studies would open up…". I think they can really say "…studies will open up…" and change "…have functional roles…" for "can acquire new functional roles…". Do the authors agree?

5.
Is the work clearly and accurately presented and does it cite the current literature? Yes

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results?

Yes
[RESPONSE: While further work is needed to prove the occurrence of protein amino acid conversion in the biological system, the present study provided data that is consistent with this concept. If it were true that posttranslational modification mechanisms can convert one type of amino acid to the other, this would imply that the DNA sequences are not the sole determinant of protein sequences. To ensure that our observations of the presumed occurrence of proline-glutamic acid conversion is not due to mutation of DNA, we treated cells with hydrogen peroxide for 10 min. This short treatment, during which gene transcription and translation processes should not be completed, caused a robust modification of proline 45, confirming that this event is post-translationally regulated. This new data has been included in Fig. 4 of the new version.] 4. Finally, I disagree with the use of "interchangeable" in the text. This would induce to think that a protein could have a P or a E or a R in a given position without compromising its function. It is hard to believe that changing an E for an R would result in a neutral consequence. Since the consequences of such change (increase or decrease activity, stability,…) has not been proved in the case of Prx6, it seems reasonable that the term should be removed and simply say that this could be a driving force for evolution, for instance (as suggested above).
[RESPONSE: We do mean that P, E and R can be in a given position within the protein structure. We, however, do not imply that this would not alter the function. Our theory is that such alterations would make the protein with altered functions, contributing to the diverse nature of the biological mechanisms. In fact, we performed experiments to mutate Pro45 to Glu and found that the redox interactions between peroxiredoxin 6 and hydrogen peroxide was altered. We have included this new data in Fig. 5 of the new version. We have also deleted the term "interchangeable". ] In the "Discussion" I would suggest the authors to change the sentence starting with ".Such studies would open up…". I think they can really say "…studies will open up…" and change "…have functional roles…" for "can acquire new functional roles…". Do the authors agree?
[RESPONSE: This has been modified in the new version.]